Purine/purine isoster based scaffolds as new derivatives of benzamide class of HDAC inhibitors

Article abstract of DOI:10.1016/j.ejmech.2020.112291

This study reports the design, synthesis and evaluation of a series of histone deacetylase (HDAC) inhibitors containing purine/purine isoster as a capping group and an N-(2-aminophenyl)-benzamide unit. In vitro cytotoxicity studies reveal that benzamide 14 suppressed the growth of triple-negative breast cancer cells MDA-MB-231 (IC₅₀ = 1.48 μM), MDA-MB-468 (IC₅₀ = 0.65 μM), and liver cancer cells HepG2 (IC₅₀ = 2.44 μM), better than MS-275 (5) and Chidamide (6). Compared to the well-known HDAC inhibitor SAHA, 14 showed a higher toxicity (IC₅₀ = 0.33 μM) in three leukemic cell lines, K-562, KG-1 and THP-1. Moreover, 14 was found to be equally virulent in the HDAC-sensitive and -resistant gastric cell lines, YCC11 and YCC3/7, respectively, indicating the potential of 14 to overcome HDACi resistance. Furthermore, substantial inhibitory effects more pronounced than MS-275 (5) and Chidamide (6) were displayed by 14 towards HDAC1, 2 and 3 isoforms with IC₅₀ values of 0.108, 0.585 and 0.563 μM respectively. Compound 14 also exhibited a potent antitumor efficacy in human MDA-MB-231 breast cancer xenograft mouse model, providing a potential lead for the development of anticancer agents.

Full text of DOI:10.1016/j.ejmech.2020.112291

Journal Pre-proof
Purine/purine isoster based scaffolds as new derivatives of benzamide class of HDAC
inhibitors
Kunal Nepali, Ting-Yu Chang, Mei-Jung Lai, Kai-Cheng Hsu, Yun Yen, Tony Eight
Lin, Sung-Bau Lee, Jing-Ping Liou
PII:
S0223-5234(20)30260-9
DOI:
https://doi.org/10.1016/j.ejmech.2020.112291
EJMECH 112291
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 29 November 2019
Revised Date: 17 March 2020
Accepted Date: 27 March 2020
Please cite this article as: K. Nepali, T.-Y. Chang, M.-J. Lai, K.-C. Hsu, Y. Yen, T.E. Lin, S.-B. Lee, J.-P.
Liou, Purine/purine isoster based scaffolds as new derivatives of benzamide class of HDAC inhibitors,
European Journal of Medicinal Chemistry (2020), doi: https://doi.org/10.1016/j.ejmech.2020.112291.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©
2020 Published by Elsevier Masson SAS.

Graphical Abstract

European Journal of Medicinal Chemistry
Purine/Purine isoster based scaffolds as new derivatives of benzamide class of HDAC
inhibitors
a,1
b,c,d,1
c
b,e,d
d
Kunal Nepali , Ting-Yu Chang
, Mei-Jung Lai , Kai-Cheng Hsu , Yun Yen , Tony Eight
d
*,b,c
*,a,b,e,f
Lin , Sung-Bau Lee , Jing-Ping Liou
1
These authors contributed equally to this work
*
To whom correspondence should be addressed. S. B. Lee: (Phone) 886-2-7361661 ext 6107;
(e-mail) sbl@tmu.edu.tw; J. P. Liou: (Phone) 886-2-2736-1661 ext 6130; (e-mail)
jpl@tmu.edu.tw
a
School of Pharmacy, College of Pharmacy, Taipei Medical University.
b
Ph.D. Program in Biotechnology Research and Development, College of Pharmacy, Taipei
Medical University.
c
Master Program in Clinical Pharmacogenomics and Pharmacoproteomics, College of
Pharmacy, Taipei Medical University.
d
Graduate Institute of Cancer Molecular Biology and Drug Discovery, College of Medical
Science and Technology, Taipei Medical University.
e
TMU Biomedical Commercialization Center, Taipei Medical University.
f
School of Pharmacy, National Defense Medical Center, Taipei, Taiwan

Abstract: This study reports the design, synthesis and evaluation of a series of histone
deacetylase (HDAC) inhibitors containing purine/ purine isoster as a capping group and an
N-(2-aminophenyl)-benzamide unit. In vitro cytotoxicity studies reveal that benzamide 14
suppressed the growth of triple-negative breast cancer cells MDA-MB-231 (IC50 = 1.48 µM),
MDA-MB-468 (IC50 = 0.65 µM), and liver cancer cells HepG2 (IC50 = 2.44 µM), better than
MS-275 (5) and Chidamide (6). Compared to the well-known HDAC inhibitor SAHA, 14
showed a higher toxicity (IC = 0.33 µM) in three leukemic cell lines, K-562, KG-1 and THP-1.
5
0
Moreover, 14 was found to be equally virulent in the HDAC-sensitive and -resistant gastric cell
lines, YCC11 and YCC3/7, respectively, indicating the potential of 14 to overcome HDACi
resistance. Furthermore, substantial inhibitory effects more pronounced than MS-275 (5) and
Chidamide (6) were displayed by 14 towards HDAC1, 2 and 3 isoforms with IC50 values of
0
.108, 0.585 and 0.563 μM respectively. Compound 14 also exhibited a potent antitumor
efficacy in human MDA-MB-231 breast cancer xenograft mouse model, providing a potential
lead for the development of anticancer agents.
Key words: Purines, isosters, histone deacetylase inhibitors, drug resistance.

1
.
Introduction
Epigenetic modification of histones is involved in the regulation of cell-specific expression of
genes required for development and maintenance of cellular function [1]. Histone deacetylases
HDAC) induce histone hypoacetylation by removing the acetyl groups on N-terminal lysines of
(
histone proteins, which in turn loosens the chromatin structure and ultimately facilitates the
transcriptional process [2]. HDACs as drug targets have garnered significant scientific attention
[2,3] as HDAC inhibitors induce relief of transcriptional repression in various leukemias and are
considered to be the key for epigenetic cancer therapy [4]. Recent literature clearly indicates that
there has been an explosion of research on design and synthesis of class I HDAC inhibitors as
antiproliferative agents and promising results were attained [5-9]. In this view, class I HDAC
inhibition appears to be an attractive therapeutic strategy for developing anticancer agents.
Structural fabrication of HDAC inhibitors includes a cap group (CAP), linker part and a
Zn-binding group (ZBG) and HDAC inhibitors are generally categorized in to two structural
classes on the basis of the zinc binding group: Hydroxamic acids and the aminoanilides
(
benzamides). SAHA (1) [10], PXD101 (2) [11] and LBH589 (3) [12] represent the US FDA
approved hydroxamic acid type HDAC inhibitors (Fig. 1) and Mocetinosat (4) [13], MS-275 (5)
14] and Chidamide (6) [14] (Fig. 1) are the representative examples of
N-(2-Aminophenyl)benzamide binding unit containing HDAC inhibitors. Among the
[

benzamides, only Chidamide (6) stands as the approved benzamide (CFDA approved) for the
treatment of patients with recurrent or refractory peripheral T-cell lymphoma (PTCL) [14-17]. In
preliminary studies, Mocetinostat (4) has also demonstrated significant anticancer efficacy that
has been attributed to its slow, tight binding to HDACs and its slow decomplexation [18-19]. In
addition, there were no observed signs of acquired resistance of cancer cells towards the
exposure to Mocetinostat [20]. Another N-(2-Aminophenyl)benzamide type HDAC inhibitor,
MS-275 (5) owing to its clinical promise and selective class I inhibitory potential has been the
subject of extensive structural engineering with numerous studies yielding optimistic results [21].
These exciting facts clearly ascertains the need of extensive explorations on new aminoanilide
inhibitors.
We have been actively involved in the search of small molecule HDAC inhibitors employing
diverse combinations of CAP constructs, linker and zinc binding motifs [22]. As an extension of
this ongoing drug discovery program, the present study deals with the structural optimization of
the MS-275 and Chidamide chemical architecture that involves feasible inclusion of
purine/purine isoster as the surface recognition motif and the replacement of the carbamate and
acrylamide groups as the connecting unit to the benzyl linker (Fig. 2). These attempts were
particularly influenced by the literature rationale of attaining favorable effects via placement of a
bicyclic heteroaryl ring as the CAP component in the structural template of MS-275/Chidamide

[21, 23] along with the existence of purine/purine isosters as a key structural unit in the chemical
architecture of numerous bioactive scaffolds exerting potent cell killing effects [24]. Other than
this, the formation of a toxic metabolite, 4-(aminomethyl)-N-(2-aminophenyl)benzamide which
nonspecifically inhibits HDAC proteins or other targets attributed to the presence of carbamate
functionality in the structure of MS-275 (Fig. 2) [23] led us conceive that its replacement as the
connecting unit can be a logical approach to overcome this issue. Furthermore, the CAP
constructs were fused with substituted anilines and heteroaryl amines reported to possess
anticancer effects [24]. The fusion of such antitumor pharmacophores is anticipated to induce
sensitization of the cancer cells towards the exposure of the designed synthetics. With this
background, we describe here the synthesis and preliminary antiproliferative evaluation of a new
series of aminoanilides (7-18) (Fig. 2) containing purine/purine isoster as the capping group.
2
2
. Results
.1 Chemistry
Schemes 1-4 presents the synthetic routes to the target benzamides (7-18). Benzamides 7, 11,
3 and 17 were synthesized as per the synthetic approach shown in scheme 1. Substituted
H-Purine (19, 21) and 7H-Pyrrolo[2,3-d]pyrimidine (20, 22) were subjected to potassium
1
9
carbonate mediated benzylation with methyl 4-(bromomethyl)benzoate employing a previously

reported methodology. For the purine based starting materials (19 and 21), benzylation
proceeded regioselectively towards N-9 position and the results were in resonance with the
results of the protocol [25]. The benzylated products 23-26 were hydrolyzed using lithium
hydroxide to yield the carboxylic acid (27-30). The acids (27-29) underwent EDC/HOBt
mediated amidation with benzene-1,2-diamine to get the desired benzamides (7, 11, 13) while
the acid 30 was amide coupled with (2-Amino-4-fluoro-phenyl)-carbamic acid tert-butyl ester
followed by trifloroacetic acid mediated Boc deprotection to furnish the target benzamide 17.
Scheme 2 depicts the methodology for the preparation of benzamides 8, 10, 12, 14-16. Protic
acid catalyzed condensation of substituted anilines and 4-methyl-2H-pyrazol-3-ylamine with the
starting materials (19-22) yielded intermediates 31-36. The intermediates 31-36 were then
subjected to the previously employed synthetic route as mentioned in scheme 1 i.e. benzylation,
lithium hydroxide mediated de-esterification and amidation to afford the synthesis of target
compounds (8, 10, 12, 14-16). The synthesis of the target compound 9 was accomplished as per
the synthetic protocol shown in scheme 3. The starting material 19 (2,6-dichloro-9H-purine)
underwent condensation with methyl 4-aminobenzoate catalyzed via mineral acid followed by
ester hydrolysis and EDC/HOBt mediated amidation with NH OTHP employing
2
N,N-Diisopropylethylamine as the base to yield intermediate 51. Benzylation of the intermediate
5
1 at position 9 with methyl 4-(bromomethyl)benzoate and subsequent de-esterification gave the

carboxylic acid 53. The acid 53 was treated with o-phenylene diamine to yield intermediate 54
which underwent TFA-mediated deprotection to afford the target compound 9. Scheme 4
presents the synthetic route to benzamide 18. Initially, numerous attempts (not shown in scheme
4
) were made to incorporate the 2,4,6-trichloro-3,5-dimethoxy aniline functionality via N-methyl
urea tetheration at position 4 of the 7H-pyrrolo[2,3-d]pyrimidin-2-ylamine core. However, the
condensation was not successful and an alternative strategy was employed which led to the
condensation of the desired aniline at position 2. The starting material 22 was treated with
methyl amine to give intermediate 55 which was converted to isocyanate via treatment with
triphosgene and then reacted with the desired aniline. The subsequent route to furnish target
compound 18 was similar to that employed in previous schemes.
2
2
.2. Biological evaluation
.2.1 Antiproliferative activity
The synthesized compounds (7-18) were evaluated for in vitro cytotoxicity against the
triple-negative breast cancer cells (MDA-MB-231) as this type of cancer have been found highly
malignant and lack of efficient therapeutics and HDAC2 and 3 were highly expressed in breast
tumors isolated from hormone receptor-negative patients [26]. Given HDAC1-3 were also found
to be upregulated in clinical samples of human hepatocellular carcinomas (HCCs) [27], we also

tested the effects of our compounds on the viability of HepG2. MS-275 (5) and CFDA-approved
Chidamide (6) were employed as internal standards and the results are depicted in Table 1. The
study reveals exciting insights regarding the cytotoxic effects of the synthesized aminoanilides.
In general, the cancer cell lines exhibited a similar trend of sensitivity to the exposure of the test
compounds as that to MS-275 (5) and Chidamide (6) with MDA-MB-231 cell line being the
more sensitive one. Careful observation of Table 1 clearly indicates the influence of amine group
at position 2 on the cellular activity of the benzamides as replacement of the chloro group (7, 11)
with amino group (13, 17) induced cell growth inhibitory effects against both cell lines.
Induction of cytotoxicity against MDA-MB-231 was also evidenced with the placement of
3
-chloro-4-flouro aniline functionality at position 6 in compounds bearing the purine core with
chloro substitution at position 2 (compare 7 and 8). Similarly, fusion of the purine core with
-methyl-1H-pyrazol-3-yl functionality induced cell growth inhibition towards the cell lines
4
employed (compare 7 and 10). However, the trend was not replicated with the corresponding
compounds bearing the 7H-pyrrolo[2,3-d]pyrimidine core as isosteric CAP construct (compare
1
1 and 12). Incorporation of 3-chloro-4-flouro aniline functionality (at position 6) emerged to be
most effective for the purine based benzamides with amine group at position 2 (compare
compound 13 and 14). Compound 14 exhibited the most substantial effects against the cell lines
employed with IC values of 1.48 ± 0.06 µM (MDA-MB-231) and 2.44 ± 0.18 µM (HepG2). It
5
0

was interesting to observe that the cell growth inhibitory potential of compound 14 was higher
than MS-275 (5) and Chidamide (6). In light of these effects, we also observed consistent results
from another TNBC cell line, MDA-MB-468, which showed higher sensitivity to 14 in
comparison to MS-275 and Chidamide (Fig. 3). In contrast, the breast cancer cell line positive
for estrogen and progesterone receptors MCF-7 was more resistant to these compounds and did
not show higher susceptibility to 14 (Fig. 3). It is tempting to speculate that the expression of the
receptors may contribute to the cytotoxicity of these benzamide derivatives in breast cancer cells.
Further structural optimization studies involved the replacement of 3-chloro-4-flouro aniline with
3
,5- dimethoxy aniline (compare 14 with 15) at position 6 and N-(2-aminophenyl) functionality
(zinc binding motif) with N-(2-amino-5-fluorophenyl) functionality (compare 17 with 13). The
replacement led to decrease in the cellular activity for compound 15, however the placement of
N-(2-amino-5-fluorophenyl) functionality as the zinc binding group (compound 17) retained the
activity against both the cell lines. Isosteric replacement of purines as CAP construct in
benzamide 14 yielding the counterpart 16 also demonstrated a decline in cell growth inhibitory
effects. Compound 18 with 2-(3-(2,4,6-trichloro-3,5-dimethoxyphenyl)ureido) moiety at position
2
was devoid of cell killing effects. In general, benzamides with purine scaffold as the CAP
construct caused greater inhibition of the cell growth and also favored the inclusion of
substituted anilines and amines at position 6. Overall, presence of amine group,

3
-chloro-4-flouro aniline functionality and the CAP construct remarkably influenced the
cytotoxic effects of the compounds.
2
.2.2 HDAC isoform inhibition
All the compounds were tested for enzyme inhibitory activity towards the HDAC1, 2, 6 and 8
isoforms as shown in the Table 2. MS-275 (5) was employed as a reference compound in the
study. Most of the benzamides were endowed with potent and selective inhibitory profile
towards the HDAC1 and 2 isoforms. Among them, compound 9 possessing weak cell growth
inhibitory effects (Table 1) stands as an exception displaying strikingly selective inhibitory
potential against the HDAC6 isoform (IC50 = 11.3 nM). Apart from HDAC6, other isoforms
were also significantly inhibited by compound 9 with IC values of 0.182 μM (HDAC1), 1.45
5
0
μM (HDAC2) and 0.352 μM (HDAC8). The selectivity towards the HDAC6 isoform could be
attributed to the structural orientation of 9 possessing hydroxamic acid functionality. Overall,
compound 9, 10, 13, 14, 15 and 16 displayed higher inhibition against the HDAC1 isoform in
comparison to the standard, 5. The most cytotoxic benzamide 14 inhibited the HDAC1 and 2
isoform with IC values of 0.108 μM and 0.585 μM. On the contrary, benzamide 15 which as
5
0
such was less potent in inhibiting the growth of cell lines caused the most substantial HDAC1
and 2 isoform inhibition with IC50 values of 0.0239 μM (HDAC1) and 0.179 μM (HDAC2). A
1
5-fold higher selectivity was displayed by 15 towards the HDAC1 isoform in comparison to

HDAC2. Conclusively, both the benzamides (14 and 15) were more potent than MS-275 (5)
against the HDAC1 as well as HDAC2 isoform. Another benzamide 16 with moderate inhibitory
potential against the cancer cell lines displayed substantial selective inhibition towards HDAC1
over the HDAC2 isoform (639 folds, IC = 93 nM). Collectively, the results indicate that the
5
0
isosteric replacement of purine core (benzamide 14) with 7H-pyrrolo[2,3-d]pyrimidine
furnishing benzamide 16 was found to be effective in potentiating the inhibition as well as
selectivity towards the HDAC1 isoform. Unlike the cytotoxic profile, 3,5 –dimethoxy aniline
ring (15) was found to be the most instrumental in inducing potent inhibitory effects towards the
HDAC1 and 2 isoforms.
2
.2.3 Evaluation of inhibitory effects of benzamide 14 towards all the HDAC isoforms
In view of the remarkable antiproliferative effects of benzamide 14 against the human cancer
cell lines employed along with substantial inhibition of HDAC1 and 2 isoforms, the compound
was further evaluated against all the HDAC isoforms and the results are depicted in Table 3. The
results of evaluation indicate a more pronounced inhibition of HDAC1, 2 and 3 by 14 with IC50
values of 0.108, 0.585 and 0.563 μM in comparison to MS-275. Among these isoforms,
benzamide 14 preferentially inhibited HDAC1 over HDAC2 and HDAC3 and was 5 folds more
potent than the standard against the HDAC1 isoform. The results presented in Table 3 ascertain
the potential of 14 as class I HDAC inhibitor.

2
.2.4 In vitro cytotoxicity studies against HDAC inhibitors resistant and sensitive in gastric
cancer cell lines
Cancers display variable sensitivity to HDAC inhibitors. Previous studies have identified a
series of gastric cancer cell lines exhibiting differential response to HDAC inhibitors [28].
Consistent to this report, the YCC3/7 cells showed more resistant to MS-275 (5) in comparison
to the sensitive cell line YCC11 (Table 4). The values of 50% growth inhibition concentration
(
IC ) of MS-275 in YCC3/7 and YCC11 were 12.98 ± 1.02 µM and 6.03 ± 0.44 µM,
50
respectively. Similar trend was also observed in the two cell lines treated with compounds 10, 11,
3 and 17. In contrast, compounds 8, 14, 15, 16 and Chidamide (6) showed non-differential
1
toxicities to two cell lines. Among these, 14 was the most virulent in both cells lines, with IC50
values of 4.79 ± 0.37 µM and 4.77 ± 0.29 µM in YCC3/7 and YCC11 cells, respectively. Taken
together, these results reveal 14 as a potent anti-cancer agent that can overcome HDACi
resistance.
2
.2.5 In vitro cytotoxicity studies against leukemic cell lines
To gain further insights into the application of compound 14, we compared the potency of 14,
MS-275, Chidamide and SAHA against three leukemic cell lines that have been tested in
previous studies [29-34]. We found that these compounds exhibited strong toxicities in all three
tested leukemic cell lines, with IC values below micromolar (Fig. 4), lower than that of most of
5
0

the solid tumor cell lines we used in the study (Tables 1, 4 and Fig. 3). It's worth noting that the
IC values of 14 in these cell lines were highly similar (~0.33 μM). All three leukemic cell lines
5
0
showed significantly higher sensitivities to 14 than SAHA. In contrast, MS-275 and Chidamide
did not possess higher activity than SAHA in K-562 cells. These differential sensitivities provide
important information for selection of HDAC inhibitors in leukemia treatment.
2
.2.6 Interaction analysis of compound 14 in HDAC1
To provide a proof of concept of our molecular docking protocol, we first re-docked the
co-crystal ligand into its target protein active site similar to that performed in the literature [35,
6]. The docked ligand poses showed similar conformations with the co-crystal ligands of
HDAC1 and HDAC2 (Supporting information, Fig. 1S). The structure of HDAC3 (PDB ID:
A69) was not used for validation because the co-crystal ligand is not an HDAC inhibitor.
3
4
Nevertheless, the results suggest that favorable accuracy is obtained with the docking protocol
and methodology used in this study. For clarity, analysis of the docked compound is separated
into three sections: Zinc binding group (ZBG), linker and cap group. These sections are observed
with traditional HDAC inhibitor [37]. The ZBG of compound 14 contains
N-(2-aminophenyl)formamide moiety. Chelating the zinc ion is crucial for HDAC inhibition
38]. This is achieved with the carbonyl oxygen of benzamide 14 which also forms a hydrogen
bond with residue Y303 (Fig. 5). The ZBG contains a primary amine and an amide NH. A
a
[

primary amine located on the aromatic ring forms hydrogen bonds to two residues, His140 and
His141. The amide NH forms a hydrogen bond to residue Gly149 (Fig. 5). Both the amine and
the amide NH function as hydrogen bond donors. Hydrophobic interactions with residues M30,
L139, H140, H141 and G300 sandwich the aromatic ring of the ZBG. Together, these
interactions, combined with zinc chelation, acts as an anchor for compound 14 in the HDAC1
active site. The benzyl moiety functions as the linker and spans a hydrophobic tunnel to connect
the ZBG to the cap group [39]. The aromatic ring of benzyl functionality facilitates hydrophobic
interactions with ring residues F150, H178 and F205 (Fig. 5). The cap group of compound 14
lies at the periphery of the HDAC1 active site. This group consists of
a
N6-(3-chloro-4-fluorophenyl)-9H-purine-2,6-diamine moiety. This moiety is bulky and contains
multiple ring structures. We observed one hydrogen bond between the amine moiety and residue
D99 (Fig. 5). The chlorine and fluorine halogens are positioned outside of the active site
entrance. A hydrophobic interaction is present between residue L271 and the chlorine moiety.
This allows the cap to occupy a pocket without steric clashes along the HDAC1 active site.
Together, we found that compound 14 can form favorable interactions within the HDAC1 active
site.
2
.2.7 Binding interactions of compound 14 in HDAC isozymes

Next, we sought to compare the interactions of compound 14 in HDAC isozymes. The HDAC
isozymes are typically separated into four families: Class I, IIa, IIb, and IV [38]. Compound 14
was found to have the greatest inhibitory effect on HDAC1 (Class I) (Table 2). When docked
into HDAC4 (Class IIa), the position of compound 14 is flipped with respect to its pose in
HDAC1 (Fig. 6A). This is attributed to differences with two amino acids in the active site. First,
the ring structure of residue Y303 is pointed into the active site in HDAC1, while the ring
structure of residue H976 in HDAC4 is pointed away. This residue change in HDAC4 removes a
hydrogen bond in the active site, compared to the interaction formed between compound 14 and
residue Y303 in HDAC1. Furthermore, the ring structure of residue P800 of HDAC4 reduces a
pocket within the active site. As a result, compound 14 does not have favorable binding to
HDAC4.
Compound 14 also shows poor inhibition of HDAC6 (Table 2). The carbonyl oxygen, which
functions as the chelator in HDAC1, is not in a favorable distance to chelate the HDAC6 zinc ion
(
Fig. 6B). When the docking results were superimposed, we observed that compound 14 does not
penetrate the HDAC6 active site as deep as that seen in HDAC1. This may be due to HDAC6
residue P608, which creates a shallower pocket due to its ring pointing towards the active site,
compared to residue L139 found in HDAC1. Furthermore, HDAC6 residue S568 does not form a
hydrogen bond with the cap group, which may reduce the potency of the compound (Fig. 6B).

HDAC1 and HDAC8 are both grouped as class I HDACs. However, compound 14 produces a
significantly higher inhibition of HDAC1 isoform (Table 2). Compared to HDAC4 and HDAC6,
the pose of compound 14 in HDAC8 is closer to the one observed in HDAC1 (Fig. 5C). This can
be attributed to the class similarity of HDAC1 and HDAC8. However, two differences were
observed that may account for the weaker inhibition of the compound. The cap forms a hydrogen
bond with HDAC1 residue D99. In HDAC8, the cap is rotated away from the respective residue,
D101. This may also force the compound 14 cap to occupy a different position in the periphery
of the HDAC active site. In addition, HDAC8 contains residue M274 instead of residue L271
found in HDAC1. As a result, HDAC8 lacks a hydrophobic interaction with the CAP construct
of compound 14 at this position. Together, these interactions show the preference of compound
1
4 towards HDAC1.
Overall, the docking results supports the substantial HDAC1 inhibition exerted by benzamide
4 in comparison to the other HDAC isozymes. Some key interactions with the amino acid
1
residues of HDAC1 isoform were figured out to rationalize the experimental findings such as i)
ZBG of compound 14 chelates the HDAC1 zinc ion and forms hydrogen bonds to residues H140,
H141 and G149. (Fig. 5) ii) The linker of compound 14 consists of a benzyl moiety that spans
the HDAC1 hydrophobic tunnel. iii) Cap construct of 14 forms hydrophobic interactions with
residue L271 and a hydrogen bond to residue D99 (Fig. 5). Overall, the interactions along the

rim of the active site may explain a greater inhibitory activity of compound 14 towards HDAC1
isoforms. Moreover, these results are in resonance with the results of studies revealing key
insights regarding the catalytic domains of HDAC isoforms [40-43] and suggests the importance
of the cap group in determining HDAC selectivity. Together, these interactions rationalize the
selectivity of compounds 14 towards HDAC1.
We also docked compound 14 into HDAC2 and HDAC3 to determine its binding interactions.
Overall, we observed similar binding poses of Compound 14 in HDAC1-3 (Fig. 5). Importantly,
the ZBG coordinated to the metal zinc ion and no steric hindrance of the linker group was
observed. These poses suggest Compound 14 has inhibitory activity towards HDAC1, HDAC2
and HDAC3.
Next, we performed a molecular dynamics (MD) simulation to determine important residues
for binding interactions between the ligand and HDAC1 protein structure. The final
conformation of the Compound 14-HDAC1 complex was similar to its docked pose (Fig. 7A-B).
Important interacting residues from the MD simulation were identified (Fig. 7C). The metal
coordination is key for HDAC inhibition and was observed throughout the MD simulation.
Hydrogen bonds occur with greater frequency with residues D99 and D176. Prominent
hydrophobic interactions were observed with residues M30, C151, Y204, F205, L271 and C273.
Many of these hydrophobic interactions occur with the hydrophobic tunnel and the compound 14

linker, however, hydrophobic interactions occur between residues L271 and C273 and the cap
group. A halogen interaction between the cap group and residue C273 also occurred with greater
frequency. Together, the MD simulation suggests that compound 14 binds effectively to the
HDAC1 active site.
2
.2.8 HDAC inhibitory activity in human breast cancer cell line.
Histones are wildly used substrates to determine the activity of most HDACs in vitro.
However, only HDAC1, 2 and 3, as well as Sir2-like enzymes, have been validated to
deacetylate histones in cells [44-45]. In contrast, the main targets of other HDACs in vivo are
non-histone proteins, such as α-tubulin for HDAC6 and structural maintenance of chromosomes
protein 3 (SMC3) for HDAC8 [46-47]. Compound 14 and 16 were thus further selected for the
Western blot analysis to ascertain their ability to modulate the protein levels of important
biomarkers associated with intracellular HDAC inhibition. This could help establish the effects
of subtle structural variation (purine and its isosteric scaffold) in the surface recognition part on
the HDAC inhibition. We determined the acetylation levels of different HDAC substrates in
MDA-MB-231 cells and exploited Tubastatin A and PCI-34051 as controls for suppression of
HDAC6 and HDAC8i, respectively. (Fig. 8). Similar to the effect of MS-275, we observed a
th
dose-dependent upregulation of the acetylation levels on the 9 lysine of histone H3 (ac-H3 K9)
th
and the 5/8/12/16 lysine of histone H4 (ac-H4 K5/8/12/16) in cells treated with the compound

1
4 and 16. In contrast, the level of aectylated α-tubulin and SMC3 were not affected (Fig. 8A).
These results were consist with in vitro HDAC inhibition study and molecular docking analysis,
supporting that benzamide 14 behaves as a class I HDAC inhibitor targeting to HDAC1, 2 and 3.
2
.2.9 G phase arrest and apoptosis in human breast cancer cell line.
1
The data present in this study indicate that the benzamide derivative 14 possesses strongly
inhibitory activity against HDAC1, 2 and 3 (Table 2 and 3). HDACs are known to control cell
proliferation through regulation of cell cycle progression, apoptosis and other cellular activities
[
48]. Consistent to the previous studies [32, 49, 50], our Western blot analysis showed that
HDAC inhibition by 14, 16 or MS-275 increased the expression of the cyclin-dependent kinase
CDK) inhibitor p21 (Fig. 8A) in a dose-dependent manner, perhaps through suppressing
(
HDAC1 and 2 bound on p21 promoter [51, 52]. Given that excess expression of p21 can disturb
cell cycle progression and MS-275 has been known to elicit p53-dependent cell death [53, 54],
we further determined the cell status of cells treated with these compounds to establish the
influence of amino group as well as isosteric replacement of the purine core.
To this, we treated MDA-MB-231 cells with a low dose (4 μM) of the standard compound (5)
and the benzamides (14 and 16) for 24 hours, followed by analysis of cell cycle distribution with
flow cytometry. Cells treated with individual compound exhibited a cell cycle profile similar to
control but lost the population of S phase, the stage between G to G /M phases (Fig. 9A),
1
2

suggesting that those cells are unable to initiate DNA replication. We further evaluated the
efficiency of cell cycle progression by measuring the rate of accumulation of mitotic cells in the
presence of nocodazole, an inhibitor of microtubule polymerization. Compared to the control
cells that accumulated in G2/M after nocodazole treatment, a large portion of cells treated with
tested compounds stayed in G (Fig. 9A). These results clearly demonstrate that the benzamide
1
1
4 and 16 function as HDAC1 inhibitors that are able to trigger cell cycle arrest in G1.
As expected, the higher dose (16 μM) of MS-275 induced cell death detected by accumulation
of subG population at 48 hours after treatment (Fig. 9B). This effect was more pronounced
1
when cells were treated with 14 (Fig. 9B). We thus further investigated their efficiency of
apoptosis induction by measuring the level of cleavage of caspases and PARP-1, indicatives of
apoptosis activation. Our results from Western blot analysis showed that cleaved forms of
caspase-3, -8 and PARP-1 were significantly increased in MDA-MB-231 cells treated with the
test compounds (Fig. 9C). Notably, benzamide 14 was endowed with more potent effect in subG1
accumulation and activation of apoptotic pathway, as compared to 5 and 16 (Fig. 9B-C). These
results highlight the apoptosis inducing ability of the selected benzamides with most substantial
effects displayed by benzamide 14, consistent with the growth suppression observed (Table 1).
The outcome of the assay also indicates the favorable effects attained via placement of amine
group on the purine core linked to 3-chloro-4-floro aniline functionality.

2
.2.10 Antitumor efficacy of compound 14 in human MDA-MB-231 breast cancer
xenograft model
Compound 14 was evaluated for in vivo efficacy against tumor xenografts in human
MDA-MB-231 breast cancer cells. The tumor growth curve and animal body weight change for
each treatment group are shown in Fig. 10. As shown in Fig. 10A, Taxol at 10 mg/kg
(intravenous injection, every other day) (p<0.05) and compound 14 at 50 mg/kg (intraperitoneal
injection, every day) (p<0.01) produced significant antitumor activity and suppressed the growth
of MDA-MB-231 breast cancer xenografts by tumor growth inhibition values (TGI) of 32.5%
and 56.3%, respectively. Moreover, there were no significant changes in body weight at all doses
tested (Fig. 10B). Overall, the result demonstrates that compound 14 is endowed with potent in
vivo antitumor efficacy in human MDA-MB-231 breast cancer xenograft model.
3
. Conclusion
This study dealt with the structural optimization of MS-275 (5) and Chidamide (6) to furnish
more potent antiproliferative agents. Specifically, attempts were made towards the inclusion of
purine/purine isoster as a bicyclic heteroaryl CAP construct and the replacement of carbamate
and acrylamide group as a connecting unit to the linker in the design of class I HDAC inhibitors.

The approach was further extended to link the CAP constructs with substituted anilines and
appropriate amines reported to exert anticancer effects. The results of the in-vitro assays led us
generate structure-cytotoxicity as well as structure-HDAC isoform inhibition relationship which
in turn established the impact of variations in the CAP component on the bioactivity (Fig. 11). It
was observed that the placement of substituted aniline at position 4 was favored for
antiproliferative effects as well as significant Class I HDAC inhibition. Specifically,
3
-chloro-4-fluoroaniline was found to be the preferred substitution for cell growth inhibitory
effects and for class I HDAC inhibition, 3,5-dimethoxyaniline was the substituent of choice.
Other than this, factors such as utilization of the purine core in the design of the target constructs
and the presence of amine group at position 2 were identified to be instrumental for inducing the
antiproliferative effects (Fig. 11). The structural notions generated in this study provides us
valuable information that can be further utilized for the extension of our research work centered
at the design of purine/purine isoster based HDAC inhibitor. Subsequently, the results of
biological evaluation led us identify a potent class I HDAC inhibitor 14 which was found to be
the most virulent in HDACi resistant (YCC3/7) as well as sensitive (YCC11) gastric cell lines.
These are optimistic findings as they are indicative of the capacity of benzamide 14 to overcome
HDACi resistance in gastric cancer which is the one of the most common cause of cancer-related
deaths and fabrication of treatment strategies via design of small molecule inhibitors is the need

of the hour. To add on, the in vitro cytotoxicity studies demonstrating the efficacy of
benzamide 14 against the triple negative breast cancer cell line (TNBC, MDA-MB-231 and
MDA-MB-468) is an equally important finding in view of the lack of biomarkers for their
targeted therapy. As such, TNBC demonstrates lack of sensitivity towards the existing systemic
therapies and have also been evidenced to incur a high rate of recurrence [55]. In this context, the
revelations of this study are significantly promising. Moreover, HDAC inhibitor 14 exhibited
significant inhibitory potential against K-562, THP-1, KG-1 and HL-60 cancer cell lines, exerted
a dose-dependent upregulation of ac-H3K9 in MDA-MB-231 cells and also triggered cell cycle
arrest in G1 phase. The results of the docking study highlight the importance of the cap group in
determining HDAC selectivity and identified key interactions that forms the basis of compounds
1
4’s selectivity towards HDAC1. Furthermore, benzamide 14 also demonstrated potent in vivo
antitumor efficacy in human MDA-MB-231 breast cancer xenograft model. These evidences and
the lack of clinically approved benzamides with the exception of Chidamide (6) as the only
benzamide type HDAC inhibitor clearly ascertains the need of detailed investigation of
benzamide 14. In a nut shell, compound 14 could emerge as a potential lead compound for the
development of anti-breast and gastric cancer agent. Further lead modifications on 14 such as
synthesis of compounds with diverse substitutions on the purine and its isoster core along with
utilization of other linkers is under progress.

4
4
. Experimental
.1. Chemistry
Nuclear magnetic resonance spectra were obtained with Bruker DRX-500 spectrometer
(
operating at 300 MHz), with chemical shift in parts per million (ppm, d) downfield from TMS
as an internal standard. High-resolution mass spectra (HRMS) were measured with a JEOL
JMS-700) electron impact (EI) mass spectrometer. Purity of the final compounds were
(
determined using an Hitachi 2000 series HPLC system using C-18 column (Agilent ZORBAX
Eclipse XDB-C18 5 mm. 4.6mm_ 150 mm) and were found to be > 95%. Flash column
chromatography was done using silica gel (Merck Kieselgel 60, No. 9385, 230e400 mesh
ASTM). All reactions were carried out under an atmosphere of dry nitrogen.
4
.1.1 Synthesis of methyl 4-((2,6-dichloro-9H-purin-9-yl) methyl) benzoate (23)
A mixture of 4-chloro-7H-pyrrolo[2,3-d] pyrimidin-2-amine (19) (0.5 g, 2.64 mmol), methyl
-(bromomethyl) benzoate (0.605 mg, 2.64 mmol) and potassium carbonate (0.547 mg, 3.96
4
mmol) in DMF was stirred at 60 °C for 4h. The reaction mixture was quenched with H O and
2
extracted with EtOAc (50 mL x 3). The combined organic layer was dried over anhydrous
MgSO , concentrated under reduced pressure and purified by silica gel chromatography to give
4

1
compound 23 in 73 % yield; H NMR (300 MHz, CD OD): 8.21 (s, 1H), 7.89 (d, J = 8.7 Hz,
3
2
H), 7.49 (d, J = 8.7 Hz, 2H), 5.21 (s, 2H), 3.81 (s, 3H).
Synthesis of methyl 4-((2,4-dichloro-7H-pyrrolo[2,3-d] pyrimidin-7-yl)methyl) benzoate
24)
The
(
title
compound
24
was
synthesized
from
compound
2
,4-Dichloro-7H-pyrrolo[2,3-d]pyrimidine (20) in a manner similar to that described for the
1
synthesis of compound 23; H NMR (300 MHz, CD OD): 7.89 (d, J = 8.1 Hz, 2H), 7.28 (d, J =
3
8
.1 Hz, 2H), 7.08 (d, J = 3.9 Hz, 1H), 6.31 (d, J = 3.9 Hz, 1H), 5.23 (s, 2H), 3.79 (s, 3H).
Synthesis of methyl 4-((2-amino-6-chloro-9H-purin-9-yl) methyl) benzoate (25)
The title compound 25 was synthesized from compound 6-chloro-9H-purin-2-amine (21) in a
1
manner similar to that described for the synthesis of compound 23; H NMR (300 MHz,
CD OD): 8.21 (s, 1H), 7.76 (d, J = 8.7 Hz, 2H), 7.46 (d, J = 8.7 Hz, 2H), 5.35 (s, 2H), 3.79 (s,
3
3
H).
Synthesis of methyl 4-((2-amino-4-chloro-7H-pyrrolo[2,3-d] pyrimidin-7-yl) methyl)
benzoate (26)
The title compound 26 was synthesized from 4-chloro-7H-pyrrolo[2,3-d] pyrimidin-2-amine (22)
1
in a manner similar to that described for the synthesis of compound 23; H NMR (300 MHz,

CD OD): 7.85 (d, J = 8.0 Hz, 2H), 7.21 (d, J = 8.0 Hz, 2H), 7.14 (d, J = 3.5 Hz, 1H), 6.28 (d, J =
3
3
.5 Hz, 1H), 5.24 (s, 2H), 3.81 (s, 3H).
Synthesis of 4-((2,6-dichloro-9H-purin-9-yl) methyl) benzoic acid (27)
A mixture of compound 23 (0.6 g, 1.77 mmol), 1M LiOH aq. (10 ml) and dioxane (20 mL) was
stirred at room temperature for 2 h. The reaction was concentrated under reduced pressure and
H O was added. The mixture was acidified with 3N HCl and extracted with EtOAc (50 mL x 3).
2
The combined organic layer was dried over anhydrous MgSO and concentrated under reduced
4
1
pressure to yield the acid (27) in 96% yield; H NMR (300 MHz, CD OD): 8.29 (s, 1H), 7.92 (d,
3
J = 8.1 Hz, 2H), 7.57 (d, J = 8.1 Hz, 2H), 5.34 (s, 2H).
Synthesis of 4-((2,4-dichloro-7H-pyrrolo[2,3-d] pyrimidin-7-yl) methyl) benzoic acid (28)
The title compound 28 was synthesized in 97 % yield from compound 24 in a manner similar to
1
that described for the synthesis of compound 27. H NMR (300 MHz, CD OD): 7.98 (d, J = 8.1
3
Hz, 2H), 7.29 (d, J = 8.1 Hz, 2H), 7.06 (d, J = 3.9 Hz, 1H), 6.39 (d, J = 3.9 Hz, 1H), 5.29 (s, 2H).
Synthesis of 4-((2-amino-6-chloro-9H-purin-9-yl) methyl) benzoic acid (29)
The title compound 29 was synthesized in 94 % yield from compound 25 in a manner similar to
1
that described for the synthesis of compound 27. H NMR (300 MHz, CD OD): 8.27 (s, 1H),
3
7
.89 (d, J = 8.7 Hz, 2H), 7.41 (d, J = 8.7 Hz, 2H), 5.33 (s, 2H).

Synthesis of 4-((2-amino-4-chloro-7H-pyrrolo[2,3-d] pyrimidin-7-yl) methyl) benzoic acid
30)
(
The title compound 30 was synthesized in 96 % yield from compound 26 in a manner similar to
1
that described for the synthesis of compound 27. H NMR (300 MHz, CD OD): 7.98 (d, J = 8.0
3
Hz, 2H), 7.43 (d, J = 8.0 Hz, 2H), 7.14 (bs, 1H), 6.39 (d, J = 3.5 Hz, 1H), 5.23 (s, 2H).
Synthesis of N-(2-aminophenyl)-4-((2,6-dichloro-9H-purin-9-yl) methyl) benzamide (7)
A mixture of compound 27 (0.5 g, 1.54 mmol), EDC.HCl (0.591 g, 3.09 mmol), HOBt (0.311 g,
2
.31 mmol), benzene-1,2-diamine (0.166 g, 1.54 mmol) and DIPEA (0.718 ml, 4.13 mmol) in
DMF (5 mL) was stirred at RT for 5 h. After being stirred for a further 5 h, the reaction mixture
was quenched with H O and extracted with EtOAc (50 mL x 3). The combined organic layer was
2
dried over anhydrous MgSO , concentrated under reduced pressure and purified by silica gel
4
1
chromatography (hexane: EtOAc = 1:1) to give 7 in 71 % yield; mp: 180 -182 °C; H NMR (500
MHz, CD OD): 8.15 (s, 1H), 7.95 (d, J = 8.5 Hz, 2H), 7.45 (d, J = 8.0 Hz, 2H), 7.17 (d, J = 7.0
3
1
3
Hz, 1H), 7.05 (m, 1H), 6.89 (d, J = 8.5 Hz, 1H), 6.76 (t, J = 8.5 Hz, 1H), 5.42 (s, 2H). C (125
MHz, DMSO-d ): 164.90, 159.91, 159.36, 158.76, 156.61, 154.05 149.52, 143.18, 143.04,
6
1
39.88, 134.11, 128.11, 127.05, 126.62, 123. 28, 119.86, 116.06, 109.39, 45.88. HRMS (ESI) for
+
C H Cl N O (M+H ): calcd, 413.0684; found, 413.0928; LRMS - MS (ESI): 413.05 [M+H],
19
15
2
6
4
11.05 [M-H].

Synthesis of N-(2-aminophenyl)-4-((2,4-dichloro-7H-pyrrolo[2,3-d] pyrimidin-7-yl) methyl)
benzamide (11)
The title compound 11 was synthesized in 75 % yield using compound 28 in a manner similar to
1
that described for the synthesis of compound 7; mp: 190 -191 °C; H NMR (500 MHz, CD OD):
3
7
1
5
1
.81 (d, J = 8.0 Hz, 2H), 7.22 (d, J = 8.0 Hz, 2H), 7.18 (d, J = 3.5 Hz, 1H), 7.14 (d, J = 7.5 Hz,
H), 6.91 (m, 1H), 6.71 (d, J = 8.0 Hz, 1H), 6.51 (t, J = 7.5 Hz, 1H), 6.37 (d, J = 3.5 Hz, 1H),
1
3
.31 (s, 2H). C (75 MHz, DMSO-d ): 165.50, 159.97, 154.05, 151.82, 143.57, 142.05, 141.52,
6
34.35, 133.85, 131.70, 128.37, 127.33, 126.94, 123.69, 116.71, 116.56, 109.09, 99.53, 47.34.
+
HRMS (ESI) for C H Cl N O (M+H ): calcd, 412.0932; found, 412.0963.
2
0
16
2
5
Synthesis of 4-((2-amino-6-chloro-9H-purin-9-yl) methyl)-N-(2-aminophenyl) benzamide
13)
(
The title compound 13 was synthesized in 74 % yield using compound 29 in a manner similar to
1
that described for the synthesis of compound 7; mp: 194 -195 °C; H NMR (500 MHz, CD OD):
3
8
1
1
4
.19 (s, 1H), 8.09 (d, J = 8.7 Hz, 1H), 7.88 (d, J = 8.1 Hz, 2H), 7.65 – 7.66 (m, 2H), 7.55 (m,
1
3
H), 7.46 (d, J = 8.4 Hz, 2H), 5.47 (s, 2H). C (125 MHz, CD OD): 167.99, 161.89, 158.12,
3
51.73, 144.71, 144.59, 141.63, 135.72, 129.64, 128.99, 127.99, 126.60, 119.22, 118.49, 110.65,
+
8.72. HRMS (ESI) for C H ClN O (M+H ): calcd, 394.1183.; found, 394.1183.
1
9
16
7