Discovery of Highly Selective and Potent HDAC3 Inhibitors Based on a 2-Substituted Benzamide Zinc Binding Group

Article abstract of DOI:10.1021/acsmedchemlett.0c00462

The selectivity of histone deacetylase inhibitors (HDACis) is greatly impacted by the zinc binding groups. In an effort to search for novel zinc binding groups, we applied a parallel medicinal chemistry (PMC) strategy to quickly synthesize substituted benzamide libraries. We discovered a series containing 2-substituted benzamides as the zinc binding group which afforded highly selective and potent HDAC3 inhibitors, exemplified by compound 16 with a 2-methylthiobenzamide. Compound 16 inhibited HDAC3 with an IC50 of 30 nM and with unprecedented selectivity of >300-fold over all other HDAC isoforms. Interestingly, a subtle change of the 2-methylthio to a 2-hydroxy benzamide in 20 retains HDAC3 potency but loses all selectivity over HDAC 1 and 2. This significant difference in selectivity was rationalized by X-ray crystal structures of HDACis 16 and 20 bound to HDAC2, revealing different binding modes to the catalytic zinc ion. This series of HDAC3 selective inhibitors served as tool compounds for investigating the minimal set of HDAC isoforms that must be inhibited for the HIV latency activation in a Jurkat 2C4 cell model and potentially as leads for selective HDAC3 inhibitors for other indications.

Full text of DOI:10.1021/acsmedchemlett.0c00462

pubs.acs.org/acsmedchemlett
Letter
Discovery of Highly Selective and Potent HDAC3 Inhibitors Based on
a 2‑Substituted Benzamide Zinc Binding Group
Jian Liu,* Younong Yu, Joseph Kelly, Deyou Sha, Abdul-Basit Alhassan, Wensheng Yu,
Milana M. Maletic, Joseph L. Duffy, Daniel J. Klein, M. Katharine Holloway, Steve Carroll,
Bonnie J. Howell, Richard J. O. Barnard, Scott Wolkenberg, and Joseph A. Kozlowski
Cite This: https://dx.doi.org/10.1021/acsmedchemlett.0c00462
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: The selectivity of histone deacetylase inhibitors
(HDACis) is greatly impacted by the zinc binding groups. In an
effort to search for novel zinc binding groups, we applied a parallel
medicinal chemistry (PMC) strategy to quickly synthesize
substituted benzamide libraries. We discovered a series containing
2-substituted benzamides as the zinc binding group which afforded
highly selective and potent HDAC3 inhibitors, exemplified by
compound 16 with a 2-methylthiobenzamide. Compound 16
inhibited HDAC3 with an IC50 of 30 nM and with unprecedented
selectivity of >300-fold over all other HDAC isoforms. Interest-
ingly, a subtle change of the 2-methylthio to a 2-hydroxy benzamide
in 20 retains HDAC3 potency but loses all selectivity over HDAC 1
and 2. This significant difference in selectivity was rationalized by X-ray crystal structures of HDACis 16 and 20 bound to HDAC2,
revealing different binding modes to the catalytic zinc ion. This series of HDAC3 selective inhibitors served as tool compounds for
investigating the minimal set of HDAC isoforms that must be inhibited for the HIV latency activation in a Jurkat 2C4 cell model and
potentially as leads for selective HDAC3 inhibitors for other indications.
KEYWORDS: HDAC, HDAC3, inhibitors, 2-methylthiobenzamide, induced pocket
istone deacetylases (HDACs) and histone acetyl trans-
Hferases (HATs) are two enzyme families which regulate
histone acetylation level, causing structural changes to
chromatin, hence controlling the transcription of DNA.
Histone acetylation diminishes the charge−charge interaction
between histones and DNA, and this relaxed conformation
induces gene expression.1,2 Eighteen unique isoforms of
HDACs have been discovered and are separated into four
distinct classes.3 The Zn2+-dependent isoforms are class I
(HDACs 1, 2, 3, and 8), class II (HDACs 4, 5, 6, 7, 9, and 10),
and class IV (HDAC11); the NAD+-dependent isoforms are
the class III Sirtuins (SirT1−7). HDACs have been the targets
for various therapeutics. Small molecule HDAC inhibitors
(HDACis) were reported to demonstrate capabilities to
regulate gene transcription, regulate cell cycle progression,
and induce differentiation and/or apoptosis in cancer cells with
a few HDACis being approved for the treatment for
oncology.4−7 HDACis have also been broadly explored as
HIV latency-reversing agents (LRAs) to activate the latently
infected CD4+ T cells. Clinically, several HDACi drugs
approved for the treatment of cancer have been shown to
induce HIV RNA transcription in resting CD4+ T cells from
the combination antiretroviral therapy (cART) suppressed
HIV patients.8−10 This HIV latency reversal activity by
HDACis was thought to be driven by HDAC3 inhibition,11,12
which led to the investigation of HDAC3 selective inhibitors.
In the attempt to search for HDAC isoform selective
inhibitors, we recently reported two series of HDACs 1, 2, 3
inhibitors: the ethyl ketone series exemplified by compound 1
and the aryl ketone series exemplified by compound 2 (Figure
1).13−15 These class I HDAC inhibitors showed efficacy for
HIV latency activation in a Jurkat 2C4 cell model and
induction of HIV gag p24 protein in patient latent CD4+ T
cells.14 In the process of determining the required HDAC
subtype inhibition for HIV latency activation, we demonstrated
that inhibition of only HDAC1, 2 by compound 316 did not
show any cellular activity in the HIV latency activation.13 We
invested our effort to search for highly selective inhibitors of
HDAC3. Due to the high homology between HDACs 1 and 2,
it has proven very difficult to find a selective inhibitor that
Received: August 20, 2020
Accepted: October 8, 2020
https://dx.doi.org/10.1021/acsmedchemlett.0c00462
© XXXX American Chemical Society
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX
A

ACS Medicinal Chemistry Letters
pubs.acs.org/acsmedchemlett
Letter
a
Scheme 1
a
Reagents and conditions: (i) Cs2CO3, DMF, rt, 1 h, 97%; (ii)
NH4OAc, toluene, 110 °C, 18 h, 87%; (iii) TFA/DCM, rt, 99%; (iv)
HATU, DIEA, DCM, rt, 16 h, 98%; (v) HBr/AcOH, rt, 16 h, 93%;
(vi) HATU, TEA, DMF, rt, 16 h.
to provide an excellent yield of 10. The benzylcarbamate
protection was successfully removed after 10 was stirred in a
solution of hydrobromide in acetic acid at room temperature
overnight to generate the common intermediate 11. The PMC
focus library was designed using the common intermediate 11
with an amide coupling reaction to selected monomer of
commercially available or readily prepared benzoic acids with
2-substituents which could potentially coordinate with the
HDAC enzyme zinc ion as a bidentate chelation along with the
amide carbonyl. These focus library sizes ranged from 24−100
monomers, and the optimization of the structure activity
relationship (SAR) was carried out in six rounds of libraries
with each round applying different substrates of substituted
benzoic acids, pyridinyl carboxylic acids, and 5- and 6-member
heteroaryl carboxylic acids.
All compounds from the focus library were tested in in vitro
HDAC inhibition assays against HDACs 1, 2, 3, 6, or 8, while
at the same time each compound was evaluated for Latency
Reversal Activity (LRA) in the Jurkat 2C4 T-cell HIV latency
model cell line.19 The representative compounds are listed in
Table 1. We discovered a series of highly selective HDAC3
inhibitors from the PMC libraries. Subtle variation of the
substituents led to dramatic changes in subtype selectivity. The
2-methylamino benzamide 13 is a highly HDAC3 selective
inhibitor with an IC50 of 41 nM toward HDAC3 and selectivity
against other subtypes of over 366-fold (HDAC1/HDAC3).
The HDAC3 selective inhibitor 13 showed no LRA activity in
the 2C4 Jurkat T-cell model system. When the methyl on the
aniline is absent, 14 as a 2-amino-6-fluorobenzamide became a
HDAC1, 2, 3 inhibitor with only 4-fold selectivity between
HDAC1 and HDAC3 inhibition activity. The HDAC1, 2, 3
inhibitor 14 demonstrated comparable LRA activity to that
which we observed in the previous reported class I HDAC
inhibitors. However, the 3-aminopyridinyl-2-carboxamide 15
dramatically reduced the inhibition of all HDAC enzymes
except HDAC3 with still an IC50 of 1,100 nM. The selectivity
of HDAC3 over HDAC1 for 15 is greater than 40-fold,
although it is much less active. The most selective HDAC3
inhibitor discovered from this series is 16 with the 2-
methylthiobenzamide. This compound afforded an IC50 of
29 nM toward HDAC3 and 690-fold selectivity over HDAC1.
The removal of the methyl on the sulfur in 17 significantly
reduced the HDAC3 inhibition; however, it also brought back
Figure 1. HDAC selective inhibitors with various zinc binding groups.
distinguishes between these two isoforms. However, there are
many reports of HDAC3 selective inhibitors, especially the
series of N-(2-aminophenyl)carboxamides represented by
compound 4 (RGFP966).17 A close analog 5 to our reported
ethyl ketone and aryl ketone HDAC inhibitors with a
methylcarboxamide was described to have improved HDAC3
inhibition compared to 4, with 50-fold selectivity over
HDAC1.18 In this letter, we report our work, taking advantage
of a parallel medicinal chemistry (PMC) strategy to quickly
synthesize and screen a series of HDAC inhibitors. This effort
resulted in the discovery of a series of highly potent and
selective HDAC3 inhibitors based on 2-substituted benza-
mides as the zinc binding group.
Parallel medicinal chemistry (PMC) has been widely applied
in drug discovery across the industry. Recent technology
advancements in automation for synthesis and purification,
data analysis, modeling, and in-silico prediction have driven the
discovery chemistry effort in our company to become more
data guided, knowledge based, and highly collaborative with
extensive application of PMC. A PMC library was designed
using the common intermediate 11 to carry out an amide
coupling using selected carboxylic acids which also contain a
potential second zinc binding group in addition to the amide
carbonyl. We hypothesized that a zinc binding group capable
of bidentate coordination with the HDAC zinc ion might
enhance potency; however, it was less clear what, if any, role
the zinc binding group may play in HDAC isoform selectivity.
Synthesis of the common intermediate 11 is depicted in
Scheme 1, which started by reacting a bromoketone 6 and the
protected lysine 7 to form a ketone ester intermediate in DMF
with cesium carbonate as base. The imidazole ring in 8 was
then prepared by condensation of this ketone ester
intermediate with ammonium acetate in toluene under heating.
The Boc protection in 8 was selectively removed by the
treatment of trifluoroacetic acid in dichloromethane, and the
freed amine was then coupled with thiazol-5-carboxylic acid 9
B
https://dx.doi.org/10.1021/acsmedchemlett.0c00462
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
pubs.acs.org/acsmedchemlett
Letter
Table 1. PMC Library with Various 2-Substituted Benzamide Zinc Binding Groups
a
b
Reported values are the average of ≥2 independent measurements with standard deviation less than 3-fold of the reported mean. NT = not
tested.
some inhibition activity toward HDAC1 and 2 for much less
HDAC3 selectivity. The 2-methyl ether benzamide 18 also
showed HDAC3 selectivity with 44-fold over HDAC1,
although it is 5-fold less active than the methylthio benzamide
C
https://dx.doi.org/10.1021/acsmedchemlett.0c00462
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
pubs.acs.org/acsmedchemlett
Letter
16. On the other hand, when this methyl ether is replaced by a
hydroxyl, compound 19 became a HDAC1, 2, 3 inhibitor with
high inhibition potencies without any HDAC3 selectivity. The
6-fluoro substituent did not have much impact on the potency
and selectivity for both phenol (20) and methyl ether (21).
Again interestingly, when the two ortho positions of the
benzamide were substituted by hydroxy and methyl ether in
22, it is a HDAC3 selective inhibitor like the methyl ether
benzamide 18. Both 4- and 5-methyl ether-2-hydroxybenza-
mides (23, 24) exhibited the same inhibition pattern of 2-
hydroxybenzamide 19 as HDAC1, 2, 3 inhibitors. The 3-
methyl ether-2-hydroxybenzamide (25) lost all inhibition
activity to all HDACs probably due to unfavorable interactions
in the binding pocket. Unlike the aminopyridine compound
15, the 3-hydroxypyridinyl-2-carboxamide compound 26 was a
HDAC1, 2, 3 inhibitor. Consistently, all the HDAC3 selective
inhibitors in Table 1 displayed no activity in the cellular assay
to activate the HIV latency, while all HDAC1, 2, 3 inhibitors
showed a certain level of potency and activation in the cellular
assay.
Table 2. PMC Library with Various 2-Substituted
Benzamide Zinc Binding Groups
The first round PMC focus library led to the discovery of a
series of highly HDAC3 selective inhibitors with 2-substituted
benzamide. We noticed that when the 2-substituents were
good zinc binding groups such as hydroxy and amino, the
compounds were mostly likely HDAC1, 2, 3 inhibitors.
However, when the 2-substituents were not good zinc binders
and even nitrogen, oxygen, and sulfur with substituents, the
compounds were HDAC3 selective inhibitors with various
potencies and selectivities. To further explore the SAR, PMC
focus libraries were designed using benzamide with bulkier 2-
substituents to investigate the impact on HDAC3 inhibition
and selectivity. Representative compounds are listed in Table
2. HDAC5, 6 inhibition and cellular assay data were not listed
since these compounds were not active against those isoforms.
The 2-dimethylamino 27 and the 2-N-pyrolyl 28 were much
less active in HDAC3 inhibition and less HDAC3 selective
compared to the methylamino compound 13. The larger 2-
ethylthio substituent (29) also led to a 20-fold loss of HDAC3
inhibition potency relative to the methylthio compound 16,
though the HDAC3 selectivity is probably maintained since 29
was completely inactive in HDAC1, 2 inhibition. The electron-
withdrawing and bigger substituents trifluoromethyl ether (30)
and trifluoroethyl ether (31) led to a significant loss of
HDAC3 inhibition activity compared to the methyl ether
substituent (18), although 30 and 31 maintained the HDAC3
selectivity. The small electron-withdrawing substituents 2-
chloro (33), cyano (34), and sulfinyl (35) also led to a loss of
HDAC3 inhibition but still retain certain levels of HDAC3
selectivity. The 2-alkyl substituents such as 2-ethyl (35), 2-
cyclopropyl (36), and the fused dihydroindene carboxamide
(37) all demonstrated high HDAC3 selective inhibition. The
2-cyclopropylbenzamide 36 had an HDAC3 inhibition activity
with an IC50 of 170 nM and 147-fold selectivity over HDAC1.
The cyclopropyl (36) may have the shape fits best to the
pocket compared to ethyl (35) and fused cyclopentyl (37).
This 2-substituted benzamide SAR confirmed that installing
nonzinc binding substituents on the 2-position of the
benzamide led to HDAC3 selective inhibitors with the
methylthio compound 16 as the most potent HDAC3 selective
inhibitor.
a
Reported values are the average of ≥2 independent measurements
with standard deviation less than 3-fold of the reported mean.
nM) and HDAC11 (IC50 = 2,000 nM), although this is 69-fold
less active relative to HDAC3.
The HDAC3 selective inhibitor 16 was characterized in a
global counter screen and pharmacokinetic (PK) tests (Table
3). Compound 16 showed very strong activity toward the
potassium ion channel with an hERG IC50 of 16 nM.20 In the
pharmacokinetic profiling when dosed at 0.05 mg/kg intra-
venously (IV) in rats, 16 did not show any drug concentration
above the detectable level, presumably due to its high
clearance. A similar SAR strategy to the other imidazole core
HDAC inhibitors we reported13−15,21 was applied here to the
We profiled the HDAC3 selective inhibitor 16 against all
subtypes of histone deacetylases 1 to 11. Besides HDAC1 and
2, 16 also showed weak inhibition of HDAC10 (IC50 = 10,000
D
https://dx.doi.org/10.1021/acsmedchemlett.0c00462
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
pubs.acs.org/acsmedchemlett
Letter
Table 3. HDAC Screen for HDAC3 Selective Inhibitor 16
HDAC1
HDAC2
HDAC3
HDAC4
HDAC5
HDAC6
HDAC7
HDAC8
HDAC9
HDAC10
HDAC11
IC50 (nM)
Comp# IC50 (nM) IC50 (nM) IC50 (nM) IC50 (nM) IC50 (nM) IC50 (nM) IC50 (nM) IC50 (nM) IC50 (nM) IC50 (nM)
16 20,000 31,000 29 >40,000 >40,000 >45,000 >40,000 0.45,000 >40,000 10,000
2,000
Table 4. SAR of HDAC3 Selective Inhibitors to Improve hERG Inhibition and Pharmacokinetic Profile
a
b
Winstar Han rats were the PK test species, dosing vehicle = DMSO/PEG400/water: 20/60/20. BLQ = Below the lower limit of quantification of
0.00220 μM (i.e., 1.00 ng/mL).
HDAC3 selective inhibitor series, to improve the off-target
selectivity and PK profile (Table 4). Replacing the 2-
naphthalenyl with 2-methoxy-3-quinolinyl in compound 38
retained the HDAC3 selectivity and inhibition and at the same
time, significantly reduced the hERG activity22 with an IC50 of
1,329 nM. HDAC3 inhibitor 38 also showed limited drug
exposure of AUC 0.049 μM·h in the PK test23 when dosed IV
at 0.05 mg/kg. Upon further modification by replacing the
imidazole with a 1,3,4-oxodiazole in 39, the HDAC3 inhibition
was weaker with an IC50 of 857 nM. The hERG off target
activity is also much weaker for 39. Compound 39 also
demonstrated a similar PK profile to 38 in the IV PK test in
rats. The SAR here proved that this HDAC3 inhibitor lead can
be optimized to improve the off-target selectivity and the
pharmacokinetic profile.
To further investigate the observed HDAC isoform
selectivity of 2-substituted benzamide HDACis, we determined
the X-ray crystal structures of human HDAC2 bound to
compounds 16 and 20 at resolutions of 2.7 and 1.3 Å,
respectively (Figure 2).24 Crystals of the wild-type human
HDAC2 were used for inhibitor 20, while crystals of human
HDAC2 with a single amino acid insertion of residue Tyr205,
designed to better reflect the amino acid insertion in HDAC3,
were used for inhibitor 16. Interestingly, the interactions of 16
and 20 with HDAC2 are quite similar on the surface and both
inhibitors bind with their hydrophobic alkyl linkers in the same
conformation, consistent with the interactions previously
reported for the crystal structures of HDAC2 bound to 1
and 2.13,14 However, there are significant differences in the
binding mode of the 2-substituted benzamides with HDAC2 at
the zinc binding site. The HDAC2 enzyme with compound 20
has a similar conformation as reported before for other
HDAC1, 2, 3 inhibitors.13−15 The phenol oxygen and the
benzamide carbonyl chelate the zinc ion in a bidentate mode.
Unexpectedly, the HDAC3 selective inhibitor 16 bound to
HDAC2 with the amide carbonyl making a monodentate
contact to the zinc ion, and the methylthiophenyl group
rotated in a manner that induced a rotamer change of the
conserved catalytic site tyrosine (Tyr305) from an “in” to “out”
position, significantly expanding the volume of the inhibitor
binding pocket (Figure 2B). The energetic cost of the
conformational rearrangement of the HDAC2 binding pocket
to accommodate inhibitor 16 with only monodentate zinc
binding may explain why the methylthiobenzamide is much
less active against HDAC2 and HDAC1. The structural data
therefore suggests that inhibitor 16 maintains potency against
HDAC3 either by allowing a high affinity bidentate zinc
binding mode in the HDAC3 active site or possibly by other
structural features of HDAC3 that would stabilize the flipped
out conformation of the catalytic site tyrosine. Interestingly, we
note the previously reported HDAC3 crystal structure (Figure
2C) was determined as a complex with inositol tetraphosphate
(IP4) and the DAD domain of the SMRT corepressor (PDB
4A69).25 The catalytic pocket of HDAC3 is quite similar to
HDAC2 with the equivalent tyrosine side chain phenol at the
“in” position. However, the HDAC3 crystal structure also
showed the proximity of Tyr298 and the nearby Arg265 to IP4.
If compound 16 binds HDAC3 in the monodentate zinc
binding mode, the induced conformation of Tyr298 and
concomitant changes to Arg265 would be expected to be
influenced by the presence of IP4 in the HDAC3/IP4/SMRT
complex. While a crystal structure of HDAC3 bound to
inhibitor 16 will be required to resolve these possibilities, these
results nonetheless reveal unexpected conformational plasticity
and an induced subpocket dependent on a rotamer change of
the catalytic site tyrosine that is exploited by exquisitely
E
https://dx.doi.org/10.1021/acsmedchemlett.0c00462
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
pubs.acs.org/acsmedchemlett
Letter
curve analysis (GPCA) method26 to measure the on rate for
both inhibitors to associate with HDAC3. Interestingly, the
HDAC3 selective inhibitor 16 had a fast on rate, although it
needs to induce a pocket with quick conformational change of
the protein, with a kon of 1.5 × 104 M−1 S−1 while the HDAC1,
2, 3 inhibitor 20 also had a fast on rate, but relatively slower
than 16 with kon of 1.7 × 103 M−1 S−1 (detailed measurement
can be found in the Supporting Information). The koff of both
compounds can be calculated based on the enzyme binding
equation KD = koff/kon, in which KD is usually the same as the
IC50 of the inhibitors against HDAC3. koff for the HDAC1, 2, 3
inhibitor 20 is calculated to be 4.0 × 10−6 s−1 which results in a
calculated residence time = 1/koff of 69.4 h. However, the
calculated koff for the HDAC3 selective inhibitor 16 is 4.4 ×
10−4 s−1 which results in a much shorter calculated residence
time of 38 min. Furthermore, this significant difference of off
rate and residence time for 20 and 16 was also confirmed by a
jump dilution assay of both compounds in HDAC3 (Figure 3).
Figure 3. Jump dilution assays to measure off rates: (A) jump dilution
for HDAC1, 2, 3 inhibitor 20; (B) jump dilution for HDAC3
inhibitor 16. RFU = relative fluorescence units.
Figure 2. X-ray crystal structures of HDAC inhibitors binding to the
HDAC2 enzyme with the pocket shown in gray: (A) zinc binding
region of HDAC1, 2, 3 selective inhibitors 20; (B) zinc binding region
of HDAC3 selective inhibitor 16; (C) HDAC3 X-ray crystal structure
as a complex with the IP4/SMRT catalytic pocket (PDB code 4a69).
The HDAC inhibitors with a concentration of 10 × IC50, 0.1 ×
IC50, and DMSO were treated with HDAC3 enzyme and the
enzyme activity was monitored by the relative fluorescence
units. The jump dilution of the assay solutions by 100 folds
were performed on the 10 × IC50 assay solution which resulted
in the inhibitor concentration to 0.1 × IC50 concentration. The
HDAC1, 2, 3 inhibitor 20 had a very slow off rate with quite
long residence time. Although the inhibitor concentration is
diluted to a 0.1 × IC50 concentration of 0.7 nM, the inhibitor
still binds to the HDAC3 enzyme within the experimental
time, which resulted in the continuous inhibition of enzyme
activity. However, the HDAC3 selective inhibitor 16 has a fast-
off rate, when jump dilution was performed on the assay
solution with 10 × IC50 concentration to 0.1 × IC50
concentration of 3 nM, inhibitor 16 was released off the
HDAC3 enzyme. Thus, the HDAC3 enzyme activity is
recovered to a similar rate to that of the assay with 3 nM
inhibitor. The off rate for 16 was measured from this jump
dilution to be koff = 3.6 × 10−04 s−1 which matches very well
selective HDAC3 inhibitors. In the meantime, SAR of 2-
substitution on the benzamide demonstrated that several
substituents unable to coordinate the zinc ion, such as
cyclopropyl (36), remain as potent HDAC3 selective
inhibitors. Compound 36 must be monodentate and therefore
should prefer the induced pocket to fit the 2-substituents. The
SAR is more consistent with and supports the hypothesis that
this series of 2-substituted benzamide HDAC3 selective
inhibitors bind to zinc in a monodentate fashion in the
induced pocket.
We also investigated binding kinetics for the two
representative compoundsHDAC1, 2, 3 inhibitor 20 and
HDAC3 selective inhibitor 16. We applied the global progress
F
https://dx.doi.org/10.1021/acsmedchemlett.0c00462
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
pubs.acs.org/acsmedchemlett
Letter
with that calculated from the assay with the GPCA method.
Since the off rate of 20 is very slow, the koff could not be
obtained from the jump dilution assays. This also confirmed
the calculated slow off rate for 20 from data with the GPCA
method.
In the HIV latency in the Jurkat 2C4 cell model, we reported
that HDAC1, 2, 3 inhibitors such as compounds 1 and 2 can
induce HIV transcription.13,14 However, the HDAC1, 2
selective inhibitor 3 and the HDAC3 selective inhibitor 16
are all inactive in this cellular assay. To further check if the
inhibition of all HDAC1, 2, and 3 are required for the
activation of the cellular HIV latency model, we carried out a
test with the combinations of HDAC3 selective inhibitor 16
and HDAC1, 2 selective inhibitor 3 (Figure 4). Three titration
confirm that inhibition alone on HDAC3 does not activate
HIV latency in the Jurkat 2C4 cells, but combination of 16
with HDAC1,2 inhibitor 3 is sufficient to activate HIV latency.
The HDAC3 selective inhibitor 16 and HDAC1,2,3 inhibitor
20 demonstrated different binding kinetic with dramatically
distinctive length of residence time. The off target and
pharmacokinetic profile of this series of HDAC3 inhibitors
can be optimized by SAR of the surface binding groups.
Overall, we discovered a series of highly HDAC3 selective
inhibitors with 2-substituted benzamides as the zinc binding
groups which can serve as tool compounds to explore
therapeutic indications.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acsmedchemlett.0c00462.
Synthetic methods, characterization of key compounds,
biology assay protocols, and data (PDF)
AUTHOR INFORMATION
■
Corresponding Author
Jian Liu − Merck & Co., Inc., Kenilworth, New Jersey 07033,
United States; orcid.org/0000-0002-8408-7689;
Phone: +1-8483913294; Email: jianliu9600@gmail.com
Figure 4. Cellular 2C4 activation by combinations of 16 (HDAC3
inhibitor) and 3 (HDAC1, 2 inhibitor).
Authors
curves were obtained with a constant concentration of
HDAC1, 2 selective inhibitor 3 at 0 (red), 20000 (purple),
and 40000 nM (green), combining with various concentrations
of HDAC3 selective inhibitor 16 at 0, 78, 156, 1250, 2500,
5000, 10000, and 20000 nM. No cellular activity was observed
when there is no HDAC1, 2 inhibitor 3 in the test. There is
enhanced activation with a combination of 16 at low
concentration, especially at 156 nM of 16 when the activation
percentage reached the maximum. However, with an increasing
concentration of 16, the percentage of activation dropped due
to the cytotoxicity of the combination of high concentrations
of both 3 and 16. This experiment confirmed again that
inhibition of HDAC1, 2, and 3 is sufficient to cause activation
of HIV latency in the Jurkat 2C4 cell model. The combination
of the HDAC1, 2 inhibitor (inactive in 2C4 cells alone) and
HDAC3 inhibitor (inactive in 2C4 cells alone) can lead to HIV
latency activation. These data demonstrate that HDAC1/2 and
3 inhibition is required for HIV latency reversal and that
inhibition of HDAC3 alone is not sufficient.
In summary, taking advantage of a parallel medicinal
chemistry strategy, we discovered a series of HDAC inhibitors
with 2-substituted benzamide as the zinc binding group. When
the 2-substituents are involved in the zinc bidentate
coordination, the inhibitors are HDAC1, 2, and 3 selective
represented by 20. However, if the 2-substituents do not
chelate with the zinc ion, it could induce a binding pocket by
flipping the Tyr305 side chain from an “in” to an “out”
conformation which was observed in the X-ray crystal structure
of 16 bound to the HDAC2 enzyme. The selectivity of 16 with
inhibition to HDAC3 is rationalized by this induced pocket
protein conformation of HDAC3 which is speculated to be
stabilized by the HDAC3/IP4/SMRT complex while no
comparable stabilization exists for HDAC1 and 2. This highly
HDAC3 selective inhibitor 16 served as the tool compound to
Younong Yu − Merck & Co., Inc., Kenilworth, New Jersey
07033, United States
Joseph Kelly − Merck & Co., Inc., Kenilworth, New Jersey
07033, United States
Deyou Sha − Merck & Co., Inc., Kenilworth, New Jersey 07033,
United States
Abdul-Basit Alhassan − Merck & Co., Inc., Kenilworth, New
Jersey 07033, United States
Wensheng Yu − Merck & Co., Inc., Kenilworth, New Jersey
07033, United States; orcid.org/0000-0001-5978-9454
Milana M. Maletic − Merck & Co., Inc., Kenilworth, New Jersey
07033, United States
Joseph L. Duffy − Merck & Co., Inc., Kenilworth, New Jersey
07033, United States
Daniel J. Klein − Merck & Co., Inc., West Point, Pennsylvania
19486, United States
M. Katharine Holloway − Merck & Co., Inc., West Point,
Pennsylvania 19486, United States
Steve Carroll − Merck & Co., Inc., West Point, Pennsylvania
19486, United States
Bonnie J. Howell − Merck & Co., Inc., West Point,
Pennsylvania 19486, United States
Richard J. O. Barnard − Merck & Co., Inc., West Point,
Pennsylvania 19486, United States
Scott Wolkenberg − Merck & Co., Inc., West Point,
Pennsylvania 19486, United States
Joseph A. Kozlowski − Merck & Co., Inc., Kenilworth, New
Jersey 07033, United States
Complete contact information is available at:
https://pubs.acs.org/10.1021/acsmedchemlett.0c00462
Notes
The authors declare no competing financial interest.
G
https://dx.doi.org/10.1021/acsmedchemlett.0c00462
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
pubs.acs.org/acsmedchemlett
Letter
Holloway, K.; Wu, J.; Wu, G.; Howell, B. J.; Barnard, R. J. O.;
Kozlowski, J. A. Selective Class I HDAC inhibitors Based on Aryl
Ketone Zinc Binding Induce HIV-1 Protein for Clearance. ACS Med.
Chem. Lett. 2020, 11 (7), 1476−1483.
(15) Clausen, D. J.; Liu, J.; Yu, W.; Duffy, J.; Chung, C. C.; Myers,
R. W.; Klein, D. J.; Fells, J.; Holloway, K.; Wu, J.; Wu, G.; Howell, B.
J.; Barnard, R. J. O.; Kozlowski, J. A. Development of a selective
HDAC inhibitor aimed at reactivating the HIV latent reservoir. Bioorg.
Med. Chem. Lett. 2020, 30 (17), 127367.
(16) Methot, L. J.; Hoffman, M. D.; Witter, J. D.; et al. Delayed and
Prolonged Histone Hyperacetylation with a Selective HDAC1/
HDAC2 Inhibitor. ACS Med. Chem. Lett. 2014, 5, 340−345.
(17) For a review of HDAC3 inhibitors, see: Sarkar, R.; Banerjee, S.;
Amin, S. A.; Adhikari, N.; Jha, T. Histone deacetylase 3 (HDAC3)
inhibitors as anticancer agents: A review. Eur. J. Med. Chem. 2020,
192, 112171.
(18) Bresciani, A.; Ontoria, J. M.; Biancofiore, I.; Cellucci, A.;
Ciammaichella, A.; Di Marco, A.; Ferrigno, F.; Francone, A.;
Malancona, S.; Monteagudo, E.; Nizi, E.; Pace, P.; Ponzi, S.;
Rossetti, I.; Veneziano, M.; Summa, V.; Harper, S. Im-proved
Selective Class I HDAC and Novel Selective HDAC3 Inhibitors:
Beyond Hydroxamic Acids and Benzamides. ACS Med. Chem. Lett.
2019, 10 (4), 481−486.
ABREVIATIONS
■
HDAC, histone deacetylase; HATs, histone acetyl transferases;
HDACi, histone deacetylase inhibitor; HIV, human immuno-
deficiency viruses; LRA, latency reversing agency; PMC,
parallel medicinal chemistry; cART, combined antiretroviral
therapy; SAR, structure activity relationship; DMF, dimethyl-
formaldehyde; THF, tetrahydrofuran; DCM, dichloromethane;
HATU, (1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo-
[4,5-b]pyridinium 3-oxide hexafluorophosphate; TFA, tri-
fluoroacetic acid; DIEA, diisopropylethylamine; hERG,
̀
human ether-a-go-go-related gene; GPCA, global progress
curve analysis; PK, pharmacokinetic.
REFERENCES
■
(1) Grunstein, M. Histone acetylation in chromatin structure and
transcription. Nature 1997, 389, 349−352.
(2) Pazin, M. J.; Kadonaga, J. T. What’s Up and Down with Histone
Deacetylation and Transcription? Cell 1997, 89, 325−328.
(3) Gregoretti, L.; Lee, Y.; Goodson, H. V. Molecular Evolution of
the Histone Deacetylase Family: Functional Implications of
Phylogenetic Analysis. J. Mol. Biol. 2004, 338, 17.
(19) Pearson, R.; Kim, Y. K.; Hokello, J.; Lassen, K.; Friedman, J.;
Tyagi, M.; Karn, J. Epigenetic Silencing of Human Immuno-deficiency
Virus (HIV) Transcription by Formation of Re-strictive Chromatin
Structures at the Viral Long Terminal Repeat Drives the Progressive
Entry of HIV into Latency. J. Virol. 2008, 82, 12291−12303.
(20) The hERG activity was measured by inhibition of radiolabeled
MK-0499, a known hERG blocker, binding to the hERG channel.
(21) Yu, W.; Liu, J.; Clausen, D.; Yu, Y.; Duffy, J. L.; Wang, M.; Xu,
S.; Deng, L.; Takao, S.; Chung, C. C.; Myers, R. W.; Klein, D. J.; Fells,
J.; Holloway, M. K.; Wu, J.; Wu, G.; Howell, B. J.; Barnard, R. J. O.;
Kozlowski, J. Lead optimization of ethyl ketone based HDAC1, 2 and
3 selective inhibitors for HIV latency reactivation. J. Med. Chem.,
submitted.
(22) hERG activity is measured by the inhibition of the compounds
to the binding of MK-499 to potassium ion channel (IKr).
(23) All pharmacokinetic studies in preclinical species were conducted
according to the highest ethical standards in accordance with and using
procedures approved by the Institutional Animal Care and Use;
Committee of Merck and Co., Inc, West Point, PA, USA.
(24) The data of X-ray crystal structures for HDAC3 inhibitor 16
and HDAC1, 2, 3 inhibitor 20 is uploaded to the Protein Data Bank
(www.pdb.org) with codes 7KBH and 7KBG, respectively. The
detailed procedure for crystallization, data collecting, and processing
can be found in the Supporting Information.
(4) Miller, T. A.; Witter, D. J.; Belvedere, S. Histone deacetylase
inhibitors. J. Med. Chem. 2003, 46, 5097−5116.
(5) Paris, M.; Porcelloni, M.; Binaschi, M.; Fattori, D. Histone
deacetylase inhibitors: from bench to clinic. J. Med. Chem. 2008, 51,
1505−1529.
(6) Bieliauskas, A. V.; Pflum, M. K. H. Isoform-selective histone
deacetylase inhibitors. Chem. Soc. Rev. 2008, 37, 1402−1413.
(7) Spiegel, S.; Milstien, S.; Grant, S. Endogenous modulators and
pharmacological inhibitors of histone deacetylases in cancer therapy.
Oncogene 2012, 31, 537−551.
(8) Archin, N. M.; Liberty, A. L.; Kashuba, A. D.; Choudhary, S. K.;
Kuruc, J. D.; Crooks, A. M.; Parker, D. C.; Anderson, E. M.; Kearney,
M. F.; Strain, M. C.; Richman, D. D.; Hudg-ens, M. G.; Bosch, R. J.;
Coffin, J. M.; Eron, J. J.; Hazuda, D. J.; Margolis, D. M.
Administration of vorinostat disrupts HIV-1 latency in patients on
antiretroviral therapy. Nature 2012, 487 (7408), 482−5.
(9) Rasmussen, T. A.; Tolstrup, M.; Brinkmann, C. R.; Olesen, R.;
Erikstrup, C.; Solomon, A.; Winckelmann, A.; Palmer, S.; Dinarello,
C.; Buzon, M.; Lichterfeld, M.; Lewin, S. R.; Ostergaard, L.; Sogaard,
O. S. Panobinostat, a histone deacetylase inhibitor, for latent-virus
reactivation in HIV-infected patients on suppressive antiretroviral
therapy: a phase 1/2, single group clinical trial. Lancet HIV 2014, 1
(1), e13−21.
̈
(10) Tapia, G.; Højen, J. F.; Okvist, M.; Olesen, R.; Leth, S.; Nissen,
(25) Watson, P. J.; Fairall, L.; Santos, G. M.; Schwabe, J. W. R.
Structure of HDAC3 Bound to Co-Repressor and Inositol
Tetraphosphate. Nature 2012, 481, 335.
(26) Zhang, R.; Windsor, W. T. In vitro kinetic profiling of hepatitis
C virus NS3 protease inhibitors by progress curve analysis. Methods
Mol. Biol. 2013, 1030, 59−79.
S. K.; VanBelzen, D. J.; O’Doherty, U.; Mørk, A.; Krogsgaard, K.;
Søgaard, O. S.; Østergaard, L.; Tolstrup, M.; Pantaleo, G.;
Sommerfelt, M. A. Sequential Vacc-4x and romidepsin during
combination antiretroviral therapy (cART): Immune responses to
Vacc-4x regions on p24 and changes in HIV reservoirs. J. Infect. 2017,
75, 555−571.
(11) Barton, K. M.; Archin, N. M.; Keedy, K. S.; Espeseth, A. S.;
Zhang, Y.-L.; Gale, F. J.; Wagner, L. F.; Holson, B. E.; Margolis, D. M.
Selective HDAC Inhibition for the Disruption of Latent HIV-1
Infection. PLoS One 2014, 9 (8), e102684.
(12) Huber, K.; Doyon, G.; Plaks, J.; Fyne, E.; Mellors, J. W.; Sluis-
Cremer, N. Inhibitors of Histone Deacetylases − Correlation between
isoform specificity and reactivation of HIV Type 1 (HIV-1) from
Latently Infected Cells. J. Biol. Chem. 2011, 286 (25), 22211−22218.
(13) Yu, W.; Liu, J.; Yu, Y.; Zhang, V.; Clausen, D.; Kelly, J.;
Wolkenberg, S.; Beshore, D.; Duffy, L. J.; Chung, C. C.; Myers, W. R.;
Klein, D. J.; Fells, J.; Holloway, K.; Wu, J.; Wu, G.; Howell, B. J.;
Barnard, R. J. O.; Kozlowski, J. A. Discovery of ethyl ketone-based
HDACs 1, 2, and 3 selective inhibitors for HIV latency reactivation.
Bioorg. Med. Chem. Lett. 2020, 30 (13), 127197.
(14) Liu, J.; Kelly, J.; Yu, W.; Clausen, D. J.; Yu, Y.; Kim, H.; Duffy,
J. L.; Chung, C. C.; Myers, R. W.; Carroll, S.; Klein, D. J.; Fells, J.;
H
https://dx.doi.org/10.1021/acsmedchemlett.0c00462
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX