Molecular Basis for the N-Terminal Bromodomain-and-Extra-Terminal-Family Selectivity of a Dual Kinase-Bromodomain Inhibitor

Article abstract of DOI:10.1021/acs.jmedchem.8b01248

As regulators of transcription, epigenetic proteins that interpret post-translational modifications to N-terminal histone tails are essential for maintaining cellular homeostasis. When dysregulated, "reader" proteins become drivers of disease. In the case of bromodomains, which recognize N-?-acetylated lysine, selective inhibition of individual bromodomain-and-extra-terminal (BET)-family bromodomains has proven challenging. We describe the >55-fold N-terminal-BET bromodomain selectivity of 1,4,5-trisubstituted-imidazole dual kinase-bromodomain inhibitors. Selectivity for the BRD4 N-terminal bromodomain (BRD4(1)) over its second bromodomain (BRD4(2)) arises from the displacement of ordered waters and the conformational flexibility of lysine-141 in BRD4(1). Cellular efficacy was demonstrated via reduction of c-Myc expression, inhibition of NF-κB signaling, and suppression of IL-8 production through potential synergistic inhibition of BRD4(1) and p38α. These dual inhibitors provide a new scaffold for domain-selective inhibition of BRD4, the aberrant function of which plays a key role in cancer and inflammatory signaling.

Full text of DOI:10.1021/acs.jmedchem.8b01248

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Article
Molecular Basis for the N-Terminal Bromodomain and Extra Terminal
(BET) Family Selectivity of a Dual Kinase-Bromodomain Inhibitor
Anand Divakaran, Siva Kumar Talluri, Alex M. Ayoub, Neeraj Mishra, Huarui
Cui, John C. Widen, Norbert Berndt, Jin-Yi Zhu, Angela S Carlson, Joseph J.
Topczewski, Ernst Schönbrunn, Daniel A. Harki, and William C. K. Pomerantz
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b01248 • Publication Date (Web): 25 Sep 2018
Downloaded from http://pubs.acs.org on September 25, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Molecular Basis for the NꢀTerminal Bromodomain
and Extra Terminal (BET) Family Selectivity of a
Dual KinaseꢀBromodomain Inhibitor
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†‡
§‡
§
§
§
Anand Divakaran, Siva K. Talluri, Alex M. Ayoub, Neeraj K. Mishra, Huarui Cui, John
†
#
#
§
§
C. Widen, Norbert Berndt, Jin-Yi Zhu, Angela S. Carlson, Joseph J. Topczewski, Ernst K.
#
†*
‡*
Schonbrunn, Daniel A. Harki, and William C.K. Pomerantz
†
Department of Medicinal Chemistry, University of Minnesota, 2231 6th St SE, Minneapolis,
MN 55455, United States
§ Department of Chemistry, University of Minnesota, 207 Pleasant St. SE, Minneapolis, MN
5455, United States
5
#
Drug Discovery Department, H. Lee Moffitt Cancer Center and Research Institute, 12902
Magnolia Drive, Tampa, FL 33612, United States
KEYWORDS
1
9
Epigenetics, BET Bromodomain, Inflammation,
Inhibitor
FꢀNMR, KinaseꢀBromodomain
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT: As regulators of transcription, epigenetic proteins that interpret postꢀtranslational
modifications to Nꢀterminal histone tails are essential for maintaining cellular homeostasis.
When dysregulated, ‘reader’ proteins become drivers of disease. In the case of bromodomains,
which recognize Nꢀεꢀacetylatedꢀlysine, selective inhibition of individual Bromodomain and
Extra Terminal (BET) family bromodomains has proven challenging. We describe the >55ꢀfold
Nꢀterminal BET bromodomain selectivity of 1,4,5ꢀtrisubstituted imidazole dual kinaseꢀ
bromodomain inhibitors. Selectivity for the BRD4 Nꢀterminal bromodomain (BRD4(1)) over its
second bromodomain (BRD4(2)) arises from the displacement of ordered waters, and the
conformational flexibility of Lysineꢀ141 in BRD4(1). Cellular efficacy was demonstrated via
reduction of cꢀMyc expression, inhibition of NFꢀκB signaling, and suppression of ILꢀ8
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
production through potential synergistic inhibition of BRD4(1) and p38α. These dualꢀinhibitors
provide a new scaffold for domainꢀselective inhibition of BRD4, the aberrant function of which
plays a key role in cancer and inflammatory signaling.
Introduction
How our genetic information is manipulated and ultimately expressed as a heritable phenotype
is a central question in the field of epigenetics. Nꢀterminal modifications on conserved histone
1
,2
proteins are key regulators of transcription. As such, proteins that interpret these complex
modifications through molecular recognition are known to be essential for maintaining cellular
3
,4
homeostasis, or when dysregulated, they become drivers of disease. Bromodomains are a
subset of epigenetic 'reader' or effector domainꢀcontaining proteins, which function via binding
5
to distinct lysine acylation states, most commonly Nꢀεꢀacetylatedꢀlysine (K ) on histones. This
ac
similar mode of molecular recognition for all 61 human bromodomains presents a significant
6
hurdle for finding selective inhibitors as both therapeutic agents and chemical probes.
ACS Paragon Plus Environment
2

Page 3 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
One family of bromodomains with clinical relevance are the bromodomain and extra terminal
BET) proteins, BRD2, BRD3, BRD4, and testis specific BRDT, each of which contain a pair of
(
bromodomains. Recent studies have shown divergent roles of these domains for targeting both
acetylated histones and acetylated transcription factors, through the Nꢀ and Cꢀterminal
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
bromodomains (D1 and D2, respectively). Prominent roles of the BRD4 BET bromodomains
8
include regulation of cꢀMyc expression at superꢀenhancers, and enhancement of NFꢀκBꢀ
7
mediated transcription through interactions with acetylated K310 of the κB RelA subunit.
Due to the biological significance of bromodomain function in cellular proliferation and
inflammatory pathways, over 20 clinical trials are underway for studying the effects of BET
family inhibition as antiꢀcancer and antiꢀinflammatory therapies. Because of the high sequence
identity among BET proteins, most reported inhibitors act as panꢀinhibitors of BET
bromodomains, such as (+)ꢀJQ1 (1, Fig.1), and do not discern the function of individual
9
bromodomains. Ouyang et al. reported FLꢀ411 (2, Fig.1), which is an exception to the norm in
10
being selective for BRD4 over other BET proteins. Nevertheless, the potential offꢀtarget
effects, doseꢀlimiting thrombocytopenia, and concerns over latent viral reactivation associated
with panꢀBET inhibition highlight the need for novel scaffolds with improved selectivity and/or
11–13
potency.
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1: Reported BET bromodomain inhibitors and a new 1,4,5 trisubstituted imidazole, V.
In lieu of isoform specific inhibitors, domain selectivity among the BET family of proteins has
been explored. D2ꢀselective molecules RVXꢀ208 (3, Fig.1) and ꢀ297 have allowed the study of
D2ꢀdependent processes, but have reduced efficacy mitigating BETꢀrelated transcription in
ACS Paragon Plus Environment
4

Page 5 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
14,15
comparison to panꢀBET inhibition.
Alternatively, specific deletion of only BRDT(1) in mice
1
6
was sufficient to impair spermatogenesis. In the case of BRD4, fluorescence recovery after
photobleaching experiments identified similar levels of BRD4 displacement from chromatin
between a D1 defective mutant and (+)ꢀJQ1 (1) treated cells, whereas chromatin binding by a D2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
7
defective mutant was less affected. In relation to NFꢀκB signaling, interactions of BRD4(1) are
1
8
likely more relevant for targeting transcriptional activity. Although molecules MSꢀ436 (4,
1
9–21
Fig.1) and Olinone (5, Fig.1) have been reported to possess some level of D1 selectivity,
BET bromodomain research continues to be dominated by panꢀBET inhibitors. Despite being
potent BRD4(1) inhibitors, 4 and structural analog MSꢀ611 are not selective for D1 over D2 in
other BET proteins. In contrast, D1ꢀselective molecule 5 is a modest inhibitor of BRD4(1) (K =
d
3
.3 µM), but represents the highest level of selectivity. Alternative chemicalꢀbiology strategies
2
2,23
like the ‘bumpꢀandꢀhole’ approach
have considerably advanced our understanding of
individual BETꢀbromodomain and ꢀprotein function. However, improved D1 selective inhibitors
are still required to further dissect the relevant interactions of BRD4 with chromatin and
acetylated transcription factors. The development of such small molecules may have significant
implications for studying the roles of relevant BET proteins in transcription and the treatment of
disease.
In this report we describe dual kinaseꢀbromodomain inhibitors with the strongest affinity for
BRD4(1) among BET bromodomains. While other BET bromodomain inhibitors have been
reported in the literature, a notable feature of this molecule is its selectivity for the first
bromodomains of BET proteins with simultaneous kinase inhibitory activity against mitogenꢀ
activated protein (MAP) kinase p38α. The atomicꢀlevel insights from structural analysis and
structureꢀactivity relationships will inform the future development of BET D1ꢀselective
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
inhibitors, while optimization of this dual inhibitor should afford a useful tool compound to test
synergistic
effects
in
treating
inflammation
and
cancer.
A
B
100
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
IV (D1)
IV (D2)
50
V
V
(D1)
(D2)
BRD4
0
-
2
0
2
100
50
0
BRDT
-
2
0
2
log[Competitor] ( M)
Figure 2: Competitive inhibition experiments via fluorescence anisotropy of SBꢀ284851ꢀBT (I)
and analogs against BET bromodomains indicates selectivity for BETꢀD1s. A. Mean ± SEM IC50
values of analogs determined by competitive fluorescence anisotropy assays (*indicates
incomplete inhibition observed, ND = not determined). B. Competition binding isotherms of
compounds IV and V against the first and second bromodomains of BRD4 and BRDT.
Results
The 1,4,5ꢀtrisubstituted imidazole, SBꢀ284851ꢀBT (I) and related analogs, are known p38α
inhibitors and were recently discovered to bind to BET bromodomains in three separate reports
using diverse libraries, including the published kinase inhibitor set library from
24–26
19
GlaxoSmithKline.
This study tested for BET bromodomain selectivity by using a F NMR
ACS Paragon Plus Environment
6

Page 7 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
26
assay which simultaneously assessed binding to BRD4(1) and the bromodomain of BPTF. SBꢀ
2
84851ꢀBT was selective for BRD4(1) over BPTF in the assay and was further validated as a
ligand in a thermal stability assay, differential scanning fluorimetry (DSF), and a fluorescence
anisotropy (FA) assay displacing fluorescentlyꢀlabeled panꢀBET ligand, BIꢀ6727, referred to
here as BIꢀBODIPY. While our previous study identified selectivity for the BET family, we did
not further evaluate selectivity within the BET family of bromodomains.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1: Synthesis of Common Intermediate 17
2
7,51
Procedure adapted from Boehm et al. and Adams et al.
We first turned our attention to two BET bromodomains, BRDT(1) and BRD4(1). Based on
significant differences in affinity arising from methyl substitution patterns on the phenyl ring of I
2
6
for BRD4(1), we synthesized several new analogs and determined their affinity to these two
27
domains. The synthesis for SBꢀ284851ꢀBT (I) has already been established, and we took
advantage of a late stage intermediate (Scheme 1) for analog synthesis using a nucleophilic
aromatic substitution reaction with various phenolic and anilinic derivatives (Schemes 2ꢀ4), of
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
which several are described here (Fig.2A). Using FA to quantify the activity of these newly
synthesized compounds against both BRDT(1) and BRD4(1), we found that compound I binds to
both domains. The observed affinity for BRD4(1) was 3ꢀfold weaker than the previously
reported data of the compound used directly from the library stock, indicating a possible
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
26
discrepancy in concentration or purity. Consistent with our prior report, the paraꢀsubstituted
analog, II, bound substantially weaker. However, a 3,4ꢀdimethyl substituted analog with an Oꢀ
substituted pyrimidine, compound III, bound with increased affinity (IC50 = 2.9 ± 0.2 µM).
Finally, when the pyrimidine oxygen was changed to a nitrogen, we observed an increase in
potency with IC50 values of 1.3 ± 0.2 µM and 1.7 ± 0.4 µM for the 3,4ꢀ and 3,5ꢀdimethyl
analogs, compounds IV and V, respectively, against BRD4(1). Molecule VI, with a 3,5ꢀdifluoro
substituted aromatic ring (Scheme 4) versus methyl groups did not interact with high affinity
(
IC > 250 µM), supporting the need for an electron rich ring. A direct binding titration of
50
compounds IV and V by isothermal titration calorimetry (ITC) determined entropically favorable
binding events leading to a K of 1.8 µM and 1.2 ꢀM, respectively (Fig.S7).
d
Scheme 2: Synthesis of Compounds I, II and III
Scheme 3: Synthesis of Compounds IV and V
ACS Paragon Plus Environment
8

Page 9 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 4: Synthesis of Compound VI
To evaluate if these molecules would function as panꢀBET inhibitors based on binding to both
BRDT(1) and BRD4(1), we next tested D1 vs. D2 selectivity, as our fluorescent tracer binds to
D2s of BRDT and BRD4 with high affinity (K = 270 and 183 nM). Surprisingly, in all cases,
d
the analogs tested were unable to fully displace BIꢀBODIPY from either D2 construct (Fig.2B).
2
8
Alternatively, panꢀBET inhibitor BI2536 displaced the fluorescent tracer from both domains of
BRDT and BRD4. A D2ꢀselective inhibitor RVXꢀ208 was tested in a similar fashion to further
validate the experimental outcomes seen with our constructs (Fig.S2 and S3). Additionally, a
dual bromodomain construct of BRD4 was tested in our FA assay, and only partial displacement
was observed with compound V, consistent with only fully displacing BIꢀBODIPY from the first
bromodomain (Fig.S4). Molecules IꢀIII also maintained their selectivity for D1 over D2 when
tested against BRDT(2), indicating that neither the substitution pattern on the aromatic ring nor
the O to N substitution on the pyrimidine ring was responsible for the observed selectivity.
The observed preference towards BRD4(1) led us to evaluate selectivity against the two other
BET proteins, BRD2 and BRD3, using an orthogonal assay. Compound V was assayed
2
9
commercially against seven of the eight BET bromodomains in an ALPHAScreen format.
While BRDT(2) was not available for testing by ALPHAScreen, D1 over D2ꢀselectivity was
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
maintained for compound V (Table 1 and Fig.S8). Additionally, compound V displayed 6ꢀ and
1
6ꢀfold selectivity for BRD4 over ubiquitously expressed BET proteins BRD2 and BRD3,
respectively (Fig.S8). Together, these results support a valuable scaffold for developing D1ꢀ
selective inhibitors of BET bromodomains, albeit with the highest affinity for BRD4(1) and
BRDT(1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Structural Analysis of Binding
Due to the paucity of D1ꢀselective inhibitors and the generality of these results within the
trisubstituted imidazole scaffold, we sought to structurally characterize their interactions.
Proteinꢀobserved fluorine NMR (PrOF NMR) is an emerging method for ligand discovery that
30
has been used to characterize the binding interface of small molecules with bromodomains, as
well as quantify the affinity of weak binding molecules via chemical shift perturbation
3
1
experiments over a range of ligand concentrations. PrOF NMR was used to characterize the
binding interactions of compound V with both BRD4(1) and BRDT(1) (Fig.3A), and to further
test selectivity against additional bromodomains at higher small molecule concentrations than
was tolerated in earlier experiments. Our focus in these experiments was primarily on the WPF
shelf tryptophan residue adjacent to the Kac binding site, W81 for BRD4(1) and W50 for
BRDT(1), to see if an unusual effect on resonance perturbation was observed relative to our prior
26
reports. In both BRD4(1) and BRDT(1), using compound V, an intermediateꢀtoꢀslow exchange
perturbation of the WPF shelf resonance was observed that reached the fully bound state at
approximately 1 eq. of small molecule. Such NMR behavior is indicative of a residence time on
the protein that is sufficiently long on the NMR time scale to partially resolve both the bound
and unbound states and is consistent with binding site and affinity of the single digit micromolar
3
2
K exhibited for these domains by compounds IV and V. W44 of BRDT(1) is outside the
d
ACS Paragon Plus Environment
10

Page 11 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
binding site and was also minimally perturbed, indicating a conformational change of the
protein. However these data did not support binding to a new binding site, and was similar to
2
6
observations in our prior reports. We further tested nonꢀBET bromodomains of BPTF, and a
yeast GCN5ꢀlike malarial protein that also contains a WPFꢀshelf. In these cases, we did not
detect binding to either protein, consistent with selectivity for the BET family of bromodomains
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Fig.S9).
We turned to Xꢀray crystallography to obtain higher resolution structural information to
understand the mode of molecular recognition and to rationalize the domain selectivity.
Trisubstituted imidazoles, SBꢀ251527 and SBꢀ284847ꢀBT, have previously been crystallized
2
5
with BRD4(1) (PDB ID: 4O7F and 4O7B, respectively) . In addition, we crystallized
compounds IV and V, with BRD4(1) (Fig.3C). All four compounds engage the protein in a
similar fashion with the imidazole making direct hydrogen bonding interactions with N140,
which is displaced by 0.8 Å relative to the apostructure along with a displacement of the Cα of
the adjacent K141 by 2.2 Å (Fig.3D, E and S11). However, unlike panꢀ BET inhibitors, these
molecules failed to hydrogen bond to Y97 via a bridging structured water. The OꢀN substitution
pattern did alter the orientation of the aromatic ring slightly, which may favor edge to face πꢀπ
interactions with W81 of BRD4(1) in a lipophilic region adjacent to the WPF shelf (Fig.S12) to
impact affinity but not selectivity. The methyl groups were necessary for binding affinity, as a
3,5ꢀdifluoro substituted analog VI only bound weakly to both bromodomains of BRD4. Overlay
of a BRD4(2) crystal structure shows potential steric incompatibilities at multiple regions of the
small molecule, including a steric clash of the C2 imidazole portion of the ring of compounds IV
and V with N433 in the binding site which may not be as conformationally flexible as the
BRD4(1) N140 due to the adjacent P434. These incompatibilities may preclude a high affinity
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
interaction with BRD4(2); however, such a steric effect could not be directly proven from our
1
9
structures. Together FꢀNMR and Xꢀray structural analysis could be used to rationalize binding
interactions but was not sufficient to describe selectivity.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment
12

Page 13 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 3: Confirmation of compound V binding BRD4(1) and BRDT(1), and possible structural
basis for D1 over D2 selectivity of inhibitors. A. PrOF NMR titrations with bromodomains of
1
9
BRD4 and BRDT. Samples contain 50 µM Fꢀlabelled BRD4(1) or BRDT(1) with increasing
concentrations of ligand V. B. Electron density maps (2FoꢀFc; 1σ) from Xꢀray structures of
BRD4(1) liganded with IV (yellow, 1.74 Å, PDB ID: 6MH7) and V (orange, 1.60 Å, PDB ID:
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6MH1). C. Superposition of compounds IV and V in the K site of BRD4(1) reveals subtle
ac
conformational changes of the dimethylphenyl moieties nested in the WPF shelf. D. In the
unliganded state of BRD4(1) (grey, PDB ID: 4IOR), the K site around Asn140 is in a relaxed
Ac
state. Upon binding of inhibitor (orange structure) Asn140 gives way (ꢁ = 0.8 Å) due to steric
hindrance with the imdiazole moiety of the inhibitor (red dotted line), and positions itself for
optimal Hꢀbonding interaction. As a result, the adjacent Lys141 undergoes a large
conformational change (ꢁ = 2.2 Å). E. In unliganded BRD4(2) (green, PDB ID: 2OUO), the
region around Asn433 is highly similar to BRD4(1) (grey) except for the presence of Pro434 for
lysine. The geometric constraints of the AsnꢀProꢀPro sequence in BRD4(2) likely renders the
peptide more rigid and less compatible with the inhibitorꢀinduced conformational changes
observed in BRD4(1).
As an alternative possibility for selectivity, we considered the sequence similarity between
BETꢀD1 and D2. Of the three nonꢀconserved amino acids between the binding sites, we
hypothesized the proximity of the piperdyl group to D144 in our crystal structures was
potentially responsible for the observed selectivity trends. This BET D1ꢀconserved aspartic acid
is replaced by a histidine (H437 of BRD4) in BETꢀD2s. At physiological pH, H437 will be
weakly protonated, and may result in potential electrostatic repulsions in addition to steric effects
that prevent binding to D2. To determine whether the origins of D1 vs D2 selectivity are due to
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
H437 in D2, we expressed a H437D mutant of BRD4(2) construct to restore binding of
compound V. For this experiment, we used a new fluorescent tracer for a FA assay, based on a
fluoresceinꢀlabeled (+)ꢀJQ1 molecule. (+)ꢀJQ1 bound to H437D BRD4(2) with comparable
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
affinity to the WT D2 protein in direct binding assays (Fig.S5; K BRD4(2) = 120 nM, BRD4(2)
d
H437D = 200 nM), with a slightly attenuated IC in competition assays (IC BRD4(2) = 490
5
0
50
nM, BRD4(2)ꢀH437D = 590 nM). Binding of V to BRD4(2)ꢀH437D however was not observed
Fig.S6), suggesting the observed selectivity trends are not entirely attributable to H437 of
BRD4(2).
(
ACS Paragon Plus Environment
14

Page 15 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 4: Displacement of Structured Waters from BRD4(1). Crystal structure of BRD4(1)
with DMSO (left, yellow PDB ID: 4IOR) and Compound V (right, yellow PDB ID: 6MH1),
where structurally conserved water molecules are shown in cyan.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Displacement of Structured Waters
Recent computational studies suggest a second possible explanation for the observed
selectivity based on differential affinities for the four structured waters within the BET
bromodomain binding sites. In these reports, the stability of bromodomain water networks have
identified the water molecules most likely to be displaced are in BRD4(1) and BRDT(1) amongst
3
3,34
the other BET bromodomains,
whose water molecules are more tightly bound to the protein.
Such an effect would also explain the observed selectivity of compound V for BRD4(1) and
BRDT(1) over the other BETs measured in the ALPHAScreen assay. We compared the coꢀ
crystal structure of compound V with BRD4(1) and a DMSO molecule bound to the same
protein (Fig.4). Whereas in the DMSO coꢀstructure six water molecules are present including the
four conserved structured waters, three water molecules are displaced in the coꢀcrystal structure
with compound V, and the water network is reorganized by the fluorophenyl group of compound
V. Similar displacements were observed in the coꢀcrystal structure with compound IV (not
shown). Displacement of water in the binding site is also consistent with the positive entropy of
binding in ITC experiments with compounds IV and V (5.0 and 4.2 cal/mol/deg respectively).
Importantly, this mode of binding and selectivity are not observed in BETꢀD1 selective
molecules 4 (Olinone) and 5 (MSꢀ436), however a remarkably similar binding mode has been
35
observed in a selective TRIM24 bromodomain inhibitor. These results may in part give rise to
our observed selectivity, but the generality of such a finding will be useful to study in future
investigations.
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1: ALPHAScreen analysis of compounds V, VII, VIII and IX against 7 of 8 BET
bromodomains
IC by ALPHAScreen (µM)
5
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
NH
R
BRD4
D1 D2
BRD3
D1 D2
BRD2
D1 D2
BRDT
D1
R
X
N
HN
N
X N
F
F
V
C
1.8 >100 11.2 >100 29
67
30
5.3
36
VII
VIII
IX
N
N
N
27
75
51
110
43
160 102
1.3 0.56
200 310
0.77 2.4
>250 260 110
HO
H C
3
1.8
12
1.6
To further evaluate the displacement of structured waters by the fluorophenyl group and
additional effects around the imidazole ring which removes steric effects from a CꢀH at position
C2, we synthesized three new molecules (VII, VIII, and XI), based on a 1,4,5ꢀtrisubsititued
triazole scaffold (Scheme 5). These molecules either maintain the fluorophenyl group for a direct
comparison with V, or replace this phenyl group with a smaller hydroxymethyl or methyl group
ACS Paragon Plus Environment
16

Page 17 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
to maintain the structured waters. Compound VII bound BRD4(1) by PrOF NMR in a manner
similar to compound V (Fig.S10). But due to solubility challenges at high concentrations and the
low sensitivity of FA, ALPHAScreen was used to test for BET selectivity. Replacement of the
imidazole ring with a triazole in compound VII resulted in a moderate loss of potency to
BRD4(1) (IC = 1.8 ꢂM for V vs. 27 ꢂM for VII) likely due to the weakened hydrogen bonding
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
with conserved N140 of BRD4 and a significant loss of selectivity (27 ꢂM D1 vs 76 ꢂM D2)
Replacement of the fluorophenyl group with a methoxy (VIII) group resulted in further reduced
binding with IC > 100 µM, possibly due to the hydrophobic nature of the binding pocket.
5
0
Finally, in contrast, truncation of the fluorophenyl group to a methyl in compound IX enhanced
binding 20ꢀfold relative to VII (IC = 1.3 µM for IX vs 155 µM for VII). The large gain in
5
0
potency indicates an energetic penalty must be overcome for displacing the structured waters
with a large aromatic group. In the case of compound IX, which lacks a large aromatic group
displacing structured waters, an inversion of BRD4 domain selectivity favoring BRD4(2) (IC =
5
0
0
.56 µM), and overall panꢀBET inhibition were observed.
Scheme 5: Synthesis of triazoles VII, VIII and IX
OMs
N
N
3
CuSO
4
5 H
2
O
NBoc
NaN
C, 20 h
3
, DMF
R
Sodium Ascorbate
R
N
N
N
+
7
5
ο
R = p-C
R = CH
R = CH
6
H
4
F
t-BuOH:H O (1:1)
2
N
2
OTBS
rt, 18 h
Boc
Boc
6 4
25 R = p-C H F (77%)
2
4 (quant.)
3
26 R = CH
7 R = CH
2
3
OTBS (83%)
(quant.)
2
S
N
N
O
S
O
S
R
N
N
Cl
n-BuLi, ZnCl
2
, THF
N
N
6
0
ο
C, 18 h
NBoc
Oxone, THF
NBoc
R
N
rt, 20 h
N
N
N
N N
2
2
8 R = p-C
9 R = CH
6
H
4
F (92%)
6 4
31 R = p-C H F (89%)
HN Pd OMs
Cy
MeO
32 R = CH OTBS (66%)
3
33 R = CH (quant.)
2
2
OTBS (88%)
(52%)
2
P
30 R = CH
3
OMe
SPhos Pd G3
HN
R
N
HN
N
N
TFA, DCM
N
NH
2
NBoc
NH
R
N
N
a. or b.
N
N
N N
3
4 R = p-C
35 R = CH
6 R = CH
6
H
4
F (59%)
OTBS (59%)
(68%)
VII R = p-C
VIII R = CH
IX R = CH
6
H
4
F (quant.)
OTBS (quant.)
3
(quant.)
2
2
3
3
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ο
ο
Reagents & Conditions: (a) 100 C, 3 h for synthesis of 34 and 36, or (b) NaHMDS, ꢀ78 C, 0.5
h for 35
From these data, we conclude that the D1 over D2 selectivity cannot be attributed to the
methylꢀsubstitution pattern of the pyrimidineꢀsubstituted aromatic ring, the phenoxyꢀpyrimidine
to anilinoꢀpyrimidine substitution, or even steric incompatibilities with the piperdyl group.
Rather, the observed selectivity is due to a combination of effects including the imidazole ring
itself, and from the fluorophenyl ring displacing structured waters.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Bromodomain and Kinase Selectivity Profiling
To more broadly test the bromodomain selectivity of compound V, we used a commercial
assay platform to characterize the activity against 32 of the 61 bromodomains provided by
DiscoverX. Based on a measured K of 1.2 µM, we tested the selectivity of V at 10 µM
d
(
approximately 10ꢀfold above the K ). Consistent with our FA and ALPHAScreen results, the
d
two most significant interactions of the eight BET bromodomains with V were with BRD4(1)
and BRDT(1), with 86% and 50% inhibition at that concentration, respectively (Table S3). The
BET D2s were the most unaffected (7ꢀ27% inhibition) relative to D1. Overall, only
bromodomains, SMARCA2, PCAF, p300, and SMARCA4 had inhibitory activity between 40ꢀ
5
0%. These results demonstrate the selectivity of compound V towards BET D1s and given the
modular synthesis of the 1,4,5ꢀtrisubstituted imidazole scaffold, represent a useful starting point
for domainꢀselective inhibitor development.
As compound I is a reported MAPꢀkinase inhibitor, we were interested in verifying the
selectivity of these analogs against other kinases. Similar analogs of I either maintaining
pyrimidylꢀaniline substitutions but lacking dimethyl groups, or containing methylated piperidinyl
groups have been previously reported, where both compounds are highly potent and selective for
3
6
p38α and β at 100 nM. As such, we expected similar affinity for p38α and a similar kinase
ACS Paragon Plus Environment
18

Page 19 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
selectivity profile. Indeed, the K of analog V against p38α was determined to be 0.47 nM
d
(Fig.S13). Due to the weaker D1 potency, we tested for selectivity amongst kinases at a
concentration of 1 µM using the DiscoverX platform. In this case, we measured high inhibitory
activity against p38α and β (Table S4). However, additional kinases were inhibited at this
concentration, giving a selectivity score S(1) of 0.037, similar to p38α chemical probe
SB203580 (0.024 at 100 nM). Due to available crystal structure data of related compounds
bound to p38α and the wealth of data available on this inhibitor class, kinase selectivity may be
further improved and potentially even abolished through rational design.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Cellular Evaluation of Lead Compounds
Figure 5: Cellular evaluation of lead compounds confirms target engagement. A. Cellular
thermal shift assay with compound V indicates thermal stabilization of BRD4 and p38α in A549
cells, with βꢀactin used as a loading control. B. Western blots of expressed cꢀMyc levels in
MM.1S cells, with GAPDH used as a loading control.
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Given the BET D1 selectivity exhibited by compounds IV and V, we proceeded by testing
these molecules in cells. First, we verified target engagement in the native chemical environment
by using a CEllular Thermal Shift Assay (CETSA) in A549 cells, where thermal stability of the
3
7
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
target protein is altered upon binding to small molecules. Here, an isothermal denaturation
experiment indicated compound V stabilized BRD4 and p38α in a doseꢀdependent manner at
concentrations above 3 µM and 1 nM, respectively. Protein melt temperatures correlated well
38,39
with previous CETSA experiments performed on BRD4 and p38α.
Importantly, as this
experiment does not measure cellular outcomes downstream of target inhibition, these results
confirm unambiguously that compound V enters cells and maintains interactions with BRD4 and
p38α in a cellular context (Fig.5A).
After verifying protein target engagement, we proceeded to test the efficacy of our D1ꢀ
selective inhibitors in MM.1S cells by measuring reduction of cꢀMyc, an oncogene highly
sensitive to BET inhibition, in several leukemia cell lines. Similar to (+)ꢀJQ1 used as a positive
control, compounds IV and V inhibited expression of the cꢀMyc oncoprotein in MM.1S cells at
concentrations approaching their respective BRD4(1) K values, as determined by Western blot
d
(Fig.4B). These results also mirrored cell viability data in the cell line studied (Fig.S14).
Together, these are characteristic hallmarks of BET inhibitors, and suggest effective cellular
bromodomain inhibitory activity. To verify inhibition of the p38α kinase, we determined the
phosphorylation states of its downstream phosphorylation target, MSK1. Here we observed a
decrease in levels of MSKꢀ1 phosphorylation at T581 upon treatment with compounds IV and V,
although at levels significantly higher than their affinity for p38α. This effect was not observed
with (+)ꢀJQ1 treatment (Fig.S16). Together, these results are consistent with BETꢀD1 inhibition
ACS Paragon Plus Environment
20

Page 21 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
using our 1,4,5ꢀtrisubstiuted imidazole scaffold being sufficient to regulate transcription in cells
in a BRD4ꢀspecific manner.
Inhibition of Inflammation
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The role BRD4 plays in coupling NFꢀκB signaling to gene transcription via K recognition
ac
has been disputed. In certain systems, transcriptional activation of RelA target genes has been
4
0
observed independent of BRD4, whereas Chen and coworkers have reported on BRD4
maintaining activity of NFꢀκB signaling via recognition of K310 acetylation at the RelA
18
subunit. In the case of NFꢀκB dependent activity, it remains unclear whether D1 or D2 plays
the dominant role in driving transcription; the two bromodomains of BRD4 maintain interactions
7
with chromatin and the presence of both domains enhances transcription. In contrast, while Zou
and coꢀworkers showed a 10ꢀfold stronger interaction of RelA with D2 by a fluorescenceꢀ
1
8
41
competition assay, Jung et al. did not detect any binding interactions in a TRꢀFRET assay,
leading to uncertainty of a direct interaction between RelA and BRD4. Together, these data
highlight a need for confirmation of the relevant interactions of each bromodomain of BRD4.
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6: Inhibition of NFꢀκB signaling and a NFꢀκB target gene. A. NFꢀκBꢀluciferase reporter
assay with A549 cells. (+)ꢀJQ1 = 20, 5, 1 µM; IV, V and RVXꢀ208 = 10, 5, 1 µM. B. Sandwich
ILꢀ8 ELISA (+)ꢀJQ1, V, SBꢀ203580 and RVXꢀ208 = 10, 1, 0.1 µM. Data reported as Mean ±
SEM values (n = 3 biological replicates) normalized to no compoundꢀinduced (I) control values.
All samples except NI control were induced with 10 ng/mL TNFα. See figure S3 for compound
mediated cellular cytotoxicity data. (ANOVA comparison to the three RVXꢀ208 treatments, ns =
not statistically significant; *** = p < 0.001 and **** = p < 0.0001)
To further assess cellular efficacy of our D1ꢀselective inhibitors, we employed a luminescenceꢀ
based NFꢀκB reporter assay in A549ꢀNFꢀκBꢀluc cells. Compound V (51% inhibition at 10 µM)
produced the most effective knockdown of TNFαꢀinduced transcription in this assay, compared
ACS Paragon Plus Environment
22

Page 23 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
to inhibition of BRD4(2) using RVXꢀ208 (25% inhibition at 10 µM) (Fig.6A). As other p38
MAPK inhibitors SBꢀ203580 and VXꢀ702 showed limited inhibitory effects in this assay
(Fig.S17), and compounds IV and V exhibited minimal cytotoxicity towards A549 cells up to
concentrations of 25 µM and 10 µM, respectively (Fig.S14), these experiments suggest BRD4(1)
inhibition may be crucial to NFꢀκB signaling in A549 cells and support earlier results using BET
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
1
selective inhibitor MSꢀ611.
Inhibition of BET bromodomains has been shown to decrease production of cytokines;
however, the role of individual BET bromodomains here is also yet to be deconvoluted. To
assess the effect of compound V on inflammation at the protein level, we quantitated the
suppression of cytokine production using an immunosorbent assay (ELISA). We chose to
measure the concentrations of secreted cytokine ILꢀ8, since its role as an inflammatory stimulus
42
for neutrophil recruitment has been well characterized. At concentrations of 1 µM in this assay,
BRD4(1) inhibition using compound V reduced levels of cytokine production (62% inhibition at
1
µM) more effectively than BRD4(2) inhibition by RVXꢀ208 (40% inhibition at 1 µM).
Moreover, the effects of compound V were comparable to panꢀBET inhibition by (+)ꢀJQ1 (63%
inhibition at 1 µM) (Fig.6B). This increased potency of V is likely due to the additional
inhibition of p38α. However, other offꢀtargets cannot be ruled out at this point. Interestingly,
complete inhibition of cytokine production was not observed with any of these molecules,
whereas an irreversible NFꢀκB inhibitor, parthenolide, decreased ILꢀ8 production to baseline at
43
2
0 µM. These results, coupled with confirmation of target occupancy by CETSA, establish a
potential mode of synergy via dual targeting of p38α and BET bromodomains.
Discussion
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Through a late stage structureꢀactivity relationship (SAR) investigation of 1,4,5ꢀtrisubstituted
imidazoles, we identified a small series of BET inhibitors. Biophysical characterization of lead
compounds against the entire BET family revealed an unexpected but high selectivity towards
BET D1s over D2s, with highest affinity for BRD4(1). Our experimental data ruled out
selectivity towards Nꢀterminal BET bromodomains due to steric hindrance by a histidine present
in Cꢀterminal BET bromodomains, where the corresponding position is occupied by aspartic acid
in BET D1s (D144 in the case of BRD4). Although H437 appears within 3Å of the piperidyl
group in overlays of BRD4(2) crystal structures (eg. PDB ID: 2YEM), this amino acid adopts
alternate conformations that may accommodate the ligand (eg. PDB ID: 5UEU).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Rather, our data with compounds VII to IX indicate selectivity was largely due to the
fluorophenyl moiety and the imidazole ring itself. Structural analysis indicates a similar binding
mode of suitably substituted imidazole rings to that of acetyl groups on acetylated histones, with
hydrogen bonding to a conserved asparagine residue (N140 in the case of BRD4(1)) by the
imidazole ring nitrogen, although N140 and K141 are slightly displaced to accommodate the
imidazole ring. Additionally, the fluorophenyl group sits much deeper in the binding pocket than
previously reported BET inhibitors and is able to displace three water molecules present in the
coꢀcrystal structure of BRD4(1) with DMSO, in which one of these is a structurally conserved
water. Our data and previous computational studies on the stability of bromodomain water
3
3,34
networks
lead us to believe our observed selectivity profile also arises in part from the
displacement of a conserved water from BRD4(1) and BRDT(1), but not other BET
bromodomains.
Cellular data indicates our compounds are highly cell permeable and capable of engaging
BRD4 in various cell lines. However, whereas effects on the MM.1S cell line appear to be
ACS Paragon Plus Environment
24

Page 25 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
largely BRD4ꢀdriven due to the Myc sensitivity of this cell line, the A549 lung cancer cell line
appears to be increasingly sensitive to both p38α and BRD4 inhibition, based on decreases in
cytokine secretion upon treatment with our molecules. While such cooperative effects can only
be cautiously assigned due to additional offꢀtarget kinase engagement, our preliminary
evaluations suggest selective inhibition of BRD4(1) may be more efficacious than BRD4(2)
inhibition. These cellular data confirm previous chromatin binding studies and work targeting
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
4,45
inhibition of BET bromodomains in inflammation,
however bromodomain inhibition alone
failed to completely inhibit signaling pathways in the system studied here.
Although kinase activity was secondary to our aim of developing selective BET inhibitors, we
endeavored to fully characterize the kinase activity of our lead molecules. Previous reports on
substituted imidazoles and pyrazoles of the ‘DFGꢀout’ class of hingeꢀbinding kinase inhibitors
suggest modest additions to the periphery of the scaffold, such as through Nꢀalkylation, may
4
6,47
improve selectivity.
Additionally, by replacing the pꢀfluorophenyl moiety from similar
trisubstituted imidazoles with ethylꢀ or chloroꢀ groups, Gallagher et al. observed a loss of p38α
48
kinase inhibition. We thus envision the selectivity profile of these molecules may be readily
improved by further analog synthesis and SAR studies.
Conclusion
Through a systematic investigation of a recently characterized 1,4,5ꢀsubstituted imidazole
scaffold, and newly reported 1,4,5ꢀsubstituted triazole scaffold, we have described a new BET
inhibitor class with high selectivity for D1 over D2 for BRD4 and BRDT. From a therapeutic
standpoint, this class of 1,4,5ꢀtrisubstituted imidazoles and triazoles may be a valuable starting
point for dual kinase/bromodomain inhibition, as simultaneous BRD4 and MAP/PI3 kinase
inhibition has recently been suggested for synergistic treatment of colon cancer, avoiding the
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
28,49
development of resistance to BET or kinase inhibitor monotherapy,
and resulting in
50
decreased toxicity relative to a cocktail of (+)ꢀJQ1 and PI3K inhibitors. As demonstrated here,
the optimization of further inhibitors of individual BET D1s may enable selective control of
BRD4ꢀassociated inflammatory signaling. Finally, the displacement of structured waters may
also be a general strategy that can be applied to other known panꢀBET inhibitors which we are
actively investigating.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Experimental Section
General Synthetic Methods: All chemical reagents were purchased from commercial vendors
and used without further purification. Flash chromatography of compounds was performed on a
1
TeledyneꢀIsco Rfꢀplus CombiFlash instrument across RediSep Rf columns. H (400 MHz) and
13
C (125 MHz) NMR spectra were recorded in Chloroformꢀd or Methanolꢀd and chemical shifts
4
(
δ) are reported in ppm. All coupling constants (J) are reported in hertz (Hz), where s = singlet, d
= doublet, t = triplet, and m = multiplet. Purity of final compounds diluted in DMSO was
determined by reverseꢀphase high performance liquid chromatography (HPLC) analysis using an
Agilent 1200 Series HPLC (detector: 215 and 245 nm DAD; column: Agilent Zorbax SBC18,
4
.6 × 150 mm, 5.0 ꢂm; mobile phase gradient: 90:10 to 15:85 0.1% TFA in H O: Acetonitrile
2
over 22 minutes; flow rate, 1.0 mL/min), where all compounds used were verified to be >95%
purity by peak area. Highꢀresolution mass spectrometry was performed using positive mode
electrospray ionization methods (ESMS) with a Bruker BioTOF II spectrometer.
Synthesis of Common Intermediate 12:
Synthesis of 4ꢀmethylbenzenesulfinic acid (7): To a suspended solution of pꢀtoluenesulfinic
acid sodium salt 6 (3.0 g, 16 mmol) in water (10.0 mL) was added methyl tertꢀbutylether (5.0
mL) followed by dropꢀwise addition of conc. HCl (1.5 mL). After stirring the reaction for 5 min.,
ACS Paragon Plus Environment
26

Page 27 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the organic phase was separated, and the aqueous phase was extracted with methyl tertꢀ
butylether (2 x 25 mL). The combined organic phase was dried over anhydrous magnesium
sulfate and concentrated under reduced pressure. The solid obtained was reꢀsuspended in
hexanes (50 mL) and was filtered and dried to give the desired compound 2 as a white solid (2.4
g, 91%).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis of Nꢀ((4ꢀfluorophenyl)(tosyl)methyl)formamide (8): A mixture of 4ꢀ
methylbenzenesulfinic acid 7 (0.50 g, 3.2 mmol), 4ꢀfluorobenzaldehyde (0.58 g, 4.7 mmol),
°
formamide (0.52 g, 12 mmol) and camphorsulfonic acid (0.09 g, 0.4 mmol) was stirred at 60 C
for 18 h. The resulting solid was stirred with a mixture of methanol (0.8 mL) and hexanes (1.9
mL). The solid was filtered and was reꢀsuspended in methanol/hexanes (1:3, 4.5 mL) and was
stirred to obtain a fine suspension. The suspension was filtered and dried to give the pure
1
compound 8 as a white solid (0.64 g, 65%). H NMR (500 MHz, CDCl ) δ: 8.06 (s, 1H), 7.69 (d,
3
2H, J = 8.3 Hz), 7.43ꢀ7.4 (m, 2H), 7.31 (d, 2H, J = 7.8 Hz), 7.06 (t, 2H, J = 8.3 Hz), 6.29 (s, 1H),
2
.43 (s, 3H).
Synthesis of 1ꢀfluoroꢀ4ꢀ(isocyano(tosyl)methyl)benzene (9): To a solution of Nꢀ((4ꢀ
fluorophenyl)(tosyl)methyl)formamide 8 (0.20 g, 0.65 mmol) in 1,2ꢀdimethoxyethane (3.5 mL)
°
at ꢀ10 C was added POCl (0.15 mL, 1.6 mmol) followed by Et N (0.46 mL, 3.3 mmol) keeping
3
3
the internal temperature below ꢀ5 °C. The reaction was warmed to room temperature and was
stirred for 2 h. The reaction mixture was quenched by addition of iceꢀcold water and the aqueous
phase was extracted with methyl tertꢀbutylether (2 x 15 mL). The combined organic phase was
washed with saturated aqueous NaHCO (1 x 10 mL), brine (1 x 10 mL) and was concentrated
3
under reduced pressure. The residue was triturated with hexanes to give the desired product as a
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 79
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
brown solid (0.11 g, 55%). H NMR (500 MHz, CDCl ) δ: 7.65 (d, 2H, J = 8.0 Hz), 7.38ꢀ7.37
3
(
m, 4H), 7.12 (t, 2H, J = 8.3 Hz), 5.61 (s, 1H), 2.5 (s, 3H).
Synthesis of 4ꢀ(dimethoxymethyl)ꢀ2ꢀ(methylthio)pyrimidine (13): A mixture of pyruvic
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
aldehyde dimethylacetal 10 (6.0 mL, 50 mmol) and N,Nꢀdimethylformamide dimethylacetal (6.0
mL, 45 mmol) was heated at 100 °C for 18 h. After cooling the mixture to room temperature,
methanol (30.0 mL), thiourea (6.96 g, 91.4 mmol) and sodium methoxide (25% in MeOH, 23.1
mL, 106 mmol) were added. The reaction mixture was then stirred at 70 °C for 2 h. The reaction
mixture was then cooled, iodomethane (14.4 mL, 231 mmol) was added dropꢀwise, and the
reaction was stirred at room temperature for 3 h. The reaction mixture was then diluted with
ethyl acetate (200 mL) and water (100 mL). The organic phase was separated, and the aqueous
phase was extracted with ethyl acetate (2 x 100 mL). The combined organic phase was dried
over anhydrous magnesium sulfate and concentrated under reduced pressure to give the
1
corresponding product 13 as a brown oil (7.5 g, 75%). H NMR (400 MHz, CDCl ) δ: 8.56 (d,
3
1
H, J = 5.1 Hz), 7.19 (d, 1H, J = 4.9 Hz), 5.19 (s, 1H), 3.42 (s, 6H), 2.58 (s, 3H).
Synthesis of 2ꢀ(methylthio)pyrimidineꢀ4ꢀcarbaldehyde (14): solution of 4ꢀ
dimethoxymethyl)ꢀ2ꢀ(methylthio)pyrimidine 13 (2.0 g, 10 mmol) in 3N HCl (8.4 mL, 25 mmol)
A
(
was stirred at 48 °C for 16 h. The reaction was cooled and quenched by addition of solid
Na CO . The residue was washed with ethyl acetate (3 x 50 mL). The combined organic phase
2
3
was dried and concentrated to give the crude product which was purified by filtration over a pad
of silica gel eluting with dichloromethane to give the pure product as a pale yellow solid (0.6 g,
1
4
0%). H NMR (400 MHz, CDCl ) δ: 9.96 (d, 1H, 0.6 Hz), 8.77 (dd, 1H, J = 4.8, 0.5 Hz), 7.45
3
(d, 1H, J = 4.8 Hz), 2.64 (s, 3H).
ACS Paragon Plus Environment
28

Page 29 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Synthesis of tertꢀbutyl 4ꢀ(((2ꢀ(methylthio)pyrimidinꢀ4ꢀyl)methylene)amino)piperidineꢀ1ꢀ
carboxylate (15): To a solution of 2ꢀ(methylthio)pyrimidineꢀ4ꢀcarbaldehyde 9 (0.20 g, 1.3
mmol) and tertꢀbutyl 4ꢀaminopiperidineꢀ1ꢀcarboxylate (0.26 g, 1.3 mmol) in dichloromethane (3
mL) was added anhydrous magnesium sulfate (0.31 g, 2.6 mmol) and the reaction was stirred at
room temperature for 24 h. After confirmation of the completion of reaction by TLC, the
reaction mixture was filtered, and the solvent was evaporated to give the desired product which
was taken directly to the next step.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis of tertꢀbutyl 4ꢀ(4ꢀ(4ꢀfluorophenyl)ꢀ5ꢀ(2ꢀ(methylthio)pyrimidinꢀ4ꢀyl)ꢀ1Hꢀ
imidazolꢀ1ꢀyl)piperidineꢀ1ꢀcarboxylate
(16):
A
mixture
of
tertꢀbutyl
4ꢀ(((2ꢀ
(methylthio)pyrimidinꢀ4ꢀyl)methylene)amino)piperidineꢀ1ꢀcarboxylate 15 (0.44 g, 1.3 mmol), 1ꢀ
fluoroꢀ4ꢀ(isocyano(tosyl)methyl)benzene 4 (0.38 g, 1.3 mmol) and potassium carbonate (0.18 g,
1
.3 mmol) in DMF at room temperature was stirred for 48 h. The reaction mixture was quenched
by the addition of water and was extracted with ethyl acetate (3 x 25 mL). The combined organic
phase was washed with 5% aqueous LiCl and concentrated under reduced pressure to give the
crude product which was purified by column chromatography over silica gel (60ꢀ120 mesh)
using ethyl acetate and hexanes as eluent (0ꢀ60%) to give the desired product as a yellow solid
1
(
0.17 g, 28%). H NMR (400 MHz, CDCl ) δ: 8.35 (d, 1H, J = 5.2 Hz), 7.76 (s, 1H), 7.45ꢀ7.41
3
(m, 2H), 7.03 (t, 2H, J = 8.6 Hz), 6.79 (d, 1H, J = 5.2 Hz), 4.85 (tt, 1H, J = 12.0, 3.6 Hz), 4.31
(br s, 2H), 2.84ꢀ2.78 (m, 2H), 2.6 (s, 3H), 2.19ꢀ2.16 (m, 2H), 1.86 (dq, 2H, J = 12.3, 4.1 Hz),
1
.49 (s, 9H).
Synthesis of tertꢀbutyl 4ꢀ(4ꢀ(4ꢀfluorophenyl)ꢀ5ꢀ(2ꢀ(methylsulfonyl)pyrimidinꢀ4ꢀyl)ꢀ1Hꢀ
imidazolꢀ1ꢀyl)piperidineꢀ1ꢀcarboxylate (17): A solution of OXONE (160 mg, 0.25 mmol) in
water (1.4 mL) was added dropꢀwise to a solution of tertꢀbutyl 4ꢀ(4ꢀ(4ꢀfluorophenyl)ꢀ5ꢀ(2ꢀ
ACS Paragon Plus Environment
29