Discovery of a Small-Molecule Bromodomain-Containing Protein 4 (BRD4) Inhibitor That Induces AMP-Activated Protein Kinase-Modulated Autophagy-Associated Cell Death in Breast Cancer

Article abstract of DOI:10.1021/acs.jmedchem.7b00275

Upon the basis of The Cancer Genome Atlas (TCGA) data set, we identified that several autophagy-related proteins such as AMP-activated protein kinase (AMPK) were remarkably downregulated in breast cancer. Combined with coimmunoprecipitation assay, we demonstrated that BRD4 might interact with AMPK. After analyses of the pharmacophore and WPF interaction optimization, we designed a small-molecule inhibitor of BRD4, 9f (FL-411) which was validated by cocrystal structure with BD1 of BRD4. Subsequently, 9f was discovered to induce ATG5-dependent autophagy-associated cell death (ACD) by blocking BRD4-AMPK interaction and thus activating AMPK-mTOR-ULK1-modulated autophagic pathway in breast cancer cells. Interestingly, the iTRAQ-based proteomics analyses revealed that 9f induced ACD pathways involved in HMGB1, VDAC1/2, and eEF2. Moreover, 9f displayed a therapeutic potential on both breast cancer xenograft mouse and zebrafish models. Together, these results demonstrate that a novel small-molecule inhibitor of BRD4 induces BRD4-AMPK-modulated ACD in breast cancer, which may provide a candidate drug for future cancer therapy.

Full text of DOI:10.1021/acs.jmedchem.7b00275

Subscriber access provided by READING UNIV
Article
Discovery of a small-molecule bromodomain-containing protein
(BRD4) inhibitor that induces AMP-activated protein kinase-
4
modulated autophagy-associated cell death in breast cancer
Liang Ouyang, Lan Zhang, Jie Liu, Leilei Fu, Dahong Yao,
Yuqian Zhao, Shouyue Zhang, Guan Wang, Gu He, and Bo Liu
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00275 • Publication Date (Web): 27 Nov 2017
Downloaded from http://pubs.acs.org on November 29, 2017
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 free 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 accessible to all
readers and 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.
Journal of Medicinal Chemistry 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 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of a small-molecule bromodomain-containing
protein 4 (BRD4) inhibitor that induces AMP-activated
protein kinase-modulated autophagy-associated cell
death in breast cancer
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
#
#
Liang Ouyang , Lan Zhang , Jie Liu, Leilei Fu, Dahong Yao, Yuqian Zhao, Shouyue
*
*
Zhang, Guan Wang, Gu He , Bo Liu
State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, Sichuan
University, and Collaborative Innovation Center of Biotherapy, Chengdu 610041, China
ABSTRACT
Based upon The Cancer Genome Atlas (TCGA) dataset, we identified that several
autophagyꢀrelated proteins such as AMPꢀactivated protein kinase (AMPK) was
remarkably downregulated in breast cancer. Combined with coꢀimmunoprecipitation
assay, we demonstrated that BRD4 might interact with AMPK. After analyses of the
pharmacophore and WPF interaction optimization, we designed a smallꢀmolecule inhibitor
of BRD4 9f (FLꢀ411) which was validated by coꢀcrystal structure with BD1 of BRD4.
Subsequently, 9f was discovered to induce ATG5ꢀdependent autophagyꢀassociated cell
death (ACD) by blocking BRD4ꢀAMPK interaction, and thus activating
AMPKꢀmTORꢀULK1ꢀmodulated autophagic pathway in breast cancer cells. Interestingly,
the iTRAQꢀbased proteomics analyses revealed that 9f induced ACD pathways, involved
in HMGB1, VDAC1/2 and eEF2. Moreover, 9f displayed a therapeutic potential on both
breast cancer xenograft mouse and zebrafish models. Together, these results
demonstrate
that
a
novel smallꢀmolecule
inhibitor
of
BRD4
induces
BRD4ꢀAMPKꢀmodulated ACD in breast cancer, which may provide a candidate drug for
future cancer therapy.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 72
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
Key words: Bromodomainꢀcontaining protein 4 (BRD4); BRD4ꢀAMPK interaction;
Smallꢀmolecule inhibitor of BRD4; Autophagyꢀassociated cell death (ACD); Breast cancer
INTRODUCTION
Bromodomainꢀcontaining protein 4 (BRD4) is a member of the bromodomain and
extraꢀterminal (BET) protein family, functioning as a central element in transmission of
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
ꢀ2
epigenetic memory across cell division and transcription regulation. Besides BRD4,
there are other three members of BET family, such as BRD2, BRD3 and BRDT, which
3,4
may regulate many cellular processes by the interactions of bromodomain. BRD4 is a
ubiquitously expressed nuclear protein of 200 kDa that contains two tandem
bromodomains (BD1 and BD2) which are structurally conserved but not functionally
5
ꢀ7
equivalent, and an extraꢀterminal domain.
BRD4 may participate in oncogenic
rearrangements, leading to highly oncogenic fusion proteins, and thus playing a key role
8
ꢀ10
in development of many types of cancer, including breast cancer.
Recently, BRD4 has
been reported to regulate breast cancer cell metastasis through modulating the enzymatic
11,12
activity of Sipa1 in ER+ breast cancer.
In addition, BRD4 is reported to modulate
extracellular matrix gene expression, frequently presented in metastasisꢀpredictive gene
signature; thereby robustly predicting progression and survival through its activation
13
signature in breast cancer. Therefore, targeting BRD4 inhibition would be a potential
strategy in breast cancer therapy.
Of note, autophagy is an evolutionarily conserved, multiꢀstep lysosomal degradation
1
4ꢀ17
process for the clearance of damaged or superfluous proteins and organelles.
In the
autophagy process, mammalian target of rapamycin (mTOR) and AMPꢀactivated protein
kinase (AMPK) are able to oppositely regulate UNCꢀ51ꢀlike kinase 1 (ULK1) activity by
1
8
direct phosphorylation. AMPKꢀmTORꢀULK1 axis may play a key role in autophagosome
19
formation and autophagy regulation. The function of AMPK, mTOR and ULK1 and their
20ꢀ22
associations with breast cancer, has been gradually elucidated.
More recently,
23
compound 1 (JQ1), as the famous and first inhibitor of BET, has been reported to induce
the AMPKꢀULK1ꢀmodulated cytoprotective autophagy in acute myeloid leukemia (AML)
2
4
stem cells. On the contrary, 1 has been also reported to inhibit cytoprotective autophagy
ACS Paragon Plus Environment

Page 3 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
5
via increasing NPM1 and HEXIM1 expression in AML cells.
For better elucidation of the role of BRD4 inhibition in autophagy and discovery of a
novel BRD4 inhibitor in breast cancer, we firstly identified the potential BRD4ꢀAMPK axis
in autophagy by the combination of TCGA analyses and coꢀ immunoprecipitation assays.
Based on the structureꢀbased design, we finally discovered a new smallꢀmolecule inhibitor
of BRD4 that induced BRD4ꢀAMPKꢀmodulated autophagy/autophagic cell death of breast
cancer in vitro and in vivo.
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
RESULTS
Identification of BRD4-AMPK interaction in autophagy to breast cancer therapy
We constructed the proteinꢀprotein interactions (PPI) network of BRD4 by the
predicted and experimentally determined proteinꢀprotein interactions (PrePPI), in which
3
82 hub proteins were identified to potentially interact with BRD4. Based upon the PPI
network, BRD4 was shown to interact with 28 autophagyꢀrelated genes/proteins, which
are annotated by Gene Ontology (GO), indicating that BRD4, as an attractive cancer
target is related to some autophagic genes/proteins by regulation of the autophagic
process (Figure 1A). Subsequently, to attest the impact of autophagy in breast cancer, the
expression differences of autophagic genes were identified between normal and cancer
tissues in breast cancer dataset from The Cancer Genome Atlas (TCGA). Some key
autophagic genes, which were involved in the PPI network of BRD4, such as MAP1LC3A
(p=1.10eꢀ4), MAP1LC3B (p=2.02eꢀ11), PRKAA1 (p=2.06eꢀ11) and ATG4C (p=1.46eꢀ8)
were identified to have lower expression in tumor tissues than normal ones (Figure 1B).
Amongst them, we demonstrated that AMPK (PRKAA1), which is a key autophagic
regulator triggering AMPKꢀmTORꢀULK1 pathway, might interact with BRD4 by reciprocal
coꢀimmunoprecipitation (Figure 1C). Interestingly, we found that AMPK did not interact
with cꢀMyc, a targeted transcription factor regulated by BRD4 (Figure S1), indicating there
may be a direct interaction between BRD4 and AMPK. Recently, decreased expression of
some autophagyꢀrelated genes, such as ULK1 and Beclinꢀ1 has been revealed to be
2
6,27
closely associated with poor prognosis in breast cancer.
It is wellꢀknown that BRD4,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 72
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
as an attractive cancer target, can be combined with PI3K inhibition to induce cell death
and overcomes resistance in breast cancer, indicating that combination of BRD4 and
28
other kinase may be also a therapeutic avenue in breast cancer. Interestingly, BRD4
has been demonstrated as a negative regulator of autophagy gene expression, and its
dissociation from autophagy gene promoters can lead to transcriptional activation of
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
29
autophagy genes. Thus, induction of ACD is a possible strategy to sensitize breast
cancer cells to therapy or kill them. As mentioned above, we identified the BRD4ꢀAMPK
interaction, which might play a key autophagyꢀmodulating role in breast cancer. Thus, we
put forward a hypothesis to design a smallꢀmolecule inhibitor of BRD4 that induces
BRD4ꢀAMPKꢀmodulated ACD of breast cancer in vitro and in vivo (Figure 1D).
Figure.1 Identification of the BRD4ꢀAMPK interaction in autophagy to breast cancer therapy. (A) The
BRD4 related proteinꢀprotein network and autophagyꢀrelated enrichment analyses. (B) Comparison of
autophagy related genes expression level between normal and cancer in TCGA breast cancer dataset.
(
(
C) Cell extracts of MCFꢀ7 or MDAꢀMBꢀ231 cells were subjected to immunoprecipitation with BRD4
upper, left panel) and AMPKα (lower, left panel) antibodies, then detected by western blot using AMPKα
and BRD4 antibodies. The expressions of BRD4 and AMPKα were determined in the input. CꢀMyc and
p53 were immunoprecipitated with BRD4 as the positive control. (D) Model of BRD4 inhibitorꢀmodulating
autophagy, via inhibiting BRD4 and activating AMPKꢀmTORꢀULK complex axis.
ACS Paragon Plus Environment

Page 5 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Design and synthesis of candidate BRD4 inhibitors
We performed twoꢀstep molecular docking on the candidate smallꢀmolecule
compounds from ZINC database, which is a free public resource for ligand discovery that
contains over twenty million commercially available molecules. We utilized the crystal
structure of BRD4(1) from the RSCB (PDB ID: 3MXF) and defined the binding site
referred to that of 1. Other parameters were set as default values. In the first step, the top
500 hits were selected by LibDock protocol mainly according to the LibDock Score.
Subsequently, the top 20 hits determining by CDOCKER protocol were selected by the
interaction energy (Figure S2). And ZINC96907903 was shown to exhibit the best score
and binding mode, suggesting this compound may have the priority as a candidate
inhibitor of BRD4.
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
To obtain more potent analogues, we designed various substituents on the phenyl
group to explore the chemical space for affinity improvement (Figure 2A). We
preferentially determined the mꢀmethoxyphenyl, mꢀbromophenyl, pꢀhydroxyphenyl when
preserving the nꢀpropyl on thiophene (Scheme 1), revealing that 4c, with a
pꢀhydroxyphenyl, showing the most potent inhibitory activity (IC50 = 8.59 ꢁM). Whereas
other compounds (4a, 4b, 4d) showed a weak inhibitory activity, suggesting that a
hydrogen bond acceptor group forming a key hydrogen bond with Asn140 plays an
important role for inhibitory potency. Subsequently, a cyclopentyl was incorporated into
the thiophene and 7 compounds with different substituted phenyls were synthesized
(
Scheme 2). The AlphaScreen assay indicated that most of these compounds showed a
certain extent improvement in potency. Especially 5a, 5b and 5f have 1 to 10 fold
improvement compared with 4c. We speculated that the replacement of cyclopentyl with
nꢀpropyl might have a favorable contribution to compound activity. To further confirm SAR,
on one hand, some greater groups in size were induced into the thiophene, including
cyclohexyl,
cycloheptyl,
4ꢀmethylcyclohexyl,
nitromethylpiperidine
and
nitrobenzylpiperidine; on the other hand, various substituents on the phenyl group were
preserved (Scheme 2 and 3). The AlphaScreen assay presented that all compounds
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 72
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
(
except for 10b) containing 4ꢀhydroxyꢀ3,5ꢀdimethylbenyl group had potent inhibitory
activity against BRD4, and the change of substituents on thiophene from nꢀpropyl to
nitromethylpiperidine, led to a remarkable potency improvement, such as 5f (IC50 = 6.38
ꢁ
M), 8h (IC50 = 3.23 ꢁM) and 9f (FLꢀ411) (IC50 = 0.43 ꢁM) (Table 1). However, further
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
extension of the substituent decreased the activity, such as 10b (IC50 > 10 ꢁM). We
demonstrated that 4ꢀhydroxyꢀ3,5ꢀdimethylbenyl was a favorable group to fit the
acetylꢀlysine binding site, the OH group as a hydrogen bond accepter could interact with
Asn140 to form a conserved hydrogen bond, and the ortho methyl of 9f could function as
the methyl of acetylated lysine to occupy the small hydrophobic pocket. Thus, the
4
ꢀhydroxyꢀ3,5ꢀdimethylbenyl seems to be an optimal acetyl mimic fragment. The
substituents on thiophene made hydrophobic interactions or hydrogen bonds with
residues located in the ZA loop, especially the WPF shelf (Trp81, Pro82 and Phe83),
which is key to the affinity and selectivity. Based upon the multiple biological evaluation
screening, we eventually discovered that 9f possessed the best BRD4(1) inhibition activity
(
IC50=0.43 ꢁM), antiꢀproliferative activity (MCFꢀ7: IC50=1.62 ꢁM, MDAꢀMBꢀ231: IC50=3.27
M) and autophagic activity (42.29% in MCFꢀ7 cells), as well as displayed a low toxicity
against MCF10A cells (Figure 2B and C, Table 1, Figure S2). In addition, 9f had a better
ꢁ
3
0
antiꢀproliferative activity than 1 and 2a (RVXꢀ208) although it showed lower BRD4(1)
inhibition activity than 1 (Table 1).
O
O
O
NH
a
b
CHO
R
NH2
N
S
S
3a
4a-d
Scheme 1. Reagents and conditions: (a) NCCH
2
CO
2
Et, S
8
, EtOH, Et N, reflux, 12 h; (b) nitrile,
3
dioxane/HCl, 100 ℃, 12 h.
Scheme 2. Reagents and conditions: (a) NCCH
2
CO
2
Et, S
8 3
, EtOH, Et N, reflux, 12 h; (b) nitrile,
ACS Paragon Plus Environment

Page 7 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
dioxane/HCl, 100 ℃, 12 h.
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 3. Reagents and conditions: (a) NCCH
2
CO
2
Et, S
8
, EtOH, Et
3
N, reflux, 12 h; (b) nitrile,
12 h.
dioxane/HCl, 100
℃
,
Figure.2 Design and synthesis of candidate BRD4 inhibitors. (A) Structure guided optimization towards
9f. (B) Heat map of mean IC50 values of antiꢀproliferative activity and BRD4(1) inhibition activity. Cell
viabilities were measured for each compound with various concentrations by MTT assay; BRD4(1)
inhibition activity was measured by AlphaScreen assay; The IC50 values were calculated by Prism 6.0. (C)
MDC fluorescence staining analysis (Flow cytometry analysis) was performed on candidate compounds,
all compounds were tested at 5 µM; 9f was selected as the best MDC positive ratio. (D) BRD4ꢀ9f
interactions include hydrogen bonds (green dash), PiꢀPi (pink dash) and hydrophobic interaction (orange
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 72
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
solid).
Table 1. Biochemical and biological evaluations of candidate compounds
BRD4(1)
Antiꢀproliferative activity
MDC positive
b
No.
Structure
AlphaScreen
(IC50, ꢁM)
c
ratio% (5 ꢁM)
a
(
IC50, ꢁM)
>10.00
MCFꢀ7
MDAꢀMBꢀ231
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
4a
>30.00
>30.00
9.58±1.02
4b
>10.00
>30.00
>30.00
>30.00
6.35±0.59
4c
8.59±0.32
28.35±1.66
18.77±0.78
4d
>10.00
>30.00
>30.00
5.90±0.68
5a
0.78±0.04
1.53±0.08
>10.00
2.22±0.82
8.39±1.34
>30.00
5.34±0.37
14.69±0.72
>30.00
26.66±0.92
25.21±0.41
11.79±0.95
5b
5c
ACS Paragon Plus Environment

Page 9 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5d
>10.00
>30.00
>30.00
9.05±0.43
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
5e
8.94±0.11
6.38±0.22
21.29±1.95
18.47±1.27
>30.00
20.31±0.87
26.98±1.32
5f
23.12±1.81
5g
>10.00
>10.00
>30.00
28.11±1.28
>30.00
>30.00
>30.00
>30.00
>30.00
>30.00
10.43±0.74
13.11±0.81
11.70±0.49
7.51±0.32
18.81±0.99
6a
6b
>10.00
6c
>10.00
>30.00
6d
7.32±0.37
22.75±2.02
6e
6.93±0.28
24.91±1.85
27.76±2.11
25.62±1.05
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 72
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
6
f
>10.00
>30.00
>30.00
14.48±0.60
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
6g
>10.00
>10.00
>30.00
>30.00
>30.00
>30.00
>30.00
12.59±0.83
10.55±0.97
19.79±1.38
7a
7b
5.73±0.12
18.52±1.01
7c
3.98±0.27
13.23±1.06
17.42±1.76
25.99±0.82
7d
0.56±0.03
>10.00
1.56±0.26
>30.00
3.29±0.29
>30.00
33.17±0.47
19.21±0.37
7e
7f
0.49±0.06
0.82±0.08
2.81±0.32
4.50±0.81
8.31±0.99
5.80±1.08
38.37±0.97
27.36±0.85
7g
ACS Paragon Plus Environment

Page 11 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7
h
>10.00
>30.00
>30.00
9.79±0.69
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
8
a
>10.00
>10.00
>30.00
>30.00
>30.00
>30.00
14.78±0.77
17.55±0.68
27.22±0.51
8b
8
c
1.02±0.09
2.69±0.37
16.26±1.11
8d
>10.00
>10.00
>30.00
>30.00
>30.00
>30.00
13.59±0.94
15.24±0.89
8
e
8f
4.31±0.22
>10.00
12.19±0.81
>30.00
22.85±1.21
>30.00
20.47±0.34
15.26±0.57
8g
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 72
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
8
h
3.24±0.23
>10.00
5.59±0.76
9.42±0.48
>30.00
30.79±0.91
8.38±0.50
O
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
NH
N
8i
>30.00
>30.00
S
8j
>10.00
>30.00
10.32±0.31
8
k
>10.00
>30.00
>30.00
9.64±0.73
9
a
2.47±0.31
>10.00
5.60±0.43
>30.00
8.19±0.63
22.21±0.88
8.31±1.08
22.47±1.49
10.78±0.71
21.26±0.45
9b
9
c
1.04±0.06
2.17±0.28
9d
>10.00
>30.00
>30.00
13.61±0.32
ACS Paragon Plus Environment

Page 13 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
9
e
0.72±0.08
0.43±0.09
5.48±0.84
8.29±0.71
3.27±0.14
33.18±0.58
42.29±0.91
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
9
f
1.62±0.06
9g
>10.00
>30.00
>30.00
10.34±0.29
10a
>10.00
>10.00
>30.00
>30.00
>30.00
>30.00
8.62±0.63
10b
12.58±0.58
10c
>10.00
>10.00
>30.00
>30.00
>30.00
>30.00
>30.00
5.87±0.48
11.30±0.39
6.74±0.56
10d
10e
6.37±0.11
17.01±1.26
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 72
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
0.073±0.006
28.75±1.97
33.04±2.66
12.30±1.23
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
2a
>10.00
140.10 ± 2.71
155.40 ± 1.89
8.70±0.95
a
b
c
Each compound was tested in triplicate; the data are presented as the mean ± SEM (n=3).
IC50 values obtained with cell viability assay for 24 h.
MDC positive ratio was detected by flow cytometry analysis using MDC staining on MCFꢀ7
cells, the data are presented as the mean ± SEM (n=3).
Identification of 9f as a selective BRD4 inhibitor
To demonstrate the direct binding of this smallꢀmolecule inhibitor, we set up a
31
coꢀcrystallization assay and elucidated the tertiary structure for 9f (PDB ID: 4ZW1) in
complex with the BD1 of BRD4 (Table S1). The 3,5ꢀdimethylphenol moiety of 9f
penetrated deeply into the Nꢀacetylated lysine pocket and occupied the same position as
the acetyl head group of the acetylated lysine present in the histone tail peptides, initiating
a direct hydrogen bond with the conserved asparagine (Asn140). One water molecule
was located at the bottom of the active pocket, as observed in the Nꢀacetylated lysine
32
peptide complex structure (PDB ID: 3UVW). The water molecule hydrogen bond
network contributed to stabilizing the ligand through additional hydrogen bonds. The
pyrimidone moiety occupied the WPF shelf site composed of the residues Trp81, Pro82,
and Phe83. The pyrimidone formed a PiꢀPi interaction with the indole side chain of Trp81
and two hydrogen bonds were observed in amide scaffold with residues Pro82, Gln85.
The carbonyl group establishes a water bridged hydrogen bond interaction with residues
Pro86 and Asp88 stabilizing the complex (Figure 3A). Subsequently, 100ns molecular
dynamics (MD) simulations were performed on BRD4(1)ꢀ9f complex. The fluctuations of
ACS Paragon Plus Environment

Page 15 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
RMSD (rootꢀmeanꢀsquare deviation) of MD trajectories were relative low, as well as the
energies, temperatures, and pressures of the systems suggested the systems were stable
and robust (Figure S4).To further characterize the selectivity of 9f to other BET family
proteins, we determined the binding affinity of 9f to BRD2(1), BRD2(2), BRD3(1),
BRD3(2), BRD4(1), BRD4(2) and BRDT(1) by using TRꢀFRET. 9f displayed a high
selectivity for BRD4 instead of other BET family proteins, with IC50 of 0.47 ꢁM and 0.93
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
ꢁ
M against BRD4(1) and BRD4(2), respectively (Figure 3B). To explain the selectivity of
9f over BRD4, we superimposed the structures of BRD2(1) (PDB ID: 1X0J), BRD3(1)
(PDB ID: 3S91), BRD4(1) (PDB ID: 4ZW1) and BRDT(1) (PDB ID: 4FLP) and performed
the molecular docking (Figure 3C and Figure S5). The results revealed that the overlay of
structures were almost consistent with the docked pose. Intriguingly, despite the highly
conservative sequences, we found that there were two potential key residues (Gln85,
Pro82) contributing to the selectivity of 9f. The Gln residue is highly conserved in BD1 of
BRD2ꢀ4, but the amide side chain of Gln residues possess different orientations which
result in the failure of the hydrogen bond formed by the oxygen atom of 9f and Gln
residues in BRD2(1) and BRD3(1). In BRDT(1), the corresponding Gln residue is replaced
by Arg residue, which may reduce the binding affinity of 9f (Figure 3C and Figure S5). In
addition, the conformation of Pro residue is also different in BD1 of BRD2ꢀ4 and BRDT. A
hydrogen bond could be observed in BRD2(1), BRD4(1) and BRDT(1), but the deviation
of Pro residue in BRD3(1) resulted in the failure of a hydrogen bond (Figure 3C and Figure
S5). Collectively, the Gln85 residue plays a key role in selectivity of 9f over BRD4. To
further illuminate the selectivity of 9f over BRD4, we performed the MD simulations on the
docked pose of 9f to monitor the RMSD of Cα atoms for the residues in 5 Å around ligand,
and it can be seen that this system is stable within 20 ns (Figure S6). Subsequently, we
calculated the binding free energy of the four systems, the results revealed that 9f showed
the strongest binding affinity against BRD4 (Table 2), which is consistent with the assay
results. Taken together, 9f is a potent and selective BRD4 inhibitor.
Moreover, we determined whether 9f had direct modulatory effect on
autophagyꢀassociated kinases such as AMPK and ULK1 by using ADPꢀGlo kinase assay.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 72
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
As expected, 9f did not activate or inhibit the kinase activities of AMPK and ULK1 in a
series of concentrations (Figure S7A and B). To demonstrate that 9f inhibits BRD4 and
blocks the interaction of BRD4ꢀAMPK, we examined the expression and interaction status
of BRD4 and AMPK. Interestingly, we found that 9f markedly inhibited the expression of
BRD4 and its downstream target cꢀMyc, while activating the phosphorylation of AMPKα at
thr172 site (Figure 4A). Further coꢀimmunoprecipitation results showed that 9f could
reduce the binding capacity between BRD4 and AMPKα (Figure 4B). Since p53 can
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
33,34
interact with BRD4 and regulated by AMPK,
we next employed
coꢀimmunoprecipitation assay to examine whether 9f disrupts the interaction between
BRD4 and AMPK via p53. Intriguingly, we found that 9f attenuated the interaction
between BRD4 and p53, whereas the interaction between p53 and AMPK was
augmented after treatment of 9f in MCFꢀ7 cells (Figure 4C). However, 9f did not disrupt
the interaction between BRD4 and p53, as well as the interaction between p53 and AMPK.
Thus, we deduce that the p53 is not a key factor for BRD4ꢀAMPK interaction. Taken
together, these results suggest that 9f, as a potent and selective BRD4 inhibitor, might
block the interaction of BRD4ꢀAMPK; thereby activating AMPKꢀmodulated autophagy in
breast cancer.
ACS Paragon Plus Environment

Page 17 of 72
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
Figure.3 Identification of 9f as a selective BRD4 inhibitor. (A) Binding model of 9f with the BRD4
(
white)/BD1 (deep blue). 9f (light yellow) in receptor residues is illustrated. (B) Binding affinities of 9f
were measured by TRꢀFRET against the first and second bromodomains of BRD2, BRD3, BRD4 and
BRDT. (C) The comparisons of the docked pose of 9f with the coꢀcrystal structures of BRD2(1), BRD3(1),
BRD4(1) and BRDT(1), the docked conformation of 9f was shown in green and the coꢀcrystalized
conformation was shown in yellow.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 72
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.4 9f activates AMPK and blocks the interaction between BRD4 and AMPK. (A) MCFꢀ7 or
MDAꢀMBꢀ231 cells were treated with 1.5 or 3 ꢁM of 9f for indicated times, respectively. Then, the
expression levels of BRD4, cꢀMyc, AMPKα and pꢀAMPKα were detected by western blot analysis. (B)
MCFꢀ7 or MDAꢀMBꢀ231 cells were treated without or with 9f. Cell extracts were subjected to
immunoprecipitation with BRD4 antibody, then detected by western blot using AMPKα antibody. The
expressions of BRD4 and AMPKα were determined in the input. (C) MCFꢀ7 or MDAꢀMBꢀ231 cells were
treated without or with 9f. Cell extracts were subjected to immunoprecipitation with BRD4 and p53
antibodies, then detected by western blot using p53 and AMPKα antibody. BRD4, p53 and AMPKα
expressions were determined in the input.
ACS Paragon Plus Environment

Page 19 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2. Binding free energy predicted by MM/GBSA methods (kcal/mol)
BRD4
ꢀ13.15
ꢀ31.60
ꢀ44.75
ꢀ4.19
BRD3
ꢀ9.98
BRD2
ꢀ11.69
ꢀ30.66
ꢀ42.35
ꢀ4.10
BRDT
ꢀ9.09
∆
E
ele
∆
E
vdw
ꢀ30.04
ꢀ40.02
ꢀ4.10
ꢀ31.21
ꢀ40.30
ꢀ3.95
∆EMM
∆
G
solꢀnp
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
∆
G
solꢀele
20.67
16.48
7.52
19.15
15.05
9.17
20.42
16.32
8.73
21.16
17.21
12.07
ꢀ35.16
ꢀ23.09
∆G
sol
a
∆Gpolar
nonpolar
bind
b
∆G
ꢀ35.79
ꢀ28.27
ꢀ34.14
ꢀ24.97
ꢀ34.76
ꢀ26.03
∆
H
a
b
∆Gnonpolar =∆Evdw + ∆Gsolꢀnp
.
∆Gpolar = ∆Eele +∆Gsolꢀele
.
9f induces ACD in breast cancer cells
We observed 9f could induce massive cytoplasmic vacuolization in the treated
MCFꢀ7 and MDAꢀMBꢀ231 cells (data not shown), which is closely associated with
autophagy. Thus, we examined whether 9f could induce autophagy in breast cancer cells.
Monodansylcadaverine (MDC), a probe for the detection of autophagic vacuoles, was
applied to 9f treated cells, and the green fluorescent dots were observed and
photographed under fluorescence microscopy (Figure 5A). Subsequently, we observed
the morphologic signs of autophagy under electron microscope, characterized by the
extensive vacuoles in the cytoplasm (Figure 5B). The induction of autophagy was further
confirmed by GFPꢀLC3 transfection and immunofluorescence of LC3, which showed an
increase in the LC3 puncta in the cells treated with 9f (Figure 5C and D). To further
confirm that 9f induces autophagy, we found that treatment with 9f resulted in increased
expression of Beclinꢀ1, downregulation of autophagy substrate p62/SQSTM1, as well as a
timeꢀdependent elevation of LC3ꢀII in MCFꢀ7 and MDAꢀMBꢀ231 cells (Figure 5E).
Intriguingly, our results suggested that 9f did not increase the expression of Beclinꢀ1,
whereas increased the transformation of LC3ꢀI to LC3ꢀII in MDAꢀMBꢀ468 cells (Figure S8).
This phenomenon may be due to 9f induced autophagy through a nonꢀcanonical pathway
in MDAꢀMBꢀ468 cells, which is independent on Beclinꢀ1. Next, we used the lysosomal
inhibitor, Bafilomycin A1 (BafA1), to validate the induction of autophagic flux. We found
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 72
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
that LC3ꢀII and p62/SQSTM1 were significantly accumulated in the presence of BafA1,
indicating that autophagic flux is enhanced by treatment with 9f (Figure 5F). To further
determine whether 9fꢀinduced autophagy is cytotoxic or cytoprotective autophagy, we
applied the autophagy inhibitor 3ꢀmethyladenine (3ꢀMA), which inhibits autophagy by
blocking autophagosome formation via inhibiting type III phosphatidylinositol 3ꢀkinases
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
(
PI3KIII), to block the induction of autophagy. The MTT assay was carried out to assess
the cell viability. We found that the cell viabilities were significantly increased following
treatment with 2 mM 3ꢀMA (p<0.01) (Figure 5G). These results suggest that 9f induces
ACD in both MCFꢀ7 and MDAꢀMBꢀ231 cells.
ACS Paragon Plus Environment

Page 21 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure.5 9f induces ACD in breast cancer cells. (A) MCFꢀ7 or MDAꢀMBꢀ231 cells were treated with 1.5 or
3
ꢁM of 9f for indicated times, respectively. Then, the cells were detected by MDC fluorescence staining
under a fluorescence microscope. Scale bar: 50 ꢁm (left panel), 20 ꢁm (right panel). (B) MCFꢀ7 or
MDAꢀMBꢀ231 cells were treated with 1.5 or 3 ꢁM of 9f for indicated times, respectively. Then cells were
processed for electron microscopy. Scale bar=1 ꢁm. (C) MCFꢀ7 or MDAꢀMBꢀ231 cells were transfected
with a GFPꢀLC3 plasmid, followed by treatment with 1.5 or 3 ꢁM of 9f, respectively. Then, the cells were
inspected under a fluorescence microscope. Scale bar=20 ꢁm. (D) MCFꢀ7 or MDAꢀMBꢀ231 cells were
treated with 1.5 or 3 ꢁM of 9f for 24h, respectively. Then, the expression of LC3B puncta were detected
by immunocytochemistry. Green: antiꢀLC3B; Blue: DAPI. Scale bar=20 ꢁm. (E) MCFꢀ7 or MDAꢀMBꢀ231
cells were treated with 9f for the indicated times. The expression levels of LC3, p62 and Beclinꢀ1 were
examined by western blot analysis. (F) MCFꢀ7 or MDAꢀMBꢀ231 cells were treated with 9f for 24 h in the
presence or absence of BafA1, the autophagy flux was detected by western blot analysis of LC3 and p62,
βꢀactin was used as a loading control. (G) MCFꢀ7 or MDAꢀMBꢀ231 cells were treated with 1.5 and 3 ꢁM
of 9f, respectively. 3ꢀMA (1 mM) was added 1 h before treated with 9f. After treatment, cell viabilities were
detected by MTT assay. ***p<0.001 compared with 9fꢀtreated group.
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
9f induces autophagy via BRD4-AMPK-mTOR-ULK complex axis
Next, we examined the upstream of autophagy signaling to illuminate the mechanism
for 9fꢀinduced autophagy. Firstly, we detected the expressions of Akt, pꢀAkt (ser473),
pꢀAkt (thr308), mTOR and pꢀmTOR (ser2448) after 9f treatment. And 9f markedly
reduced the expression levels of pꢀAkt (ser473), pꢀAkt (thr308) and pꢀmTOR (ser2448),
while the expressions of Akt and mTOR were unchanged (Figure 6A). Because the ULK
complex (ULK1ꢀmATG13ꢀFIP200) is closely related with autophagosome formation, we
subsequently examined the expression of ULK1, pꢀULK1 (ser317), ATG13, pꢀATG13
(ser318) and FIP200. The results showed that after treatment with 9f the phosphorylation
of ULK1 and ATG13 were increased, while ULK1, ATG13 and FIP200 were slightly
downregulated, suggesting that 9f may regulate autophagy through the ULK1 complex
(
Figure 6B). To demonstrate whether 9fꢀinduced autophagy is only associated with BRD4
inhibition, we used two specific siRNAs to knockdown the expression of BRD4.
Interestingly, we found that silence of BRD4 decreased the expression of cꢀMyc, which
was consistent with 9f treatment. However, different from 9f treatment, the expression
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 72
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
and phosphorylation of AMPKα and ULK1, as well as the LC3ꢀII accumulation and p62
degradation were merely changed after knocking down of BRD4, indicating silencing of
BRD4 cannot induce obvious activation of autophagy (Figure 6C). To explore whether
9fꢀinduced autophagy is dependent on ATG5, we used specific siRNA to knockdown the
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
expression of ATG5 in cells (Figure 6D). Interestingly, we found that silencing of ATG5 did
not disrupt the activation of AMPK but could block the degradation of p62 and lipidation of
LC3 (Figure 6D). Taken together, these results indicate that 9f induces ATG5ꢀdependent
autophagy via BRD4ꢀAMPKꢀmTORꢀULK complex axis.
Figure.6 9f induces autophagy via BRD4ꢀAMPKꢀmTORꢀULK1 axis. (A) MCFꢀ7 or MDAꢀMBꢀ231 cells
were treated with 1.5 or 3 ꢁM of 9f for indicated times, respectively. Then the expression levels of AKT,
pꢀAKT, mTOR, pꢀmTOR, p70S6K and pꢀp70S6K were detected by western blot analysis. (B) MCFꢀ7 or
MDAꢀMBꢀ231 cells were treated with 1.5 or 3 ꢁM of 9f for indicated times, respectively. Then the
expression levels of the ULK1, pꢀULK1, ATG13, pꢀATG13 and FIP200 were detected by western blot
analysis. (C) MCFꢀ7 or MDAꢀMBꢀ231 cells were transfected with two specific siRNAs of BRD4 or
negative control siRNA for indicated times, respectively. After treatment, the expression levels of BRD4,
cꢀMyc, AMPKα, pꢀAMPKα, ULK1, pꢀULK1, p62 and LC3 were detected by western blot analysis. (D)
MCFꢀ7 or MDAꢀMBꢀ231 cells were transfected with specific siRNA of ATG5 or negative control siRNA for
ACS Paragon Plus Environment

Page 23 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
indicated times, respectively. After treatment, the expression levels of ATG5, AMPKα, pꢀAMPKα, p62 and
LC3 were detected by western blot analysis. βꢀactin was used as a loading control.
Proteomics-based identification of BRD4-modulating autophagic mechanisms in 9f
treatment
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
Previous studies reported that 1 induced AMPKꢀULK1 modulated cytoprotective
2
4
autophagy, thereby conferring resistance to AML stem cells. As a consequence, in our
study, we found that 9f showed more potent autophagic activity than 1 and 2a at
concentration of 5 ꢁM (Figure 2C and Table 1). Interestingly, 3 ꢁM of 9f could markedly
increase the LC3 puncta and LC3 lipidation, as well as the degradation of p62. However,
1
0 ꢁM of 1 induced weak autophagy in both MCFꢀ7 and MDAꢀMBꢀ231 cells. And high
concentration of 2a (150 ꢁM) also lead to obvious autophagy in breast cancer cells,
especially in MCFꢀ7 cells (Figure 7A and B). To further reveal the potential mechanisms
underlying 9fꢀinduced ACD with BRD4, iTRAQꢀbased proteomics analysis was employed
to profile differentially expressed proteins in MCFꢀ7 and MDAꢀMBꢀ231 cells treated with 9f,
respectively. Hundreds of proteins were retrieved from the proteomics results, among
which only differentially expressed proteins (>1.2, <0.8) were selected (Table S2 and S3).
We identified some differentially expressed proteins that interacted with BRD4. In MCFꢀ7
cells, three differential expression proteins HMGB1, VDAC1 and eEF2, were predicted to
interact with BRD4 or to be affected by BRD4. While in MDAꢀMBꢀ231 cells, two differential
expression proteins VDAC2 and eEF2, were predicted to interact with BRD4 or to be
affected by BRD4 (Figure 8A and B). Next, in the MCFꢀ7 cells, we found that 9f induced
the downregulation of HMGB1, as well as a decreased phosphorylation of eEF2 and
upregulation of VDAC1. Otherwise, we found that the phosphorylation of eEF2 was also
downregulated after 9f treatment in the MDAꢀMBꢀ231 cells, while the expression of
VDAC2 was simultaneously increased (Figure 8C and D). To further validate the possible
mechanisms under BRD4 inhibition, we used 1 to compare with 9f. Interestingly, we found
that 1 only displayed some similar results with 9f on the expression of HMGB1, but not
affected the phosphorylation of eEF2, as well as the expressions of VDAC1 and VDAC2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 72
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
(
Figure 8E and F). These results indicate that the 9fꢀinduced autophagy is involved
HMGB1, VDAC1, VDAC2 and eEF2, which still needs to be further explored whether
these proteins can be directly regulated by BRD4.
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 7. 9f induces more obvious autophagy compared with 1 and 2a. (A) MCFꢀ7 or MDAꢀMBꢀ231 cells
were transfected with GFP/mRFPꢀLC3 plasmid, followed by treatment with 3.0 ꢁM of 9f, 10 ꢁM of 1 or
1
50 ꢁM of 2a for indicated times, respectively. Then, the GFPꢀLC3 puncta were observed by
fluorescence microscope. Scale bar = 20 ꢁm. (B) MCFꢀ7 or MDAꢀMBꢀ231 cells were treated with 3.0 ꢁM
of 9f, 10 ꢁM of 1 or 150 ꢁM of 2a for indicated times, respectively. Then the expression levels of LC3 and
p62 were detected. βꢀactin was measured as loading control.
ACS Paragon Plus Environment

Page 25 of 72
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
Figure.8 Proteomicsꢀbased identification of 9fꢀinduced ACD mechanisms. (A) Venn diagram showed the
consensus proteins that were remarkably changed in both 9fꢀtreated MCFꢀ7 and MDAꢀMBꢀ231 cells. (B)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 72
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
Correlation of log10 fold changes between the two samples across all changed proteins in both MCFꢀ7
and MDAꢀMBꢀ231 cells by Deming linear regression. The proteins that were validated by subsequent
experiments were highlighted. (C) MCFꢀ7 cells were treated with 1.5 ꢁM of 9f for indicated times, then
the expression levels of HMGB1, VDAC1, pꢀeEF2 and eEF2 were detected by western blot analysis. (D)
MDAꢀMDꢀ231 were treated with 3.0 ꢁM of 9f for indicated times, then the expression levels of VDAC2,
pꢀeEF2 and eEF2 were detected by western blot analysis. (E) MCFꢀ7 cells were treated with 1.5 ꢁM of 9f
or 10 ꢁM of 1 for indicated times, respectively. Then the expression levels of HMGB1, VDAC1, pꢀeEF2
and eEF2 were detected by western blot analysis. (F) MDAꢀMBꢀ231 cells were treated with 3.0 ꢁM of 9f
or 10 ꢁM of 1 for indicated times, respectively. Then the expression levels of VDAC2, pꢀeEF2 and eEF2
were detected by western blot analysis. βꢀactin was used as a loading control.
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
9f has a therapeutic potential by targeting BRD4-modulated autophagy in vivo
To evaluate the antiꢀtumor activity of 9f in vivo, two breast tumor xenograft models namely
MCFꢀ7, and MDAꢀMBꢀ231 cell lines models were used. We conducted the in vivo study
using three different doses of 9f: 25 mg/kg, 50 mg/kg and 100 mg/kg. In all the models, 9f
showed significant tumor growth inhibition in a doseꢀdependent manner as determined by
8
0% and 76% tumor growth inhibition ratio in MCFꢀ7 and MDAꢀMBꢀ231 cell models,
respectively (Figure 9A and B). And we observed a remarkable loss of tumor weights in all
dose groups (p<0.001) (Figure 9C). Moreover, 9f displayed no obvious effects on the
body weight of all the treatment groups (Figure 9D). To examine whether 9fꢀmediated
inhibition of tumor growth in vivo was associated with reduced cell proliferation and the
increased autophagyꢀassociated cell death, tumor tissues from control and 9fꢀtreated
mice were processed for the immunohistochemical analysis of Kiꢀ67 and LC3. We found
that 9f treatment obviously reduced the number of Kiꢀ67 (p<0.001) positive cells as well as
increased autophagy levels, which was determined by increased LC3 expression
(
p<0.001) (Figure 9E). Next, we carried out western blot analysis to further clarify the
mechanism of 9f in vivo. We found that 9f increased LC3ꢀII accumulation and cleavage of
caspase3 in vivo, as well as inhibit expressions of BRD4 and cꢀMyc, which was almost
consistent with the immunohistochemical analysis and the results obtained in vitro (Figure
9
F). Moreover, we used a zebrafish xenograft model to further assess the antiꢀtumor
ACS Paragon Plus Environment

Page 27 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
effect of 9f. Interestingly, we found that all three doses of 9f caused obvious decrease of
tumor growth (12.5 ꢁM: p<0.01, 25 and 50 ꢁM: p<0.001) compared to the control group
(
Figure 10A and B). Together, these results suggest that 9f bears a good antiꢀtumor
activity in both mouse and zebrafish xenograft models.
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.9 9f has an antiꢀtumor effect in mouse xenograft models via inducing ACD. (A) Mice bearing
MCFꢀ7 (n = 7 per group) or MDAꢀMBꢀ231 (n = 6 per group) tumor xenograft were treated with vehicle
control or indicated dose of 9f once a day. Tumor volume data were presented as mean ± SEM. (B)
Analysis of tumor volume change on final study day. *p<0.05, ***p<0.001, compared with control group.
(
(
C) Tumor weight change of mice in different group. *p<0.05, ***p<0.001, compared with control group.
D) Body weight of mice during 9f treatment. (E) Immunohistochemical analysis of proliferative marker
Kiꢀ67, and LC3ꢀII. Representative images and quantitative analysis of percentage of positive staining are
shown. Tumor tissues were excised from the control and 9f (100mg/kg) treated mice; ***p<0.001
compared with control group. Scale bar=200 ꢁm. (F) Tumor tissues excised from the MCFꢀ7 or
MDAꢀMBꢀ231 xenograft mice were lysed. Then the expression levels of BRD4, cꢀMyc, LC3, and
caspase3 were detected by western blot analysis, βꢀactin was used as a loading control.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 72
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.10 9f has an antiꢀtumor effect on zebrafish xenograft model. (A) Antiꢀtumor activity of 9f in
zebrafish xenograft model. Representative images for bright field and fluorescent images of zebrafish
implanted with CMꢀDilꢀlabeled MCFꢀ7 cells after treatment with vehicle control or different concentrations
of 9f (12.5, 25, 50 ꢁM). (B) Tumor fluorescence intensities (mean ± SEM) and percentages of inhibition in
tumor volumes of treated embryos after 48 h. **p<0.01, ***p<0.001, compared with control group. Scale
bar=200 ꢁm.
DISCUSSION
BRD4, which is an attractive epigenetic target for cancer therapy, engages in direct
regulatory interactions with some DNAꢀbinding transcription factors, such as Mediator,
35
Jmjd6, and PꢀTEFb to influence their functions. Some PPI employed by BRD4 have
been shown to be crucial to the potential therapeutic effects of a BRD4 inhibitor; however,
its regulatory mechanisms for PPI still remain in its infancy. In our study, we constructed
the PPI network of BRD4 with a function enrichment, and identified BRD4 could interact
ACS Paragon Plus Environment

Page 29 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with 28 autophagic proteins including AMPK that were annotated by GO, suggesting
BRD4 may function as a regulator in the autophagic pathways. In TCGA dataset, we
identified that some autophagic proteins such as AMPK, LC3A, LC3B and ATG4C were
remarkably downregulated in breast cancer tissues, which is different from the known
overexpression of BRD4, suggesting that BRD4 might play an inhibitory role in autophagy
and thus inhibition of BRD4 may activate AMPKꢀmodulated autophagy/ACD. Our results
of network predication, TCGA analyses and coꢀimmunoprecipitation together revealed
that BRD4 might interact with AMPK. However, we consider that aboveꢀmentioned result
is insufficient to confirm the direct interaction between BRD4 and AMPK, additional
studies such as structural biology study would help us to clarify the intrinsic mechanism of
BRD4ꢀAMPKꢀmodulated autophagy. More recently, a study have been revealed that
BRD4 represses autophagy and lysosome gene expression. This repression is alleviated
during nutrient deprivation through AMPKꢀSIRT1 signaling, allowing autophagy
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
29
activation. These findings might indicate that disrupting the interaction between BRD4
and AMPK by smallꢀmolecule inhibitor of BRD4, thereby activating AMPKꢀmodulated
autophagy will be a new therapeutic approach to breast cancer therapy.
Of note, AMPK, mTOR and ULK1 together form a complex, which comprises both
1
8
key positiveꢀ and negativeꢀfeedback regulation in autophagy. mTOR and AMPK can
1
9
oppositely regulate ULK1 kinase activity by direct phosphorylation. In our study, we
found that BRD4 might interact with AMPK and thus indicating that there might be the
interaction
between
BRD4
and
AMPK,
which
subsequently
regulates
mTORꢀULK1ꢀmodulated autophagy. Accumulating evidence for ACD has been emerging
21,22
now.
To our knowledge, not any BRD4 inhibitor has been reported to induce ACD in
Thus, targeting BRD4 inhibition and thereby activating
cancer.
BRD4ꢀAMPKꢀmTORꢀULK1ꢀmodulated ACD will fill in the gap in breast cancer therapy.
We designed smallꢀmolecule inhibitor of BRD4 (9f) according to the
a
pharmacophore of the BRD4 (1) and WPF interaction, which was further validated by the
coꢀcrystal structure of 9fꢀBRD4 and other multiple biological evaluation. In addition, we
found that the part B of 9f was a pharmacophore of BRD4, which was similar to the
ACS Paragon Plus Environment