The Optimization of a Novel, Weak Bromo and Extra Terminal Domain (BET) Bromodomain Fragment Ligand to a Potent and Selective Second Bromodomain (BD2) Inhibitor

Article abstract of DOI:10.1021/acs.jmedchem.0c00796

The profound efficacy, yet associated toxicity of pan-BET inhibitors is well documented. The possibility of an ameliorated safety profile driven by significantly selective (>100-fold) inhibition of a subset of the eight bromodomains is enticing, but challenging given the close homology. Herein, we describe the X-ray crystal structure-directed optimization of a novel weak fragment ligand with a pan-second bromodomain (BD2) bias, to potent and highly BD2 selective inhibitors. A template hopping approach, enabled by our parallel research into an orthogonal template (15, GSK046), was the basis for the high selectivity observed. This culminated in two tool molecules, 20 (GSK620) and 56 (GSK549), which showed an anti-inflammatory phenotype in human whole blood, confirming their cellular target engagement. Excellent broad selectivity, developability, and in vivo oral pharmacokinetics characterize these tools, which we hope will be of broad utility to the field of epigenetics research.

Full text of DOI:10.1021/acs.jmedchem.0c00796

Subscriber access provided by University of Birmingham
Article
The Optimisation of a Novel, Weak Bromo and Extra
Terminal Domain (BET) Bromodomain Fragment Ligand to a
Potent and Selective Second Bromodomain (BD2) Inhibitor
Jonathan Thomas Seal, Stephen J. Atkinson, Helen Aylott, Paul Bamborough, Chun-
wa Chung, Royston C. B. Copley, Laurie J. Gordon, Paola Grandi, James R. J. Gray,
Lee A Harrison, Thomas G Hayhow, Matthew Lindon, Cassie Messenger, Anne-Marie
Michon, Darren J Mitchell, Alex Preston, Rab K. Prinjha, Inmaculada Rioja, Simon
Taylor, Ian D. Wall, Robert J Watson, James M. Woolven, and Emmanuel H Demont
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00796 • Publication Date (Web): 23 Jul 2020
Downloaded from pubs.acs.org on July 24, 2020
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 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The Optimisation of a Novel, Weak Bromo and
Extra Terminal Domain (BET) Bromodomain
Fragment Ligand to a Potent and Selective Second
Bromodomain (BD2) 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
†
⁋
†
§
§
Jonathan T. Seal* , Stephen J. Atkinson* , Helen Aylott , Paul Bamborough , Chun-wa Chung ,
§
§
¥
‖
†
Royston C. B. Copley , Laurie Gordon , Paola Grandi , James R. J. Gray , Lee A. Harrison , Thomas G.
†
⁋
†#
§
¥
†
Hayhow , Matthew Lindon , Cassie Messenger , Anne-Marie Michon , Darren Mitchell , Alex
†
†
†
‖⁑
§
†
Preston , Rab K. Prinjha , Inmaculada Rioja , Simon Taylor , Ian D. Wall , Robert J. Watson , James
§
†
M. Woolven , Emmanuel H. Demont
†
‖
Epigenetics Discovery Performance Unit; Quantitative Pharmacology, Immunoinflammation
§
Therapy Area Unit; Platform Technology and Science, GlaxoSmithKline, Medicines Research Centre,
¥
Gunnels Wood Road, Stevenage, Hertfordshire, SG1 2NY, U.K; IVIVT Cellzome, Platform Technology
and Science, GlaxoSmithKline, Meyerhofstr. 1, 69117 Heidelberg, Germany.
Abstract:
The profound efficacy, yet associated toxicity of pan-BET inhibitors is well documented. The possibility
of an ameliorated safety profile driven by significantly selective (>100 fold) inhibition of a sub-set of
the 8 bromodomains is enticing, but challenging given the close homology. Herein, we describe the X-
ray crystal structure-directed optimisation of a novel weak fragment ligand with a pan-second
bromodomain (BD2) bias, to potent and highly BD2 selective inhibitors. A template hopping approach,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 101
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
enabled by our parallel research into an orthogonal template (15, GSK046) was the basis for the high
selectivity observed. This culminated in two tool molecules 20 (GSK620) and 56 (GSK549) which
showed an anti-inflammatory phenotype in human whole blood, confirming their cellular target
engagement. Excellent broad selectivity, developability and in vivo oral pharmacokinetics characterize
these tools, which we hope will be of broad utility to the field of epigenetics research.
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
Introduction:
The bromodomain and extra terminal (BET) family of proteins, comprising the ubiquitous BRD2, 3
and 4 and the testis restricted BRDT, are epigenetic readers whose tandem bromodomains (BDs: N
terminus, BD1 and C-terminus, BD2) recognise and bind acetylated lysine histone tail peptides of
nucleosomes. Once bound to chromatin these proteins recruit transcriptional complexes and
1
therefore play a key role in gene transcription. The therapeutic potential of a pan-Bromo and Extra
Terminal Domain inhibitor (iBET), binding equipotently to the eight bromodomains of the family, has
now been extensively reported both in oncology2-11 and immunoinflammation.12-23 Moreover, iBET are
actively progressing in the clinic for oncology indications, illustrating the tremendous potential of this
epigenetic reader family as a therapeutic target.24-26 Nevertheless it is also well known that a number
of non-clinical toxicities and dose-limiting clinical findings have been associated with pan-inhibition of
the BET family.25, 27-29 There have been far fewer reports of selective BET bromodomain inhibitors,
although these are now starting to emerge in the patent30-37 and publication literature. Both BD1 (1-
7
)38-44 and BD2 (8-12)45-50 selective (>5 fold) inhibitors , binding to either the four BD1 or the four BD2
domains of the BET proteins have been published and the structures are shown in Figure 1 with the
acetylated lysine mimetics highlighted in blue. Despite the emergence of such inhibitors, there is some
debate as to whether 50-100 fold selective inhibitors are really capable of teasing out the relative roles
of BD1 and BD2, due to the profound effects offered by iBET. Notwithstanding this, the recent
5
1
disclosure that BD2-biased inhibitors such as RVX-297 (9b), maintain a range of immune relevant
efficacies in vitro and in vivo is noteworthy.52 In parallel to our own research,53 AbbVie recently
ACS Paragon Plus Environment

Page 3 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
published ABBV-744 (12),48-49 a highly selective BD2-inhibitor which retains some of the anti-
proliferative activity of pan-BET inhibitors, albeit restricted largely to acute myeloid leukaemia and
certain prostate cancer cell lines. Furthermore, there have been a number of studies detailing the
different roles in chromatin binding that the BD1 and BD2 domains have.54-57 The multivalent
interaction between the BET family and chromatin or other binding partners is complex and context-
dependent, and remains poorly understood. It is difficult to predict what different efficacy or safety
profiles might be found when domain-selective inhibitors are compared to pan-BET inhibitors. We
hope that the publication of potent, highly selective BD1 and BD2 domain-selective BET inhibitors will
enable advances in our biological understanding of the mechanism of these highly important and
disease-relevant targets.
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 101
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
BD1-Biased Inhibitors
BD2-Biased Inhibitors
O
N
Olinone (1)
BRD4 BD1 K = 3.4 M
N
O
NH
8
N
d
O
BRD4 BD1 K = 520 nM
d
BRD4 BD2 Kd -
BRD4 BD2 K = 50 nM
d
N
(ITC)
O
N
H
(ITC)
N
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
CN
O O
S
N
H
N
H
MS611 (2)
O
N
R
N
BRD4 BD1 K = 0.41 M
BRD4 BD2 K = 41.3 M
i
RVX-208 (9a)
O
N
i
BRD4 BD1 K = 1.1 M
BRD4 BD2 K = 135 nM
RVX-297 (9b)
BRD4 BD1 K = 1.4 M
d
HO
(FAB)
NH
d
O
O
d
O
O S
N
3
R=OH, 9a
BRD4 BD2 Kd = 180 nM
BRD4 BD1 K = 240 nM
R=
, 9b
(ITC)
d
N
BRD3 BD1 K = 2.9 M
O
d
O
BRD3 BD2 Kd = 2.4 M
BRD4 BD2 K = ND
N
H
d
O
HO
NH
OH
(ITC)
N
H N
2
GSK268 (10)
BRD4 BD1 IC50 = 2.0 M
BRD4 BD2 IC50 = 50 nM
N
N
N
O
N
nBu
N
N
N
N
S
(FRET)
HN
N
N
4
O
O
BRD4 BD1 IC50 = 5.0 M
BRD4 BD2 IC50
-
N
(HTRF)
N
N
BY27 (11)
BRD2 BD1 K = 45 nM
O
O
d
N
N
BRD2 BD2 K = 1.2 nM
d
N
HN
(BROMOscan)
H
O
N
H
MS402 (5)
Brd4 BD1 = 77 nM
Cl
Cl
Brd4 BD2 = 718 nM (FAB)
O
H
N
O
N
6
HN
N
HN
Brd4 BD1 IC50 = 1.8 M
Brd4 BD2 IC50 >100 M
(ALPHAscreen)
ABBV-744 (12)
BRD4 BD1 K = 520 nM
BRD4 BD2 K = 1.6 nM
O
N
N
i
F
HN
HN
i
F
(TR-FRET)
OH
N
Ph
N
O
7
Brd4 BD1 IC50 = 40 nM
Brd4 BD2 IC50 = 6.3 M
(FRET)
N
O
N
N
O
Figure 1. Published Domain Selective BET Bromodomain Inhibitors and their Literature Potencies /
Affinities to the BD1 and BD2 Bromodomain of BRD4 with the acetylated lysine mimetic coloured
blue.
ACS Paragon Plus Environment

Page 5 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Results
As part of our strategy to understand and profile more pan-BD2 inhibitors, a programme of hit
finding was undertaken targeting inhibitors of the second bromodomain ultimately aiming for a pIC50
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 and selectivity over the first bromodomain of at least 100 fold. This approach was driven by the
high degree of sequence homology in the acetylated lysine (KAc) binding region for the BD2 (second)
domain across the BRD 2,3,4 and T proteins. This is in contrast to reduced homology between the first
(BD1) and second domains for the same proteins (Figure S3 a-c, Supporting Information); it was
reasoned that pan-BD2 selectivity had a higher probability of success (versus single isoform
selectivity). Accordingly, an HTS screen was initiated where the entire GSK compound collection (over
2M compounds) was screened at single concentration versus the BRD4 dual domain truncated protein
construct, considered as representative of the BET family. Here the BD1 domain was mutated (Y390A)
to prevent binding of acetyl lysine mimetics at this site and allow evaluation of potency exclusively at
the second domain. Following the single concentration screen (TR-FRET assay format), a triaging
strategy, including an orthogonal BD2 biochemical assay (FP assay format), full curve follow up against
BRD4 BD2 and BD1 and SPR biophysical characterisation, led to real and robust hits. A focus on high
ligand efficiency (LE58-59/LLEAT60), a minimum of 3 fold selectivity over BD1 and CHROMlogD <4 led to
the identification of 13a (Table 1) as one of our preferred hits. Given the molecule contained only 2
61
aromatic rings and had a BRD4 BD2 derived LE of 0.35 and LLE of 0.34, this was an appealing start
AT
point, especially given the level of BRD4 BD2 selectivity was already >10 fold. Whilst 13a contained a
potentially metabolically vulnerable methyl amide warhead, initial microsomal clearance in both rat
and human was <1 mL/min/g tissue. The BD2 selectivity for BRD3 and BRDT was similar for 13a, but
the selectivity profile for BRD2 was somewhat lower and would therefore be monitored as we
optimised this start point. The physicochemical profile was further improved by removal of the Br and
F atoms to give 13b, with no deleterious effect on potency and a large improvement in LLE (0.43).
AT
As part of generating an understanding of the key core functionality required for potency and
selectivity, screening of the corresponding 2-pyridone 14a led to 10 fold increase in potency for BRD4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 101
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
BD2, albeit accompanied with a similar rise in the BRD4 BD1 potency. Notably, an equivalent potency
and selectivity was observed against BRD2, 3, T BD2, suggestive that this series was capable of
generating pan-BD2 inhibitors. With a BRD4 BD2 derived LE of 0.45 and LLE of 0.51, this represented
AT
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
a highly efficient start point to begin optimisation. When the methyl group of the amide was replaced
with an ethyl (14b), a reduction of almost 1 log in potency was observed, together with reduced BD2
selectivity, suggestive that the methyl group plays an important and specific role in the binding of this
template.
Table 1. Structure and Properties of BD2 HTS hit 13a and Optimised Fragments 13b, 14a and 14b.
R1
HN
O
HN
O
X
3
O
N
N
Y
5
O
X = F, Y = Br, 13a
X = Y = H, 13b
R = Me, 14a
1
R = Et, 14b
1
Compound
13a
13b
14a
14b
a
BRD4 BD2 / BD1 pIC (n)
5.1 (6) / 4.0
4.9 (5) / <4.3
(5)
5.9 (9) / 5.1
(9)
5.2 (4) / 4.8 (4)
5
0
b
(1 )
a
c
f
BRD2 BD2 / BD1 pIC (n)
4.6 (6 ) / 4.5
(2 )
4.6 (8 ) / 4.8
(2 )
5.6 (8) / 4.8
(8)
-
-
-
5
0
d
e
a
a
BRD3 BD2 / BD1 pIC (n)
5.0 (6) / <4.3
(6)
5.0 (4) / 4.8
6.3 (4) / 4.9
(4)
50
c
(2 )
c
c
BRDT BD2 / BD1 pIC (n)
4.9 (5 ) / 4.4
4.7 (6 ) / 4.5
5.7 (6) / 5.0
(6)
5
0
e
e
(
2 )
(2 )
BRD4 BD2 selectivity (fold)
BRD4 BD2 LE / LLEAT
13
>4
7
3
0.35 / 0.34
0.37 / 0.43
1.6 / 0.6
≥108 / 58
0.45 / 0.51
3.0 / 0.6
78 / 595
0.37 / 0.40
3.7 / 1.1
- / 790
CHROMlogD (pH 7.4) / clogP 2.6 / 1.6
g
h
CLND (µg/mL) / AMP
≥143 / 260
(nm/sec)
Microsomal Cli (rat, hu,
mL/min/g tissue)
0.76 / 0.53
-
-
-
a
In all tables, the activity of the inhibitors has been determined using a TR-FRET assay, see
b
experimental section for details. The average standard deviation for pIC is 0.1; also tested <4.3
5
0
ACS Paragon Plus Environment

Page 7 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
c
d
e
f
(
n=5); also tested <4.3 (n=2); also tested <4.3 (n=8); also tested <4.3 (n=6); also tested <4.3 (n=1);
g
h
CLND = chemiluminescent nitrogen detection; AMP = artificial membrane permeability.
There is a wealth of structural information concerning the binding mode of a variety of pan-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
inhibitors to their target proteins.3, 62-64 Indeed, an X-ray crystal structure of 14a in BRD2 BD2 was soon
solved (Figure 2, a) revealing the key interactions. The pendant methyl amide serves as the KAc-
mimetic, its carbonyl oxygen forming hydrogen-bonds to Asn429 and the conserved water network in
an analogous manner to the carbonyl of the acetylated histone. The amide hydrogen-bonds internally
to the pyridone 2-carbonyl oxygen, while its methyl group fills the KAc-methyl pocket. Ligand-efficient
fragments containing related aryl amides have been shown to bind relatively weakly to other
bromodomains in a similar way, but so far these have been limited to examples outside the BET family
6
5
such as BAZ2B. The benzyl group occupies the hydrophobic “WPF” shelf, packing against Val435 and
Trp370. It also contacts a histidine residue (His433) in an edge-to-face interaction. A similar interaction
4
5
has been noted between His433 and the shelf group of other inhibitors. In BD1 this histidine is
replaced with an aspartic acid residue, and this “Asp/His switch” and the differing interactions of the
shelf aromatic ring as a result could account for the BD2-domain selectivity observed for 14a.
WPF Shelf
W370
(
a)
(b)
W370
P371
H433
H433
N429
P371
F372
N429
F372
Y386
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 101
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 2. (a) X-ray Structure of 14a in BRD2 BD2 (PDB 6ZB0) showing key interactions with Asn429
and a network of 5 water molecules, the benzyl group occupies the “WPF” shelf; (b) 14a (green)
5
3
overlaid with alternate tool 15 (magenta) in BRD2 BD2 (PDB 6SWP ) highlighting the additional
vector and interaction with Asn429 utilised by 15.
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
Despite the excellent efficiency of 14a, the compound remained a relatively weak inhibitor, with a
large increase in BD2 potency required. Amongst the published pan-BET inhibitors, almost all series
gain potency by occupying three key regions of the BET binding site: a potent, often heteroaromatic
warhead which serves as the KAc-mimetic, the WPF shelf and/or the ZA channel. In this case, we
believed we had achieved BD2 bias by starting from a relatively weak non-aromatic warhead and
engaging the shelf of the protein. Whilst the benzylic methylene offered a vector to enter the ZA
channel, this did not lead to the required level of potency and selectivity (data not shown) and an
alternative growth vector was sought.
Table 2. Structure and BRD4 Potency and Selectivity of GSK046 (15).53, 66
O
HN
O
F
3
OH
1
O
N
H
Profile
15
BRD4 BD2 pIC (n)
7.3 (14)
4.2 (3a)
1400
5
0
BRD4 BD1 pIC (n)
5
0
Selectivity (fold)
aalso tested <4.3 (10)
In parallel to our efforts on this series, a second HTS hit was being elaborated and SAR-driven
optimisation had led to the tool 15 (GSK046, Table 2).53, 66 The acetamide warhead of 15 and its “WPF
shelf” benzyl group bound in an analogous manner to 14a, but 15 offered >1000 fold selectivity for
ACS Paragon Plus Environment

Page 9 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
BRD4 BD2 over BD1. The increase in potency and selectivity was believed to be due to the addition of
a second amide substituent entering a cleft in the binding site formed by Leu383, Tyr428 and Asn429
(see Figure 2, b). Additionally, the NH of the C1 amide substituent made a hydrogen bond with the
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
carbonyl of Asn429 (which also interacts with the warhead carbonyl, vide supra). Comparing the X-ray
structures of bound compounds 14a with 15, we considered that adding an amide group to the 5-
position of the pyridone core may lead to a similar increase in potency and selectivity.
The installation of an amide group at C5 of 14a (cf. Table 1 & 3) delivered potent and highly selective
inhibitors, whilst also improving the physicochemical properties of the series (i.e. lower CHROMlogD
and higher CLND solubility vs 14a). Table 3 shows that increasing the size of the amide from primary
2
2
2
(
16, R = H) to a secondary methyl (17, R = Me) and ethyl amide (18, R = Et) offers a stepwise increase
in both potency and selectivity compared to 13. 16-18 maintain excellent binding efficiency, with
similar LE values and an increase in LLE vs fragment 13. Ethyl amide 18 in particular is close to 100
AT
nM, with >100 fold selectivity and an LE of 0.41 and LLE of 0.52. These data validated both our hit
AT
finding approach of starting with a small and efficient fragment, and of working on multiple templates
in parallel during the hit-to-lead phase.
2
To assess the SAR at R a range of logical modifications were investigated. For example, branching
of the ethyl amide to iPr (19) was less well tolerated, however, small cyclic alkyl groups proved optimal,
with cPr and cBu (20 and 21) boasting selectivity levels of 200 fold. Larger rings were less well tolerated
(e.g. THP 22). This is in contrast to acetamide 15 (Table 2) and represents divergent SAR between the
6
6
series. Searching for additional affinity, the effect of growing from the promising ethyl amide was
also investigated; the installation of polarity from the beta-carbon, such as for 23 and 24, was poorly
tolerated with a loss of >3 fold of potency relative to 18. Indeed, branching from this carbon of the
amide was also deleterious (25). Potency could be restored through gamma-branched examples 26
and 27 although this appeared to offer no advantage to the smaller derivatives of similar potency
already identified where both the LE and LLE values were significantly higher (cf. 18 and 20). Whilst
AT
a basic amine was not tolerated in example 28, extending the carbon chain further did provide a BRD4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 101
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
BD2 potency boost and led to the most selective compounds prepared yet (>300 fold over BD1).
Piperidine 29, expectedly showed poor predicted permeability in an artificial membrane high
6
7
throughput assay, by virtue of its high basicity, although this was somewhat attenuated in the less-
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
basic morpholine 30. Nevertheless it was clear that the small and neutral ethyl and cyclopropyl amides
1
8 and 20 offered the best balance of potency, selectivity and physicochemical properties. Indeed, cPr
analogue 20 had a low clogP and CHROMlogD, was predicted to be permeable and was soluble in our
high throughput CLND kinetic solubility assay (conducted from DMSO solution).68
Table 3. SAR for the amide vector (R2)
HN
O
3
O
N
R2
5
O
Compound R2
BRD4 BD2 /
Selectivity BRD4
BD2 LE (pH 7.4) /
/ LLEAT clogP
CHROMlogD CLND
(µg/mL)
/ AMP
nm/s)
BD1 pIC (n) (fold)
50
(
1
1
1
1
2
2
6
7
8
9
0
1
NH2
6.4 (4) / 5.1
6)
20
0.42 /
0.58
2.0 / -0.7
71 / 110
(
NHMe
NHEt
6.8 (3) / 5.1
3)
50
0.42 /
0.56
2.4 / -0.5
3.0 / 0.0
3.6 / 0.3
3.1 / 0.1
3.8 / 0.4
2.8 / -0.9
≥118 /
130
(
6.8 (5) / 4.8
120
15
0.41 /
0.51
≥144 /
245
a
(
4 )
NHiPr
NHcPr
NHcBu
6.2 (3) / 5.0
0.35 /
0.44
≥136 /
330
b
(
1 )
7.1 (16) / 4.8 220
(14)
0.40 /
0.51
≥154 /
200
7.0 (8) / 4.7
220
0.39 /
0.47
28 / 427
c
(5 )
22
H
N
6.6 (4) / 4.9
60
0.34 /
0.49
≥174 /
110
(3)
O
ACS Paragon Plus Environment

Page 11 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
2
2
2
3
4
5
6
NH(CH ) OMe
6.2 (4) / 5.1
3)
10
30
0.34 /
0.46
2.7 / -0.3
1.9 / -1.1
4.3 / 1.0
4.8 / 1.5
≥164 /
110
2
2
(
NH(CH ) OH
6.2 (4) / 4.7
3)
0.35 /
0.52
≥140 /
22
2
2
(
H
N
6.3 (3) / <4.3 >100
3)
0.34 /
0.40
43 / 510
10 / 530
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
(
H
N
6.9 (3) / 4.5
(3)
210
0.36 /
0.39
2
7
H
N
6.7 (3) / 4.6
120
0.31 /
0.41
3.6 / 0.3
1.6 / 0.3
≥127 /
170
(3)
O
28
H
N
4.6 (3) / <4.3 >2
(3)
0.22 /
0.32
≥146 /
<3
HCl
HCl
N
H
2
3
9
0
H
N
7.8 (3) / 5.2
3)
370
0.36 /
0.43
2.0 / 0.8
1.4 / -0.6
156 / <3
2
2
(
N
H
H
N
7.7 (2) / 5.1
(2)
340
0.35 /
0.48
≥194 /
15
O
N
H
a
b
c
also tested <4.3 (n=1); also tested <4.3 (n=2); also tested <4.3 (n=3)
20 was soaked into the BRD2 BD2 construct and bound as shown in Figure 3, a. In addition to the
interactions of the KAc mimetic and shelf groups, already described for 14a, the additional hydrogen
bond between the amide NH and Asn429 that we had hoped to engage, was indeed observed. The cPr
group shows excellent shape complementarity with a narrow groove created by His433 and Pro430
on one side and Tyr428 and Leu383 on the other (Figure 3,b). This may explain why larger groups in
this position are less well tolerated. The origin of BD2 selectivity arises from residue differences
between BD1 and BD2 as highlighted in Figure 3, d. The key residue changes are His433 and Pro430
(
BRD2 BD2 numbering) to Asp144 and Lys141 (BRD4 BD1 numbering). Both changes affect the shape
of the groove that the cyclopropyl group occupies. As described earlier, in BD2 His433 makes a key
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 101
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
edge-to-face interaction with the shelf benzyl group. In addition, the tight-packed arrangement of the
cyclopropyl amide and His433 adds significantly to the surface area of this residue buried by 20. The
combination of all these positive interactions which are absent in BD1 are likely to drive the observed
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
Figure 3. (a) Key interactions of GSK620 (20) with BRD2 BD2 (PDB 6ZB1, cyan); (b) GSK620 shown
with the protein surface representation of BRD2 BD2; (c) Key interactions of GSK620 (20) with BRD4
BD1 (PDB 6ZB3, white); (d) Overlay of GSK620 (20) in BRD2 BD2 (cyan) and BRD4 BD1 (white)
providing a rationale for the domain selectivity.
It is important to realise that this “Asp/His switch” involves conformational differences as well as an
amino acid exchange. In BRD2 BD2, His433 points towards the KAc site and the bound inhibitor 20,
ACS Paragon Plus Environment

Page 13 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
whereas in BRD4 BD1 the equivalent Asp144 sidechain points away (Figures 3 a,c). This is consistent
with systematic differences found in deposited Protein Data Bank crystal structures, where the BET
BD1 domain aspartic acids typically lie in gauche(+) chi1 rotameric states while the BD2 histidines are
predominantly seen in the trans rotamer with minor gauche(+) components (Figure S3 e-g, Supporting
Information). In this conformation, the imidazole of BRD4 His433 is able to accept a helix-capping
hydrogen-bond from the Val435 backbone NH. In BRD4 BD1, the Asp144 sidechain forms a conserved
hydrogen-bond with the free backbone NH of Lys141, perhaps explaining the prevalence of its
gauche(+) conformation. The substitution of Lys141 for Pro430 in BRD2 BD2 makes this interaction
impossible and may sterically restrict the loop, encouraging the histidine to lie in its trans
conformation. The Asp144/Lys141 and His433/Pro430 combinations are conserved in all BET BD1 and
BD2 domains (Figure S3 d, Supporting Information) so this rationalisation of BD2 over BD1 selectivity
extends across the family.
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 potency and physicochemical properties of amides 18 and 20 looked highly promising and
indeed the CLND solubility was towards the maximum limit of the high throughput assay for the vast
majority of analogues. However, upon crystallisation of cPr amide 20 [confirmed by X-Ray Powder
Diffraction (XRPD) analysis] and evaluation of its fasted-state simulated intestinal fluid (FaSSIF)
solubility, a far more relevant end point for an orally-targeted inhibitor, its solubility was only
moderate (13-35 µg/mL, Table 4). Given the attractive physicochemical space occupied by this
inhibitor, it was likely this low solubility was being driven by its highly crystalline nature. To investigate
this further, a small molecule crystal structure was determined for 20 (Figure 4, a). The Me-amide
…
group forms an intramolecular N-H O hydrogen bond with the pyridone oxygen atom, O14. The cPr-
…
amide forms an intermolecular N-H O hydrogen bond with the oxygen atom of the Me-amide, O16.
These intermolecular hydrogen bonds between the two pivotal amides connect molecules of 20 into
columns (Figure 4, b) that extend in the direction of the crystallographic b-axis. There are no pi-
stacking interactions in this structure. Given the amide groups of 20 make important hydrogen bonds
to the BD2 domain and were therefore considered non-replaceable, we interrogated the other
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 101
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
proximal interactions within the small molecule crystal structure. It was noteworthy that the benzylic
methylene resided in close proximity to the carbonyl oxygen of the cPr-amide in an adjacent molecule,
whilst the 6-position of the pyridone core also sits close to the same CH of its neighbour (Figure 4, c).
We hypothesised that adding substituents to these positions may be sufficient to disrupt the packing
and lead to increased solubility. As shown in Table 4, this hypothesis proved correct and crystalline
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
(by XRPD) single enantiomer 31, not only showed a markedly increased FaSSIF solubility of 815 µg/mL,
but was well tolerated in terms of potency and selectivity. Interestingly, further additions to the cleft
region offered no advantage to 20 and 31 in terms of potency or selectivity. Unfortunately, whilst
crystalline (by XRPD) 6-methylpyridone 32 also had an excellent FaSSIF solubility, this substituent
impacted binding at BD2, with a low BRD4 BD2 pIC of 5.7 resulting and was not progressed further.
5
0
Table 4. The Potency, Selectivity and Solubilities of Methylated Derivatives of 20
HN
O
R = H, R = H, 20
3
4
O
R = Me, R = H, 31
3
4
H
N
N
R = H, R = Me, 32
3
4
R3 R4
O
Compound
20
31
32
BRD4 BD2 / BD1 pIC (n)
7.1 (16) / 4.8 (14)
220
7.1 (4) / 4.9 (4)
170
5.7 (5) / <4.3 (4)
>30
5
0
BRD4 BD2 Selectivity (fold)
CHROMlogD (pH 7.4) / clogP
3.1 / 0.1
3.5 / 0.4
≥168 / 300
815
2.9 / 0.2
≥126 / 110
577
a
b
CLND (µg/mL) / AMP (nm/s)
≥154 / 200
13 – 35
c
FaSSIF (µg/mL)
a
b
CLND = chemiluminescent nitrogen detection solubility; AMP = artificial membrane permeability;
FaSSIF = fasted state simulated intestinal fluid solubility
c
ACS Paragon Plus Environment

Page 15 of 101
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
(a)
(b)
(c)
Figure 4. (a) A view of a molecule of 20 from the crystal structure, showing the numbering scheme
employed. Anisotropic atomic displacement ellipsoids for the non-hydrogen atoms are shown at the
50% probability level. Hydrogen atoms are displayed with an arbitrarily small radius. (b) Molecules of
2
0 forming part of a hydrogen bonded column within the crystal structure. Hydrogen bonds are
displayed as dashed lines with the donors and acceptors labelled. (c) Packing plot of the crystal
structure of 20 viewed down the crystallographic b-axis showing the close proximity of the benzylic
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 101
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
methylene and 6-pyridone hydrogens to an adjacent molecule (circled). Hydrogen bonds are displayed
as dashed lines with the donors and acceptors labelled.
Further profiling of 20 and 31 will be discussed subsequently, but it was also important to evaluate
the SAR of the shelf region of the template. Due to its key role in BD2 binding, further affinity may be
possible in this region, whilst diversity could also be valuable for physicochemical properties and
metabolic stability. As shown in Table 5, the replacement of the phenyl ring of 20 with cyclohexyl 33
was tolerated. Given the edge-to-face interactions observed for 20, this was somewhat surprising –
but later rationalised by crystallography: the cyclohexyl ring adopts a different conformation to the
phenyl group and packs nicely against the tryptophan. Unfortunately, capitalising on this observation
was problematic. Cyclohexyl 33 itself offered no benefit in terms of lipophilicity (CHROMlogD = 4 vs
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
3
.1 for 20) and attempts to insert polarity either onto or into this ring led to a significant reduction in
potency (e.g. THPs 34-36). It would have been similarly desirable to increase the polarity of the
aromatic ring, but unfortunately pyridyls 37-39 showed a 10 fold reduction in potency, with no obvious
means of obtaining this back elsewhere in the inhibitor. A range of electron withdrawing groups (EWG)
and electron donating groups (EDG) were then prepared but as 40-46 show, the potency and
selectivity SARs were flat with no compound offering an advantage in these terms over the
unsubstituted lead 20. The incorporation of a 3-CN group (47) proved detrimental, whilst in contrast,
the addition of solubilising groups such as secondary alcohol 48, or morpholine 49 were well tolerated
and led to a large increase in FaSSIF solubility. Morpholine 49 had sub-optimal in vitro clearance data
in rat (14.7 mL/min/g tissue), however alcohol 48 was worthy of further evaluation and will be
discussed further alongside our other preferred inhibitors.
The addition of a meta-aniline (50) appeared to offer a small selectivity advantage over 20. Hoping
to capitalise on this observation and given that a free aniline was not desired from a genetic toxicity
6
9
viewpoint, it was hypothesised that bicyclic motifs bearing an appropriate H-bond donor at this
position were worthy of investigation. Cyclisation of the meta- and para-positions to provide
benzimidazole 51 showed a reduction in potency relative to aniline 50, however cyclisation of the
ACS Paragon Plus Environment

Page 17 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ortho- and meta-positions looked more encouraging, with benzimidazole 52 equipotent to the aniline.
Preparation of regioisomeric indazoles 53 and 54 was also tolerated, though in both cases selectivity
was reduced relative to the benzimidazole. Benzofuran 55, lacking any H-bond donor was only
moderately potent and selective. This was in contrast to indole 56 which was not only 3 fold more
potent at BRD4 BD2 than any other example tested, but also exquisitely selective versus BD1 (>1000
fold). Interestingly, the same potency increase was not observed with regioisomeric indole 57. A
crystal structure in BRD2 BD2 for 56 was solved and provided a rationale for its high potency. As shown
in Figure 5, the indole group, as expected, still occupies the WPF shelf. However, the additional pi-
system of the pyrrole ring now appears to engage in a more optimal edge-to-face interaction with
His433 (which remains in the same place as observed for 20 (Figure 3). The NH of the indole also makes
a through water H-bond with Asp434. It is hypothesised that it is this combination of interactions
which drives the additional BD2 potency of 56.
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
D432
D434
P430
W370
H433
H433
P431
P430
W370
P371
N429
L381
Y428
L383
Y386
Figure 5. X-ray Structure of 56 in BRD2 BD2 (PDB 6ZB2) showing: (a) the key interactions with
His433, Asp434 and Trp370 on the WPF shelf; (b) the binding mode with a surface to highlight its
optimal position on the WPF shelf.
Given the excellent potency of 56, it was postulated that further polarity might be able to be
incorporated in the Ph ring. 5-aza-indole 58 was prepared and whilst a 1 log drop-off in biochemical
potency was observed (in accord with o-37 vs 20), 58 still had a reasonable potency / selectivity profile,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 101
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 highly polar nature of the compound. Finally, to understand the effect of the second pi-
system on the enhanced potency of indole 56, the corresponding indoline 59 was prepared. In this
instance, 3 fold potency was lost versus BD2, supporting the binding hypothesis in Figure 5, that the
extended pi-system, as well as the meta-hydrogen bond donor, has a role in the high potency of 56.
Nevertheless, given indoline 59 has one less aromatic ring, has a reduced CHROMlogD of 2.1 (vs 3.1
for 20 and 2.9 for 56), and a BD2 selectivity of 415 fold and an enhanced solubility, this analogue was
also worthy of further investigation.
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
Table 5. SAR for the “WPF shelf” vector (R5).
HN
O
O
R5
N
R2
O
CLND
µg/mL) /
FaSSIF
(µg/mL) /
AMP
(
BRD4 BD2 /
BD1 pIC50
(n)
CHROMlog
D (pH 7.4)
/ clogP
Selectivity
fold)
Compound
R2
R5
(
(nm/s)
7
4
.1 (16) /
.8 (14)
≥154 / 13-
35 / 200
2
2
3
0
1
3
NHcPr
NHcBu
NHcPr
Ph
220
220
85
3.1 / 0.1
3.8 / 0.4
4.0 / 1.3
2.6 / -0.5
2.0 / -1.1
1.5 / -1.1
2.2 / -1.1
7
(5 )
.0 (8) / 4.7
28 / 24 /
427
Ph
a
6
(3)
.9 (3) / 5.0
≥148 / - /
190
cHex
6.4 (3) / 4.5
≥149 / - /
46
o-34
m-35
p-36
o-37
70
b
(1 )
O
5
.6 (3) / 4.4
≥131 / - /
22
NHcPr
NHcBu
14
b
(1 )
5
.5 (3) / 4.6
≥165 / - /
12
10
(3)
5
<
.9 (3) /
4.3 (3)
≥103 / 920
/ 51
>40
ACS Paragon Plus Environment

Page 19 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CLND
µg/mL) /
FaSSIF
(µg/mL) /
AMP
(
BRD4 BD2 /
CHROMlog
D (pH 7.4)
/ clogP
Selectivity
(fold)
Compound
R2
R5
BD1 pIC50
n)
(
(nm/s)
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
6
<
.0 (3) /
4.3 (3)
≥99 / 189 /
130
N
m-38
p-39
o-40
m-41
p-42
m-43
p-44
m-45
p-46
>50
>20
120
170
120
250
100
240
190
1.8 / -1.1
1.7 / -1.1
3.0 / 0.2
3.2 / 0.2
3.2 / 0.2
3.6 / 0.6
3.5 / 0.6
3.0 / 0.0
3.0 / 0.0
5
<
.6 (3) /
4.3 (3)
≥118 / 121
/ 130
6
.9 (3) / 4.8
≥145 / 88 /
220
(3)
F
6
.9 (4) / 4.7
≥133 / 31 /
300
NHcPr
NHcPr
(3)
7
.1 (5) / 5.0
≥162 / 84 /
300
(4)
7.2 (5) / 4.8
≥205 / 88 /
440
Me
c
(
4 )
7.0 (3) / 5.0
3)
24 / 11 /
350
(
6
.8 (4) / 4.4
79 / 582 /
220
OMe
CN
c
(1 )
NHcPr
NHcPr
NHcPr
NHcPr
7
.0 (4) / 4.7
≥143 / 263
/ 220
(4)
6
.1 (5) / 4.4
≥148 / - /
100
4
7
50
2.6 / -0.5
2.2 / -0.6
2.3 / 0.2
b
(2 )
OH
7
.0 (4) / 4.1
≥131 / 495
/ 34
48
690
800
a
(2 )
N
O
7.0 (3) / 4.1
≥185 / 887
/ 68
4
9
a
(
2 )
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 101
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
CLND
µg/mL) /
FaSSIF
(µg/mL) /
AMP
(
BRD4 BD2 /
CHROMlog
D (pH 7.4)
/ clogP
Selectivity
(fold)
Compound
R2
R5
BD1 pIC50
n)
(
(nm/s)
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
NHMe
7
.2 (3) / 4.6
≥155 / - /
430
5
0
NHcBu
NHcPr
450
180
3.6 / -0.1
1.2 / -0.5
c
(2 )
HN
N
6
.6 (4) / 4.4
≥163 / 225
/ <3
51
b
(2 )
Me
N
HN
7.3 (4) / 4.4
4)
≥198 / - /
24
5
5
2
3
NHcPr
NHcPr
800
390
1.5 / -0.2
1.5 / -0.4
(
HN N
7.1 (3) / 4.5
10 / 4 / 38
b
(
1 )
N
NH
6
<
.8 (3) /
4.3 (3)
≥181 / 131
/ 190
5
5
5
5
4
5
6
7
NHcPr
NHcPr
NHcPr
NHcPr
>320
300
2.2 / -0.4
3.4 / 0.6
2.9 / 0.1
3.7 / 0.1
O
6
.9 (4) / 4.4
≥110 / 74 /
410
c
(4 )
HN
7
.8 (8) / 4.7
78 / 17-23
/ 178
1150
200
(8)
NH
7.0 (3) / 4.6
≥144 / 11 /
480
c
(
3 )
ACS Paragon Plus Environment

Page 21 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CLND
µg/mL) /
FaSSIF
(µg/mL) /
AMP
(
BRD4 BD2 /
CHROMlog
D (pH 7.4)
/ clogP
Selectivity
(fold)
Compound
R2
R5
BD1 pIC50
n)
(
(nm/s)
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
HN
6
<
.7 (4) /
4.3 (4)
58
NHcPr
NHcPr
>230
415
1.7 / -0.9
2.1 / -0.3
73 / - / 7
N
HN
7
(4 )
.2 (7) / 4.6
≥184 / 100
/ 140
5
9
a
a
b
c
also tested <4.3 (n=3); also tested <4.3 (n=2); also tested <4.3 (n=1)
With the desired level of potency and selectivity in hand, and having controlled the
physicochemical properties closely (lipophilicity, permeability and solubility) a diverse set of optimised
inhibitors remained. In search of an orally bioavailable molecule, a selection of lead compounds were
evaluated for their pharmacokinetics in vivo. Before committing to such studies, all compounds were
evaluated in vitro using isolated hepatocytes. As indicated in Table 6, the ethyl and cPr amides 18 and
20 have low measured in vitro Cl in hepatocytes for rat, human and dog (all <1 mg/mL/g tissue). This
i
translated into low in-vivo clearance for both compounds in the rat and when dosed orally, excellent
oral bioavailability. Given 20 had slightly superior selectivity, this example was also profiled in the dog,
with a blood clearance (Cl ) of 2 mL/min/kg and an oral bioavailability (F%) of 57. It was also notable
b
that both compounds had high free fractions in rat and dog, illustrating that unbound clearances are
70
also low and high free blood concentrations in vivo are achievable. A number of the more selective
biaryl shelf examples were also profiled. Whilst indazole 53 had raised Cl in vivo (in contrast to its in
b
vitro rat profile) and poor oral bioavailability (likely driven by poor absorption), the highly selective
indole 56 and indoline 59 both had low/moderate rat Cl of 23 and 16 mL/min/kg respectively. In the
b
case of indoline 59 this translated into an oral bioavailability of 53%. Whilst the oral leg of the rat study
for 56 was run in a different species (female Lewis) and so the bioavailability cannot be calculated, 56
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 101
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
also had good oral exposure as shown by an AUC/D of 23.6 min.kg/L. Finally, two examples where
increased FaSSIF solubility had been achieved were tested. Secondary alcohol 48 unfortunately
showed high Cl in rat in vivo, which correlated with raised in vitro clearance in hepatocytes in that
b
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
species. Interestingly, this compound is predicted to have low clearance in dog and human, however,
given the improved rat Cl profile of other analogues, this was not tested further. Finally, the addition
b
of the methyl group to the benzylic methylene to provide 31 unfortunately resulted in high rat Clb.
Overall, the excellent pharmacokinetics of 20, together with the enhanced potency and selectivity of
56 allowed us to select these as our two preferred BD2-selective and in vivo capable tools.
Table 6. The pharmacokinetic profile of selected BD2 inhibitors following intravenous infusion and
oral administration in the Male Wistar Han Rat and Beagle Dog
Compound Species Hepatocyte Dose
CLb
Vss
t½
Fpo
fub
CLint
(mL/mi
n/kg)
iv (n) / po (n)
(mg/kg)
(L/kg) (h) (%)
(mL/min/g
tissue)
Rat
0.95
1 (1) / 3 (3)
12
0.8
1.1 67
0.16
2
1
0
8
Human <0.45
Dog
Rat
<1.26
<0.80
0.5 (1) / 1 (1)
1 (3) / 3 (3)
2
1.0
0.8
5.7 57
0.38
0.16
12
0.6 104
Human <0.87
Dog
Rat
<1.26
<0.80
1.5 (1) / 3 (3)
1 (1) / 10 (3a)
42
23
1.2
1.3
0.4 14
0.22
0.18
53
Human <0.45
Rat 1.70
Human <0.45
0.8 NDa
5
5
6
9
Dog
Rat
<1.26
<0.80
1 (1) / 3 (3)
16
0.9
0.6 53
0.53
Human <0.45
Dog <1.26
ACS Paragon Plus Environment

Page 23 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Rat
3.0
1.4 (1) / -
1.4 (1) / -
74
72
1.7
1.6
0.4
0.4
-
-
0.48
0.27
4
8
Human <0.45
Dog
Rat
<1.26
1.50
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
31
Human <0.87
aF% cannot be calculated as po dosing carried out using alternate female Lewis rat strain, po AUC/D
= 23.6 min.kg/L.
Chemistry
In order to synthesise this series of analogues, a plan was envisaged which commenced from
2
3
commercial pyridone 60 (Scheme 1). To vary the R amide substituent, the R = benzyl “shelf” group
can be installed at an early stage. Accordingly, selective N-benzylation and a subsequent two-step
process converts the ester into the corresponding methyl amide 61 in 66% yield for the three steps.
At this stage, two alternate routes were utilised in parallel. Bromide 61 can be carbonylated either in
the presence of 2,4,6-trichlorophenyl formate using palladium acetate and xantphos catalyst-ligand
7
1
system. The resultant activated ester 62 is formed in 87% yield and can be easily displaced by the
o
desired amines in the presence of catalytic DMAP and triethylamine in THF at 45 C to form analogues
1
6-17, 19, 21, 25-27 and 63. Alternatively, carbonylation under similar palladium-catalysed conditions
7
2
(
Pd(OAc) , dppb) but using carbon monoxide and EtOH affords the ethyl ester. This transformation
2
can be practically carried out on a small scale using a balloon of CO. However, we were able to scale
o
this reaction using a Parr reactor vessel purged with CO to 50 psi and heated to 100 C overnight. This
approach delivered >20 g of this key intermediate. After sodium hydroxide-mediated hydrolysis, this
provided carboxylic acid 64 in 75% yield over the two steps. Whilst both processes were similarly
effective, employing 2,4,6-trichlorophenyl formate, which itself had to be prepared in a single step,
was less atom efficient than preparing the ethyl ester. Finally, carboxylic acid 64 could be converted
to the desired amides 18, 20, 22-24 and 65-66 using a HATU-mediated amide coupling in either DMF
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 101
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
or DCM. For piperidine and morpholine containing analogues 63, 65 and 66 a subsequent Boc-
deprotection step was required to afford the final free amines 28-30.
3
For analogues where the R shelf group needed to be varied, direct aminocarbonylation was
3
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
utilised which meant only a single step was required after the different R diversity had been installed.
As such, methyl ester 60 could be converted to the corresponding methyl amide 67 in a single step
3
with an excess of methylamine in refluxing THF in 44% yield. A variety of R groups (68-81) could then
be incorporated via N-alkylation of the pyridone using potassium carbonate and heating in either DMF
or MeOH. Finally, the key aminocarbonylation step directly installs the amide using dicobalt
octacarbonyl as the CO source, in the presence of palladium acetate and either Xantphos or CataCXium
7
3
A as the ligand. The reaction was normally conducted under microwave irradiation in the presence
o
of DMAP, in 1,4-dioxane at 75 - 80 C and delivers the final compounds (33-36, 40-44, 46-47 and 53-
54) in low to moderate yield. Indole 57 was prepared as its N-tosyl derivatives 82 and the final inhibitor
o
was obtained by deprotection with cesium carbonate in MeOH / THF at 70 C.
ACS Paragon Plus Environment

Page 25 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
HN
O
O
O
HN
O
x
i-iii
O
O
O
HN
HN
N
Br
Br
Br
6
7
60
61
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
Cl
xi
For R = H
iv-v
For R =
vi
Cl
Cl
HN
O
HN
O
O
O
N
O
O
R3
N
R
Br
6
8-81
When:
R = H (64)
when: Cl
(62)
xii
R =
viii Cl
vii
Cl
HN
O
HN
O
HN
O
O
O
O
H
H
N
H
R3
N
N
N
N
N
R2
R2
O
O
O
33-36, 40-44, 46, 47,
53-54, 82
18, 20, 22-24,
65-67
16, 17, 19, 21,
25-27, 63
8
2
7
65, 67
28, 29
63
xiii
ix
ix
5
30
o
o
i. NaH, DMF/THF (1:1) 0 C, 30 min, then BnBr, 0 C to rt, 2 h (89%); ii. LiOH, THF/H O/MeOH (1:1:1),
rt, 2 h (78%); iii. (COCl) , DMF (0.5 mol%), DCM, rt, 2 h, then MeNH (2M in THF), Et N, THF, 0 C, 1.5 h
(95%); iv. Pd(OAc) (10 mol%), dppb (10 mol%), Et N, EtOH, DMSO, CO, 100 C, 50 psi, 16 h (94%); v.
NaOH, MeOH/THF (1:1), rt, 2 h (80%); vi. Pd(OAc) (5 mol%), Xantphos (10 mol%), Et N, toluene, 80
C, 20 min, then 2,4,6-trichlorophenyl formate 80 C, 2.5 h (87%); vii. R NH , HATU, DIPEA or Et N,
DMF or DCM, rt, 2 h; viii. R NH , DMAP (0.2 eq.), Et N, THF, 45 C, 3 h; ix. TFA, DCM, rt, 1 h; x. MeNH
(2M in THF), THF, reflux, 23 h (44%); xi. K CO , R CH Br, MeOH or DMF, 65 – 90 C,; xii. cPrNH ,
Co (CO) , Pd(OAc) , xantphos or cataXium A, DMAP, 1,4-dioxane, µw 75 – 80 C; xiii. Cs CO ,
2
o
2
2
3
o
2
3
2
3
o
o
2
2
3
2
o
2
3
3
2
o
2
3
2
2
o
2
8
2
2
3
o
MeOH/THF (1:2), 70 C, 1 h (41%).
2
Scheme 1. Divergent synthetic routes for the variance of R and R3
Variations of this synthetic route were followed for additional analogues 37-39, 48-52, 55-56,
8-59 as detailed in Scheme 2. The key advantage to this alternative approach was the use of pyridone
5
O-alkylation as a protecting group strategy to allow preparation of advanced amides 86-88, allowing
3
for a late stage introduction of the R “shelf” substituent either by Mitsunobu or alkylation chemistry.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 101
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
Overall, this late stage diversification was attractive for analogue synthesis, especially where the shelf
substituents required bespoke preparation. Additional late stage functional group interconversions
were required to afford alcohol 48 and morpholine 49 as detailed in Scheme 2. Whilst this route was
3
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
excellent for R diversification it is less step efficient than the Scheme 1 route for scale up.
Finally, separate synthetic routes were utilised to prepare 6-methylpyridone 32 and benzylic-
methyl derivative 31. Their preparation is outlined in Scheme 3 and 4. 6-Methylpyridone 32 was
7
4
prepared from known intermediate 94 in 4 steps using similar chemistry to that already discussed.
On the other hand, pyran 96 was used as a novel start point for the preparation of pyridone 31, and
allowed the commercial (R)-1-phenylethanamine to be used as the source of the chiral methyl group.
Furthermore, selective bromination of the resultant pyridone 97 provided the functional handle to
allow access to the carbonylation chemistry already discussed.
ACS Paragon Plus Environment

Page 27 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
HO
MeO
O
O
O
HN
RO
O
i-iii
iii
O
R = Bn
N
HN
N
Br
Br
Br
R = Bn, 85
6
0
R = Me, 84
83
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
HN
O
HN
O
xv-xvi
iv-vi
O
O
H
HN
O
HN
N
R2
O
O
2
R = cPr, 87
86
2
R = cBu, 88
vii or
viii
xvii-xix
HN
O
HN
O
O
O
H
N
H
R3
N
R3
N
N
R2
O
O
45
37-39, 50-51,
55, 58, 89-93
89
90
ix-x
xiii
xi-xii
4
8
49
91, 92
2, 59
93
56
xiv
5
Reagents and conditions: i. BnBr, Ag CO , CHCl , reflux, 16 h (67%); ii.LiOH, THF/MeOH (1:1), rt, 2 h
2
3
3
(
94%); iii. (COCl) , DMF (0.5 mol%), DCM, rt, 2 h, then MeNH (2M in THF), THF or 2-MeTHF, rt, 6 h (54-
2
2
o
9
5%); iv. Pd(OAc) (5 mol%), Xantphos (10 mol%), Et N, toluene, 80 C, 20 min, then 2,4,6-
2
3
o
2
o
trichlorophenyl formate 80 C, 2-18 h (57%); v. R NH , Et N, DMAP (5 mol%), THF, 45 C, 45 min – 16
2
3
o
o
h (85-86%); vi. TFA, 80 – 90 C, 1 h (83-99%); vii. CMTP, toluene, µw 120 C, or DIAD, PPh , toluene, rt,
3
o
o
3
-24 h viii. K CO , THF or DMF, rt – 50 C; ix. MeMgBr (3M in Et O), THF, -78 C, 45 min (43%); x. chiral
2
3
2
o
separation (Chiralpak AD-H); xi. DMP, DCM, 0 C – rt, 48 h (quant.); xii. morpholine, Et N, DCM, rt, 45
3
min, then STAB, rt 24 h (25%); xiii. TFA, DCM, rt, 30-90 min (80%) or HCl, IPA, rt, 3 days (48%);
o
xiv.Cs CO , MeOH/THF (1:2), 70 C, 1 h (68%); xv. Pd(OAc) (5 mol%), Xantphos (6 mol%), Et N, MeOH,
2
3
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 101
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
o
o
DMF, 60 C, 21 h (60%); xvi. NaI, MeCN, rt, 16 h (25%); xvii. K CO , DMF, 90 C, 1 h (67%); xviii. LiOH,
2
3
2
1
,4-dioxane / water (1:1), rt, 30 min (90%); xix. R NH , HATU, DIPEA, DMF, rt, 7 h (quant).
2
Scheme 2. Alternative routes for the preparation of pyridone analogues
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
O
O
HN
O
HN
O
i-ii
O
iii-iv
O
O
H
HN
O
N
O
N
N
Me
O
Me
95
O
Me
32
O
94
o
o
Reagents and conditions:i. K CO , BnBr, DMF, 40 C, 4 h (35%); ii. MeNH (2M in THF), MeOH, 50 C, 30
2
3
2
min (82%); iii. LiOH, 1,4-dioxane / water (1:1), rt, 1 h 45 min (99%); ,iv. cPrNH , HATU, DIPEA, DMF, rt,
2
1 h (74%).
Scheme 3. Preparation of 6-methylpyridone 32
HN
O
O
O
HN
O
O
O
O
i
ii-iii
iv-vi
O
O
O
O
H
N
N
N
N
Br
O
96
97
98
31
Reagents and conditions: i. (R)-1-phenylethanamine, DMF / THF (1:4), rt, 30 min, then EDC, DMAP (6
o
mol%), rt, 64 h (75%); ii. NBS, 2-MeTHF, rt, 15 min; iii. MeNH (2M in THF), MeOH, 50 C, 90 min (78%,
2
o
2 steps); iv. Pd(OAc) (10 mol%), Xantphos (7 mol%), Et N, MeOH, DMF, 60 C, 4 h (60%); v. NaOH,
2
3
MeOH / H O, rt, 16 h (90%); vi. cPrNH , HATU, DIPEA, DMF, rt, 1 h (89%).
2
2
Scheme 4. Homochiral preparation of pyridone 31
CONCLUSION
In summary, we have described the lead optimisation of a weak fragment-like hit into a range of
potent pan-BD2 selective inhibitors, possessing excellent physicochemical properties and with
selectivities from 200 - >1000 fold. Analogues 20 (GSK620) and 56 (GSK549) were selected for further
profiling and as shown in Table 7 both compounds show agreement in their selectivity for all BET
ACS Paragon Plus Environment

Page 29 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
family members (BRD2, BRD3, BRD4 and BRDT BD2 vs BD1), with indole 56 maintaining its extremely
high selectivity against BD1 for the other family members. This illustrates the high sequence homology
between the BD2 domains of these 4 proteins. Both 20 and 56 were evaluated using a BET surface
7
5
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
plasmon resonance (SPR) assay. Here, the same mutant protein constructs are used as for the FRET
assay, but binding affinity can be assessed. Pleasingly, both compounds were similarly potent and
selective as observed in the FRET assay. The selectivity of 56 was also evaluated against the wider
bromodomain family using the DiscoverX BromoScan screening platform. As illustrated by Figure 6
and Supplementary Table S12, 20 is highly selective for the BET-BD2 family of proteins with >200 fold
selectivity over all other bromodomains. Both 20 and 56 have also been profiled against a range of
known liability targets as shown in Supplementary Table S13 and both show minimal activity against
all targets tested. 20 is also Ames76 and mouse lymphoma assay (MLA)77-78 negative, whilst both
compounds showed low levels of inhibition against indicators of hepatoxicity such as the bile salt
export pump (BSEP) assay (pIC50 <3.7 for 20 and 4.3 for 56).79 They were also negative in the
8
0
glutathione (GSH) trapping assay and both showed no toxicity in the cell health evaluation of hepG2
cells against three end points. Both 20 and 56 do not have a CYP3A4 metabolism-dependant inhibition
(MDI) or direct inhibition liability as tested in house and for 20 this was confirmed in an external time-
dependant inhibition (TDI) assay at high concentration.81-82 These data are particularly encouraging
for the unsubstituted indole as this is often implicated in toxicities caused by reactive metabolites,83
though at least in our in vitro screens, this has not been observed to date.
Table 7. Profile of Lead Compounds 20 (GSK620) and 56 (GSK549).
Compound
HN
O
HN
O
O
O
H
H
N
N
N
N
HN
O
O
20
56
ACS Paragon Plus Environment