Methylpyrrole inhibitors of BET bromodomains

Article abstract of DOI:10.1016/j.bmcl.2017.02.057

An NMR fragment screen for binders to the bromodomains of BRD4 identified 2-methyl-3-ketopyrroles 1 and 2. Elaboration of these fragments guided by structure-based design provided lead molecules with significant activity in a mouse tumor model. Further mo

Full text of DOI:10.1016/j.bmcl.2017.02.057

Accepted Manuscript
Methylpyrrole inhibitors of BET bromodomains
Lisa A. Hasvold, George S. Sheppard, Le Wang, Steven D. Fidanze, Dachun
Liu, John K. Pratt, Robert A. Mantei, Carol K. Wada, Robbert Hubbard, Yu
Shen, Xiaoyu Lin, Xiaoli Huang, Scott E. Warder, Denise Wilcox, Leiming Li,
F. Greg Buchanan, Lauren Smithee, Daniel H. Albert, Terrance J. Magoc, Chang
H. Park, Andrew M. Petros, Sanjay C. Panchal, Chaohong Sun, Peter Kovar,
Nirupama B. Soni, Steven W. Elmore, Warren M. Kati, Keith F. McDaniel
PII:
S0960-894X(17)30200-7
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.02.057
BMCL 24729
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
5 January 2017
17 February 2017
23 February 2017
Please cite this article as: Hasvold, L.A., Sheppard, G.S., Wang, L., Fidanze, S.D., Liu, D., Pratt, J.K., Mantei, R.A.,
Wada, C.K., Hubbard, R., Shen, Y., Lin, X., Huang, X., Warder, S.E., Wilcox, D., Li, L., Buchanan, F.G., Smithee,
L., Albert, D.H., Magoc, T.J., Park, C.H., Petros, A.M., Panchal, S.C., Sun, C., Kovar, P., Soni, N.B., Elmore, S.W.,
Kati, W.M., McDaniel, K.F., Methylpyrrole inhibitors of BET bromodomains, Bioorganic & Medicinal Chemistry
Letters (2017), doi: http://dx.doi.org/10.1016/j.bmcl.2017.02.057
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Graphical Abstract
Leave this area blank for abstract info.
Methylpyrrole inhibitors of BET
bromodomains
Lisa A. Hasvold*, George S. Sheppard*, Le Wang, Steven D. Fidanze, Dachun Liu, John K. Pratt, Robert A.
Mantei, Carol K. Wada, Robert Hubbard, Yu Shen, Xiaoyu Lin, Xiaoli Huang, Scott E. Warder, Denise Wilcox,
Leiming Li, F. Greg Buchanan, Lauren Smithee, Daniel H. Albert, Terrance J. Magoc, Chang H. Park, Andrew
M. Petros, Sanjay C. Panchal, Chaohong Sun, Peter Kovar, Nirupama B. Soni, Steven W. Elmore, Warren M.
Kati and Keith F. McDaniel

Bioorganic & Medicinal Chemistry Letters
journal homepage: www.els evier.com
Methylpyrrole inhibitors of BET bromodomains
Lisa A. Hasvolda*, George S. Shepparda*, Le Wanga, Steven D. Fidanzea, Dachun Liua, John K. Pratta,
Robert A. Manteia, Carol K. Wadaa, Robbert Hubbardb, Yu Shena, Xiaoyu Lina, Xiaoli Huanga, Scott E.
Wardera, Denise Wilcoxa, Leiming Lia, F. Greg Buchanana, Lauren Smitheeb, Daniel H. Alberta, Terrance J.
Magocb, Chang H. Parka, Andrew M. Petrosa, Sanjay C. Panchala, Chaohong Suna, Peter Kovara, Nirupama
B. Sonia, Steven W. Elmorea, Warren M. Katia and Keith F. McDaniela
aAbbVie Inc., Oncology Discovery, 1 North Waukegan Rd., North Chicago, IL 60064, USA
bFormer AbbVie employee
ARTICLE INFO
ABSTRACT
Article history:
Received
An NMR fragment screen for binders to the bromodomains of BRD4 identified 2-methyl-3-
ketopyrroles 1 and 2. Elaboration of these fragments guided by structure-based design provided
lead molecules with significant activity in a mouse tumor model. Further modifications to the
methylpyrrole core provided compounds with improved properties and enhanced activity in a
mouse model of multiple myeloma.
Revised
Accepted
Available online
2009 Elsevier Ltd. All rights reserved
.
Keywords:
BET proteins
Bromodomain
Fragment Screening
O
O
O
The BET (Bromodomain and Extraterminal) family of
proteins, consisting of BRD2, BRD3, BRD4 and BRDT, has
emerged as a promising epigenetic target for drug discovery.1
These proteins act as epigenetic ‘readers’ that recognize
acetylated lysines on the N-terminal tails of histones and recruit
transcription factors to chromatin. This recognition occurs
HN
O
HN
H2N
through interaction with
a
structural motif called
a
bromodomain,2 with each BET protein containing two
bromodomains. Many of the interacting transcription factors and
their target genes are known to play roles in the pathogenesis of
cancer and inflammation. Accordingly, small molecules which
bind to the bromodomains present in the BET family proteins
have shown activity in a wide range of preclinical disease
models, and multiple compounds have advanced into clinical
trials for the treatment of cancers.3
1
2
Ki = 21 µM (BEI 22)
Ki = 25 µM (BEI 21)
Figure 1. NMR fragment hits 1 and 2.
In an effort to discover novel BET bromodomain inhibitors,
we undertook a fragment-based approach using NMR screening4
with a TR-FRET competitive binding assay to confirm binding.
A number of structural types which had not been previously
reported to bind to bromodomains were identified.5 Two such
fragment hits were the 2-methyl-3-keto pyrroles 1 and 2, which
showed a large degree of spectral perturbation,4 moderate binding
affinities and high binding efficiency indices (BEI)6 (Figure 1).
Figure 2. X-ray co-crystal structures of fragments 1 and 2 in the BRD4 BDII
protein. Residues Pro378 and Val379 are not displayed for clarity. Dotted
lines represent H-bonds predicted by Schrödinger Maestro.
The X-ray co-crystal structures of compounds 1 (PDB code:
5UEV) and 2 (PDB code: 5UEP) (Figure 2) with bromodomain II
of BRD4 revealed several key interactions between the ligands

and the protein which consists of four α-helices (αZ, αA, αB and
αC) and two loops, one linking helices αZ and αA (ZA loop) and a
second linking helices αB and αC (BC loop).7 The compounds
bind in the acetyl lysine recognition site of the bromodomain,
with the 3-keto functionality of both compounds accepting a
hydrogen bond from the Asn433 sidechain amide. The 2-methyl
group on each pyrrole ring resides in an ordered water network
forming an amphoteric pocket. In addition to the Asn433
interaction, the 3-keto functionality of each compound also
hydrogen bonds with a nearby water molecule in the amphoteric
pocket. In each structure the pyrrole NH interacts with the
Pro375 carbonyl group. For compound 1, a water molecule
bridges between the aniline and the amide backbone of the ZA
loop.
selectivity versus non-BET bromodomains.10 Additionally, there
is an extended cleft between the ZA loop and the adjacent helix
that neither compound fully occupies. By placing substituents on
the phenyl ring at the appropriate positions with vectors directed
toward these unoccupied regions of the protein, it may be
possible to gain additional interactions with the protein and
improve potency. Several compounds related to fragments 1 and
2 were prepared to investigate these areas of the protein (Table
1).
O
OCH3
N
N
N
N
S
Co-crystal structures of related analogs provided additional
insight into the binding of the fragments. Compounds 3 and 4,
containing a primary amide in place of the methyl ketone moiety
in 1, displayed similar binding affinity (Figure 3); however, a
second hydrogen bond between the amide and the Asn433
residue of the BRD4 BDII protein was observed for compound 3
(PDB code: 5UET) (Figure 4). Replacing the ester moiety of
compound 2 with the 2-aniline moiety (compound 5) provided a
small improvement in activity (Figure 3).
Cl
6
Ki = 0.017 µM
O
O
O
NH2
NH2
HN
H2N
HN
HN
H2N
ZA loop
ZA cleft
3
4
5
Ki = 12 µM
Ki = 22 µM
Ki = 9.5 µM
Figure 3. Compounds 3, 4 and 5.
WPF shelf
Figure 5. X-ray co-crystal structure overlay of fragment 1 (orange) and
compound 6 (MS417) (purple) in the BRD4 BDII protein.
Compounds were evaluated using a binding assay and two
cellular assays.
Binding to a construct containing both
bromodomains of BRD4 (K57-K550) was determined using a
TR-FRET assay as described in the supplementary material.
Binding determinations to other bromodomains of the BET
family showed no meaningful selectivity within the family; these
data are reported for key compounds 15, 25 and 38 in the
supplementary material.
Single concentration screening
experiments against selected non-BET bromodomains are also
reported for 15, 25 and 38 in the supplementary material and
indicate low affinity for most of the non-BET bromodomains
examined. Cellular proliferation was measured using the triple
negative breast cancer cell line MX-1 (ATCC) in a 3-day
proliferation assay using the Cell Titer Glo assay kit (Promega).
Target engagement in cells was measured with a luciferase
reporter assay based on the contribution of BRD4 to human
papilloma virus (HPV) E2-mediated transcriptional repression,
where BRD4 is part of the HPV long control region (LCR)
promoter repression complex with E2 and EP400.11 An H1299-
derived cell line was generated with a stably integrated E2
expression cassette and an HPV-LCR-driven luciferase reporter.
A luciferase signal occurs only when a BET inhibitor blocks
BRD4, thus providing a measure of target engagement. Good
Figure 4. X-ray co-crystal structure of compound 3 in the BRD4 BDII
protein. Residues Pro378 and Val379 are not displayed for clarity. Dotted
lines represent H-bonds predicted by Schrödinger Maestro.
An overlay of the X-ray co-crystal structures of compound 1
and known BET inhibitor 6 (MS417)8 (PDB code: 5UEU) shown
in Figure 5 revealed additional exploitable binding sites within
the protein, such as the WPF shelf, which is comprised of
residues W374, P375 and F376.9 The WPF shelf is occupied by
the 4-chlorophenyl moiety of compound 6, but is not occupied by
compound 1. Occupation of the WPF shelf is reported to provide

correlation was observed between the two cellular assays, an
indication that cell killing is due to target engagement.
binding activity for compound 15 compared to fragment 2.
Sizable improvements in activity were also achieved in both the
cellular proliferation and luciferase reporter assays. Simple ester
(compound 16) and amide (compound 17) moieties were also
investigated at the same position. The methyl ester showed
decreased TR-FRET activity compared to the sulfonamide
moiety (91 nM vs. 33 nM), and an even more significant loss of
TR-FRET activity was observed with the ethyl amide moiety
(250 nM vs. 33 nM). Cellular potency also declined for both the
methyl ester and ethyl amide compounds.
A co-crystal structure of compound 15 (PDB code: 5UER)
bound to BRD4 BDII (Figure 7) was also obtained. As with
compound 9, the phenoxy group resides in the WPF pocket. An
additional interaction was also achieved with one oxygen atom of
the sulfonamide group being hydrogen bonded to the Asp381
backbone amide of the ZA loop. Surprisingly, in the fully
elaborated molecule 15, the isoindolone ring system was rotated
compared to the X-ray co-crystal structure observed for fragment
2. The rotated ring system retains the interaction between the
To access the WPF shelf region, phenylamino and phenoxy
groups were placed on the phenyl ring of compound 3. It can be
seen in Table 1 that a significant gain in binding activity was
observed upon addition of phenylamino (compound 7, 50-fold) or
phenoxy substituents (compound 8, 80-fold; compound 9, 50-
fold). The phenoxy substituents provided greater cellular
potency in both the MX-1 and luciferase target engagement
assays than the phenylamino moiety. The X-ray co-crystal
structure of compound 9 (PDB code: 5UES) confirmed that the
phenoxy group occupies the WPF shelf. It also indicates
retention of the bidentate interaction between the primary amide
and Asn433, the pyrrole NH interaction with Pro375 and the
placement of the 2-methyl group among the waters in the
amphoteric pocket (Figure 6).
In an effort to further enhance potency by reestablishing the
aniline interactions with Pro375 and the amide backbone of the
ZA loop, compound 10 which incorporates both the 2-amino and
6-phenoxy moieties was examined. This combination proved to
be less active in the TR-FRET assay and both cellular assays
compared to compound 8. The amino moiety was then moved to
the adjacent carbon, situating it para to the phenoxy moiety to
provide a different vector for access to the ZA loop and adjacent
cleft. Some activity was regained by the repositioning of the
amine (compound 11) and with incorporation of the
methanesulfonamide moiety (compound 12), but neither
compound surpassed the potency of compound 8 with the single
phenoxy substituent.
isoindolone carbonyl group and Asn433.
However, the
methylene groups alpha and beta to the carbonyl group now
occupy nearly the same space in the amphoteric pocket that the
methyl group on the pyrrole ring occupied in all of the previously
obtained X-ray co-crystal structures for the 2-methyl-3-keto
pyrroles 1 and 2 and analogs containing those cores. An overlay
of compounds 2 and 15 shown in Figure 8 illustrates how these
compounds occupy the BRD4 BDII protein relative to each other.
Figure 7. X-ray co-crystal structure of compound 15 in the BRD4 BDII
protein. Dotted lines represent H-bonds predicted by Schrödinger Maestro.
Figure 6. X-ray co-crystal structure of compound 9 in the BRD4 BDII
protein.
Concurrently, similar substitutions on compound 5 were
investigated. Similar to the potency gains observed with the
fragment 1 structure, a greater than 20-fold gain in TR-FRET
activity was obtained by replacement of amino with phenoxy on
the compound 5 isoindolone structure (compound 13). Cellular
potency was substantially improved as well. In this case, a
modest improvement in binding affinity was observed upon the
addition of an amino group para to the phenoxy moiety
(compound 14) and an even larger 13-fold improvement in
binding affinity was obtained when the methanesulfonamide
moiety was added (compound 15). Together these changes
combined to provide an overall 760-fold gain in TR-FRET

vehicle
100 mg/kg, PO, QD
(53% TGI)
3,000
2,000
1,000
0
100 mg/kg/day, PO, BID
(73% TGI)
Rx
-7
0
7
14
21
28
days post initiation of dosing
Figure 9. Tumor growth inhibition data for compound 15 in an OPM-2
xenograft model of multiple myeloma. Values represent mean SE (n = 10
per group).
A metabolite identification study of compound 15 in both
mouse and rat liver microsomes revealed the saturated
isoindolone ring and the terminal phenoxy ring to be the primary
sites of metabolism. As a possible means for improving activity
and ADME properties, a variety of additional fused bicyclic
pyrrole cores capable of producing a bidentate interaction with
Asn433 similar to that observed with the amide-substituted
monocyclic pyrrole core were prepared (Table 3). Cores III-VI
introduced 1-2 nitrogen atoms into the fused ring and varied the
ring saturation. For core VII, the nitrogen atom of core IV was
moved over 1 atom, placing the methyl group on nitrogen instead
of on carbon. In addition to the core changes, sulfonamide
replacements that would potentially retain the ZA loop backbone
Figure 8. Overlay of X-ray co-crystal structures of 2 (blue) and 15 (orange)
in the BRD4 BDII protein.
Among these early compounds, compound 15 stood out as
having superior activity both in the TR-FRET binding assay and
in the cellular assays. Mouse pharmacokinetic information for
compound 15 is reported in Table 2. Compound 15 demonstrated
high unbound i.v. clearance at 84 L/hr/kg and low unbound oral
AUC of 0.09 µM*hr with 30% oral bioavailability. Though
improvement in pharmacokinetics would be necessary, exposure
was sufficient to take compound 15 forward into a xenograft
tumor model for proof of concept studies.
Prior to the initiation of the xenograft model, a study to
determine the maximum tolerated dose (MTD) indicated that
compound 15 was well-tolerated at 100 mg per kg per day (mkd).
Using this as a guide for dose selection, compound 15 was then
evaluated in an OPM-2 mouse xenograft model of multiple
myeloma dosed orally once a day (QD) at 100 mkd or twice a
day (BID; 50 mkd x 2) over 14 days (Figure 9). Administration
of the same total dose in a twice daily fashion provided improved
tumor growth inhibition (TGI) of 73% compared to 53% TGI
observed for daily QD dosing during the treatment period, due to
longer coverage of the target. Both dosing regimens were well-
tolerated. The moderate TGI observed confirmed the use of
pyrrole-based BET inhibitors as possible cancer therapeutics and
prompted further optimization of these cores to discover BET
inhibitors with improved potency and ADME properties.
interaction were also investigated.
Additionally, 2,4-
difluorophenoxy was investigated as a phenoxy replacement to
block sites of metabolism and perhaps reduce the high unbound
clearance observed in liver microsomes and in mouse
pharmacokinetic experiments.
Prototypical synthetic schemes for the seven core types are
shown in Schemes 1-7. Halogen substituted cores I, IV and V
were prepared from the common starting material 41 and then
coupled with the lower aryl piece via Suzuki-Miyaura cross
coupling (Schemes 1, 4, and 5). Similarly, core VII was prepared
through coupling of the iodoazaindole intermediate 67 (Scheme
7). Core II was prepared in a similar manner through elaboration
of pyrrole ketone 2 (Scheme 2).12 Alternatively, C-H activation
of the pyrrole ring13 was used for the coupling of core III to the
lower aryl piece (Scheme 3). Construction of the fused rings in
cores III, IV and V utilized formylation at the 4-position of the
pyrrole, achieved by temporary blocking of the 5-position
through TIPS protection of the pyrrole nitrogen, followed by
formation of the fused ring (Schemes 3, 4 and 5). In some cases,
the phenoxy moiety was introduced after cross coupling through
nucleophilic aromatic substitution. Core VI was assembled onto
the lower aryl piece rather than coupling the completely formed
core with the lower aryl piece (Scheme 6). Experimental
procedures and spectral data for select compounds can be found
in the supplementary material.

Scheme 1. Reagents and conditions: a) LiOH, 1,4-dioxane/H2O,
73%; b) HATU, NH3, DIPEA, 1,4-dioxane, DMF, 73%; c) NIS,
THF, 71%; d) Pd2(dba)3, 1,3,5,7-tetramethyl-8-phenyl-2,4,6-
trioxa-8-phosphaadamantane (PA-Ph), K3PO4, THF/H2O, 60 °C,
37%; e) (i) CH3SO2Cl, DIPEA, DCM; (ii) NaOH (1 M aq.), THF,
50 °C, 70%.
Scheme 4. Reagents and conditions: a) ethyl 2-isocyanoacetate,
NaH, THF, 70 °C, 35%; b) NBS, THF, -78 °C, 78%; c) (2-
phenoxyphenyl)boronic acid, Pd2(dba)3, PA-Ph, Na2CO3, 1,4-
dioxane/H2O, 60 °C, 63%; d) LiOH, THF/MeOH/H2O, 50 °C,
98%; e) CuO, 1,10-phenanthroline, NMP/quinoline, 190 °C
microwave; f) (i) TFA, RT; (ii) NaOAc, CH3CN/H2O, 50 °C,
15% e-f.
O
O
HO
OH
B
F
a-c
d
+
HN
HN
O2N
CO2Et
2
Br
44
45
O
O
O
HN
HN
HN
f,g
e
O
F
O
O
O
S
N
O2N
O2N
H
46
47
15
Scheme 2. Reagents and conditions: a) LiOH, THF/H2O,
81%; b) HCl, EtOH/H2O, heat, 62%; c) NBS, THF, -78 °C,
76%; d) Pd2(dba)3, PA-Ph, K3PO4, THF/H2O, 60 °C, 71%; e)
phenol, Cs2CO3, DMF, 110 °C, 85%; f) Fe, NH4Cl,
THF/EtOH/H2O, reflux, 95%; g) (i) CH3SO2Cl, Et3N, DCM;
(ii) NaOH (1 M aq.), 1,4-dioxane, 50 °C, 86%.
Scheme 5. Reagents and conditions: a) H2NNH2, H2O, EtOH,
68%; b) NBS, DMF/THF, 54%; c) Pd2(dba)3, PA-Ph, K3PO4,
THF/H2O, 60 °C, 85%; d) 2,4-difluorophenol, Cs2CO3, DMF, 50
°C, 66%; e) H2, Pd/C, EtOAc, 73%; f) (i) CH3SO2Cl, Et3N,
DCM; (ii) NaOH (1 M aq.), THF/1,4-dioxane, 75%.
Scheme 3. Reagents and conditions: a) NaH, TIPSCl, THF,
88%; b) oxalyl chloride, DMF, DCM, 0 °C, 74%; c) NaH,
SEMCl, THF, 89%; d) NH4OAc, CH3NO2, 90 °C, 85-95%; e) H2,
washed Raney nickel, THF, 67-95%; f) LiOH, 95% EtOH, 120
°C, 60-77%; g) allyl PdCl dimer, KOAc, DMA, 130 °C, 38-43%;
h) Cs2CO3, DMSO, 80 °C, 40%; i) (i) TFA; (ii) K2CO3, MeOH,
87%.

H2N
CN
34) consistent with the suggestion from the binding mode of 15
that the hydrogen bond between the pyrrole NH and Pro375 does
not substantially contribute to the activity. For cores III, V and
VI, analogs incorporating exact comparisons were made to
evaluate vs. compound 15. Compounds 23 (core III) and 28
(core V) have similar binding activity compared to compound 15,
but compound 33 (core VI) has diminished activity (Ki = 150
nM) compared to compound 15. In the cellular assays,
compounds 23 and 28 appear to be slightly more active in the
MX-1 proliferation assay, but have somewhat diminished activity
in the luciferase reporter assay compared to compound 15. As
seen with the binding assay results, compound 33 is less potent
than compound 15 in both the MX-1 proliferation assay and the
luciferase reporter assay. Core III does, however, have the
advantage with regard to unbound intrinsic clearance in mouse
liver microsomes (<7.1 L/hr/kg for compound 23 vs. 55 L/hr/kg
for compound 15). Two additional sulfonyl-containing moieties,
ethanesulfonamide (compound 36) and methylene-linked
methylsulfone (compound 40) were incorporated into compounds
of core VII. Both replacements were well-tolerated showing
similar or slightly improved activity in the TR-FRET assay
compared to the typical methanesulfonamide moiety (Ki = 11 nM
for compound 36 and Ki = 11 nM for compound 40 vs. Ki = 17
nM for compound 35).
CHO
CHO
b
F
a
O
O
O2N
O2N
O2N
61
62
63
CO2Et
CN
HN
O
c
d
63
+
O
CO2Et
O2N
64
65
O
O
OEt
NH
NH
HN
HN
CH3CO2H
NH2
NH2
O
N
HCl
e
+
O
O2N
O2N
66
67
68
O
O
NH
NH
HN
HN
N
O
N
f
g
O
O
O
S
N
H2N
69
33
H
Scheme 6. Reagents and conditions: a) phenol, K2CO3, 1,4-
dioxane, 70 °C, 99%; b) NaCN, NH4Cl, NH4OH, MeOH, 28%; c)
p-TsOH, toluene, 80 °C, 99%; d) NaOEt, EtOH, 44%; e) EtOH,
80 °C, 38%; f) H2, Pd/C, MeOH, 78%; g) (i) CH3SO2Cl, Et3N,
DCM; (ii) K2CO3, 1,4-dioxane, 50 °C, 56%.
Scheme 7. Reagents and conditions: a) NIS, DMF, 93% ; b)
NaH, CH3I, DMF, 95%; c) Pd2(dba)3, PA-Ph, K3PO4, THF/H2O,
60 °C, 88%; d) 2,4-difluorophenol, Cs2CO3, DMSO, 110 °C, 94-
99%; e) 4 M HCl, 1,4-dioxane, 71-87%.
Figure 10. X-ray co-crystal structure of compound 36 in the BRD4 BDII
protein. Dotted lines represent H-bonds predicted by Schrödinger Maestro.
Additional X-ray co-crystal structures were obtained to
determine whether fully elaborated molecules based on the new
cores would bind to the BRD4 BDII protein in the ring flipped
conformation observed for the isoindolone ring system of
compound 15 or if they would retain the original binding
conformation observed for fragment 2, which would allow a
bidentate interaction with ASN433. It can be seen in Figures 10
and 11 that fully elaborated compounds 36 (core VII) (PDB code:
5UEO) and 29 (core V) (PDB code: 5UEQ) both bind to the
BRD4 BDII protein with the original conformation observed for
fragment 2, with each making the intended bidentate Asn433
interaction. Among the X-ray co-crystal structures obtained,
compound 15 was the only compound for which rotation of the
fused bicyclic ring system was observed.
Table 3 summarizes the activity and ADME data for the new
cores compared to the original pyrrole and isoindolone cores. In
general, for the compounds of Table 3 and others not shown,
compounds based on cores III-V showed similar or diminished
activity compared to the isoindolone core II-based compounds in
both the TR-FRET binding assay and the cellular assays. It was
also typically found for cores I-V and VII that replacement of -
NHSO2Me with -SO2Me and replacement of phenoxy with 2,4-
difluorophenoxy resulted in a loss of activity in both the TR-
FRET and cellular assays. Core VII, which differs from the other
cores by the lack of hydrogen bonding capability between the
pyrrole NH and Pro375 resulting from relocation of the pyrrole
nitrogen, stood out with improved activity compared to the
isoindolone core. It was also somewhat more active in the TR-
FRET assay compared to core IV (compare compounds 26 and

vehicle
3,000
2,000
1,000
0
25, 150 mg/kg, PO, QD
(68% TGI)
38, 30 mg/kg, PO, QD
(64% TGI)
*
*
**
Rx
-7
0
7
14
21
days post initiation of dosing
Figure 12. Tumor growth inhibition data for compounds 25 and 38 in an
OPM-2 xenograft model of multiple myeloma. Values represent mean SE
(n = 10 per group). * indicates animal removed from study.
In summary, the discovery of novel proprietary BET inhibitors
was initiated through an NMR fragment screen against BRD4.
Two ketopyrrole-based hits 1 and 2 were utilized as the basis for
hit-to-lead activities. Structurally-enabled optimization efforts of
ketopyrrole fragment 2 to compound 15 improved the TR-FRET
activity nearly 760-fold (25 µM vs. 33 nM). Isoindolone 15
exhibited moderate mouse oral bioavailability (30%) and in vivo
tumor growth inhibition in an OPM-2 xenograft model of
multiple myeloma. Further core optimization led to compounds
with TR-FRET activity improved in excess of three log units for
select examples relative to compound 2. Gains were also
achieved in cellular potency and ADME properties allowing for
Figure 11. X-ray co-crystal structure of compound 29 in the BRD4 BDII
protein. Dotted lines represent H-bonds predicted by Schrödinger Maestro.
Three compounds, 25, 38 and 39, were further evaluated in
mouse pharmacokinetic studies (Table 4). Although high
potency is very desirable, elements such as structural diversity
and ADME properties also played a large role in the selection
process for pharmacokinetic experiments. Compound 25 is a
core III compound with the methylsulfonyl and 2,4-
difluorophenyl moieties that exhibited very low unbound intrinsic
clearance in mouse liver microsomes and good permeability.
Compound 25 demonstrated a 6-fold improvement in both
unbound i.v. clearance (14 L/hr/kg vs. 84 L/hr/kg) and unbound
oral AUC (0.53 µM*hr vs. 0.09 µM*hr) compared to compound
15. However, the oral bioavailability was nearly the same (32%
vs. 30%). Compound 38 also incorporates the methylsulfonyl
and 2,4-difluorophenyl moieties, but is a core VII compound with
greater potency and permeability than compound 25. It also
demonstrated improvements in both unbound i.v. clearance and
unbound oral AUC compared to compound 15, and had nearly
double the oral bioavailability (57% vs. 30%). Incorporation of
core VII with an ethanesulfonamide moiety and just a single
fluoro group at the 4-position of the phenoxy ring in compound
39 resulted in very good potency (Ki = 5.5 nM). Unfortunately,
this combination resulted in higher unbound i.v. clearance at 130
L/hr/kg and lower unbound oral AUC at 0.06 µM*hr compared to
compounds 15, 25 and 38. Blocking both the 2- and 4-positions
of the terminal phenoxy ring appears to be necessary to reduce
metabolism at that site (compound 39 vs. compound 36).
additional xenograft studies.
Compounds 38 and 25
demonstrated tumor growth inhibition of 64% and 68%,
respectively, in an OPM-2 model of multiple myeloma which
validates pursuit of BET inhibitors as potential cancer
therapeutics.
Acknowledgements and Disclosures
This research used resources at the Industrial Macromolecular
Crystallography Association Collaborative Access Team (IMCA-
CAT) beamline 17-ID, supported by the companies of the
Industrial Macromolecular Crystallography Association through
a contract with Hauptman-Woodward Medical Research Institute.
This research used resources of the Advanced Photon Source,
a U.S. Department of Energy (DOE) Office of Science User
Facility operated for the DOE Office of Science by Argonne
National Laboratory under Contract No. DE-AC02-06CH11357.
All authors are current or former AbbVie employees. AbbVie
was involved in the design, conduct and financial support for this
research. AbbVie participated in the interpretation of data,
review, and approval of the publication.
Since compounds 38 and 25 both achieved improved drug
exposure relative to 15, they were evaluated in the OPM-2 model
(Figure 12). The compounds when dosed QD displayed
improved tumor growth inhibition relative to 15 QD (64% and
68%, respectively). Tolerability was viewed as adequate with
minimal mean weight loss (1-3%) throughout the study for both
treatment groups.
Supplementary Material
Supplementary data associated with this article can be found,
in the online version, at
References
1) (a) Mujtaba, S.; Zeng, L.; Zhou, M-M. Oncogene 2007,
26, 5521. (b) Sanchez, R.; Zhou, M-M. Curr. Opin. Drug
Disc. Dev. 2009, 12, 659. (c) Shi, J.; Vakoc, C. R. Mol.
Cell 2014, 54, 728. (d) Fu, L. L.; Tian, M.; Li, X.; Li, J.
J.; Huang, J.; Ouyang, I.; Zhang, Y.; Liu, B. Oncotarget
2015, 6, 5501.

2) (a) Filippakopoulos, P.; Knapp, S. Nat. Rev. Drug Disc.
2014, 13, 337. (b) Dhalluin, C.; Carlson, J. E.; Zeng, L.;
He, C.; Aggarwal, A. K.; Zhou, M.-M. Nature 1999, 399,
491. (c) Dey, A.; Chitsaz, F.; Abbasi, A.; Misteli, T.;
Ozato, K. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 8758.
7) Dhalluin, C.; Carlson, J. E.; Zeng, L.; He, C.; Aggarwal,
A. K.; Zhou, M-M. Nature 1999, 399, 491.
8) Mitsubishi
Tanabe
Pharma
Corporation
WO
2009/2009084693.
9) Nicodeme, E.; Jeffrey, K. L.; Schaefer, U.; Beinke, S.;
Dewell, S.; Chung, C.-W.; Chandwani, R.; Marazzi, I.;
Wilson, P.; Coste, H.; White, J.; Kirilovsky, J.; Rice, C.
M.; Lora, J. M.; Prinjha, R. K.; Lee, K.; Tarakhovsky, A.
Nature 2010, 468, 1119.
3) (a) For a recent Perspectives article on bromodomain
inhibitors, see Romero, F. A.; Taylor, A. M.; Crawford,
T. D.; Tsui, V.; Cote, A.; Magnuson, S. J. Med. Chem.
2016, 59, 1271. (b) For a recent review on the clinical
development of BET bromodomain inhibitors, see
Theodoulou, N. H.; Tomkinson, N. C. O.; Prinjha, R. K.;
Humphreys, P. G. Curr. Opin. Chem. Biol. 2016, 33, 58.
10) (a) Filippakopoulos, P.; Knapp, S. FEBS Lett. 2012, 586,
2692. (b) Filippakopoulos, P.; Picaud, S.; Mangos, M.;
Keates, T.; Lambert, J-P.; Barsyte-Lovejoy, D.; Felletar,
I.; Volkmer, R.; Müller, S.; Pawson, T.; Gingras, A-C.;
Arrowsmith, C. H.; Knapp, S. Cell 2012, 149, 214.
4) Shuker, S. B.; Hajduk, P. J.; Meadows, R. P.; Fesik, S.
W. Science 1996, 274, 1531.
5) (a) For information on an additional core we identified
through the NMR fragment screen, see Wang, L.; Pratt, J.
K.; Soltwedel, T.; Sheppard, G. S.; Fidanze, S.; Liu, D.;
Hasvold, L. A.; Mantei, R. A.; Holms, J.; McClellan, W.;
Wendt, M.; Wada, C.; Frey, R.; Hansen, T. M.; Hubbard,
R.; Park, C.; Li, L.; Magoc, T.; Albert, D. H.; Lin, X.;
Warder, S. E.; Kovar, P.; Huang, X.; Wilcox, D.; Wang,
R.; Rajaraman, G.; Petros, A.; Hutchins, C.; Torrent, M.;
Panchal, S. C.; Sun, C.; Elmore, S. W.; Rosenberg, S.;
Shen, Y.; McDaniel, K. F.; Kati, W. Manuscript in
preparation. (b) Subsequent to our discovery and
11) Smith, J. A.; White, E. A.; Sowa, M. E.; Powell, M. L.
C.; Ottinger, M.; Harper, J. W.; Howley, P. M. Proc.
Natl. Acad. Sci. U.S.A. 2010, 107, 3752.
12) Clezy, P. S.; Fookes, C. J. R.; Mirza, A. H. Aust. J. Chem.
1977, 30, 1337.
13) Bheeter, C. B.; Bera, J. K.; Doucet, H. Tetrahedron Lett.
2012, 53, 509.
submission of
a patent application, another group
independently reported the discovery of 3-keto-2-
methylpyrrole binders of BET-bromodomains through a
virtual library screen. Lucas, X.; Wohlwend, D.; Hügle,
M.; Schmidtkunz, K.; Gerhardt, S.; Schüle, R.; Jung, M.;
Einsle, O.; Günther, S. Angew. Chem. Int. Ed. 2013, 52,
14055. (c) Recently another group independently reported
the discovery of bromodomain scaffolds through the use
of biosensor fragment screening.
Navratilova, I.;
Aristotelous, T.; Picaud, S.; Chaikuad, A.; Knapp, S.;
Filappakopoulos, P.; Hopkins, A. L. ACS Med. Chem.
Lett. 2016, 7, 1213.
6) Abad-Zapatero, C.; Metz, J. T. Drug Discovery Today
2005, 10, 464.
Table 1. SAR of fragment analogs.
MX-1 Proliferation
EC50 (nM)
Luciferase Reporter
(10% FBS) EC50 (nM)
Core
ID
R
TR-FRET Ki (nM)
22,000
4
>10,000
>10,000
7
8
450
270
6,900
970
2,800
450
9
410
1,200
>3,000
690
280
10
11
1,600
420
2,100
880
12
280
1,000
1,400
5
9,500
440
>10,000
1,400
7,100
280
13

14
15
210
33
230
160
300
51
16
17
91
1,200
1,400
2,000
1,400
250
TR-FRET BRD4 Ki values are reported as the geometric mean derived from 2 or more measurements. running MSR = 1.2 and test-retest MSR = 2.1.
EC50 values are reported as the mean derived from two measurements.
MX-1 proliferation assay: running MSR = 1.2 and test-retest MSR = 4.7.
Luciferase reporter assay: running MSR = 1.3 and test-retest MSR = 2.6.
Table 2. Mouse PK of compound 15.
I.V. unbound
P.O. unbound
I.V. t½
(hr)
F
(%)
ID
CLp (L/hr/kg) AUC (µM*hr)
15
84
0.09
0.6
30
I.V. dose: 3 mg/kg; P.O. dose: 10 mg/kg.

Table 3. Data comparison of cores I-VII.
I
II
III
IV
V
VI
VII
A
B
C
TR-FRET
Ki (nM)
MX-1
EC50 (nM)
Reporter 10% FBS
EC50 (nM)
Clint, u Mouse
(L/hr/kg)
Permeability
(x10-6 cm/s)
Core
ID
R
H
Ar
8
A
A
B
B
A
A
B
B
A
A
B
B
A
A
A
A
B
B
A
A
A
B
B
A
B
C
B
270
280
370
2,200
440
33
970
450
1,400
2,300
4,500
280
51
540
21
29
12
12
18
19
13
15
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
NHSO2Me
NHSO2Me
SO2Me
H
1,000
660
3,200
1,400
160
250
570
190
55
I
8.1
6.0
10,000
55
NA
NA
9.5
27
NHSO2Me
NHSO2Me
SO2Me
H
II
97
75
49
19
460
340
41
1,400
280
120
420
290
700
220
180
75
84
NA
NA
15
NA
<7.1
19
NHSO2Me
NHSO2Me
SO2Me
H
III
IV
84
230
900
220
180
59
8.7
22
300
450
37
<6.9
1,900
520
37
61
H
30
NHSO2Me
NHSO2Et
NHSO2Me
SO2Me
H
30
10
V
10
52
48
9.5
6.1
12
16
92
280
750
1,600
410
1,100
46
>15
37
77
390
1,200
460
580
38
1,100
150
120
17
>370
36
100
13
VI
NHSO2Me
H
6,400
29
13
NHSO2Me
NHSO2Et
SO2Me
SO2Me
NHSO2Et
CH2SO2Me
21
11
56
56
12
14
VII
17
68
91
22
48
98
110
21
150
47
12
35
5.5
11
20
32
20
40
24
40
TR-FRET BRD4 Ki values are reported as the geometric mean derived from 2 or more measurements: running MSR = 1.2 and test-retest MSR = 2.1.
EC50 values are reported as the mean derived from two measurements.
MX-1 proliferation assay: running MSR = 1.2 and test-retest MSR = 4.7.
Luciferase reporter assay: running MSR = 1.3 and test-retest MSR = 2.6.
Clint, u mouse: scaled unbound intrinsic clearance in mouse liver microsomes.

Table 4. Mouse PK data for select compounds.
I.V. unbound
CLp (L/hr/kg)
P.O. unbound
AUC (µM*hr)
I.V. t½
(hr)
F
(%)
ID
15
25
84
14
0.09
0.53
0.6
1.3
30
32
38
39
16
0.86
0.06
2.5
5.6
57
33
130
I.V. dose: 3 mg/kg; P.O. dose: 10 mg/kg.