From a novel HTS hit to potent, selective, and orally bioavailable KDM5 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2017.05.016

A high-throughput screening (HTS) of the Genentech/Roche library identified a novel, uncharged scaffold as a KDM5A inhibitor. Lacking insight into the binding mode, initial attempts to improve inhibitor potency failed to improve potency, and synthesis of

Full text of DOI:10.1016/j.bmcl.2017.05.016

Accepted Manuscript
From a novel HTS hit to potent, selective, and orally bioavailable KDM5 in-
hibitors
Jun Liang, Sharada Labadie, Birong Zhang, Daniel F. Ortwine, Snahel Patel,
Maia Vinogradova, James R. Kiefer, Till Mauer, Victor S. Gehling, Jean-
Christophe Harmange, Richard Cummings, Tommy Lai, Jiangpeng Liao,
Xiaoping Zheng, Yichin Liu, Amy Gustafson, Erica Van der Porten, Weifeng
Mao, Bianca M. Liederer, Gauri Deshmukh, Le An, Yingqing Ran, Marie
Classon, Patrick Trojer, Peter S. Dragovich, Lesley Murray
PII:
S0960-894X(17)30491-2
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.05.016
BMCL 24961
Reference:
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
24 March 2017
1 May 2017
3 May 2017
Please cite this article as: Liang, J., Labadie, S., Zhang, B., Ortwine, D.F., Patel, S., Vinogradova, M., Kiefer, J.R.,
Mauer, T., Gehling, V.S., Harmange, J-C., Cummings, R., Lai, T., Liao, J., Zheng, X., Liu, Y., Gustafson, A., Van
der Porten, E., Mao, W., Liederer, B.M., Deshmukh, G., An, L., Ran, Y., Classon, M., Trojer, P., Dragovich, P.S.,
Murray, L., From a novel HTS hit to potent, selective, and orally bioavailable KDM5 inhibitors, Bioorganic &
Medicinal Chemistry Letters (2017), doi: http://dx.doi.org/10.1016/j.bmcl.2017.05.016
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.

From a novel HTS hit to potent, selective, and orally
bioavailable KDM5 inhibitors
a*
a
a
a
a
Jun Liang , Sharada Labadie , Birong Zhang , Daniel F. Ortwine , Snahel Patel , Maia
a
a
a
b
b
Vinogradova , James R. Kiefer , Till Mauer , Victor S. Gehling , Jean-Christophe Harmange ,
b
c
c
c
a
Richard Cummings , Tommy Lai , Jiangpeng Liao , Xiaoping Zheng , Yichin Liu , Amy
a
a
c
a
a
Gustafson , Erica Van der Porten , Weifeng Mao , Bianca M. Liederer , Gauri Deshmukh , Le
a
a
a
b
a
a
An , Yingqing Ran , Marie Classon , Patrick Trojer , Peter S. Dragovich , Lesley Murray
a. Genentech Inc., 1 DNA Way, South San Francisco, California, 94080, USA
b. Constellation Pharmaceuticals Inc., 215 First Street, Suite 200, Cambridge, MA 02142,
USA
c. WuXi AppTec Co., Ltd., 288 Fute Zhong Road, Waigaoqiao Free Trade Zone, Shanghai,
2
00131, China
AUTHOR EMAIL ADDRESS liang.jun@gene.com
RECEIVED DATE (to be automatically inserted after your manuscript is accepted)
*
CORRESPONDING AUTHOR FOOTNOTE Corresponding Author. Tel: 650-467-3567, E-
mail: liang.jun@gene.com
1

Table of contents graphic
ABSTRACT: A high-throughput screening (HTS) of the Genentech/Roche library identified a
novel, uncharged scaffold as a KDM5A inhibitor. Lacking insight into the binding mode, initial
attempts to improve inhibitor potency failed to improve potency, and synthesis of analogs was
further hampered by the presence of a C-C bond between the pyrrolidine and pyridine. Replacing
this with a C-N bond significantly simplified synthesis, yielding pyrazole analog 35, of which we
obtained a co-crystal structure with KDM5A. Using structure-based design approach, we
identified 50 with improved biochemical, cell potency and reduced MW and lower lipophilicity
(LogD) compared with the original hit. Furthermore, 50 showed lower clearance than 9 in mice.
In combination with its remarkably low plasma protein binding (PPB) in mice (40%), oral dosing
of 50 at 5 mg/kg resulted in unbound Cmax ~2-fold of its cell potency (PC9 H3K4Me3 0.96 M),
meeting our criteria for an in vivo tool compound from a new scaffold
2

It is now widely accepted that DNA-histone interactions play an important role in
1
modulating chromatin structure and, as a result, regulation of gene transcriptions. An excellent
example is the histone lysine acetyl modification, modulated by histone acetyltransferases
(HATs) and histone deacetylases (HDACs), that have led to the discovery of several drugs
2
approved for treating cancers. More recently, histone lysine methylation states, which are
determined by the interplay between histone methyltransferases and histone demethylases, have
started to generate similar attention. The KDM5 family of histone demethylases (KDM5A-D) are
Fe(II)- and 2-oxoglutarate (2-OG)-dependent oxygenases that specifically remove methyl groups
3
from trimethylated lysine residues at the 4-position of histone 3 (H3K4Me3). Disruption of
KDM5 activity, either through RNAi or by a selective small molecule KDM5 inhibitor, reduced
4
survival of drug-tolerant cancer cells in vitro. Attracted by the therapeutic potential of reducing
drug resistance in the clinic, several groups have recently disclosed small molecule KDM5
5
inhibitors (1-5, Figure 1) as in vitro probes. We aimed to discover potent, selective, and orally
bioavailable KDM5 inhibitors with excellent pharmacokinetics profiles in mice in order to
further test the therapeutic hypothesis in vivo. Accordingly, we recently reported a promising
[1,5-a]pyrimidin-7(4H)-one-containing lead molecule (6) that met our in vivo “tool compound”
6
criteria.
Despite the identification of 6, we sought to identify a structurally and mechanistically
distinct second scaffold to test our therapeutic hypothesis both in vitro and in vivo. And the
rationales were two-fold. First, we reasoned that a structurally distinct scaffold might mitigate
any potential safety liabilities associated with the [1,5-a]pyrimidin-7(4H)-one scaffold.
Secondly, we noted that all the published KDM5 inhibitors (1-6) are weakly acidic and displayed
a competitive mechanism of inhibition relative to co-substrate 2-OG. This is consistent with their
3

binding mode, as highlighted with a co-crystal structure of N-oxalylglycine (NOG) 1 with
5f
KDM5B (PDB code 5FV3), in which the distal carboxylic acid achieved strong ionic
interactions with Lys501 in the active site. We sought to identify an uncharged inhibitor series,
preferably competitive with histone substrate alone or in combination with 2-OG competition
that might display smaller enzyme-cell potency shift in cells. Therefore, a HTS was conducted
with the Genentech/Roche small molecule library, in search of additional structurally and
mechanistically distinct hits against KDM5.
Figure 1.
Figure 1. Reported KDM5 inhibitors and the measured pKa values.
From the HTS campaign, compound 7 (as a racemic mixture) emerged as an interesting
hit with reasonable physiochemical properties (Figure 2) and excellent KDM5 potency (IC50
7
0
.26 M, LE 0.33, LLE 4.0). Additionally, 7 was found to be quite selective for KDM5,
exhibiting IC > 25 M against other KDMs (KDM1A, 2B, 3B, 4C, 6A and 7B were selected
50
from each KDM family to assess a compound’s KDM-selectivity. Only the IC values for
5
0
KDM4C are shown in Figure 2, since it is closest to KDM5A by sequence and structure). More
4

importantly, the measured basic pKa for 7 was 6.1, a weak basic due to the pyridyl moiety.
Therefore, 7 was expected to be mostly uncharged at physiological pH (7.4), in contrast to the
majority of KDM5 inhibitors (1-6) reported in the literature which were weakly acidic and
negatively charged. The measured LogD of 7 at pH 7.4 was 2.8, which was relatively high in the
commonly accepted range of < 3.0 and something we hoped to address in our SAR optimization.
1
9
In a biophysical confirmation of binding with the KDM5A enzyme via F NMR, 7 was found to
be competitive with a 17-amino acid histone-derived peptide, in which an additional p-CF3-
substituted phenylalanine was added to the C-terminus resulting in the sequence ART-KMe3-
8
QTARKSTGGKAPRKQLA-p-CF . Given that 7 fulfilled the criteria desired for a structurally
3
and mechanistically distinct hit relative to 6, we decided to undertake systematic SAR in order to
improve its biochemical and cell potency. As the first step toward this goal, a racemic mixture of
7
was subjected to chiral SFC separation to yield a pair of enantiomers which were subsequently
profiled. One of the two enantiomers (9) was found to be more active than 7, while maintaining
the excellent KDM-selectivity of the original hit. However, even the more active enantiomer 9
was inactive when tested in the cell assay (PC9 H3K4Me3 EC > 30 M).
5
0
Figure 2.
Physical properties:
Measured pKa 6.1 (pyridine)
MW
378
2.8
LogD
tPSA
79
Kinetic solu.
56 M
7
8
9
(racemic)
(enantiomer 1) (enantiomer 2)
a
KDM5A IC (M)
0.26
0.77 0.16
50
5

b
KDM4C IC (M)
>25
>30
ND
ND
13.3
>30
50
c
PC9 H3K4Me3 EC (M)
50
a
Figure 2: Structure of the racemic HTS hit 7, and the biochemical and physical properties.
Biochemical assay measuring the demethylation of a H3K4Me3 peptide substrate has been
4
c
described. Results shown are the arithmetic mean of at least 3 separate runs. On average, the
b
coefficients of variation were less than 0.10 times the mean for biochemical assay. Biochemical
assays for KDM1A, 2B, 3B, 4C, 6A and 7B have been established to assess KDM-selectivity
4c
and have been described. Only the IC values for KDM4C are shown in this table, since it is
5
0
closest to KDM5A by sequence and structure. Results shown are the arithmetic mean of at least
3
separate runs. On average, the coefficients of variation were less than 0.21 times the mean for
c
biochemical assay. Cell-based assay measuring the H3K4Me3 level in PC9 cells relative to that
4
c
of DMSO control has been described. Results are the arithmetic mean of at least 3 separate
runs. On average, the coefficients of variation were less than 0.24 times the mean. ND, not
determined.
Despite numerous attempts, we were unsuccessful in obtaining a co-crystal structure of
the HTS hit 7 or its more active enantiomer 9 with KDM5A enzyme. Therefore, we relied on
docking models to guide our initial SAR optimization. In the docking model of the more active
enantiomer 9 with KDM5A enzyme (Figure 3), the amide carbonyl was proposed to bind to the
active site Fe(II), while the isopropyl-pyrazole N was envisioned to have a H-bond interaction
with Lys501. Furthermore, two additional H-bond interactions were proposed between 9 and
KDM5A: one between Tyr472 and pyridine N atom and a second one between Arg73 and the N-
methyl pyrazole N atom.
Figure 3.
6

Tyr472
Lys501
Arg73
Glu485
His571
His483
Figure 3: Docking model of 9 in complex with KDM5A enzyme, adopted from a previously
published KDM5A structure (PDB code 5CEH).
Even though the binding mode of 9 with KDM5A seemed reasonable, we sought to
confirm the binding hypothesis through SAR. Initially we focused on the pyrrolidine amide
which was proposed to play a key role in binding to KDM5A via coordination with the active
site Fe(II) (Table 1). First, we wanted to simplify the stereochemistry of 7 by introducing an
olefin in the pyrrolidine ring. This exercise led to analog 10 which was inactive. Replacement of
the amide bond in 7 either by a methylene (11) or a sulfonamide (12) also led to complete loss of
potency. Additionally, the pyrrolidine moiety appeared to be critical for activity as both C4- or
C3-substituted piperidines (13 or 14) were not tolerated. Accordingly, we decided to vary other
areas of 7 for optimization.
7

Table 1. SAR for analogs containing pyrrolidine amide modifications
KDM5A
1
Example*
R
a
IC50 (nM)
7
260
1
0
>25000
1
1
1
2
>25000
>25000
1
1
3
4
>25000
>25000
a
*
Unless specified, a racemic mixture was used for testing. Biochemical assay. Results shown
are the arithmetic mean of at least 3 separate runs. On average, the coefficients of variation were
less than 0.10 times the mean for biochemical assay.
We next shifted SAR studies to the isopropyl-pyrazole. The associated SAR is
summarized in Table 2. Complete loss of enzyme potency was noted when the isopropyl-
8

pyrazole in 7 was replaced with other 5-membered heterocyles (examples 15-17), despite the fact
that a heteroatom (N or O) was maintained at a location which could maintain a H-bond
interaction with Lys501 as proposed in our docking model (Figure 3). We were also somewhat
surprised to find that the two N-methyl pyrazole regio-isomers 18 and 19 were completely
inactive. Additionally, replacement of isopropyl group with N,N-dimethylamino moiety (20),
mimicking the dimethylated lysine substrate, also resulted in >10-fold loss in biochemical
potency relative to 7. To confirm that the isopropyl substituent on the pyrazole was optimal, we
systematically examined other alkyl groups at this position (examples 21-25). Linear or branched
alkyls were determined to be less potent than the isopropyl. When a polar functional group was
introduced at this position in the form of a methyl ketone (26), complete loss of potency was
observed. This result suggested that polar substituents were not tolerated in this area.
Table 2. SAR for analogs containing modification at the left-hand amide.
KDM5A
2
Example*
R
a
IC50 (nM)
7
260
1
5
>25000
9

1
1
1
6
7
8
>25000
>25000
>25000
1
2
9
0
>25000
3300
2
2
1
2
2600
878
2
2
3
4
458
357
1
0

2
2
5
6
897
>25000
a
*
Unless specified, a racemic mixture was used for testing. Biochemical assay. Results shown
are the arithmetic mean of at least 3 separate runs. On average, the coefficients of variation were
less than 0.10 times the mean for biochemical assay.
Subsequently, we explored SAR of the right hand pyrido-methyl-pyrazole moiety (Table
3
). We initially prepared analog 27 which contained a phenyl group instead of the pyridine
moiety present in 7. However, analog 27 was found to be biochemically inactive suggesting the
pyridine N atom was critical for binding with KDM5A. Benzyl-substituted pyrazole and other 5-
or 6-membered aromatic groups (28-32) were tolerated with comparable potency as N-methyl
pyrazole 7. Interestingly, 3-methyl phenyl analog (31) was tolerated, suggesting the proposed H-
bond with Arg73 was not important or our proposed binding model might be incorrect for this
area. With an eye towards improving lipophilic ligand efficiency (LLE), we also designed small
and polar groups in the form of amides (examples 33 and 34), in an attempt to replace the
methyl-pyrazole moiety. However, neither of these modifications was tolerated. The methyl-
pyrazole moiety appeared to have limited impact on potency, with benzyl pyrazole (28) and a
couple of related analogs (31 and 32) only slightly more potent than 7, albeit at the cost of
significantly greater MW and lipophilicity than 7.
1
1

Table 3. SAR for analogs varying the methyl-pyrazole group
KDM5A
3
a
Example*
R
IC50
nM)
(
7
260
2
2
2
7
8
9
>25000
105
469
3
3
0
1
265
122
1
2

3
3
2
3
128
7200
3
4
1500
a
*
Unless specified, a racemic mixture was used for testing. Biochemical assay. Results shown
are the arithmetic mean of at least 3 separate runs. On average, the coefficients of variation were
less than 0.10 times the mean for biochemical assay.
In light of the unproductive SAR outlined above, we made more drastic changes to 7,
particularly with goal towards structural simplification. A significant hurdle associated with the
synthesis of 7 and its analogs was the formation of the C-C bond between pyrolidine and
pyridine. As a result, we sought to replace the synthetically more challenging C-C bond with a
C-X bond in which X could be a heteroatom such as N or O. We envisioned S 2 displacement,
N
Buchwald coupling, or amide coupling reactions would access these types of modifications.
While unclear such changes would be tolerated by the KDM5A enzyme, we reasoned that the
ease of synthesis would allow us to rapidly evaluate the new designs. The resulting SAR is
summarized in Table 4.
1
3

Table 4. SAR for analogs varying the methyl-pyrazole pyridine moiety
KDM5A
4
a
Example*
R
IC50
(
nM)
7
260
3
3
5
6
727
>25000
3
3
7
8
14300
8900
3
4
9
0
14600
99
a
*
Unless specified, a racemic mixture was used for testing. Biochemical assay. Results shown
are the arithmetic mean of at least 3 separate runs. On average, the coefficients of variation were
less than 0.10 times the mean for biochemical assay.
We began by replacing the pyridine present in 7 with 3-Br pyrazole (35), with the
pyrazole N2 mimicking the pyridine N atom. This change was tolerated, with only 2-3 fold loss
in potency. However, saturated heterocycles, exemplified by morpholine (36), were completely
inactive. We also examined 2-O- and 2-NH-pyridine derivatives (examples 37-38) and found that
1
4

both changes led to significant loss of potency. Similarly, we synthesized amide analogs,
assuming that the carbonyl O might fulfill the same role as the pyridine N atom in 7.
Unfortunately, methyl-pyrazole amide 39 was not active, while the “reverse amide” 40 showed
improved potency against KDM5A relative to 7.
We obtained a co-crystal structure of the compound 35 in complex with KDM5A at 3.0 Å
resolution (Figure 4). Despite the fact that a racemic mixture was used in the co-crystallization
with KDM5A, one enantiomer with R-configuration at C3-position on the pyrrolidine bound in
2
+
the KDM5A active site. In this structure, Ni was used as a catalytically inactive surrogate for
2
+
Fe . Contrary to our initial docking model, the crystal structure showed that the isopropyl-
pyrazole and amide were co-planar with the N1 and carbonyl O atoms both coordinated to the
metal. In addition, the N2 atom of the isopropyl-pyrazole was positioned 2.5 Å away from
Glu485, suggesting a strong H-bond between the two atoms. These observations suggested that
the isopropyl-pyrazole bound to KDM5A in the tautomer depicted in Figure 4, in order to
achieve two highly favorable interactions with the protein. Thus, the orientation of the compound
in the actual binding site was approximately 180° rotated from the docking model. At the
opposite end of the inhibitor, the Br-pyrazole N1’ atom was positioned 2.5 Å away from the
sidechain of Lys501, suggesting another potential, though long, H-bond interaction that could be
further optimized by replacing the Br-pyrazole ring.
While the majority of compound 35 would overlap the binding pocket of the co-substrate
2
-OG, and thus expected to be 2-OG competitive, superposition of the KMe3 histone tail peptide
9
complex from a co-crystal structure of KDM6A (PDB code 3AVR) demonstrated that the
isopropyl group occupied the same region of the binding pocket as the trimethylated lysine
(Supplementary Figure S1b). The crystal structure was consistent with the competition of 7
1
9
against histone-derived peptide with binding to KDM5A as determined by F NMR described
1
5

earlier. This suggested that the pyrolidine amide inhibitors competed with histone peptides for
binding with KDM5A, a novel mechanism that has not been reported for KDM5 inhibitors.
Figure 4.
Figure 4. X-ray crystal structure of 35 in complex with KDM5A enzyme. In this crystal
2+
2+
structure, Ni was used as a catalytically inactive surrogate for Fe . The resolution of the X-ray
structure is 3.0 Å (PDB code 5V9P). Hydrogen bonds are indicated as dashed lines, and the
2+
active site metal ion (Ni ) is shown as a tan sphere.
With the understanding the true binding mode of the series, we continued SAR
explorations of benzamide 40 by preparing and testing a small library of different amides.
Representative examples are shown in Table 5. Starting with acetyl amide (41), a significant
drop of potency was observed, compared to benzamide 40. Ethyl (42) and isopropyl amide (43)
were 3-4-fold more potent than the acetyl amide (41). However, no significant improvement was
observed with tert-butyl amide (44). We also evaluated cyclic alkyl amides and were gratified to
note that cyclobutyl (45) and cyclopropyl amide (48) showed the improved biochemical potency
1
6

(
KDM5A IC = 80 and 65 nM, respectively). Therefore, we proceeded to separate the two
50
enantiomers of each racemic mixture. One of the cyclobutyl enantiomers (47) was more active
than the other (48) by approximately 3-fold. Similarly, we found that for the cyclopropyl amide
pairs (49 and 50), one was more potent than the other (45 vs. 90 nM). The more active
enantiomer (50) also showed cell EC of 960 nM, meeting the cell potency criteria for an in vivo
50
probe. Additionally, as a result of its improved biochemical potency, lower MW and reduced
lipophilicity (measured LogD = 1.3 at pH 7.4), 50 showed improved LE (0.49 vs. 0.33) and LLE
(5.9 vs. 4.0) when compared to that of the HTS hit 7.
Table 5. SAR of different amides.
KDM5A
PC9 H3K4Me3
5
Example*
R
a
b
IC50 (nM)
EC50 (nM)
4
4
4
4
4
4
0
1
2
3
4
5
Ph
Me
Et
99
1100
251
204
198
80
6500
ND
ND
ND
ND
ND
iPr
tBu
cBu
cBu
4
6
180
ND
(
(
enantiomer 1)
cBu
4
4
7
8
60
65
1500
1600
enantiomer 2)
cPr
1
7

cPr
4
5
9
0
90
45
ND
960
(
(
enantiomer 1)
cPr
enantiomer 2)
a
*
Unless specified, a racemic mixture was used for testing. Biochemical assay. Results shown
are the arithmetic mean of at least 3 separate runs. On average, the coefficients of variation were
b
less than 0.10 times the mean for biochemical assay. Cell-based assay measuring the H3K4Me3
level in PC9 cells relative to that of DMSO control. Results are the arithmetic mean of at least 3
separate runs. On average, the coefficients of variation were less than 0.24 times the mean. ND,
not determined.
Subsequently, we obtained a crystal structure of 50 bound to KDM5A at 2.7 Å resolution
which helped us define the absolute stereochemistry of the more active enantiomer as R-
configuration at the C3-position on the pyrrolidine. As shown in Figure 5, the overall binding
mode of 50 was quite similar to that observed with pyrazole 35, particularly at the active site in
2
+
which Ni resided. As in the previous structure, the isopropyl-pyrazole was co-planar with the
2
+
amide carbonyl, resulting in bidentate coordination to the Ni , while maintaining the H-bond
interaction between the NH and Glu485 of KDM5A.
Figure 5.
1
8

Figure 5. X-ray crystal structure of 50 in complex with KDM5A enzyme. In this crystal
2+
2+
structure, Ni (tan sphere) was used as a catalytically inactive surrogate for Fe . The resolution
of the X-ray structure is 2.7 Å (PDB code 5V9T).
While optimizing the biochemical and cell potencies of the pyrolidine amide analogs, we
also monitored their metabolic stability both in vitro and in vivo (mice) in order to identify a
potential in vivo chemical probe. Those data are summarized in Table 6. The more active
enantiomer (9) of the HTS hit 7 exhibited moderate to high in vitro metabolic instability, as
measured with human, rat, and mouse liver microsomes (LM) and hepatocytes (Hep). Consistent
with the in vitro metabolic stability data in mouse, 9 showed high clearance (CLp 99.9
mL/min/kg) when dosed in mice. MetID studies of 9 with human LM showed the main
metabolites formed were due to N-demethylation of the methyl pyrazole and oxidation of the
methyl group on the pyridine moiety. Consistent with the metID results, replacement of N-
methylpyrazole-pyridine with cyclopropyl amide (50) led to significant improvement in in vitro
metabolic stability across species. When tested in vivo, 50 showed moderate clearance (28
1
9

mL/min/kg) in mice with good oral bioavailability (F% 34). Given that 50 showed remarkably
low plasma protein binding in mice (40%), the unbound Cmax observed for the compound
following a 5 mg/kg oral dose was ~2x of its cell EC (960 nM).
5
0
Table 6. In vitro metabolic stability and in vivo PK parameters for select analogs
a
PPB
In vitro metab. stability
In vivo metab. stability
Human, rat, mouse
d
Mouse
Example
b
c
Human, rat, mouse
LM (mL/min/kg)
Human, rat, mouse
(
%)
CLp
t1/2 Vd
F%
Hepatocyte (mL/min/kg)
(
mL/min/kg) (hr) (L/kg)
9
95.5 97.3 95.7
44.0 37.6 40.0
7
37
69
12
12
39
66
0.6
99.9
28
0.3
2.1
0.7
94
34
5
0
4.2 13
1.2
6.3
0.4
a
b
c
Plasma protein binding. Hepatic clearance predicted from liver microsomes. Hepatic
d
clearance predicted from hepatocytes. Female CD-1 mice were given an intravenous dose of 1
mg/kg in PEG/EtOH/H O as a solution and an oral dose of 5 mg/kg in MCT as a suspension.
2
Based on this cell potency and in vivo PK profile, 50 met our in vivo tool compound
criteria and was selected for additional profiling. Compound 50 was shown to be a pan-KDM5
enzyme inhibitor, with IC50 of 56 and 55 nM against KDM5B and KDM5C isoforms,
respectively. However, it was significantly less potent against other KDM enzymes (1A, 2B, 3B,
4
C, 6A, 7B), inhibiting KDM4C the strongest with an IC of 4.1 M. Thus 50 displayed 91-fold
5
0
selectivity for KDM4C vs. KDM5A. When tested at 10 M in Invitrogen kinase (300 kinases)
and Cerep (40 enzymes/receptors/ion channels) panels, 50 did not show > 50% inhibition against
any target.
In conclusion, we were attracted to a novel HTS hit 7 as a structurally distinct scaffold,
with different mechanism of KDM5A inhibition relative to the molecules we had previously
2
0

disclosed. After comprehensive SAR studies that were guided by structural biology, we were
able to identify a biochemically and cellular more potent compound 50, with promising in vivo
PK profiles. Remarkably, we significantly reduced MW and lipophilicity (LogD) for the top lead
(50) of this scaffold from the original hit (7). As a result, 50 was selected from the second
scaffold for in vivo biological studies, results from which will be disclosed in the future.
ACKNOWLEDGEMENT
We thank, Mengling Wong, Michael Hayes, Baiwei Lin, Yutao Jiang, and Yanzhou Liu
for purification and analytical support; Daniel Hascall, Grady Howes, Gigi Yuen, and Garima
Porwal for compound management support; Leslie Wang, Hoa Le and Qin Yue for ADME
support; Emile Plise and Jonathan Cheng for MDCK data; Ning Liu for microsomal stability;
Jasleen Sodhi for reversible CYP inhibition measurement; Quynh Ho for plasma protein binding
measurement. We thank the BME group at Genentech for insect cell expression. We thank the
staff at LS-CAT at the APS and Shamrock Inc for data collection. 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. Use of the LS-CAT Sector 21 was supported by the Michigan
Economic Development Corporation and the Michigan Technology Tri-Corridor (Grant
0
85P1000817). We also thank the staff at beamline 5.0.2 at ALS. This research used resources
of the Advanced Light Source, which is a DOE Office of Science User Facility under contract
no. DE-AC02-05CH11231.
A. SUPPLEMENTARY DATA
2
1

Supplementary data associated with this article can be found, in the online version, at http://xxx.
REFERENCES
1
. (a) Margueron, R.; Reinberg, D. Nat. Rev. Genet. 2010, 11, 285. (b) Itoh, Y.; Suzuki, Miyata,
N. Mol. BioSyst. 2013, 9, 873. (c) Black, J. C.; Van Rechem, C.; Whetstine, J. R. Mol. Cell 2012,
8, 491. (d) Greer, E. L.; Shi, Y. Nat. Rev. Genet. 2012, 13, 343. (e) Lohse, B.; Kristensen, J. L.;
Kristensen, L. H.; Agger, K.; Helin, K.; Gajhede, M.; Clausen, R. P. Bioorg. Med. Chem. 2011,
4
1
9, 3625. (f) Suzuki, T.; Miyata, N. J. Med. Chem. 2011, 54, 8236.
2
. (a) Marks, P. A.; Rifkind, R. A.; Richon, V. M.; Breslow, R.; Miller, T.; Kelly, W. K. Nature
Rev. Cancer 2001, 1, 195; (b) Manal, M.; Chandrasekar, M. J. N.; Priya, J.; Nanjan, M. J.
Bioorg. Chem. 2016, 67, 18.
3
. (a) Kooistra, S. M.; Helin, K. Nat. Rev. Mol. Cell Biol. 2012, 13, 297. (b) Højfeldt, J. W.;
Agger, K.; Helin, K. Nat. Rev. Drug Discov. 2013, 12, 917.
. (a) Sharma, S. V.; Lee, D. Y.; Li, B.; Quinlan, M. P.; Takahashi,F.; Maheswaran, S.;
4
McDermott, U.; Azizian, N.; Zou, L.; Fischbach, M. A.; Wong, K.; Brandstetter, K.; Wittner, B.;
Ramaswamy, S.; Classon, M.; Settleman, J. Cell 2010, 141, 69. (b) Roesch, A.; Vultur, A.;
Bogeski, I.; Wang, H.; Zimmermann, K. M.; Speicher, D.; Körbel, C.; Laschke, M. W.; Gimotty,
P. A.; Philipp, S. E.; Krause, E.; Pätzold, S.; Villanueva, J.; Krepler, C.; Fukunaga-Kalabis, M.;
Hoth, M.; Bastian, B. C.; Vogt, T.; Herlyn, M. Cancer Cell 2013, 23, 811. (c) Vinogradova, M.;
Gehling, V. S.; Gustafson, A.; Arora, S.; Tindell, C. A.; Wilson, C.; Williamson, K. E.;
Gangurde, P.; Manieri, W.; Busby, J.; Flynn, M.; Lan, F.; Kim, H.-J.; Odate, S.; Cochran, A. G.;
Liu, Y.; Wongchenko, M.; Yang, Y.; Cheung, T.; Maile, T. M.; Lau, T.; Costa, M.; Hegde, G.
2
2

V.; Jackson, E.; Pitti, R.; Arnott, A..; Bailey, C.; Bellon, S.; Cummings, R. T.; Albrecht, B. K.;
Harmange, J. C.; Kiefer, J. R.; Trojer, P.; Classon, M. Nat. Chem. & Biol., 2016, 12, 531.
5
. (a) Cloos, P. A. C.; Christensen, J.; Agger, K.; Mailica, A.; Rappsilber, J.; Antal, T.; Hansen,
K. H.; Helin, K. Nature 2006, 442, 307. (b) McAllister, T. E.; England, K. S.; Hopkinson, R. J.;
Brennan, P. E.; Kawamura, A.; Schfield, C. J. J. Med Chem. 2016, 59, 1308. (c) Westaway, S.
M.; Preston, A. G. S.; Baker, M. D.; Brown, F.; Brown, J. A.; Campbell, M.; Chung, C.; Diallo,
H.; Douault, C.; Drewes, G.; Eagle, R.; Gordon, L.; Haslam, C.; Hayhow, T. G.; Humphreys, P.
G.; Joberty, G.; Katso, R.; Kruidnenier, L.; Leveridge, M.; Liddle, J.; Mosley, J.; Muelbaier, M.;
Randle, R.; Rioja, I.; Rueger, A.; Seal, G. A.; Sheppard, R. J.; Singh, O.; Taylor, J.; Thomas, P.;
Thomson, D.; Wilson, D. M.; Lee, K.; Prinjha, R. K. J Med. Chem. 2016, 59, 1357. (d)
Westaway, S. M.; Preston, A. G. S.; Barker, M. D.; Brown, F.; Brown, J. A.; Campbell, M.;
Chung, C.; Drewes, G.; Eagle, R.; Garton, N.; Gordon, L.; Haslam, C.; Hayhow, T. G.;
Humphreys, P. G.; Joberty, G.; Katso, R.; Kruidnenier, L.; Leveridge, M.; Pemberton, M.; Seal,
G. A.; Shipley, T.; Singh, O.; Suckling, C. J.; Taylor, J.; Thomas, P.; Wilson, D. M.; Lee, K.;
Prinjha, R. K. J Med. Chem. 2016, 59, 1370. (e) Bavetsias, V.; Lanigan, R. M.; Ruda, G. F.;
Atrash, B.; McLaughlin, M. G.; Tumber, A.; Mok, N. Y.; Le Bihan, Y.-V.; Dempster, S.; Boxall,
T. J.; Jeganathan, F.; Hatch, S. B.; Savitsky, P.; Velupillai, S.; Krojer, T.; England, K. S.;
Sejberg, J.; Thai, C.; Donovan, A.; Pal, A.; Scozzafava, G.; Bennett, J. M.; Kawamura, A.;
Johnsson, C.; Szykowska, A.; Gileadi, C.; Burgess-Brown, N. A.; Delft, F.; Oppermann, U.;
Walters, Z.; Shipley, J.; Raynaud, F. I.; Westaway, S. M.; Prinjha, R. K.; Fedorov, O.; Burke, R.;
Schofield, C. J.; Westwood, I. M.; Bountra, C.; Muller, S.; Montfort, R. L. M.; Brennan, P. E.;
Blagg, J. J Med. Chem. 2016, 59, 1388. (f) Johansson, C.; Velupillai, S.; Tumber, A.;
Szykowska, A.; Hookway, E. S.; Nowak, R. P.; Strain-Damerell, C.; Gileadi, C.; Phipott, M.;
Burgess-Brown, N.; Wu, N.; Nuzzi, A.; Steuber, H.; Egner, U.; Badock, V.; Munro, S.;
2
3

LaThangue, N. B.; Westaway, S.; Brown, J.; Athanasou, N.; Prinjha, R.; Brennan, P. E.;
Oppermann, U. Nat. Chem. & Biol., 2016, 12, 539. (g) Horton, J. R.; Liu, X.; Gale, M.; Wu, L.;
Shanks, J. R.; Zhang, X.; Webber, P. J.; Bell, J. S. K.; Kales, S. C.; Mott, B. T.; Rai, G.; Jansen,
D. J.; Henderson, M. J.; Urban, D. J.; Hall, M. D.; Simeonov, A.; Maloney, D. J.; Johns, M. A.;
Fu H.; Jadhav, A.; Vertino, P. M.; Yan, Q.; Cheng, X. Cell Chem. Biol. 2016, 23, 769. (h)
Labadie, S.; Dragovich, P. S.; Ackerman, L.; Cummings, R.; Deshmukh, G.; Gustafson, A.;
Harmange, J. C.; Kiefer, J. R.; Liang, J.; Liederer, B. M.; Liu Y.; Manieri, W.; Mao, W.; Murray,
L.; Ortwine, D. F.; Trojer, P.; VanderPorten, E.; Vinogradova, M. Bioorg. Med. Chem Lett.
2
016, 26, 4492. (i) Gehling, V. S.; Bellon, S.; Harmange, J. C.; Leblanc, Y.; Odate, S.; Buker, S.;
Lan F.; Sandy, P.; Bergeron, L.; Mao, W.; Gustafon, A.; Liu Y.; VanderPorten, E.; Trojer, P.;
Albrecht, B. K. Bioorg. Med. Chem Lett., 2016, 26, 4350. (j) Tumber, A.; Nuzzi, A.; Hookway,
E. S.; Hatch, S. B.; Velupillai, S.; Johansson, C.; Kawamura, A.; Savitsky, P.; Yapp, C.;
Szykowska, A.; Wu, N.; Bountra, C.; Strain-Damerell, C.; Burgess-Brown, N. A.; Ruda, G. F.;
Fedorov, O.; Munro, S.; England, K. S.; Nowak, R. P., Schofield, C. J.; La Thangue, N. B.;
Pawlyn, C.; Davies, F.; Morgan, G.; Athanasou, N.; Muller, S.; Oppermann, U.; Brennan, P. E.
Cell Chem. Biol. http://dx.doi.org/10.1016/j.chembiol.2017.02.006.
6
. Liang, J.; Zhang, B.; Labadie, S. Ortwine, D. F.; Vinogradova, M.; Kiefer, J. R.; Gehling, V.
S.; Hamange, J.-C.; Cummings, R.; Lai, T.; Liao, J.; Zheng, X.; Liu, Y.; Gustafson, A.; Van der
Porten, E.; Mao, W.; Liederer, B. M.; Deshmukh, G.; Classon, M.; Trojer, P.; Dragovich, P. S.;
Murray, L. Bioorg. Med. Chem Lett. 2016, 26, 4036.
7
. (a) Hopkins, A. L.; Groom, C. R.; Alex A. Drug Discov. Today 2004, 9, 430. (b) Leeson, P.
D.; Springthorpe, B. Nature Rev. Drug Discov. 2007, 6, 881.
. Dalvit, C.; Papeo, G.; Mongelli, N.; Giordano, P.; Saccardo, B.; Costa, A.; Veronesi, M.; Ko,
S-Y. Drug Develop. Res. 2005, 64, 105.
8
2
4

9
. Sengoku, T.; Yokoyama, S. Gene & Develop. 2011, 25, 2266.
2
5