Discovery of First-in-Class, Potent, and Orally Bioavailable Embryonic Ectoderm Development (EED) Inhibitor with Robust Anticancer Efficacy

Article abstract of DOI:10.1021/acs.jmedchem.6b01576

Overexpression and somatic heterozygous mutations of EZH2, the catalytic subunit of polycomb repressive complex 2 (PRC2), are associated with several tumor types. EZH2 inhibitor, EPZ-6438 (tazemetostat), demonstrated clinical efficacy in patients with acceptable safety profile as monotherapy. EED, another subunit of PRC2 complex, is essential for its histone methyltransferase activity through direct binding to trimethylated lysine 27 on histone 3 (H3K27Me3). Herein we disclose the discovery of a first-in-class potent, selective, and orally bioavailable EED inhibitor compound 43 (EED226). Guided by X-ray crystallography, compound 43 was discovered by fragmentation and regrowth of compound 7, a PRC2 HTS hit that directly binds EED. The ensuing scaffold hopping followed by multiparameter optimization led to the discovery of 43. Compound 43 induces robust and sustained tumor regression in EZH2^MUT preclinical DLBCL model. For the first time we demonstrate that specific and direct inhibition of EED can be effective as an anticancer strategy.

Full text of DOI:10.1021/acs.jmedchem.6b01576

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Article
Discovery of First-in-Class, Potent and Orally Bioavailable
EED inhibitor with Robust Anti-cancer Efficacy
Ying Huang, Jeff Zhang, Zhengtian Yu, Hailong Zhang, Youzhen Wang, Andreas Lingel, Wei Qi,
X. Justin Gu, Kehao Zhao, Michael David Shultz, Long Wang, Xingnian Fu, Yongfeng Sun, Qiong
Zhang, Xiangqing Jiang, Jiang-wei Zhang, Chunye Zhang, Ling Li, Jue Zeng, Lijian Feng, Chao
Zhang, Yueqin Liu, Man Zhang, Lijun Zhang, Mengxi Zhao, Zhenting Gao, Xianghui Liu, Douglas
Fang, Haibing Guo, Yuan Mi, Tobias Gabriel, Michael P Dillon, Peter Atadja, and Counde Oyang
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01576 • Publication Date (Web): 16 Jan 2017
Downloaded from http://pubs.acs.org on January 16, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
Dillon, Michael; Ideaya Biosciences ,
Atadja, Peter; Novartis Institutes for Biomedical research, Biology
Oyang, Counde; China Novartis Institutes for BioMedical Research Co., Ltd.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 34
1
2
3
4
TITLE
5
6
7
8
9
Discovery of FirstꢀinꢀClass, Potent and Orally Bioavailable EED inhibitor with Robust Antiꢀ
cancer Efficacy
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ying Huang*†, Jeff Zhang†, Zhengtian Yu†, Hailong Zhang†¥, Youzhen Wang†§, Andreas
Lingel‡, Wei Qi†, Justin Gu†, Kehao Zhao†§, Michael D. Shultz†§, Long Wang†, Xingnian Fu†,
Yongfeng Sun†, Qiong Zhang†, Xiangqing Jiang†, Jiangwei Zhang†, Chunye Zhang†, Ling Li†,
†
†
†
†
†
†
†
£
£
Jue Zeng , Lijian Feng , Chao Zhang , Yueqin Liu , Man Zhang , Lijun Zhang , Mengxi Zhao ,
†
†€
†
†
†
†¶
£
Zhenting Gao , Xianghui Liu , Douglas Fang , Haibing Guo , Yuan Mi , Tobias Gabriel ,
Michael P. Dillon‡&, Peter Atadja†, Counde Oyang†
†Novartis Institutes for BioMedical Research, 4218 Jinke Road, Shanghai 201203, China
‡Novartis Institutes for BioMedical Research, 5300 Chiron Way, Emeryville, California 94608,
United States
*Corresponding author:
Ying Huang
Address: 4218 Jinke Road, Zhangjiang HiꢀTech Park, Pudong New Area, Shanghai
201203, China
Tel: 86ꢀ18621082695
Keywords:
Epigenetic, allosteric inhibition, fragmentꢀbased drug discovery, PRC2, oncology, EED
ACS Paragon Plus Environment
1

Page 3 of 34
Journal of Medicinal Chemistry
1
2
3
4
ABSTRACT
5
6
7
8
9
Overexpression and somatic heterozygous mutations of EZH2, the catalytic subunit of Polycomb
repressive complex 2 (PRC2), are associated with several tumor types. EZH2 inhibitor, EPZꢀ
6438 (Tazemetostat), demonstrated clinical efficacy in patients with acceptable safety profile as
monotherapy. EED, another subunit of PRC2 complex, is essential for its histone
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
methyltransferase activity through direct binding to trimethylated lysine 27 on histone 3
(H3K27Me3). Herein we disclose the discovery of a firstꢀinꢀclass potent, selective and orally
bioavailable EED inhibitor compound 43 (EED226). Guided by Xꢀray crystallography,
compound 43 was discovered by fragmentation and regrowth of Compound 7, a PRC2 HTS hit
that directly binds EED. The ensuing scaffold hopping followed by multiꢀparameter
optimization led to the discovery of 43. Compound 43 induces robust and sustained tumor
regression in EZH2MUT preꢀclinical DLBCL model. For the first time we demonstrate that
specific and direct inhibition of EED can be effective as an antiꢀcancer strategy.
ACS Paragon Plus Environment
2

Journal of Medicinal Chemistry
Page 4 of 34
1
2
3
4
ꢀ INTRODUCTION
5
6
7
8
Tumorigenesis is believed to involve multiple epigenetic alterations, in addition to genetic
aberrations, that contribute to the progressive transformation of normal cells towards a malignant
phenotype. As a result, novel cancer therapies that work by reversing epigenetic effects are being
increasingly explored.1 Postꢀtranslational modifications of core histone proteins of chromatin are
one of the major epigenetic mechanisms regulating gene expression, and paramount among them
are methylation events at lysine and arginine residues, catalyzed by histone methyltransferases
(HMTs).2
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Polycomb repressive complex 2 (PRC2) is a multiprotein complex that catalyzes the methylation
of histone H3 at lysine 27 (H3K27). Trimethylated H3K27 (H3K27Me3) is a repressive postꢀ
translational modification.3 Overexpression, gainꢀofꢀfunction mutations of EZH2 and
hypertrimethylation of H3K27 have been implicated in a myriad of cancers.4 The polycomb
group (PcG) proteins SUZ12, EED, EZH2 (or its homolog EZH1), RBBP4, and RBBP7
constitute the ‘‘core PRC2’’; removal of core subunits by genetic or RNAiꢀbased approaches
destabilizes EZH2 and results in the abrogation of all PRC2 functions.5 More importantly, it was
known that the recognition of H3K27Me3 by WD40ꢀrepeats containing βꢀpropeller protein EED
is essential in stimulating basal PRC2 activity and propagating H3K27 methylation in repressive
chromatin for gene silencing 6
. EED selectively binds the positively charged quaternary amine of
the trimethylated lysine via a soꢀcalled “aromatic cage” formed by Phe97, Tyr148 and Tyr365,
while the neighboring Trp364 interacts with the aliphatic sidechain of the lysine residue.6a
EZH2 or EZH1 acts as the catalytic subunit of PRC2.7 Despite high sequence identity, EZH2
and EZH1 are not functionally redundant and have different expression patterns. While EZH2 is
found only in actively dividing cells, EZH1 is found in both dividing and nonꢀdividing cells.
ACS Paragon Plus Environment
3

Page 5 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
1 (DZNep)
2 (EPZ005687)
3 (GSK126)
9
SAM competitive
50ꢀfold selective over Ezh1
SAM competitive
150ꢀfold selective over Ezh1
SAH hydrolase inhibitor
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5 (EPZꢀ6438)
Tazemetostat
SAM competitive
4 (UNC1999)
SAM competitive
<10ꢀfold selective for Ezh1
6 (CPI1205)
SAM competitive
~25ꢀfold selective for Ezh1
35ꢀfold selective over Ezh1
Figure 1: Selected PRC2 inhibitors
Although PRC2 containing EZH1 (PRC2−EZH1) has lower catalytic activity compared to that
containing EZH2 (PRC2−EZH2), both complexes contribute to the maintenance of cellular
H3K27 methylation states.
Given the association of PRC2/EZH2 with cancer, multiple biotech and pharmaceutical
companies have been actively pursuing compounds that can effectively inhibit PRC2 activity
(Figure 1). The first such compound was 3ꢀdeazaneplanocin A (1, DZNep) which interferes with
Sꢀadenosylꢀlꢀhomocysteine (SAH).8 Although the compound has very short plasma halfꢀlife,
significant antitumor activity in various cancer types was reported, along with toxicities possibly
due to nonspecific inhibition of histone methylation. This result further spurred interest in
development of specific PRC2 inhibitors. In 2012, Epizyme and GSK reported the development
of Sꢀadenosylmethionine (SAM)ꢀcompetitive inhibitors 2 (EPZ005687)9 and 3 (GSKꢀ126),10
respectively, resulting from optimization of hits identified from highꢀthroughput screenings.
Compound 3 markedly inhibits the growth of lymphomas carrying activating EZH2 mutations in
vivo. Following this, compound 4 (UNC1999) was reported as the first orally bioavailable EZH2
ACS Paragon Plus Environment
4

Journal of Medicinal Chemistry
Page 6 of 34
1
2
3
4
5
6
7
8
9
inhibitor that was highly selective for both wildꢀtype and Y641 mutant EZH2, and as well as
EZH1.11 Shortly after, compound 5 (EPZꢀ6438) was reported as Epizyme’s second generation
EZH2 inhibitor with better potency and good oral bioavailability.12 Based on the encouraging
efficacies in preclinical studies, phase 1/2 clinical trials of compound 5 in advanced solid tumors
and B cell lymphomas was launched in June 2013. In the meantime, compound 3 and compound
6 (CPIꢀ1205)13 from Constellation also entered clinical trials in EZH2ꢀmutant B cell NHL and
SMARCB1ꢀdeficient tumors, respectively. With three independent EZH2 inhibitors in clinical
trials, a trove of data is expected on both potential toxicity and efficacy of this approach.
However, all three PRC2 inhibitors currently in clinic (compound 5: NCT01897571,
NCT02601937, NCT02601950; compound 3: NCT02082977; compound 6: NCT02395601) are
pyridoneꢀderived SAMꢀcompetitive inhibitors targeting EZH2 with relative weaker activity
against EZH1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
It has been reported that binding of H3K27Me3 to EED allosterically activates the
methyltransferase activity of PRC2.6a This could partially be explained by the recently published
high resolution crystal structures of PRC2 complex, which suggested that this interaction induces
conformational change of stimulationꢀresponsive motif (SRM) in EZH2, leading to enhanced
catalytic efficiency.14 We hypothesized that low molecular weight (LMW) compounds
interacting with PRC2 via the “Me3 pocket” of EED may inhibit the methyltransferase activities
of both PRC2ꢀEZH2 and PRC2ꢀEZH1, and therefore may provide therapeutic(s) similar or
complementary to the EZH2 inhibitors in clinic.
ꢀ RESULTS and DISCUSSION
In order to find LMW compounds that antagonize PRC2 complex activity, we initiated a highꢀ
throughput screening (HTS) campaign, which used the recombinant 5ꢀmember PRC2 complex as
ACS Paragon Plus Environment
5

Page 7 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
enzyme, H3[21ꢀ44, K27Me0] as a peptide substrate, and homogeneous time resolved
fluorescence (HTRF) method to detect the dimethylated product. The mechanisms of action of
the screening hits were further elucidated by various competition studies using SAM,
H3K27Me0 substrate peptide, and/or K27Me3 stimulation peptide. Compound 7 (Figure 2a) was
identified as a K27Me3ꢀcompetitive inhibitor in biochemical assays, and its binding to EED
through K27Me3 pocket via the “aromatic cage” was further confirmed by Xꢀray crystallography
(PDB accession code: 5H19, details of the hits finding are reported in a separate manuscript).15,18
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2: (a) Chemical structure of HTS hit compound 7. (b) More significant movement of key residues upon
binding of compound 7 compared to binding of K27Me3 (green: compound 7, PDB accession code: 5H19; blue:
EED residues binding to compound 7; yellow: EED residues binding to K27Me3; K27Me3 peptide is not showed
for the purpose of clarity; movement of the residues are indicated by red arrows).
Although the dynamic nature of the Me3 pocket, especially the flexibility of the sideꢀchains of
Trp364 and Arg367, has been previously reported by comparing the K27Me3 bound structure to
apo EED protein structure,6c binding of compound 7 induced more substantial conformational
change of sideꢀchains of Trp364 and Tyr365, in addition to Arg367 (Figure 2b). As a result, a
much deeper pocket was induced with the calculated druggability score improved from
previously reported 0.6416 to 0.96. Encouraged by this finding, we started our discovery efforts
to optimize compound 7 in order to discover potential LMW therapeutic agents targeting the
K27Me3 pocket of EED.
ACS Paragon Plus Environment
6

Journal of Medicinal Chemistry
Page 8 of 34
1
2
3
4
5
6
7
8
9
During the SAR investigation, biological activities of our compounds were assessed in a cascade
of assays (Supporting Information). To determine the binding affinity of the compounds to EED,
we developed an EEDꢀH3K27me3 peptide alphascreen binding assay and performed competition
studies with our compounds. At the same time, the biochemical inhibitory activities of the
compounds were evaluated with a liquid chromatography mass spectrometry (LCꢀMS) based
assay which detects SAH formation using either H3K27Me0 peptide or nucleosome as
substrate.17 Since results from the two assays correlate reasonably well, in this publication the
SAR is discussed only with biochemical activities although alphascreen assay was routinely ran
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
as part of SAR support 18
. Furthermore, the ability of compounds to reduce global K27Me3 at
cellular level was assessed in G401 cells with an ELISA assay. Antiꢀproliferative activity of
selected compounds was determined in KARPASꢀ422 cells, a diffuse large Bꢀcell lymphoma
(DLBCL) cell line harboring a monoallelic Y641N EZH2 mutation. Since we also observed the
reported slow kinetics of the antiproliferative activity for SAMꢀcompetitive EZH2 inhibitors,10,
12b IC50’s at 14 day were required to differentiate in vitro antiꢀtumor activity for the compounds.
Compound 7, with IC50 of 0.62 µM in alphascreen binding assay, IC50 of 4.03 µM in biochemical
assay with H3K27Me0 peptide as substrate, and ELISA IC50 of 3.28 µM in G401 cells,
represented an excellent starting point for us to start our optimization efforts.
Fragmentation and regrowth: bicyclic triazolopyridine as scaffold I
The key interactions in the binding of compound 7 include two hydrogen bonds and a set of πꢀπ
interactions. The two hydrogen bonds, one between hydrogen on 5’ꢀamino and the side chain
carbonyl of Asn194 (red dash, Figure 3a), and another one between nitrogen at 2ꢀpostion with
sideꢀchain of Lys211, appear to offer key polar interactions. The πꢀπ interactions, taking full
advantage of the aromatic cage residues, also play indispensable roles. Firstly, the electronꢀ
ACS Paragon Plus Environment
7

Page 9 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
deficient bicyclic [1,2,4]triazolo[4,3ꢀa]pyridine core is held in place by πꢀπ stacking interactions
with the electronꢀrich Tyr148 and Tyr365. Secondly, the electronꢀrich furan forms a πꢀπ stacking
interaction with guanidinium group of Arg367, and an edgeꢀtoꢀface interaction with Tyr365,
which we referred to as “deep pocket” interactions, since the furan group is located in the inner
part of the induced pocket. Additionally, the CN group at C6 seems interaction with multiple
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3: Structural basis for fragmentation and regrowth: (a) Key interactions of compound 7 with EED ((PDB
accession code: 5H19) (b) Compound 8 bounds to EED (PDB accession code: 5H24) (c) Overly of compound 7 and
11 bound to EED (PDB accession code: 5H25) (green: compound 7; magenta: compound 11)
residues through a water network. (Figure 3a, see Supporting Information (Figure S1) for details
of the interactions). Upon closer inspection of the interactions between compound 7 and EED, it
appeared to us that the entire piperidine ring connecting C7 to C8 and the benzyl group attached
ACS Paragon Plus Environment
8

Journal of Medicinal Chemistry
Page 10 of 34
1
2
3
4
5
6
7
8
9
to nitrogen at 9ꢀposition did not contribute much to the interaction, and likely reduced efficiency
of binding due to the nonꢀessential lipophilicity. Compound 8, a fragment of compound 7 (Table
1), which was confirmed to retain most of the key interactions with EED (Figure 3b), is
equipotent as compound 7. This transformation led to dramatic improvements of both ligand
efficiency (LE) and lipophilic efficiency (LipE) (Table 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
With what appeared to be the minimal pharmacophore required for binding, we set out to
improve potency for 8. The coꢀcrystal structure of compound 8 with EED suggested C8 to be a
more attractive vector for additional substitution as compared to C7 (red arrow, Figure 3b).
Meanwhile, the surprising activities of compound 11 (Figure 3c), another HTS hit, caught our
attention. The two key features critical to the binding to EED, the electronꢀrich “deep pocket”
aryl group and hydrogen bond donating NH at 5’ꢀposition, are missing in compound 11; and the
importance of “the deep pocket” was attested by total loss of activities in compound 12 (Figure
3c). We postulate that an edgeꢀtoꢀface interaction between the phenyl ring of compound 11 and
Phe97 (Figure 3c) possibly contributed most to its activity in the absence of “the deep pocket”
group. Furthermore, when the coꢀcrystal structure of compound 11 was overlaid with that of
compound 7, we hypothesized that the aryl substitution at 8ꢀposition of compound 8 would
mimic that of compound 11 and help us pick up this edgeꢀtoꢀface interaction. Compound 9
(Table 1) was therefore designed and synthesized, and a close to 20ꢀfold improvement in
biochemical activity was achieved, with slight decrease in LE and LipE. Substitutions at C8
position are mostly solvent exposed and achieving additional interactions with the protein was
challenging, however some interesting SAR were observed. Addition of
a
(dimethylamino)methyl group at 4’ꢀposition of C8 aryl group (compound 10), improved
ACS Paragon Plus Environment
9

Page 11 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
biochemical potency 7ꢀfold, and this effect was also observed with similar amino tails and at
both 4’ꢀ and 3’ꢀposition of the C8 aryl group (unpublished results).
7
8
9
Compounds in table 1 were synthesized according to the reaction sequence depicted in Scheme
1. 2ꢀCyanoacetamide was condensed with ethyl propiolate under basic condition to afford
dihydroxylated pyridine compound 14, which was brominated to afford compound 15.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1: Discovery of bicyclic 6ꢀCN triazolopyridine lead compound 9: fragmentation of hit compound 7
and regrowth
BioChem
LipE
(BioChem)d
LE
(BioChem)c
compound
Structure
clogPb
(IC50, ꢀM)a
4.03
1.00
2.87
0.98
2.87
0.26
2.6
5.0
4.4
7
8
9
0.46
0.42
CN
H
O
N
0.062
N
N
N
0.009
2.70
0.40
5.3
10
aIC50 values reported as an average ≥2 determinations; see Supporting Information for further details; bclogP values
generated via ChemBioDraw version 14.0.0.117; cLE = ligand efficiency = pIC50 – HAC; dLipE = lipophilic
efficiency = pIC50 − cLogP .
compound 15 was treated with POCl3, followed by aqueous hydrazine solution and then cyclized
to give the desired bicyclic 8ꢀbromoꢀ[1,2,4]triazolo[4,3ꢀa]pyridine derivative 16. Displacement
with furanꢀ2ꢀylmethanamine at room temperature in ethanol afforded compound 17, which was
converted to desired target compound 9 after Suzuki coupling with phenylboronic acid.
ACS Paragon Plus Environment
10

Journal of Medicinal Chemistry
Page 12 of 34
1
2
3
4
5
6
7
8
9
Scheme 1: Synthesis of 5ꢀ((furanꢀ2ꢀylmethyl)amino)ꢀ8ꢀphenylꢀ[1,2,4]triazolo[4,3ꢀa]pyridineꢀ6ꢀ
carbonitrile
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Reagents and conditions: (a) NaOMe, ethyl propiolate, 90°C, 2h, 79%; (b) Br2, acetic acid, 63%; (c) POCl3, 160°C,
2h, 27%; (d) NH2NH2, 0°C, 93%; (e) CH(OMe)3, TFA, 0°C, 67%; (f) furanꢀ2ꢀylmethanamine, EtOH, rt, 73%; (g)
phenylboronic acid, PdCl2(dppf), NaHCO3, dioxaneꢀH2O, 90oC overnight, 10%.
Core permutation: [1,2,4]triazolo[4,3-c]pyrimidine core as scaffold II
To continue our SAR exploration, we carried out systematic scaffold hopping efforts of the [6, 5]
bicyclic core (Table 2). Removal of the cyano group at the 6ꢀposition of compound 9 and
replacing it with nitrogen led to compound 18 with similar potency and slightly improved LE
and LipE, which is possibly due to a newly formed waterꢀmediated hydrogen bond network
(supporting information, Figure S1). Having the nitrogen atom at 6ꢀposition seems to be critical
to the binding. Shifting the nitrogen to 7ꢀposition as in compound 22 decreased activity over 60ꢀ
fold vs. compound 18, and removing this nitrogen resulted in a potency decrease over 50ꢀfold
(compound 21 vs. compound 20). The accessibility of nitrogen at the 2ꢀposition as a hydrogen
bond acceptor to Lys211 also plays a pivotal role in binding; removing this nitrogen resulted in a
400ꢀfold drop in potency (compound 19 vs. compound 18). Although the nitrogen at 1ꢀposition
does not seem to contribute to meaningful polar interaction judging from Xꢀray structures,
ACS Paragon Plus Environment
11

Page 13 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
removing the nitrogen let to a 5ꢀfold drop in potency from compound 18 to 21, which is possibly
due to less effective πꢀπ interactions with a less electronꢀdeficient bicyclic core. In addition to
pyrimidineꢀbased bicyclic core, we also investigated their pyrazineꢀbased counterparts, such as
compounds 22 and 23. In all the cases we examined, alteration of the heterocyclic core of 18 was
detrimental to potency. Nonetheless, we had two very promising scaffolds in our hand, namely,
8ꢀarylꢀ[1,2,4]triazolo[4,3ꢀa]pyridine (Scaffold I) and 8ꢀarylꢀ[1,2,4]triazolo[4,3ꢀc]pyrimidine
(Scaffold II) for further optimization, and we were intrigued to determine if the SAR at the Cꢀ
ring and Dꢀring were transferable between the two scaffolds.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2: Discovery of bicyclic 6ꢀCN triazolopyridine lead compound 9: fragmentation of hit compound 7
and regrowth
BioChem
LipE
(BioChem)d
LE
(BioChem)c
compound
Structure
clogPb
2.87
(IC50, ꢀM)a
0.062
0.052
0.41
0.45
4.3
4.5
9
2.78
18
0.21
0.22
19
20
20.49
14.80
3.54
4.18
1.2
0.55
ACS Paragon Plus Environment
12

Journal of Medicinal Chemistry
Page 14 of 34
1
2
3
4
5
6
7
8
21
22
23
0.26
3.52
2.18
3.54
2.99
3.75
0.30
0.25
0.26
3.0
2.5
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1.91
aIC50 values reported as an average ≥2 determinations; see Supporting Information for further details; bclogP values
generated via ChemBioDraw version 14.0.0.117; cLE =ligand efficiency=pIC50 – HAC; dLipE = lipophilic efficiency
= pIC50 − cLogP .
Synthesis of the analogs for 8ꢀarylꢀ[1,2,4]triazolo[4,3ꢀc]pyrimidine (Scaffold II) is detailed in
Scheme 2. Commercially available 5ꢀbromoꢀ4ꢀchloroꢀ2ꢀ(methylthio)pyrimidine (compound 24)
was treated with hydrazine, followed by trimethyl orthoformate to afford bicyclic compound 26,
which was transformed to compound 27 after selective displacement with furanꢀ2ꢀ
ylmethanamine at the 5ꢀposition. Compound 27 was further converted to the 8ꢀphenyl analog 28
after Suzuki reaction. In some cases when stronger bases were used, Dimroth rearrangement19
product 29 was also observed. By using NaHCO3, we were able to prepare compound 28 in
multiꢀgram scales with minimum formation of side product 29. Compound 29 can be separated,
and its identity was distinctive from singlet at δ 8.50 in 1H NMR accounting for the proton at C2.
When profiled in our biochemical assay, compound 29 lost most of the activity of its isomer 28,
which is consistent with the notion that the Hꢀbond acceptor to Lys211 at the 1ꢀposition of the
bicyclic system is important to maintain the interaction with EED.
ACS Paragon Plus Environment
13

Page 15 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Syntheses of the compounds in Table 2 with bicyclic cores other than [1,2,4]triazolo[4,3ꢀ
a]pyridineꢀ6ꢀcarbonitrile and [1,2,4]triazolo[4,3ꢀc]pyrimidine are reported in the supporting
information.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Nꢀ(furanꢀ2ꢀylmethyl)ꢀ8ꢀphenylꢀ[1,2,4]triazolo[4,3ꢀc]pyrimidinꢀ5ꢀamine
Reagents and conditions: (a) hydrazine, rt, 4h, 92%; (b) CH(OMe)3, reflux, 3h, 92%; (c) furanꢀ2ꢀylmethanamine, rt,
16.7%; (d) phenylboronic acid, PdCl2(dppf), NaHCO3, dioxaneꢀH2O, 90oC overnight, 10%.
SAR investigation of “the deep pocket”
Next, we turned our attention to the deep pocket or Cꢀring SAR for both scaffold I and II, in
search for a replacement for the potentially metabolically labile furan ring.20 Electronꢀrich 5ꢀ
membered rings, such as furan, are preferred, regardless of the position of the oxygen as seen in
compounds 10a & 10b vs. 31a & 31b (Table 3). Thiophene isosteres are generally tolerated,
albeit slightly less potent than their furan counterparts as in compound 30a & 30b. Compounds
bearing less electronꢀrich Cꢀrings, such as oxazole, isoxazole and thiazole, are generally less
potent. Overall, SAR of the Cꢀring is quite similar in both scaffolds until substitutions were
introduced onto the ring. In both scaffolds, small substitution at the 5’ꢀposition of C ring, such
as F (39a & b) is more tolerated than the larger substitutions (40a & b). The two scaffolds
started to behave quite differently when substitutions were introduced at 3’ꢀposition; the
ACS Paragon Plus Environment
14

Journal of Medicinal Chemistry
Page 16 of 34
1
2
3
4
5
6
7
8
9
substitutions were tolerated in Scaffold II, but not tolerated in Scaffold I, when comparing 37b
vs 37a, and 38b vs 38a. The same differences were also observed in 6ꢀmembered ring, as in
compound 39a-40a vs 39a-40b. We believe in Scaffold I there is a steric clash between the CN
and the 3’ꢀsubstituted on the C ring that resulted in the subtle difference in SAR between
Table 3: Investigation of SAR in “the deep pocket” (Cꢀring)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CN
CN
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Scaffold II
Scaffold II
Scaffold I
Scaffold I
BioChem
(IC50, µM)a
BioChem
(IC50 ,µM)a
BioChem
(IC50 ,µM)a
BioChem
(IC50 ,µM)a
R1
Cmpd
10a
Cmpd
R1
Cmpd
Cmpd
0.013
0.020
0.013
0.093
0.51
2.22
0.15
12.4
10b
30b
35a
36a
35b
36b
30a
0.009
0.15
0.019
0.29
21.8
2.34
0.049
0.053
31a
32a
31b
32b
37a
38a
37b
38b
S
H
N
Cl
H
0.069
0.069
0.77
0.13
21.8
14.9
0.049
0.094
33a
34a
33b
34b
39a
40a
39b
40b
N
aIC50 values reported as an average ≥2 determinations; see Supporting Information for further details.
scaffold I and II. At this point, we concluded this particular quest without finding a viable
replacement for furan in the deep pocket.
Selection of compound for pharmacology studies
ACS Paragon Plus Environment
15

Page 17 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
We next turned our attention to optimization of the Dꢀring. For reasons unclear to us, the gain of
potency in biochemical assay from varying solventꢀexposed Dꢀring mostly diminished in the 14ꢀ
day antiꢀproliferative activities in KARPASꢀ422 (Table 4). We therefore relied on overall
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4. Selection of in vivo candidate based on compound potency and pharmacokinetic profile
Compound
41
42
43
5
structures
IC50
0.007
0.37
0.030
0.41
0.022
0.37
0.002
0.28
(ꢀM)a
LEb
BioChem
LipEc
4.4
4.5
0.21
0.076
6.4
0.22
0.080
5.3
G401Elisa
0.032
0.088
0.040
0.012
(IC50, ꢀM)d
KARPAS 14D IC50
(ꢀM)e
MLM (hER )f
0.52
38.1
0.80
2.81
0.36
14.4
0.77
1.4
fu (% free)g
dnAUCꢀiv (nM*hr)h
765
1751
534
1303
Wholei
Freej
531
202
714
5648
813
254
4
dnAUCꢀ
po
(nM*hr)
20.1
Cl (ml/min/kg)k
Vdss (L/kg)l
T 1/2 (hr)m
57
1.9
0.6
71
29
0.7
0.4
42
8
22
1.2
1.6
20
0.8
2.2
109
F% (po)n
aIC50 values reported as an average of more than determinations see Supporting Information for further details; bLE
= ligand efficiency = pIC50 – HAC; cLipE = lipophilic efficiency = pIC50 – cLogP; dGlobal H3K27me3 levels
measured in G401 cells and IC50 values reported as an average of more than two determinations; eIC50 value reported
for compound 41 and 42 as single determination, while that of compounds 43 reported as an average of 10
measurements with SD = 0.019 and that of 5 as an average of 2 measurement with SD = 0.006 ; f Hepatic extraction
ratio of compounds in mouse liver microsomes; gFraction unbound to plasma protein (expressed in % unbound);
hDoseꢀnormalized total exposure following intravenous (iv) dosing of 1 mg/kg in male CDꢀ1 mice, formulated in
PEG300:Soltuol HS15:pH4.65 acetate buffer (20:10:70, v/v/v); iDoseꢀnormalized total exposure following per os
ACS Paragon Plus Environment
16

Journal of Medicinal Chemistry
Page 18 of 34
1
2
3
4
5
6
(po) dosing of 2 mg/kg in male CDꢀ1mice, formulated in PEG300:Soltuol HS15:pH4.65 acetate buffer (20:10:70,
k
v/v/v); jDoseꢀnormalized free AUC following po dosing; Cl = plasma clearance; lVdss = volume of distribution at
n
steady state; mTerminal halfꢀlife; Oral bioavailability.
7
8
9
pharmacokinetic profiles to select our compound that would enable headꢀtoꢀhead comparison
with the reported EZH2 inhibitor(s) in mouse efficacy studies.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compounds were evaluated in mouse liver microsome (MLM) assays and also in male CDꢀ1
mice for in vivo PK properties at dosed at 1 mg/kg intraveinous (iv) and 2 mg/kg orally (po).
Overall, hepatic extraction rates in MLM predicted fairly well the in vivo total clearance of the
compounds, suggestive of hepatic clearance being the major clearance pathway. Among the three
EED compounds shown in Table 4, compound 43 (Nꢀ(furanꢀ2ꢀylmethyl)ꢀ8ꢀ(4ꢀ
(methylsulfonyl)phenyl)ꢀ[1,2,4]triazolo[4,3ꢀc]pyrimidinꢀ5ꢀamine) stood out as having very low
in vivo and in vitro clearance, and approximately 100% oral bioavailability. Compound 43 has
low volume of distribution (0.8 L/kg), reasonable terminal t1/2 (2.2 h), and moderate plasma
protein binding (PPB), approximately 10ꢀfold greater free fraction than that of compound 5 at
14.4% & 1.5%, respectively. Together this led to free normalized AUC free in CDꢀ1 mice at 813
nM*hr in comparison to that at 4 nM*hr for compound compound 5. This 200ꢀfold higher freeꢀ
exposure of compound 43 counters well against the 8ꢀfold lower antiꢀproliferation activity in the
14ꢀday assay than compound 5. In CDꢀ1 mouse, compound 43 achieved sufficient targets
coverage for at least 8 hours (Supporting Information, Figure S2) at 2 mg/kg. Taken together, the
data suggest that compound 43 constitutes a reasonable candidate for pharmacology studies.
Efficacy studies
Compound 43 was then profiled at high doses in mouse pharmacokinetic studies. In the
subsequent tolerability test in CDꢀ1 mice, dosing of compound 43 for 14 days at 300 mg/kg bid
was well tolerated with no apparent adverse effects (Supporting Information, Figure S39(a)).
Further pharmacology studies were carried out in the KARPASꢀ422 derived DLBCL model in
ACS Paragon Plus Environment
17

Page 19 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Balb/C nude mice. Complete tumor regression (10/10 CR) was observed with treatment of
compound 43 at 300 mg/kg bid for 34 days (Figure 4). After treatment of compound 43 was
stopped, no tumor regrowth was observed in 4 out of the 5 animals kept for observation during
the entire 5 months. With the single regrowth that was observed at approximately 130 days,
spontaneous regression occurred (Figure 4).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Pharmacology studies of compound 43 in KARPASꢀ422 derived DLBCL model in Balb/C nude mice at
300 mg/kg (bid, suspension, n =10)
We next performed dose deꢀescalation studies to determine the minimal efficacious dose and to
establish dose dependent efficacy (Supporting Information, Figure S3(b)). After two weeks of
treatment, differentiated growth patterns started to emerge between the lower dose groups (1.5 &
4 mg/kg) and higher dose group (12 & 40 mg/kg). Gobal H3K27Me3 levels in tumors was used
as our pharmacodynamics markers, which decreased doseꢀdependently along with the target
genes upregulation after 21 days of compound dosing, and all showing good correlation with the
plasma exposure of compound 43.18
Compound 43 is a highly selective EED inhibitor with drug-like properties
To confirm the onꢀtarget binding of compound 43 to the H3K27Me3 pocket, we obtained high
resolution crystal structures of EED (76ꢀ441) in complex with compound 43 and EZH2 (40ꢀ68)
peptide (EBD) (PDB accession code: 5GSA). Compound 43 retained all the interactions of its
parent compound 7 (supra vide) with EED, while in addition the phenyl ring of the mostly
ACS Paragon Plus Environment
18

Journal of Medicinal Chemistry
Page 20 of 34
1
2
3
4
5
6
7
8
9
solvent exposed pꢀmethylsulfonylphenyl group picks up edgeꢀtoꢀface πꢀπ interaction with sideꢀ
chain of Phe97. Compound 43 was subsequently profiled in our inꢀhouse histone
methyltransferase panel including close to 30 HMT targets. Compound 43 is highly selective
against all other targets (>10,000 fold) except EZH1, with almost identical potency as EZH2. In
our safety pharmacology panel including 59 targets,21 compound 43 shown <50% inhibition at
the highest concentration tested (10 µM or 30 µM). IC50’s in hERG binding assay and QPatch
assay are both well above 30 µM.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4. Characterization of compound 43
Solubility
Cacoꢀ2
MP (oC)
with crystalline material (mg/mL)
HTꢀ
MW
clogPa
1.22
(A>B, 10ꢀ6cm/s /
ratio)
Sol
SGFc
Fassifd
(mM)b
pH7
pH3
369.4
207.6
3.0 / 7.6
0.012
0.025
0.026
0.055
0.035
In vitro ADME
In vivo PK
Mouse
Rat
Dog
In vitro LM
dnAUC(nM*h)
(p.o.)h
36/14/41/27/25
5648
1871 1907
(hER%, M/R/D/CM/H)e
In vitro hepatocytes (hER%,
M/R/D/CM/H)f
25/22/51/30/20
15/8/24/13
Cl (mL/min/kg)i
F (%)j
8
14
60
16
68
fu (M/R/D/H)g
109
b
c
aclogP values generated via ChemBioDraw version 14.0.0.117; Equilibrium solubility; simulated gastric fluid;
dfasted state simulated intestinal fluid; hHepatic extraction ratio of compound in mouse (M), rat (R), dog (D), cynoꢀ
monkey (CM) and human (H) liver microsomes; Hepatic extraction ratio of compound in mouse (M), rat (R), dog
f
g
(D), cynoꢀmonkey (CM) and human (H) hepatocytes; Unbound fraction of compound to mouse (M), rat (R), dog
(d) and human (h) plasma protein; hDoseꢀnormalized free AUC after po dosing, 2 mg/kg in male Sprague Dawley rat
and 2 mg/kg in male Beagle dog, formulated with PEG300:Soltuol HS15:pH4.65 acetate buffer (10:10:80, v/v/v);
j
iCl = total clearance; Oral bioavailability.
Compound 43 was further evaluated for its physicochemical properties and other DMPK
parameters (Table 4). The solubility was relatively low, and with little dependency on the pH of
the medium. Compound 43 has moderate permeability as the measured in Cacoꢀ2 cells at A→B
ACS Paragon Plus Environment
19

Page 21 of 34
Journal of Medicinal Chemistry
1
2
= 3.0 x 10ꢀ6 cm/s, with an efflux ratio at 7.6. When iv/oral PK studies were performed in rat and
dogs (Table 5), we found that the in vivo clearances measured were well predicted by in vitro
incubation experiments with liver microsome and hepatocytes. Given low in vitro clearance
across all preclinical species, and the high predictive value of in vitro hepatocytes assays for
compound 43, clearance in human was projected to be low (< 5 mL/min/kg, ER <25%). The
promising pharmacokinetic profile of compound 43 in all animal species tested, together with the
results from its in vitro safety profile and superb efficacy in the xenograft model, support
compound 43 for further evaluation.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ CONCLUSIONS
Herein, we report discovery of compound 43, a firstꢀinꢀclass EED inhibitor. Compound 43 is a
highly potent, efficient and selective inhibitor of EZH2 and EZH1 evaluated against a broad
range of epigenetic and nonꢀepigenetic targets. This inhibitor potently reduced global
H3K27Me3 mark in cells and demonstrated selectively cell killing effects in cells carrying a
heterozygous Y641N mutation. With favorable pharmacokinetic properties, compound 43
demonstrated very impressive antiꢀtumor activities in mouse xenograft model. For the first time
we have demonstrated that inhibiting methyltransferase activity of PRC2 through binding to the
K27Me3 pocket of EED could constitute a viable strategy for developing antiꢀcancer
therapeutics. Together with extensive preclinical profiling results of compound 43, especially its
favorable pharmacokinetic profile and tolerability in preclinical animal species, compound of
this type may be suitable for assessing longꢀterm effects of pharmacologically inhibiting both
EZH2 and EZH1 in preclinical models, and potentially in clinic, in terms of therapeutic benefits
and potential toxicity.
ꢀ EXPERIMENTAL SECTION
ACS Paragon Plus Environment
20

Journal of Medicinal Chemistry
Page 22 of 34
1
2
3
4
5
6
7
8
9
Synthesis and Characterization: Unless otherwise mentioned, all reagents and solvents were
obtained from commercial sources and used without purification. In some cases, intermediates
were characterized by LCꢀMS to confirm that the mass matched the structure and carried on to
the next step without further purification. Air sensitive procedures were performed under an
atmosphere of nitrogen or argon. Purification of the final compounds to > 95% purity was
carried out either using prepacked silica gel cartridge (Analogix. Biotage or ISCO) or reverse
phase C18 column. 1H (300 MHz) and 13C (100 MHz) NMR spectra were recorded on a Brucker
AC300 spectrometer. NMR chemical shift (δ) are quoted in parts per million (ppm) referenced
to the residual solvent peak [DMSOꢀd6] set at 2.49 ppm or [CDCl3] set at 7.26 ppm. Purity of all
compounds was determined to be >95% by analytical HPLC.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5-((Furan-2-ylmethyl)amino)-8-phenyl-[1,2,4]triazolo[4,3-a]pyridine-6-carbonitrile (9): 8ꢀ
Bromoꢀ5ꢀ((furanꢀ2ꢀylmethyl)amino)ꢀ[1,2,4]triazolo[4,3ꢀa]pyridineꢀ6ꢀcarbonitrile (compound 17)
(40 mg, 0.13 mmol), phenylboronic acid (18.40 mg, 0.15 mmol), PdCl2(dppf) (4.60 mg, 6.3
µmol) and NaHCO3 (31.7 mg, 0.38 mmol) in dioxane (2 mL) and H2O (1 mL) was heated to at
90oC under N2 overnight. The resulting mixture was cool to room temperature, and H2O was
added to the mixture and extracted with EtOAc. The combined organic layers were dried
(Na2SO4), filtered and concentrated. The residue was purified by basic prepꢀHPLC to afford 3.8
mg (10%) of the title compound. 1HꢀNMR (400 MHz, DMSOꢀd6) δ ppm 9.45 (s, 1 H), 8.87 (t, 1
H), 8.12 (d, J = 1.2 Hz, 1 H), 8.11 (s, 1 H), 7.68 (d, J = 0.8 Hz, 1 H), 7.66 (s, 1 H), 7.38 ꢀ 7.48
(m, 3 H), 6.46 ꢀ 6.54 (m, 2 H), 5.03 ꢀ 5.05 (d, J = 6 Hz, 2 H). LCꢀMS (m/z): 315.9 [M+H]+
2,6-Dihydroxynicotinonitrile (14): To a suspension of 2ꢀcyanoacetamide (250 g, 2.55 mol) in
methanol (1000 mL) was added NaOMe (137.8 g, 2.55 mol) at 0°C under N2. The reaction
mixture was stirred at the same temperature for 30 min. Ethyl propiolate (214.4 g, 2.55 mol) was
ACS Paragon Plus Environment
21

Page 23 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
added dropwise at 0°C. The reaction mixture was stirred at 0°C for 2 h, then at 90°C for 2 h. The
reaction mixture was then cooled to rt, and the resulting yellow paste was filtered. The filter cake
was suspended with water and neutralized with concentrated HCl till pH reached ~6. A mixed
solvent of DCMꢀMeOH (v/v = 20/1) was added and stirred at 0°C for 30 min. The solid was
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
filtered and washed with Et2O to give 274 g (79 %) of the title compound as yellow solid. Hꢀ
NMR (400 MHz, DMSOꢀd6) δ ppm 9.88 (s, 1 H), 7.16 (d, J = 9 Hz, 1 H), 5.06 (d, J = 9 Hz, 1 H).
LCꢀMS (m/z): 137.1 [M+H]+
5-Bromo-2,6-dihydroxynicotinonitrile (15): A suspension of 2,6ꢀdihydroxynicotinonitrile (274
g, 2 mol) in acetic acid (1500 mL) was stirred at 70 °C for 30 min. To the mixture at 10°C was
added a solution of Br2 (100 mL, 2 mol) in acetic acid (100 mL). The resulting mixture was
stirred for 20 min. The solid was filtered to give 271 g (63%) of the title compound as light
yellow solid. 1HꢀNMR (400 MHz, DMSOꢀd6) δ ppm 7.86 (s, 1 H). LCꢀMS (m/z): 212.9 [M+H]+
8-Bromo-5-chloro-[1,2,4]triazolo[4,3-a]pyridine-6-carbonitrile (16): A solution of compound
15 (130 g, 0.6 mol) in POCl3 (500 mL) was sealed and heated at 160oC for 2 h. The mixture was
cooled to room temperature and concentrated to remove extra POCl3. Ethyl acetate (2000 mL)
and iceꢀwater (2000 mL) was added to the residue and the resulting mixture was neutralized with
aqueous NaHCO3 until pH reached ~7. The aqueous layer was separated and extracted with ethyl
acetate. The combined organic layers were washed with brine, dried (Na2SO4), filtered and
concentrated. The mixture was purified with silica gel column chromatograph (eluted with PEꢀ
EA: 50/1) to give 41 g (27%) of the 5ꢀbromoꢀ2, 6ꢀdichloronicotinonitrile compound as light
yellow solid. 1HꢀNMR (400 MHz, DMSOꢀd6) δ ppm 9.10 (s, 1 H). LCꢀMS (m/z): 250.5 [M+H]+
8-Bromo-5-((furan-2-ylmethyl)amino)-[1,2,4]triazolo[4,3-a]pyridine-6-carbonitrile (17): To
a mixture of compound 16 (1.03 g, 4 mmol) in EtOH (7 mL) at room temperature was added
ACS Paragon Plus Environment
22

Journal of Medicinal Chemistry
Page 24 of 34
1
2
3
4
5
6
7
8
9
furanꢀ2ꢀylmethanamine (0.855 g, 8.80 mmol) in EtOH (1 mL) dropwise. The resulting mixture
was stirred at room temperature for 2 h under N2. After the starting material was consumed,
DCM (20 mL) was added. Solvent was removed under vaccum after silica gel (~4 g) was added.
The silica gel coated with crude product was loaded on a silica gel column and eluted with
DCMꢀMeOH to give 1.2 g (94%) of the title compound as a light green solid. 1HꢀNMR
(400MHz, DMSOꢀd6) δ ppm 9.63 (s, 1 H), 8.85 (s, 1 H), 7.73 (s, 1 H), 7.66 (d, J = 0.8 Hz, 1 H),
6.51 (d, J = 2.8 Hz, 1 H), 6.44 (dd, J = 3.2, 2 Hz, 1 H), 4.97 (s, 2 H). LCꢀMS (m/z): 318.0.
[M+H]+
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5-Bromo-4-hydrazinyl-2-(methylthio)pyrimidine (25): To a solution of 5ꢀbromoꢀ4ꢀchloroꢀ2ꢀ
(methylthio)pyrimidine (24, 49.0 g, 0.205 mol) in ethanol (1000 mL) was added hydrazine (21.5
g, 0.43 mol). The reaction was stirred at room temperature for 4 h. The resulting suspension was
filtered, washed with hexane and dried in vacuo to give the title compound (44.1 g, 92%) as a
1
white solid. HꢀNMR (400 MHz, DMSOꢀd6) δ ppm 8.08 (s, 1H), 2.42 (s, 3H). LCꢀMS (m/z):
234.9, 236.9 [M+H]+
8-Bromo-5-(methylthio)-[1,2,4]triazolo[4,3-c]pyrimidine (26): Compound 25 (40.0 g, 0.17
mol) was dissolved in 200 mL triethoxymethane. The mixture was heated at reflux and stirred for
3 h. The reaction mixture was concentrated under reduced pressure, and the residue was purified
by flash chromatography (eluted with EA: PE =1:15~1:1) to give the title compound (38.3 g,
92%) as a white solid. 1HꢀNMR (400 MHz, methanolꢀd4) δ ppm 8.87 (s, 1H), 8.03 (s, 1H), 2.82
(s, 3H). LCꢀMS (m/z): 245.0, 247.0 [M+H]+
8-Bromo-N-(furan-2-ylmethyl)-[1,2,4]triazolo[4,3-c]pyrimidin-5-amine (27): 8ꢀBromoꢀ5ꢀ
(methylthio)ꢀ[1,2,4]triazolo[4,3ꢀc]pyrimidine (26) (100 mg, 0.408 mmol) was dissolved in furanꢀ
2ꢀylmethanamine (1.5 mL). The mixture was stirred at rt for 1 h. The solvent was removed and
ACS Paragon Plus Environment
23

Page 25 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the residue was purified with silica gel column chromatography, eluted with DCMꢀMeOH
(20:1), to give 20 mg (16.7%) of the title compound. 1HꢀNMR (400 MHz, CD3OD) δ ppm 9.28
(s, 1H), 7.87 (s, 1H), 7.47 (s, 1H), 6.38 ꢀ 6.40 (m, 2H), 4.77 (s, 2H). LCꢀMS (m/z): 294.0, 296.0
[M+H]+
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N-(Furan-2-ylmethyl)-8-phenyl-[1,2,4]triazolo[4,3-c]pyrimidin-5-amine (28): To a solution
of 8ꢀbromoꢀNꢀ(furanꢀ2ꢀylmethyl)ꢀ[1,2,4]triazolo[4,3ꢀc]pyrimidinꢀ5ꢀamine (20.0 mg, 0.068
mmol) in mixed solvent (dioxane: MeCN: water =3 mL: 0.3 mL: 0.3 mL) was added potassium
carbonate (28.2 mg, 0.204 mmol), phenylboronic acid (8.3 mg, 0.068 mmol) and Pd(PPh3)4 (7.86
mg, 6.80 µmol). The resulting mixture was stirred in microwave under N2 at 110oC for 20 min.
The mixture was then cooled to room temperature and solvent was removed in vacuo. The
residue was purified with silica gel chromatography (eluted with DCMꢀMeOH (10:1)) to give 5
mg (22.3%) of the title compound 28: 1HꢀNMR (400 MHz, Methanolꢀd4) δ 9.30 (s, 1H), 7.94 –
7.88 (m, 3H), 7.52 – 7.44 (m, 3H), 7.39 (t, J = 7.4 Hz, 1H), 6.44 (d, J = 3.0 Hz, 1H), 6.41 – 6.38
(m, 1H), 4.84 (s, 2H). LCꢀMS (m/z): 292.1 [M+H]+ 29: 1HꢀNMR (400 MHz, Methanolꢀd4) δ
8.30 (s, 1H), 8.04 (s, 1H), 7.81 – 7.72 (m, 2H), 7.40 – 7.32 (m, 3H), 7.30 – 7.23 (m, 1H), 6.26 (d,
J = 1.7 Hz, 2H), 4.74 (s, 2H). LCꢀMS (m/z): 292.1 [M+H]+
N-(Furan-2-ylmethyl)-8-(4-(methylsulfonyl)phenyl)-[1,2,4]triazolo[4,3-c]pyrimidin-5-amine
(43): To a solution of 8ꢀbromoꢀNꢀ(furanꢀ2ꢀylmethyl)ꢀ[1,2,4]triazolo[4,3ꢀc]pyrimidinꢀ5ꢀamine
(20.0 mg, 0.068 mmol) in mixed solvent (dioxane: MeCN: water =3 mL: 0.3 mL: 0.3 mL) was
added potassium carbonate (28.2 mg, 0.204 mmol), (4ꢀ(methylsulfonyl)phenyl)boronic acid
(17.7 mg, 0.088 mmol) and Pd(PPh3)4 (7.86 mg, 6.80 µmol). The resulting mixture was stirred in
microwave under N2 at 110oC for 20 min. The mixture was then cooled to rt, and solvent was
removed in vacuo. The residue was purified with silica gel chromatography (eluted with DCMꢀ
ACS Paragon Plus Environment
24

Journal of Medicinal Chemistry
Page 26 of 34
1
2
MeOH (10:1)) to give 10 mg (35.8%) of the title compound. 1HꢀNMR (400 MHz, MeOHꢀd4) δ
ppm 9.32 (s, 1H), 8.27 (d, J = 8.5 Hz, 2H), 8.13 (s, 1H), 8.06 (d, J = 8.5 Hz, 2H), 7.49 (d, J =
1.1 Hz, 1H), 6.47 – 6.38 (m, 2H), 3.36 (s, 2H), 3.18 (s, 3H). 13CꢀNMR (101 MHz, DMSOꢀd6) δ
155.8, 155.1, 153.4, 152.3, 147.3, 145.6, 143.8, 143.5, 140.3, 132.9, 132.2, 128.7, 128.6, 123.6,
120.8, 120.6, 113.4, 113.2, 107.9, 107.8, 105.7, 71.0, 61.4, 39.7, 36.7, 36.6, 28.0, 24.9, 18.
HRMS (ESI): m/z calcd. for C17H15N5O3S [M+ H]+ 370.0896, obsd 370.0975.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Xenograft tumor models
All experiments conducted were performed in female athymic balb/c nude mice (Vital
River) in an AAALAC certificated facility. All procedures and protocols were approved by the
Institutional Animal Care and Use Committee of China Novartis Institute of Biomedical
Research. Karpas 422 human B cell lymphoma cells were obtained from the DSMZ cell bank
(Germany) and cultured in RPMIꢀ1640 medium (Gibco; 11875ꢀ093) supplemented with 15%
FBS (Gibco; 10099ꢀ141) and 1% Pen Strep (Gibco; 15140ꢀ122) at 37ºC in an atmosphere of 5%
CO2 in air. Cells were maintained in suspension cultures at concentrations between 0.5 ꢀ 2×106
cells/mL. To establish Karpas 422 xenografts cells, were reꢀsuspended in PBS and mixed with
Matrigel (BD Bioscience) (1:1, v/v) before injecting 100 µL containing 5×106 cells
subcutaneously into right flanks of Balb/c nude mice.
In vivo pharmacokinetic/pharmacodynamics and efficacy studies
Compound 43 was used in pharmacokinetic/pharmacodynamics relationship and efficacy studies
in animals bearing Karpas 422 xenograft tumors. Tumor volume was determined by measuring
the length (L) and perpendicular width (W), with calipers and calculated using the formula:
0.5×L×W2. Tumor bearing mice were sorted into treatment groups. Treatment was initiated when
the average tumor volume reached 200ꢀ400 mm3. The compound was formulated as a suspension
ACS Paragon Plus Environment
25

Page 27 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
in 0.5% PHMC+0.5% Tween 80 in water and administered orally by gavage at a dose volume of
10 mL/kg. At the end point, the animals was given the first dose administration (i.e., the second
dose was not given). For PK analysis 100 µl of blood samples were collected from each animal
by orbital sinus bleeding. For analysis of compound levels and PD in tissues, tumors were
collected 4 hr post treatment and frozen immediately in liquid nitrogen. When applicable, results
are presented as mean ± SEM. Graphing and statistical analysis was performed using GraphPad
Prism 5.00 (GraphPad Software). Tumor and body weight change data were analyzed
statistically. If the variances in the data were normally distributed (Bartlett's test for equal
variances), the data were analyzed using oneꢀway ANOVA with post hoc Dunnet’s test for
comparison of treatment versus control group. The post hoc Tukey test was used for intragroup
comparison. Otherwise, the KruskalꢀWallis ranked test post hoc Dunn’s was used.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As a measure of efficacy the %T/C value is calculated at the end of the experiment according to:
(ꢁtumor volumetreated/ꢁtumor volumecontrol)*100
Tumor regression was calculated according to:
ꢀ(ꢁtumor volumetreated/tumor volumetreated at start)*100
Where ꢁtumor volumes represent the mean tumor volume on the evaluation day minus the mean
tumor volume at the start of the experiment.
ꢀ ASSOCIATED CONTENT
Supporting information is available free of charge on ACS Publication website: Synthetic
methods for compounds 19 to 23; characterization data for selected compounds; experimental
conditions for crystallization and structure determination; key interactions with EED protein;
experimental protocols of biological assays, including alphascreen binding assay, biochemical
ACS Paragon Plus Environment
26

Journal of Medicinal Chemistry
Page 28 of 34
1
2
3
4
and cellular assay; mouse pharmacokinetic study and MTD determination for compound 43;
5
6
molecular formula strings and compound data (CSV)
7
8
9
ꢀ PDB ACCESSION CODES
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 7: 5H19; compound 8: 5H24; compound 11: 5H25. Authors will release the atomic
coordinates and experimental data upon article publication.
ꢀ AUTHOR INFORMATION
Corresponding Author
Phone +86 21 186 2108 2695. Eꢀmail: yingꢀ2.huang@novarits.com
Present Addresses
§Novartis Institutes for BioMedical Research, 250 Massachusetts Avenue, Cambridge,
Massachusetts 02139, United States
¶Novartis Institutes for BioMedical Research, Fabrikstrasse 2, 4056 Basel, Switzerland
&Ideaya Biosciences, 280 Utah Avenue, South San Francisco, CA 94080, United States
¥ Array BioPharma, 3200 Walnut Street, Boulder, CO 80301, United States
£ Janssen (China) R&D Center, 3F Cheng Kai International Tower, 355 Hong Qiao Road,
Shanghai 200030, China
€Singapore Centre for Environmental Life Sciences Engineering, 60 Nanyang Drive, SBSꢀ01Nꢀ
27, Singapore 637551
Notes:
The authors declare no competing financial interest
ꢀ ACKNOWLEDGMENT
The authors would like to thank Jing Sun, Homan Chan, Chenhui Zeng, Karin Briner, William
Sellers and En Li for their support and insightful discussions.
ACS Paragon Plus Environment
27

Page 29 of 34
Journal of Medicinal Chemistry
1
2
3
4
ꢀ ABBREVIATIONS USED
5
6
7
8
PRC, Polycomb repressive complex; EED, Embryonic ectoderm development; H3K27, histone
H3 lysine 27; SAM, Sꢀadenosylmethionine; HTRF, homogeneous time resolved fluorescence;
SAH, Sꢀadenosylꢀlꢀhomocysteine hydrolase; HTS, highꢀthroughput screen; NMR, nuclear
magnetic resonance; DMSO, dimethyl sulfoxide; HAC, heavy atom count; LE, ligand efficiency;
LipE, lipophilic efficiency; SAR, structureꢀactivity relationship; HPLC, High Performance
Liquid Chromatography.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ REFERENCES
(1) (a) Liu, F.; Wang, L.; Perna, F.; Nimer, S. D. Beyond transcription factors: how oncogenic
signaling reshapes the epigenetic landscape. Nat. Rev. Cancer 2016, 16, 359ꢀ372. (b) Jones, P.
A.; Baylin, S. B. The fundamental role of epigenetic event in cancer. Nat. Rev. Genet. 2002, 3,
419ꢀ428.
(2) (a) Greer. E. L.; Shi, Y. Histone methylation: a dynamic mark in health, disease and
inheritance. Nat. Rev. Genet. 2012, 13, 343ꢀ357. (b) Copeland, R. A.; Solomon, M. E.; Richon,
V. M. Protein methyltransferases as a target class for drug discovery. Nat. Rev. Drug Discovery
2009, 8, 725ꢀ732. (c) Zagni C.; Chiacchio U; Rescifina A. Histone methyltransferase inhibitors:
novel epigenetic agents for cancer treatment. Curr. Med. Chem. 2013, 20, 167ꢀ185.
(3) (a) Margueron, R.; Reinberg, D. The Polycomb complex PRC2 and its mark in life. Nature
2011, 469, 343–349. (b) Xu, C.; Bian, C.; Yang, W.; Galka, M.; Ouyang, H.; Chen, C.; Qiu, W.;
Liu, H.; Jones, A. E.; MacKenzie, F.; Pan, P.; Li, S.ꢀC.; Wang, H.; Min, J. Binding of different
histone marks differentially regulates the activity and specificity of Polycomb Repressive
Complex 2 (PRC2). Proc. Natl. Acad. Sci. 2010, 107, 19266–19271.
ACS Paragon Plus Environment
28