Discovery of betrixaban (PRT054021), N-(5-chloropyridin-2-yl)-2-(4-(N,N-dimethylcarbamimidoyl)benzamido)-5-methoxybenzamide, a highly potent, selective, and orally efficacious factor Xa inhibitor

Article abstract of DOI:10.1016/j.bmcl.2009.02.111

Systematic SAR studies of in vitro factor Xa inhibitory activity around compound 1 were performed by modifying each of the three phenyl rings. A class of highly potent, selective, efficacious and orally bioavailable direct factor Xa inhibitors was discovered. These compounds were screened in hERG binding assays to examine the effects of substitution groups on the hERG channel affinity. From the leading compounds, betrixaban (compound 11, PRT054021) has been selected as the clinical candidate for development.

Full text of DOI:10.1016/j.bmcl.2009.02.111

Bioorganic & Medicinal Chemistry Letters 19 (2009) 2179–2185
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of betrixaban (PRT054021), N-(5-chloropyridin-2-yl)-2-
(4-(N,N-dimethylcarbamimidoyl)benzamido)-5-methoxybenzamide,
a highly potent, selective, and orally efficacious factor Xa inhibitor
||
Penglie Zhang , Wenrong Huang à, Lingyan Wang §, Liang Bao –, Zhaozhong J. Jia ||, , Shawn M. Bauer ,
*
Erick A. Goldman , Gary D. Probst àà, Yonghong Song ||, Ting Su ||, Jingmei Fan, Yanhong Wu, Wenhao Li,
John Woolfrey , Uma Sinha ||, Paul W. Wong §§, Susan T. Edwards ––, Ann E. Arfsten ||||, Lane A. Clizbe
James Kanter ààà, Anjali Pandey ||, Gary Park §§§, Athiwat Hutchaleelaha ||, Joseph L. Lambing ||,
Stanley J. Hollenbach ||, Robert M. Scarborough ||, Bing-Yan Zhu –
,
Millennium Pharmaceuticals, Inc., 256 East Grand Avenue, South San Francisco, CA 94080, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Systematic SAR studies of in vitro factor Xa inhibitory activity around compound 1 were performed by
modifying each of the three phenyl rings. A class of highly potent, selective, efficacious and orally bio-
available direct factor Xa inhibitors was discovered. These compounds were screened in hERG binding
assays to examine the effects of substitution groups on the hERG channel affinity. From the leading com-
pounds, betrixaban (compound 11, PRT054021) has been selected as the clinical candidate for
development.
Received 13 January 2009
Revised 25 February 2009
Accepted 26 February 2009
Available online 3 March 2009
Dedicated to the memory of Dr. Robert M.
Scarborough (1953–2006), co-founder and
former Senior VP of Chemistry at Portola
Pharmaceuticals, Inc.
Ó 2009 Elsevier Ltd. All rights reserved.
Serine protease factor Xa (fXa), positioned at the juncture of the
intrinsic and the extrinsic pathways, plays a pivotal role in the
blood coagulation cascade. In the prothrombinase complex consist-
ing of fXa, factor Va and Ca2+ assembled on the platelet surface, fXa
catalyzes the conversion of prothrombin to thrombin, the final en-
zyme responsible for fibrin clot formation in this cascade.1 To safely
interrupt blood coagulation without disrupting primary hemosta-
sis, inhibition of fXa should be more effective than direct inhibition
of thrombin, which has made fXa a particularly attractive target for
the treatment of severe cardiovascular diseases.2 Selective inhibi-
tion of fXa without affecting the existing thrombin levels may cause
less impairment of primary hemostasis and thus should be a safer
anticoagulant therapy than direct inhibition of thrombin. Clinical
findings have confirmed the potential of fXa inhibition for produc-
ing excellent antithrombotic efficacy with minimal bleeding risk
when compared to direct thrombin inhibitors.3,4
In the preceding communication, we reported the discovery of
anthranilamide-based compound 1, N-(5-chloropyridin-2-yl)-2-(4-
(N,N-dimethylcarbamimidoyl)-benzamido)benzamide, as a potent
fXa inhibitor (IC50 3 nM; Ki 1.4 nM).5,6 This compound has strong
anticoagulant activity in our human plasma based thrombin genera-
tion assay (2 Â TG 0.54
lM) and good oral bioavailability in rat
* Corresponding author at present address: Portola Pharmaceuticals, Inc., 270
East Grand Ave, Ste 22, South San Francisco, CA 94080, USA. Tel.: +1 650 246 7073;
fax: +1 650 246 7776.
(F 31%). The 4-(N,N-dimethyl-carbamimidoyl)benzamidine motif,
binding in the fXa S4 pocket, delivers the most potent fXa inhibitory
activity among all the N-substituted benzamidines investigated. To
further improve its in vitro fXa potency and anticoagulant activity,
we decided to systematically examine aromatic rings A, B and C of
compound 1 to expand the SAR scope of this class of fXa inhibitors.
E-mail address: zjia@portola.com (Z.J. Jia).
Present address: ChemoCentryx, Inc., Mountain View, CA 94043, USA.
Present address: Foley & Lardner, Inc., Palo Alto, CA 94304, USA.
Present address: Schering-Plough Research Institute, Kenilworth, NJ 07033, USA.
Present address: Genentech, Inc., South San Francisco, CA 94080, USA.
à
§
–
||
Present address: Portola Pharmaceuticals, Inc., South San Francisco, CA 94080, USA.
Present address: Exelixis, Inc., South San Francisco, CA 94080, USA.
Present address: Elan Pharmaceuticals, Inc., South San Francisco, CA 94080, USA.
Present address: Berkeley HeartLab, Inc., Alameda, CA 94501, USA.
Present address: Medtronic Vascular, Inc., Santa Rosa, CA 95403, USA.
Present address: Genelabs Technologies, Inc., Redwood City, CA 94063, USA.
Present address: Cell Biosciences, Inc., Palo Alto, CA 94304, USA.
Present address: Nanosyn, Inc., Menlo Park, CA 94025, USA.
Present address: Rigel Pharmaceuticals, Inc., South San Francisco, CA 94080, USA.
C
àà
B
§§
HN
N
O
HN
O
––
A
||||
N
ààà
HN
Cl
§§§
1
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.02.111

2180
P. Zhang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2179–2185
Table 1
B ring substitution SAR
HN
N
O B3
B4
N
HN
N
O
HN
B5
HN
S
O
O
N
1-14
15
HN
Cl
HN
Cl
Compound
B3
B4
B5
fXa IC50 (nM)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
H
Cl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
4
21
2
0.5
0.7
1.3
1.0
<0.5
1.2
1.4
1.5
<0.5
1.5
1.3
2
Me
CF3
Et
C„CH
CN
F
OMe
SMe
SO2Me
SEt
Figure 2. Computer modeling of compound 4 docked in fXa active enzyme pocket.
binding pocket, with the N@C bond resting parallel to the phenyl
ring planes and one methyl group inserting into the narrow hydro-
phobic cleft between the phenyl rings of Phe174 and Trp215. A p-
cation interaction presumably exists between the protonated ami-
dino group and the phenyl rings. Furthermore, as shown in Figure
2, B3 and B4 are in the open space out of the active binding pocket,
thus substitutions at B3 and B4 are expected to be tolerated, but to
have little impact on potency. The B5-chloro group points to a
small hydrophobic pocket, bordered partially by a disulfide bridge
between Cys191 and Cys220; thus a small hydrophobic B5 group in
this pocket could boost the inhibitor’s fXa binding affinity. As con-
firmation of this hypothesis, compounds 5–14, each bearing a
small B5 group capable of hydrophobic interactions, display strong
anti-fXa activity with IC50 values all in the range of <0.5–1.5 nM.
Through analog 15 (IC50 2 nM), we learned that replacement of
phenyl B ring with thiophene maintains excellent fXa potency. This
docking study also revealed that substitution on the C ring should
be tolerated, especially at the position (C2) next to the carbonyl.
It is worthy noting that highly potent acetylenyl analog 8 (IC50
We started our investigation of compound 1 by modifying the cen-
tral phenyl B ring with a chloro substitution group as B3, B4 or B5 (2–
4, Table 1). Among these three analogs, compound 4 with B5-chloro
is the most potent (IC50 0.5 nM). Its high potency is retained in the
human plasma based thrombin generation assay (2 Â TG 0.27
lM).
In rat, compound 4 has a clearance (CL) of 29.7 mL/min/kg, half-life
(t1/2) of 12.8 h, and an oral bioavailability (F) of 26.1% when dosed at
0.2 mg/kg IV and 6 mg/kg PO. Replacement of the B5 chloro with
methyl and trifluoromethyl yielded analogs 5 and 6 with IC50 values
of 0.7 nM and 1.3 nM, respectively, indicating minimal electronic
effects.
7
Computer-based docking studies using GOLD as well as both
<0.5 nM; 2 Â TG 0.21
lM) and methylthio analog 12 (IC50 <0.5 nM;
2 Â TG 0.24 M) were found to be stable in rat and human in vitro
published and in-house fXa crystal data were used to analyze the
binding mode of compound 4 at the active site of fXa. As shown
in Figure 1, the chloropyridine A ring extents into the fXa S1 site
with the chloro atom positioned above the phenyl ring of Tyr228
and occupies a hydrophobic pocket formed by Ala190, Val213
and Tyr228. This result is consistent with the disclosed X-ray struc-
ture of fXa co-crystallized with an anthranilamide compound sim-
ilar in structure to compound 4.8 The N,N-dimethylamidine moiety
is flanked by the phenyl groups of Tyr99 and Phe174 of the S4 aryl
l
liver metabolic experiments (S9 incubation t1/2 >2 h).9 In rat PK
studies, compound 8 (CL 74.8 mL/min/kg; t1/2 8.4 h; F 27.9%) dis-
played higher clearance and shorter half-life than compound 4,
and compound 12 (CL 77.1 mL/min/kg; t1/2 16.5 h; F 26.7%) showed
higher clearance. Thus we have concluded that chloro at B5 is the
optimal substitution on B ring based upon the desired PK parame-
ters and superior fXa binding activity of compound 4.
Once chloro had been established as the optimal B5 substitu-
tion, we synthesized compounds 16–23 (Table 2) to scan various
B3 groups for different physicochemical properties and improved
Table 2
B ring substitution SAR
HN
N
O B3
HN
Cl
O
N
4,16-23
HN
Cl
Compound
B3
fXa IC50 (nM)
4
16
17
18
19
20
21
22
23
H
OMe
OEt
OiPr
NMe2
NMeEt
NMe(CH2CH2OMe)
1-piperidinyl
0.5
0.6
0.6
0.5
0.5
0.5
1.2
0.9
1.3
4-morpholinyl
Figure 1. Computer modeling of compound 4 docked in fXa active enzyme pocket.

P. Zhang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2179–2185
2181
Table 4
C ring substitution SAR
PK profiles. Consistent with the modeling results discussed above,
a flat anti-fXa SAR was observed regardless of bulkiness of the B3
group, with IC50 values in the range of 0.5–1.3 nM. However, no
improvement was achieved in regard to PK parameters in rat. For
example, the bioavailability of compounds 16 and 21 was 11.2%
and 8.1%, respectively in rat, compared to that of 26.1% of com-
pound 4.
NH
C3
C2
HN
N
O
HN
O
HN
N
N
O
HN
O
O
HN
O
N
Cl
S
Cl
Cl
N
N
N
HN
Cl
45
HN
Cl
46
HN
Cl
4, 31-44
With B5-chloro locked, we focused on the A ring SAR to investi-
gate the S1 binding motif and prepared compounds 24–30 (Table
3). Replacement of chloro by fluoro (24) and methyl (25) slightly
compromises fXa potency, probably due to weaker interaction
with Tyr228. Replacement of chloro by acetylene (26, IC50
Compound
C3
C2
fXa IC50 (nM)
4
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
H
F
OMe
H
H
H
H
H
H
H
H
H
H
F
0.5
1.1
7
<0.5
0.5
1.2
<0.5
3
0.8 nM; 2 Â TG 0.38
lM) retains excellent anti-fXa activity. Though
OMe
OCH2CH2OMe
OCH2CH2NMe2
NMe2
it was stable in rat and human in vitro liver metabolic experiments
(S9 incubation t1/2 >2 h), compound 26 (CL >100 mL/min/kg; t1/2
6.3 h) displayed higher clearance and shorter half-life than com-
pound 4 in rat PK studies. With the A ring nitrogen moved to the
neighboring location, 2-chloro-5-pyridinyl 27 (IC50 3 nM) decreases
in fXa potency compared to 5-chloro-2-pyridinyl 4. With 4-chloro-
aniline (28, IC50 0.7 nM) and 4-acetylenylaniline (29, IC50 0.9 nM)
as the A ring motifs, in vitro fXa activities comparable to 2-amino-
pyridines 4 and 26 have been achieved. Due to the concern of
aniline’s potential toxicity, 2-amino-5-chloropyridine was chosen
as the preferred S1 binding motif over the corresponding aniline.
Replacement of chloropyridine with chlorothiophene (30) reduces
fXa potency significantly.
NMe(CH2CH2OMe)
1-azetidinyl
1-pyrrolidinyl
1-piperidinyl
4-HO2C-1-piperidinyl
4-morpholinyl
1-homopiperidinyl
0.5
0.4
2
<0.5
0.7
0.7
1.0
3
H
H
H
H
H
1.0
compounds 35–44. Their high fXa potency demonstrates that the
C2 position tolerates a wide range of functional groups, as pre-
dicted by the computer-based modeling studies (Figs. 1 and 2).
Both pyridine C ring (45, IC50 3 nM) and thiophene C ring (46,
IC50 1.0 nM) slightly reduce fXa potency from lead compound 4.
Among all the analogs in Table 4, the C2-fluoro compound 33
With compound 4 as the starting point, we moved on to the C
ring modifications (Table 4). Compounds 31 and 32 with C3 substi-
tutions are less potent than the corresponding C2 substituted ana-
logs 33 and 34. Further modifications at C2 led to the synthesis of
(IC50 <0.5 nM; 2 Â TG 0.38
lM) is the most attractive. It has fXa po-
Table 3
A ring SAR
tency very similar to that of lead compound 4, and displays consid-
erably improved PK profiles in rat with lower clearance (CL
15.3 mL/min/kg) and higher oral bioavailability (F 38.0%). The
other C2-substitution groups have resulted in compromises in rat
PK parameters. For example, the bioavailability of compounds 40
and 41 was 6.9% and 21.3%, respectively.
HN
N
O
HN
O
Cl
4, 24-30
A
After the A, B and C ring SAR had been examined, we went back
to modification of the P4 benzamidine N-substitution. Following
the previous SAR exploration in this region with compound 1,5a
we prepared analogs 47–54 (Table 5). All these compounds are
highly potent fXa inhibitors, and the SAR overall mirrors that
reported in the preceding communication. Judged by the in vitro
anti-Xa activity, the N,N-dimethylbenzamidine as in compound 4
is clearly still the best P4 moiety.
Compound
A
fXa IC50 (nM)
0.5
4
N
HN
Cl
24
25
26
27
N
3
HN
F
N
1.1
0.8
3
HN
Me
Table 5
P4 benzamidine N-substitution SAR
N
O
HN
O
HN
R1
O
HN
O
N
N
HN
N
Cl
Cl
R2
R2
N
N
N
HN
HN
HN
Cl
Cl
4, 31-36
HN
Cl
HN
Cl
37-38
28
29
30
0.7
0.9
6
Compound
R1
R2
fXa IC50 (nM)
4
47
48
49
50
51
52
53
54
Me
Et
1-azetidinyl
1-pyrrolidinyl
1-piperidinyl
4-HO2C-1-piperidinyl
4-morpholinyl
Me
Me
0.5
0.9
1.2
2
1.3
2
2
1.5
6
Cl
S
HN
Me
Et

2182
P. Zhang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2179–2185
B
Cl
B
Cl
Cl
B
b or c
a
d
H2N
O2N
O
b
a
O2N
HO2C
O2N
O
Cl
O2N
Cl
Cl
A
A
O
(N)
(N)
N
HN
57
HN
56
HO2C
HO2C
HN
62
Cl
55
C
C
c
d
HN
N
O
O
e
B
B
NC
HN
O
HN
O
1R2RN
O2N
O
RO
O2N
A
A
(N)
(N)
59
58
HN
HN
Cl
Cl
f or g
O
N
N
h or i
C
HN
Cl
HN
R1
O
HN
64
Cl
B
63
C
N
HN
O
R2
O
N
N
B
A
Scheme 2. Reagents and conditions: (a) concd HNO3 (1.2 equiv), concd H2SO4, 0 °C,
70–80%; (b) 2-amino-5-chloropyridine (1 equiv), POCl3 (1 equiv), pyridine, 0 °C, 80–
90%; (c) ROH (1.2 equiv), NaH (1.2 equiv), NMP, 110 °C, 50–70%; (d) NHR1R2
(2 equiv), Cs2CO3 (2 equiv), DMSO, 110 °C, 60–80%.
(N)
HN
O
60
HN
R2
A
(N)
61
HN
Scheme 1. Reagents and conditions: (a) substituted aminopyridine or aniline
(1 equiv), POCl3 (1 equiv), pyridine, 0 °C, 60–95%; (b) SnCl2Á2H2O (4 equiv), EtOAc,
reflux, 50–70%; (c) H2 (balloon), 5% Pt(S)/C or 10% Pd/C, EtOAc, 50–90%; (d)
substituted 4-cyanobenzoyl chloride (1 equiv), THF, rt, 80–95%; (e) LiNMe2
(4 equiv), THF, 0 °C to rt, 70–90%; (f) (1) H2S (g), pyridine, Et3N, rt; (2) MeI
(2 equiv), acetone, reflux; (3) NHR1R2 (2 equiv), HOAc (3 equiv), MeOH, reflux, 50–
90%; (g) (1) HCl (g), dry MeOH, 0 °C; (2) NHR1R2 (2 equiv), MeOH, reflux, 50–80%;
(h) (1) H2S (g), pyridine, Et3N, rt; (2) MeI (2 equiv), acetone, reflux; (3)
H2NCH2CH2NHR2 (2 equiv), HOAc (5 equiv), MeOH, reflux, 50–80%; (i) (1) HCl (g),
dry MeOH, 0 °C; (2) H2NCH2CH2NHR2 (2 equiv), MeOH, reflux, 50–70%.
c
a,b
O2N
HO2C
Cl
O2N
O
Cl
O2N
Cl
Cl
H2NOC
S
HN
66
65
Scheme 3. Reagents and conditions: (a) (COCl)2 (3 equiv), DMF (cat), DCM, rt; (b)
conc. NH4OH (10 equiv), DCM, 0 °C, 90% for 2 steps; (c) 2-chloro-5-iodothiophene
(1 equiv), Cs2CO3 (2 equiv), CuI (0.1 equiv), trans-1,2-diaminocyclohexane
(0.1 equiv), dioxane, reflux, 50–60%.
The general synthesis of the compounds shown in Tables 1–5 is
described in Scheme 1. Substituted 2-nitrobenzoic acid 55 was
coupled with substituted aminopyridine or aniline in pyridine
using phosphoryl chloride, or by other standard synthetic method-
ologies. Reduction of the nitro group of compound 56 was accom-
plished using tin(II) chloride or hydrogenation with sulfided
platinum or palladium on carbon, depending on the compatibility
of the substitution groups. The resulting aniline 57 was then trea-
ted with substituted 4-cyanobenzoyl chloride, commercially avail-
able or freshly produced from the corresponding benzoic acid
using oxalyl chloride, to prepare benzonitrile 58. For target 59 with
inert A, B and C substitution groups, it was synthesized by reacting
58 with excessive lithium dimethylamide (commercial 5 wt % sus-
pension in hexane). More generally, target 60 was synthesized via
methyl thioimidate or methyl imidate as described in our previous
communications.5a,10 Imidazoline target 61 was prepared from
benzonitrile 58 via methyl thioimidate or methyl imidate as well.
The compounds shown in Table 2 were prepared according to
Scheme 2. 3,5-Dichlorobenzoic acid was nitrated to produce 3,5-di-
chloro-2-nitrobenzoic acid, which was next coupled with amino-
pyridine to obtain 62. Reaction of 62 with alkoxide, newly
produced from the corresponding alcohol and sodium hydride,
afforded compound 63. Treatment of 62 with excessive amount
of amine yielded compound 64. Final targets 16–23 were then fin-
ished from these two intermediates by the same chemistry illus-
trated in Scheme 1.
The key steps toward preparation of A ring SAR target 30 (Table
3) are shown in Scheme 3. Chlorothiophene derivative 66 was
synthesized by copper-catalyzed coupling reaction of 2-chloro-5-
iodothiophene and benzamide 65, which had been produced from
3-chloro-6-nitrobenzoic acid. Target compound 30 was completed
from 66 via the same synthetic route revealed in Scheme 1.
The key steps toward C ring SAR targets (Table 4) are described
in Scheme 4. The C2 side-chains were installed by treatment of flu-
oro precursor 67 with cesium carbonate and the corresponding
alcohols or amines. 4-Cyano-2-thiophenecarboxylic acid was made
through cyanation and oxidation of commercial 4-bromo-2-thio-
phenecarboxaldehyde. It was coupled with aniline 70 to produce
71. All the final targets were then prepared from nitriles 68, 69
and 71 by the general route shown in Scheme 1.
R
O
F
O
HN
O
O
HN
O
NC
NC
a
Cl
Cl
N
N
HN
Cl
HN
Cl
67
68
R1
b
N R2
O
NC
HN
O
Cl
N
HN
Cl
69
NC
Cl
Br
NC
d
c
CO2H
CHO
CHO
S
S
S
H2N
e
O
N
HN
Cl
70
NC
O
S
HN
O
Cl
N
HN
Cl
71
Scheme 4. Reagents and conditions: (a) ROH (10 equiv), Cs2CO3 (5 equiv), NMP,
110 °C, 60–70%; (b) NHR1R2 (5 equiv), Cs2CO3 (3 equiv), NMP, 110 °C, 70–85%; (c)
CuCN (3 equiv), CuI (0.3 equiv), DMF, reflux, 70%; (d) NaClO2 (9 equiv), NaH2PO4
(6 equiv), 2-methyl-2-butene, tBuOH, acetone, water, 0 °C, 90%; (e) POCl3 (1 equiv),
pyridine, 0 °C, 80%.

P. Zhang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2179–2185
2183
Table 6
0.1 mg/kg IV and 1 mg/kg PO, they had bioavailability of 73.0%
and 86.1%, respectively in monkey. The outstanding fXa potency,
excellent in vitro anticoagulant activity and good PK characteristics
of compounds 4 and 33 led to the evaluation of their safety profiles.
Unfortunately, both compounds were found to exhibit potent affin-
ity toward the hERG (human ether-a-go-go related gene) ion chan-
C ring and B ring substitution SAR for compound 11
C2
HN
N
O
HN
O
O
R
N
11, 72-77
nel, with Ki values of 0.11 lM and 0.42 lM, respectively in a
HN
Cl
radioligand (dofetilide or astemizole) hERG transfected membrane
binding assay.11,12 In recent years, the interaction of pharmaceuti-
cal agents with the hERG-encoded cardiac potassium channel has
represented a significant safety concern and regulatory hurdle for
new molecular entities in development.13,14 Inhibition of the hERG
channel can cause delayed ventricular cell repolarization, seen as
QT interval prolongation on the electrocardiogram and associated
with a potentially fatal cardiac arrhythmia. Thus, the potential risk
of this class of excellent fXa inhibitors to cause undesired cardiac
repolarization forced us to reexamine our SAR exploration
directions.
Compound
R
C2
fXa IC50
(nM)
2 Â TG
M)
hERG Ki
(lM)
(l
11
72
73
74
75
76
77
Me
CF3
CH2CF3
CH2CO2H
Me
H
H
H
H
F
1.5
17
10
6
0.7
2
0.33
nd
nd
1.5
0.34
1.8
nd
nd
>10
2.1
0.20
>10
Me
Me
1-piperidinyl
4-HO2C-1-
piperidinyl
0.9
0.32
An extensive screening of the compounds in Tables 1–5 to
understand the hERG binding SAR revealed that modifications to
the A, B, C rings and the amidine P4 moieties overall have little ef-
fect on the hERG channel affinity with the exception of two carbox-
ylic acid-containing compounds 42 and 51 (both have hERG Ki
Compounds 4 (fXa Ki 44 pM) and 33 (fXa Ki 60 pM) have dis-
played good PK profiles in beagle dog and cynomolgus monkey.
Dosed at 0.1 mg/kg IV and 1 mg/kg PO, compounds 4 and 33 had
bioavailability of 64.9% and 68.8% respectively in dog; dosed at
values >10 lM). Their reduced hERG affinity is accredited to the
Table 7
Biological and animal pharmacokinetic data for the leading fXa inhibitors
Compound
4
33
11 (betrixaban)
75
HN
O
HN
O
F
HN
N
O
HN
O
F
HN
N
O
HN
N
O
HN
O
N
Cl
O
HN
O
Cl
O
N
N
HN
Cl
HN
Cl
N
N
HN
Cl
HN
Cl
fXa IC50 (nM)
fXa Ki (nM)
0.5
<0.5
0.060
0.29
1.1
87.6
0.42
nd
1.5
0.7
0.044
0.27
0.83
88.1
0.11
0.33
0.117
0.33
1.7
80.0
1.8
0.105
0.34
1.5
63.4
2.1
2 Â TG (
rabbit 2 Â PT (
human plasma protein binding (%)
hERG Ki ( M)
patch clamp hERG IC50
lM)
lM)
l
(lM)
8.9
4.2
thrombin IC50 M)
(
l
>10
>10
>10
>10
>10
0.51
>10
>10
>10
>10
>10
0.54
>10
>10
>10
>10
>10
6.3
>10
>10
>10
>10
>10
7.2
trypsin IC50 M)
(l
t-PA IC50
(lM)
aPC IC50 (l
M)
plasmin IC50
(
lM)
kallikrein IC50
(
lM)
Rat PK
F (%)
t1/2 (h)
CL (mL/min/kg)
Vd (L/kg)
AUCPO (ng h/mL)
26.1
12.8
29.7
26.1
885
38.0
5.2
15.3
6.9
2425
0.2; 6
23.8
8.8
43.6
32.9
456
1; 5
48.6
5.5
57.5
26.5
770
1; 5
dose IV; PO (mg/kg)
0.2; 6
Dog PK
F (%)
t1/2 (h)
CL (mL/min/kg)
Vd (L/kg)
AUCPO (ng h/mL)
dose IV; PO (mg/kg)
64.9
15.3
34.0
45.7
320
68.8
13.1
30.5
35.1
381
51.6
21.2
26.5
48.8
825
0.5; 2.5
88.9
7.9
24.6
16.8
1500
0.5; 2.5
0.1; 1
0.1; 1
Monkey PK
F (%)
t1/2 (hr)
73.0
19.4
26.7
36.5
512
86.1
22.8
11.2
20.5
1301
0.1; 1
58.7
9.6
18.7
13.4
4180
0.75; 7.5
77.6
11.4
16.2
16.3
6170
0.75; 7.5
CL (mL/min/kg)
Vd (L/kg)
*
AUCPO (ng hr/mL)
dose IV; PO (mg/kg)
0.1; 1

2184
P. Zhang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2179–2185
carboxylic acid residue, which couples with the P4 basic benzam-
idine moiety to cause compounds 42 and 51 to stay in a zwitter-
ionic form.11,14 However, this zwitterionic property is strongly
detrimental to oral absorption. The oral bioavailability of com-
pound 42, for example, in rat is <5%. Fortunately, among the
remaining compounds shown in Tables 1–5, compound 11 (hERG
dosing, but there is no difference in PK when administered intrave-
nously. Compound 75 is relatively more potent in the patch clamp
hERG assay and commands much higher manufacturing cost than
compound 11. Based upon the overall biological, toxicological and
PK/PD profiles, hERG liability concern and manufacturing cost,
compound 11 (betrixaban) was chosen as the clinical candidate
for this class of anthranilamide-based fXa inhibitors. In addition,
a safety window of greater than 30-fold was maintained for the ra-
tio of patch clamp hERG IC50 to the expected compound Cmax ad-
justed for its unbound fraction.13c The Phase II study of
betrixaban (PRT054021) as an oral fXa inhibitor for prevention of
venous thromboembolic (VTE) events after total knee replacement
has been successfully completed in 2008 and has validated its po-
tential to be a safe and effective oral anticoagulant.19 The ongoing
Phase II EXPLORE Xa study will assess the safety, tolerability and
efficacy of betrixaban compared with dose-adjusted warfarin in
patients with non-valvular atrial fibrillation (AF).
Ki 1.8
lM) exhibits significantly lower hERG activity than all the
others (hERG Ki 6 0.5
lM).
To follow up on compound 11 (fXa Ki 117 pM), compounds 72–
77 (Table 6) were prepared to fine-tune fXa potency and hERG
channel affinity. Replacement of methoxy with trifluoromethoxy
(72) or 2,2,2-trifluoroethoxy (73) reduces anti-fXa activity consid-
erably. While the C2-fluoro analog 75 (IC50 0.7 nM; fXa Ki 105 pM)
displays similarly mild hERG activity (hERG Ki 2.1 lM) as com-
pound 11, the corresponding 1-piperidinyl analog 76 has strong
hERG binding potency. Zwitterionic compounds 74 and 77 exhibit
greatly decreased hERG affinity, but their oral bioavailability in rat
is very low (<5%).
From these systematic SAR explorations, compounds 11 and 75
have been selected for further evaluations (Table 7). They both
show excellent in vitro anticoagulant potency, judged by their
References and notes
1. (a) Mann, K. G.; Jenny, R. J.; Krishnaswamy, S. Annu. Rev. Biochem. 1988, 57, 915;
(b) Mann, K. G.; Butenas, S.; Brummel, K. Arterioscler. Thromb. Vas. Biol. 2003, 23,
17.
2 Â TG values of 0.33
lM and 0.34 lM. Compounds 4, 33, 11 and
75 all are dose-dependently efficacious in our rabbit deep vein
2. (a) Schaffer, L. W.; Davidson, J. T.; Vlasuk, G. P.; Dunwiddie, C. T.; Siegl, P. K. S.
Arterioscler. Thromb. 1992, 12, 879; (b) Zhu, B.-Y.; Scarborough, R. M. Annu. Rep.
Med. Chem. 2000, 35, 83; (c) Linkins, L.-A.; Julian, J. A.; Rischke, J.; Hirsh, J.;
Weitz, J. I. Thromb. Res. 2002, 105, 241; (d) Walenga, J. M.; Jeske, W. P.;
Hoppensteadt, D.; Fareed, J. Curr. Opin. Invest. Drugs 2003, 4, 272; (e) Samama,
M. M. Thromb. Res. 2002, 106, V267; (f) Kaiser, B. Cell. Mol. Life Sci. 2002, 59, 189.
3. (a) McBride, B. F. J. Clin. Pharm. 2005, 45, 1004; (b) Saiah, E.; Soares, C. S. Curr.
Top. Med. Chem. 2005, 5, 1677.
4. (a) Roehrig, S.; Straub, A.; Pohlmann, J.; Lampe, T.; Pererstorfer, J.; Schlemmer,
K.-H.; Reinemer, P.; Perzborn, E. J. Med. Chem 2005, 48, 5900; (b) Perzborn, E.;
Strassburger, J.; Wilmen, A.; Pohlmann, J.; Roehrig, S.; Schlemmer, K.-H.;
Straub, A. J. Thromb. Haemostasis 2005, 3, 514; (c) Eriksson, B. L.; Borris, L.; Dahl,
O. E.; Haas, S.; Huisman, M. V.; Kakkar, A. K. J. Thromb. Haemostasis 2006, 4, 121;
(d) Quan, M. L.; Lam, P. Y. S.; Han, Q.; Pinto, D. J. P.; He, M. Y.; Li, R.; Ellis, C. D.;
Clark, C. G.; Teleha, C. A.; Sun, J.-H.; Alexander, R. S.; Bai, S.; Luettegn, J. M.;
Knabb, R. M.; Wong, P. C.; Wexler, R. R. J. Med. Chem 2005, 48, 1729; (e) Wong,
P. C.; Crain, E. J.; Watson, C. A.; Wexler, R. R.; Lam, P. Y. S.; Quan, M. L.; Knabb, R.
M. J. Thromb. Thrombolysis 2007, 24, 43; (f) Pinto, D. J. P.; Orwat, M. J.; Koch, S.;
Rossi, K. A.; Alexander, R. S.; Smallwood, A.; Wong, P. C.; Rendina, A. R.;
Luettgen, J. M.; Knabb, R. M.; He, K.; Xin, B.; Wexler, R. R.; Lam, P. Y. S. J. Med.
Chem 2007, 50, 5339; (g) Furugohri, T.; Isobe, K.; Honda, Y.; Kamisato-
Matsumoto, C.; Sugiyama, N.; Nagahara, T.; Morishima, Y.; Shibano, T. J.
Thromb. Haemostasis 2008, 6, 1542; (h) Turpie, A. G.; Bauer, K. A.; Erilsson, B. I.;
Lassen, M. R. Arch. Intern. Med 2002, 162, 1833.
5. (a) Zhang, P.; Bao, L.; Zuckett, J. F.; Jia, Z. J.; Sinha, U.; Park, G.; Hutchaleelaha,
A.; Scarborough, R. M.; Zhu, B.-Y. Bioorg. Med. Chem. Lett. 2009, 19, 2186, the
preceding communication, doi:10.1016/j.bmcl.2009.02.114.; The fXa IC50
values were determined by the method described in: (b) Sinha, U.; Ku, P.;
Malinowski, J.; Zhu, B. Y.; Scarborough, R. M.; Marlowe, C. K.; Wong, P. W.; Lin,
P. H.; Hollenbach, S. J. Eur. J. Pharmacol. 2000, 395, 51; The fXa Ki values were
determined by the method described in: (c) Betz, A.; Wong, P. W.; Sinha, U.
Biochemistry 1999, 38, 14582; For description of our human plasma-based
thrombin generation assay (2ÂTG), see: (d) Sinha, U.; Lin, P. H.; Edwards, S. T.;
Wong, P. W.; Zhu, B.; Scarborough, R. M.; Su, T.; Jia, Z. J.; Song, Y.; Zhang, P.;
Clizbe, L.; Park, G.; Reed, A.; Hollenbach, S. J.; Malinowski, J.; Arfsten, A. E.
Arterioscler. Thromb. Vas. Biol. 2003, 23, 1098.
thrombosis model.15 The concentration required to double the rab-
bit prothrombin time (2 Â PT)16 is below 2
lM for each fXa inhib-
itor. Owing to the N,N-dimethylbenzamidine functionality which is
charged in the biological pH range, these four compounds have
good solubility in water and exhibit low human plasma protein
binding (63–88%).17 Detailed in vitro biological study and in vivo
anticoagulant efficacy study of these compounds will be disclosed
in a separate biology-focused publication. To benchmark our lead
compounds in anticoagulant activity, we prepared and studied riv-
aroxaban (BAY59-7939, reported fXa Ki 0.4 nM)4a (78) and apix-
aban (BMS-562247, reported fXa Ki 0.08 nM)4f (79), two oral fXa
inhibitors in advanced clinical trials. We found the in-house
2 Â TG values for them to be 0.41
l
M and 1.8
lM, respectively.
O
N
H2N
O
O
O
N
N
O
O
O
N
N
N
H
Cl
S
HN
78
79
O
O
In patch clamp hERG assays,18 compounds 11 and 75 have IC50
values of 8.9 M and 4.2 M respectively, while that is 0.33 M for
l
l
l
compound 4. All four compounds are uniformly selective for fXa
with very poor activity toward thrombin, trypsin, t-PA, aPC or plas-
min inhibition (IC50 >10
lM) and only weak activity toward plasma
6. Several other laboratories have also investigated anthranilamide-derived fXa
inhibitors: (a) Yee, Y. K.; Tebbe, A. L.; Linebarger, J. H.; Beight, D. W.; Craft, T. J.;
Gifford-Moore, D.; Goodson, T., Jr.; Herron, D. K.; Klimkowski, V. J.; Kyle, J. A.;
Sawyer, J. S.; Smith, G. F.; Tinsley, J. M.; Towner, R. D.; Weir, L.; Wiley, M. R. J.
Med. Chem. 2000, 43, 873; (b) Shrader, W. D.; Young, W. B.; Sprengler, P. A.;
Sangalang, J. C.; Elrod, K.; Carr, G. Bioorg. Med. Chem. Lett. 2001, 11, 1801; (c)
Chou, Y.-L.; Davey, D. D.; Eagen, K. A.; Griedel, B. D.; Karanjawala, R.; Philips, G.
B.; Sacchi, K. L.; Shaw, K. J.; Wu, S. C.; Lentz, D.; Liang, A. M.; Trin, L.; Morrissey,
M. M.; Kochanny, M. J. Bioorg. Med. Chem. Lett. 2003, 13, 507; (d) Kochanny, M.
J.; Alder, M.; Ewing, J.; Griedel, B. D.; Ho, E.; Karanjawala, R.; Lee, W.; Lentz, D.;
Liang, A. M.; Morrissey, M. M.; Phillips, G. B.; Post, J.; Sacchi, K. L.; Sakata, S. T.;
Subramanyam, B.; Vergona, R.; Walters, J.; White, K. A.; Whitlow, M.; Ye, B.;
Zhao, Z.; Shaw, K. J. Bioorg. Med. Chem. 2007, 15, 2127; (e) Ye, B.; Arniaz, D. O.;
Chuo, Y.-L.; Griedel, B. D.; Karanjawala, R.; Lee, W.; Morrissey, M. M.; Sacchi, K.
L.; Sakata, S. T.; Shaw, K. J.; Wu, S. C.; Zhao, Z.; Adler, M.; Cheeseman, S.; Dole,
W. P.; Ewing, J.; Fitch, R.; Lentz, D.; Liang, A.; Light, D.; Morser, J.; Post, J.;
Rumennik, G.; Subramanyam, B.; Sullivan, M. E.; Vergona, R.; Walters, J.; Wang,
Y.-X.; White, K. A.; Whitlow, M.; Kochanny, M. J. J. Med. Chem. 2007, 50, 2967;
(f) Mendel, D.; Marquart, A. L.; Joseph, S.; Waid, P.; Yee, Y. K.; Tebbe, A. L.; Ratz,
A. M.; Herron, D. K.; Goodson, T.; Masters, J. J.; Franciskovich, J. B.; Tinsley, J. M.;
Wiley, M. R.; Weir, L. C.; Kyle, J. A.; Klimkowski, V. J.; Smith, G. F.; Towner, R. D.;
kallikrein inhibition. The plasma kallikrein IC50 and Ki values for
compound 11 are 6.3
lM and 3.5
l
M respectively, and those for
compound 75 are 7.2
lM and 3.5
lM, respectively. These com-
pounds are stable in rat, dog, monkey and human liver S9 incuba-
tion, and have displayed a profile of good oral bioavailability and
oral exposure, long half-life, moderate to high clearance, and high
volume of distribution (Vd). Dosed at 0.5 mg/kg IV and 2.5 mg/kg
PO, compounds 11 and 75 had bioavailability of 51.6% and 88.9%
respectively in dog; dosed at 0.75 mg/kg IV and 7.5 mg/kg PO, they
had bioavailability of 58.7% and 77.6% respectively in monkey. The
C2-fluoro group improves oral absorption and lowers CL and Vd
parameters in rat, dog and monkey, as shown for compounds 33
and 75 compared to 4 and 11.
Comparison of compounds 11 and 75 head-to-head in various
animal model studies revealed little difference in efficacy. Com-
pound 75 has higher exposure than compound 11 following oral

P. Zhang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2179–2185
2185
Froelich, L. L.; Buben, J.; Craft, T. J. Bioorg. Med. Chem. Lett. 2007, 17, 4832; (g)
Corte, J. R.; Fang, T.; Pinto, D. J. P.; Han, W.; Hu, Z.; Jiang, X.-J.; Li, Y.-L.; Gauuan,
J. F.; Hadden, M.; Orton, D.; Rendina, A. R.; Luettgen, J. M.; Wong, P. C.; He, K.;
Morin, P. E.; Chang, C.-W.; Cheney, D. L.; Knabb, R. M.; Wexler, R. R.; Lam, P. Y.
S. Bioorg. Med. Chem. Lett. 2008, 18, 2845.
Hammond, T. G. Cardiovas. Res. 2003, 58, 32; (d) Shah, R. R. Drug Saf. 2005, 28,
115.
14. (a) Waring, M. J.; Johnstone, C. Bioorg. Med. Chem. Lett. 2007, 17, 1759; (b)
Jamieson, C.; Moir, E. M.; Rankovic, Z.; Wishart, G. J. Med. Chem. 2006, 49, 5029.
15. (a) Sinha, U.; Edwards, S. T.; Wong, P. W.; Delaney, S. M.; Nanda, N.; Marzec, U.
M.; Pak, Y.; Hollenbach, S. J.; Hanson, S. R.; Scarborough, R. M. Blood 2006, 108.
Abstract 907; (b) Abe, K.; Siu, G.; Edwards, S.; Lin, P. H.; Zhu, B.-Y.; Marzec, U.
M.; Hanson, S. R.; Pak, Y.; Hollenbach, S. J.; Sinha, U. Blood 2006, 108. Abstract
901.
7. Jones, G.; Willet, P.; Glen, R. C.; Leach, A. R.; Taylor, R. J. Mol. Biol. 1997, 267, 727.
8. Adler, M.; Kochanny, M. J.; Ye, B.; Rumennik, G.; Light, D. R.; Biancalana, S.;
Whitlow, M. Biochemistry 2002, 41, 15514.
9. In vitro liver metabolic stability was determined using liver S9 fractions and
expressed as half-life: Compounds (1
l
M) were incubated with liver S9
16. The rabbit 2ÂPT values were determined by the method described in Ref. 5(b).
17. Human plasma protein binding, expressed as unbound fraction, was
determined by this protocol: The compound stock solutions were diluted in
1.0 M HEPES (pH 7.4) to yield a 1.0 mM working solution. The working solution
was added to plasma samples (EDTA was used as anticoagulant) in a ratio of 1/
proteins (1 mg/mL) and a NADPH regenerating system (NADP + glucose-6-
phosphate) suspended in phosphate buffer (100 mM). Samples were taken at
t = 0, 5, 10, 15 and 60 min. and added to an organic solvent to terminate the
reaction. After adding an internal standard, the mixtures were centrifuged and
the supernatant transferred to HPLC vials. Samples were analyzed by tandem-
mass spectrometry. The change in area ratio (substrate/internal standard)
between t = 0 and 15 min was used to estimate half-life. Coumarin 7-
hydroxylase (CYP2A6), caffeine demethylase (CYP1A2), verapamil (CYP3A4)
and/or propanolol (CYP3A4) were used as positive control. For negative
controls, substrates were incubated with heat inactivated (60 min at 60 °C)
liver S9 fractions.
100 yielding a final concentration of 10
incubated at 37 °C for 30 min. At the conclusion of the incubation 3 aliquots
(450 L) each were added to a centrifugal filter device fitted to a 96 well plate.
lM. The mixture was gently mixed and
l
Standards were prepared in protein free human plasma and transferred to the
centrifugal filter device fitted to the same plate. The plates were centrifuged for
25 min at 32 °C in a centrifuge. 15
round bottom 96 well plate and 15
l
l
L each of the filtrate was transferred to a
L of acetonitrile including KN1022 (1 g/
mL) as internal standard was added followed by 60 lL of DI water. Plates were
l
10. (a) Jia, Z. J.; Wu, Y.; Huang, W.; Zhang, P.; Song, Y.; Woolfrey, J.; Sinha, U.;
Arfsten, A. E.; Edwards, S. T.; Hutchaleelaha, A.; Hollenbach, S. J.; Lambing, J. L.;
Scarborough, R. M.; Zhu, B.-Y. Bioorg. Med. Chem. Lett. 2004, 14, 1229; (b) Jia, Z.
J.; Su, T.; Zuckett, J. F.; Wu, Y.; Goldman, E. A.; Li, W.; Zhang, P.; Clizbe, L. A.;
Song, Y.; Bauer, S. M.; Huang, W.; Woolfrey, J.; Sinha, U.; Arfsten, A. E.;
Hutchaleelaha, A.; Hollenbach, S. J.; Lambing, J. L.; Scarborough, R. M.; Zhu, B.-
Y. Bioorg. Med. Chem. Lett. 2004, 14, 2073.
11. Zhu, B.-Y.; Jia, Z. J.; Zhang, P.; Su, T.; Huang, W.; Goldman, E.; Tumas, D.;
Kadambi, V.; Eddy, P.; Sinha, U.; Scarborough, R. M.; Song, Y. Bioorg. Med. Chem.
Lett. 2006, 16, 5507.
12. Finalayson, K.; Turnbull, L.; January, C. T.; Sharkey, J.; Kelly, J. S. Eur. J.
Pharmacol. 2001, 430, 147.
13. (a) De Ponti, F.; Poluzzi, E.; Cavalli, A.; Recanatini, M.; Montanaro, N. Drug Saf.
2002, 25, 263; (b) Mitcheson, J. S.; Perry, M. D. Curr. Opin. Drug Disc. Dev. 2003,
6, 667; (c) Redfern, W. S.; Carlsson, L.; Davis, A. S.; Lynch, W. G.; Mackenzie, I.;
Palethorpe, S.; Siegl, P. K. S.; Strang, I.; Sullivan, A. T.; Wallis, R.; Camm, A. J.;
placed on a Multi-Tube Vortexer for 30 s and vortexed. Concentrations in the
filtrate were determined by LC/MS/MS and using standard curves prepared in
ultra filtrated plasma.
18. The patch clamp assay utilizes cultured, whole Human Embryonic Kidney
(HEK) 293 cells stably transfected with the hERG gene to evaluate the potential
of a compound to interfere with the potassium-selective IKr current generated
under normoxic conditions. Interference of the potassium channel suggests
that the compound could modify (shorten or prolong) the electrocardiographic
(ECG) QT interval. (a) Zhou, Z.; Gong, Q.; Ye, B.; Fan, Z.; Makielski, J. C.;
Robertson, G. A.; January, C. T. Biophys. J. 1998, 74, 230; (b) Zhou, Z.; Gong, Q.;
Ye, B.; Fan, F.; Makielski, J. C.; Robertson, G. A.; January, C. T. Biophys. J. 1997, 72,
A225; (c) Hancox, J. C.; Levi, A. J.; Witchel, H. J. Pflugers Arch. 1998, 436, 843; (d)
Jurkat-Rott, K.; Lehmann-Horn, F. Curr. Pharm. Biotechnol. 2004, 5, 387.
19. Turpie, A. G. G.; Bauer, K. A.; Davidson, B. L.; Fisher, W. D.; Gent, M.; Huo, M. H.;
Sinha, U.; Gretlet, D. D. Thromb. Haemostasis 2009, 101, 68.