Cyanoguanidine-based lactam derivatives as a novel class of orally bioavailable factor Xa inhibitors

Article abstract of DOI:10.1016/j.bmcl.2009.06.014

The N,N′-disubstituted cyanoguanidine is an excellent bioisostere of the thiourea and ketene aminal functional groups. We report the design and synthesis of a novel class of cyanoguanidine-based lactam derivatives as potent and orally active FXa inhibitor

Full text of DOI:10.1016/j.bmcl.2009.06.014

Bioorganic & Medicinal Chemistry Letters 19 (2009) 4034–4041
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Cyanoguanidine-based lactam derivatives as a novel class of orally bioavailable
factor Xa inhibitors
à
*
Yan Shi , Jing Zhang , Mengxiao Shi , Stephen P. O’Connor, Sharon N. Bisaha, Chi Li, Doree Sitkoff,
Andrew T. Pudzianowski, Saeho Chong, Herbert E. Klei, Kevin Kish, Joseph Yanchunas Jr.,
Eddie C.-K. Liu, Karen S. Hartl, Steve M. Seiler, Thomas E. Steinbacher, William A. Schumacher,
Karnail S. Atwal, Philip D. Stein
Research and Development, Bristol-Myers Squibb Company, PO Box 5400, Princeton, NJ 08543-5400, USA
a r t i c l e i n f o
a b s t r a c t
The N,N0-disubstituted cyanoguanidine is an excellent bioisostere of the thiourea and ketene aminal func-
tional groups. We report the design and synthesis of a novel class of cyanoguanidine-based lactam deriv-
atives as potent and orally active FXa inhibitors. The SAR studies led to the discovery of compound 4
Article history:
Received 23 February 2009
Revised 5 June 2009
Accepted 8 June 2009
Available online 13 June 2009
(BMS-269223, Ki = 6.5 nM, EC2xPT = 32 lM) as a selective, orally bioavailable FXa inhibitor with an excel-
lent in vitro liability profile, favorable pharmacokinetics and pharmacodynamics in animal models. The
X-ray crystal structure of 4 bound in FXa is presented and key ligand–protein interactions are discussed.
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Factor Xa inhibitor
Cyanoguanidine
Bioisostere
Thiourea
Ketene aminal
Bioavailable
The trypsin-like serine protease factor Xa (FXa) has been one of
the major targets for antithrombotic agent development for some
time.1 Within the blood coagulation cascade, factor Xa functions
at the point where the intrinsic and extrinsic coagulation pathways
converge,2 and FXa is the key enzyme responsible for thrombin
activation. Thrombin plays a major role in thrombosis and hemos-
tasis by mediating a feed-back loop that amplifies its production,
inducing platelet aggregation and converting fibrinogen to fibrin
which ultimately leads to clot formation. Selective FXa inhibitors
are believed to have a wider therapeutic window than direct
thrombin inhibitors, since FXa inhibition could significantly reduce
thrombin generation and prevent thrombus formation, while the
basal levels of thrombin would still facilitate hemostasis and re-
duce unwanted bleeding risk.2c Recent preclinical and clinical data
indicate that selective inhibition of factor Xa is a highly effective
approach to the prevention and treatment of arterial and venous
thromboembolism.3
Factor Xa contains a deep S1 cavity and a box-like S4 enclosure
near the enzyme’s active site. Potent FXa inhibitors generally re-
quire an S1 and an S4 binding element which are connected
through an L-shaped scaffold. We have previously disclosed a ser-
ies of caprolactam-based FXa inhibitors containing a thiourea 14 or
a ketene aminal 25 as linkers of the P1 and P4 pharmacophores.
While these compounds are selective and orally active FXa inhibi-
tors, their in vivo efficacy was limited by their moderate in vitro
potencies and short plasma half lives. In this Letter, we describe
the cyanoguanidine as an alternative bioisostere to the thiourea
and ketene aminal motifs that can adopt the desired L-shaped con-
formation, and the detailed SAR studies of both P1 and P4 groups.
The X-ray structure of compound 4 (BMS-269223) bound to FXa
and the detailed analysis of its interactions with FXa, as well as
the PK/PD data and in vivo pharmacology of compound 4, will also
be discussed.
The N,N0-disubstituted cyanoguanidine is known to be an excel-
lent bioisostere of N,N0-disubstituted thiourea from previous SAR
studies of histamine H2-receptor antagonists.6 For example, the
cynaoguanidine moiety in Cimetidine is a bioisostere of the thio-
urea functional group in Metiamide6a (Scheme 1). Similar to its
thiourea and 1-nitroketene aminal analogs4,5, the 1-methyl-3-phe-
nyl-2-cyanoguanidine (compound 3) is predicted to prefer the
anti–syn1 conformation by gas phase ab initio calculations.7
* Corresponding author. Tel.: +1 609 818 4124; fax: +1 609 818 3450.
E-mail address: yan.shi@bms.com (Y. Shi).
Present address: Hoffman-La Roche Inc., 340 Kingsland Street, Nutley, NJ 07110,
USA.
à
Present address: Wyeth Research, 401 N Middletown Road, Pearl River, NY 10965,
USA.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.06.014

Y. Shi et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4034–4041
4035
O
O
H
N
H
N
H
N
H
N
N
N
Me
N
N
Me
O
O
S
O
NC
CN
2
1
IC50 = 30 nM
IC50 = 110 nM
CN
N
N
N
S
S
HN
HN
S
N
H
N
H
N
H
N
H
Cimetidine
Scheme 1. The thiourea 1, ketene aminal 2, Metiamide and Cimetidine.
Metiamide
syn-anti1
anti-anti1
anti-syn1
anti-anti2
CN
NC
N
NC
N
CN
N
N
H
H
H
Me
Ph
H
H
H
N
N
N
Ph
N
H
N
H
N
N
N
Ph Me
2.4 kcal/mol
Scheme 2. Relative energies of conformers of 1-methyl-3-phenyl-2-cyanoguanidine 3.
Me
Ph Me
0.0 kcal/mol
3.3 kcal/mol
1.9 kcal/mol
Scheme 2 shows the four lowest energy conformers for 3. The other
possible conformers are disfavored due to the internal steric inter-
actions between the nitrogen substituents and the cyano group.
The anti–syn1 conformation affords the best combination of relief
from steric crowding between the cyano, the phenyl and methyl
groups.
IC50 = 84 nM) is sevenfold less potent than 4, while increasing the
size of the 2-substituent from methyl to ethyl, resulted in a loss
in activity of >1000-fold (16, IC50 = 15,538 nM). Changing the 2-
methyl- to a 2-cyano-benzofuran led to a 20-fold less active com-
pound 17 (IC50 = 2600 nM), while relocating the methyl group to
the
3 position resulted in a 245-fold loss in activity (18,
Replacement of the 1,1-dicyanoketene aminal functionality in 2
with a cyanoguanidine gave compound 4 with an IC50 of 12 nM
IC50 = 2939 nM). The 2,3-dimethyl substituted analog 19 is about
30-fold less active than 4.
against FXa and EC2xPT of 30
l
M (Table 1).8 This compound is about
The SAR for the substituents of the phenyl ring of the 2-methyl
benzofuran was also explored. While the 7-fluoro-2-methyl benzo-
furan analog (20, IC50 = 11 nM) has similar anti-FXa activity as
compound 4, both the 4- and 6-fluoro analogs (21 and 22) are less
active. Increasing the size of the 7-substituent beyond a fluorine
atom also produced less active compounds, with the 7-Cl (23)
and 7-methyl (24) analogs showing about twofold loss of activity
compared to 4. Incorporation of bulkier 7-trifluoromethyl (25)
and 7-methoxy (26) substituents led to significant losses of
activity.
threefold more potent than 2 (IC50 = 30 nM), indicating that both
compounds likely adopt similar conformations when bound to
FXa, and that the cyanoguanidine may even be a better linker than
the ketene aminal in 2. Similar SARs are observed for both series
when the P1 group is a mono- or disubstituted phenyl group (data
not shown).5 In addition, analogs with a P1 group derived from
alkylamines were uniformly poor FXa inhibitors, similar to results
found in the thiourea series (data not shown).4 Thus, we focused
our efforts in SAR optimization on compound 4.
Early efforts were concentrated on the possible isosteres of the
2-methylbenzofuran-7-yl pharmacophore to improve the potency.
All efforts to replace the 2-methylbenzofuran with other bicyclic
aryls resulted in significant losses in activity (Table 1). For exam-
ple, the 2-napthyl analog (5, IC50 = 780 nM) is about 60-fold less
potent than 4; replacement of the oxygen atom in 2-methyl benzo-
We also studied the SAR of the central caprolactam ring by
varying its ring size and the chirality of its amino group. The results
are summarized in Table 3. Compound 4, the S-enantiomer with a
seven-membered lactam ring, is the most potent analog among
these compounds. It is about 77-fold more potent than its R-enan-
tiomer 30 (IC50 = 940 nM) and the corresponding (S)-six-mem-
bered valerolactam analog 29 (IC50 = 928 nM). Both enantiomers
of the five-membered lactam are very weak inhibitors of FXa.
The racemic form of the eight-membered lactam analog 31 is also
less potent than 4 with an IC50 of 195 nM.
With the optimized P1 pharmacophore and central lactam ring
in hand, we next turned our attention to the SAR studies of the P4
binding groups. Similar to the thiourea series,4 the replacement of
the pyrrolidine with non-cyclic amines, or modest changes to the
ring size (azetidine and piperidine) led to significant losses in po-
tency (data not shown). Substitutions at the C-2 and C-3 positions
on the pyrrolidine ring were well tolerated but none of these ana-
furan with either
a sulfur (6, IC50 = 2882 nM) or NH (10,
IC50 > 34,000 nM) led to analogs with much poorer activity com-
pared to 4. The isomeric 2-methylbenzofuran-6-yl analog 7 is com-
pletely inactive; while changing the 3-CH in the furan ring to a
nitrogen atom led to compound 8 (IC50 = 2677 nM) with a 223-fold
loss of activity. Other isosteric heteroaryl replacements (com-
pounds 11–14) are much less potent than 4.
We next studied the SAR of substitutions on the 2-methyl ben-
zofuran pharmacophore. Most changes to the substitution pattern
on the furan ring resulted in significant losses in activity (Table 2).
For example, the analog without the 2-methyl substituent (15,

4036
Y. Shi et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4034–4041
Table 1
Table 2
SAR of the aromatic binding element
SAR of the substituted benzofurans
R2
O
O
H
H
Ar
N
H
N
H
N
R1
2
N
N
N
N
N
N
N
O
O
O
CN
7
CN
4, 15-26
4-14
Compound
R1
2-H
2-Me
2-Et
2-CN
2-Me
2,3-Dimethyl
2-Me
2-Me
2-Me
2-Me
2-Me
R2
IC50 (nM)a
Compound
Ar
IC50 (nM)a
4
15
16
17
18
19
20
21
22
23
24
25
26
H
H
H
H
H
H
7-F
4-F
6-F
12
84
15,538
2600
2939
349
11
2728
76
27
4
12
Me
O
5
6
780
7-Cl
7-Me
7-CF3
21
167
14,818
2882
Me
2-Me
2-Me
S
7-OMe
a
IC50’s are measured against human factor Xa utilizing the cleavage of a syn-
thetic substrate S-2222. Data are the average of two independent determinations.
O
7
>34,000
2677
Me
Me
N
O
Table 3
8
SAR of the central lactam core
O
H
N
H
N
N
O
N
N
Me
9
18,759
>34,000
11,879
>34,000
1300
Me
Me
N
O
O
CN
(CH2)1-4
27-31
10
11
12
13
N
H
Compound
Lactam core
IC50 (nM)a
>32,000
N
O
Me
Me
H
N
27
N
H
N
O
O
N
H
N
28
29
30,000
928
N
N
H
N
H
N
Me
Me
N
N
N
O
H
N
14
2400
N
30
940
195
N
a
IC50’s are measured against human factor Xa utilizing the cleavage of a syn-
thetic substrate S-2222. Data are the average of two independent determinations.
O
H
N
31
logs were significantly more potent than the unsubstituted ana-
log 4. For example, compound 32 with a cyclopropane fused to
the pyrrolidine has a similar IC50 (15 nM) to 4 but with a weaker
EC2xPT (56 lM). The 2-benzyl substituted pyrrolidine 33
(IC50 = 19 nM) is slightly more potent than the 3-substituted ana-
a
IC50’s are measured against human factor Xa utilizing the cleavage of a syn-
thetic substrate S-2222. Data are the average of two independent determinations.
log 34 (IC50 = 53 nM), but both show diminished EC2xPT
The tight SAR observed for both the P1 and P4 binding pharma-
cophores can be rationalized by examining the X-ray crystal struc-
ture of compound 4 bound to FXa resolved at 2.2 Å shown in Figure
1.11 As expected, the cynanoguanidine adopts the anti–syn confor-
mation with the CN group positioned atop the disulfide bond. The
central caprolactam carbonyl forms a hydrogen bond with NH of
Gly216 (3.16 Å). The 2-methylbenzofuran moiety binds to FXa in
the S1 pocket and the pyrrolidine is placed squarely in the S4
pocket.
(>50 l
M).9 Among the 2-aminomethyl substituted pyrrolidine ana-
logs 35–38, the R-enantiomers 36 and 38 are more potent than
their S-enantiomers 35 and 37, although the magnitude of differ-
ence is less dramatic in the 37/38 than in the 35/36 pair. Finally,
the bicyclic analog 39, which replaces the pyrrolidine in 4 with a
3,8-diazabicyclo[3.2.1]octane, provides the only compound more
potent than 4 in this series with an IC50 of 5 nM and EC2xPT of
11.5 l
M.10

Y. Shi et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4034–4041
4037
Table 4
SAR of the P4 binding pharmacophores
R1
N
Gln192
O
H
N
H
N
R2
N
Me
O
N
O
Tyr99
CN
Cys191
Cys220
3.2 Å
4, 32-39
NR1R2
IC50 (nM)a
EC2ꢀPT
b
Compound
Gly216
Val213
Trp215
4
12
30
N
Phe174
Asp189
Tyr228
32
33
34
35
36
37
15
19
53
386
22
54
56
N
Figure 1. Crystal structure of 4 bound in FXa. The hydrogen bond between the
ligand’s lactam carbonyl and Gly216 backbone amide NH is highlighted. The
conserved crystallographic water molecule in the far end of the P4 pocket is shown
(red sphere). This and the following figures were created using PyMOL (http://
www.pymol.org).
Ph
>50
>50
—
N
N
Ph
2a is actually above, rather than in the plane of the 2-methyl ben-
zofuran ring, correlating with the reduced activity of 22. The losses
in activity seen in the 2-methylbenzofuran isosteres in Table 1
(compounds 6–14) can also be understood to some degree by
examining the crystal structure. The 2-methylbenzofuran pharma-
cophore is likely favored due to two attributes in addition to the
space-filling nature of its shape: (1) it is overall relatively hydro-
phobic; (2) the slight polarization of the 2-methylbenzofuran ring
is such that C3 is likely to have a small net partial positive charge
due to the furan oxygen, allowing the 3-CH to interact favorably
with Asp189. All of the isosteres differ in either one or both of
these attributes.
H2N
N
H2N
O
27
N
>50
N
H
N
Figure 2b shows the detail of the S4 binding pocket with the
favorable interaction volume, calculated by a GRID methyl probe
in the absence of 4.12 The carbonyl group of the pyrrolidine fills a
pocket lined by the aliphatic portion of the side chain of Glu217.
The pyrrolidine itself has a complementary shape to the favorable
interaction volume, correlating with the observed difficulties in
further optimizing this group. The single compound (39) showing
an improvement in activity over compound 4, may have a P4 group
that augments the pyrrolidine-S4 shape complementarity with
favorable electrostatic interactions between the ligand’s charged
NH and several potential partners near the S4 binding pocket,
namely the backbone carbonyls of Thr97 or Glu98, or the con-
served S4 water molecule, or some combination thereof; such
interactions are seen in multiple examples of FXa-ligand crystal
structures in the Protein Data Bank.13
O
38
28
5
33
N
H
N
NH
39
11.5
N
a
IC50’s are measured against Human Factor Xa utilizing the cleavage of a syn-
thetic substrate S-2222.
Concentration of inhibitor required to double the prothrombin based clotting
time in human plasma. Data are the average of two independent determinations.
b
Cyanoguanidines 4–31 were synthesized by the stepwise dis-
placement of both phenoxy groups of diphenyl cyanocarbonimi-
date 40 as shown in Scheme 3. Alternatively, compound 4 can
also be prepared in a one-pot sequence according to a literature
procedure.14 Treatment of the 2-methyl benzofuran-7-isothiocya-
nate 425 with sodium cyanoamide gave the intermediate 47, which
was reacted with caprolactam amine 43S4 and EDCI to afford 4 in
90% yield. Compounds 32–39 were prepared from the common
intermediate acid 50 according to Scheme 4. Reaction of 47 and
amine 48 with the assistance of EDCI provided methyl ester 49.
Hydrolysis of 49 afforded the acid 50, which is then coupled with
amines to produce 32–34 and BOC protected compounds 51a,
51b, and 52. Compounds 51a, 51b and 52 were treated with TFA
to yield 35, 36 and 39. Acetylation of 35 and 36 with acetic anhy-
dride and triethylamine furnished 37 and 38, respectively (Table
4).
Figure 2a shows the detail of the S1 binding pocket with the
favorable interaction volume, calculated by a GRID methyl probe
in the absence of 4.12 The 2-methyl substituent of the benzofuran
occupies the pocket at the base of S1, and there appears to be little
room for increasing the size of this substituent; this correlates with
the loss in activity seen for the 2-ethylbenzofuran compound 16.
Substitutions at other positions of the benzofuran were also found
to be detrimental for the most part, in accordance with the view
shown in Figure 2a. One exception is that small substituents at po-
sition 7 were tolerated (compounds 20, 23, 24), in contrast to the
expectation based on the apparent tight fit of 2-methyl benzofuran
to the favorable interaction volume at this position. Position 7 is di-
rected at the Ser195 side chain, which has been observed in some
crystal structures to rotate and free up space for slightly larger li-
gands;8 this may explain the tolerance for slight increases in the
size of the 7-substituent. The contoured space near C6 in Figure

4038
Y. Shi et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4034–4041
a
b
Gln192
Tyr99
Gly218
Cys191
Cys220
Ser195
Val213
Gly216
Ala190
Trp215
Phe174
Asp189
Glu217
Tyr228
Figure 2. Views of the (a) S1 and (b) S4 pockets of the FXa crystal structure complex with 4. Available favorable interaction volumes, calculated with a GRID15 methyl probe in
the absence of 4, are shown in brown contours (mapped at interaction level of ꢁ2.5 kcal/mol).
O
H
N
H
N
H
N
N
PhO
OPh
OPh
b
a
N
Ar
Ar
O
N
N
N
CN
CN
CN
40
41
4-31
O
O
N
SCN
H2N
N
N
H2N
Me
N
O
O
O
(CH2)n
42
43S
44S: n = 1
44R: n = 1
43S: n = 3
43R: n = 3
45S: n = 2
46: n = 4
O
H
H
N
H
N
N
N
SNa
d
c
N
Me
42
O
Me
N
N
O
O
CN
CN
47
4
Scheme 3. The synthesis of 4-31. Reagents and conditions: (a) ArNH2, CH3CN, 80 °C; (b) amine from 43 to 46, 80 °C (15–65% yield over two steps); (c) sodium cyanoamide,
DMF 60 °C; (d) 43S, 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide hydrchloride (EDCI) (90% yield).
Similar to the thiourea 14 and ketene aminal 25, compound 4 is
a selective FXa inhibitor relative to related trypsin-like serine pro-
teases (570-fold vs t-PA and >1000-fold vs other serine proteases
tested) as shown in Table 5. In addition, compound 4 is not a sig-
nificant inhibitor of a panel human cytochrome P450 enzymes; it
shows no significant activity in hERG, sodium channel, PXR and
HHA assays.15 It is negative in an Ames microbial mutagenicity
2 + 2, 5 + 5, 10 + 10 and 20 + 20 which produced dose-dependent
prolongation of ex vivo prothrombin time (Fig. 3) and inhibition
of vena cava and carotid artery thrombosis (Fig. 4). Compound 4
was more potent in the venous thrombosis model
(ED50 = 6 1 mg/kg total dose, Fig. 4a) compared to the arterial
thrombosis model (ED50 = 25 5 mg/kg total dose, Fig. 4b). Anti-
thrombotic activity was also evidenced by improved blood flow
in arterial thrombosis. Whereas a high proportion of arteries were
occluded with vehicle treatment (13/14 rats), occlusion rates were
significantly reduced at compound 4 dose levels of 5 + 5 (4/8 rats),
10 + 10 (1/8 rats) and 20 + 20 (0/8 rats), but not at the 2 + 2 dose
level (9/9rats).
Compound 4 was also effective in prolonging prothrombin time
when administered as an intraduodenal injection (Fig. 5). A 30 mg/
kg intraduodenal injection of BMS-269223 increased prothrombin
time by a factor of P1.8 for up to 120 min. In comparison, the
10 mg/kg + 10 mg/kg/h intravenous infusion increased prothrom-
study16 up to 400
of activity against 35 receptors and enzymes at 10
l
g/mL, and is inactive in an in vitro assessment
M (Cerep,
l
Paris, France).17 Furthermore, compound 4 exhibits good oral bio-
availability in rats (50%), dogs (77%) and monkeys (49%). The ter-
minal elimination half life ranged from 2.4 to 10.3 h and the
volume of distribution was consistently low in preclinical animal
models as shown in Table 6.
The anticoagulant and antithrombotic effects of 4 were deter-
mined in anesthetized rats.18 Compound 4 was administered as
initial loading plus sustaining infusions (mg/kg + mg/kg/h) of

Y. Shi et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4034–4041
4039
O
H
N
H
N
H
N
OMe
SNa
b
a
N
42
Me
Me
O
N
N
O
O
CN
CN
47
49
R1
N
O
O
H
N
H
N
H
N
H
OH
R2
N
d
N
N
c
Me
Me
O
O
N
N
O
O
CN
CN
50
32-34, 51a, 51b, 52
BOC
N
H2N
H2N
51a: NR1R2
51b: NR1R2
=
=
35: NR1R2
36: NR1R2
=
H
N
N
BOC
e
N
=
=
H
N
N
BOC
N
NH
52: NR1R2
=
39: NR1R2
N
N
O
O
H
N
H
N
N
OMe
H2N
f
N
N
Me
O
35, 36
O
O
HN
N
O
CN
37, 38
48
Scheme 4. The synthesis of cyanoguanidine 32–39. Reagents and conditions: (a) sodium cyanoamide, DMF, 60 °C; (b) 48, 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide
hydrchloride (EDCI), 90% yield; (c) LiOHꢂ2H2O, THF–H2O 98% yield; (d) HNR1R2, EDCI, Et3N, CH2Cl2, (75–90% yield); (e) TFA, CH2Cl2, 80–98% yield; (f) acetic anhydride, Et3N,
CH2Cl2, 50–85% yield.
Table 5
Selectivity profile of 4
4.5
= SEM
(n=4 per dose level)
4.0
3.5
3.0
2.5
2.0
1.5
1.0
Human enzyme
Ki (nM)a
Factor Xa
Trypsin
Thrombin
Tryptase
Activated protein C
u-PA
6.5
>10,000
6600
>16,000
>18,000
>7800
>16,000
3700
Plasmin
t-PA
a
Ki’s are calculated from the IC50 values assuming competitive inhibition versus
the low molecular weight synthetic substrates using the relationship, Ki = IC50
/
0
10+10
20+20
30+30
40+40
(1 + [S]/Km), where S is the substrate concentration in the assay and Km is the
Michaelis constant for that substrate. Data are the average of two independent
determinations.
BMS-269223 Dose
(mg/kg + mg/kg/hr, i.v.)
Figure 3. Prolongation of prothrombin time by intravenous infusion of 4 (BMS-
269223).
Table 6
Pharmacokinetic parameters for 4 (n = 3, fasted)a
bin time by a comparable factor of 1.8 and inhibited both venous
and arterial thrombosis by P50%. This suggests that efficacy could
also be achieved by oral dosing.
Parameter (unit)
Rat
Dog
Monkey
iv Dose (mg/kg)
CLtotal (L/kg)
Vdss (L/kg)
t1/2
MRT (h)
po Dose (mg/kg)
Cmax
7
8.3
0.12
0.11
10.3
16
7
75
0.6
77
10
2.2
1.8
2.9
1.5
10
40
—
1.7
0.19
2.4
1.9
10
In summary, we have described a series of cyanoguanidine-
based lactam derivatives as a novel class of orally bioavailable
FXa inhibitors. The N,N0-disubstituted cyanoguanidines are excel-
lent bioisosteres of their corresponding ketene aminal and thiourea
functional groups. The SAR studies led to the discovery of
compound 4 (BMS-269223) as a selective, orally bioavailable FXa
inhibitor with an excellent in vitro liability profile, favorable
pharmacokinetics and pharmacodynamics in animal models.
23
Tmax
BA (%)
0.9
50
49
a
Dosing vehicle: 10 mg/mL in PEG400/water (1:1, v/v).

4040
Y. Shi et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4034–4041
a
b
*
p<0.05 to Vehicle
50
40
30
20
10
0
6
5
4
3
2
1
0
= SEM
-39%
*
-52%
-45%
-55%
*
*
*
*
-63%
*
-83%
*
Vehicle 2+2 5+5 10+10
(n=7) (n=5) (n=5) (n=5)
Vehicle 2+2 5+5 10+10 20+20
(n=14) (n=9) (n=8) (n=8) (n=8)
BMS-269223
(mg/kg + mg/kg/hr, i.v.)
BMS-269223
(mg/kg + mg/kg/hr, i.v.)
Figure 4. Inhibition of venous and arterial thrombosis by 4 (BMS-269223).
A. W.; Brandt, J. T. ASH Ann. Meet. Abstr. 2005, 106, 278; (i) Hampton, T. JAMA J.
Am. Med. Assoc. 2006, 295, 743; (j) Liebeschuetz, J. W.; Jones, S. D.; Wiley, M. E.;
Young, S. C. Struct.-Based Drug Discovery 2006, 173.
4
3.5
3
4. Bisacchi, G. S.; Stein, P. D.; Gougoutas, J. Z.; Hartl, K. S.; Lawrence, R. M.; Liu, E.
C.; Pudzianowski, A. T.; Schumacher, W. A.; Sitkoff, D.; Steinbacher, T. E.;
Sutton, J.; Zhang, Z.; Seiler, S. M. Lett. Drug Des. Discovery 2005, 2, 625.
5. Shi, Y.; Zhang, J.; Stein, P. D.; Shi, M.; O’Connor, S. P.; Bisaha, S. N.; Li, C.; Atwal,
K. S.; Bisacchi, G. S.; Sitkoff, D.; Pudzianowski, A. T.; Liu, E. C.; Hartl, K. S.; Seiler,
S. M.; Youssef, S.; Steinbacher, T. E.; Schumacher, W. A.; Rendina, A. R.; Bozarth,
J. M.; Peterson, T. L.; Zhang, G.; Zahler, R. Bioorg. Med. Chem. Lett. 2005, 15, 5453.
6. (a) Silverman, R. B. In The Organic Chemistry of Drug Design and Drug Action;
Academic: New York, 1992; pp 159–165; (b) Durant, G. J.; Emmett, J. C.;
Ganellin, C. R.; Miles, P. D.; Parsons, M. E.; Prain, H. D.; White, G. R. J. Med. Chem.
1977, 20, 901; (c) Black, J. W.; Durant, G. J.; Ganellin, C. R. Nature 1974, 248, 65;
(d) Bradshaw, J. C.; Ganellin, C. R.; Parsons, M. E. J. Int. Med. Res. 1975, 3, 86; (e)
Domschke, W.; Lux, G.; Domschke, D. Lancet 1979, 8111, 320.
7. All calculations were done at the RI-MP2/6-31+G(d,p) level of theory with the Q-
CHEM 3.0 program system: (a) Feyereisen, M.; Fitzgerald, G.; Komornicki, A.
Chem. Phys. Lett. 1993, 208, 359; (b) Shao, Y.; Fusti-Molnar, L.; Jung, Y.;
Kussmann, J.; Ochsenfeld, C.; Brown, S. T.; Gilbert, A. T. B.; Slipchenko, L. V.;
Levchenko, S. V.; O’Neill, D. P.; Distasio, R. A. Jr.; Lochan, R. C.; Wang, T.; Beran,
G. J. O.; Besley, N. A.; Herbert, J. M.; Lin, C. Y.; Van Voorhis, T.; Chien, S. H.; Sodt,
A; Steele, R. P.; Rassolov, V. A.; Maslen, P. E.; Korambath, P. P.; Adamson, R. D.;
Austin, B.; Baker, J.; Byrd, E. F. C.; Dachsel, H.; Doerksen, R. J.; Dreuw, A.;
Dunietz, B. D.; Dutoi, A. D.; Furlani, T. R.; Gwaltney, S. R.; Heyden, A.; Hirata, S.;
Hsu, C.-P.; Kedziora, G.; Khalliulin, R. Z.; Klunzinger, P.; Lee, A. M.; Lee, M. S.;
Liang, W.; Lotan, I.; Nair, N.; Peters, B.; Proynov, E. I.; Pieniazek, P. A.; Rhee, Y.
M.; Ritchie, J.; Rosta, E.; Sherrill, C. D.; Simmonett, A. C.; Subotnik, J. E.;
Woodcock, H. L. III; Zhang, W.; Bell, A. T.; Chakraborty, A. K.; Chipman, D. M.;
Keil, F. J.; Warshel, A.; Hehre, W. J.; Schaefer III, H. F.; Kong, J.; Krylov, A. I.; Gill,
P. M. W.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2006, 8, 3172.
I.D. (30 mg/kg, n=5)
I.V. (10 mg/kg, n=4)
BMS-269223
2.5
2
1.5
1
0
20
40
60
80
100 120
Time (min)
Figure 5. Prolongation of prothrombin time by intravenous and intraduodenal
injection of 4 (BMS-269223).
Acknowledgments
We thank Drs. P. T. Cheng, P. Y. S. Lam, R. R. Wexler for insight-
ful review of this work.
References and notes
8. For detailed assay methods, see: Shi, Y.; Sitkoff, D.; Zhang, J.; Klei, H. E.; Kish, K.;
Liu, E. C.-K.; Hartl, K. S.; Seiler, S. M.; Chang, M.; Huang, C.; Youssef, S.;
Steinbacher, T. E.; Schumacher, W. A.; Grazier, N.; Pudzianowski, A.; Apedo, A.;
Discenza, L.; Yanchunas, J., Jr.; Stein, P. D.; Atwal, K. S. J. Med. Chem. 2008, 51,
7541.
9. Several literature reports suggest that highly nonpolar coagulation factor
inhibitors can have a tendency toward high protein binding, and this tendency
could be associated with the relatively high concentrations required to achieve
efficacy in plasma-based assays and presumably in vivo. Most of compounds in
1. For reviews, see: (a) Kunitada, S.; Nagahara, T.; Hara, T. Handbook Exp.
Pharmacol. 1998, 397; (b) Vacca, J. P. Ann. Rep. Med. Chem. 1998, 33, 81; (c) Zhu,
B.-Y.; Scarborough, R. M. Curr. Opin. Cardiovacs., Pulmon., Ren. Investig. Drugs
1999, 1, 63; (d) Ewing, W. R.; Pauls, H. W.; Spada, A. P. Drugs Future 1999, 24,
771; (e) Fevig, J. M.; Wexler, R. R. Ann. Rep. Med. Chem. 1999, 34, 81; (f) Betz, A.
Exp. Opin. Ther. Patent 2001, 11, 1007.
2. (a) Davie, E. W.; Fujikawa, K.; Kisiel, W. Biochemistry 1991, 30, 10363; (b)
Butenas, S.; Van ’t Veer, C.; Cawthern, K.; Brummel, K. E.; Mann, K. G. Blood
Coagul. Fibrin. 2000, 11, S9; (c) Wong, P. C.; Watson, C. A.; Crain, E. J.; Ogletree,
M. L.; Wexler, R. R.; Lam, P. Y. S.; Pinto, D. J.; Knabb, R. M. J. Thromb. Haemost.
2007, 5. Abstr. 1315.
this series have the concentrations to double the prothrombin time (EC2ꢀPT
) on
the order of 103 times higher than the corresponding IC50’s probably due to
plasma protein bounding. For example, compound 4 is 98.7% plasma protein
bound in human plasma. See Ref. 8.
3. For Apixaban, see: (a) Lassen, M. R.; Davidson, B. L.; Gallus, A.; Pineo, G.; Ansell,
J.; Deitchman, D. J. Thromb. Haemost. 2007, 5, 2368; (b) 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; (c) Wong, P. C.; Crain, E. J.; Xin, B.; Wexler, R. R.; Lam, P.
Y.; Pinto, D. J.; Luettgen, J. M.; Knabb, R. M. J. Thromb. Haemost. 2008, 6, 820; For
Rivaroxaban, see: (d) Roehrig, S.; Straub, A.; Pohlmann, J.; Lampe, T.;
Pernerstorfer, J.; Schlemmer, K.-H.; Reinemer, P.; Perzborn, E. J. Med. Chem.
2005, 48, 5900; (e) Eriksson, B. I.; Borris, L.; Dahl, O. E.; Haas, S.; Huisman, M. V.;
Kakkar, A. K.; Misselwitz, F.; Kaelebo, P. J. Thromb. Haemost. 2006, 4, 121; (f)
Kubitza, D.; Becka, M.; Wensing, G.; Voith, B.; Zuehlsdorf, M. Eur. J. Clin.
Pharmacol. 2005, 61, 873; (g) Turpie, A. G. G.; Fisher, W. D.; Bauer, K. A.; Kwong,
L. M.; Irwin, M. W.; Kalebo, P.; Misselwitz, F.; Gent, M. J. Thromb. Haemost. 2005,
3, 2479; For LY517717, see: (h) Agnelli, G.; Haas, S. K.; Krueger, K. A.; Bedding,
10. Compound 39 was synthesized much later after the discovery of compound 4,
and it was not fully characterized in vivo.
11. The X-ray crystal structure coordinates of 4 in human factor Xa have been
deposited in the Protein Data Bank (PDB code 3HPT).
12. Contours calculated using GRID (version 22a, Molecular Discovery) with
methyl probe C3, contoured at interaction energy level of ꢁ2.5 kcal/mol:
Goodford, P. J. J. Med. Chem. 1985, 28, 849.
13. Some examples are ascension codes 1lpk, 1nfu, 1nfy, 1xka, 1xkb, 1wu1, 2h9e.
14. Atwal, K. S.; Ahmed, S. Z.; O’Reilly, B. C. Tetrahedron Lett. 1989, 30, 7313.
15. Compound
4
has following safety profile: CYP (inhibition) 3A4-BFC
M, 2C9 IC50 = 29.2 M, others (3A4-BZR, 1A2, 2B6, 2C8, 2C19,
M; PXR EC50 > 25 M; hERG (flux) IC50 > 80 M; HHA
(2C19, 2C9, 2D6, 3A4, TC5) all IC50 > 200
M. Spectral data for compound 4: 1
NMR (500 MHz, CDCl3) d 7.42 (d, 1H, J = 8.1 Hz), 7.32 (d, 1H, J = 2.2 Hz), 7.07–
IC50 = 18.5
l
l
2D6) all IC50s > 40
l
l
l
l
H

Y. Shi et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4034–4041
4041
7.03 (m, 2H), 6.46 (d, 1H, J = 5.5 Hz), 6.38 (s, 1H), 4.61 (dd, 1H, J = 10.0, 5.5 Hz),
4.13–4.03 (m, 2H), 3.72–3.65 (m, 1H), 3.43 (t, 2H, J = 7.15 Hz), 3.37–3.30 (m,
2H), 3.24–3.19 (m, 1H), 2.46 (s, 3H), 2.18–2.12 (m, 1H), 2.0–1.92 (m, 3H), 1.84–
1.71 (m, 5H), 1.57–1.49 (m, 1H) ppm; Anal. Calcd for C23H28N6O3: C, 62.74; H,
6.61; N, 18.76, Found: C, 62.89; H, 6.69; N, 18.55; tR = 5.9 min with the HPLC
condition: Column-Zorbax SB C18 4.6 ꢀ 75 mm; flow-2.5 mL/min detected at
220 nm; solvent A = 10% CH3OH + 90% H2O + 0.2% H3PO4, solvent B = 90%
CH3OH + 10% H2O + 0.2% H3PO4; gradient-linear, 50% A to 0% A over 8 min
then hold.
transporter (r), Choline transporter (r). For details regarding the specific
receptors and experimental conditions see: http://www.cerep.fr/Cerep/Users/
pages/catalog/binding/catalog.asp.
18. Arterial and venous thrombosis were performed in pentobarbital-anesthetized
male Sprague Dawley rats as described in: Schumacher, W. A.; .Heran, C. L.:
Steinbacher, T. E. J. Cardiovas. Pharmacol. 1996, 28, 19. To measure ex vivo
clotting time, rats were fasted overnight and anesthetized with sodium
pentobarbital (50 mg/kg, ip). Catheters were placed in a jugular vein for iv
dosing, and into the duodenum at the level of the bile duct for intraduodenal
dosing. Arterial blood samples were withdrawn from a carotid artery catheter
into 3.8% Na–citrate (1/10; v/v) to measure prothrombin time by the standard
procedure described for Dade Thromboplastin-C reagent (Baxter Healthcare
Corp., Miami, FL) (Baxter Healthcare Corp., Miami,FL), which had an
international sensitivity index of 2.0 for the mechanical clot detection
method employed. The dosing solution for these studies was 10% ethanol/
90% PEG300 in a volume of 1 mL/kg for bolus iv or intraduodenal dosing, and at
1 mL/kg/h for sustaining iv infusion.
16. Maron, D. M.; Ames, B. N. Mutat. Res. 1983, 113, 173.
17. Cerep in vitro pharmacology screening package including the following assays:
A1 (h),
(r), CCKA (CCK1) (h), D1 (h), D2S (h), AMPA (r), Kainate (r), NMDA (r), H1
(central) (guinea-pig), H2 (guinea-pig), MAO-A (r), MAO-B (r), (non-
selective) (r), Opiate (non-selective) (r), PCP (r), Rolipram (m), 5-HT (non-
a1 (non-selective) (r), a2 (non-selective) (r), b1 (h), b2 (h), BZD (central)
M
selective) (r), 5-HT1A (h), 5-HT2B (h), sst (non-selective), GR (h), ERa (h), PR (h),
Ca2+ channel (L, DHP site) (r), Ca2+ channel (L, verapamil site) (r), Na+ channel
(site 2) (r), Cl-channel (r), NE transporter (h), DA transporter (r), GABA