Aminobenzisoxazoles with biaryl P4 moieties as potent, selective, and orally bioavailable factor Xa inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.01.010

We have previously reported on a series of aminobenzisoxazoles as potent, selective, and orally bioavailable factor Xa inhibitors, which culminated in the discovery of razaxaban. Herein, we describe another approach to improve factor Xa inhibitory potency

Full text of DOI:10.1016/j.bmcl.2006.01.010

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1795–1798
Aminobenzisoxazoles with biaryl P4 moieties as potent,
selective, and orally bioavailable factor Xa inhibitors
Mimi L. Quan,* Qi Han,* John M. Fevig, Patrick Y. S. Lam, Steve Bai, Robert M. Knabb,
Joseph M. Luettgen, Pancras C. Wong and Ruth R. Wexler
Discovery Chemistry, Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, NJ 08543, USA
Received 2 December 2005; revised 4 January 2006; accepted 5 January 2006
Available online 24 January 2006
Abstract—We have previously reported on a series of aminobenzisoxazoles as potent, selective, and orally bioavailable factor Xa
inhibitors, which culminated in the discovery of razaxaban. Herein, we describe another approach to improve factor Xa inhibitory
potency and pharmacokinetic profile by incorporating basic and water soluble functionalities on the terminal ring of the P4 biaryl
group found in our earlier Xa inhibitors. This approach resulted in a series of potent, selective, and orally bioavailable factor Xa
inhibitors.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Thrombosis remains the major cause of cardiovascular
disorders in the Western countries. The continued im-
pact of thrombotic diseases on morbidity and mortality
has led to extensive efforts to discover and develop new
antithrombotic agents.1 Factor Xa has become a major
focus of pharmaceutical intervention in the past decade
because of its central and unique position in the coagu-
lation cascade.2 Preclinical and clinical studies have
shown that inhibitors of factor Xa are effective in both
venous and arterial thrombosis.3,4
macokinetic profile of 1 by incorporating basic and
water soluble functionalities on the terminal ring. This
approach resulted in a series of potent, selective, and
orally bioavailable factor Xa inhibitors (2) as shown
in Figure 1.
The general synthesis for compounds with structure 2 is
described in Scheme 1. Pyrazole-4-carboxylic acid 35
was converted to the acyl chloride with thionyl chloride
F3C
F
Our efforts to discover and develop orally active inhibi-
tors of factor Xa led to our first clinical candidate
DPC423.5 In an effort to improve trypsin selectivity,
the m-benzylamine P1 moiety in DPC423 was replaced
with 3-aminobenzisoxazole which resulted in compound
1 with >4000-fold selectivity. However, compound 1
demonstrated poor solubility and permeability which
resulted in very low oral bioavailability in dogs. Re-op-
timization of the P4 moiety by replacing the terminal
ring with solubilizing and less lipophilic heterocycles
culminated in the discovery of clinical candidate razax-
aban.6a Razaxaban was studied in phase II clinical tri-
als and was shown to be efficacious in the treatment of
deep vein thrombosis.6b Herein, we describe another
approach to improve the factor Xa potency and phar-
F3C
H
N
F
H
N
SO2NH2
N
SO2NH2
N
N
N
O
O
NH2
NH2
O
N
1
DPC423
Xa Ki = 0.15 nM
dog F% = 57
Xa Ki = 0.09 nM
dog F% = 2
F3C
F
H
N
N
F3C
N
F
N
O
H
R
O
N
N
N
N
N
O
NH2
N
NH2
Razaxaban
Xa Ki = 0.19 nM
dog F% = 84
O
N
2
Keywords: Factor Xa inhibitors; Anticoagulants.
*
Corresponding authors. Tel.: +1 609 818 5301; fax: +1 609 818 3331;
e-mail: mimi.quan@bms.com
Figure 1.
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.010

1796
F3C
M. L. Quan et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1795–1798
F
F3C
NH2
F
CHO
B(OH)2
H
NH2
F
CHO
N
a
N
N
CO2H
CN
N
F
+
N
F
a
O
Br
Br
6
10
NH2
b.
4
CN
F
b
5
F
3
OTMS
CHO
B(OH)2
H2N
c
Br
F
OH
c
H2N
4
F3C
12
6
N
11
CO2H
CN
N
F
F3C
F
F3C
F
H
N
d
NR'R''
H
N
N
CHO
d
N
N
N
F
O
O
3
F
F3C
CN
H
OH
CN
F3C
F
N
N
F
H
OTMS
N
O
N
N
O
8
7
N
F
e, f
O
e
NH2
N
CN
13
F3C
F
14
g
H
NR'R''
N
N
N
O
F
F3C
F3C
F
H
H
N
R
Br
N
N
N
N
O
N
h
NH2
O
N
O
O
N
2b-2h
NH2
NH2
O
N
Scheme 1. Reagents and conditions: (a) SOCl2/rt; (b) 4, Et3N/CH2Cl2,
two steps: 80–90%; (c) Pd(PPh3)4, 2 M Na2CO3, H2O–THF, reflux,
4 h, 80–90%; (d) NHR0R00, NaBH(OAc)3, THF, 70–80%; (e)
CH3CONHOH, K2CO3, DMF/H2O, 80–90%.
15
2i-2v
Scheme 2. Reagents and conditions: (a) Pd(PPh3)4, 2 M Na2CO3,
H2O–THF, reflux, 4 h, 80–90%; (b) NaBH4, 95%; (c) TMSCl,
imidazole, 90%; (d) 1—SOCl2; 2—Et3N/CH2Cl2, 85%; (e) CH3CON-
HOH, K2CO3, DMF/H2O, 80–90%; (f) HCl/EtOH, reflux 3 h, 80%; (g)
PPh3, CBr4, 0 ꢁC to rt, 85%; (h) excess R1R2NH, 40–50%.
and then coupled with 4-bromo-2-fluoroaniline (4) to
give amide 5. Suzuki–Miyaura coupling of 5 with boron-
ic acid 6 using Pd(PPh3)4 as the catalyst produced biaryl
compound 7. Reductive amination with various amines
generated compound 8. Aminobenzisoxazole forma-
tion6a with acetohydroxamic acid and potassium
carbonate afforded the final product 2b–2h.
The SAR for this series is shown in Table 1. Replacing
the aminosulfonamide moiety of 1 with an aminomethyl
group resulted in compound 2a with a Xa Ki of 2 nM,
which represents a 20-fold loss in Xa affinity. However,
alkylation of the amino group in 2a regained Xa potency
to subnanomolar affinity, except for 2r. Compared with
2a, mono- and di-alkylation of the amino group in-
creased Xa affinity by 5- to 20-fold. Cyclic amines also
retained Xa potency. The most potent compounds pyr-
rolidine 2i and piperidine 2o showed a 15- to 20-fold
enhancement in Xa potency. Hydroxymethyl, hydroxyl,
and amino-substitutions on the pyrrolidine ring were all
well tolerated (2j–n). Piperidine is more potent than
morpholine and piperazine. Methylation of the pipera-
zine (2q) did not result in improved Xa potency. An aro-
matic heterocycle such as imidazole (2s) was found to be
threefold less potent compared with alkylamines.
Compounds 2a, 2i–2v were prepared by the sequence
shown in Scheme 2. 4-Bromo-2-fluoroaniline 4 was cou-
pled with boronic acid 6 under Suzuki–Miyaura condi-
tions to yield biaryl aldehyde 10. Aldehyde 10 was
reduced with sodium borohydride to give alcohol 11,
which was then protected with TMSCl to give com-
pound 12. Compound 12 was coupled with acid 3 using
the same conditions described previously to give amide
13. Aminobenzisoxazole formation, followed by de-pro-
tection of the TMS group, produced alcohol 14. Com-
pound 14 was converted to the corresponding bromide
15, which was reacted with pyrrolidine, 2(S)-CH2OH-
pyrrolidine, 3(R)-OH-pyrrolidine, 3(S)-OH-pyrrolidine,
piperidine, morpholine, 4-methylpiperazine, imidazole,
pyridine, N-methylmorpholine, and triethylamine to
yield the final compounds 2i–2l, 2o–2q, and 2s–2v,
respectively, or reacted with 3(R)-Boc-aminopyrrolidine,
3(S)-Boc-aminopyrrolidine, and Boc-protected pipera-
zine, followed by de-protection of the Boc group with
TFA, to give compounds 2m, 2n, and 2r, respectively.
Compound 2a was prepared from reacting bromide 15
with sodium azide, followed by Staudinger reduction.7
Interestingly, quaternary amines (2t–2v) were the most
potent Xa inhibitors in this series with double digit pico-
molar affinity as demonstrated by their Ki values. They
also exhibited the most potent in vitro anticoagulant
activity with an EC2· of 300 nM in the aPTT assay.
However, the quaternary amines are less permeable than
other compounds in this series as exemplified by
compounds 2t and 2v with permeability of 0.35 · 10ꢀ6
and 0.25 · 10ꢀ6 cm/s in the Caco-2 assay, respectively.

M. L. Quan et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1795–1798
1797
Table 1. In vitro activity
F
F3C
H
N
R
N
N
O
NH2
O
N
2
Compound
R
Xa Ki (nM)
Caco-2 · 10ꢀ6 (cm/s)
aPTT EC2· (lM)
2a
2b
2c
2d
2e
2f
–NH2
–NHMe
2.0
nd
6.0
4.5
0.27
0.25
0.34
0.20
0.30
0.21
0.40
0.10
0.20
0.16
0.53
0.52
0.54
0.15
0.55
0.86
1.14
0.66
3.48
nd
–NHCH2-c-Pr
–NH-i-Pr
–NH-c-Bu
19.1
5.5
nd
0.83
nd
10.9
12.6
2.8
–NH-c-Pen
–NMe2
2g
2h
2i
4.20
nd
Aziridinyl
Pyrrolidinyl
10.6
15
3.84
nd
2j
2(S)-CH2OH-Pyrrolidinyl
3(R)-OH-Pyrrolidinyl
3(S)-OH-Pyrrolidinyl
3(R)-NH2-Pyrrolidinyl
3(S)-NH2-Pyrrolidinyl
Piperidinyl
3.9
7.3
2k
2l
3.7
nd
16
2m
2n
2o
2p
2q
2r
2s
4.86
4.64
8.89
nd
12.0
7.2
8.7
Morpholinyl
60.8
6.4
4-Me-Piperazinyl
Piperazinyl
1H-Imidazolyl
2.32
nd
10.0
19.4
nd
N+
Br-
Br-
2t
0.08
0.02
0.35
nd
0.29
Me
N+
O
2u
0.35
2v
Razaxaban
–N+(Me)3Brꢀ
0.05
0.19
0.25
5.56
0.29
6.1
Xa Kis obtained from purified human enzymes and are averaged from multiple determinations (n = 2). See Ref. 6a for details on the Xa, Caco-2, and
aPTT assays.
nd, no data.
The permeability and in vitro anticoagulant activity of
these compounds varied depending on the substituents.
For compounds with similar Ki values, the anticoagulant
activity difference in the aPTT assay is probably due to
the difference in protein binding of the compounds.
(Table 3). Compounds 2g, 2i, and 2k were dosed at
0.4 mg/kg intravenously and 0.2 mg/kg orally. Com-
pared with razaxaban, these compounds have a slightly
lower permeability in the Caco-2 assay and therefore it
is not surprising that they showed lower bioavailability.
The clearance for these compounds is similar to razax-
aban, but these compounds do show a higher volume
of distribution, which resulted in a longer halflife. Over-
all these compounds are very much comparable to
razaxaban in terms of the pharmacokinetic profile in
that they have halflife in the range of 5–10 h and show
good oral bioavailability.
We earlier reported that compounds with the aminoben-
zisoxazole P1 moiety exhibit excellent selectivity relative
to many key serine proteases.6a As anticipated, com-
pounds in the current series also were very selective.
As exemplified in Table 2, compounds 2g, 2i, and 2k
showed >10,000-fold selectivity relative to trypsin,
aPC, factor IXa, factor VIIa, plasmin, tPA, plasma kal-
likrein, urokinase, and chymotrypsin. They are less
selective relative to thrombin, but still show >1000-fold
selectivity. The overall selectivity profile for these com-
pounds is very similar to razaxaban.
These three compounds were also studied in the rabbit
arterio-venous (A-V) shunt thrombosis model.5b Upon
intravenous dosing, compounds 2g, 2i, and 2k inhibited
thrombus formation in a dose-dependent manner with
an IC50 of 920, 180, and 50 nM, respectively (Table 4).
The correlation between in vitro anticoagulant activity
and in vivo antithrombotic activity is not very good.
However, the in vivo antithrombotic activity observed
The pharmacokinetic profile of selected potent
compounds with good permeability in the Caco-2
assay was studied in dogs via a cassette dosing format

1798
M. L. Quan et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1795–1798
Table 2. Human enzyme selectivity profile
Enzyme Ki (nM)
2g
2i
2k
Razaxaban
Xa
0.21
300
0.10
250
0.16
400
0.19
540
Thrombin
Trypsin
aPC
>1600
15,000
>12,000
>15,000
>15,000
>33,000
>2300
>13,000
1500
>2500
13,000
>12,000
>15,000
>15,000
>33,000
>2300
>13,000
1700
>2500
5700
>10,000
19,700
9000
IXa
VIIa
>12,000
>15,000
>27,000
>35,000
nd
>15,000
>15,000
>33,000
>2300
>13,000
8500
Plasmin
tPA
Plasma Kallikrein
Urokinase
Chymotrypsin
nd
nd
All Kis obtained from purified human enzymes and are averaged from multiple determinations (n = 2). See Ref. 6a for more details.
nd, no data.
Table 3. Dog pharmacokinetic profiles
Acknowledgments
Compound
Cl (L/kg/h)
Vdss (L/kg)
t1/2 (h)
F (%)
We thank Frank Barbera, Tracy Bozarth, Joseph Buri-
ak, Jr., Earl Crain, and Carol Watson for technical
assistance.
2g
1.1
1.7
0.76
1.1
9.0
23
6.6
10
38
64
43
84
2i
2k
5.8
5.3
6.5
3.4
Razaxaban
Compounds were dosed as the TFA salts in an N-in-1 format at
0.4 mg/kg iv and 0.2 mg/kg po (n = 2).
References and notes
1. (a) Hirsh, J.; O’Donell, M.; Weitz, J. I. Blood 2005, 105,
453; (b) Golino, P.; Loffredo, F.; Riegler, L.; Renzullo, E.;
Cocchia, R. Curr. Opin. Invest. Drugs 2005, 6, 298; (c)
Quan, M. L.; Smallheer, J. Curr. Opin. Drug Disc. Dev.
2004, 7, 460.
2. (a) Drout, L.; Bal ditSollier, C. Eur. J. Clin. Invest. 2005, 35,
21; (b) Mann, K. G.; Butenas, S.; Brummel, K. Arterioscler.
Thromb. Vasc. Biol. 2003, 23, 17; (c) Leadley, R. J., Jr. Curr.
Table 4. Anticoagulant activity in rabbits
Compound Xa Ki
(rabbit) nM (rabbit)
EC2·, lM IC50 (nM)
aPTT/PT Rabbit
Protein
A-V shunt5b binding8
(rabbit)
2g
2i
0.22
0.46
0.17
5.3/1.5
7.5/0.8
7.1/1.4
3.5/1.9
920
180
50
94.4
82.0
68.0
93.4
Top. Med. Chem. 2001, 1, 151; (d) Hauptmann, J.; Sturzeb-
echer, J. Thromb. Res. 1999, 93, 203.
¨
2k
3. (a) Wong, P. C.; Crain, E. J.; Watson, C. A.; Zaspel, A. M.;
Wright, M. R.; Lam, P. Y. S.; Pinto, D. J.; Wexler, R. R.;
Knabb, R. M. J. Pharmacol. Exp. Ther. 2002, 303, 993; (b)
Wong, P. C.; Pinto, D. J.; Knabb, R. M. Cardiovasc. Drug
Rev. 2002, 20(2), 137.
Razaxaban 0.19
340
4. (a) Rajagopal, V.; Bhatt, D. L. J. Thrombosis Haemostasis
2005, 3, 436; (b) Viles-Gonzalez, J. F.; Gaztanaga, J.; Zafar,
U. M.; Fuster, V.; Badimon, J. J. Am. J. Cardiovasc. Drugs
2004, 4(6), 379.
5. (a) Pinto, D. J. P.; Orwat, M. J.; Wang, S.; Fevig, J. M.;
Quan, M. L.; Amparo, E.; Cacciola, J.; Rossi, K. A.;
Alexander, R. S.; Smallwood, A. M.; Luettgen, J. M.;
Liang, L.; Aungst, B. J.; Wright, M. R.; Knabb, R. M.;
Wong, P. C.; Wexler, R. R.; Lam, P. Y. S. J. Med. Chem.
2001, 44, 566; (b) Wong, P. C.; Quan, M. L.; Crain, E. J.;
Watson, C. A.; Wexler, R. R.; Knabb, R. M. J. Pharmacol.
Exp. Ther. 2000, 292, 351.
6. (a) Quan, M. L.; Lam, P. Y. S.; Han, Q.; Pinto, D. J.; He, M.;
Li, R.; Ellis, C. D.; Clark, C. G.; Teleha, C. A.; Sun, J. H.;
Alexander, R. S.; Bai, S. A.; Luettgen, J. M.; Knabb, R. M.;
Wong, P. C.; Wexler, R. R. J. Med. Chem. 2005, 48, 1729; (b)
Lessen, M. R.; Davidson, B. L.; Gallus, A.; Pineo, G.; Ansell,
J.; Deitchman, D. Blood 2003, 102, 15a, Abstract 41.
7. Staudinger, H.; Meyer, J. Helv. Chim. Acta 1919, 2, 635.
8. Pacific, G. M.; Viani, A. Clin. Pharmacokinet. 1992, 23,
449.
in the rabbit A-V shunt model correlates with the rabbit
protein binding shown in Table 4. With similar Xa affin-
ity, compound 2g has higher protein binding and is less
potent in the rabbit A-V shunt thrombosis model. Com-
pound 2k has 32% free fraction and, therefore, exhibits
the most potent anticoagulant activity in the A-V shunt
model (IC50 of 50 nM). Overall, compound 2k is six
times more potent than razaxaban in the rabbit A-V
shunt thrombosis model.
In summary, optimization of the P4 moiety by incorpo-
rating basic and water solubilizing groups on the termi-
nal ring of the biphenyl group has led to a series of
potent, selective, and orally bioavailable factor Xa
inhibitors. Compounds 2g, 2i, and 2k show comparable
potency, selectivity, and pharmacokinetic profile to
razaxaban. Compound 2k showed a sixfold enhance-
ment in potency in the rabbit A-V shunt thrombosis
model compared to razaxaban.