Design, synthesis, and SAR of monobenzamidines and aminoisoquinolines as factor Xa inhibitors

Article abstract of DOI:10.1016/S0960-894X(02)00234-2

Monoamidine FXa inhibitors 3 were designed and synthesized. SAR studies and molecular modeling led to the design of conformationally constrained diaryl ethers 4 and 5, as well as benzopyrrolidinone 7 as potent FXa inhibitors. The monoamidines show high efficacy in a DVT model, but lack desirable oral bioavailability. The benzopyrrolidinone-based aminoisoquinolines 8 do not show significant improvement in oral bioavailability.

Full text of DOI:10.1016/S0960-894X(02)00234-2

Bioorganic & Medicinal Chemistry Letters 12 (2002) 1657–1661
Design, Synthesis, and SAR of Monobenzamidines and
Aminoisoquinolines as Factor Xa Inhibitors
Penglie Zhang,a,* Jingmei F. Zuckett,a John Woolfrey,a Katherine Tran,b Brian Huang,b
Paul Wong,b Uma Sinha,b Gary Park,b Andrea Reed,b John Malinowski,b
Stan Hollenbach,b Robert M. Scarborougha and Bing-Yan Zhua,*
aDepartment of Medicinal Chemistry, Millennium Pharmaceuticals, Inc., 256 E. Grand Avenue, South San Francisco,
CA 94080, USA
bDepartment of Biology, Millennium Pharmaceuticals, Inc., 256 E. Grand Avenue, South San Francisco, CA 94080, USA
Received 3 November 2001; accepted 29 March 2002
Abstract—Monoamidine FXa inhibitors 3 were designed and synthesized. SAR studies and molecular modeling led to the design of
conformationally constrained diaryl ethers 4 and 5, as well as benzopyrrolidinone 7 as potent FXa inhibitors. The monoamidines
show high efficacy in a DVT model, but lack desirable oral bioavailability. The benzopyrrolidinone-based aminoisoquinolines 8 do
not show significant improvement in oral bioavailability. # 2002 Elsevier Science Ltd. All rights reserved.
Due to the central role of factor Xa (FXa) in both
intrinsic and extrinsic pathways in the blood coagula-
tion cascade, the inhibition of FXa is believed to be an
effective approach for the treatment of thrombotic dis-
orders.1,2 Previously disclosed small molecule FXa
inhibitors are mostly dibasic.3 A P1 amidino group in
the inhibitor interacts with the Asp189 side chain in the
FXa S1 specificity pocket, resulting in a strong ionic
interaction and hydrogen bonding. A P4 group, posi-
tively charged under physiological conditions, binds to
the cation sink in the FXa S4 site. The unfavorable
pharmacokinetic properties associated with the strongly
basic nature of the amidine moieties have led to the active
search for less basic compounds as orally efficacious
anticoagulants.4
the synthesis and structure–activity relationship (SAR)
studies of compound 3. The initial SAR results and
molecular modeling led us to design two types of novel,
conformationally constrained FXa inhibitors, as repre-
sented by the diaryl ether 4a and benzopyrrolidinone 7.
Compound 1 has been reported as a potent FXa inhi-
bitor. The biphenylsulfonamide group interacts with the
S4 pocket of FXa as suggested by molecular modeling
results.5 Compound 2, disclosed by Ajinomoto,6 has an
ethanolamine linker between two benzamidine moieties.
Based on the assumption that the meta-benzamidine
moiety is the S1 binding group, compound 3a was
designed as a FXa inhibitor (Fig. 1). Herein, we report
*Corresponding author. Fax: +1-650-244-9287; e-mail: penglie.
zhang@mpi.com
Figure 1. Design of conformationally constrained FXa inhibitors.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00234-2

1658
P. Zhang et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1657–1661
Chemistry
intermediates were then used for the amidine formation.
The dimethylamine and methylamines 5d, 5k and/or 5l
were found as the major products when 5b and 5i were
treated with 10% Pd/C under H2 atmosphere (1 atm) in
methanol. Debromination of 5h was achieved with cat-
alytic hydrogenation to give compound 5a.
The syntheses of compound 3a and its analogues are
described in Scheme 1. The N-Boc protected amino
alcohols were reacted with 3-cyanophenol under Mitsu-
nobu conditions.7 Removal of the Boc protecting group
was followed by the coupling with 4-(2-tert-butylami-
nosulfonyl)phenyl benzoic acid. Alkylation of the amide
nitrogen was effected with RX (X=OTs, Br, Cl, or I).
Pinner conditions8 followed by ammonolysis gave the
desired amidines with the simultaneous removal of the
tert-butyl group.
The syntheses of aminoisoquinolines 6 are described in
Scheme 4. A unique three step sequence was used for
the conversion of isoquinoline to aminoisoquinoline.11
The tert-butyl group was removed with trifluoroacetic
acid at reflux.
The diaryl ethers 4 were synthesized according to
Scheme 2. Substituted 1-fluoro-2-nitrobenzenes were
reacted with 3-cyanophenol under basic conditions to
afford the ether intermediate. The nitro group was
reduced with tin(II) chloride hydrate. The coupling of
the resulting amines with 4-(2-tert-butylaminosulfonyl)-
phenylbenzoic acid followed by the amidine formation
gave the desired compound 4.
Compound 7 was synthesized according to Scheme 5.
The amine intermediate obtained from Scheme 1
(R1=H) was reacted with the benzylic bromide to give
the lactam in a single step. Scheme 6 describes the
syntheses of the aminoisoquinoline compounds 8.
Results and Discussion
The syntheses of compound 5 is described in Scheme 3.
The Weinreb conditions9 were employed for the amide
bond formation. The conversion of the amino group to
N-acetyl or sulfonamide was performed under standard
conditions. The Sandmeyer reaction (tBuONO, CuBr or
CuCl, CH3CN, reflux, 1 h) was employed for the halo-
genation of the central phenyl ring.10 The cyano
Compound 3a is a potent FXa inhibitor (IC50=7.5 nM,
Table 1). It is also selective for FXa over other serine
proteases. It has IC50 values of 321 mM for thrombin,
1.2 mM for trypsin, 109 mM for tissue plasminogen acti-
vator (tPA), 3.7 mM for activated protein C (APC),
13 mM for plasmin, and 2.1 mM for kallikrein. Substitu-
tion on the ethanolamine template was carried out to
Scheme 3. (a) TMSCHN2, MeOH; (b) 3-cyanophenol, K2CO3, DMF;
(c) 4-(2-tert-butylaminosulfonyl)biphenylamine, AlMe3, CH2Cl2; (d)
(1) HCl(g), MeOH; (2) NH4OAc, MeOH.
Scheme 1. (a) BocNHCHR1CH2OH, PPh3, DEAD, THF; (b) TFA,
CH2Cl2; (c) 4-carboxyphenyl boronic acid, PdCl2(PPh3)2, K2CO3,
dioxane, H2O; (d) BOP reagent, Et3N, DMF; (e) RX (X=Br, Cl, I,
OTs), Cs2CO3, DMF; (f) (1) HCl(g), MeOH; (2) NH4OAc, MeOH.
Scheme 4. (a) 7-hydroxy-isoquinoline, K2CO3, DMF; (b) SnCl2–
2H2O, EtOAc; (c) 4-(2-tert-butylaminosulfonyl)phenylbenzoic acid,
BOP reagent, Et3N, DMF; (d) (1) MCPBA, acetone; (2) TsCl, –Py; (3)
ethanolamine; (e) TFA.
Scheme 2. (a) 3-cyanophenol, K2CO3, DMF; (b) SnCl2–2H2O,
EtOAc; (c) 4-(2-tert-butylaminosulfonyl)phenylbenzoic acid, BOP
reagent, Et3N, DMF; (d) (1) HCl(g), MeOH; (2) NH4OAc, MeOH.

P. Zhang et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1657–1661
1659
Scheme 5. (a) 2-(tert-butylaminosulfonyl)phenyl boronic acid, TEA,
PdCl2(dppf), Dioxane; (b) NBS, CCl4; (c) NH2(CH2)2O–(3-CN)Ph,
TEA, benzene; (d) (1) HCl(g), MeOH; (2) NH4OAc, MeOH.
Figure 2. Compound 4a docked in the FXa active site. The Connolly
surface indicates more concentrated red color in areas of increased par-
tial negative charge, and blue color in areas of partial positive charge.
Table 2. In vitro potency of 2-aminophenol-based FXa inhibitors
Scheme 6. (a) BocNHCHR1CH2OH, PPh3, DEAD, THF; (b) (1)
MCPBA, acetone; (2) TsCl, Py; (3) ethanolamine; (c) TFA, CH2Cl2;
(d) methyl 4-(2-tert-butylaminosulfonyl)phenyl-2-bromomethylbenzo-
ate, TEA, benzene; (e) TFA.
Compd
R2
R3
IC50 (nM)
4a
4b
4c
4d
4e
4f
4g
4h
4i
H
F
Br
CF3
SO2Me
CO2Me
CO2H
H
H
H
H
H
H
H
H
H
H
3.6
4.5
1.6
2.5
1.8
2.5
49
9.9
1.6
2.1
1.6
3.9
Table 1. In vitro potency of ethanolamine-based FXa inhibitors
H
H
CH3
Cl
Br
OCH3
OH
4j
4k
4l
Compd
R
R1
IC50 (nM)
Molecular modeling indicates an interaction between
the benzamidine moiety of compound 3a and the car-
boxylate side chain of Asp189. The biphenyl portion of
compound 3a occupies the FXa S4 pocket. The N1–C2
and C3–O4 bonds of the central ethanolamine linker are
roughly eclipsed in one of the favorable binding con-
formations. This conformation can be fixed by incor-
porating the C2 and C3 atoms into an aromatic ring
(Fig. 1, path a). The diaryl ether compound 4a with a
2-aminophenol as the central template is projected to fit
well in the FXa active site (Fig. 2).
3a
3b
3c
3d
3e
3f
3g
3h
3i
H
H
H
H
Me
CH2Ph
H
Me
Ph
Bn
H
H
H
H
H
7.5
16
3.4
5.6
24
14
11
7.8
13
CH2-(3-pyridinyl)
CH2-(3-furyl)
CH2CO2Me
CH2CO2H
3j
H
34
The high potency of compound 4a (IC50=3.6 nM)
against FXa validated our design of diaryl ethers as
novel FXa inhibitors (Table 2). Compound 4a offers
diverse opportunities for exploring SARs by substitu-
tion in the template region. However, the potencies of
its analogues 4b–4f and 4h–4l are very close to that of
compound 4a. This indicates that the substitution group
at the central ring does not play a critical role in the
binding of compounds 4 with FXa. The lone exception
is the carboxylic acid compound 4g, which loses an
order of magnitude of potency.
optimize the interaction between the inhibitor and FXa.
Compounds 3c and 3d containing a bulkier phenyl and
benzyl, respectively, are more potent than compound 3b
(racemic) which only bears a methyl group. Compounds
3e–3j are alkylated at the amide nitrogen and are mar-
ginally less potent than compound 3a. These results
clearly show that the alkylation does not jeopardize the
anti-FXa activity. As an interesting note, the carboxylic
acid compound 3j is nearly 3 times less potent than the
corresponding methyl ester 3i.

1660
P. Zhang et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1657–1661
Table 5. In vitro potency of aminoisoquinolines as FXa inhibitors
Table 3. In vitro potency of 2-OH-benzoic acid-based FXa inhibitors
R2
R3
IC50 (nM)
Compd
R1
IC50 (nM)
Compd
8a
8b
8c
8d
H
Me
Ph
CH2Ph
22
25
623
226
5a
5b
5c
5d
5e
5f
5g
5h
5i
5j
5k
5l
5m
5n
5o
5p
H
NO2
NH2
N(CH3)2
NHAc
NHSO2Me
H
H
H
H
H
H
H
H
7.3
1.9
12
10
7.7
5.4
7.7
4.4
4.8
19
12
16
61
31
Cl
Br
H
H
H
H
H
H
H
H
The antithrombotic activity of selected compounds in
these series was evaluated. Compound 4a inhibited
thrombosis by 41%, which is close to the maximum
inhibition (ꢀ50%) in our modified rabbit deep vein
thrombosis model,12 at 0.8 mM plasma concentration
after an intravenous dose of 2 mg/kg. Selected com-
pounds were also evaluated for pharmacokinetic prop-
erties in Sprague–Dawley rats. For example, compound
4a is undetectable after oral administration in rats at a
dose of 6 mg/kg in a vehicle of 25% PEG-300. The
absolute oral bioavailability of compound 4g and
compound 5i in rats is only 5.5% and 0.3%, respectively.
NO2
NH2
NHCH3
N(CH3)2
NHAc
NHSO2Me
Cl
1.5
1.2
Br
Table 4. In vitro potency of aminoisoquinolines as FXa inhibitors
The poor pharmacokinetic profile of compound 4a and
its analogues is not unpredicted and is probably asso-
ciated with the strongly basic amidino group.3 Less
basic derivatives of benzamidines including 2-aminoiso-
quinolines have been reported as thrombin and FXa
inhibitors.13,14 To explore the effect of an aminoisoquin-
oline replacement in the diaryl ether series, compounds
6a–6d were synthesized. Unfortunately, they are poor
FXa inhibitors (Table 4).
Compd
R2
IC50 (nM)
6a
6b
6c
6d
H
F
CF3
SO2Me
646
1320
1810
6060
Previously, we noted that the alkylation of the amide
nitrogen in compound 3a gave compounds that retained
their potency. This encouraged us to design another
type of FXa inhibitor with conformational constraints
by tethering the amide nitrogen to the ortho position of
the proximal phenyl ring of the P4 moiety (Fig. 1, path
b). The prototypical benzopyrrolidinone compound 7 is
potent against FXa (IC50=3.6 nM) and it is two times
more active than the uncyclized compound 4a.
The reversal of the amide bond in compound 4a was
then investigated to address whether the linkage
between the central phenyl ring and P4 moiety is
tolerant to changes. To our delight, compound 5a
retains the anti-FXa potency (Table 3). Similar to what
was observed in the series of compound 4, the impact of
central ring substitution on potency was negligible, as
demonstrated by compound 5b–5p.
The P1 aminoisoquinoline analogues of compound 7
were then pursued to obtain FXa inhibitors with
improved oral bioavailability. Compounds 8c and 8d
with bulkier R1 groups are >10 times less active than
compounds 8a and 8b (Table 5). The different potencies
of compounds 8a and 6a are speculated to arise from
their distinctive binding conformations with FXa.
Compound 8a was tested for pharmacokinetic property
profiling. At a dose of 6 mg/kg in 25% PEG-300, it has
an oral bioavailability of less than 5% in rats.
The binding conformation of compound 4a (Fig. 2)
indicates that R2 is mostly solvent exposed, which may
explain the minimal effect of R2 substitution on
potency. The R3 group points to the entrance wall of the
FXa S1 pocket. What leads to the flat SARs with R3
variations is unknown.
The diaryl ethers shown in Tables 2 and 3 have very
good enzyme selectivity. For example, compound 4a has
IC50 values of >11 mM for thrombin, 0.97 mM for tryp-
sin, 18 mM for tPA, >181 mM for activated protein C
(APC), 43 mM for plasmin, and 0.25 mM for kallikrein.
In summary, monoamidine FXa inhibitors with an
ethanolamine template were designed and synthesized.
SAR studies and molecular modeling led us to design

P. Zhang et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1657–1661
1661
two series of novel conformationally constrained FXa
inhibitors. The diaryl ethers, as represented by
compound 4a, are very potent and highly efficacious in
a rabbit deep vein thrombosis model. The benzopyrro-
lidinone compound 7 is also highly potent against FXa.
The poor pharmacokinetic properties of the benzami-
dines, especially their low oral bioavailability, led us to
synthesize the less basic aminoisoquinolines. Although
the diaryl ether-based aminoisoquinolines are not
potent FXa inhibitors, the benzopyrrolidinones have
decent potency. However, their insufficient oral bioa-
vailability does not warrant further development. We
have utilized these SAR results to subsequently design
potent, non-benzamidine FXa inhibitors with excellent
pharmacokinetic properties. These details will be reported
in due course.
Knabb, R. M.; Lam, G.; Wright, M. R.; Wong, P. C.; Wexler,
R. R. J. Med. Chem. 1999, 42, 2760.
6. Takayanagi, M.; Sagi, K.; Nakagawa, T.; Yamanashi, M.;
Karahara, T.; Takehana, S.; Fukuda, Y.; Takahashi, M.;
Shoji, M. WO 09831661, 1998.
7. For a review: Hughes, D. L. The Mitsunobu Reaction. Org.
React. 1992, 42, 335. For procedure: Kim, H.-O.; Kahn, M.
Synlett 1999, 8, 1239.
8. Pinner, A. In Die Imidoaether und ihre Derivate; Oppen-
heim: Berlin, 1892; p 1.
9. Levin, J. I.; Turos, E.; Weinreb, S. M. Synth. Commun.
1982, 12, 989.
10. Boger, D. L.; Kim, S. H.; Miyazaki, S.; Strittmatter, H.;
Weng, J.-H.; Mori, Y.; Rogel, O.; Castle, S. L.; McAtee, J. J.
J. Am. Chem. Soc. 2000, 122, 7416.
11. Ueda, I.; Shiokawa, Y.; Manabe, T. U.S. patent
4,673,676, 1987.
12. Hollenbach, S.; Sinha, U.; Lin, P. H.; Needham, K.; Frey,
L.; Hancock, T. E.; Wong, A.; Wolf, D. L. Thromb. Haemost.
1994, 71, 357.
References and Notes
13. (a) Rewinkel, J. B. M.; Lucas, H.; van Galen, P. J. M.;
Noach, A. B. J.; van Dinther, T. G.; Rood, A. M. M.; Jenne-
boer, A. J. S. M.; van Boeckel, C. A. A. Bioorg. Med. Chem.
Lett. 1999, 9, 685. (b) Rewinkel, J. B. M.; Lucas, H.; Smit,
M. J.; Noach, A. B. J.; van Dinther, T. G.; Rood, A. M. M.;
Jenneboer, A. J. S. M.; van Boeckel, C. A. A. Bioorg. Med.
Chem. Lett. 1999, 9, 2837.
1. Davie, E. W.; Fujikawa, K.; Kisiel, W. Biochemistry 1991,
30, 10363.
2. Vlasuk, G. P.; Ramjit, D.; Fujita, T.; Dunwiddie, C. T.;
Nutt, E. M.; Smith, D. E.; Shebuski, R. J. Thromb. Haemost.
1991, 65, 257.
3. Scarborough, R. M. J. Enz. Inhib. 1998, 14, 15.
4. (a) Zhu, B.-Y.; Scarborough, R. M. Curr. Opin. Cardiov.
Pulmon. Ren. Investig. Drugs 1999, 1, 63. (b) Zhu, B.-Y.;
Scarborough, R. M. Annu. Rep. Med. Chem. 2000, 35, 83.
5. Quan, M. L.; Ellis, C. D.; Liauw, A. Y.; Alexander, R. S.;
14. Choi-Sledeski, Y. M.; Becker, M. R.; Green, D. M.;
Davis, R.; Ewing, W. R.; Mason, H. J.; Ly, C.; Spada, A.;
Liang, G.; Cheney, D.; Barton, J.; Chu, V.; Brown, K.;
Colussi, D.; Bentley, R.; Leadley, R. J.; Dunwiddie, C.; Pauls,
H. W. Bioorg. Med. Chem. Lett. 1999, 9, 2539.