Design, synthesis, and biological evaluation of peptidomimetic inhibitors of factor XIa as novel anticoagulants

Article abstract of DOI:10.1021/jm060978s

Human coagulation factor XIa (FXIa), a serine protease activated by site-specific cleavage of factor XI by thrombin, FXIIa, or autoactivation, is a critical enzyme in the amplification phase of the coagulation cascade. To investigate the potential of FXIa

Full text of DOI:10.1021/jm060978s

J. Med. Chem. 2006, 49, 7781-7791
7781
Design, Synthesis, and Biological Evaluation of Peptidomimetic Inhibitors of Factor XIa as
Novel Anticoagulants
Jian Lin,* Hongfeng Deng, Lei Jin, Pramod Pandey, Jesse Quinn, Susan Cantin, Michael J. Rynkiewicz, Joan C. Gorga,
Frank Bibbins, Cassandra A. Celatka, Pamela Nagafuji, Thomas D. Bannister, Harold V. Meyers, Robert E. Babine,
Neil J. Hayward, David Weaver, Howard Benjamin, Frans Stassen, Sherin S. Abdel-Meguid, and James E. Strickler*
Daiichi Asubio Medical Research Laboratories LLC (DAIAMED), One Kendall Square Building 700, Cambridge, Massachusetts 02139
ReceiVed August 11, 2006
Human coagulation factor XIa (FXIa), a serine protease activated by site-specific cleavage of factor XI by
thrombin, FXIIa, or autoactivation, is a critical enzyme in the amplification phase of the coagulation cascade.
To investigate the potential of FXIa inhibitors as safe anticoagulants, a series of potent, selective
peptidomimetic inhibitors of FXIa were designed and synthesized. Some of these inhibitors showed low
nanomolar FXIa inhibitory activity with >1000-fold FXa selectivity and >100-fold thrombin selectivity.
The X-ray structure of one of these inhibitors, 36, demonstrates its unique binding interactions with FXIa.
Compound 32 caused a doubling of the activated partial thromboplastin time in human plasma at 2.4 µM
and was efficacious in a rat model of venous thrombosis. These data suggest that factor XIa plays a significant
role in venous thrombosis and may be a suitable target for the development of antithrombotic therapy.
Introduction
new trend in the development of antithrombotic agents is to
explore inhibitors targeting the factor VIIa/tissue factor com-
plex,8,9 factor IXa,10 factor XIa,11,12 carboxypeptidase U (throm-
bin activatable fibrinolysis inhibitor, TAFIa),13 or agents that
enhance activated protein C that in turn inactivates factor Va
and VIIIa.14
A well-balanced and highly regulated hemostatic process is
essential to normal physiology. The overproduction of thrombin
may lead to vascular occlusion as a consequence of many
cardiovascular diseases such as acute myocardial infarction
(AMI), stroke, pulmonary embolism (PE), and deep vein
thrombosis (DVT). The inappropriate down-regulation of throm-
bin generation may also cause hemorrhagic problems.1,2 Thus
far, none of the available therapies for the prevention and
treatment of thromboembolic disorders are ideal. The most
commonly used anticoagulants, heparin and warfarin, require
very careful and costly monitoring to ensure safe therapeutic
drug levels for the duration of treatment because of the high
risk of bleeding. Warfarin and other coumarins, the only orally
administered anticoagulants, often exhibit drug-drug and food-
drug interactions due to their erratic plasma concentration and
anticoagulant effects.3 Heparin also has variable effects due to
its nonspecific binding with multiple coagulation factors in the
system.4 In addition, a small population of patients can develop
an autoimmune syndrome known as heparin-induced thromb-
ocytopenia (HIT).5 Clearly, there is a strong medical need for
a well-tolerated, efficacious anticoagulant with a wide thera-
peutic window. To address this need, a significant focus within
the pharmaceutical industry for decades has been the develop-
ment of potent and selective inhibitors of the coagulation
enzymes to replace heparin and warfarin. Among the coagulation
enzymes, thrombin (factor IIa) and factor Xa (FXa) have
received the most attention due to their multiple functions and
key roles in the final common pathway in the blood coagulation
cascade. Some direct thrombin inhibitors and FXa inhibitors
have been evaluated in phase III clinical trials or have been
licensed in North America or Europe for limited indications.
However, all of these new anticoagulants are still associated
with bleeding problems and have no significant improvement
in efficacy-to-safety index compared with heparin or warfarin
in many studies (reviewed in refs 6 and 7). More recently, a
The roles of factor XI in the intrinsic pathway of coagulation
and its thrombin-mediated feedback activation were not fully
understood until the early 1990s.15 Thrombin formation can
occur via activation of the extrinsic or intrinsic pathway of
coagulation. In the extrinsic pathway, thrombin generation is
initiated by exposure of blood to the tissue factor/activated factor
VII (FVIIa) complex. This complex activates clotting factors
X and IX (proteolysis rate of FX . FIX), leading to the
generation of thrombin and the formation of fibrin.16 This
pathway is usually rapidly shut down by tissue factor pathway
inhibitor (TFPI) and anti-thrombin III. To maintain normal
hemostasis, thrombin-mediated activation of FXI initiates an
amplification phase via the intrinsic pathway. Activated factor
XI (FXIa) continues to activate factor IX and then factor X,
resulting in the formation of additional thrombin, which takes
place inside the fibrin clot. This additional thrombin may be
required for the activation of thrombin activatable fibrinolysis
inhibitor (TAFI) to protect fibrin clots against lysis.17 Therefore,
the role of FXI in hemostasis can be considered as a combination
of procoagulant and antifibrinolytic actions, which is critical to
the consolidation of blood coagulation when tissue factor is not
available.18
Hypothetically, a specific inhibition of FXIa, which is
involved only in the amplification phase via the intrinsic
pathway, represents an attractive strategy for novel antithrom-
botic agents. Specific inhibitors of FXIa might inhibit thrombosis
without completely interrupting the function of normal hemo-
stasis and, thus, might prevent or minimize the risk of hemostasis
complication. Deficiencies in factors VIII and IX can cause
hemophilia A and B, respectively, whereas genetic FXI defi-
ciency results in only a mild-to-moderate bleeding disorder.19
High levels of FXI have been demonstrated as risk factors for
venous thrombosis20 and acute myocardial infarction.21 Recently,
our hypothesis has been supported by a ferric chloride (FeCl3)
* Authors to whom correspondence should be addressed [J.L., e-mail
jlin@epixpharma.com, telephone (781) 761-7639, fax (781) 372-3267;
J.E.S., e-mail jimstrickler@comcast.net].
10.1021/jm060978s CCC: $33.50 © 2006 American Chemical Society
Published on Web 12/21/2006

7782 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
Chart 1. Factor XIa Inhibitor Design
Lin et al.
Chart 2. Natural Cleavage Sites on FIX Performed by FXIa
Scheme 1.a Method A
induced arterial injury mouse model,22 where both FXI- and
FIX-deficient mice were fully or partially protected from
occlusion depending on the concentration of FeCl3. However,
FXI-deficient and wild-type mice were shown to have similar
bleeding time, whereas FIX deficiency was associated with
severely prolonged bleeding time. Furthermore, anti-FXI anti-
bodies were shown to reduce both clot size and bleeding time
in a baboon model of thrombosis23 and to enhance fibrinolysis
in a rabbit model.17b Interestingly, therapy using recombinant
VIIa on a patient with a FXI/FXIa inhibitor has been found to
be effective and safe, ensuring adequate hemostasis without
thrombotic complication.24
a Reagents and conditions: (a) 20% piperidine/DMF; (b) EDCI, HOBt,
DIEA, (L)-(Fmoc)NH-Xaa-OH or (D)-(Fmoc)NH-Xaa-OH; (c) pyridine or
triethylamine, RCOCl, RSO2Cl, RNCO; (d) 50% TFA/CH2Cl2.
Chemistry
A combination of solid- and solution-phase synthesis was
utilized to prepare a series of peptidomimetic inhibitors as shown
in Schemes 1-4. The dipeptide cores, designed as P2-P3
components, were readily constructed on Wang resin by
employing a 9-fluorenylmethoxycarbonyl (Fmoc) coupling
strategy. The elaboration of the N terminus with various
acylating reagents such as acid chlorides, sulfonyl chlorides and
isocyanates was carried out in the final step on solid support,
followed by a standard TFA cleavage protocol to yield a variety
of dipeptido precursors, 5a-5c (Scheme 1). As an alternative
to the isocyanate approach (method A), N-terminal ureas were
also synthesized by reaction of the amine terminus of the resin
with p-nitrophenylchloroformate (5-10 equiv) in the presence
of diisopropylethylamine (DIEA), followed by the addition of
excess (5-10 equiv) primary or secondary amine. In one
example, the resin-supported p-nitrophenylcarbonate dipeptide
was treated with (R)-1-(4-bromophenyl)ethylamine to provide
resin-supported compound 4d. Optically pure isomer 5d was
readily obtained after cleavage with 50% TFA in dichlo-
romethane (method B shown in Scheme 2).
The R-ketothiazole derivative, 6, was prepared in three steps
from commercial Boc-NH-Arg(Mtr)-OH as described by Ber-
ryman et al.29 (Scheme 3). Coupling of acids 5a-d with
R-ketothiazole, 6, using 1-hydroxybenzotriazole (HOBt)/1-(3-
dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDCI)
chemistry and subsequent deprotection of the 2,3,6-trimethyl-
4-methoxybenzenesulfonyl (Mtr) group under strong acidic
conditions (95% TFA/2.5% thioanisole/2.5% water) yielded
peptidomimetics 1a-d (Scheme 4).
Inhibitor Design
Due to the lack of literature precedents of FXIa inhibitors,
the starting point of our search for potent and selective inhibitors
of FXIa was initially based on two known thrombin inhibitors,
leupeptin (L-Leu-Leu-ArgCHO)25 and PPACK (D-Phe-Pro-
ArgCH2Cl).26 We used tripeptides as templates to interact with
the S1, S2, and S3 pockets of the enzyme and further elaborated
the C terminus and N terminus of the tripeptides to gain more
specific interactions with our target enzyme. We first chose to
synthesize a number of arginine derivatives, 1, as shown in Chart
1, with activated electrophilic carbonyl groups such as an
R-ketothiazole that has a well-characterized covalent interaction
with Ser195,27 in which the Arg-R-ketothiazole component
served as the P1-P1′ template. P2 and P3 components were
initially selected from hydrophobic residues, including Phe, Val
(as replacement of Thr), Leu, Ile, Pro, Ala, and Tyr, as the
corresponding non-prime sequences of the two scissile FIX
substrate sites contain hydrophobic P2 and P3 subunits (Chart
2). The X-ray crystal structure of a FXIa-ecotin complex28a
also provided us with a comprehensive understanding of the
FXIa active site to guide our inhibitor design.

Peptidomimetic Inhibitors of Factor XIa
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7783
Scheme 2. Method Ba
a Reagents and conditions: (a) 20% piperidine/DMF; (b) 4-nitrophenyl chloroformate, DIEA, THF/DCM (1:1); (c) triethylamine, (R)-1-(4-bromophenyl)
ethylamine (5 equiv), DMF; (d) 50% TFA/CH2Cl2.
Scheme 3a
a Reagents and conditions: (a) Me(MeO)NH‚HCl, EDCI, HOBt, DIEA, THF; (b) thiazole, n-BuLi, TMEDA, -78 to -40 °C, then -78 °C; (c) 25%
TFA/CH2Cl2.
Scheme 4a
a Reagents and conditions: (a) EDCI, HOBt, DIEA, THF; (b) TFA/thioanisole/water (95:2.5:2.5), 1-3 h.
Results and Discussion
Another lead compound, 15, from the Leu-Val-Arg-R-
ketothiazole sublibrary, also displayed potent FXIa activity with
an IC50 of 63 nM (Table 2). Starting from lead compound 15,
we explored the SAR of P2 groups by fixing the 3,4-
dichlorobenzyl urea-linked R4 component. As shown in Table
2, FXIa appears to prefer a relatively small, branched hydro-
phobic group such as Val, Leu, or Ileu in its S2 pocket.
Compounds 15 and 16 showed not only potent inhibitory
properties against FXIa (IC50 ) 63 and 76 nM, respectively)
but also decent selectivity over FXa (for 16, >1000-fold) and
thrombin (for 16, 34-fold). Highly branched, larger R2 groups
such as cyclohexyl, Phe, and tBu (18, 19, and 21, respectively)
or a very small group such as Ala (20) displayed weaker FXIa
affinity. However, the larger Tyr residue at the P2 position (17)
showed unexpectedly good binding affinity for FXIa with an
IC50 of 200 nM. The p-hydroxyl group in the Tyr residue of 17
The lead compound, 7, emerged from the early exploration
of the P4 component using Phe-Val-Arg-R-ketothiazole as P3-
P2-P1-P1′ template (Table 1). Incorporation of various isocy-
anates, acid chlorides, and sulfonyl chlorides on the N terminus
generated urea-linked, amide-linked, and sulfonamide-linked R4
substitutions, respectively. Compounds with urea-linked R4
substitutions such as compounds 7-10 showed superior potency
against FXIa, compared with those compounds with amide- or
sulfonamide-linked R4 substitutions (12-14). Compound 7, with
a 3,4-dichlorobenzyl urea link, exhibits potent FXIa activity with
an IC50 of 90 nM and good selectivity over thrombin (72-fold)
and FXa (153-fold). The potency loss of compound 11, which
has a rigid phenyl urea R4 component, indicates that a flexible
spacer group such as -CH2- or -CH2CH2- in urea-linked
R4 elements is needed for FXIa binding affinity.

7784 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
Table 1. Early Lead Exploration
Lin et al.
a nd, not determined.
may interact with Glu 98 in the small S2 loop of FXIa (see the
detailed structure information of FXIa28).
10-20-fold enhanced compared with the corresponding (S)-
diastereomeric isomers 29 and 31, respectively (Table 4). The
selectivity of 28 and 30 over FXa and thrombin is also 2-17-
fold better than that of the (S)-isomers 29 and 31.
Further efforts to explore the SAR of the R3 elements are
shown in Table 5. Either hydrophobic or hydrophilic residues
at the R3 position demonstrated excellent potency against FXIa.
Hydrophobic residues had better selectivity over FXa and
thrombin such as 22 and 28. Compounds with less hydrophobic
R3 residues such as 32 and 37 showed less selectivity over FXa
(∼200-fold); however, residues with the most hydrophilicity at
R3, for example, 33, significantly reduced the selectivity against
both FXa and thrombin (Table 5).
Taking advantage of solid- and solution-phase synthesis
technologies, we were able to quickly optimize R3 and R4
elements in parallel (Tables 3-5). As shown in Table 3, the
racemic 1-(4-bromophenyl)ethyl urea at the R4 position (22)
gave the best FXIa potency (IC50 ) 10 nM) and good FXa
(1500-fold) and thrombin (140-fold) selectivity in this series.
The dichloro substitution at the benzyl phenyl ring was found
to be favored by FXIa when the activities of compounds 15
and 23 were compared to those of 24 and 25. Replacement of
the urea linkage of compound 25 with carbamate analogue 26
resulted in a 2-3-fold loss in potency. Truncation of the R4
element, such as in compound 27, reduced the potency dramati-
cally. The stereochemistry of the R-methyl benzyl substituent
of 22 was also studied. The inhibitory activity against FXIa of
the compounds with an (R)-configuration such as 28 and 30 is
Some of the above potent compounds were selected and tested
in in vitro clotting assays. The concentration needed to cause a
doubling of the activated partial thromboplastin time (aPTT)
and the prothrombin time (PT) in human plasma in vitro was

Peptidomimetic Inhibitors of Factor XIa
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7785
Table 2. SAR Exploration of R2 Elements
determined for each compound (see Table 6). Most compounds
required only submicro- to low micromolar concentrations in
coagulation (2× aPTT) assays and 10-50-fold higher concen-
trations in the 2× PT assay. The results indicated that potency
in the aPTT versus the PT is consistent with the high level of
specificity for FXIa over thrombin or FXa as determined by
the biochemical assay. Compound 28 resulted in a significant
loss of potency in both coagulation assays, likely due to poor
solubility rendered by the lipophilic cyclohexyl group at R3.
substantially more potent and selective than the (S)-isomers
(Table 4). In addition, there are hydrogen-bonding interactions
between the urea CdO group, the Val CdO group (P2), and
Nú of Lys192 (see Figure 1). These key interactions are most
likely responsible for the observed improvement in potency and
selectivity. Compared with other trypsin-like coagulation en-
zymes, only FXIa and FVIIa have a lysine at position 192.28
As shown in the X-ray crystal structure of 36, Lys192 may be
a key selectivity element for FXIa.
The IC50 values of 32 versus selected serine proteases are
shown in Table 7. Compound 32 showed good selectivity versus
thrombin (340-fold), FXa (267-fold), FVIIa (>6000-fold), and
activated protein C (APC, >3000-fold), but poor selectivity over
plasma kallikrein and trypsin (both only ∼2-fold). In addition
to potency and selectivity, we evaluated the pharmacokinetic
profile of 32 via intravenous administration to rats. Compound
32 displayed relatively high clearance (32 mL/kg/min), a short
half-life (t1/2 of 45 min), and a low volume of distribution (Vdss
∼ 236 mL/kg).
The X-ray crystal structure of 36 in complex with FXIa
exemplifies how this series of potent inhibitors interact with
FXIa (Figure 1). A detailed scheme of the key binding
interactions of 36 with FXIa is described in Chart 3. As
expected, the arginine of 36 forms a bidentate interaction with
Asp189 in the S1 pocket similar to the FXIa-ecotinM84R
structure.28a The C-terminal thiazole binds in the S1′ pocket:
the nitrogen of thiazole forms a hydrogen bond with His57;
the ketocarbon is covalently bound to the hydroxyl group of
Ser195. The valine residue of 36 fits in the S2 pocket. The P3
component, a citrulline side chain, unexpectedly binds in the
S4 pocket, instead of the S3 site, but does not have any direct
polar interactions with the S4 pocket of FXIa. Interestingly, the
(R)-1-(4-bromophenyl)ethyl urea substituent interacts with
Lys192 of FXIa by hydrophobic packing between the phenyl
group and the lysine side chain. The (R)-configuration of this
substituent enables the right orientation for the phenyl group to
make such contact. This explains why the (R)-isomers are
To test the importance of factor XI in clot formation in vivo,
we used an established rat model of venous thrombosis.30,31 In
this model, a hypotonic insult is used to induce clot formation
in the inferior vena cava. To test for potential species specificity,
we investigated the ability of 32 to affect the aPTT and PT in
rat plasma (data in human plasma, see Table 6). Compound 32
was ∼10-fold less effective at prolonging rat aPTT than human
aPTT. In rat plasma, 25 µM 32 was required to yield a doubling

7786 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
Table 3. Optimization of R4 Elements
Lin et al.
a nd, not determined.
of the aPTT, whereas >120 µM was required to double the
PT. This was most likely a direct effect on factor XI as shown
by the fact that the IC50 of this compound for recombinant rat
FXI catalytic domain was also 10-fold higher than the IC50 for
the human enzyme.32 Despite the lower potency against the rat
enzyme, continuous intravenous administration of 32 suppressed
thrombus formation in the inferior vena cava induced by rapid
infusion of hypotonic saline followed by partial stenosis in the
rat (Figure 2). Administration of 32 commenced 5 min prior to
thrombus induction, with determination of thrombus mass at
10 min postchallenge. Efficacy was observed at a total dose
exposure of 0.25 mg/kg. The extent of the reduction is
comparable to that obtained with heparin given as an intravenous
bolus of 50 units/kg followed by continuous infusion at 25 units/
kg/h. Similar results were obtained with 37 (data not shown).
Compound 37 was also tested in a rat mesenteric arteriole
bleeding model.31,35,36 When administered at up to 4 times the
efficacious dose (1 mg/kg), 37 did not alter the bleeding time,
whereas heparin, at a dose comparable to the clinically relevant
dose, significantly increased bleeding (Figure 3).
showed low nanomolar FXIa inhibitory activity with >1000-
fold FXa selectivity and >100-fold thrombin selectivity. The
X-ray structure of the inhibitor, 36, demonstrates its unique
binding with FXIa. In vivo efficacy of 32 was demonstrated in
a rat model of venous thrombosis. The results presented in this
paper clearly suggest that FXIa plays a significant role in clot
formation and/or maintenance and that it is a suitable target for
the development of novel antithrombotic agents.
Experimental Section
Chemistry. General Methods. The preloaded amino acid
Wang resin [100-200 mesh, 1% divinylbenzene (DVB) cross-
linking, 0.63 mmol/g] and Wang resin (100-200 mesh, 1%
DVB cross-linking, 1.2 mmol/g) were obtained from Novabio-
chem Co. The amino acids were purchased either from Nova-
biochem Co. or from Advanced Chemtech, Inc. Other chemicals
and anhydrous solvents were obtained from Aldrich, Fluka,
Lancaster, and Fisher. All reagents and solvents were used
without further purification. All synthetic compounds and
intermediates gave satisfactory MS. Mass spectrometry analyses
were performed on a Finnigan LCQduo system. HPLC analyses
were recorded on a Gilson HPLC system consisting of a 170
diode array detector, a 215 liquid handler, and a 322 pump. 1H
and 13C NMR spectra were recorded at 300 and 75 MHz on a
Conclusion
A series of potent, selective peptidomimetic inhibitors of FXIa
have been designed and synthesized. Some of these inhibitors

Peptidomimetic Inhibitors of Factor XIa
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7787
Table 4. Stereochemistry Studies of R4 Elements
a nd, not determined.
Bruker 300 NMR spectrometer with CDCl3, DMSO-d6, and
MeOH-d4 as solvents.
h at room temperature. The resin was filtered and washed with
DMF f MeOH (four times) and CH2Cl2 (two times) and then
dried under vacuum overnight to give resin 4a.
General Procedure for Preparation of Resin 3. Deprotec-
tion of Fmoc Group of Resin 2. The preloaded Fmoc-Val-
Wang resin 2 (10.0 g, 6.3 mmol, 0.63 mmol/g) was treated with
20% piperidine in DMF (120 mL) at room temperature for 2 h.
The resin-bound amine was filtered and washed with DMF f
MeOH (four times) and CH2Cl2 (two times) and then dried under
vacuum.
Coupling with (L)-Fmoc-Xaa-OH. The above resin-bound
amine (6.3 mmol) was preswelled in 30 mL of dry CH2CH2.
HOBt (3.65 g, 27.0 mmol) and 4-(dimethylamino)pyridine
(DMAP) (0.39 g, 3.2 mmol) were then added. To this mixture
was added a solution of (L)-Fmoc-Xaa-OH (10.5 g, 27.0 mmol)
in dry DMF (70 mL), followed by the addition of 1,3-
diisopropylcarbodiimide (DIC) (4.23 mL, 27.0 mmol). The
mixture was agitated using an orbital shaker at room temperature
for 16 h. The resin was filtered, washed with DMF f MeOH
(four times) and CH2Cl2 (two times) and dried under vacuum
to give resin 3.
General Procedure for Preparation of Acid 5. Deprotec-
tion of Fmoc Group of Resin 3. The above resin (4.0 g, 2.4
mmol, 0.60 mmol/g) was treated with 20% piperidine in DMF
(50 mL) at room temperature for 2 h. The resin was filtered
and washed with DMF f MeOH (four times) and CH2Cl2 (two
times) and then dried under high vacuum, yielding 3.1 g of the
resin-bound amine.
Formation of the Resin-Bound Amide 4a. The resin-bound
amine from 3 (200 mg, 0.12 mmol) was treated with acid
chloride (6 equiv, 0.72 mmol) in 2 mL of the mixture of
pyridine/CH2Cl2 (1:9). The resulting slurry was agitated for 3
Formation of the Resin-Bound Sulfonamide 4b. The resin-
bound amine from 3 (300 mg, 0.18 mmol) was treated with
sulfonyl chlorides (6 equiv, 1.2 mmol) in 3 mL of the mixture
of 2,6-lutidine/CH2Cl2 (1:9). The resulting slurry was agitated
for 3 h at room temperature. The resin was filtered and washed
with DMF f MeOH (four times) and CH2Cl2 (two times) and
then dried under vacuum overnight to give resin 4b.
Formation of the Resin-Bound Urea 4c (Method A). The
resin-bound amine from 3 (200 mg, 0.12 mmol, loading ) 0.63
mmol/g) was treated with isocyanate (0.72 mmol, 6.0 equiv)
and DMAP (0.5 equiv) in 2 mL of the mixture DMF/CH2Cl2
(2:1). The reaction mixture was shaken overnight at room
temperature. The resin was filtered and washed with DMF f
MeOH (four times) and CH2Cl2 (two times) and then dried under
vacuum overnight to give resin 4c.
Formation of the Resin-Bound Urea 4d (Method B). The
resin-bound amine from 3 (3.0 g, 1.9 mmol) was treated with
a freshly prepared 0.5 M solution of p-nitrophenyl chloroformate
(3.81 g, 19 mmol, 10 equiv) and DIEA (3.3 mL, 19 mmol, 10
equiv) in 38 mL of THF/CH2Cl2 (1:1). The resulting slurry was
agitated for 1 h at room temperature. The resin was filtered
and washed with THF/CH2Cl2 (1:1) (three times), followed by
the addition of (R)-1-(4-bromophenyl)ethylamine (3.8 g, 19
mmol, 0.5 M) and triethylamine (2.6 mL, 19 mmol, 0.5 M) in
38 mL of DMF. After 2 h of mixing, the resin was washed
with DMF (four times) and THF/CH2Cl2 (1:1) (four times), then
dried under vacuum overnight to give resin 4d.

7788 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
Table 5. SAR Exploration of R3 Elements
Lin et al.
Table 6. In Vitro Anticoagulant Activity in Human Plasma
product 5 for the next coupling step without further purifi-
cation.
compd
IC50 (nM)
2× aPTT (µΜ)
2× PT (µΜ)
General Procedure of Coupling of Acid 5 and Amine 6
To Afford Product 1. Coupling Reaction of 5 with 6 (Scheme
3). The acid, 5, (500 mg, 1.02 mmol), HOBt (168 mg, 1.24
mmol), and EDCI (257 mg, 1.34 mmol) were dissolved in dry
THF (10 mL), followed by the addition of DIEA (0.53 mL, 3.1
mmol). Sonication of the mixture may be needed to assist
solubility. After 5-10 min of stirring, a solution of the
thiazolylarginol trifluoroacetic acid salt 629,33 (511 mg, 0.9
mmol) in THF (4 mL) was added along with a THF (0.5 mL)
rinse. The reaction mixture was stirred at room temperature for
3 h and then diluted with 100 mL of ethyl acetate. The organic
layer was washed sequentially with 5% aqueous acetic acid (40
mL × 2), saturated aqueous sodium bicarbonate (40 mL × 2),
water (40 mL), and brine. The organic layer was dried over
anhydrous sodium sulfate, filtered, and concentrated in vacuo
32
34
37
36
28
6
8
12
30
10
2.4
0.56
2.4
1.35
15.4
25.0
26.3
31.0
48.6
178.6
Cleavage. Resin 4 was swollen with 15 mL of CH2Cl2. The
solvent was then drained. The resin was treated with 50% TFA/
DCM (38 mL) at room temperature for 3 h. The resin was
filtered and washed with CH2Cl2 (two times). The collected
cleavage product was combined with washes and mixed with
1.5 mL of iPrOH. After concentration of solvents in vacuo, the
collected yellow residue was triturated with Et2O (50 mL), then
filtered, and dried under vacuum, yielding a sufficiently pure

Peptidomimetic Inhibitors of Factor XIa
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7789
Figure 2. Efficacy of 32 in a rat model of venous thrombosis. The
model used stenosis of the inferior vena cava with rapid infusion of
hypotonic saline followed by partial stasis of blood flow to induce clot
formation.
Figure 1. Active site of the crystal structure of FXIa-36 complex.
FXIa is in a green ribbon representation with S1, S2, and S4 pockets
labeled. Ligand 36 is in stick representation and colored according to
atom type: oxygen in red, nitrogen in blue, sulfur in yellow, bromine
in dark red, and carbon in gray. The catalytic triad (S195, H57, and
D102) of FXIa and residues D189 and K192 are in stick representation
with carbon atoms in green. Hydrogen-bonding interactions that are
discussed in the text are shown in black lines with the distance indicated
in angstroms
Chart 3. Schematic Representation of the Binding of 36 in the
FXIa Substrate Site As Seen by X-ray Crystallography
Figure 3. Effect of 37 on rat mesenteric arteriole bleeding time. The
bleeding time of punctured mesenteric arterioles was determined prior
to compound administration (light gray) and at 15 min after initiation
of compound administration (dark gray). Compound 37 was adminis-
tered as a continuous infusion into the femoral vein at 0.25 and 1 mg/
kg per 15 min. Heparin was administered as a bolus injection followed
by a continuous infusion into the femoral vein at a dose of 25 unit/kg
+ 50 units/kg/h (total dose ) 38 units/kg). Results are expressed as
mean bleeding times ( SEM (n ) 6-8 per group: *, p < 0.05,
Student’s t test).
Deprotection of Mtr Group. The above coupling product
(728 mg) was dissolved in 95% TFA (14.3 mL), 2.5% water
(375 µL), and 2.5% thioanisole (375 µL). The resulting dark
brown solution was stirred for 3 h at room temperature. The
solution was concentrated under reduced pressure to a dark
brown gum. The gum was triturated with diethyl ether. The ether
supernatant was decanted. The ether trituration was repeated
two more times. The yellow solid was dried under vacuum
overnight to yield 720 mg of deprotected product 1 with a ratio
of 2.3:1.0 mixture of the desired and undesired isomers. This
crude mixture was purified by preparative HPLC to give the
desired isomer with purity >95%.
Table 7. Pharmacokinetics Parameters and Selectivity Profile of 32
plasma
Enzyme Affinity and Anticoagulant Activity. Inhibition
Assay (IC50). Peptidolytic activities for thrombin, trypsin, and
factor XIa were measured using a fluorogenic reporter group,
7-methyl-4-aminocoumarin (AMC), attached via an amide
linkage to the custom tripeptide pyroglutamic acid-proline-
arginine (Glp-Pro-Arg). Commercially available substrate, H-
Pro-Phe-Arg-AMC (BACHEM I-1295), was used for FXa
activity. All assays were carried out in a Tris buffer (100 mM
compd FXIa thrombin FXa FVIIa kallikrein APC trypsin
32
6a
2,040
1600 38000
10
20700
12
a All IC50 values are nanomolar.
to yield Mtr-protected 1 as yellow foam. LC-MS analysis
indicated the presence of a trace amount of undesired diaster-
eomeric isomers.

7790 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
Lin et al.
Tris-HCl/200 mM NaCl and 0.02% Tween 20) at pH 7.4.
Enzyme and substrate concentrations were 50 nM and 50 µM,
respectively, for thrombin, trypsin, and factor XIa. Factor Xa
required twice the concentrations. For all of the enzymes but
factor Xa, the 10-point dose-response assays were run as
endpoints after 15-30 min of incubation at 30 °C. Factor Xa
required a longer, 60 min incubation time. Released AMC was
measured at an emission wavelength of 460 nm on a Wallac
1420 Multilabel Counter excited at 355 nm using a continuous
source. The data were fit to a dose-response one-site model
using a nonlinear least-squares algorithm to determine the IC50
and Hill coefficient.
30G needle while being superfused with saline at 37 °C.
Bleeding time was determined as cessation of blood flow from
the puncture sites. After the intestines were replaced, vehicle
or heparin was administered as an intravenous bolus injection
followed by continuous infusion via the femoral vein. Com-
pound 37 was administered as a continuous infusion only.
Fifteen minutes after the start of infusion, another small portion
of the small intestine was exteriorized, and again three arterioles
were punctured with a 30G needle and flushed with saline at
37 °C to determine bleeding time. The animals were then
sacrificed by an anesthetic overdose.
Bleeding time from mesenteric arterioles was determined in
each animal prior to compound administration and then after
15 min of continuous infusion. Compound infusion was
continued during the determination of bleeding time. Animals
with bleeding times of >210 s prior to compound administration
were excluded from the study. If bleeding time was >9 min,
the time was assumed to be 9 min for calculation purposes.
Protein Purification and X-ray Crystallography. The
detailed expression and purification procedure of the mutated
catalytic domain of FXIa (rhFXI370-607-S434A,T475A,C482S,-
K437A, hereinafter FXIac) and the crystallization method for
the FXIac-benzamidine complex has been published.28b Com-
pound 36 was added to a drop of crystallization solution (2.0
M ammonium sulfate, 0.1 M Tris-HCl, pH 8.5) in powder form.
A FXIac-benzamidine crystal was soaked in the compound 36
containing solution overnight at 10 °C. After soaking, the crystal
was briefly put into cryoprotection solution containing crystal-
lization solution with 20% (v/v) glycerol before flash freezing
in liquid nitrogen. X-ray diffraction data were collected on a
RAXIS-IV++ image plate with an RU-H3R rotating anode X-ray
source (Rigaku) equipped with blue Osmic mirrors and X-stream
2000 system (Rigaku). Data indexing, processing, and scaling
were performed with HKL2000.34 The structure of FXIac-
compound 36 was solved using the structure of FXIac com-
plexed with benzamidine (PDB code 1ZHR) with benzamidine
and water molecules removed from the initial model. Refinement
and difference maps were calculated by CNX (Accelrys), and
ligand placement and manual rebuilding of FXIac were per-
formed using QUANTA (Accelrys). The final model contains
residues 370-606 of factor XI (residues 16-244 using chy-
motrypsin numbering), 128 water molecules, and the bound
ligand. The crystal structure of 36 was refined to an R of 20.7%
and an Rfree of 22.6% at 2.25 Å resolution (PDB code 1ZOM).
Human Plasma Based Clotting Assay. aPTT Assay. To
determine the effect of compounds on coagulation pathway,
activated partial thromboplastin time was determined using a
Thromboscreen 400C instrument. This pathway involves FXII,
kallikrein, and FXI, which activates factors IX and FVIII,
leading to activation of FX and FV and then activation of factor
II to form a blood clot. Plasma (50 µL) and different concentra-
tions of compounds were added to cuvettes. After incubation
for 2 min, aPTT reagent (ALEXIN, Sigma) was added (50 µL)
and incubated for an additional 3 min. The cuvettes were
transferred to a measuring position,and prewarmed CaCl2
reagent (50 µL, 30 mM) was added. Readings were then taken
over a maximum of 300 s. A dose-response curve was
generated, and the concentration at which the clotting time was
doubled (2× aPTT) was determined.
PT Assay. The PT assay was also used to measure inhibition
of coagulation. In this assay, the factor XI dependent steps are
bypassed. Hence, the assay measures inhibition of factor VIIa,
factor Xa, and thrombin, but not FXI. This assay measures the
ability of factor VIIa to activate factor X, which activates factor
II to form a blood clot. The assay was performed using a
Thromboscreen 400C instrument and thromboplatin reagent
(ThromboMax with calcium, Sigma). Plasma (50 µL) and
compounds were prewarmed at 37 °C for 3 min in the cuvettes.
The cuvettes were transferred to the measuring position. The
prewarmed thromboplastin reagent (100 µL) was then added,
and readings were taken over 300 s. A dose-response curve
was generated, and the concentration at which the clotting time
was doubled (2× PT) was determined.
Rat Model of Venous Thrombosis. An established rat model
of venous thrombosis was used to assess the antithrombotic
potential of FXI inhibitors. Briefly, under anesthesia a stenosis
of the inferior vena cava with rapid infusion of hypotonic saline
followed by reduced blood flow was used to induce clot
formation.29,30 Rats were treated with vehicle alone (0.1 M
NaOAc, pH 4.0), 32 at 0.0625 or 0.25 mg/kg intravenous
infusion, or heparin (50 units/kg by intravenous bolus followed
by a continuous infusion of 25 units/kg/h). Following a treatment
time of 10 min, the clot was isolated and weighed. The thrombus
weight in both the 0.25 mg/kg 32 and heparin-treated groups
was significantly reduced compared to that found in the vehicle-
treated animals (p < 0.05, n ) 9, Tukey-Kramer multiple-
comparison test) (see Supporting Information for complete
experimental details).
Supporting Information Available: Detailed synthetic proce-
dures and a table of characterization data of final R-ketothiazole
products 7-37. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) Mann, K. G.; Butenas, S.; Brummel, K. The dynamics of thrombin
formation. Arterioscler. Thromb. Vasc. Biol. 2003, 45, 1-10.
(2) Jenny, N. S.; Mann, K. G. Coagulation Cascade: an Overview. In
Thrombosis and Hemorrhage; Loscalzo, J., Schafer, A. I., Eds.;
Williams and Wilkins: Baltimore, MD, 1998; pp 3-27.
(3) Hirsh, J.; Dalen, J.; Anderson, D. R.; Poller, L.; Bussey, H.; Ansell,
J.; Deykin, D. Oral anticoagulants: mechanism of action, clinical
effectiveness, and optimal therapeutic range. Chest 2001, 119 (1,
Suppl.), 8S-21S.
(4) Hirsh, J.; Warkentin, T. E.; Shaughnessy, S. G.; Anand, S. S.;
Halperin, J. L.; Raschke, R.; Granger, C.; Ohman, E. M.; Dalen, J.
E. Heparin and low-molecular-weight heparin: mechanisms of action,
pharmacokinetics, dosing, monitoring efficacy and safety. Chest 2001,
119 (1, Suppl.), 64S-94S.
Rat Mesenteric Arteriole Bleeding Time. The rat model
used was modified from established techniques.31,35,36 Briefly,
male Sprague-Dawley rats were anesthetized with ketamine/
xylazine. Body temperature was maintained at 37 °C. Com-
pounds were administered through a catheter implanted in the
left femoral vein. A small portion of the small intestine was
exteriorized to display the mesenteric artery. Three arterioles
branching from the mesenteric artery were punctured with a
(5) Kelton, J. G. Heparin-induced thrombocytopenia: an overview. Blood
ReV. 2002, 16, 77-80.
(6) Hirsh, J.; O’Donnell, M.; Weitz, J. I. New anticoagulants. Blood 2005,
105, 453-463.

Peptidomimetic Inhibitors of Factor XIa
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7791
(7) Weitz, J. I. New anticoagulants for treatment of venous thromboem-
bolism. Circulation 2004, 110 (Suppl. I), I-19-26.
(20) Meijers, J. C. M.; Tekelenburg, W. L. H.; Bouma, B. N.; Bertina, R.
M.; Rosendaal, F. R. High levels of coagulation factor XI as a risk
factor for venous thrombosis. New Engl. J. Med. 2000, 342, 696-
701.
(21) (a) Merlo, C.; Wuillemin, W. A.; Redondo, M.; Furlan, M.; Sulzer,
I.; Kremer-Hovinga, J.; Binder, B. R.; Lammle, B. Elevated levels
of plasma prekallikrein, high molecular weight kininogen and factor
XI in coronary heart disease. Atherosclerosis 2002, 161, 261-267.
(b) Minnema, M. C.; Peters, R. J. G.; de Winter, R.; Lubbers, Y. P.
T.; Barzegar, S.; Bauer, K. A.; Rosenberg, R. D.; Hack, C. E.; ten
Cate, H. Activation of clotting factors XI and IX in patients with
acute myocardial infarction. Arterioscler. Thromb. Vasc. Biol. 2000,
20, 2489-2493.
(22) Wang, X.; Cheng, Q.; Xu, L.; Feuerstein, G. Z.; Hsu, M. Y.; Smith,
P. L.; Seiffert, D. A.; Schumacher, W. A.; Ogletree, M. L.; Gailani,
D. Effect of factor IX or factor XI deficiency on ferric chloride-
induced carotid artery occlusion in mice. J. Thromb. Haemost. 2005,
3, 695-702.
(23) Gruber, A.; Hanson, S. R. Factor XI-dependence of surface-tissue
factor-initiated thrombus propagation in primates. Blood 2003, 102,
953-955.
(24) Bern, M. M.; Sahud, M.; Zhukov, O.; Qu, K.; Mitchell, W., Jr.
Treatment of factor XI inhibitor using recombinant activated factor
VIIa. Haemophilia 2005, 11, 20-25.
(25) Witting, J. I.; Pouliott, C.; Catalfamo, J. L.; Fareed, J.; Fenton, J.
W. Thrombin inhibition with dipeptidyl argininals. Thromb. Res.
1988, 50, 461-467.
(26) (a) Kettner, C.; Shaw, E. D-Phe-Pro-Arg-CH2Cl, a selective affinity
label for thrombin. Thromb. Res. 1979, 14, 969-973. (b) Hauptmann,
J.; Markwardt, F. Studies on the anticoagulant and antithrombotic
action of an irreversible thrombin inhibitor. Thromb. Res. 1980, 20,
347-351.
(27) (a) Tsutsumi, S.; Okonogi, T.; Shibahara, S.; Patchett, A. A.;
Christensen, B. G. R-Ketothiazole inhibitors of polyl endopeptidase.
Bioorg. Med. Chem. Lett. 1994, 4, 831-834. (b) Akiyama, Y.;
Tsutumi, S.; Hatsushiba, E.; Ohuchi, S.; Okonogi, T. Peptidyl
R-ketothiazole as potent thrombin inhibitors. Bioorg. Med. Chem.
Lett. 1997, 7, 533-538.
(28) (a) Jin, L.; Pandey, P.; Babine, R.; Gorga, J. C.; Seidl, K. J.; Gelfand,
E., Weaver, D. T.; Abdel-Meguid, S. S.; Strickler, J. E. Crystal
structures of the FXIa catalytic domain in complex with ecotin
mutants reveal substrate-like interactions. J. Biol. Chem. 2005, 280,
4704-4712. (b) Jin, L.; Pandey, P.; Babine, E. R.; Weaver, D. T.;
Abdel-Meguid, S. S.; Strickler, J. E. Mutation of surface residues to
promote crystallization of activated factor XI as a complex with
benzamidine: an essential step for the iterative structure-based design
of factor XI inhibitors. Acta Cryst. D: Biol. Cryst. 2005, 61, 1418-
1425.
(29) Berryman, K. A.; Doherty, A. M.; Edmunds, J. J.; Plummer, J. S.
Process for the preparation of chiral keto-heterocycles of basic amino
acids. WO 97/48687, A1, 1997.
(30) Millet, J.; Theveniaux, J.; Pascal, M. A new experimental model of
venous thrombosis in rats involving partial stasis and slight endot-
helium alterations. Thromb. Res. 1987, 45, 123-133.
(31) Schumacher, W. A.; Heran, C. L.; Steinbacher, T. E. Low-molecular-
weight heparin (fragmin) and thrombin active-site inhibitor (arga-
troban) compared in experimental arterial and venous thrombosis and
bleeding time. J. CardioVasc. Pharm. 1996, 23, 19-25.
(32) Strickler, J. E.; Pandey, P.; Bibbins, F.; Weaver, D. Data not shown.
(33) The modified preparative procedure for 6 is described in the
Supporting Information.
(34) Otwinowski, Z.; Minor, W. Processing of x-ray diffraction data
collected in oscillation mode. In Macromolecular Crystallography,
Part A; Carter, J. C. W., Sweet, R. M., Eds.; Academic Press: New
York, 1997; Vol. 276, pp 307-326.
(35) Zawilska, K. M.; Born, G. V. R.; Begent, N. A. Effect of ADP-
utilizing enzymes on the arterial bleeding time in rats and rabbits.
Br. J. Haematol. 1982, 50 (2), 317-325.
(36) Hsiao, G.; Yen, M. H.; Lee, Y. M.; Sheu, J.-R. Antithrombotic effect
of PMC, a potent R-tocopherol analogue on platelet plug formation
in vivo. Br. J. Haematol. 2002, 117 (3), 699-704.
(8) Lee, A. V.; Vlasuk, G. P. Recombinant nematode anticoagulant
protein c2 and other inhibitors targeting blood coagulation factor VIIa/
tissue factor. J. Intern. Med. 2003, 254, 313-321.
(9) Lee, A.; Agnelli, G.; Buller, H.; Ginsberg, J.; Heit, J.; Rote, W.;
Vlasuk, G.; Costantini, L.; Julian, J.; Comp, P.; van der Meer, J.;
Piovella, F.; Raskob, G.; Gent, M. Dose-response study of recom-
binant VIIa/tissue factor inhibitor recombinant nematode anticoagu-
lant protein c2 in prevention of postoperative VTE in patients
undergoing total knee replacement. Circulation 2001, 104, 74-78.
(10) Refino, C. J.; Jeett, S.; DeGuzman, L.; Bunting, S.; Kirchhofer, D.
A human antibody that inhibits factor IX/IXa function potently
inhibits arterial thrombosis without increasing bleeding. Arterioscler.
Thromb. Vasc. Biol. 2002, 22, 517-522.
(11) Abdel-Meguid, S. S.; Babine, R. E.; Deng, H.; Jin, L.; Lin, J.; Magee,
S. R.; Meyers, H. V.; Pandey, P.; Rynkiewicz, M. J.; Weaver, D. T.
Blood coagulation factor XI inhibitors and methods for treatment of
thrombosis. WO 04089297 A2, 2004.
(12) Schumacher, W. A.; Seiler, S. M.; Belfield, J. S. Methods of treating
thrombosis with reduced risk of bleeding. U.S. 2004/0180855 A1,
2004.
(13) An inhibitor of TAFIa is used for suppression of fibrinolysis thus
enhancing clot lysis. (a) Bouma, B. N.; Marx, P. F.; Mosnier, L. O.;
Meijers, J. C. M. Thrombin-Activatable Fibrinolysis Inhibitor (TAFI,
Plasma Procarboxypeptidase B, Procarboxypeptidase R, Procarbox-
ypeptidase U). Thromb. Res. 2001, 101, 329-354. (b) Nagashima,
M.; Yin, Z.-F.; Zhao, L.; White, K.; Zhu, Y.; Lasky, N.; Halks-Miller,
M.; Broze, G. J., Jr.; Fay, W. P.; Morser, J. Thrombin-activatable
fibrinolysis inhibitor (TAFI) deficiency is compatible with murine
life. J. Clin. InVest. 2002, 109, 101-110. (c) Polla, M. O.; Tottie,
L.; Norden, C.; Linschoten, M.; Musil, D.; Trumpp-Kallmeyer, S.;
Aukrust, I. R.; Ringom, R.; Holm, K. H.; Neset, S. M.; Sandberg,
M.; Thurmond, J.; Yu, P.; Hategan, G.; Anderson, H. Design and
synthesis of potent, orally active, inhibitors of carboxypeptidase U
(TAFIa). Bioorg. Med. Chem. 2004, 12, 1151-1175.
(14) (a) Esmon, C. T.; Ding, W.; Yasuhiro, K.; Gu, J. M.; Ferrell, G.;
Regan, L. M.; Stearns-Kurosawa, D. J.; Kurosawa, S.; Mather, T.;
Laszik, Z.; Esmon, N. L. The protein C pathway: new insights.
Thromb. Haemost. 1997, 78, 70-74. (b) Esmon, C. T. The protein
C pathway. Chest 2003, 124 (3, Suppl.), 26S-32S.
(15) (a) Naito, K.; Fujikawa, K. Activation of human blood coagulation
factor XI independent of factor XII: factor XI is activated by
thrombin and factor XIa in presence of negatively charged surface.
J. Biol. Chem. 1991, 266, 7353-7358. (b) Gailani, D.; Broze, G.
Factor XI activation in a revised model of coagulation. Science 1991,
253, 909-912. (c) Walsh, P. N. Roles of platelets and factor XI in
the initiation of blood coagulation by thrombin. Thromb. Haemost.
2001, 86, 75-82.
(16) Davie, E. W.; Fujikawa, K.; Kisiel, W. The coagulation cascade:
initiation, maintenance, and regulation. Biochemistry 1991, 30,
10363-10370.
(17) (a) von dem Borne, P. A. K.; Meijers, J. C. M.; Bouma, B. M.
Feedback activation of factor XI by thrombin in plasma results in
additional formation of thrombin that protects fibrin clots from
fibrinolysis. Blood 1995, 86, 3035-3042. (b) Minnema, M. C.;
Friederich, P. W.; Levi, M.; von dem Borne, P. A. K.; Mosnier, L.
O.; Meijers, J. C. M.; Biemond, B. J.; Hack, C. E.; Bouma, B. M.;
ten Cate, H. Enhancement of rabbit jugular vein thrombolysis by
neutralization of factor XI: in vivo evidence for a role of factor FXI
as an anti-fibrinolytic factor. J. Clin. InVest. 1998, 101, 10-14. (c)
Bouma, B. M.; Meijers, J. C. M. Role of blood coagulation of factor
XI in downregulation of fibrinolysis. Curr. Opin. Hemotol. 2000, 7,
266-272.
(18) Wielders, S. J. H.; Be´guin, S.; Hemker, H. C.; Lindhout, T. Factor
XI-dependent reciprocal thrombin generation consolidates blood
coagulation when tissue factor is not available. Arterioscler. Thromb.
Vasc. Biol. 2004, 24, 1-6.
(19) (a) Askai, R.; Chung, D. W.; Davie, E. W.; Seligsohn, U. Factor XI
deficiency in Ashkenazi Jews in Israel. New Engl. J. Med. 1991,
325, 153-158. (b) Berliner, S.; Horowitz, I.; Martinowitz, U.;
Brenner, B.; Seligsohn, U. Dental surgery in patents with severe factor
XI deficiency without plasma replacement. Blood 1992, 3, 465-
468. (c) Ragni, M. V.; Sinha, D.; Seaman, F.; Lewis, J. H.; Spero, J.
A.; Walsh, P. N. Blood 1985, 65, 719-724.
JM060978S