Regular Article

THROMBOSIS AND HEMOSTASIS

Factor XII inhibition reduces thrombus formation in a primatethrombosis modelAnton Matafonov,1,2 Philberta Y. Leung,3,4 Adam E. Gailani,1 Stephanie L. Grach,1 Cristina Puy,3 Qiufang Cheng,1

Mao-fu Sun,1 Owen J. T. McCarty,3 Erik I. Tucker,3,4 Hiroaki Kataoka,5 Thomas Renne,6,7 James H. Morrissey,8

Andras Gruber,3,4 and David Gailani1

1Department of Pathology, Microbiology and Immunology, Vanderbilt University, Nashville, TN; 2Department of Bioengineering and Organic Chemistry,

Tomsk Polytechnic University, Tomsk, Russia; 3Departments of Biomedical Engineering and Medicine, Oregon Health and Science University, Portland,

OR; 4Aronora, Inc., Portland, OR; 5Department of Pathology, University of Miyazaki, Miyazaki, Japan; 6Department of Molecular Medicine and Surgery,

Karolinska Institutet and University Hospital, Stockholm, Sweden; 7Institute of Clinical Chemistry, University Hospital Hamburg-Eppendorf, Hamburg,

Germany; and 8Department of Biochemistry, University of Illinois at Urbana-Champaign, Urbana, IL

Key Points

• Factor XII can contributeto thrombus formation inhuman and nonhumanprimate blood.

• An antibody that blocksfactor XII activation (15H8)produces an antithromboticeffect in a primate thrombosismodel.

The plasma zymogens factor XII (fXII) and factor XI (fXI) contribute to thrombosis in

a variety ofmousemodels. These proteins serve a limited role in hemostasis, suggesting

that antithrombotic therapies targeting them may be associated with low bleeding risks.

Although there is substantial epidemiologic evidence supporting a role for fXI in human

thrombosis, the situation is not as clear for fXII. We generated monoclonal antibodies

(9A2and15H8)against thehuman fXII heavychain that interferewith fXII conversion to the

protease factor XIIa (fXIIa). The anti-fXII antibodieswere tested inmodels inwhich anti-fXI

antibodies are known to have antithrombotic effects. Both anti-fXII antibodies reduced

fibrin formation in human blood perfused through collagen-coated tubes. fXII-deficient

mice are resistant to ferric chloride–induced arterial thrombosis, and this resistance can

be reversed by infusion of human fXII. 9A2 partially blocks, and 15H8 completely blocks,

the prothrombotic effect of fXII in this model. 15H8 prolonged the activated partial

thromboplastin time of baboon and human plasmas. 15H8 reduced fibrin formation in

collagen-coated vascular grafts inserted into arteriovenous shunts in baboons, and reduced fibrin and platelet accumulation

downstream of the graft. These findings support a role for fXII in thrombus formation in primates. (Blood. 2014;123(11):1739-1746)

Introduction

There is considerable interest in the possibility that targetedinhibition of the plasma protease factor XIIa (fXIIa) and factor XIa(fXIa) may be useful for preventing or treating thrombosis.1-5 Micelacking factor XII (fXII) or factor XI (fXI), the zymogens of fXIIaand fXIa, respectively, are resistant to injury-induced arterial andvenous thrombosis,6-9 and to ischemia-reperfusion injury in thebrain10

and heart.11 fXI inhibition also reduces experimental thrombus for-mation in primates.8,12,13 Humans lacking fXII do not bleed ab-normally,5,14 and fXI-deficient patients have a relativelymild bleedingdisorder,14,15 raising the prospect that drugs targeting these proteinswill produce antithrombotic effects without significantly compromis-ing hemostasis. In the cascade/waterfallmodels of plasma coagulation,sequential proteolytic reactions initiated by conversion of fXII to fXIIalead to thrombin generation and fibrin formation.16,17 fXIIa catalyzesconversion of fXI to fXIa during this process.Autoactivation of fXII inplasma is readily induced by adding minerals such as silica or kaolinthat provide a surface on which the reaction occurs, and amplified byreciprocal activation of fXII and prekallikrein (PK) in a process called

contact activation.4,14 In vivo, polymers such as collagen,18 laminin,19

RNA,20 DNA,21 and polyphosphate,9 as well as misfolded proteinaggregates (such as occur in systemic amyloidosis),22 may facilitatefXII activation by similar processes, contributing to thrombosis. fXIImay also be activated on membranes of vascular endothelial cells bya distinct mechanism initiated by prolylcarboxypeptidase-mediatedconversion of PK to a-kallikrein.23,24

Consistent with data from mouse studies, there is substantialevidence supporting a role for fXI in arterial and venous thrombosisin humans.25-30 However, available datamake a less compelling casefor a role for fXII in human thrombosis. Indeed, 2 large surveysreported the counterintuitive observation that plasma fXII levels areinversely correlated with risk of myocardial infarction26 and deathfrom all causes.31 This suggests that the contribution of fXII to throm-bus formation observed in rodents may not reflect processes inhumans.To address this issue,wedevelopedmonoclonal antibodies tohuman fXII that inhibit fXII activation, specifically to determinewhether blocking fXII in a primate thrombosis model produces an

Submitted April 25, 2013; accepted December 27, 2013. Prepublished online

as Blood First Edition paper, January 9, 2014; DOI 10.1182/blood-2013-04-

499111.

A.M. and P.Y.L. contributed equally to the study and to manuscript preparation

and should be considered cofirst authors.

The online version of this article contains a data supplement.

There is an Inside Blood commentary on this article in this issue.

The publication costs of this article were defrayed in part by page charge

payment. Therefore, and solely to indicate this fact, this article is hereby

marked “advertisement” in accordance with 18 USC section 1734.

© 2014 by The American Society of Hematology

BLOOD, 13 MARCH 2014 x VOLUME 123, NUMBER 11 1739

For personal use only.on April 29, 2017. by guest www.bloodjournal.orgFrom

antithrombotic effect similar to the reported effects of antibodies thatblock fXI.

Materials and methods

Proteins

Human fXII, fXIIa, PK, and high-molecular-weight kininogen (HK) werepurchased from Enzyme Research Laboratories. Human fXI and fXIa werepurchased from Haematologic Technologies.

Anti-fXII monoclonal antibodies

The murine fXII null genotype (C57Bl/6 background)7 was crossed onto theBalb-C background through 7 generations. fXII-deficient Balb-C mice weregiven 25mg of a mixture of human fXII and fXIIa by intraperitoneal injectionin Freund complete adjuvant on day 0 and Freund incomplete adjuvant on day28. A 25-mg booster dose in saline was given on day 70. On day 73, spleenswere removed and lymphocytes were fused with P3X63Ag8.653 myelomacells using a standard polyethylene glycol–based protocol. Antibodies weretested for capacity to recognize human fXII by enzyme-linked immunosor-bent assay (ELISA) and western blot, and to prolong the activated partialthromboplastin time (aPTT) of human plasma. Clones 9A2 and 15H8 weresubcloned, expanded in a CL1000 bioreactor (Integra Biosciences), andpurified by cation exchange and thiophilic agarose chromatography.

Expression of recombinant fXII and antibody mapping

The human fXII complementary DNA (cDNA) was inserted into vectorpJVCMV.32 Sequence encoding individual domains from the fXII homologhepatocyte growth factor activator (HGFA) were amplified from the humanHGFA cDNA by polymerase chain reaction (PCR),33 and used to replacecorresponding sequence in the fXII cDNA (Figure 1A and supplementalTables 1-2, available on the BloodWeb site). HEK293 fibroblasts (ATCC-CRL1573) were transfected with pJVCMV/fXII-HGFA constructs as des-cribed.32 Conditioned serum-free media (Cellgro Complete; Mediatech) fromexpressingcloneswere size-fractionatedon10%polyacrylamide–sodiumdodecylsulfate gels, and chemiluminescentwestern blotswere prepared using 9A2, 15H8,or goat polyclonal–anti-human fXII immunoglobulin G (IgG) for detection.

Clotting assays

aPTT assays were performed by mixing 65 mL of normal plasma (0.32%sodium citrate weight to volume ratio [w/v] with an equal volume ofphosphate-buffered saline (PBS) with or without 8mM anti-fXII IgG. After 5minutes at room temperature (RT), 65 mL of aPTT reagent (DiagnosticaStago) was added, followed by a 5-minute incubation at 37°C. CaCl2(25mM–65mL)was added and time to clot formation determined on an ST4fibrometer (Diagnostica Stago). In separate assays, 65 mL of fXIIa (50 nM)in PBS was incubated with an equal volume of antibody (1mM) or vehicle for15 minutes prior to addition of 65 mL of plasma. CaCl2 was added and time toclot formation determined.

fXII activation

Polyphosphate (75-100 phosphate units) was prepared by gel electrophoresisas described.9 fXII (100 nM) was incubated with aPTT reagent (2.5% of totalvolume) or polyphosphate (2 mM) at 37°C in the presence of 1 mM 9A2,15H8, both antibodies, or vehicle in reaction buffer (50 mMTris-HCl pH 7.4,100 mM NaCl, and 1 mg/mL polyethylene glycol 8000). At various times,aliquots were removed into Polybrene (5 mM final). fXIIa activity wasidentified by adding chromogenic substrate S-2302 (500 mM; Diapharma)and following changes in optical density (OD) 405 nmon amicroplate reader.Results were compared with a control curve prepared with pure fXIIa.

PK activation

fXIIa (1 nM) was incubated with PK (50 nM) and HK (70 nM) in reactionbuffer containing 250 mMCS-31(02) (Biophen) at RT, with or without aPTT

reagent (5% volume to volume ratio [v/v]), and with or without anti-fXIIIgG (100 nM). Changes in OD 405 nm reflecting conversion of PK toa-kallikrein were followed on a microplate reader.

Thrombin generation

Normal plasma (0.32% sodium citrate w/v) was supplemented with 415 mMZ-Gly-Gly-Arg-AMC, 5 mM phosphatidylcholine/phosphatidylserine ves-icles, and 4mM IgG anti-fXII IgG. Supplemented plasma (40mL) was mixedwith aPTT reagent (1% v/v) or type I fibrillar collagen (100 mg/mL; Chrono-Log). Tenmicroliters of 20mMHEPES (N-2-hydroxyethylpiperazine-N9-2-ethanesulfonic acid), pH 7.4, 100 mM CaCl2, 6% bovine serum albumin(BSA)was added andfluorescence (excitationl 390 nm, emission l 460 nm)was monitored at 37°C on a Thrombinoscope.34 In a separate experiment,fXII-deficient plasma (George King) was treated in a similar manner, exceptthat 5 nM fXIIa was added with aPTT reagent (5% v/v). Each conditionwas tested 3 times in duplicate. Peak thrombin generation and endogenousthrombin potential (ETP) were determined (Thrombinoscope Analysissoftware, version 3.0).

Flow model

Bloodwas collected from healthy volunteers (0.32% sodium citrate w/v), andplatelets were labeled by adding 1,19-dimethyl-3,3,39,39-tetramethylindodi-carbocyanine iodide (DiIC15) (2 mM).13 Blood was supplemented with Alexa

Figure 1. Antibodies to human fXII. (A) Schematic diagrams comparing the

domain structures of fXII and its homolog HGFA. Arrowed numbers indicate the

locations of amino acid pairs that were used to create splice sites for introduction

of HGFA domains into fXII to create fXII/HGFA chimeras. (B) Western blots of

nonreduced human (H) and baboon (B) plasma size-fractionated by sodium dodecyl

sulfate-PAGE. The primary anti-factor XII antibodies used for detection are indicated

at the top of each panel. (C) Western blots of wild-type human fXII (FXII) and human

fXII with the first or second epidermal growth factor (EGF1 or EGF2), fibronectin

type 1 (Fib-1), kringle (KNG), or proline-rich (ProR) domains replaced with the corres-

ponding HGFA domain. Primary antibodies are indicated to the left of each blot. Poly,

Polyclonal goat IgG against human fXII.

1740 MATAFONOV et al BLOOD, 13 MARCH 2014 x VOLUME 123, NUMBER 11

For personal use only.on April 29, 2017. by guest www.bloodjournal.orgFrom

594–labeledfibrinogen(20mg/mL)and4mManti-fXII IgG,and incubated for30minutes at 37°C prior to use. Glass capillary tubes (0.23 2.03 50mm;VitroCom) were coated with 100mg/mL type I fibrillar collagen (Chrono-Log)overnight at 4°C, then blocked with 0.5% BSA. Blood was perfusedthrough tubes at an initial shear rate of 300 s21 using a syringe pump. Prior toentering the capillary tube, blood was mixed with 20 mM Tris-HCl pH 7.4,154 mM NaCl with 37.5 mM CaCl2, 19.8 mMMgCl2 via a second pump andpassed through a coiled 12-cmmixing tube.Blood is diluted;20%by this step,with final free [Ca21] and [Mg21];2.5 and 1.2 mM, respectively. Tubes weresubsequently perfused with 3.5% formaldehyde/PBS solution and imaged bylaser-scanning microscopy, using a Zeiss LSM 710 microscope.

Capillary occlusion assay

Glass35 capillary tubes (0.23 2 mm; VitroCom) were incubated for 1 hour atRT with fibrillar collagen (100 mg/mL), washed with PBS, blocked with5 mg/mL denatured BSA for 1 hour, then placed in a vertical position. Thetop of the tube was connected to reservoir, and the bottom was immersedin PBS. Human blood (0.32% sodium citrate w/v) supplemented with7.5 mM CaCl2 and 3.75 mMMgCl2 was added to the reservoir. Blood flowsthrough the tube under the force of gravity. The height of the sample reservoirwas maintained to produce an initial shear rate of 300 s21.

Mouse thrombosis model

fXII-deficient (fXII2/2) C57Bl/6 mice were anesthetized with pentobarbital.PBS (100mL)with orwithout 10mg of human fXII, andwith orwithout 100mgof anti-fXII IgG, was infused into the right jugular vein. Thrombus formationwas induced in the right carotid artery by applying 3.5% ferric chloride (FeCl3),asdescribed.8Arterial bloodflowwasmonitored for30minutes using aDopplerflow probe (Model 0.5 VB; Transonic System). Studies with mice were ap-proved by the Institutional Animal Care and Use Committee of VanderbiltUniversity.

Baboon thrombosis model

Nonterminal studies were performed on 2 male baboons (Papio anubis) withexteriorized femoral arteriovenous shunts, as described.8,12,13 Thrombusformation was initiated by deploying a thrombogenic graft (Figure 2) into theshunt for 60 minutes. The graft comprised a 20 3 4 mm ePTFE (Gortex)segment coatedwithcollagen, a2034mmsilicon rubber linker, anda2039mmsilicon rubber expansion chamber. Flow through the shunt was restricted to100mL perminute, producing an initial wall shear rate in the graft of 265 s21.Platelet deposition in the graft and expansion chamber was assessed in realtime by quantitative imaging of 111In-labeled platelet accumulation usinga GE-400T g scintillation camera interfaced to a NuQuest InteCam computersystem. Endpoint fibrin deposition was determined by direct measurement

of 125I-labeled fibrinogen, as described.8,12,13 Plasma thrombin-antithrombin(TAT) complex levelsweremeasuredwith anEnzygnostTATELISA(Siemens).Studies with baboons were approved by the Institutional Animal Care and UseCommittee of Oregon Health and Sciences University.

Results

Anti-fXII antibodies

Antibodies 9A2 and 15H8 recognize fXII in human and baboonplasma (Figure 1B) on western blots. The fXII gene arose froma duplication of the HGFA gene.33,36 fXII and HGFA have similardomain structures, except that fXII has a proline-rich region not foundin HGFA (Figure 1A). We prepared fXII with individual domainsreplaced by corresponding HGFA domains.With the exception of thefXII/HGFA-fibronectin type II domain chimera, all proteins weresecreted by a human fibroblast line (Figure 1C top). 9A2 and 15H8recognize distinct epitopes on fXII,with 9A2binding to thefibronectintype I and/or EGF2 domains (Figure 1C, middle), and 15H8 to theEGF2 /kringle domains (Figure 1C, bottom).

9A2 and 15H8 prolong the aPTT of human plasma (Figure 3A),with 15H8 having a greater effect. Antibodies at ;2 to 3 times theplasma fXII concentration achieved maximum inhibition. Combining9A2 and 15H8 produced a greater degree of inhibition (Figure 3B),consistent with the 2 antibodies recognizing distinct epitopes. Thecurve (open circles) in Figure 3C shows the relationship between fXIIconcentration and the aPTT. Using this curve for comparison, theeffect of 15H8 on the aPTT of normal plasma corresponds to.95%inhibition of fXII activity,while 9A2 achieves;50% reduction. 15H8prolonged the aPTT of baboon plasma (Figure 3D) to a greater degreethan human plasma (Figure 3A). The curve with closed circles inFigure 3C was prepared by mixing baboon plasma with fXII-deficienthuman plasma.When data in Figure 3D are compared with this curve,

Figure 2. Schematic diagram of the arteriovenous shunt and thrombogenic

device used to study thrombosis in baboons. Thrombogenic devices contain

a 20-mm long segment of ePTFE graft tubing (4-mm diameter) coated with collagen

and a 20-mm expansion chamber made of silicon rubber tubing (9-mm diameter,

20-mm length) downstream of the collagen-coated segment. The collagen-coated

segment and expansion chamber are connected by a 20-mm linker made of

uncoated silicon tubing (4-mm diameter). Flow through the arterio-venous shunt is

adjusted to 100 mL per minute by a flow restrictor, producing an initial shear rate in

the collagen-coated portion of the thrombogenic device of 265 s21.

Figure 3. Effects of anti-fXII antibodies on plasma coagulation. (A) Results of

a standard aPTT assay using a silica-based reagent for normal human plasma

supplemented with different concentrations of IgG 9A2 (s) or 15H8 (d). Data are

averages of 2 clotting times. (B) aPTT assay of normal human plasma supplemented

with control vehicle (C), 4 mM 9A2 or 15H8, or 4 mM of both antibodies. Each circle

represents a single clotting time, and the bar indicates the mean for the group. (C)

aPTT results for human fXII-deficient plasma mixed with normal human (s) or

baboon (d) plasma in various ratios. (D) Effect of different concentrations of IgG

15H8 on the aPTT of baboon plasma.

BLOOD, 13 MARCH 2014 x VOLUME 123, NUMBER 11 fXII AND THROMBOSIS 1741

For personal use only.on April 29, 2017. by guest www.bloodjournal.orgFrom

it appears that 15H8 inhibits .99% of the fXII activity in baboonplasma. 9A2 did not prolong the aPTT of baboon plasma.

Effects of anti-fXII antibodies on fXII and PK activation in vitro

fXII undergoes autoactivation in the presence of a variety of surfacesand polymers.9,14,19 We tested the capacity of anti-fXII antibodies toinhibit fXII autoactivation in the presence of a silica-based aPTT re-agent (Figure4A)orpolyphosphate (Figure4B).Bothantibodies reducedfXII activation with the aPTT reagent, with 15H8 having a greatereffect, while both had roughly similar effects with polyphosphate.

When PK is incubated with fXIIa in solution, a-kallikrein isformed (Figure 4C, open circle). The rate of PK activation by fXIIa isenhanced slightly by adding HK (Figure 4C, open square) or aPTTreagent (Figure 4C, filled square), while adding both significantlyincreases activation (Figure 4C, filled triangle). This is consistent withthe well-described surface-dependent cofactor effect of HK on PKactivation by fXIIa. 9A2 has little effect on surface-dependent PKactivation by fXIIa (Figure 4D, open triangle), while 15H8 reduced therate of the reaction (Figure 4D, open diamond).

Effects of anti-fXII antibodies on thrombin generation in plasma

Addition of aPTT reagent to normal plasma leads to a burst ofthrombin generation (Figure 5A, ETP 1805 nM 3 minutes) that isalmost completely blocked by 15H8. 9A2 reduces ETP by ;50%(961 nM 3 minutes), with a delay in peak thrombin generation.Similar results were obtained using collagen to induce coagulation(Figure 5B). In contrast, neither antibody blocks thrombin generationinduced by adding fXIIa to fXII-deficient plasma supplemented with

aPTT reagent (Figure 5C). These data indicate that 15H8 and 9A2mostly inhibited fXII autoactivation to fXIIa while 15H8, but not9A2, also reduces fXIIa activation of PK. Neither antibody affectssurface-dependent fXI activation by fXIIa in plasma appreciably.

Anti-fXII antibodies in flow models

Anti-fXI antibodies inhibit fibrin formation in recalcified humanblood perfused across collagen-coated surfaces.13 Figure 6A showsimages from collagen-coated tubes perfused with human blood ata shear rate of 300 s21. Platelet aggregates appear green and fibrinstrands orange.The anti-fXI antibodyO1A613 blocksfibrin generationin this system. 9A2 and 15H8 substantially reduce fibrin deposition,although some fibrin does form. These data indicate that the anti-fXIIantibodies have an effect in flowing human blood that is similar to theeffect reported for anti-fXI antibodies.13 9A2and 15H8also prolongedthe time it takes for blood to occlude collagen-coated capillary tubesin which flow is induced by gravity (Figure 6B). At a concentration(1.3 mM) ;3.5-fold higher than the plasma fXII concentration, 9A2increased time to occlusion twofold, while 15H8 increased it nearlythreefold.

Anti-fXII antibodies in a murine thrombosis model

Exposing blood vessels in mice to FeCl3 results in changes to theblood vessel endothelium that lead to thrombus formation in a fXII-and fXI-dependent manner.7,8,37 fXII-deficient C57Bl/6 mice do notdevelop carotid artery occlusion when exposed to 3.5% FeCl3, whilewild-type mice reproducibly develop occlusion in 10 to 15 minutes.8

Infusing human fXII into fXII-deficient animals to raise the plasmalevel to;20% of normal restores the wild-type phenotype (n5 5, allwith vessel occlusion). Coadministration of fXII and a 10-fold molarexcess of 9A2 reduced the incidence of arterial occlusion to 50%,while 15H8 prevented arterial occlusion (n5 6 for each antibody).

Anti-fXII antibodies in a baboon thrombosis model

We tested the effects of 15H8 on platelet (Figure 7A) and fibrin(Figure 7B) deposition in thrombogenic devices deployed intoarteriovenous shunts in 2 baboons (Figure 2). Thrombus formation istriggered in the collagen-coated segment of the graft where the initialwall shear rates is ;265 s21. A distal expansion chamber made ofsilicon rubber is incorporated to assess thrombus propagation underlower shear (,30 s21; Figure 2). 15H8 (5-6 mg/kg IV) prolonged theaPTT from 29.5 to 50 seconds in 1 baboon, and from 33.5 to 78seconds in a second animal. Based on the curve in Figure 3C (blackcircles), these results indicate substantial (;99%) inhibition of fXIIactivity. The inhibitory effect lasted.24 hours. Results were obtainedfor 9 thrombogenic devices tested before, and 4 devices tested after,15H8 administration. Devices were tested sequentially, and at least 30minutes were allowed to lapse between removal of a device andplacement of a new device. Data and means are shown in Figure 7,however, the design of the study is not conducive to a detailedstatistical analysis.15H8 did not affect platelet deposition (Figure 7A)within the collagen-coated graft, had a modest effect in the linkerdownstream from the collagen, and caused a substantial reduction(;80%) compared with control in the expansion chamber. 15H8reducedfibrin deposition by 70%6 5% in the collagen-coated portionof the graft (Figures 7B, left panel), and by 95% 6 1% in the linker-expansion chamber (Figure 7B, right panel). The grafts promotethrombin generation that is detectable in the systemic circulation asa TAT complex.8,13 15H8 reduced TAT levels by;50% (Figure 7C).

Figure 4. Effects of anti-fXII antibodies on fXII and PK. fXII activation: conversion

of fXII to fXIIa in the presence of (A) a silica-based aPTT reagent or (B) poly-

phosphate (2 mM) and vehicle (s), 9A2 (d), 15H8 (4), or the combination of 9A2

and 15H8 (▼). Results are means of 3 separate runs 6 1 standard deviation (SD).

Prekallikrein activation: (C) PK (50 nM) was incubated in reaction buffer containing

250 mM CS-3102 at RT, in the absence (d) or presence (s) of fXIIa (1 nM).

Cleavage of CS-3102 was monitoring by following changes in OD 405 nm. fXIIa at

the concentration used does not cleave CS-3102 at an appreciable rate in the

absence of PK (not shown). Addition of HK (N, 70 nM) or aPTT reagent (■, 5% v/v)

to the reaction with PK and fXIIa had modest effects on the rate of activation, while

addition of both aPTT reagent and HK (:) had a greater effect. (D) The effects of

vehicle (:) 100 nM 9A2 (4) and 15H8 (♢) on activation of PK by fXIIa in the presence of

aPTT reagent (5% v/v) and HK (70 nM).

1742 MATAFONOV et al BLOOD, 13 MARCH 2014 x VOLUME 123, NUMBER 11

For personal use only.on April 29, 2017. by guest www.bloodjournal.orgFrom

Discussion

Anticoagulants currently used for treating or preventing thrombo-embolism directly inhibit thrombin or factor Xa activity, or limitproduction of their precursors. Although effective, this strategy

increases bleeding risk because the targeted proteases are central tohemostasis. This places limits on the types of patients who can betreated safely with anticoagulants, and the clinical scenarios inwhichtreatment is applied. The intuitive notion that thrombosis reflects“hemostasis in the wrong place” has been brought into question bydata from rodent models demonstrating prothrombotic roles for theproteases fXIa and fXIIa.6-10,38-40 These observations suggest that itmay be possible to develop therapies in which antithrombotic effectsare largely or completely dissociated from antihemostatic effects.

Both fXIa and fXIIa have features that make them attractivetherapeutic targets. There is substantial evidence supporting a rolefor fXI in human thrombosis. Plasma fXI levels at the upper end ofthe normal range increase risk for myocardial infarction,26 stroke,29

and venous thromboembolism (VTE)25,28 relative to the remainder ofthe population, while severe fXI deficiency reduces incidence ofstroke27 and VTE.30 The major function of fXIa, activation of fIX,appears to serve a limited role in hemostasis, primarily directed atpreventing excessive trauma-induced bleeding in tissues with highfibrinolytic activity such as the oropharynx and urinary tract.14,15

In patients with severe fXI deficiency, some types of surgery41

and normal child birth42 are associated with relatively low rates ofexcessive bleeding in the absence of factor replacement. Indeed, manyfXI-deficient individuals do not experience abnormal hemostasis, andsymptomatic patients rarely bleed spontaneously (with the exceptionof menorrhagia),14,15 indicating that drugs targeting fXIa would beassociated with less bleeding than drugs that inhibit thrombin or fXa.The absence of a bleeding diathesis in fXII-deficient individualssuggests that drugs specifically targeting this protein would notcompromise hemostasis, allowing them to be used in patients with themost restrictive contraindications for current anticoagulation thera-pies. Enthusiasm for developing fXIIa inhibitors, however, istempered by 2 considerations. First, while numerous functions areattributed to fXIIa, the physiologic roles of the protease are in-completely understood. Perhaps as important, a clear link between fXIIand thrombosis in humans is not established.

Anecdotal reports suggesting that fXII deficiency actuallypredisposes to VTE date back to the death of the first personidentifiedwith severe fXII deficiency from a pulmonary embolism.43

Subsequent investigation did not confirm an association betweenfXII levels and VTE,44,45 and an analysis of case reports concludedthatmost thrombotic events in fXII-deficient patients are unrelated tothe deficiency.46 However, 2 recent studies have returned the issue offXII and thrombotic risk to the forefront. Doggen et al reported aninverse relationship between plasma fXII levels and risk ofmyocardialinfarction,26 with an odds ratio of 0.4 for individuals in the highestquartile for fXII level compared with those in the lowest quartile. Thisstudy examined fXII within the broad normal range, and not theconsequences of severe fXII deficiency, which may be more relevant

Figure 5. Effects of anti-fXII antibodies on throm-

bin generation. Shown are the effects of 4 mM 9A2,

15H8, or vehicle (V) on thrombin generation in normal

plasma triggered with (A) 1% v/v aPTT reagent or (B)

100 mg/mL type I collagen. No thrombin is generated

in the absence of aPTT reagent or collagen. (C)

Thrombin generation in fXII-deficient plasma supple-

mented with 5 nM fXIIa in the presence of 500 nM 9A2,

15H8, or vehicle (V). –XIIa indicates that no thrombin

was generated in the absence of fXIIa.

Figure 6. Effect of anti-fXII antibodies on fibrin formation in human blood

under flow. (A) Immunofluorescent images (Zeiss LSM 710, objective lenses:

203/0.80 Plan-Apochromat, 320 magnification) showing the effects of the anti-fXI

IgG O1A6 (300 nM) or the anti-fXII IgGs 9A2 and 15H8 (4 mM) on fibrin deposition

over time in recalcified human blood flowing across collagen-coated surfaces with

an initial average shear rate of 300 s21. Direction of flow is indicated at the bottom of

the image. Fibrin appears orange in these images and platelet aggregates appear

green. (B) Collagen-coated glass capillary tubes were perfused with recalcified

human blood driven by a constant pressure gradient under the force of gravity.

Shown are times to capillary occlusion in the presence of varying concentrations of

9A2 (N) or 15H8 (n). Each bar represent means for 3 separate measurements

6 standard error.

BLOOD, 13 MARCH 2014 x VOLUME 123, NUMBER 11 fXII AND THROMBOSIS 1743

For personal use only.on April 29, 2017. by guest www.bloodjournal.orgFrom

for anticipating effects of therapeutic fXII inhibition. Endler et al alsoobserved an inverse relationship between plasma fXII and all-causemortality,31withparticipantswith10% to20%of the normal fXII levelhaving a hazard ratio of 4.7 compared with those with fXII .100%of normal. Curiously, there was no significant increase in mortalityfor subjects with fXII levels of 1% to 10% of normal, suggestinga fundamental difference between severe and moderate deficiency.Data from the studies fromDoggen et al and Endler et al seem at oddswith work showing that elevated plasma fXIIa is associated with anincreased risk of coronary events.47,48 It is difficult to draw unifyingconclusions from this data, but there seems to be grounds for concernthat fXII may not contribute to thrombosis in humans and rodents inthe same manner, or to the same extent.

The current study examined the contribution of fXII in human/primate models that require fXI for normal thrombus formation. Anassumption here is that fXIIa contributes to thrombosis through fXIactivation, and ultimately thrombin generation, although other fXIIa-mediated activities could be involved. For example, fXIIa can havedirect effects on fibrin structure that could contribute to thrombusstability.49 Antibodies 9A2 and 15H8 recognize epitopes on the fXIIheavy chain. They reduce fXII activity in plasma in which contactactivation is triggered by inhibiting fXII activation, and not fXIIaactivity, supporting work showing the fXII heavy chain is requiredfor binding to polyanions.50-52 In baboons, 15H8 reduced fibrindeposition in thrombogenic grafts, and limited platelet-rich thrombusgrowth downstream from the collagen-coated graft segment. Theantibody reduced thrombin-antithrombin complex levels induced bythe grafts, consistent with the premise that a reduction in thrombingeneration was behind the antithrombotic effect. It is illustrative tocompare this performance to those of anti-fXI antibodies in thismodel.

The antibody O1A6 is a potent inhibitor of factor IX activation byfXIa.13,32 O1A6 reduces systemic TAT levels by $80%, and sig-nificantly reduces platelet andfibrin accumulationwithin the collagen-coated segments of grafts. Platelets adhere to collagen in the presenceof O1A6, but there is a defect in 3-dimensional thrombus growth,13

consistent with the thrombus instability observed in fXI- and fXII-deficient mice.7,8 The considerably greater effect of O1A6 comparedwith 15H8maybe explained by its larger effect on thrombin generation.The anti-fXI antibody 14E11 inhibits fXI activation by fXIIa, but doesnot affect fXIa activity.8 Similar to 15H8, 14E11 had relatively littleeffect on platelet accumulation in the collagen-coated graft segment, butdid limit downstream platelet deposition. The performance of theseantibodies in human blood in a collagen-based flow system are, ingeneral, consistent with those for the baboon model. Taken as a whole,the data suggest that inhibiting fXI activation by fXIIa, either by in-hibiting fXIIa generation (15H8) or by blocking fXIIa activation offXI (14E11) can inhibit thrombus formation by reducing thrombingeneration. The effect, however, is not as pronounced as the oneproduced by inhibiting fXIa activation of factor IX (O1A6).Interestingly, O1A6,13 14E118 and 15H8 prolong the aPTT to similarextents in treated baboons, demonstrating that it is the mechanismtargeted, and not the absolute value of the aPTT, that determines theantithrombotic effect. The primate models used in this study (thebaboon and human ex vivoflowmodels) are collagen-based.Althoughmechanistic details for the process(es) responsible for fXII activationcannot be definitively established in these models, collagen maycontribute directly to fXII activation through a process similar oridentical to contact activation. The findings of this study, therefore,may only be applicable to surface-induced thrombosis, such as a clotassociated with an intravascular catheter, and may not be applicable to

Figure 7. Effect of 15H8 on platelet and fibrin deposition in a baboon arteriovenous shunt thrombosis model. Thrombogenic devices depicted in Figure 2 were

inserted into femoral arteriovenous shunts in olive baboons as described.9,13,14 Flow through the grafts was maintained at 100 mL per minute, producing an initial average wall

shear rate of 265 s21 within the 4-mm diameter portions of the graft. (A) Platelet accumulation in the collagen-coated, silicon linker, and silicon expansion chamber segments

of grafts was assessed in real time by imaging of local 111In-labeled platelet accumulation using a GE-400T camera with NuQuest InteCam. The curves composed of closed

circles (d) with error bars (61 SD) represent mean values for 9 devices inserted into arterio-venous shunts in 2 untreated animals (control results). Individual results for 4

devices tested in the same 2 animals at least 1 hour after administration of anti-fXII antibody 15H8 (5-6 mg/kg IV) are indicated by the symbols s, 4, ,, and ♢. (B) Endpoint

determinations of total 125I-labeled fibrin deposition during the experiments in panel A. Fibrin deposition in the collagen-coated graft segment (left panel) and silicon expansion

chamber (right panel) were determined for 8 of the 9 devices inserted before animals received 15H8 (d) and for the 4 devices tested after animals received 15H8 (s). Large

bars indicate mean values and smaller bars6 1 SD (C) Plasma TAT complex measured in blood obtained at various times after graft insertion from the arteriovenous shunt upstream

of the site of graft insertion for 6 of 8 control grafts (d), and the 4 grafts placed after 15H8 administration (s). Large bars indicate mean values and smaller bars 6 1 SD.

1744 MATAFONOV et al BLOOD, 13 MARCH 2014 x VOLUME 123, NUMBER 11

For personal use only.on April 29, 2017. by guest www.bloodjournal.orgFrom

thrombus formation triggered by other stimuli. However, it should benoted that 15H8 is effective in preventing FeCl3-induced thrombosis inmice. Here, thrombus formation is initiated by changes in vascularendothelium that promote red blood cell and platelet adhesion,37 withexposure of subendothelial collagen probably playing a relativelyminor role.

The results suggesting that fXII inhibition has a smaller anti-thrombotic effect than fXI inhibition in primates contrast with thoseobtained with mouse thrombosis models. We observed thatfXII-deficient mice are somewhat more resistant to carotid arterythrombosis induced by FeCl3 or laser injury than are fXI-deficientmice.8 Thrombin-mediated feedback activation of fXI may explainthe observation that fXII deficiency does not cause a hemorrhagictendency.34 Perhaps this mechanism plays a more prominent role inthrombus formation in primates than in mice, accounting for thegreater effectiveness of O1A6 compared with 15H8 and 14E11 inbaboons. This scenario is consistentwith themoremodest reduction inTAT levels in baboons treated with 15H8 compared with O1A6.13

Alternatively, it appears that relatively small amounts of fXII havesignificant effects on the aPTT in primate plasmas. Therefore, a highdegree of fXII inhibition (not easily obtainedwith an antibody)may berequired to produce an antithrombotic effect. Indeed, while 15H8reduced fXII activity in the aPTT, it did not block it completely.Consistent with this, Pixley and coworkers reported that an antibodythat neutralized ;60% of fXII activity in baboons did not affectendotoxin-induced disseminated intravascular coagulation.53 In con-trast, reducing fXI levels in baboons by as little as 50% affects throm-bus formation in the arteriovenous shunt model.54 Furthermore,inhibiting fXII with an antibody to produce a similar effect to total fXIIdeficiency is made difficult by the relatively high plasma fXII con-centration (400 nM). The plasma fXI concentration is only 30 nM.These observations have implications for developing therapeutic fXII/XIIa inhibitors,whichmayneed to inhibit a highpercentageof proteaseactivity to produce a therapeutic effect.

Acknowledgments

Wegratefully acknowledge JenniferGreisel’s expert technical supportwith the primate studies, andDrWilliamDupont for recommendationson the use of statistical analysis.

The authors acknowledge support from National Institutes of Healthawards (Heart, Lung and Blood Institute) HL81326, HL58837 (D.G.),HL106919 (I.M.V.), HL047014 (J.H.M.), and HL101972 (A.G.,O.J.T.M.); (National Institute of Allergy and Infectious Diseases)AI088937 (A.G., E.I.T.); (National Institute of Neurological Disordersand Stroke) NS077600 (P.Y.L.); UL1TR000128 to the Oregon Clinicaland Translational Research Institute, and RR000163 to the OregonNational Primate Research Center from the National Institutes ofHealth.

Authorship

Contribution: A.M. performed flow model studies, designed experi-ments for characterization of effects of antibodies on fXII, andcontributed to thewriting of themanuscript; P.Y.L. performed baboonstudies and contributed to antibody development and to writing of themanuscript; A.E.G. and S.L.G. characterized the effects of antibodieson fXII activation and fXIIa activity in vitro; C.P. conducted capillaryflow studies;Q.C. designed and conductedmurine thrombosis studies;M-f.S. prepared the fXII/HGFAmutants and conducted studies tomapthe binding sites of 9A2 and 15H8 on fXII; O.J.T.M. contributedto design and interpretation of flow model experiments; E.I.T.contributed to antibodygeneration and characterization, and tobaboonthrombosis studies; H.K. contributed to the design of XII/HGFAchimeras; T.R. contributed to the design ofmurine thrombosismodels,and to the writing of the manuscript; J.H.M. contributed to the designof experiments with poly-P, and to the writing of themanuscript; A.G.oversaw studies involving baboons, generation of antibodies, and thewriting of themanuscript; andD.Gwas responsible for oversight of theproject and preparation of the final manuscript.

Conflict-of-interest disclosure: D.G. is a consultant, and receivesconsultant’s fees from several pharmaceutical companies. P.Y.L.,E.I.T., and A.G. are employees of the company Aronora. A.M.,P.Y.L., A.G., D.G., Oregon Health and Sciences University, andVanderbilt University may have financial interest in the results of thisstudy. The remaining authors declare no competing financial interests.

Correspondence: David Gailani, Hematology/Oncology Divi-sion, Vanderbilt University, 777 Preston Research Building, 2220Pierce Ave, Nashville, TN 37232; e-mail: dave.gailani@vanderbilt.edu.

References

1. Schumacher WA, Luettgen JM, Quan ML, SeiffertDA. Inhibition of factor XIa as a new approach toanticoagulation. Arterioscler Thromb Vasc Biol.2010;30(3):388-392.

2. Lowenberg EC, Meijers JC, Monia BP, Levi M.Coagulation factor XI as a novel target forantithrombotic treatment. J Thromb Haemost.2010;8(11):2349-2357.

3. Woodruff RS, Sullenger B, Becker RC. The manyfaces of the contact pathway and their role inthrombosis. J Thromb Thrombolysis. 2011;32(1):9-20.

4. Stavrou E, Schmaier AH. Factor XII: what does itcontribute to our understanding of the physiologyand pathophysiology of hemostasis & thrombosis.Thromb Res. 2010;125(3):210-215.

5. Renne T, Schmaier AH, Nickel KF, Blomback M,Maas C. In vivo roles of factor XII. Blood. 2012;120(22):4296-4303.

6. Wang X, Cheng Q, Xu L, et al. Effects of factor IXor factor XI deficiency on ferric chloride-induced

carotid artery occlusion in mice. J ThrombHaemost. 2005;3(4):695-702.

7. Renne T, Pozgajova M, Gruner S, et al. Defectivethrombus formation in mice lacking coagulationfactor XII. J Exp Med. 2005;202(2):271-281.

8. Cheng Q, Tucker EI, Pine MS, et al. A role forfactor XIIa-mediated factor XI activation inthrombus formation in vivo. Blood. 2010;116(19):3981-3989.

9. Muller F, Mutch NJ, Schenk WA, et al. Plateletpolyphosphates are proinflammatory andprocoagulant mediators in vivo. Cell. 2009;139(6):1143-1156.

10. Kleinschnitz C, Stoll G, Bendszus M, et al.Targeting coagulation factor XII providesprotection from pathological thrombosis incerebral ischemia without interfering withhemostasis. J Exp Med. 2006;203(3):513-518.

11. Nickel K, Frantz S, Esmon C, Gailani D, Renne T.Targeting the polyphosphate-coagulation factorXII pathway prevents cardiac ischemia-reperfusion injury without influencing bleeding

risk. In: Proceedings from the InternationalSociety on Thrombosis and Haemostasis; June29-July 4, 2013; Amsterdam, The Netherlands.Abstract OC53.5.1.

12. Gruber A, Hanson SR. Factor XI-dependenceof surface- and tissue factor-initiated thrombuspropagation in primates. Blood. 2003;102(3):953-955.

13. Tucker EI, Marzec UM, White TC, et al.Prevention of vascular graft occlusion andthrombus-associated thrombin generation byinhibition of factor XI. Blood. 2009;113(4):936-944.

14. Gailani D, Neff AT. Rare coagulation factordeficiencies, In: Hoffman RH, Benz EJ, SilbersteinLE, Heslop H, Weitz JI , Snastasi J, eds.Hematology, Basic Principles and Practice. 6thed. Philadelphia, PA: Saunders-Elsevier; 2013:1971-1986.

15. Seligsohn U. Factor XI deficiency in humans.J Thromb Haemost. 2009;7(suppl 1):84-87.

BLOOD, 13 MARCH 2014 x VOLUME 123, NUMBER 11 fXII AND THROMBOSIS 1745

For personal use only.on April 29, 2017. by guest www.bloodjournal.orgFrom

16. MacFarlane RG. An enzyme cascade in theblood clotting mechanism, and its function asa biochemical amplifier. Nature. 1964;202:498-499.

17. Davie EW, Ratnoff OD. Waterfall sequence forintrinsic blood clotting. Science. 1964;145(3638):1310-1312.

18. van der Meijden PE, Munnix IC, Auger JM, et al.Dual role of collagen in factor XII-dependentthrombus formation. Blood. 2009;114(4):881-890.

19. White-Adams TC, Berny MA, Patel IA, et al.Laminin promotes coagulation and thrombusformation in a factor XII-dependent manner.J Thromb Haemost. 2010;8(6):1295-1301.

20. Kannemeier C, Shibamiya A, Nakazawa F,et al. Extracellular RNA constitutes a naturalprocoagulant cofactor in blood coagulation. ProcNatl Acad Sci U S A. 2007;104(15):6388-6393.

21. Fuchs TA, Brill A, Duerschmied D, et al.Extracellular DNA traps promote thrombosis. ProcNatl Acad Sci USA. 2010;107(36):15880-15885.

22. Maas C, Govers-Riemslag JW, Bouma B, et al.Misfolded proteins activate factor XII in humans,leading to kallikrein formation without initiatingcoagulation. J Clin Invest. 2008;118(9):3208-3218.

23. Shariat-Madar Z, Mahdi F, Schmaier AH.Assembly and activation of the plasma kallikrein/kinin system: a new interpretation. IntImmunopharmacol. 2002;2(13-14):1841-1849.

24. Adams GN, LaRusch GA, Stavrou E, et al. Murineprolylcarboxypeptidase depletion inducesvascular dysfunction with hypertension and fasterarterial thrombosis. Blood. 2011;117(14):3929-3937.

25. Meijers JC, Tekelenburg WL, Bouma BN, BertinaRM, Rosendaal FR. High levels of coagulationfactor XI as a risk factor for venous thrombosis.N Engl J Med. 2000;342(10):696-701.

26. Doggen CJ, Rosendaal FR, Meijers JC. Levelsof intrinsic coagulation factors and the risk ofmyocardial infarction among men: opposite andsynergistic effects of factors XI and XII. Blood.2006;108(13):4045-4051.

27. Salomon O, Steinberg DM, Koren-Morag N,Tanne D, Seligsohn U. Reduced incidence ofischemic stroke in patients with severe factor XIdeficiency. Blood. 2008;111(8):4113-4117.

28. Cushman M, O’Meara ES, Folsom AR, HeckbertSR. Coagulation factors IX through XIII and therisk of future venous thrombosis: the LongitudinalInvestigation of Thromboembolism Etiology.Blood. 2009;114(14):2878-2883.

29. Suri MF, Yamagishi K, Aleksic N, Hannan PJ,Folsom AR. Novel hemostatic factor levels andrisk of ischemic stroke: the Atherosclerosis Risk inCommunities (ARIC) Study. Cerebrovasc Dis.2010;29(5):497-502.

30. Salomon O, Steinberg DM, Zucker M, Varon D,Zivelin A, Seligsohn U. Patients with severe factorXI deficiency have a reduced incidence of

deep-vein thrombosis. Thromb Haemost. 2011;105(2):269-273.

31. Endler G, Marsik C, Jilma B, Schickbauer T,Quehenberger P, Mannhalter C. Evidence of aU-shaped association between factor XII activityand overall survival. J Thromb Haemost. 2007;5(6):1143-1148.

32. Geng Y, Verhamme IM, Messer A, et al. Asequential mechanism for exosite-mediated factorIX activation by factor XIa. J Biol Chem. 2012;287(45):38200-38209.

33. Miyazawa K, Shimomura T, Kitamura A, Kondo J,Morimoto Y, Kitamura N. Molecular cloning andsequence analysis of the cDNA for a humanserine protease reponsible for activation ofhepatocyte growth factor. Structural similarity ofthe protease precursor to blood coagulation factorXII. J Biol Chem. 1993;268(14):10024-10028.

34. Kravtsov DV, Matafonov A, Tucker EI, et al.Factor XI contributes to thrombin generation in theabsence of factor XII. Blood. 2009;114(2):452-458.

35. Puy C, Tucker EI, Wong ZC, et al. Factor XIIpromotes blood coagulation independent of factorXI in the presence of long-chain polyphosphates.J Thromb Haemost. 2013;11(7):1341-1352.

36. Ponczek MB, Gailani D, Doolittle RF. Evolution ofthe contact phase of vertebrate blood coagulation.J Thromb Haemost. 2008;6(11):1876-1883.

37. Barr JD, Chauhan AK, Schaeffer GV, Hansen JK,Motto DG. Red blood cells mediate the onset ofthrombosis in the ferric chloride murine model.Blood. 2013;121(18):3733-3741.

38. Colman RW. Are hemostasis and thrombosis twosides of the same coin? J Exp Med. 2006;203(3):493-495.

39. Hagedorn I, Schmidbauer S, Pleines I, et al.Factor XIIa inhibitor recombinant human albuminInfestin-4 abolishes occlusive arterial thrombusformation without affecting bleeding. Circulation.2010;121(13):1510-1517.

40. Chen JW, Figueiredo JL, Wojtkiewicz GR, et al.Selective factor XIIa inhibition attenuates silentbrain ischemia: application of molecular imagingtargeting coagulation pathway. JACC CardiovascImaging. 2012;5(11):1127-1138.

41. Salomon O, Steinberg DM, Seligshon U. Variablebleeding manifestations characterize differenttypes of surgery in patients with severe factor XIdeficiency enabling parsimonious use ofreplacement therapy. Haemophilia. 2006;12(5):490-493.

42. Salomon O, Steinberg DM, Tamarin I, Zivelin A,Seligsohn U. Plasma replacement therapy duringlabor is not mandatory for women with severefactor XI deficiency. Blood Coagul Fibrinolysis.2005;16(1):37-41.

43. Ratnoff OD. The George M. Kober lecture.The legacy of John Hageman: new dividends.Trans Assoc Am Physicians. 1985;98:cli-clxi.

44. Koster T, Rosendaal FR, Briet E, VandenbrouckeJP. John Hageman’s factor and deep-veinthrombosis: Leiden thrombophilia Study. Br JHaematol. 1994;87(2):422-424.

45. Zeerleder S, Schloesser M, Redondo M, et al.Reevaluation of the incidence of thromboemboliccomplications in congenital factor XIIdeficiency—a study on 73 subjects from 14Swiss families. Thromb Haemost. 1999;82(4):1240-1246.

46. Girolami A, Randi ML, Gavasso S, Lombardi AM,Spiezia F. The occasional venous thrombosesseen in patients with severe (homozygous) FXIIdeficiency are probably due to associated riskfactors: a study of prevalence in 21 patients andreview of the literature. J Thromb Thrombolysis.2004;17(2):139-143.

47. Grundt H, Nilsen DW, Hetland O, Valente E,Fagertun HE. Activated factor 12 (FXIIa) predictsrecurrent coronary events after an acutemyocardial infarction. Am Heart J. 2004;147(2):260-266.

48. Siegerink B, Govers-Riemslag JW, RosendaalFR, Ten Cate H, Algra A. Intrinsic coagulationactivation and the risk of arterial thrombosis inyoung women: results from the Risk of ArterialThrombosis in relation to Oral contraceptives(RATIO) case-control study. Circulation. 2010;122(18):1854-1861.

49. Konings J, Govers-Riemslag JW, Philippou H,et al. Factor XIIa regulates the structure of thefibrin clot independently of thrombin generationthrough direct interaction with fibrin. Blood. 2011;118(14):3942-3951.

50. Pixley RA, Stumpo LG, Birkmeyer K, Silver L,Colman RW. A monoclonal antibody recognizingan icosapeptide sequence in the heavy chain ofhuman factor XII inhibits surface-catalyzedactivation. J Biol Chem. 1987;262(21):10140-10145.

51. Clarke BJ, Cote HC, Cool DE, et al. Mappingof a putative surface-binding site of humancoagulation factor XII. J Biol Chem. 1989;264(19):11497-11502.

52. Citarella F, te Velthuis H, Helmer-Citterich M,Hack CE. Identification of a putative binding sitefor negatively charged surfaces in the fibronectintype II domain of human factor XII—animmunochemical and homology modelingapproach. Thromb Haemost. 2000;84(6):1057-1065.

53. Pixley RA, De La Cadena R, Page JD, et al. Thecontact system contributes to hypotension but notdisseminated intravascular coagulation in lethalbacteremia. In vivo use of a monoclonal anti-factor XII antibody to block contact activation inbaboons. J Clin Invest. 1993;91(1):61-68.

54. Crosby JR, Marzec U, Revenko AS, et al.Antithrombotic effect of antisense factor XIoligonucleotide treatment in primates. ArteriosclerThromb Vasc Biol. 2013;33(7):1670-1678.

1746 MATAFONOV et al BLOOD, 13 MARCH 2014 x VOLUME 123, NUMBER 11

For personal use only.on April 29, 2017. by guest www.bloodjournal.orgFrom

online January 9, 2014 originally publisheddoi:10.1182/blood-2013-04-499111

2014 123: 1739-1746  

Morrissey, Andras Gruber and David GailaniCheng, Mao-fu Sun, Owen J. T. McCarty, Erik I. Tucker, Hiroaki Kataoka, Thomas Renné, James H. Anton Matafonov, Philberta Y. Leung, Adam E. Gailani, Stephanie L. Grach, Cristina Puy, Qiufang thrombosis modelFactor XII inhibition reduces thrombus formation in a primate 

http://www.bloodjournal.org/content/123/11/1739.full.htmlUpdated information and services can be found at:

(1065 articles)Thrombosis and Hemostasis    Articles on similar topics can be found in the following Blood collections

http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requestsInformation about reproducing this article in parts or in its entirety may be found online at:

http://www.bloodjournal.org/site/misc/rights.xhtml#reprintsInformation about ordering reprints may be found online at:

http://www.bloodjournal.org/site/subscriptions/index.xhtmlInformation about subscriptions and ASH membership may be found online at:

  Copyright 2011 by The American Society of Hematology; all rights reserved.of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society

For personal use only.on April 29, 2017. by guest www.bloodjournal.orgFrom