NZ619385B2 - Inhibitory anti -factor xii/xiia monoclonal antibodies and their uses - Google Patents

Abstract

Disclosed is an anti-Factor Xll/XIIa monoclonal antibody or antigen-binding fragment thereof that has a more than 2 fold higher binding affinity to human Factor XII a-beta than to human Factor XII and that is capable of inhibiting the amidolytic activity of human Factor Xlla, wherein the binding affinity is measured by surface plasmon resonance-based technology, and wherein inhibiting the amidolytic activity of human Factor Xlla means an inhibition of at least 80% of the activity observed in a control experiment without any inhibitor present. Also disclosed is an anti-Factor XII/XIIa monoclonal antibody or antigen-binding fragment thereof that has a more than 2 fold higher binding affinity to human Factor XIIa-beta than to human Factor XII, and that inhibits the amidolytic activity of Factor XIIa-alpha by more than 50% in an in vitro amidolytic activity assay when used at a molar ratio of FXIIa-alpha to antibody of 1:0.2. inity is measured by surface plasmon resonance-based technology, and wherein inhibiting the amidolytic activity of human Factor Xlla means an inhibition of at least 80% of the activity observed in a control experiment without any inhibitor present. Also disclosed is an anti-Factor XII/XIIa monoclonal antibody or antigen-binding fragment thereof that has a more than 2 fold higher binding affinity to human Factor XIIa-beta than to human Factor XII, and that inhibits the amidolytic activity of Factor XIIa-alpha by more than 50% in an in vitro amidolytic activity assay when used at a molar ratio of FXIIa-alpha to antibody of 1:0.2.

Description

INHIBITORY ANTI -FACTOR XII/XIIA ONAL ANTIBODIES AND THEIR USES The ion relates to inhibitory anti—factor Xll/FXIla antibodies and methods of their use.

Factor XII (Hageman Factor) is a serum glycoprotein with a molecular weight of about 80 kDa. s an autoactivation by exposure to negatively charged surfaces, factor Xll is additionally activated by kallikrein by proteolytic cleavage to form alpha—factor Xlla, which is then further converted, for example by n, into beta-factor Xlla (FXlla-B). Alpha-factor Xlla is composed of the inal heavy chain of about 50 kDa, which contains the contact g domain, and the C- terminal light chain of about 28 kDa, which contains the catalytic center. The heavy and light chains are connected by a disulfide bond. FXlla-B is an active form of FXll of about 30 kDa, consisting of the complete light chain and a 2000 Da fragment of the heavy chain linked by a disulfide bond.

Vessel wall injury triggers sudden adhesion and aggregation of blood platelets, followed by the activation of the plasma coagulation system and the formation of fibrin—containing thrombi, which occlude the site of injury. These events are crucial to limit post—traumatic blood loss but may also occlude diseased vessels leading to ischemia and infarction of vital organs. in the all model, blood coagulation proceeds by a series of reactions involving the tion of zymogens by limited proteolysis culminating in tion of thrombin, which ts plasma fibrinogen to fibrin and activates platelets. in turn, collagen- or fibrin-adherent platelets facilitate thrombin generation by several orders of magnitude via exposing procoagulant olipids (mainly phosphatidyl serine) on their outer surface, which propagates assembly and activation of coagulation protease complexes and by direct interaction between platelet receptors and coagulation factors.

Two converging pathways for coagulation exist that are triggered by either extrinsic (vessel wall) or intrinsic (blood—borne) components of the vascular system. The "extrinsic" pathway is initiated by the x of the plasma factor VII (FVII) with the integral membrane protein tissue factor (TF), an essential coagulation or that is absent on the luminal e but strongly expressed in subendothelial layers of the vessel and which is accessible or liberated via tissue . TF expressed in circulating microvesicles might also contribute to thrombus propagation by sustaining in generation on the surface of activated platelets. The "intrinsic" or contact activation pathway is initiated when factor Xll (FXII, Hageman factor) comes into contact with negatively charged es in a reaction ing high molecular weight kininogen and plasma kallikrein. FXII can be activated by macromolecular constituents of the subendothelial matrix such as glycosaminoglycans and ens, sulfatides, nucleotides, polyphosphates and other soluble polyanions or non-physiological material such as glass or polymers.

One of the most potent contact activators is kaolin and this on serves as the mechanistic basis for the major clinical clotting test, the activated partial oplastin time (aPTT), which measures the coagulation capacity via the "intrinsic" pathway. In reactions propagated by platelets, activated FXll then activates FXI to FXIa and subsequently FXIa activates factor IX. The complex of FVllla, which FVllla has been previously activated by traces of FXa and/or thrombin, and FIXa (the tenase complex) uently activates FX.

Despite its high potency to induce blood clotting in vitro, the (patho—) physiological significance of the FXII-triggered intrinsic coagulation pathway is questioned by the fact that hereditary deficiencies of FXlI as well as of high molecular weight kininogen and plasma kallikrein are not associated with bleeding complications.

Together with the observation that humans and mice lacking extrinsic pathway tuents such as TF and FVII suffer from severe bleeding this has led to the current hypothesis that the cessation of bleeding in vivo requires exclusively the sic cascade (Mackman, N. 2004. Role of tissue factor in hemostasis, thrombosis, and vascular development. Arteriosc/er. Thromb. Vasc. Biol. 24, 101 5- 1 022).

In pathological conditions, the coagulation cascade may be activated inappropriately which then results in the formation of haemostatic plugs inside the blood vessels. Thereby, vessels can be occluded and the blood supply to distal organs limited. This process is known as thrombosis, and, if the thrombus embolizes, as thromboembolism which is associated with high mortality. in addition, the use of prosthetic devices, which come into contact with blood, is severely limited because of activation of the intrinsic coagulation cascade. Suitable g of the prosthetic surface may avoid said problem in some cases but may compromise its function in others. Examples of such prosthetic devices are hemodialysers, cardiopulmonary bypass circuits, heart valves, vascular stents and in-dwelling catheters. in cases where such s are used, anticoagulants, such as heparin, are administered to prevent fibrin formation on the surface. However, some patients are intolerant of heparin, which can cause heparin-induced thrombocytopenia (HIT) resulting in et aggregation and |ife~threatening thrombosis. Furthermore, an inherent antage of all anticoagulants used in clinics is an increased risk of serious bleeding events. Therefore, a strong need for new types of anticoagulants exist, which are not associated with such cations and that can be used in affected patients or as superior prophylaxis/ therapy concept preventing thrombosis without increased bleeding risks.

For more than five decades it has been known that deficiency of coagulation factor Xll is not associated with sed spontaneous or -related bleeding cations (Ratnoff OD & Colopy JE 1955. A familial hemorrhagic trait associated with a deficiency of a clot-promoting fraction of . J Clin Invest 34:602-613). Indeed, although readily detected by a pathological value measured in the aPTT (a clinical clotting test that addresses the intrinsic pathway of ation) humans that are deficient in FXll do not suffer from abnormal ng even during major surgical ures (Colman RW. Hemostasis and Thrombosis.

Basic principles & clinical practice (eds. Colman RW, Hirsch J, Mader VJ, Clowes AW, & George J) 103—122 (Lippincott Williams & Wilkins, Philadelphia, 2001)). in contrast, deficiency of FXll had been associated with increased risk of venous thrombosis (Kuhli C et al. 2004. Factor XII deficiency: a thrombophilic risk factor for l vein occlusion. Am. J. Ophthalmol. 137:459-464; Halbmayer WM at al. 1993.

Factor XII (Hageman factor) deficiency: a risk factor for development of thromboembolism. Incidence of FXll deficiency in patients after recurrent venous or arterial thromboembo/ism and myocardial infarction. Wien. Med. Wochenschr. 143:43-50). Studies and case reports supporting this idea refer to the index case ’IO for FXll deficiency, Mr. John Hageman, who died of pulmonary sm. The hypothesis that FXll deficiency is associated with an increased prothrombotic risk is challenged by a recent reevaluation of several case reports the original reports of which linked FXll deficiency with thrombosis ami A et al. 2004. The occasional venous thromboses seen in ts with severe ygous) FXll deficiency are probably due to associated risk factors: A study of prevalence in 21 patients and review of the literature. J. Thromb. Thrombolysis 17:139-143). In most cases the authors identified concomitant congenital or acquired prothrombotic risk factors in ation with factor FXll deficiency that could be responsible for the thrombotic event independently of FXll. The t epidemiological studies using well characterized patients (Koster T et al. 1994. John Hageman’s factor and deep—vein thrombosis: Leiden thrombophilia Study. Br. J. ol. 87:422-424) and FXll- deficient families (Zeer/eder S et al. 1999. Reeve/uation of the incidence of thromboembolic complications in ital factor XII deficiency — a study on 73 subjects from 14 Swiss families. Thromb. Haemost. 82:1240-1246) indicated that there is no ation of FXll deficiency and any pro- or rombotic risk.

Surprisingly and in contrast to common believe of those skilled in the art it has been discovered that the factor XII—driven intrinsic coagulation pathway is ed in arterial us formation in vivo but is not necessary for normal —specific hemostasis (Renne T et al. 2005. Defective thrombus formation in mice lacking factor Xll. J. Exp. Med. 202:271-281; Kleinschnitz C et al. 2006. Targeting coagulation factor XII provides protection from pathological thrombosis in cerebral ischemia without interfering with hemostasis. J. Exp. Med. 203, 513-518; W02006066878). Unexpectedly, these results place factor Xll in a central position in the process of pathological thrombus formation. Hence substances capable of interfering and blocking FXIl activation or FXll activity may be suited to block pathogenic arterial thrombus formation and the clinical consequences thereof.

In 066878 the use of dies against Xlla or the use of inhibitors of FXII/FXlla is proposed. As potential inhibitors antithrombin lll (AT ill), angiotensin converting enzyme inhibitor, Cf inhibitor, aprotinin, alpha-i protease inhibitor, in ([(S)—l—CarboxyPhenylethyl]—Carbamoyl~L-Arg—L-Val-Arginal), Z-Pro— Proaldehyde-dimethyl acetate, DX88 (Dyax lnc., 300 Technology Square, Cambridge, MA 02139, USA; cited in: Williams A and Baird LG.2003. DX-88 and HAE: a developmental perspective. Transfus Apheresis Sci. 29:255-258), tin, inhibitors of prolyl oligopeptidase such as Fmoc—Ala-Pyr-CN, corn-trypsin inhibitor, mutants of the bovine pancreatic trypsin tor, ecotin, yellowfin sole anticoagulant protein, Cucurbita maxima trypsin inhibitor-V including Curcurbita maxima isoinhibitors and Hamadarin (as disclosed by lsawa H et al. 2002. A mosquito salivary protein inhibits activation of the plasma contact system by binding to factor XII and high molecular weight kininogen. J. Biol. Chem. 277:27651-27658) have been proposed.

An ideal tor of FXll/FXlla as a therapeutic agent - while exhibiting a high inhibitory activity towards FXll/FXlla - will not increase the risk of bleeding, be nonimmunogenic and have to be administered as sparingly as possible ~ ideally only once. Small molecule inhibitors like Z~Pro-Pro-aIdehyde—dimethyl acetate will have only a very short ife after administration, thus ing le injections, or would have to be developed into orally available slow release forms and then also be given constantly over a long period. Human plasma proteins like Ci inhibitor would at first sight fulfill all requirements, having a relatively high inhibitory activity towards FXll/FXlla while not increasing the risk of ng, being non- immunogenic as a human n and also having a considerably long plasma half— life. It was now surprisingly found that in an in vivo model of thrombosis 0‘] inhibitor as a prime candidate of a human FXll/FXIla inhibitor could not be used sfully to t occlusion. Another proposed FXll/FXlla inhibitor from human plasma namely AT lll inhibitor would at least not fulfill the second ement as the bleeding risk would increase (Warren BL et al. 2001. Caring for the critically ill patient. High-dose antithrombin III in severe sepsis: a randomized controlled trial.

JAMA 286:1869—1878). ln W02008098720A1 the use of KazaI—type serine protease inhibitor lnfestin or ‘IO domains thereof or modified Kazal-type serine protease inhibitors based on lnfestin homologs as inhibitors of FXll/ FXlla is proposed. Selected from this subset, recombinant lnfestin—4 fused to human albumin for prolongation of half-life (rHA- in—4) was developed demonstrating high inhibitory activity towards XIIa.

Moreover this substance trated antithrombotic efficacy without impairing (physiologic) asis while demonstrating a useful half-life after fusion to human albumin (Hagedorn et al. 2010. Factor Xl/a tor Recombinant Human Albumin lnfestin-4 Abolishes Occlusive Arterial Thrombus Formation Without Affecting Bleeding. Circulation. 121:1510-1517). However, although immunogenicity was reduced during development, there is still the risk of immunogenic responses in man. Furthermore, an even longer half-life would have additional beneficial effects.

Hence, it is apparent that there still exists a need for an improved tion for the treatment and/or prophylaxis of thrombosis and similar disorders. Therefore, the present invention provides an improved medication to satisfy such a need. A candidate for such an improved medication is an improved anti—FXll/FXlla dy with inhibitory activity.

Antibodies to Factor Xll have been disclosed. Pixley et al (J Biol Chem (1987) 262, 10140-10145) disclosed monoclonal antibody B7C9 to human Factor Xll. This antibody blocked surface-mediated coagulant activity, but not amidolytic activity of Factor Xlla. Small et al (Blood (1985), 65, 202-210) disclosed a onal antibody to human Factor Xll, which ted activation of Factor Xll, but not the coagulant or the amidolytic activity of activated FXll ). Nuijens et al (J, Biol.

Chem. (1989) 264, 12941-12949) disclosed monoclonal antibodies F1 and F3, which inhibited ation activity but not amidolytic activity of FXll. W08911865 provides monoclonal antibodies ed against the light chain of FXll (BBFS, 0687, D2E10). These dies inhibit the coagulation activity, but only show partial inhibition of the amidolytic activity of FXlla. W09008835 describes the production of monoclonal antibody that selectively binds FXlla-B over FXll, and the development of an immunoassay that specifically detects FXlla-B in blood. From example 7 in W09008835, it is clear that the antibody does not inhibit amidolytic ty of FXlla. W09117258 describes the treatment of sepsis with an anti—FXll antibody OT—2, which binds to native FXll in plasma, and inhibits activation of the contact system in plasma, as well as amidolytic activity of FXlla.

The present invention provides the development of an improved antibody which — while exhibiting a high tory activity towards FXlla - will not se the risk of bleeding, be non—immunogenic and have a long half-life. Since FXll has a multidomain structure including ectin type and EGF—like domains wed by Stavrou and Schmaier (2010) Thromb. Res, 125:210—215), it was believed that FXll should have additional important physiologic functions in addition to its role as FXlla, i.e. as the enzyme following activation. . New studies have demonstrated now that FXll contributes to cell proliferation and growth leading to angiogenesis (reviewed by Schmaier and LaRusch (2010) Thromb. Haemost., 104:915-918).

Therefore, in order not to interfere with these (and maybe other so far unknown) functions of FXll, it is preferable for a therapeutic antibody against Xlla to have a clear higher affinity towards FXlla, for example towards FXlla-B, compared to FXll.

Summary of the Invention One aspect of the ion is therefore an anti-Factor Xll/FXlla monoclonal antibody or antigen-binding fragment thereof that has a more than 2 fold higher binding affinity to human Factor Xlla-beta than to human Factor Xll and that is capable of inhibiting the amidolytic activity of human Factor Xlla, wherein the g affinity is measured by e plasmon resonance-based logy, and wherein completely inhibiting the amidolytic ty of human Factor Xlla means an inhibition of at least 80% of the activity observed in a control experiment without any tor present. Another aspect of the invention is an anti—Factor XII/Xlla monoclonal antibody or antigen-binding fragment thereof, that ts human Factor Xlla—alpha by more than 50% when used at a molar ratio of FXlla-alpha to antibody of 120.2.

Another aspect of the invention is a ned antibody of the t invention. A further aspect of the invention is a bispecific antibody comprising one Fab region of an antibody of the present invention.

Another aspect of the invention is use of an antibody or antigen—binding fragment f of the invention for the manufacture of a medicament for the treatment and/or prevention of a er selected from the group consisting of venous, al or capillary thrombus formation, thrombus formation in the heart, thromboembolism, thrombus formation during and/or after contacting blood of a human or animal subject with artificial surfaces, by preventing the formation and/or the stabilization of thrombi and y three-dimensional intraluminal thrombus growth, or by preventing and/or treating intraluminal thrombi; interstitial lung disease, inflammation, a neurological inflammatory disease, complement activation, fibrinolysis, angiogenesis and diseases related to FXll/ FXlla-induced kinin formation or FXlI/FXlla-mediated complement activation.

A further aspect of the invention is use of an antibody or antigen-binding fragment thereof of the invention for the manufacture of a medicament for prevention or treatment of a condition associated with increased retinal vascular permeability, -8A- including progressive retinopathy, threatening complication of retinopathy, macular edema, non-proliferative retinopathy, proliferative retincpathy, retinal edema, diabetic retinopathy, hypertensive retinopathy, and retinal trauma.

Preferably, the antibody or antigen-binding fragment thereof has one or more of the ing features: it binds murine FXll/FXlla; the level of binding of the antibody to a polypeptide comprising SEQ ID NO: 2 or relevant fragment thereof in which (a) the asparagine residue at position 398 of SEQ ID NO: 2 is substituted for lysine; or (b) the isoleucine residue at position 438 of SEQ ID NO: 2 is substituted for alanine, is lower than the level of binding of the protein to the corresponding polypeptide comprising SEQ ID NO: 2 or relevant fragment f without said substitution; lt comprises a heavy chain variable (vH) region which is more than 85% identical to the sequence of SEQ ID NO: 4; it comprises a light chain variable (vL) region which is more than 85% identical to the sequence of SEQ ID NO: 5; it comprises heavy chain CDRl at least 80% cal to the sequence of SEQ ID NO: 6, and/or heavy chain CDRZ at least 60% identical with SEQ ID NO: 7, and/or heavy chain CDR3 at least 80% identical to the ce of SEQ ID NO: 9; it comprises light chain CDRl at least 50% cal with SEQ ID NO: 11, and/or light chain CDR2 of SEQ lD NO: 12, and/or light chain CDR3 with the sequence A-X1-W-X2-X3-X4-X5—R—X6—X7 wherein X1 can be A or S, X5 can be L or V, the other an can be any amino acid (SEQ ID NO: 14). it binds human Factor Xlla—beta with a KB of better than 10‘8M. it competes with lnfestin, in particular with lnfestin-4, for binding to human Factor Xlla-beta. it is a human lgG or variant thereof, preferably human lgG4 or t thereof.

Another aspect of the invention is a nucleic acid encoding the antibody, or antigen- binding nt thereof, of the invention.

Yet another aspect of the ion is a vector sing the nucleic acid encoding the antibody, or antigen-binding fragment thereof, of the invention, operably linked to a suitable promoter ce.

A further aspect of the invention is a cell line or yeast cell comprising the vector of the invention.

Another aspect of the invention is a method of producing the antibody or antigen binding fragment thereof of the invention, comprising culturing the cell line or yeast cell of the invention under appropriate conditions to express the antibody or antigen g fragment thereof, and purifying the antibody or antigen binding fragment f from the culture supernatant.

Yet another aspect of the invention is the antibody or n binding fragment f for medical use.

A further aspect of the invention is the antibody or antigen binding fragment thereof for use in preventing and/or treating a disorder selected from the group consisting of venous, arterial or capillary thrombus formation, thrombus formation in the heart, thrombus formation during and/or after contacting blood of a human or animal subject with artificial surfaces, thromboembolism, by preventing the formation and/or the stabilization of thrombi and thereby three-dimensional uminal thrombus growth, or by ting and/or treating intraluminal thrombi; titial lung disease, inflammation, a neurological inflammatory disease, complement activation, fibrinolysis, angiogenesis and diseases related to FXll/ FXlla—induced kinin formation or FXll/FXlla-mediated complement activation. Yet another aspect of the ion is the antibody or antigen-binding fragment thereof for use in the treatment of intraluminal i in a human or animal subject related to a disorder selected from the group consisting of venous, arterial or capillary us formation, thrombus formation in the heart, thrombus formation during and/or after contacting blood of a human or animal subject with artificial surfaces, thromboembolism; interstitial lung disease, inflammation, a neurological inflammatory disease, complement activation, fibrinolysis, enesis and diseases related to FXll/FXlla~induced kinin formation or FXlI/FXlla—mediated complement activation. Preferably, the venous or arterial thrombus formation is stroke, myocardial infarction, deep vein thrombosis, portal vein thrombosis, renal vein thrombosis, jugular vein thrombosis, cerebral venous sinus thrombosis, Budd~ Chiari syndrome or Paget-Schroetter disease. Preferably, the diseases related to FXll/FXlla-induced kinin formation are ed from the group hereditary angioedema, bacterial infections of the lung, trypanosoma infections, hypotensive shock, atitis, chagas disease, articular gout, arthritis, inated ascular coagulation (DIC) and sepsis.

Preferably the interstitial lung disease is fibroproliferative and/or idiopathic ary fibrosis.

Preferably, the thrombus formation occurs during and/or after contacting blood of a human or animal subject with artificial surfaces during and/or after a medical procedure med on said human or animal subject and said antibody or n-binding fragment f is administered before and/or during and/or after said medical procedure, and further (i) the artificial surface is exposed to at least 80% of the blood volume of the subject and the artificial surface is at least 0.2 m2 or (ii) the artificial surface is a container for collection of blood outside the body of the subject or (iii) the artificial surface is a stent, valve, intraluminal er, or a system for internal assisted pumping of blood.

Yet a further aspect of the invention is a medical device coated with the antibody or antigen-binding fragment thereof of the ion, wherein the device is a cardiopulmonary bypass machine, an extracorporeal membrane oxygenation system for oxygenation of blood, a device for assisted pumping of blood, a blood dialysis device, a device for the extracorporeal filtration of blood, a repository for use in the collection of blood, an intraluminal catheter, a stent, an artificial heart valve, and/or accessories for any one of said devices including , cannulae, centrifugal pump, valve, port, and/or diverter. r aspect of the invention is the antibody or antigen—binding fragment thereof for use for administration in a patient receiving a medical procedure, n the medical procedure comprises contact with at least one of: (a) heart, (b) at least one blood vessel chosen from: the aorta, the aortic arch, a carotid , a coronary artery, brachiocephalic artery, vertebrobasilar circulation, ranial arteries, renal artery, a c artery, a mesenteric artery, and/or a blood vessel of the arterial system cranial to the heart, (0) a venous blood vessel if the patient has a known septal defect; and wherein the medical procedure comprises release of at least one embolus in at least one of said blood vessels in the body that could result in ischemia in at least one target organ and administration of the antibody or antigen binding nt thereof , during and/or after the medical procedure.

Another aspect of the ion is the antibody or antigen binding fragment thereof for use in the prevention or treatment of a condition associated with increased vascular permeability, in particular increased l vascular permeability, including progressive retinopathy, sight-threatening complication of retinopathy, macular edema, non-proliferative pathy, proliferative retinopathy, retinal edema, diabetic retinopathy, hypertensive retinopathy, and retinal trauma.

Another aspect of the ion is a pharmaceutical composition comprising the antibody or antigen binding fragment thereof of the invention.

Brief Description of the Figures Figure 1: Anti—FXlla phage ition ELISA using the FXlla amidolytic inhibitor infestin4. The concentrations of the competitor (rHA-lnf4) are shown on the X-axis.

Fixed concentrations of phage—expressed Fab antibody or infestin4 (pTaclnf4) used in the assay were ined using a phage ion ELlSA.

Figure 2: Concentration-dependent tion of amidolytic activity of human FXlla by monoclonal antibody 3F7 as a fully human lgG4. The anti-human GCSF receptor monoclonal antibody (31.2 (fully human lgG4) was used as a negative control and rHA-lnfestin as a positive control for the assay.

Figure 3: 3F7 heavy chain stop templates used for affinity maturation. CDR regions are shaded grey and amino acid positions in each library that were randomised are designated as u n Figure 4: 3F7 light chain stop templates used for affinity maturation. CDR regions are shaded grey and amino acid positions in each library that were randomised are designated as a ll Figure 5: Concentration-dependent inhibition of amidolytic activity of human FXlla by monoclonal antibodies 3F7 and OT—2.

Figure 6: A: Alignment of the catalytic s of FXll of mouse, rat and human, and identification of the residues that form the catalytic triad (*) and the mutations introduced (i) to identify the potential epitope of antibody 3F7. B: Western Blot 3O g the binding of SF? to the various mutants.

Figure 7: ion rate in FeClg-induced osis following treatment with MAb 3F? (n=5-25/group) Figure 8: Effect of MAb 3F7 on aPTT (n=5-25/group; meaniSD) Figure 9: Effect of MAb 3F? on PT (n=5~25/group; meaniSD) Figure 10: Effect of MAb 3F7 on FXlla-activity (n=5-25/group; meaniSD) Figure 11: Effect of MAb SF? on time to hemostasis. Data are presented as mean values (+SD). Statistics: p>0.05 (Kruskal-Wallis test). N=10/group.

Figure 12: Effect of MAb 3F? on total blood loss. Data are presented as mean values (+SD). tics: p>0.05 (Kruskal-Wallis test). N=10/group.

Figure 13: Effect of MAb 3F7 on time to hemostasis. Horizontal lines represent median values. Statistics: p>0.05 (Kruskal-Wallis test). N=10/group.

Figure 14: Effect of MAb 3F7 on total blood loss. Horizontal lines represent median 2O values. Statistics: p>0.05 al-Wallis test). N=10/group.

Figure 15: Comparison of aPTT of OT—2, MAb 3F? and affinity—matured versions of MAb 3F7 Figure 16: Comparison of inhibition of human Factor Xlla-alpha by different antibodies List of Sequences SEQ lD NO: 1: Human FXll sequence SEQ lD NO: 2: Mouse FXll sequence SEQ ID NO: Rat FXII sequence SEQ ID NO: 3F7 vH sequence SEQ ID NO: 3F7 vL sequence SEQ ID NO: 3F7 heavy chain CDR1 (HCDR1) SEQ ID NO: SEQ ID NO: 0°39??? 3F7 heavy chain CDR2 (HCDR2) : 3F7 heavy chain CDR2 with variation SEQ ID NO: 9: 3F7 heavy chain CDR3 (HCDR3) SEQ ID NO: 10: 3F7 heavy chain CDR3 with variation SEQ ID NO: 11: 3F7 iight chain CDR1 (LCDR1) 1O SEQ ID NO: 12: 3F7 light chain CDR2 ) SEQ ID NO: 13: 3F7 iight chain CDR3 (LCDR3) SEQ ID NO: 14: 3F7 light chain CDR3 with variation SEQ ID NO: 15: 3F7 heavy chain stop template H1 SEQ ID NO: 16: Oligonucleotide mutagenic trimer mix 3F7 H1 SEQ ID NO: 17: 3F7 heavy chain stop template H2 SEQ ID NO: 18: ucieotide nic trimer mix 3F7 H2 SEQ ID NO: 19: 3F7 heavy chain stop template H3.1 SEQ ID NO: 20: Oligonucleotide mutagenic trimer mix 3F7 H3.1 SEQ ID NO: 21: 3F7 heavy chain stop template H32 SEQ ID NO: 22: Oligonucleotide mutagenic trimer mix 3F7 H32 SEQ ID NO: 23: 3F7 light chain stop template L1 SEQ ID NO: 24: Oligonucleotide mutagenic trimer mix 3F7 L1 SEQ ID NO: 25: 3F7 light chain stop template L3.1 SEQ ID NO: 26: Oligonucleotide mutagenic trimer mix 3F7 L3.1 SEQ ID NO: 27: 3F7 Iight chain stop template L32 SEQ ID NO: 28: Oligonucleotide mutagenic trimer mix 3F7 L32 SEQ ID NO: 29: VR119 heavy chain CDR2 SEQ ID NO: 30: VR112 heavy chain CDR2 SEQ ID NO: 31: VR115 heavy chain CDR2 SEQ ID NO: 32: VR11O heavy chain CDR2 SEQ ID NO: 33: VR107 heavy chain CDR2 SEQ ID NO: 34: VR108 heavy chain CDR2 SEQ ID NO: 35: VR103 heavy chain CDRZ SEQ ID NO: 36: VR101 heavy chain CDRZ SEQ ID NO: 37: VR109 heavy chain CDRZ SEQ ID NO: 38: VR99 heavy chain CDR2 SEQ ID NO: 395 VR149 heavy chain CDR3 SEQ ID NO: 40: VR167 heavy chain CDR3 SEQ ID NO: 41: VR148 heavy chain CDR3 SEQ ID NO: 42: VR159 heavy chain CDR3 IO SEQ ID NO: 43: VR16O heavy chain CDR3 SEQ ID NO: 44: VR24 light chain CDR1 SEQ ID NO: 45: VROB light chain CDRI SEQ ID NO: 46: VR16 light chain CDR1 SEQ ID NO: 47: VR05 light chain CDR‘I SEQ ID NO: 48: VR12 light chain CDRI SEQ ID NO: 49: VR10 light chain CDRl SEQ ID NO: 50: VR14 light chain CDR1 SEQ ID NO: 51: VR17 light chain CDR1 SEQ ID NO: 52: VR31 light chain CDR3 SEQ ID NO: 53: VR29 light chain CDR3 SEQ ID NO: 54: VR27 light chain CDR3 SEQ ID NO: 55: VR39 light chain CDR3 SEQ ID NO: 56: VR46 light chain CDR3 SEQ ID NO: 57: VR41 light chain CDR3 SEQ ID NO: 58: VR38 light chain CDR3 SEQ ID NO: 59: VR58 light chain CDR3 SEQ ID NO: 60: VR62 light chain CDR3 SEQ ID NO: 61: VR53 light chain CDR3 SEQ ID NO: 62: VR52 light chain CDR3 3O SEQ ID NO: 63: VR63 light chain CDR3 SEQ ID NO: 64: cing primer CH1 Rev SEQ ID NO: 65: Sequencing primer pLacPCwa SEQ ID NO: 66: Sequencing primerwt Gill stump rev SEQ ID NO: 67: Sequencing primer KpaCwad SEQ ID NO: 68: cing primer LdaCwad SEQ ID NO: 69: Sequencing primer PUCrev SEQ ID NO: 70: Sequencing primer 3254 SEQ ID NO: 71: Sequencing primer Seq CL lambda SEQ ID NO: 72: Sequencing primer Seq CH1 SEQ ID NO: 73: vH sequence of VR115 SEQ ID NO: 74: vH sequence of VR112 SEQ ID NO: 75: vL sequence of VR24 SEQ ID NO: 76: vH sequence of VR11O SEQ ID NO: 77: vH sequence of VR119 Detailed description of the invention The present invention provides the development of an improved antibody which —- while exhibiting a high inhibitory activity s FXIla ~ will not increase the risk of ng, be non-immunogenic and have a long half—life.

One aspect of the invention is therefore an anti-Factor IIa monoclonal antibody or antigen-binding fragment thereof that has a more than 2 fold higher g affinity to human Factor Xlla, preferably to human Factor Xlla-beta, than to human Factor XII and that is capable of completely inhibiting the amidolytic activity of human Factor Xlla.

Another aspect of the invention is an antibody or antigen binding fragment thereof that has a more than 2 fold higher binding affinity to human Factor Xlla, preferably 3O to human Factor Xlla-beta, than to human Factor XII and that is capable of completely inhibiting the amidolytic activity of human Factor Xlla and that competes with an antibody comprising the sequences of SEQ ID NOS: 4 and 75 sed as lgG4 for the binding to FXll/FXlla.

Preferably the antibody or antigen binding fragment thereof has more than 3 fold, more preferably more than 4 fold, even more preferably more than 5 fold, more than 6 fold, more than 8 fold, more than 10 fold, more than 12 fold, more than 14 fold, more than 16 fold, most preferably more than 18 fold higher binding affinity to human Factor Xlla, ably to human FactoerIa—beta, than to human Factor Xll. 1O Preferably, the antibody or antigen-binding fragment thereof completely ts the amidolytic activity of FXlla at a concentration of less than 100 nM, more preferably less than 50 nM, even more preferably less than 40 nM, or even less than 30 nM.

Preferably the antibody or antigen—binding fragment thereof completely inhibits at a concentration of between 1 pM and 100 nM, more preferably at a concentration n 5 pM and 50 nM. Preferably the assay for the amidolytic activity of FXlla is carried out as described in Example 1(5).

Another aspect of the invention is an anti—Factor XII/FXlla monoclonal antibody or antigen—binding fragment f that inhibits Factor Xlla-alpha, ably human Factor Xlla-alpha, by more than 40%, preferably more than 50%, even more preferably more than 60%, when used at a molar ratio of alpha to antibody of 120.2. Alternatively, the antibody or antigen binding fragment thereof inhibits Factor lpha, preferably human Factor Xlla—alpha, by more than 80%, preferably more than 85%, more preferably more than 90%, at a molar ratio of alpha to antibody of 120.5; most preferably, the antibody or antigen~binding fragment thereof achieves complete inhibition of FXlla—alpha at a molar ratio of 120.5. Preferably the antibody or antigen-binding fragment thereof has an affinity to human FXlla that is at least comparable to antibody 3F7 disclosed herein.

Preferably, the antibody or antigen-binding nt thereof binds murine FXll/FXlIa; more preferably, the level of binding of the antibody to a polypeptide comprising SEQ ID NO: 2 or relevant fragment thereof in which (a) the asparagine residue at position 398 of SEQ ID NO: 2 is substituted for lysine; or (b) the isoieucine residue at position 438 of SEQ ID NO: 2 is substituted for alanine, is lower than the level of binding of the protein to the corresponding polypeptide comprising SEQ ID NO: 2 or relevant fragment thereof without said substitution. A relevant fragment of the polypeptide of SEQ ID NO: 2 comprises the catalytic center; examples are the light chain, FXIIa~beta, FXIIa-alpha, or the complete FXII.

Preferably, the antibody or antigen-binding fragment thereof comprises a heavy chain variable (vH) region which is more than 85% identical to the sequence of SEQ ID NO: 4, more preferably more than 88%, 90%, 92%, 93%, 94%, 95%, 96%, 97%, even more preferably 98%, or even 99% identical to the ce of SEQ ID NO: 4. Preferred embodiments of the invention are antibodies or antigen-binding fragments thereof comprising a heavy chain variable region with the sequence of SEQ ID NOs: 4, 73, 74, 76 or 77. ably, the antibody or antigen binding fragment thereof comprises a light chain variable (vL) region which is more than 85% cal to the sequence of SEQ ID NO: 5, more preferably more than 88%, 90%, 92%, 93%, 94%, 95%, 96%, 97%, even more preferably 98%, or even 99% cal to the sequence of SEQ ID NO: . Preferred embodiments of the ion are antibodies or antigen g fragments thereof comprising a light chain variable region with the sequence of SEQ ID NOS: 5 or 75.

Preferred embodiments of the invention are antibodies or n binding fragments thereof with a vH region described above combined with a vL region as described above. Most preferred are antibodies with the following vH/vL combinations: (a) A vH region of SEQ ID NO: 4 combined with a vL region of SEQ ID NO: 5 or SEQ ID NO: 75; (b) A vH region of any of SEQ ID NOS: 4, 73, 74, 76 or 77 combined with a vL region of SEQ ID NO: 5.

Preferably, the antibodies or antigen binding nts f comprise heavy chain CDR1 at least 80% identical to the sequence of SEQ ID NO: 6, preferably heavy chain CDR1 of SEQ lD NO: 6, and/or heavy chain CDR2 at least 60% identical to the sequence of SEQ ID NO: 7, and/or heavy chain CDR3 at least 80% identical to the sequence of SEQ lD NO: 9. More preferably, heavy chain CDR2 has the sequence GlX1X2X3X4X5X3TVYADSVKG (see SEQ ID NO: 8) wherein X1 is R, N or D, X2 is P, V, | or M, X3 is S, P or A, X4 is G, L, V, or T, X5 can be any amino acid, preferably X5 is G, Y, Q, K, R, N or M, and X3 is T, G, or S, and/or heavy chain CDR3 has the sequence ALPRSGYLX1X2X3X4YYYYALDV (see SEQ ID NO: 10), wherein X1 is I, M or V, X2 is S or K, X3 is P, K, T or H, and X4 is H, N, G, or Q.

Preferably, the antibodies or n binding fragments thereof comprise light chain CDR1 at least 50% identical with SEQ ID NO: 11, and/or light chain CDR2 of SEQ ID NO: 12, and/or light chain CDR3 with the sequence AX1WX2X3X4X5RX3X7 (shown in SEQ lD NO: 14), wherein X1 is A or S, X5 is L or V, X5 is G, L, or K, and X2, X3, X4 and X7 can be any amino acid, preferably X2 is D, Y, E, T, W, E or S, X3 is A, N, l, L, V, P, Q, or E, X4 is S, D, P, E, Q, or R, and X7 is V, A, D, T, M, or G.

Preferred embodiments of the invention are antibodies or antigen binding fragments f with the heavy chain CDRs bed above combined with the light chain CDRs as described above.

More preferably, the antibodies or antigen binding fragments thereof comprise the combinations of heavy chain CDRs ) and light chain CDRs (LCDRs) shown in Table 1, wherein the numbers in the columns underneath HCDR1, HCDR2, HCDR3, LCDR1, LCDR2 and LCDR3 are the respective SEQ ID NOs: 12 _6 7 9 11 12 13 E-6 29 9 11 12 13 1-6 so 9 n 12 _6 31 9 11 12 13 VR24 6 7 9 44 12 13 _6 32 9 11 12 13 _6 33 9 11 12 13 VR06 6 7 9 45 12 13 VR31 6 7 9 11 12 52 VR108 6 34 9 11 12 13 VR103 6 35 9 11 12 13 VR101 6 36 9 11 12 13 VR16 6 7 9 46 12 13 ___53 VROS _13 VR12 1—3 VR27 _9 11 12 54 VRIO 6 7 9 49 12 13 VR149 6 7 39 11 12 13 VR58 6 7 9 11 12 59 VR39 6 7 9 11 12 55 VR167 6 7 40 11 12 13 VR62 6 7 9 11 12 60 VR14 mo ——13 vmv —__13 R63 _—63 VR99 _13 VR38 6 7 9 11 12 58 Preferably, the antibody or n g fragments thereof of the invention binds human Factor Xlla—beta with a KB of better than 10'7M, more preferably better than 3x 10‘8M, more ably better than 10'8M, even more preferably better than 3x '9 M, most preferably 10'9M or even 0M. ably, the antibody or antigen binding fragment thereof of the invention competes with lnfestin, preferably with Intestin—4, for binding to human Factor Xlla- beta.

The antibody or antigen binding fragment thereof can be any isotype, including lgG, lgM, IgE, lgD, or lgA, and any subtype thereof. Preferably, the antibody or antigen binding fragment thereof of the invention is a human lgG or variant thereof, preferably human IgG4 or t thereof. Methods to switch the type of antibody are well known in the art. The nucleic acid molecule encoding the vH or VL region is isolated, and operatively linked to a nucleic acid sequence encoding a different cH or CL, respectively, from the constant region of a different class of immunoglobulin molecule.

The present disclosure encompasses ns and/or antibodies described herein comprising a constant region of an antibody. This includes antigen binding fragments of an antibody fused to a Fc.

Sequences of constant regions useful for producing the proteins of the present disclosure may be obtained from a number of different s. In some examples, the constant region or portion thereof of the protein is derived from a human antibody. The constant region or portion thereof may be derived from any antibody class, including lgM, lgG, lgD, lgA and lgE, and any antibody isotype, including lgG1, lgGZ, lgG3 and lgG4. in one example, the nt region is human isotype lgG4 or a stabilized lgG4 constant region.

In one example, the Fc region of the constant region has a d ability to induce effector function, e.g., compared to a native or ype human lgGi or lgGS Fc region. In one example, the effector function is antibody-dependent cell-mediated cytotoxicity (ADCC) and/or antibody—dependent cell-mediated phagocytosis (ADCP) and/or complement-dependent cytotoxicity (CDC). s for assessing the level of effector function of an Fc region containing protein are well known in the art.

In one example, the Fc region is an lgG4 Fc region (i.e., from an lgG4 nt ), e.g., a human lgG4 Fc region. Sequences of suitable lgG4 Fc regions will be apparent to the skilled person and/or ble in ally available databases (e.g., available from National Center for Biotechnology Information).

In one example, the constant region is a stabilized lgG4 constant region. The term “stabilized lgG4 constant region" will be understood to mean an lgG4 constant region that has been modified to reduce Fab arm exchange or the propensity to undergo Fab arm exchange or formation of a half-antibody or a propensity to form a half antibody. “Fab arm exchange" refers to a type of protein modification for human lgG4, in which an lgG4 heavy chain and attached light chain (half~molecule) is swapped for a light chain pair from another lgG4 molecule. Thus, lgG4 molecules may acquire two distinct Fab arms recognizing two distinct antigens (resulting in bispecific molecules). Fab arm ge occurs naturally in vivo and can be induced in vitro by purified blood cells or reducing agents such as reduced glutathione. A “half antibody” forms when an lgG4 antibody dissociates to form two molecules each containing a single heavy chain and a single light chain.

In one example, a stabilized lgG4 constant region comprises a proline at position 241 of the hinge region according to the system of Kabat (Kabat et at, Sequences of Proteins of logical Interest Washington DC United States ment of Health and Human Services, 1987 and/or 1991). This position corresponds to position 228 of the hinge region according to the EU numbering system (Kabat et a/., Sequences of Proteins of Immunological Interest Washington DC United States Department of Health and Human Services, 2001 and Edelman et al., Proc. Natl.

Acad. Sci USA, 63, 78-85, 1969). In human lgG4, this residue is generally a . Following substitution of the serine for proline, the lgG4 hinge region comprises a ce CPPC. in this regard, the skilled person will be aware that the “hinge region" is a proline—rich portion of an antibody heavy chain constant region that links the Fc and Fab regions that confers ty on the two Fab arms of an antibody. The hinge region includes cysteine residues which are involved in inter-heavy chain disulfide bonds. It is generally d as hing from Glu226 to Pr0243 of human lgG1 according to the numbering system of Kabat. Hinge regions of other lgG isotypes may be aligned with the lgG1 sequence by placing the first and last cysteine residues forming inter-heavy chain disulphide (S—S) bonds in the same positions (see for example W02010/080538).

Additional examples of stabilized lgG4 dies are antibodies in which arginine at position 409 in a heavy chain constant region of human lgG4 (according to the EU numbering system) is substituted with lysine, threonine, methionine, or leucine (e.g., as described in W02006/033386). The Fc region of the constant region may additionally or atively comprise a residue selected from the group consisting of: alanine, valine, e, isoleucine and leucine at the position corresponding to 405 (according to the EU numbering system). Optionally, the hinge region comprises a proline at position 241 (i.e., a CPPC sequence) (as described above). in another example, the Fc region is a region modified to have d effector function, i.e., a “non-immunostimulatory Fc region”. For example, the Fc region is an lgG1 Fc region comprising a substitution at one or more positions selected from the group consisting of 268, 309, 330 and 331. in another example, the Fc region is an lgG1 Fc region comprising one or more of the following changes E233P, L234V, L235A and deletion of 6236 and/or one or more of the following s A3276, A3308 and P3318 (Armour et al., Eur J Immunol. 29:2613—2624, 1999; 3O Shields et al., J Biol Chem. ."6591-604, 2001). Additional examples of non- immunostimulatory Fc regions are described, for example, in Dall'Acqua et a/., J ImmunoI. 177: 1129-1138, 2006; and/or Hezareh J 75: 12161-12168, 2001).

In another example, the Fc region is a chimeric Fc region, e.g., comprising at least one CH2 domain from an IgG4 antibody and at least one CH3 domain from an lgGi dy, wherein the Fc region comprises a substitution at one or more amino acid positions selected from the group consisting of 240, 262, 264, 266, 297, 299, 307, 309, 323, 399, 409 and 427 (EU numbering) (e.g., as described in W02010/085682). Exemplary substitutions include 240F, 262L, 264T, 266F, 297Q, 299A, 299K, 307P, 309K, 309M, 309P, 323F, 3998, and 427F.

The t disclosure also contemplates additional modifications to an antibody.

For example, the antibody comprises one or more amino acid tutions that increase the half-life of the protein. For example, the antibody comprises a Fc region comprising one or more amino acid substitutions that se the affinity of the Fc region for the neonatal Fc region (FcRn). For example, the Fc region has increased affinity for FcRn at lower pH, e.g., about pH 6.0, to facilitate Fc/FcRn binding in an endosome. In one example, the Fc region has increased affinity for FcRn at about pH 6 ed to its affinity at about pH 7.4, which facilitates the re- release of Fc (and therefore of PC region-comprising molecules) into blood following cellular recycling. These amino acid substitutions are useful for extending the half life of a protein, by reducing clearance from the blood.

Exemplary amino acid substitutions include T2500 and/or M428L or T252A, T2548 and T266F or M252Y, 8254T and T256E or H433K and N434F according to the EU numbering . Additional or alternative amino acid substitutions are described, for example, in U820070135620 or U87083784. 3O More ably, the antibody of the invention is a human lgGi or human lgG4, ered for enhanced binding to the human neonatal Fc receptor FcRn at a lower pH, e.g. pH 6, which leads to an increased half life of the antibody in human serum. Methods to screen for l Fc variants for optimizing FcRn binding have been described (e.g. Za/evsky et a/ (2010) Nature Biotech 28, 157-159).

Other preferred antibodies or antigen binding fragments thereof of the invention comprise mammalian immunoglobulin constant regions, such as the constant regions of mammalian isotypes such as lgG, lgM, lgE, IgD, or IgA, and any subtype thereof. Preferably, the antibody is a ian lgG, including mouse lgG, pig lgG, cow lgG, horse lgG, cat lgG, dog lgG and primate IgG or variants thereof.

These antibodies may be chimeric antibodies, where the human variable regions of the invention are combined with the nt region of the immunoglobulin of the selected species. Alternatively, the antibody or antigen binding fragments thereof may be ed by grafting the human CDR regions described herein into the framework residues from an globulin of the selected species.

Preferably the antibodies or antigen binding nts thereof of the invention are in their mature form, i.e. without the signal peptide; however, the antibodies or antigen binding fragments thereof including the signal peptides are also ed in the invention.

The antigen g fragment may be any fragment of an antibody of the invention that maintains the ability to bind FXlla. Preferred antigen binding fragments are an Fab fragment, an Fab’ fragment, an F(ab')2 fragment, an Fv fragment, 3 single chain antibody, a single chain Fv nt, a disulfide stabilized Fv protein, or a dimer of a single chain Fv fragment. Antibodies also ed in the invention are a chimeric antibody, a humanized antibody, a murinized antibody or a bispecific antibody. Methods for producing these fragments and dies are well known in the art (see for example, Harlow & Lane: Antibodies, A Laboratory Manual, Cold Spring Harbor Laboratory, 1988).

Also included in the invention is a fusion protein or a phage particle comprising the antigen binding fragment of the antibody of the invention. The antigen binding nt may, for example, be fused with human serum albumin or a variant f. The skilled person will be well aware of other proteins that can be used as fusion partners for antigen binding fragments. The antibody or antigen binding fragment thereof may also be fused to a tag, such as a hexa—Histidine tag. The tag may be provided with a cleavable linker e, so that it can be removed from the antibody or antigen binding fragment thereof when desired.

Another aspect of the ion is a nucleic acid encoding the antibody, or antigen- binding fragment thereof, of the invention. ably, the nucleic acid also comprises a region encoding a signal peptide, preferably the nucleic acid comprises a region encoding a signal peptide for the heavy chain and a region encoding a signal peptide for the light chain.

Nucleic acid molecules encoding the polypeptides ed by the invention can be readily produced by the skilled person, using the amino acid sequences ed, the genetic code and sequences available in public databases. In addition, a variety of functionally equivalent nucleic acids can be readily produced and are ore also included in the present invention. The nucleic acid molecules can be prepared by any suitable method, for e by direct chemical synthesis.

Methods for ing DNA are well known in the art.

Yet another aspect of the invention is a vector comprising the c acid encoding the antibody, or antigen-binding fragment thereof, of the invention, operably linked to a suitable er sequence or incorporated into a suitable expression te, which may include additional regulatory elements such as enhancer elements to increase expression levels. Preferably a strong promoter is used. For expression in E. coli, a promoter such as T7, lac, trp or lambda promoters may be used, preferably in conjunction with a ribosome binding site and a transcription termination . For mammalian cells, SV40, CMV or immunoglobulin promoters can be used to provide high expression levels. Preferably, the vector is a mammalian cell expression vector, more preferably a vector selected from Lonza’s GS SystemTM or Selexis c ElementsTM systems. Preferably, the vector also contains a selectable marker sequence such as gpt, neo, amp or hyg genes, and a gene amplification system such as glutamine synthetase or DHFR. Another preferred vector is a yeast expression vector, e.g. an sion vector optimized for Pichia pastoris. The vector may also be a viral vector, e.g. a vector based on vaccinia virus, adenovirus, or a retrovirus. The vector may also be a baculovirus for expression in insect cells.

A r aspect of the invention is a cell line or yeast cell comprising the vector of the invention. Preferably the cell line is a mammalian cell line, such as CHO, HEK293, MDCK, COS, HeLa, or myeloma cell lines such as NSO. Another embodiment is an insect cell line for use with a baculovirus, such as SF9 cells, SF21 cells, or HighFiveTM cells. Yet another cell is a yeast cell, such as Saccharomyces, e.g. S. cerevisiae, or Pichia pistoris. Bacterial host cells such as E. coli are also possible. Methods for introducing DNA into the respective host cells are well known in the art. For e, when the host cell is a mammalian cell line, techniques such as lipofection or oporation may be used.

Another aspect of the invention is a method of producing the antibody or antigen binding fragment thereof of the invention, sing culturing the host cells, such as the cell line or yeast cell, of the invention under appropriate conditions to express the antibody or antigen binding fragment thereof. The antibody of antigen g fragment thereof may then be ed. Preferably, the antibody or antigen binding fragment thereof is secreted by the host cell, and can then easily be purified from the culture supernatant. Techniques for purifying antibodies are well known in the art, and include ques such as ammonium sulfate precipitation, size ion chromatography, affinity chromatography, ion exchange chromatography 3O and others.

When expressed in E. coli, the antibodies or antigen binding fragments thereof may be produced in inclusion bodies. Methods to isolate ion bodies and refold the expressed protein are well known in the art.

Yet another aspect of the invention is the antibody or antigen-binding fragment thereof of the invention for medical use.

A further aspect of the invention is the dy or antigen—binding fragment thereof for use in the prevention of the formation and/or the ization of thrombi in a human or animal subject. Three—dimensional intraluminal thrombus growth is reduced or even prevented. Thus, this aspect of the invention relates to the dy or antigen-binding fragment thereof for use in the treatment or prevention of a disorder selected from the group consisting of venous, arterial or capillary thrombus formation, thrombus formation in the heart, thrombus formation during and/or after contacting blood of a human or animal t with artificial surfaces and thromboembolism by preventing and/or treating the formation and/or stabilization of thrombi and thereby the three—dimensional intraluminal thrombus . Yet another aspect of the ion is the antibody or antigen binding fragment thereof for use in the ent of intraluminal thrombi in a human or animal subject related to a disorder selected from the group consisting of venous, al or capillary thrombus formation, us formation in the heart, thrombus formation during and/or after contacting blood of a human or animal subject with artificial surfaces or thromboembolism. Preferably, the venous or arterial thrombus formation is stroke, dial infarction, deep vein thrombosis, portal vein thrombosis, thromboembolism, renal vein thrombosis, jugular vein thrombosis, cerebral venous sinus thrombosis, Budd-Chiari syndrome or Paget~Schroetter disease.

A further aspect of the invention relates to the antibody or antigen binding fragment f for use in the prevention and/or treatment of inflammation, a neurological inflammatory disease, interstitial lung disease, complement activation, fibrinolysis, angiogenesis and diseases d to FXIl/FXlla-induced kinin formation or XIia-mediated complement activation. Preferably, the diseases d to FXIIFXlla-induced kinin formation are selected from the group hereditary angioedema, bacterial infections of the lung, osoma infections, hypotensive shock, pancreatitis, chagas disease, articuiar gout, arthritis, disseminated intravascular coagulation (DIC) and sepsis.

Preferably the interstitiai lung disease is fibroproliferative and/or idiopathic pulmonary fibrosis.

An aspect of the invention is also a method of treatment of any of the conditions or diseases mentioned above in a subject, by administering to the t in need thereof a therapeutically effective amount of the antibody or n-binding fragment thereof.

The beneficial effect of the antibody or antigen-binding fragment thereof in the various conditions can be verified, for example, by employing a suitable animal model, for example a mouse model. By comparison of animals treated with the antibody or antigen-binding fragment thereof and a control group, the beneficial effect of the treatment of the tive disease with the antibody can be demonstrated. Alternatively, patient plasma samples can be tested for relevant parameters. For example, a cial effect in treating or preventing disease— related symptoms in ts with hereditary angioedema can be tested by employing a mouse model, for example as described in Han et al (2002) J. Clin.

Invest. 109:1057-1063. An in vitro test, using patient plasma samples can also be envisaged; treated and untreated patient plasma samples could be ed for bradykinin and/or high moiecuiar weight kininogen levels. The dy should reduce the bradykinin generation, and/or prevent a decrease in high molecular weight kininogen levels.

Preferably, the thrombus formation occurs during and/or after contacting blood of a 3O human or animal subject with artificial surfaces during and/or after a medical procedure performed on said human or animal subject and said dy or antigen binding fragment thereof is administered before and/or during and/or after said medical procedure, and further wherein (i) the artificial surface is exposed to at least 80% of the blood volume of the subject and the artificial surface is at least 0.2 m2 or (ii) the artificial surface is a container for collection of blood outside the body of the subject or (iii) the artificial surface is a stent, valve, intraluminal catheter, or a system for internal assisted pumping of blood.

Preferably, the bleeding risk of said human or animal subject (i) is not increased; and/or (ii) is ined 8) via the ear or finger tip bleeding time according to Duke and wherein said ear or finger tip bleeding time is not longer than 10 minutes or b) according to the method of Ivy and wherein the bleeding time is not longer than 10 s or 0) ing to the method of Marx and the bleeding time is not longer than 4 minutes.

The medical procedure may be i) any procedure ing a cardiopulmonary bypass or ii) the oxygenation of blood via extracorporeal membrane oxygenation or iii) the internal assisted pumping of blood or iv) the dialysis of blood or v) the extracorporeal filtration of blood or vi) the collection of blood in any tory for later use in an animal or a human subject or vii) the use of intraluminal er(s) or viii) the use of stent(s) or ix) the use of artificial heart valve(s).

The dy or antigen—binding fragment thereof of the invention may be administered before, after and/or during a medical procedure ing cardiopulmonary bypass, or a l procedure comprising the collection of blood in any repository for later use in an animal or human subject. It may also be administered by being coated on the artificial surface. Where the medical procedure involves blood donation, the antibody or antigen—binding fragment thereof may be: 1O i) administered to the blood donor before and/or during the blood donation process or ii) mixed with the blood in the collection repository or iii) stered to the blood ent before, during, and/or after the blood is administered to the human or animal recipient.

Preferably the amount of heparin or tives thereof and/or hirudin or tives thereof which is added in addition to the antibody or antigen-binding fragment thereof before and/or during and/or after the medical procedure is reduced or even completely omitted as compared to the amount of heparin or derivatives thereof and/or hirudin or derivatives thereof which is administered normally before and/or during said medical procedure when no said anti—FXll/FXlla antibody or antigen binding fragment thereof is administered.

Preferably, the ombotic risk following the postoperative antagonism of heparin or derivatives f and/or the postoperative antagonism of hirudin or derivatives thereof is prevented or reduced; the prothrombotic risk may also be caused by the administration of protamine.

A further aspect of the invention is the antibody or antigen-binding fragment thereof 3O of the invention for the prevention or the treatment of Pump Head syndrome.

Yet a further aspect of the invention is a l device coated with an antibody or antigen—binding fragment thereof of the invention, wherein the device is a pulmonary bypass machine, an extracorporeal membrane oxygenation system for oxygenation of blood, a device for assisted pumping of blood, a blood dialysis device, a device for the extracorporeal filtration of blood, a repository for use in the collection of blood, an intraluminal catheter, a stent, an artificial heart valve, and/or ories for any one of said s including tubing, cannulae, centrifugal pump, valve, port, and/or diverter.

Another aspect of the ion is the antibody or n—binding fragment thereof for use for administration in a patient receiving a medical procedure, n the medical procedure comprises contact with at least one of: (a) heart, (b) at least one blood vessel chosen from: the aorta, the aortic arch, a carotid artery, a coronary , brachiocephalic artery, vertebrobasilar circulation, intracranial arteries, renal artery, a hepatic artery, a mesenteric artery, and/or a blood vessel of the arterial system cranial to the heart, (C) a venous blood vessel if the patient has a known septal defect; and wherein the medical procedure comprises release of at least one embolus in at least one of said blood vessels in the body that could result in ischemia in at least one target organ and administration of the antibody or antigen—binding fragment thereof before, during, and/or after the medical procedure.

The embolus may be comprised of bubbles, oil, fat, cholesterol, coagulated blood, and/or debris.

The target organ may be: (a) brain, and wherein the patient has, has had, or is at risk for: (i) silent brain ia or (ii) a stroke caused by a nonthrombolysable substance; and/or (b) heart, kidney, liver; and/or gastrointestinal tract organ.

Preferably, the medical procedure comprises contact with the inside of or clamping of at least one or more of said blood vessels.

Preferably, the l procedure is a vascular ure that comprises any one or more of a catheter, a stent, a balloon, a graft, and/or administering a st agent Preferably, the medical procedure is a vascular surgery and/or is a vascular procedure that is stic. More preferably, the medical procedure is coronary angiography, carotid artery stenting, percutaneous coronary intervention, carotid endarerectomy, a cardiovascular surgery, or dilation of stenotic renal artery.

Another aspect of the ion is the antibody or antigen-binding fragment f for use in the prevention or treatment of a condition associated with increased vascular permeability, in particular increased retinal vascular permeability, including progressive retinopathy, sight-threatening complication of retinopathy, r edema, non~proliferative retinopathy, proliferative retinopathy, retinal edema, diabetic retinopathy, ensive retinopathy, and retinal trauma. r aspect of the invention is a ceutical composition comprising the antibody or antigen-binding fragment thereof of the invention. The antibody or antigen-binding fragment thereof can be ated according to known methods for preparing a pharmaceutical composition. For example, it can be mixed with one or more pharmaceutically acceptable rs, diluents or excipients. For example, sterile water or physiological saline may be used. Other nces, such as pH buffering solutions, viscosity reducing , or stabilizers may also be included.

A wide variety of pharmaceutically acceptable excipients and carriers are known in the art. Such pharmaceutical carriers and excipients as well as suitable pharmaceutical formulations have been amply described in a variety of publications 3O (see for example “Pharmaceutical Formulation Development of Peptides and Proteins", Frokjaer et at, Taylor & Francis (2000) or “Handbook of Pharmaceutical Excipients", 3rd edition, Kibbe et al., Pharmaceutical Press (2000) A. Gennaro (2000) "Remington: The Science and Practice of Pharmacy", 20th edition, Lippincott, Williams, & Wilkins; Pharmaceutical Dosage Forms and Drug ry Systems (1999) H. C. Ansel et al., eds 7th ed., Lippincott, Williams, & Wilkins; and Handbook of Pharmaceutical Excipients (2000) A. H. Kibbe et al., eds., 3rd ed.

Amer. Pharmaceutical Assoc). In particular, the pharmaceutical composition comprising the antibody of the invention may be formulated in lyophilized or stable soluble form. The polypeptide may be lyophilized by a y of procedures known in the art. Lyophilized formulations are reconstituted prior to use by the addition of one or more pharmaceutically able diluents such as e water for injection or sterile logical saline solution.

The pharmaceutical ition of the invention can be administered in dosages and by techniques well known in the art. The amount and timing of the administration will be determined by the treating physician or veterinarian to achieve the desired purposes. The route of administration can be via any route that delivers a safe and therapeutically effective dose to the blood of the subject to be treated. Possible routes of administration include ic, topical, enteral and parenteral routes, such as intravenous, intraarterial, subcutaneous, intradermal, 2O intraperitoneal, oral, transmucosal, epidural, or intrathecal. Preferred routes are intravenous or aneous.

The effective dosage and route of administration are determined by factors such as age and weight of the subject, and by the nature and therapeutic range of the dy or antigen-binding fragment thereof. The determination of the dosage is determined by known methods, no undue experimentation is required.

A therapeutically effective dose is a dose of the antibody or antigen g fragment thereof of the invention that brings about a positive therapeutic effect in the patient or subject requiring the treatment. A eutically effective dose is in the range of about 0.01 to 50 mg/kg, from about 0.01 to 30 mg/kg, from about 0.1 to mg/kg, from about 0.1 to 10 mg/kg, from about 0.1 to 5 mg/kg, from about 1 to 5 mg/kg, from about 0.1 to 2 mg/kg orfrom about 0.1 to 1 mg/kg. The treatment may comprise giving a single dose or multiple doses. If multiple doses are required, they may be administered daily, every other day, weekly, biweekly, monthly, or bimonthly or as required. A depository may also be used that slowly and continuously releases the antibody or antigen-binding fragment thereof. A therapeutically ive dose may be a dose that inhibits FXlla in the t by at least 50%, preferably by at least 60%, 70%, 80%, 90%, more preferably by at least 95%, 99% or even 100%.

A further aspect of the invention is an affinity-matured antibody or antigen-binding fragment f of the antibodies (or antigen binding nts thereof) described above. ’15 Definitions Unless otherwise stated, all terms are used according to conventional usage.

“Antibody" in its broadest sense is a polypeptide comprising an immunoglobulin variable region which specifically recognizes an epitope on an antigen. dies are usually comprised of two identical heavy chains and two identical light chains, each of which has a variable region at its N-terminus (vH and vL region). Usually a vH and a vL region will combine to form the antigen binding site. r, single domain antibodies, where only one variable region is present and binds to the antigen, have also been described.

Typically, an antibody contains two heavy and two light chains, connected by disulfide bonds. There are 5 major es of antibodies (lgG, lgM, lgE, lgA, lgD), some of which occur as multimers of the basic antibody structure. The isotype is determined by the constant region of the heavy . There are two types of light chains, lambda and kappa.

The term “antibody” as used herein includes intact antibodies, as well as variants and portions thereof that retain antigen binding. This includes fragments of antibodies such as Fab fragments, F(ab’)2 fragments, Fab’ fragments, single chain Fv fragments, or disulfide-stabilized Fv fragments. Thus, the term “antibody or antigen-binding fragment thereof” in this document is only precautionary, the term “antibody” alone is already intended to cover the antibody and antigen-binding fragments f.

Each heavy and light chain consists of a variable region and a constant region. The variable regions contain framework residues and hypervariable regions, which are also called complementarity determining regions or CDRs. The extent of the framework residues and CDRs is determined according to Kabat; the Kabat database is available online (Kabat EA, Wu TT, Perry HM, Gottesman KS, Foe/ler C (1991) ces of proteins of logical interest, 5” edn. US.

Department of Health and Human services, NIH, Bethesda, MD). The CDR regions are important in binding to the epitope and ore determine the specificity of the A “monoclonal antibody” is an antibody produced by a single clone of B lymphocytes, or by a cell line engineered to s a single antibody.

A “chimeric antibody” is an antibody with the variable s from one species grafted onto the constant regions from a different s. A “humanized” antibody is an antibody where CDR regions from a different species, e.g. a mouse monoclonal antibody, are grafted into the ork of a human antibody.

Analogously, a “murinized” antibody is an antibody where the CDR s from a different species, e.g. a human monoclonal antibody, are grafted into the framework of a mouse antibody. A human antibody is an antibody that is wholly d from human, i.e. human CDRs in a human framework and any constant region suitable for administration to a human.

A “germlined” antibody is an antibody where somatic mutations that introduced changes into the framework residues are ed to the original sequence present in the genome. en binding fragment” refers to any nt of an dy that retains the ability to specifically bind the epitope of the antigen that the antibody binds to.

These include but are not limited to Fab, F(ab’)2, or single chain Fv fragments. 1O “Binding affinity" refers to the affinity of the antibody to its antigen. It can be measured by a variety of techniques, e.g. surface plasmon resonance based technology (BiaCore).

“Epitope” is the antigenic determinant, it is defined by the residues or particular al structures that the antibody makes contact with on the antigen.

“Sequence identity” relates to the similarity of amino acid sequences. The best possible alignment of two sequences is ed, and the sequence identity is determined by the percentage of identical residues. Standard s are available for the alignment of sequences, e.g. algorithms of Needleman and Wunsch (J Mol Biol (1970) 48, 443), Smith and Waterman (Adv Appl Math (1981) 2, 482), Pearson and Lipman (Proc Natl Acad Sci USA (1988) 85, 2444), and others.

Suitable software is commercially available, e.g. the GCG suite of software (Devereux et ai (1984), Nucl Acids Res 12, 387), where alignments can be produced using, for example, GAP or BESTFlT with default parameters, or successors thereof. The Blast thm, originally described by Altschul et ai (J.

Mol. Biol. (1990) 215, 403), but further refined to include gapped alignments (Blast 2), ble from various sources such as the EBI, NCBl, will also produce alignments and calculate the % ty n two sequences.

“Specific binding" refers to the binding to substantially only a single antigen.

“FXIl/FXlla” refers to either or both of Factor Xll and activated Factor Xll (FXlla).

Thus “FXlI/FXIla inhibitor” includes inhibitors of either or both of FXII and FXlla.

Further, anti-FXlI/FXlla antibodies include antibodies that bind to and inhibit either or both of FXII and FXlla. “lnfestins” are a class of serine protease inhibitors derived from the midgut of the hematophagous insect, Triatoma infestans, a major vector for the parasite Trypanosoma cruzi, known to cause Chagas‘ disease (Campos ITN et al. 32 Insect Biochem. Mol. Bio. 991-997, 2002; Campos ITN et al. 577 FEBS Lett. 512-516, 2004). This insect uses these tors to prevent ation of ingested blood.

The infestin gene encodes 4 domains that result in ns that can inhibit different factors in the ation pathway. in particular, domain 4 encodes a protein (lnfestin—4) that is a strong inhibitor of FXlla. lnfestin-4 has been administered in mice without bleeding complications (). Intestin-4 has been d to human serum albumin (rHA-lnfestin-4).

“Complete inhibition of the amidolytic activity of FXlla" means an inhibition of 80% or more, preferably of 90% or more, more preferably of 95% or more, of the ty observed in a control experiment without any inhibitor present. “Activity of Factor Xlla” includes the activity of all forms of Factor Xlla, such as FXIla-alpha and FXlIa- beta.

The terms “treatment” or “treating" or “therapy" are intended to be interpreted broadly; an ement in any disease—related symptom in the subject or t or in a level of a relevant biomarker would be included.

Examples The following examples illustrate certain embodiments of the ion but are not intended to limit the invention to the embodiments that are exemplified. The techniques used are based on standard laboratory procedures well known to the skilled person, and bed in standard laboratory manuals.

To isolate fully human dies from the DYAX Fab-based phage display library which are able to effectively inhibit the ytic activity of human FXlla.

Materials rHA-lnfestin-4 (inhibitor of FXlla amidolytic activity) was ed by Drs. Thomas Weimer, Holger Lind, and Stefan Schmidbauer (CSL Behring). Human FXlI, FXlla, and FXlla beta were purchased from Enzyme Research Laboratories (supplied by Banksia Scientific, Qld, Australia). genic substrate 8—2303 was from Chromogenix (supplied by Abacus ALS). Sulfo-NHS—SS-Biotin and TMB ate Solution were from Pierce. Enzymes and M13—KO7 helper phage were from New England Biolabs. Maxisorp immunoplates were from Nunc. Dynabeads M-280 Streptavidin were from lnvitrogen Corp. Twin tec skirted 96-well PCR plates were from Eppendorf. Taq DNA polymerase was from ifix. ExoSAP—lt was supplied by GE Healthcare. BigDye Terminator sequencing kit was from Applied Blosystems. Anti~human FXll dy (OT—2) was from Sanquin (Amsterdam, Netherlands).

Example 1. Phage display selection 1) Phage panning method A human Fab—based phage display library (Dyax Corp. Cambridge, MA) was used to screen against biotinylated FXlla beta. Prior to initiating each round of selection, the antibody library was preincubated with 500 uL of 4% milk in PBS for 1 hr at room temperature (RT). 100 uL aliquots of M280 Streptavidin beads were coated with 3 ug of biotinylated FXlla beta overnight at 4°C, followed by washing 3 times in PBS/0.05% Tween 20 (PBST) and once in PBS using a KingFisher magnetic particle processor (Thermo Fisher Scientific). Beads were collected using a Dynal magnetic particle separator (MPS) rogen Corp), resuspended in 1mL of 2% milk in PBS, and tumbled at RT for 1 hr. Blocked beads were collected using the MP8 and Round 1 was performed by incubating 5.5 x 1012 colony forming units (cfu) of phage with lised FXlla beta in total volume of 1 mL at RT for 20 1O minutes. Following the tion the beads were collected and washed 10 times with PBST using the Kingfisher, ed by 2 manual washes in PBS. Finally, the beads were resuspended in 500 uL PBS and designated as Round 1 output (approximately 0.5 x 108 cfu total). The Round 1 output phage were then amplified by infecting 6 mls of TG1 culture with one half (250 uL) of beads at 37°C for 30 minutes, with g at 250 rpm. One mL of infected culture was removed and stored at 4°C, and 2.5 x 1010 pfu of M13KO7 helper phage were added to the remaining 5 mLs of culture, followed by an additional tion at 37°C t shaking. The amplification was completed by on of 30 mLs of 2xYT media (containing 100 ug/mL Ampicillin and 50 pg ImL Kanamycin) and an overnight 2O incubation at 30°C. Following amplification, the bacterial pellets were harvested by centrifugation for 30 min at 4000 rpm, and the phage were precipitated from the resulting medium following the addition of 1:5 volume NaCl—PEG on (20% PEG 8000, 2.5 M NaCl) and incubation on ice for 60 min. The precipitate was resuspended in 1 mL PBS, bacterial debris removed by centrifugation at 8000 rpm using a bench top centrifuge for 10 minutes and the phage precipitated again as described above. The final phage pellets were resuspended in a total volume of 1mL in PBS, and titered to be used as input for the next round of selection. Rounds 2 and 3 were performed as described for Round 1. Following Round 3, a pilot scale selection of clones and preliminary analysis for binding to FXlla beta was done by ELISA. 2) Pilot scale picking and ELISA analysis of clones from Round 3 of phage display selection The preliminary screening of Round 3 output clones was carried out by Fab-phage ELISA. Colonies were picked and inoculated into 120 pL of 2xYT medium, containing 2% glucose and 100 ug/mL ampicillin. These were shaken overnight at 37°C, 250 rpm (lnfors Supershaker) and designated “masterplate”. These cultures were used to inoculate 100 uL of 2xYT/100 pg/mL ampicillin in deep well , and plates incubated at 37°C, 700 rpm to an OD600 of approximately 0.5. 100 uL of helper phage was then added to a final concentration of 0.5 x 1010 pfu, and plates incubated without shaking for 30 min at 37°C. 2xYT media (containing 100 pg/mL Ampicillin and 100 ug/mL Kanamycin) was added to the rescued cultures to give a final concentration of 25pg/mL of kanamycin, followed by an ght incubation at 30°C with shaking (650 rpm). The resultant cultures were spun at 6009 for 30 minutes, and supernatants used for phage ELISA.

For Fab—phage ELISA, Nunc immunopiates were coated overnight at 4°C with 100 uL/well of 1 pg/mL FXIIa in PBS. Negative control wells coated with PBS alone were also included. Wells were then d for 2 hrs at 37°C with 200 uL of 5% skim milk/PBS, and washed 3x in PBST. Fifty pL of 1% skim milk/PBST and 50 uL of phage e atant were added to each well, and plates were incubated with shaking at room temperature for 2 hrs. Plates were than manually washed 5 times with PBST, and 100 pL of anti-M13 mAb diluted 1/5000 in 1% milk/PBST was added to each well, followed by 30 min incubation at RT with shaking. Plates were then washed as before, and 100 pL of TMB substrate was added to each well and the plates then incubated for 10 minutes at RT with shaking. The on was d by the addition of 50 pL of 2M phosphoric acid, and the absorbance was read at 450 nm in a microplate reader (Waiiac Victor). Twelve clones appeared positive in the single well ELISA, and were further tested in a competition ELISA. 3) Analysis of clones from Round 3 of selection: Competition phage ELISA The twelve clones found reactive to FXlla in a single well Fab-phage ELISA were further tested for reactivity to FXIla in a competition ELISA. Briefly, the phage titres from culture supernatants (see previous section) were first determined using a titration ELISA. For titration ELISA, Nunc plates were coated overnight at 4°C with 100 uL/well of 1 ug/mL FXlla in PBS. Negative control wells coated with PBS alone were also included. Wells were then blocked for 2 hrs at 37°C with 200 pL of 5% skim BS, and washed 3x in PBS/0.05% Tween 20 (PBST). Fifty pL of phage supernatants were 4-fold serially diluted in 1% skim milk/PBST, and 100 uL of each dilution were added to the blocked plate. After 1.5 hr incubation at RT with shaking, plates were manually washed 5 times in PBST, and the rest of the ELISA protocol was followed ially as described in previous section. The data was plotted using KaleidaGraph re with Sigmoidal curve fit, and E050 value was ed.

For competition ELISA, Nunc 96—well immunoplates were coated and blocked as above. Phage concentrations were fixed at a level determined from the titration ELISA, and the competitor protein (rHA~lnfestin—4) was serially diluted. Briefly, 4- fold serial dilutions of the competitor protein were made by having 100 uL of 2 times itor in the initial well (ie. 200 nM for desired 100 nM concentration) with 75 uL on buffer (1% skim milk/PBST) in remaining wells, and serially ng 25 uL of competitor down the plate. 75 uL of 2x phage stock (dilution determined from ion ELISA) were added to each well, and 100 uL from each well were transferred into a coated and blocked plate, and the rest of ELISA protocol was followed as described above. Phage expressing lnfestin domain 4 (lnf4) as a gene lll fusion were used as positive control in the competition ELISA. Phage clones designated 3F7 and 3H4 showed competition with ECsO values equivalent to the control Inf4-phage, and were selected for further is (Figure 1). The results of 3O the competition ELISA indicate that rHA-lnfestin-4 is able to compete with SF? and 3H4 Fab—phage and most likely bind to similar regions on FXlla. All other phage clones whilst able to bind to FXlla were not ed by rHA-infestin-4 (as represented by clone 3G5 in Figure 1) and hence were unlikely to bind to similar regions within the catalytic domain of FXIla. 4) Analysis of clone 3F7: Sequence analysis To determine the amino acid sequences for Fab clones 3F? and 3H4, 5 mL overnight cultures were started using 5 uL of “masterpiate” cultures, and plasmids were isolated using Qiagen miniprep kit. The Fab casette DNA was sequenced using CHtRev and pLacPCwa primers (Table 2). Sequencing reactions and electrophoresis were carried out at the DNA cing facility of Department of Pathology, Melbourne University. The sequences were analyzed using SeqMan (Lasergene), and found to be 100% identical, hence a single antibody (3F7) with the y to e with infestin-4 for binding to FXlla was obtained from panning.

Table 2: Sequencing primers used for the characterization of phage clones Primer name Sequence SEQ 1137 ___..—.__.—__I..__ CH1 Rev 5’ GTCCTTGACCAGGCAGCCCAG 3’ W5‘ GTGAGTTAGCTCACTCATTAG 3‘ wt em 5‘ TTTTCATCGGCATTTTCGGTC 3‘ stqu rev KpaCwad 5’ CCATCTGATGAGCAGTTGAAATCT 3‘ d 5‘ GTTCCCGCCCTCCTCTGAGGAGCT 3‘ PUCrev 5’ AGCGGATAACAATTTCACACAGG 3’ 69 3254 5’ GGTTCTGGCAAATATTCTG 3’ 70 Seq CL 5’ GTTGCACCGACCGAATGTA 3’ lambda Seq CH1 59 72 ACCGTGAGCTGGAACAGCGGTGC GC 3 ’ Table 3: Sequences of the variable regions and CDRs of 3F7. CDR's defined according to KABAT numbering system (Kabat EA, Wu TT, Perry HM, Gottesman KS, Foeller C (1991) Sequences of proteins of immunological interest, 5th edn. U.S.

Department of Health and Human services, NIH, Bethesda, MD) VH SGGGLVQP QWV Q LEWVSGIRPSGGTTVYADSVKGRFTISRDNSKNTLYLQMNSLRAE DTAVYYCARALPRSGYLISPHYYYYALDVWGS gGTTVTVSS VL PPSASGTPGQRVTISCSGSSSNIGRNYVYWYQQVPGTAPK LLIYSNNQRPSGVPDRFSGSKSGTSASLVISGLRSEDEADYYCAAW DASLRGVFGGGTKLTVLG HCDR 1 (Kabat KYIMQ 31—35) HCDR 2 (Kabar GIRPSGGTTVYADSVKG [__ 50—65) HCDR 3 (Kabat ALPRSGYLISPHYYYYALDV 95-1 02) LCDR 1 (Kabat SGSSSNIGRNYVY 24—34) LCDR 2 (Kabat SNNQRPS 50-56) LCDR 3 (Kabat AAWDASLRGV 89—97) ) Analysis of clone 3F7: FXlla inhibition by 3F? mAb To assess whether the BF? mAb ts FXlla amidolytic activity in an in vitro assay, 3F? age was reformatted into full length human IgG4/lambda 1O antibody and purified using protocols described in Example 3. Briefly, 1 ug of FXlla was incubated in Nunc immunoplates in presence or absence of rHA-lnfestin-4, 3F? mAb or control mAb (anti-human GCSFR antibody C12) in a volume of 160 uL for min at 37°C. Forty uL of ate (4mM 8-2302) were added, and the plate was further incubated at 37°C for 15 min. The reaction was stopped by the addition of 4O uL of 20% acetic acid, and colour change was detected at 405 nm in plate reader. The data was plotted using KaleidaGraph software with Sigmoidal curve fit, and E050 value was recorded. As shown in Figure 2, the 3F7 antibody was found to effectively inhibit FXlla ytlc activity. e 2. Affinity maturation of the 3F7 antibody The aim of the affinity maturation of 3F7 was to identify and characterise 3F7 mAb ts able to bind to human FXlla with higher affinity than the al antibody.

Higher affinity variants have the potential to show improved inhibition of FXlla amidolytic activity. The method for the generation of Fab-phage affinity tion libraries (see Library construction below) is dependent on degenerate oligonucleotides annealing to a ssDNA template which is then extended to make a double stranded form for ormation. The size of the library is dependent upon transformation efficiency, and degeneracy of the primers used. The primers used (see below) covered a 19 amino acids combination (without cysteine). Libraries targeting 6 amino acid residues at a time were designed. The theoretical diversity of using trimer ucleotides for 6 residues is 198 = 4.7 x 107. 1) Design of affinity maturation libraries For each phagemid, a germline stop template was created by replacing 18 codons (6 amino acid residues) in all CDRs, except CDR—LZ, with TAA stop codons. The linear design for the constructs is as follows: Ncol~VL-CL—|inker-VH—Sall. Flanking Ncol and Sail sites were included for cloning into phage display pTac vector, containing remaining elements for phage display. The stop template versions named 3F7 Hi, 3F7 H2, 3F7 H3.i and 3F7 H32 (heavy chain variable ) and 3F7 Li, 3F7 L3.1, 3F7 L3.2 (light chain variable region) were ed by GeneArt and are shown in Figures 3 and 4 respectively. 2) Libram construction Libraries were constructed using methods bed by Sidhu et al. (Phage display for ion of novel binding peptides. Methods in Enzymology, 2000, vol. 238, p.333-336) with “stop template” versions of pTac—3F7 Fab. Each stop template was used as template for the Kunkel mutagenesis method (Kunkel et al., Rapid and efficient site—specific mutagenesis without phenotypic selection. Methods in Enzymology, 1987, vol. 154, p. 367—382) with mutagenic oligonucleoteides (Table 4) designed to aneously repair the stop codons and introduce mutations at the designed sites. The mutagenesis reactions were introduced into E. coli 88320 by electroporation, and phage tion was ted with on of M13-KO7 helper phage. After overnight growth at 30°C, the phage were harvested by precipitation with PEG/NaCl. The mutagenesis efficiencies were assessed by sequencing of12 clones randomly picked from each library, and ranged from 50 to 100%. Each library contained 0.75 -— 3.75 X 109 individual clones. Primer 3254 (Table 2) was used to sequence clones from libraries L1, L3.1 and L3.2 and primer Seq CL lambda (Table 2) was used to sequence clones from libraries H1, H2, H3.1 and H32.

Table 4: 3F7 mutagenic trimer ollgonucleotides used for affinity maturation, where each “Nnn” designates a triplet encoding one of 19 amino acids without cysteine (produced and supplied by Ella Biotech, Germany).

’GCTGTAGCGGTAGCAGCNnnNnnNnnNnnNnnNnnTATGTGTATTGGTA TCAGCA 3’ (SEQ ID NO: 24) 3F7 L3.1 5 AGCCGATTATTATTGTNnnNnnNnnNnnNnnNnnCTGCGTGGTG TTTTTGGT 3’ (SEQ ID NO: 26) 3F7 L3.2 5 ’TTATTGTGCAGCATGGGATNnnNnnNnnNnnNnnNnnTTTGGTGGTGGC ACCAAA 3’ (SEQ ID NO: 28) ’AGCAAGCGGTTTTACCTTTNnnNnnNnnNnnNnnNnnTGGGTTCGCCAG GCAC 3’ (SEQ ID NO: 16) ’GGAATGGGTTAGCGGTATTNnnNnnNnnNnnNnnNnnACCGTTTATGCA G 3’ (SEQ ID NO: 18) 3F7 H3.1 5 ’TTATTATTGCGCACGTGCANnnNnnNnnNnnNnnNnnCTGATTTCTCCGC ATTATTA 3’ (SEQ ID NO: 20) 3F7 H3.2 5’CACTGCCTCGTAGCGGTNnnNnnNnnNnnNnnNnnTATTATTATTATGCC CTGGAT 3’ (SEQ ID NO: 22) 3) Library panning Libraries were cycled through five rounds of selection with decreasing concentration of biotinylated FXIla beta. The target concentration was reduced 10—fold with each round, from 40 nM in Round 1 to 4 pM in Round 5. Panning was carried out in on with the ylated FXlla beta. Phage samples were incubated with antigen diluted in 4% milk in PBST (or 4% milk/PBST alone to make blank samples with no target) with rotation at RT for 1 hr. Dynal M-280 Streptavidin magnetic beads were blocked in 5% skim milk/PBS for 30 min at 37°C with horizontal g. Beads were collected using MP8 and phage/antigen mixture was added for 30 minutes. Beads were then washed 10 times in PBST (KingFisher Long Wash), ed by a manual wash in PBS. Beads were finally resuspended in 500 pL 50 mM DTT and incubated at 37°C for 30 min with horizontal shaking. The eluted phage were collected, and added to 170 uL of neutralisation buffer (0.351 g L-cysteine + 5 mg BSA made up to 5 mL with 1 M Tris pH 8). 330 uL of the eluted phage were used as input for the next round. Enrichment for each round of selection was calculated as ratio of eluted phage selected on target versus blank samples. 4) Analysis of clones from 3F7 affiniy maturation At the completion of panning, a number of phage clones were selected from each enriched library and ced using the primers detailed above (library construction). Unique clones from each library were then selected based on sequence and reformatted into fully human lgG4/lambda antibodies for binding is. ty matured variants were initially screened using Biacore as unpurified cell culture supernatant to estimate binding affinities in comparison to parental 3F7 (as described in Example 4(1)). Table 5 lists the antibodies that were found to have a higher binding affinity to FXlla beta than 3F7. The highest ty clones tended to come from the heavy chain CDR2 and the light chain CDR 1 regions.

Table 5: ted binding affinities of 3F7 and affinity d variants based on binding kinetics at a single FXlla beta concentration. All antibodies were tested as fed lgG4 molecules in cell culture supematants on a Biacore A100 instrument. Only variants with better affinity to FXlla than 3F7 are shown. Refer to Figures 3 and 4 for library locations. 29 (res1dues 3—8) (residues 3—8) 1.52E-09 VR115 H2 DIPTKG 31 (residues 3—8) 1.56E—O9 VR24 L1 EMTVHH 44 (residues 5—10) 1.63E-09 VR110 H2 DMPTKG 32 (residues 3—8) 2.04E-09 VR107 H2 NPATRT 33 (residues 3-8) 24513-09 VR06 L1 FSHPHH 45 ues 510) 2.56E—09 VR31 L3.1 ASWYND 52 ues 1-6) VR108 H2 PATKT 34 (residues 3—8) 2.81E-09 VR103 H2 DVPVRG 35 (residues 3-8) 2.87E-09 VR101 H2 NPATRS 36 (residues 3—8) 3.33E-09 VR16 L1 EFVEYN 46 (residues 5-10) WE—M VR29 L3.1 ASWEIP firesidues 1-6) 3.89E-09 VROS L1 DTNSHH 47 (residues 5—10) 4.36E-09 VR12 L1 WTEQHN 48 (residues 5—10) -09 VR27 L3.1 ASWTNE 54 (residues 1-6) 4.64E-09 VR10 L1 VMVTNH 49 residues 5-10) 4.97E-09 VR149 H3.2 YLMKKN 39 ues 7-12) 5.12E-09 VR58 L32 PQVRLA 59 (residues 5—10) 5.33E—09 VR39 L31 ASWWND 55 ues 1—6) 5.63E-09 VR16? H32 LMKTG 40 (residues 7-12) 5.80E~09 VR62 L32 QQVRLD 60 (residues 5-10) 5.81E—09 VR109 H2 NPATNT 37 (residues 3~8) _15—.98E—09 VR14 L1 GMVEQN 50 ues 5—10) 62213—09 VR46 L31 ASWELP 56 (residues 1—6)—]6.67E-09 VR148 H32 YLVKKQ 41 (residues 7—12) 6.93E-09 VR159 H32 YLVKHG 42 (residues 7-12) 6.93E-09 VR53 L32 QQVRKT461 (residues 5—10) 7.05'13—09 VR52 L32 ERVRLM 62 (residues 5—10) WEE—09 VR160 H32 YLMKPG 43 (residues 7—12) 7.13E-09 VR17 L1 FKVEET 51 (residues 5—10) $5309 VR41 L3.1 ASWSIP 57 (residues 1—6) 9.14E-09 PATMT 38 (residues 3-8) 9.19E-09 ASWEVP 58 (residues 1-6) U! H Based on the results from the estimated binding affinity screening the best 5 mabs were then purified and subjected to detailed binding affinity analysis (as described in example 4(2)). As shown in Table 6, these clones showed a 24 to 57-fold improvement in binding ty compared to parental 3F7.

Table 6: Detailed Biacore analysis of the binding affinity of purified 3F? and the top affinity matured variants to FXlla beta from Table 5. All antibodies were tested as fully human lgG4 molecules. 2 Fold affinity 3F7 Variant; ka(1/MS) kd (I/s) KD (M) , l .9x10'5 Example 3. lgG production and purification of phage-derived dies of the invention 1) Mammalian expression vector construction The ian expression vectors were constructed using standard lar biology techniques by cloning the entire light chain (variable and constant s) and the variable domain of the heavy chain from the selected phage-derived Fab constructs into the pRhG4 vector as previously described (Jostock et al 2004.

Rapid generation of functional human lgG antibodies d from Fab—on-phage display ies. J Immunol Methods, 289; 65-80). 2) Cell Culture Serum-free suspension adapted 293-T cells were obtained from Genechoice Inc.

Cells were cultured in FreeStyleTM Expression Medium (lnvitrogen) supplemented with penicillin/streptomycin/fungizone reagent (lnvitrogen). Prior to transfection the cells were maintained at 37°C in humidified incubators with an atmosphere of 8% C02. 3) Transient Transfection The transient ection of the mammalian expression vectors using 293-T cells 1O was med using 293fectin transfection reagent (lnvitrogen) according to the manufacturer’s instructions. The light and heavy chain expression vectors were combined and co-transfected with the 293-T cells. Cells (1000 ml) were transfected at a final concentration of 1 x 106 viable cells/ml and incubated in a Cellbag 2L (Wave Biotech/GE Healthcare) for 5 days at 37°C with an atmosphere of 8% C02 on a 2/10 Wave Bioreactor system 2/10 or 20/50 (Wave Biotech/GE Healthcare).

The culture conditions were 35 rocks per minute with an angle of 8°. Pluronic® F- 68 rogen), to a final concentration of 0.1% v/v, was added 4 hours post- transfection. 24 hours post-transfection the cell es were supplemented with Tryptone N1 (Organotechnie, France) to a final concentration of 0.5 % v/v. The cell culture supernatants were harvested by centrifugation at 2500 rpm and were then passed through a 0.45pM filter (Nalgene) prior to purification. 4) Analysis of Protein Expression After 5 days 20ul of e supernatant was electrophoresed on a 4-20% Tris— Glycine SDS polyacrylamide gel and the antibody was visualised by staining with Coomassie Blue reagent.

) Antibody Purification onal antibodies were purified using tandem protein A affinity chromatography and desalting column chromatography. Chromatography using Hitrap MabSelect sure (1 ml, GE Healthcare, UK) and ing (HiPrep 26/10, GE Healthcare, UK) resins were ped using an AKTA express (GE Healthcare, UK) as per manufacturers recommended method. Briefly, equilibration of the Protein A affinity column was performed in 1 X MT-PBS buffer. The filtered conditioned cell culture media (500 ml) was applied to the column at 1 ml/min and washed sequentially with 1 X MT-PBS (10ml) and 10mM Tris, 0.5M Arginine,150mM NaCl pH 7.2 (80ml). The bound antibody was then eluted with 0.1M Na Acetate pH 3.0 (8ml) and immediately applied to the desalting column.

The antibody concentration was determined chromatographically by comparison to control dy standards. Protein fractions were pooled and concentrated using 1O an Amicon UltraCel 50K centrifugal device (Millipore) prior to sterile filtration using 0.22um s.

The purity of the dy was analysed by GE, where 2 pg protein in reducing Sample Buffer (Invitrogen, CA) was loaded onto a Novex NuPAGE 4-12% Bis-Tris Gel (lnvitrogen, CA) and a constant e of 200V was applied for 40 minutes in an XCell SureLock ell (lnvitrogen, CA) with NuPAGE MES SDS running buffer before being visualised using Coomassie Stain, as per the manufacturer’s instructions.

Example 4. Antibody affinity determination — Biacore analysis 1) Estimated binding affinities from unpurifed antibody supematants Anti-human (Goat anti-human lgG (gamma) mouse adsorbed, lnvitrogen, Cat No.

H10500) was chemically immobilised on a CM—5 sensor surface using amine coupling chemistry. e supematants were diluted 1/60 with running buffer before capture. Antibodies were captured for 180 seconds representing an average capture of 800 se units (RU). FXlla beta was then ed at zero and 100nM for 180 seconds, and dissociated for 180 seconds. All assays were conducted on a Biacore A100 instrument at 37 degrees Celsius and the data fitted to a 1:1 kinetic model. 2) Detailed binding ty is Anti-human (Goat ant— Human igG (gamma) mouse adsorbed, lnvitrogen, Cat No.

H10500) or anti mouse Fc specific antibody (Jackson immuno Research Labs inc.

Cat No. 515—005-071) was chemically immobilised on a CM—5 sensor surface using amine coupling chemistry. The immobilised antibodies were then used to capture anti-FXll/FXlla mAbs from on.

Human FXll or FXlla beta was then injected over captured antibody at various 1O concentrations for detailed binding kinetics. Responses from a reference flow cell (in which mAb was not ed, but otherwise treated identically), were subtracted.

The responses from a blank injection were then subtracted from the resultant sensorgrams.

The final corrected responses were fitted using non-linear regression to a model describing 1:1 cs, including a term for mass transport limitation. The Rmax value was fitted locally, to account for slight deviations in the level of mAb captured.

Association rate (ka), dissociation rate (kd) and equilibrium dissociation constant (KD) were determined.

For detailed binding cs FXll was injected at 0, 15.1, 31.25, 62.5, 125, 250, and 500nM, in duplicate and FXlla beta was injected at O, 1.25, 2.5, 5, 10, 20 and 40nM, with 10nM in ate.

For the 3F7 antibody, regeneration was performed after each cycle with a 90 second injection of 100mM H3PO4. For mab OT—2, regeneration was performed after each cycle with a 60 second injection of 25mM e, pH 1.7, followed by a second injection of 25mM glycine, pH 8.6. All assays were conducted at 25°C.

Example 5. Comparison of 3F7 with other antibody inhibitors of FXIla amidolytic activity.

A review of the relevant scientific literature revealed that although a number of antibodies have been described which can modulate FXll activity, the majority of these are either directed to the heavy chain and prevent the initial contact activation of FXll or are directed to the light chain and appear to only lly inhibit FXll amidolytic activity. The aim of this work was to compare 3F7 to antibody OT-2 which has been claimed to completely block the amidolytic activity of FXlla (Dors et al., A novel sensitive assay for functional FXll based on the generation of rein— C1—inhibitor xes in FXll deficient plasma by glass-bound Factor Xll.

Thrombosis and Haemostasis, 1992, vol. 67, p. 644-648; Citarella et al., Structure/function analysis of human factor Xll using recombinant deletion mutants. an Journal of Biochemistry, 1996, vol. 238, p. 240—249). 1) Inhibition of FXlla amidolytic activity with 3F7 and OT-2 antibodies The activity of 3F7 and OT-2 antibodies was ed in an in vitro FXIla amidolytic activity assay, ially as described in Example 1(5). Both antibodies were able to tely block the amidolytic activity of FXIla (Figure 5). 2) Biacore analysis of SF? and OT-2 mAbs binding to FXll and activated FXIla beta Whilst both 3F7 and OT-2 were shown to completely block the amidolytic activity of FXIIa, 3F7 showed a small but reproducible ~2-fold higher potency in this assay.

To determine if BF? and OT-2 share a similar epitope on FXIla we initially performed a competition ELISA with these antibodies and showed they were able to effectively compete with each other for binding to FXIla (data not shown). 3O To further terize the comparative g of these antibodies to FXll we med Biacore experiments with both antibodies against unactivated FXll and catalytically active FXlla beta. The results ofthis experiment are shown in Table 7 and demonstrate that whilst OT-2 shows equivalent binding affinity to FXll or activated FXlla beta, 3F7 shows a clear preference for binding to the activated form of FXII (FXIla). These results show that whilst both antibodies appear to bind to similar regions on the light chain of FXII they do not appear to share an identical epitope. The ability of 3F7 to preferentially bind to activated FXII may confer a pharmokinetic and/or pharmacodynamic advantage.

Table 7: Detailed Biacore analysis of the g affinity of the purified lgG monoclonal antibodies 3F7 and OT—2 to FXlla beta. mAb FXIIKB (111%) FXlla/f KD (nM) 121 $19 (N23) 6.2 i 0.2 (N=3) 0.69 :t 0.25 (N=3) 0.76 i 0.077 (N=3) Example 6: Identifying key FXlla es involved in the binding of 3F7 Having screened 3F7 for its ability to t the activity of FXlla from a number of species (data not shown), we determined 3F7 to be highly potent t mouse and human FXlla, but not rat FXlla. Using this information we investigated which key residues within the FXlla light chain may be involved in the 3F7 epitope by ting a recombinant murine FXII (which is recognised by 3F7 using Western analysis) and mutating various es that differed from the rat amino acid (see Figure 6). As the result shows, mutating either on 398 or 438 abolishes the binding of 3F7. 1. Construction and expression of wild-type and mutant murine Factor Xll (Mu-FXll) A cDNA encoding the entire Mu-FXll protein (GenBank Accession no. NM_02‘|489) was obtained from GeneART AG (Regensberg, Germany). This cDNA was used as a template to make the following separate residue changes by rd PCR techniques: a) N376D, b) A385D, c) N398K, d) W420R, e)R427H, f) I438A, g) Q450R, h)’delE451, i) S4526, j) K453R, T454K, k) G4728, l)N5168, m) T538A and n) A589D. With the exception of f) I438A, these residue changes ponded to a switch from the mouse e to its rat orthoiogue (GenBank Accession no.

NM_001014006) (see Fig. 6A). In the case of h), this involved a deletion of Glu451.

One further mutant (0) was generated, a multiple mutant ing the murine to rat amino acid changes E552D, T555V and A556T. All constructs were modified at the 3' end to encode a C—terminal 8xHis-tag, cloned into the mammalian expression vector pcDNA3.1 (lnvitrogen, Carlsbad, USA) and the sequence validated by DNA sequence is. yleTM 293 suspension cells (lnvitrogen] were grown to 1.1 x 106 cells/ml in 5m| Freestyle Expression media (lnvitrogen). 7 pL 293Fectin (lnvitrogen) transfection reagent was pre-incubated for 5 minutes with 167 uL Opti-MEM l medium (lnvitrogen), then added to 5 pg plasmid DNA encoding wild—type or mutant Mu-FXll and the mixture incubated for a further 20 minutes. The DNA-293Fectin complex was added to the cells which were cultured for 6 days at 37 °C, 8% C02 in a shaking incubator at 250 rpm. Culture supernatants were harvested by centrifugation at 2000 rpm for 5 minutes and stored at 4 °C for analysis. 2. Western Blotting Supernatants containing recombinant wild-type or mutant mu—FXII were added to equal volumes of 2x ducing sample buffer, incubated at 80°C for 10 minutes and then loaded onto pre—cast 4-12 % Bis—Tris gels (lnvitrogen) and electrophoresed for 1 hour at 200V. Proteins were then transferred ophoretically onto nitrocellulose filters and blocked for 1 hour in 5% Milk powder in uffered saline with 0.05% Tween-20 (TTBS). Filters were then incubated for 1 hour with either 3F? mAb or an anti-His mAb 3H3 (both at 1 mg/mL in TTBS with 5% Milk ), washed thoroughly with TTBS, then incubated for a further hour with uman TC or anti-mouse igG-FITC, respectively (Millipore, USA; both at 0.25 mg/ml in TTBS with 5% Milk powder). Following further washing of membranes in TTBS, Ab—FITC bound proteins were visualized using a Typhoon variable mode analyzer (GE Healthcare, USA). The results are shown in Fig. SB. The binding of BF? is abolished when residues 398 and 438 of the mouse sequence are mutated, indicating that these two residues may be part of the epitope of mAb 3F7.

Example 7: Prevention of FeCla-induced arterial thrombosis in mice with by enous treatment with monoclonal antibody 3F7 us studies (e.g. disclosed in W02006066878) have shown that inhibition of FXIIa prevented FeClg—induced arterial thrombosis in mice. The goal of this study was to explore whether mice are also protected against arterial thrombosis by treatment with a specific monoclonal antibody directed against coagulation factor Xlla (MAb 3F7).

Methods Treatment groups were as shown in Table 8: Table 8: Treatment groups N0. Treatment Dose / volume / schedule / route N (f) 1 Isotonic saline Na] / 0.1 mL/20g b.W. / t=-15 min. / i.V. 25 2 MAb 3F7 30 mg/kg / 0.1 mL/20g b.W. / t=-15 min. /i.v. 10 3 MAb 3F7 20 mg/kg / 0.1 mL/20g b.W. / t=-15 min. /i.v. 5 4 MAb 3F7 10 mg/kg / 0.1 mL/20g =-15 min. /i.v. 10 MAb 3F7 5 mg/kg / 0.1 mL/20g b.W. /t=—15 min. /i.v. 10 6 MAb 3F7 2.5 mg/kg / 0.1 mL/20g b.W. / t=-15 min. /i.v. 10 7 MAb 3F7 1 mg/kg / 0.1 mL/20g b.W. /t=-15 min. /i.v. 10 8 MAb 3F7 0.5 mg/kg / 0.1 mL/20g b.W. /t=-15 min. /i.v. 10 9 Control MAb 30 mg/kg / 0.2 mL/20g b.W. /t=-15 min. /i.V. 10 ,_._. .1 1Na. = not applicable Mice of strain NMRI, obtained from Charles River Laboratories, female, aged 6-8 weeks, ng between 25 and 39 9, received a single i.V. injection of the treatment solution as listed in Table 8 at t=~15 min in deep esia. Thereafter, the effects of the treatment on the thrombotic occlusion rate were quantified.

Baseline blood flow was determined by placing an ultrasonic flow probe around the d arteria carotis. To initiate thrombosis, a 0.5 mm2 (0.5x'l.0 mm) patch of filter paper saturated with 10 % ferric chloride solution was placed on the arteria carotis ream from the flow probe at t=0 min. After 3 minutes the filter paper was removed and blood flow was monitored for 60 minutes to determine the occurrence of thrombotic ions.

Following the 60 minutes observation period, blood samples were taken from study animals (anticoagulant: 10% citrate). Thereafter, plasma was prepared according to standard methods, and deep frozen (-80 °C i 10 °C) until ination of aPTT (activated partial thromboplastin time), PT (prothrombin time) and FXlla-activity.

Determination of the aPTT: The aPTT was determined by adding 50 uL of study plasma samples (see above) to 50 uL Pathromtin SL (Siemens HealthCare Diagnostics Products GmbH, Marburg, Germany) followed by an incubation phase of 120 seconds at 37°C.

Subsequently, 50 uL of a calcium chloride solution (25 mM, Siemens HealthCare stics Products GmbH, Marburg, Germany) was added to start the reaction.

Determination of the PT: The PT was determined by adding 50 uL of study plasma samples (see above) to 100 uL of the activation reagent Thromborel S ns HealthCare Diagnostics Products GmbH, g, Germany) after 15 seconds incubation time at 37°C.

Determination of activity: The FXlla-activity was determined by using an aPTT~based assay and compared to a reference curve ed with dilutions of standard human plasma and FXll- deficient plasma (Siemens HealthCare Diagnostics Products GmbH, Marburg, y). 50 uL of the study plasma samples (see above), which were pre-diluted 1:5 with ol buffer solution (Siemens HealthCare Diagnostics Products GmbH, Marburg, y), were added to 50 uL of FXil-deficient plasma. After an incubation time of 30 seconds at 37°C, 50 pL Pathromtin SL (Siemens HealthCare Diagnostics Products GmbH, Marburg, Germany) was added and the solution thereafter incubated for 120 seconds at 37°C. Subsequently, 50 pL of a calcium chloride solution (25 mM, s HealthCare Diagnostics Products GmbH, Marburg, Germany) was added to start the reaction.

All three analyses were performed in a BCT (Behring Coagulation Timer; s HealthCare Diagnostics Products GmbH, Marburg, Germany) in line with the conditions suggested by the supplier of tive assay reagents (Siemens HealthCare Diagnostics Products GmbH, Marburg, Germany).

Results: intravenous injection of 30 mg/kg, 20 mg/kg, 10 mg/kg and 5 mg/kg MAb 3F7 resulted in a te protection from FeCls—induced occlusion of the arteria carotis of mice (Table 9, Figure 7). At sing doses (Le. 2.5 — 0.5 mg/kg), occlusion rates increased while times to occlusion decreased dose—dependently (Table 9, Figure 7). Compared to controls, PT was unchanged (Table 10, Figure 9) while aPTT was prolonged about fourfold at the high doses (Table 10, Figure 8). FXlla— activity was nearly completely inhibited at a dose of 10 mg/kg and above, and still halved at a dose of 0.5 mg/kg (Table 10, Figure 10). Furthermore, aPTT sed while FXlla-activity sed dose-dependently at decreasing doses of the MAb 3F? (Table 10, Figures 8 and 10). The control MAb showed no protection from FeCig-induced occlusion of the arteria carotis and aPTT, PT and FXlla—activity values were unchanged (Tables 9, 10, Figures 7 to 10).

No. Treatment Occlusion rate 1 Isotonic saline 21/25 (84 %) 2 :; 0/10 (0 %) 3 20%;; 0/5 (0 %) 4 11a:5:; 0/10 (0 %) 1211:?gig7 0/10 (0 %) 6 $4392 3/10 (30 %) 7 “fig/31:: 4/10 (40 %) 8 34313:; 6/10 (60 %) Control MAb 9 8/10 (80 %) mg/kg N0. ent PT FI- activity 1 icsaline 8.99_+.1.13 31.193397 71.13i14.64 MAb3F7 2 10.58i1.12 116.70i29.96 0.63i1.16 30mg/kg MAb 3137 3 11.68i0.98 137.90i7.74 0.00i0.00 ”mg/kg MAb3F7 4 9.74:0.57 124.70i24.41 O.94i].29 10mg/kg MAb 31:7 10.43i0.92 91.72i16.89 3.]7i0.92 “mg/kg MAb3F7 6 9.14i0.33 11.05 7.68i1.59 25ng MAb3F7 7 9.51i0.61 39.84i5.83 30.02i10.00 1mg/kg MAb 31:7 8 9.61i0.60 35.89:3.73 37.22i7.92 0.5 mg/kg ControlMAb 9 .28 29.33i2.52 59,70i1254 30mg/kg Discussion: This study demonstrated that mice were fully protected against arterial thrombosis after intravenous treatment with the MAb 3F7 at a dose of 5 mg/kg or higher. At decreasing doses, occlusion rates increased while times to occlusion decreased dose—dependently. Compared to controls, PT was unchanged while aPTT and FXlIa—activity were ependently prolonged and decreased, respectively. in summary, MAb 3F7 trated a remarkable efficacy profile and a desirable dose-response relationship. e 8: s of anti-FXlla monoclonal antibody 3F7 on hemostasis in a subaquatic bleeding model in mice e 7 had demonstrated that MAb 3F7 fully prevents FeClg—induced arterial thrombosis in mice at doses of 30 - 5 mg/kg. In addition to this effect, FXlla—activity was nearly completely inhibited and aPTT prolonged up to fourfold at these protective doses. In order to clarify the question whether such effects may influence physiological hemostasis, the aim of this study is to igate MAb 3F7 with regard to its effect on hemostasis in the murine tail tip bleeding model at the ‘10 lowest fully protective dose (i.e. 5 mg/kg) as well as 5 fold beyond this dose (i.e. 25 mg/kg).

Methods Treatent No. Treatment Dose / volume / schedule / route N (m/f) i isotonic saline N.a1./ 0.1 mL/ZO g b.w. /t=-5 min. /i.v. 10 (0/10) 2 MAb 3F7 5 mg/kg / 0.1 mL/ZO g b.w. /t=~5 min. / LV. 10 (0/10) 3 MAb 3F7 25 mg/kg / 0.1 mL/ZO g b.w. /t=-5 min. / iv. 10 (0/10) Female NMRI mice were obtained from Charles River Laboratories (Kisslegg).

They were 6 to 8 weeks old and weighed 25 to 32 g.

Hemostasis was determined in a subaquatic model. in brief, tail tip bleeding parameters were determined by quantifying time to hemostasis and blood loss.

The volume of total blood loss was calculated by measuring the hemoglobin present in the saline used for submersion of the tail tip. The obin of the s was taken into consideration ingly. The tail tip cut was performed with a scalpel knife under deep anesthesia (Narcoren), removing about 3 mm of the tail tip. Immediately upon lesion, the tail tip was submerged in saline, which was kept at the physiological body temperature of the mice using a water bath. The observation period to monitor bleeding was 30 min. All test articles were administered i.v. at 5 min. prior to the start of the observation period (tail cut).

Results: ndent of group, all animals showed hemostasis within the observation period (Table 12). Time to hemostasis and total blood loss did not differ between the groups (Tables 12 and 13, Figures 11 to 14; Kruskal-Wallis test: ).

Table 12: Frequency and Time to asis Within 30 minutes following treatment with MAb 3F7 (n=10/group) Frequency of Treatment Time to hemostasis hemostasis Mean iSD Min. Med. Max. (sec.) (see) (sec.) (sec.) 1 57 i 49 Isotonic saline 10/10 (100 %) 70 125 360 178 i185 MAb 3F7 5 mg/kg 10/10 (100 %) 60 98 660 MAb 3F7 25 10/10 (100 %) 196 i144 30 163 450 ting/kg Table 13: Total blood loss following treatment with MAb 3F7 (n=10/grou13=)u Treatment Mean iSD (pL) Min. (pL) Median (uL) Max. (pL) 12.3 i9.5 2.1 10.5 270 lsotonic saline 7.0 i7.1 0.6 4.2 233 MAb 3F7 5 mg/kg MAb 3F7 25 mg/kg 9.9 1:103 2.1 5.1 30.8 Discussion: From the results of this study, it can be concluded that the two d doses of MAb 3F7 (5 and 25 mg/kg), potently preventing FeClg-induced arterial thrombosis in mice, had no effects on physiological hemostasis using the murine tail tip ng model.

Example 9: ison of aPTT of 3F7 and affinity-matured versions The activated partial thromboplastin time (aPTT) was determined in standard human plasma (SHP, Dade Behring), where different amounts of the respective inhibitor were added into logical saline to a total volume of 200 uL. 5O uL of this on were added to 50 uL Pathromtin SL (Dade Behring) and incubated for 1O 120 sec at 37°C. Subsequently, 50 uL of a calcium chloride solution (25 mM) were added to start the reaction.

The procedure was performed in a 808 XP (Behring Coagulation System) according to the conditions suggested by the manufacturer.

The aPTT of OT-2, MAb 3F7 and affinity-matured versions of MAb 3F7 was compared. The results are shown in Figure 15. The affinity-matured ns of MAb 3F7 were significantly more active than OT—2 and the original MAb 3F7.

Example 10: Comparison of the inhibition of Factor XlIa-alpha by different antibodies An inhibition assay was performed, essentially as described in Example 1(5) above.

In this case, 3F7, the affinity-matured 3F7 derivatives and OT-2 were compared in different molar ratios to human Factor Xlla-alpha, g from 1:0.1 to 1:10. The data are shown in Table 14 below and in Figure 16. 3F7 and the affinity—matured derivatives showed better tion than OT-2., and a higher amount of OT—2 was required to achieve maximal inhibition than of SF? and derivatives thereof.

Table 14: dy Ratio FXIIa-alpha:Antibody % Inhibition 3F7 1:0.1 35.5 1:0.2 62.5 1:0.5 91.9 1:1 97.9 1:2 100 1:5 100 1:10 100 3F7 1:0.1 38.3 1:0.2 66.8 1:0.5 91.4 1:1 96.1 1:2 100 1:5 100 1:10 100 VR115 0.1:1 39.4 0.2:1 72.8 0.5:1 100 1:1 100 1:2 100 1:5 100 1:10 100 VR112 1:0.1 39.9 1:0.2 68.7 1:0.5 99.7 1:1 100 1:2 100 1:5 100 1:10 100 VR110 1:0.1 33.7 1:0.2 66.9 1:0.5 100 1:1 100 1:2 100 1:5 100 1:10 100 VR24 1:0.1 34.5 1:0.2 67.3 1:0.5 99.7 1:1 100 1:2 100 1:5 100 1:10 100 OT-2 1:0.1 21.6 1:0.2 37.7 1:0.5 76.6 1:1 92.7 1:2 96.9 1:5 100 :1 100 in control 1:1 90.0

Claims (30)

Claims

1. An anti-Factor XII/XIIa monoclonal antibody or antigen-binding fragment thereof that has a more than 2 fold higher binding affinity to human Factor 5 XIIa-beta than to human Factor XII and that is capable of inhibiting the amidolytic activity of human Factor XIIa, wherein the binding affinity is measured by surface plasmon resonance-based technology, and wherein inhibiting the amidolytic activity of human Factor XIIa means an inhibition of at least 80% of the activity observed in a control ment without any tor 10 present.

2. An anti-Factor XII/XIIa monoclonal antibody or antigen-binding fragment thereof that has a more than 2 fold higher binding affinity to human Factor XIIa-beta than to human Factor XII, and that inhibits the amidolytic activity of 15 Factor XIIa-alpha by more than 50% in an in vitro amidolytic activity assay when used at a molar ratio of FXIIa-alpha to antibody of 1:0.2.

3. The dy or antigen-binding fragment thereof of claim 1 or claim 2, wherein the antibody binds murine FXII/FXIIa; and wherein the level of 20 g of the antibody to a polypeptide sing SEQ ID NO: 2 in which (a) the asparagine residue at position 398 of SEQ ID NO: 2 is substituted for lysine; or (b) the isoleucine residue at position 438 of SEQ ID NO: 2 is substituted for alanine, is lower than the level of binding of the protein to the corresponding polypeptide comprising SEQ ID NO: 2 without said substitution.

4. The antibody or antigen-binding fragment f of any of claims 1 to 3, wherein the antibody or antigen-binding fragment f comprises a heavy chain variable (vH) region which is more than 85% identical to the ce of SEQ ID NO: 4.

5. The antibody or antigen-binding fragment f of any of claims 1 to 4, wherein the antibody or n-binding fragment thereof comprises a light chain variable (vL) region which is more than 85% identical to the sequence of SEQ ID NO: 5.

6. The antibody or antigen-binding fragment thereof of any of claims 1 to 5, n the antibody or antigen-binding nt f comprises a heavy chain CDR1 at least 80% identical to the sequence of SEQ ID NO: 6, a heavy chain CDR2 at least 60% identical with SEQ ID NO: 7, and a heavy chain 5 CDR3 at least 80% identical to the sequence of SEQ ID NO: 9.

7. The dy or antigen-binding fragment thereof of any of claims 1 to 6, wherein the antibody or antigen-binding fragment f comprises a light chain CDR1 at least 50% identical with SEQ ID NO: 11, a light chain CDR2 of 10 SEQ ID NO: 12, and a light chain CDR3 of SEQ ID NO: 14.

8. The dy or antigen binding fragment thereof of any one of claims 1 to 7, wherein the antibody or antigen-binding fragment thereof binds human Factor XIIa-beta with a KD of at least 10-7M.

9. The antibody or antigen-binding fragment thereof of any one of claims 1 to 8, wherein the antibody or antigen-binding fragment thereof competes with Infestin for binding to human Factor XIIa-beta. 20

10. The antibody or antigen-binding nt thereof of any one of claims 1 to 9, wherein the antibody or antigen-binding fragment thereof is a human IgG or variant thereof.

11. The antibody or antigen-binding fragment thereof of claim 10, wherein the IgG 25 is IgG4.

12. A germlined dy according to any one of claims 1 to 11.

13. A bispecific antibody comprising one Fab region of an antibody of any one of 30 claims 1 to 12.

14. A nucleic acid encoding the antibody or antigen-binding fragment thereof of any one of claims 1 to 13. 35

15. A vector comprising the nucleic acid of claim 14, operably linked to a promoter sequence.

16. An isolated cell line or yeast cell comprising the vector of claim 15.

17. A method of producing the antibody or n-binding fragment thereof of any of claims 1 to 13, comprising culturing the cell line or yeast cell of claim 5 16 under appropriate conditions to express the antibody or antigen-binding fragment thereof and purifying the antibody or antigen-binding fragment thereof from the culture supernatant.

18. An antibody or antigen-binding fragment thereof of any of claims 1 to 13 for 10 medical use.

19. An antibody or antigen binding fragment f of any of claims 1 to 13 for use in preventing and/or treating a er selected from the group consisting of venous, al or capillary us formation, thrombus formation in the 15 heart, thromboembolism, thrombus formation during and/or after ting blood of a human or animal subject with artificial surfaces, by preventing the formation and/or the stabilization of thrombi and thereby three-dimensional intraluminal thrombus growth, or by preventing and/or treating uminal thrombi; interstitial lung disease, inflammation, a neurological inflammatory 20 disease, ment activation, olysis, angiogenesis and diseases related to FXII/ FXIIa-induced kinin formation or FXII/FXIIa-mediated complement tion.

20. The antibody or antigen-binding fragment thereof of claim 19, wherein the 25 venous or arterial thrombus formation is , myocardial infarction, deep vein thrombosis, portal vein thrombosis, thromboembolism, renal vein thrombosis, jugular vein thrombosis, cerebral venous sinus thrombosis, Budd- Chiari syndrome or Paget-Schroetter e. 30

21. The antibody or antigen-binding fragment thereof of claim 19, wherein the diseases related to FXII/FXIIa-induced kinin formation are selected from the group hereditary angioedema, bacterial infections of the lung, trypanosoma infections, hypotensive shock, pancreatitis, chagas disease, articular gout, arthritis, disseminated intravascular coagulation (DIC) and sepsis.

22. The antibody or antigen-binding fragment thereof of claim 19, wherein the interstitial lung disease is fibroproliferative and/or idiopathic pulmonary fibrosis. 5

23. The antibody or antigen-binding fragment thereof of claim 19, wherein the thrombus ion during and/or after contacting blood with artificial es occurs during and/or after a medical procedure performed on said human or animal subject and, wherein said antibody or antigen-binding fragment f is administered before and/or during and/or after said medical procedure, and 10 further n (i) the artificial surface is exposed to at least 80% of the blood volume of the subject and the artificial e is at least 0.2 m2 or (ii) the artificial surface is a container for collection of blood outside the body of the subject or 15 (iii) the cial surface is a stent, valve, intraluminal catheter, or a system for internal assisted pumping of blood.

24. A medical device coated with an antibody or antigen-binding fragment thereof of any of claims 1 to 13, wherein the device is a cardiopulmonary bypass 20 machine, an extracorporeal membrane oxygenation system for ation of blood, a device for assisted pumping of blood, a blood dialysis device, a device for the extracorporeal filtration of blood, a repository for use in the collection of blood, an intraluminal catheter, a stent, an artificial heart valve, and/or accessories for any one of said devices including tubing, cannulae, 25 centrifugal pump, valve, port, and/or diverter.

25. The antibody or antigen-binding fragment thereof of any of claims 1 to 13 for administration in a patient ing a medical procedure, n the medical procedure comprises contact with at least one of: 30 (a) heart, (b) at least one blood vessel chosen from: the aorta, the aortic arch, a carotid , a coronary artery, ocephalic artery, vertebrobasilar circulation, intracranial arteries, renal artery, a hepatic artery, a mesenteric artery, and/or a blood vessel of the arterial system cranial to the heart, (c) a venous blood vessel if the patient has a known septal defect; and wherein the medical procedure comprises release of at least one embolus 5 in at least one of said blood vessels in the body that could result in ischemia in at least one target organ and administration of the dy or nbinding fragment thereof before, during, and/or after the medical procedure.

26. The antibody or n-binding nt thereof of any of claims 1 to 13 for 10 use in the tion or treatment of a condition associated with increased retinal vascular permeability, including progressive retinopathy, sightthreatening complication of retinopathy, macular edema, non-proliferative pathy, proliferative retinopathy, retinal edema, diabetic retinopathy, hypertensive retinopathy, and retinal trauma.

27. A pharmaceutical composition comprising the antibody or antigen-binding fragment thereof of any of claims 1 to 13.

28. Use of an antibody or antigen-binding fragment thereof of any one of claims 1 20 to 13 for the manufacture of a medicament for the treatment and/or prevention of a disorder selected from the group consisting of venous, arterial or capillary thrombus formation, thrombus ion in the heart, thromboembolism, thrombus formation during and/or after contacting blood of a human or animal subject with artificial surfaces, by ting the formation and/or the 25 ization of thrombi and thereby three-dimensional intraluminal thrombus growth, or by ting and/or treating intraluminal thrombi; interstitial lung disease, inflammation, a neurological inflammatory disease, complement activation, olysis, angiogenesis and diseases related to FXII/ nduced kinin formation or FXII/FXIIa-mediated complement activation.

29. Use of an antibody or antigen-binding fragment thereof of any one of claims 1 to 13 for the manufacture of a medicament for prevention or treatment of a condition associated with increased retinal vascular permeability, including progressive retinopathy, sight-threatening complication of retinopathy, macular edema, non-proliferative retinopathy, proliferative pathy, retinal edema, diabetic retinopathy, hypertensive retinopathy, and retinal trauma.

30. An anti-Factor XII/XIIa monoclonal antibody or antigen-binding fragment 5 thereof of any one of claims 1 to 11 or claims 18 to 23 or claims 25 to 26, or a ned antibody of claim 12, or a bispecific antibody of claim 13, or a nucleic acid of claim 14, or a vector of claim 15, or a cell line or yeast cell of claim 16, or a method of producing the antibody or fragment thereof of claim 17, or a medical device coated with an antibody or antigen-binding fragment 10 thereof of claim 24, or a pharmaceutical composition of claim 27, or a use of claim 29 or 30, substantially as bed herein.