PROTEOLYTICALLY CLEAVABLE FUSION PROTEINS WITH HIGH MOLAR SPECIFIC ACTIVITY

Abstract

The invention relates to therapeutic fusion proteins in which a coagulation factor is fused to a half-life enhancing polypeptide, and in which both are connected by a linker peptide that is proteolytically cleavable. The cleavage of such linkers liberates the coagulation factor from activity-compromising steric hindrance caused by the half-life enhancing polypeptide and thereby allows the generation of fusion proteins may show relatively high molar specific activity when tested in coagulation-related assays. Furthermore, the fact that the linker is cleavable can enhance the rates of inactivation and/or elimination after proteolytic cleavage of the peptide linker compared to the rates measured for corresponding therapeutic fusion proteins linked by the non-cleavable linker having the amino acid sequence GGGGGGV.

Claims

1-33. (canceled)

34. A method of administering an effective amount of a coagulation factor fusion protein to a patient in need thereof, wherein the coagulation factor fusion protein comprises: a) a von Willebrand Factor; b) a half-life enhancing polypeptide (HLEP), and c) a peptide linker which joins the von Willebrand Factor and the half-life enhancing polypeptide; wherein the HLEP is an immunoglobulin without an antigen binding domain, wherein the peptide linker is cleavable by one or more proteases involved in coagulation or activated by coagulation enzymes, and wherein the peptide linker is cleaved during a coagulatory event; and wherein the patient suffers from a blood coagulation disorder.

35. The method of claim 34, wherein said fusion protein has at least one of the following properties, in comparison to the respective therapeutic fusion protein linked by a non-cleavable linker having the amino acid sequence GGGGGGV (SEQ ID NO: 94): i) an increased molar specific activity in at least one coagulation-related assay, ii) an increased inactivation rate of the activated coagulation factor after the peptide linker is proteolytically cleaved in a coagulation-related mode, and iii) an increased elimination rate after the peptide linker is proteolytically cleaved in a coagulation-mode.

36. The method of claim 34, wherein said fusion protein has a higher in vivo recovery compared to the in vivo recovery of a von Willebrand Factor when not fused to a half-life enhancing polypeptide.

37. The method of claim 34, wherein said fusion protein has an increased half-life in plasma compared to the half-life in plasma of a von Willebrand Factor when not fused to a half-life enhancing polypeptide.

38. The method of claim 34, wherein the molar specific activity of the fusion protein is increased at least 25% compared to that of the respective fusion protein linked by a non-cleavable linker consisting of the amino acid sequence GGGGGGV (SEQ ID NO: 94) in at least one coagulation-related assays.

39. The method of claim 34, wherein the inactivation rate of the von Willebrand Factor after cleavage of the peptide linker which links the von Willebrand Factor to the half-life enhancing polypeptide is increased at least 10% as compared to the inactivation rate of a von Willebrand Factor in a respective fusion protein linked by a non-cleavable linker consisting of the amino acid sequence GGGGGGV (SEQ ID NO: 94).

40. The method of claim 34, wherein the elimination rate of the von Willebrand Factor after cleavage of the peptide linker which links the von Willebrand Factor to the half-life enhancing polypeptide is increased by at least 10% as compared to the elimination rate of a von Willebrand Factor in a respective fusion protein linked by a non-cleavable linker consisting of the amino acid sequence GGGGGGV (SEQ ID NO: 94).

41. The method of claim 34, wherein the peptide linker is cleavable by a protease that naturally activates FVII or FVIII in vivo.

42. The method of claim 34, wherein the kinetics of linker cleavage by the protease is not delayed by more than a factor of 3 compared to the activation kinetics of the coagulation factor.

43. The method of claim 34, wherein the linker is cleaved by thrombin during coagulation.

44. The method of claim 34, wherein the blood coagulation disorder is hemophilia A.

45. The method of claim 34, wherein the peptide linker is cleavable by a protease that is activated directly or indirectly by a coagulation factor that is activated during a coagulation event.

46. A method of administering an effective amount of a coagulation factor fusion protein to a patient in need thereof, comprising: (a) administering a composition comprising said coagulation factor fusion protein, (b) administering a composition comprising a polynucleotide encoding said coagulation factor fusion protein via a gene therapy protocol, or (c) administering a composition comprising a plasmid or vector comprising a polynucleotide encoding said coagulation factor fusion protein via a gene therapy protocol, wherein the coagulation factor fusion protein comprises: a) a von Willebrand Factor; b) a half-life enhancing polypeptide (HLEP), and c) a peptide linker which joins the von Willebrand Factor and the half-life enhancing polypeptide; wherein the HLEP is an immunoglobulin without an antigen binding domain, wherein the peptide linker is cleavable by one or more proteases involved in coagulation or activated by coagulation enzymes, and wherein the peptide linker is cleaved during coagulation; and wherein the patient suffers from a blood coagulation disorder.

47. The method of claim 46, wherein the administration comprises administering via a gene therapy protocol (a) a composition comprising a polynucleotide encoding said coagulation factor fusion protein or (b) a composition comprising a plasmid or vector comprising a polynucleotide encoding said coagulation factor fusion protein.

48. The method of claim 46, comprising administering a composition comprising said fusion protein.

49. The method of claim 34, wherein the von Willebrand Factor or the immunoglobulin comprises a sequence that is 95% identical to the sequence of a wild-type human von Willebrand Factor or a wild-type human immunoglobulin, respectively.

50. The method of claim 34, wherein the von Willebrand Factor or the immunoglobulin comprises a sequence that is identical to the sequence of a wild-type human von Willebrand Factor or a wild-type human immunoglobulin respectively.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0098] FIG. 1: In vitro activation of FIX-albumin fusion proteins by FXIa at 37 C. at a molar ratio of FXIa to fusion protein of about 1:500. One fusion protein with non-cleavable linker (1478/797) and two fusion proteins with cleavable linker (1088/797 and 1089/797) were used. Samples were analyzed by SDS-PAGE under reducing conditions followed by Coomassie blue staining

[0099] FIG. 2: Pharmacokinetics of activated rec FIX and FIX-albumin fusion proteins with and without cleavable linker in comparison to non-activated fusion proteins.

[0100] FIG. 3: Inactivation of activated rec FIX or FIX-albumin fusion protein by AT. Residual FIX activity was determined after 120 min using a non-activated partial thromboplastin time assay.

EXAMPLES

Example 1: Generation of cDNAs Encoding FIX and FIX-Albumin Fusion Proteins

[0101] Factor IX coding sequence was amplified by PCR from a human liver cDNA library (ProQuest, Invitrogen) using primers We1403 and We1404 (SEQ ID NOS:5 and 6). After a second round of PCR using primers We1405 and We1406 (SEQ ID NOS:7 and 8) the resulting fragment was cloned into pCR4TOPO (Invitrogen). From there the FIX cDNA was transferred as an EcoRI Fragment into the EcoRI site of pIRESpuro3 (BD Biosciences) wherein an internal XhoI site had been deleted previously. The resulting plasmid was designated pFIX-496 and was the expression vector for factor IX wild-type.

[0102] For the generation of albumin fusion constructs the FIX cDNA was reamplified by PCR under standard conditions using primers We2610 and We2611 (SEQ ID NOS:9 and 10) deleting the stop codon and introducing an XhoI site instead. The resulting FIX fragment was digested with restriction endonucleases EcoRI and XhoI and ligated into an EcoRI/BamH1 digested pIRESpuro3 together with one XhoI/BamH1 digested linker fragment as described below.

[0103] Two different glycine/serine linker fragments without internal cleavage sites were generated: Oligonucleotides We2148 and We2150 (SEQ ID NOS: 11 and 12) were annealed in equimolar concentrations (10 pmol) under standard PCR conditions, filled up and amplified using a PCR protocol of a 2 min. initial denaturation at 94 C. followed by 7 cycles of 15 sec. of denaturation at 94 C., 15 sec. of annealing at 55 C. and 15 sec. of elongation at 72 C., and finalized by an extension step of 5 min at 72 C. The same procedure was performed using oligonucleotides We2156 and We2157 (SEQ ID NOS: 13 and 14). The resulting linker fragments were digested with restriction endonucleases XhoI and BamH1 and used separately in the above described ligation reaction. The resulting plasmids therefore contained the coding sequence for FIX and a C-terminal extension of a glycine/serine linker.

[0104] Two different cleavable linker fragments derived from the activation sites of FIX were generated: Oligonucleotides We2335 and We2336 (SEQ ID NOS:15 and 16), containing the activation cleavage site of the FIX light chain/activation peptide border region, were annealed, filled, and amplified as described above. The resulting linker fragment was digested with restriction endonucleases XhoI and BamH1 and used in the above described ligation reaction. The resulting plasmid therefore contained the coding sequence for FIX and a C-terminal extension of a cleavable FIX sequence (amino acids 136 to 154 of SEQ ID NO:2). In a subsequent site directed mutagenesis reaction with a commercially available mutagenesis kit (QuickChange XL Site Directed Mutagenesis Kit, Stratagene) using oligonucleotides We2636 and We2637 (SEQ ID NOS:17 and 18) the XhoI site was deleted.

[0105] For generation of the second cleavable linker fragment derived from FIX, the same procedure was performed using oligonucleotides We2337 and We2338 (SEQ ID NOS: 19 and 20) for linker construction. The resulting linker fragment was digested with restriction endonucleases XhoI and BamH1 and used in the above described ligation reaction. The resulting plasmid now contained the coding sequence for FIX and a C-terminal extension of a cleavable FIX sequence derived from the activation cleavage site of the FIX activation peptide/heavy chain border region (amino acids 173 to 186 of SEQ ID NO:2). Oligonucleotides We2638 and We 2639 (SEQ ID NOS:21 and 22) were used for deletion of the XhoI site as described above.

[0106] In the next cloning step the above generated plasmids were digested with BamH1 and a BamH1 fragment containing the cDNA of mature human albumin was inserted. This fragment had been generated by PCR on an albumin cDNA sequence using primers We1862 and We1902 (SEQ ID NOS:23 and 24) under standard conditions.

[0107] The final plasmids with non-cleavable glycine/serine linkers were designated pFIX-980 (SEQ ID NO:30) and pFIX-986 (SEQ ID NO:31), respectively. The final plasmids with cleavable linkers derived from FIX sequences were designated pFIX-1088 (SEQ ID NO:40) and pFIX-1089 (SEQ ID NO:49), respectively. Their linker sequences and the C-terminal FIX and N-terminal albumin sequences are outlined below. Proteolytic cleavage sites within the linkers are indicated with arrows, the FIX derived linker sequences are underlined.

TABLE-US-00005 [00001]

[0108] For expression in CHO cells the coding sequences for the FIX albumin fusion protein were transferred into vectors pIRESneo3 (BD Biosciences) or pcDNA3.1 (Invitrogen), respectively.

[0109] Using the above protocols and plasmids and by applying molecular biology techniques known to those skilled in the art (and as described e.g. in Current Protocols in Molecular Biology, Ausubel F M et al. (eds.), including Supplement 80, published October 2007, John Wiley & Sons, Inc.) other constructs can be made with insertions of different linker sequences, e.g. as described in tables 3a and 3b.

[0110] For efficient processing of the propeptide in cells expressing FIX in high amounts coexpression of furin is required (Wasley L C et al. 1993. PACE/Furin can process the vitamin K-dependent pro-factor IX precursor within the secretory pathway. J. Biol. Chem. 268:8458-8465). Furin was amplified from a liver cDNA library (Ambion) using primers We1791 and We1792 (SEQ ID NOS:25 and 26). A second round of PCR using primers We1808 and We1809 (SEQ ID NOS:27 and 28) yielded a furin fragment where the carboxyterminal transmembrane domain (TM) was deleted and a stop codon introduced; this fragment was cloned into pCR4TOPO (Invitrogen). From there the furinTM cDNA was transferred as an EcoRI/NotI Fragment into the EcoRI/NotI sites of pIRESpuro3 (BD Biosciences) wherein an internal XhoI site had been deleted previously. The resulting plasmid was designated pFu-797. This plasmid was cotransfected with all FIX constructs in a 1:5 (pFu-797:pFIX-xxx) molar ratio. The amino acid sequence of the secreted furin encoded by pFu-797 is given as SEQ-ID NO:29.

Example 2: Transfection and Expression of FIX and FIX-Albumin Fusion Proteins

[0111] Plasmids were grown up in E. coli TOP10 (Invitrogen) and purified using standard protocols (Qiagen). HEK-293 cells were transfected using the Lipofectamine 2000 reagent (Invitrogen) and grown up in serum-free medium (Invitrogen 293 Express) in the presence of 50 ng/ml Vitamin K and 4 g/ml Puromycin. Transfected cell populations were spread through T-flasks into roller bottles or small-scale fermenters from which supernatants were harvested for purification.

[0112] Alternatively, CHO K1 or DG44 cells (Invitrogen) were transfected using the Lipofectamine 2000 reagent (Invitrogen) and grown up in serum-free medium (Invitrogen CD-CHO) in the presence of 50 ng/ml Vitamin K and 500-750 ng/ml Geneticin. High expressing clones were selected and spread through T-flasks into roller bottles or small-scale fermenters from which supernatants were harvested for purification.

Example 3: Purification of FIX and FIX-Albumin Fusion Proteins

[0113] Cell culture harvest containing FIX or FIX albumin fusion protein was applied on a Q-Sepharose FF column previously equilibrated with 50 mM TrisxHCl/100 mM NaCl buffer pH 8.0. Subsequently, the column was washed with equilibration buffer containing 200 mM NaCl. Elution of the bound FIX or FIX fusion protein was achieved by a salt gradient using 50 mM TrisxHCl/200 mM NaCl buffer pH 8.0 as a basis. The eluate was further purified by column chromatography on a hydroxylapatite resin. For this purpose, the eluate of the Q-Sepharose FF column was loaded on a hydroxylapatite chromatography column equilibrated with 50 mM TrisxHCl/100 mM NaCl buffer pH 7.2. The column was washed with the same buffer and FIX or FIX-HSA were eluted using a potassium phosphate gradient at pH 7.2. The eluate was dialyzed to reduce the salt concentration and used for biochemical analysis as well as for determination of the pharmacokinetic parameters. FIX antigen and activity were determined as described in example 5.

Example 4: Alternative Purification Scheme of FIX and FIX-Albumin Fusion Proteins

[0114] As described in example 3, cell culture harvest containing FIX or FIX albumin fusion protein was purified by chromatography on Q-Sepharose FF. The Q-Sepharose eluate was further purified by chromatography on a Heparin-Fractogel column. For this purpose, the Heparin-Fractogel column was equilibrated using 50 mM Tris x HCl, 50 mM NaCl pH 8.0 buffer (EP), the Q-Sepharose FF eluate was applied and the column was washed with equilibration buffer containing 75 mM NaCl. FIX or FIX albumin fusion protein, respectively, was eluted using EP adjusted to 300 mM NaCl.

[0115] The Heparin-Fractogel eluate was further purified by chromatography on a hydroxylapatite chromatography column as described in example 3. The purified FIX resp. FIX albumin fusion protein concentrate was subjected to FIX activity and antigen determination according to example 5 and characterized by further in vitro and in vivo investigations.

Example 5: Determination of FIX Activity and Antigen

[0116] FIX activity was determined as clotting or coagulation activity (FIX:C) using commercially available aPTT reagents (Pathromtin SL and FIX depleted plasma, Dade Behring). An internal substandard calibrated against the WHO International FIX concentrate Standard (96/854) was used as a reference.

[0117] FIX antigen (FIX:Ag) was determined by an ELISA acc. to standard protocols known to those skilled in the art. Briefly, microtiter plates were incubated with 100 L per well of the capture antibody (Paired antibodies for FIX ELISA 1:200, Cedarlane, but other sources of appropriate antibodies may also be applied) overnight at ambient temperature. After washing plates three times with washing buffer B (Sigma P3563), each well was incubated with 200 L blocking buffer C (Sigma P3688) for one hour at ambient temperature. After another three wash steps with buffer B, serial dilutions of the test sample in buffer B as well as serial dilutions of a substandard (SHP) in buffer B (volumes per well: 100 L) were incubated for two hours at ambient temperature. After three wash steps with buffer B, 100 L of a 1:200 dilution of the detection antibody (Paired antibodies for FIX ELISA, peroxidase labelled, Cedarlane) in buffer B were added to each well and incubated for another two hours at ambient temperature. After three wash steps with buffer B, 100 L of substrate solution (TMB, Dade Behring, OUVF) were added per well and incubated for 30 minutes at ambient temperature in the dark. Addition of 100 L undiluted stop solution (Dade Behring, OSFA) prepared the samples for reading in a suitable microplate reader at 450 nm wavelength. Concentrations of test samples were then calculated using the standard curve with standard human plasma as reference.

Example 6: Comparison of FIX-Activity/FIX-Antigen Ratio of Different FIX-Albumin Fusion Proteins in Cell Culture Supernatant

[0118] Cell culture supernatants of HEK cells transfected with DNA constructs coding for FIX-albumin fusion proteins that contained different linker peptides were subjected to FIX activity and antigen testing as described above (see example 5). The ratio of FIX:C to FIX:Ag was calculated representing a measure directly proportional to molar specific activity of the different constructs.

[0119] The results shown in table 4 indicate that there is an increase in activity/antigen ratio upon introduction of cleavable linkers into the FIX-HSA molecule. It also shows that the cleavable linker peptide should have a length of more than two amino acids in order to provide clearly increased activity/antigen ratios.

TABLE-US-00006 TABLE4 FIX:C/FIX:AgratiosofFIX-albuminfusionproteins containingdifferentlinkerpeptides Foldincrease comparedto fusionprotein 980/797with non-cleavable linker FIX-HSA (GGGGGGV) construct Linker FIX:C/FIX:Ag (SEQIDNO:94) 1182/797 None <0.031 1366/797 RI <0.068 1478/863 GGGGGGV 0.041 (Sheffieldetal.) (SEQIDNO:94) 980/797 SS(GGS).sub.7GS 0.070 1.7 (SEQIDNO:30) 986/797 SSNGS(GGS)3NGS 0.076 1.9 (GGS)3GGNGS (SEQIDNO:31) 1483/863 SVSQTSKLTRAETVFPDVD 0.688 16.8 GSGGS(SEQIDNO:95) 1088/797 SVSQTSKLTRAETVFPDVDGS 0.832 20.3 (SEQIDNO:39) 1365/797 SVSQTSKLTRAETVFPDVD 0.630 15.4 (SEQIDNO:36) 1482/863 SVSQTSKLTRAETVFP 0.482 11.8 (SEQIDNO:99) 1087/797 SVSQTSKLTRAETVFPDVDGS 0.472 11.5 (SEQIDNO:39) (FIXdeltaKLT) 1089/797 QSFNDFTRVVGGEDGS 0.532 13.0 (SEQIDNO:49) 1091/797 PERGDNNLTRIVGGQEGS 0.111 2.7 (SEQIDNO:109)

Example 7: Comparison of FIX and FIX-Albumin Fusion Proteins in Respect to Molar Specific Activity, Terminal In Vivo Half-Life and In Vivo Recovery in Rats or Rabbits

[0120] Purified recombinant wild type FIX (rFIX 496/797) and FIX-albumin fusion proteins (rFIX 980/797, rFIX 986/797, rFIX-1088/797 and rFIX 1089/797) were tested for FIX activity in a clotting assay as described above. In parallel, the difference of the optical density at 280 and 320 nm was determined as a measure for protein concentration (OD280-320). The ratios of activity per OD280-320 were calculated and based on the molar optical densities the molar specific activities were calculated. In the following table 5 the results are summarized.

TABLE-US-00007 TABLE 5 Molar specific activities of wt FIX compared to FIX-albumin fusions (linker sequences correspond to SEQ ID NOS 94, 30, & 31, respectively, in order of appearance) FIX Molar Optical clotting Activity/ specific density activity Vol/OD activity* Linker (OD280-320) (IU/mL) (IU/mL/OD) (IU/nmol) rFIX, wt (496/797) 0.3798 21.2 55.8 4.23 rFIX-HSA GGGGGGV 2.9189 5.8 2.0 0.23 (non-cleavable, 1478/863 (Sheffield et al.) rFIX-HSA SS (GGS).sub.7 GS 1.1122 3.4 3.0 0.35 (non-cleavable, 980/797) rFIX-HSA SS IMGS (GGS)3 0.8107 3.2 4.0 0.45 (non-cleavable, 986/797) NGS (GGS)3 GGN G rFIX-HSA FXIa cleavable 0.3421 11.9 34.8 3.95 (cleavable, 1088/797) rFIX-HSA FXIa cleavable 0.4512 11.3 25.0 2.84 (cleavable, 1089/797) *Molar specific activity based on activity, optical density and the following molar optical densities: Molar optical density of FIX: OD(280 nm, 1 mol/L) = 75 810 Molar optical density of albumin: OD(280 nm, 1 mol/L) = 37 791 Molar optical density of FIX-albumin fusion protein: OD(280 nm, 1 mol/L) = 113 601

[0121] Taking the results summarized in Table 5 into account, two constructs that were generated according to the present invention show highly increased molar specific activities compared to the fusion proteins with non-cleavable linkers. In addition, the molar specific activity of these constructs was only moderately decreased compared to wild type rFIX.

[0122] In vitro investigations of the proteolytic cleavage reactions by Factor XIa (FXIa) confirmed that FIX-albumin fusion proteins containing a cleavable linker such as construct no. 1088/797 or 1089/797 are activated and in parallel the linker is cleaved resulting in release of the albumin moiety (FIG. 1). The fusion protein with non-cleavable linker did not show a corresponding release of the albumin moiety.

[0123] In the case of FVIIa as cleaving protease in the presence of tissue factor, the FIX-albumin fusion proteins 1088/797 or 1089/797 containing a cleavable linker also showed release of the albumin moiety in parallel to release of the FIX activation peptide (Data not shown).

[0124] In addition to determination of molar specific coagulation activity, the polypeptides no. 496/797, 980/797, 986/797, 1088/797 and 1089/797 described above were administered intravenously to narcotized CD/Lewis rats (6 rats per substance) and/or rabbits (4 rabbits per substance) with a dose of 50 IU/kg body weight. Blood samples were drawn prior to test substance administration and at appropriate intervals starting at 5 minutes after administration of the test substances. FIX antigen content was subsequently quantified by an ELISA assay specific for human Factor IX (see above). The mean values of the respective groups were used to calculate in vivo recovery after 5 min. Half-lives for each protein were calculated using the time points of the beta phase of elimination (terminal half-life) according to the formula t.sub.1/2=ln 2/k, whereas k is the slope of the regression line obtained upon plotting FIX:Ag levels in logarithmic scale and time in linear scale.

[0125] Calculated in vivo half-lives are summarized in table 6. In rats as well as in rabbits the in vivo half-lives of the FIX-albumin fusion proteins were found to be significantly increased in comparison to non-fused wild-type recombinant FIX prepared in-house or in comparison to the commercially available recombinant FIX product BeneFIX. The in vivo half-lives of the albumin fusion proteins compared to BeneFIX were increased to about 200-400%, depending on the animal species or construct used (Table 6).

[0126] To evaluate the in vivo recovery, the FIX antigen levels measured per mL of plasma at their maximum concentrations after intravenous administration (t=5 min) were related to the amount of product applied per kg. Alternatively, a percentage was calculated by relating the determined antigen level (IU/mL) 5 min post infusion to the theoretical product level expected at 100% recovery (product applied per kg divided by an assumed plasma volume of 40 mL per kg). The in vivo recoveries (IVR) of the FIX-albumin fusion proteins were significantly higher than the in vitro recoveries of rFIX (496/797) or BeneFIX (Table 7).

TABLE-US-00008 TABLE 6 Terminal in vivo half-lives of FIX preparations derived from recombinant expression (BeneFIX, rFIX 496/797) and FIX albumin fusion proteins (rFIX 980/797, rFIX 986/797, rFIX 1088/797, and rFIX 1089/797) after intravenous administration of 50 IU/kg into rats and/or 50 IU/kg into rabbits, respectively. Rat experiments PSR18-05, PSRC06-05, Rabbit experiment PSR02-05 PSK11-05 Terminal relative to Terminal relative to half-life BeneFIX half-life BeneFIX (h) [%] (h) [%] rFIX 496/797 4.5* 91 n.t. n.t. rFIX 980/797 11.6* 234 36.9 410 29.3 326 (2.sup.nd exp.) rFIX 986/797 10.5* 212 n.t. n.t. rFIX 1088/797 8.3* 168 30.3 337 rFIX 1089/797 10.5* 212 n.t. n.t. BeneFIX 4.95* (mean of 100 9.0 100 5.3 and 4.6) *Determined between 120 and 1440 min Determined between 4 and 96 h

TABLE-US-00009 TABLE 7 In vivo recoveries (amount of substance 5 minutes post administration) of recombinant FIX preparations (BeneFIX, rFIX 496/797) and FIX albumin fusion proteins (rFIX 1088/797, rFIX 1089/797) after intravenous administration of 50 IU/kg into rats. The percentage of in vivo recovery was calculated based on an assumed plasma volume of 40 mL/kg. rat expriment in vivo recovery relative to IU/dL per BeneFIX IU/kg/[%]* [%] rFIX 0.462/18.5 74.6 496/797 rFIX 1.034/41.4 166.5 1088/797 rFIX 1.063/42.5 171.2 1089/797 BeneFIX 0.621/24.8 100 *Calculated based on a plasma volume of 40 mL/kg

Example 8: In Vitro Activation of FIX Albumin Fusion Proteins with/without Cleavable Linker (1088/797 and 980/797) and Determination of Pharmacokinetics in Rats

[0127] FIX-albumin fusion proteins and rec FIX were activated in vitro using commercially available Factor XIa (Kordia). Briefly, identical molar amounts of FIX or FIX-albumin fusion protein (3.010.sup.6 mol/L) were activated at 37 C. in solution in the presence of FXIa (1.910.sup.8 mol/L) and CaCl.sub.2) (1.5 mmol/L) buffered at pH 6.8. After complete activation as shown by SDS-PAGE the reaction was stopped by addition of a 5 molar excess of C1-Inhibitor (Berinert P) based on the amount of FXIa. The samples were stored frozen below 70 C. until start of pharmacokinetic investigation.

[0128] A pharmacokinetic investigation of the activated FIX and the FIX-albumin fusion proteins was performed in rats as described in example 7 and the results were compared to a pharmacokinetic results covering non-activated fusion proteins. It turned out that the activated fusion proteins demonstrated significantly reduced half-lives as well as AUC's compared to the non-activated molecules (FIG. 2). Upon activation the FIX-fusion protein with cleavable linker (1088/797) showed a pharmacokinetic behaviour very similar to activated rec FIX (BeneFIX) whereas the activated fusion protein with non-cleavable linker (980/797) resulted in a higher initial as well as terminal half-life compared to activated fusion protein 1088/797 with cleavable linker. Therefore, in this example, the cleavable linker results in increased elimination of the coagulation factor after activation and, therefore, avoids accumulation of potentially thrombogenic, activated fusion proteins with extended half-lives.

Example 9: Comparison of FIX-Albumin Fusion Proteins with/without Cleavable Linker in Respect to Inactivation Rate of the Activated Coagulation Factors by Antithrombin III (AT)

[0129] FIX fusion proteins with (1088/797) and without (980/797) cleavable linker were activated by incubation with FXIa as described in example 8. The activated factors were incubated with AT for 120 min and residual FIXa activity was determined using a manual FIX clotting assay method without activation (naPTT, see below) acc. to Schnitger and Gross. As control samples the activated FIX-albumin fusion proteins were used in presence of the same amount of AT but without incubation.

[0130] The FIX activity was determined with the aid of a non-activated partial thromboplastin time assay (naPTT) using FIX deficient plasma from Dade Behring. The samples were prediluted in a buffer of pH 6.8 containing His, Gly, Sucrose, and Tween 80. The whole determination was performed using coagulometers acc. to Schnitger & Gross. A mixture of 0.1 ml F IX deficient plasma, 0.1 ml sample, and 0.1 ml of 0.1% Phospholipids (Rhone-Poulenc-Nattermann, 1:3 prediluted in imidazole buffer supplemented with 1% HSA) was incubated for 2 minutes at 37 C. The coagulation reaction was initiated by adding 0.1 ml 0.025 mol/l CaCl.sub.2) solution and the clotting time was determined.

[0131] FIG. 3 shows the results of a corresponding inactivation experiment. In the case of the fusion protein with cleavable linker (1088/797) an increase in clotting time from 210 to 540 sec (factor of 2.57) demonstrated an accelerated inactivation process of FIXa activity by AT compared to a fusion protein with non-cleavable linker (980/797) that only showed an increase from 196 to 411 sec (factor of 2.10). Most probably, the albumin residue sterically affects the AT dependent inactivation process in the case of the fusion protein with non-cleavable linker whereas in the case of the fusion protein with cleavable linker the albumin residue is cleaved off resulting in an accelerated inactivation by AT.

Example 10: Design of FIX-HSA Fusion Proteins with Reduced Immunogenicity

[0132] As there is with any fusion between two proteins a slight risk associated that a neoepitope is created around the fusion point it was investigated whether the linker region as described in table 3a and 3b could be modified in order to decrease this risk.

[0133] In the course of this investigation all proposed linker sequences and the adjacent regions of FIX and HSA were analyzed for potential T-cell epitopes by way of prediction of binding capabability to multiple MHC-II alleles. One of these approaches involved the PreDeFT analysis offered by the company EpiVax (146 Clifford St., Providence, R.I. 02903, USA) in which the input sequences were parsed into overlapping 9-mer frames where each frame overlaps the last by 8 amino acids. Each frame was then assessed for its ability to bind with a set of common HLA. These detailed findings were then summarized producing regional and overall assessments of immunogenic potential. Finally, any epitope clusters identified were screened against the non-redundant protein database at GenBank and EpiVax's own database of known MHC ligands and T-cell epitopes.

[0134] As a result of these in silico predictions the following FIX fusion proteins were cloned, expressed and purified:

TABLE-US-00010 SEQIDNO:113 FIX-PVSQTSKLTRAETVFPDV-HSA SEQIDNO:114 FIX-PSVSQTSKLTRAETVFPDV-HSA

Example 11: Neoantigenicity Test

[0135] FIX-HSA fusion proteins comprising linkers SEQ ID NO:113 or 114 can be shown to display a reduced immunogenicity compared to fusion proteins comprising different linkers or to display a comparable immunogenicity as compared to wild type factor FIX by the following neoantigenicity test.

[0136] Products to be compared are administered subcutaneously into rabbits with or in the absence of Freund's adjuvant. The endpoint assay is a native Western blot.

[0137] A suitable dose of either the FIX-HSA fusion with a linker with reduced immunogenicity or a FIX-HSA fusion with a linker of enhanced immunogenicity or FIX wild type can be administered as a slow bolus.

[0138] A sample can be taken about 30 to 40 days after the start of the immunization and be assayed by a Western blot method to ensure that the rabbits developed antibodies against each of the respective immunogens.

[0139] Antibodies against the test sample for which a potentially increased immunogenicity is to be measured are blocked by an excess of a control sample (e.g. wild type FIX) By doing so all of the antibodies present which formed against native epitopes are unable to react when that antibody is used as a probe in a Western blot. The test and the control sample are run on the Western blot membrane and the blocked antibody is used as a probe.

[0140] If there were epitopes on the test samples (here FIX-HSA fusion protein with a linker) but not in the control sample (wild type FIX) that caused antibody formation in the rabbits, these antibodies would be detected after blocking as residual antibodies which would react with the test sample, but not the control sample.

[0141] Likewise FIX-HSA with a linker having predicted reduced immunogenicity can be used as a control versus a FIX-HSA with a linker having predicted increased immunogenicity as a test sample.

[0142] Preferably the same Western blot is assayed with non-blocked antibodies raised against the test sample as a positive control. Here it is expected that test sample as well as control sample are detected in the Western blot assay.

[0143] If with blocked antibodies raised against the test sample both test and control sample are not detected it can be concluded that the test sample has no neoepitopes as compared to the control sample.

[0144] If with blocked antibodies raised against the test sample only the test sample but not the control sample is detected it can be concluded that the test sample has neoepitopes as compared to the control sample.

[0145] For doing the analysis the IgG fraction of the hyper immune pooled serum can be purified using a protein A column from Pierce or other suppliers with a bed volume for example greater than 1 ml according to the instructions of the supplier. Preferably the rabbit sera are delipidated for example with trichlorotrifluoroethane.

[0146] Blocking is done for example by mixing 0.1-1.00 mg of the purified antiserum with 0.1 to 100 mg of the control sample in for example a 0.5 to 5 ml centrifuge cup, bringing the volume up to 0.5 to 3 ml. The tubes are then rotated slowly at room temperature of a minimum of 2 hours.

[0147] Nonblocked control antibodies to be used as a positive control can be prepared in the same way except that no control sample is used for blocking.

[0148] All blocked and nonblocked antibodies can be added to 3 to 4.5 ml of 5% dry milk solution in TBS+0.1% Tween-20 before incubation with their respective membrane. Preferably Western blots are performed as native Western blots.

BRIEF DESCRIPTION OF THE FIGURES

[0149] FIG. 1: In vitro activation of FIX-albumin fusion proteins by FXIa at 37 C. at a molar ratio of FXIa to fusion protein of about 1:500. One fusion protein with non-cleavable linker (1478/797) and two fusion proteins with cleavable linker (1088/797 and 1089/797) were used. Samples were analyzed by SDS-PAGE under reducing conditions followed by Coomassie blue staining.

[0150] FIG. 2: Pharmacokinetics of activated rec FIX and FIX-albumin fusion proteins with and without cleavable linker in comparison to non-activated fusion proteins.

[0151] FIG. 3: Inactivation of activated rec FIX or FIX-albumin fusion protein by AT. Residual FIX activity was determined after 120 min using a non-activated partial thromboplastin time assay.