WO2013017555A1

WO2013017555A1 - Factor vii compositions with specific glycosylation for controlled half-life

Info

Publication number: WO2013017555A1
Application number: PCT/EP2012/064849
Authority: WO
Inventors: Guillaume CHEVREUX; Valegh FAID
Original assignee: Lfb-Biotechnologies
Priority date: 2011-08-01
Filing date: 2012-07-30
Publication date: 2013-02-07

Abstract

The present invention relates to factor VII(FVII and FVIIa)compositions wherein the relative abundance of glycan forms comprising at least one sialylated or non-sialylated N,N'-diacetyllactosamine (GalNAc(β1,4)GlcNAc) antenna is controlled.

Description

FACTOR VII COMPOSITIONS WITH SPECIFIC GLYCOSYLATION FOR CONTROLLED HALF-LIFE

The present invention relates to factor VII compositions wherein the relative abundance of glycan forms comprising at least one substituted or unsubstituted Ν,Ι - diacetyllactosamine (GahVAc(pi,4)GlcNAc) antenna is controlled.

Technological background

Blood coagulation is achieved according to cascading steps involving various zymogens present in the blood which, in the presence of certain co factors (platelet anionic phospholipids, calcium ions, zymogens, etc.), are converted by proteolytic cleavage into their activated form. This succession of coagulation steps, or cascade, is carried out according to two coagulation systems, called the extrinsic coagulation pathway and the intrinsic coagulation pathway, which result in the transformation of prothrombin into thrombin. The extrinsic pathway involves the intervention of factor VII zymogen (FVII) present in blood. However, FVII requires activation by FVIIa to initiate this coagulation cascade. FVIIa has low enzymatic activity until it is complexed, in an equimolar manner, in the presence of calcium ions, with tissue factor (TF) in contact with blood from damaged tissue. FVIIa thus complexed converts FX into FXa in the presence of calcium ions and platelet anionic phospholipids. FXa, in turn, converts prothrombin into thrombin, which converts FV into FVa. Thrombin also activates FXIII which generates FXIIIa. Thrombin, still in the presence of calcium ions, acts on fibrinogen by transforming it into fibrin. The presence of FXIIIa enables formation of a clot comprised of an adhesive, solid mesh of fibrin, which is gradually and slowly resorbed as the installation of consolidation scar tissue takes place, for which the fibrin mesh serves as a framework. This cross-linked fibrin is insoluble and cannot be attacked by fibrinolytic enzymes, at least during scar tissue development.

The intrinsic coagulation pathway also involves FVIIa. This pathway comprises a cascade of reactions resulting in the activation of thrombin by FXII. FXII, once activated, activates FXI to produce FXIa which, in turn, converts FIX into FIXa. FIXa participates in the activation of FX into FXa in the presence of its activated cofactor, FVIIIa, with which it forms a complex in the presence of platelet anion phospholipids and calcium ions. This activation cascade, resulting in the formation of FXa, is followed by the conversion of prothrombin into thrombin, a step which is common to both the intrinsic and extrinsic activation pathways. It should be noted that the presence of FXa or thrombin enables the activation of F VII into FVIIa.

Factor VII therefore plays a dominant role in mechanisms of extrinsic coagulation which result in the formation of a blood clot. FVIIa is used to treat hemophilia A and B with circulating inhibitors, i.e., in the presence of specific antibodies which limit or prevent the action of FVIII and FIX, respectively. FVIIa has the advantage of being able to act locally in the presence of TF released after tissue damage, causing hemorrhages, even in the absence of FVIII or FIX.

However, the half- life of FVIIa classically observed is several hours.

To overcome this disadvantage, there is therefore a need for a FVII composition with a controlled half-life. Summary of the invention

In this context, the inventors have unexpectedly demonstrated that a large proportion of N-glycan forms of FVIIa comprising one or more GalVAc(pi,4)GlcNAc antennae, which may be substituted or unsubstituted with a sialic acid, a fucose or a sulfate, are correlated with a very short FVIIa half-life. Thus, FVIIa compositions comprising more glycan forms of this type are characterized by distinctly shorter half- lives, which can be of the order of a few minutes.

The invention thus relates to a recombinant human factor VII composition, wherein at least 50%, preferably at least 70%, and in particular at least 90%, of the molecules of the N-glycan factor VII composition are forms of the complex type which comprise at least one GahVAc(pi,4)GlcNAc antenna substituted or unsubstituted with a sialic acid, a fucose or a sulfate.

The invention also relates to a recombinant human factor VII composition, wherein less than 15%, preferably less than 10%, and in particular less than 5% of the molecules of the N-glycan factor VII composition are forms of the complex type which comprise at least one GaLVAc(pi,4)Gl VAc antenna substituted or unsubstituted by a sialic acid, a fucose or a sulfate. Description of the figures

Figure 1: Examples of biantennary N-glycan forms linked to Asn¹⁴⁵ and Asn³²², which comprise at least one substituted or unsubstituted GaLVAc(pi,4)Gl VAc antenna.

Figure 2: Glycosylation profile of N-glycans analyzed by MALDI mass spectrometry of the inventive factor VII produced in the milk of transgenic goats.

Figure 3: Glycosylation profile of N-glycans analyzed by MALDI mass spectrometry of the reference factor VII, NovoSeven^®'

Figure 4: Glycosylation profile of N-glycans analyzed by MALDI mass spectrometry of factor VII produced in the milk of transgenic rabbits.

Figure 5: Kinetics of plasma concentrations of FVII: Ag after a single intravenous administration in rabbit (200 μg/kg, medium, n=6); comparison of milk of transgenic rabbits to NovoSeven^®.

Figure 6: Kinetics of plasma concentrations of FVII: Ag after a single intravenous administration in rabbit (200 μg/kg, medium, n=6); comparison of milk of transgenic goats to NovoSeven^®.

Detailed description of the invention

The term "FVII" includes polypeptides comprising the wild-type human FVII sequence (SEQ ID NO: 1 of 406 amino acids, corresponding to the FVII sequence without the signal peptide, and the propeptide of 60 amino acids, see SEQ ID NO: 2, for the complete sequence of 466 amino acids). FVII further comprises the natural allelic variations of factor VII which may exist.

Natural FVII comprises four distinct structural domains: the N-terminal γ- carboxylic (Gla) domain, followed by two epidermal growth factor (EGF)-like domains, and finally the C-terminal serine protease catalytic domain. The amino acid positions indicated below relate to the sequence SEQ ID NO: 1. Activation of FVII into FVIIa is characterized by breaking of the Arg¹⁵²-Ile¹⁵³ bond. Plasma FVIIa comprises several post-translational modifications (PTMs): the first 10 glutamic acid residues are γ- carboxylated; aspartic acid residue 63 (Asp⁶³) is partially β-hydroxylated; serine residues 52 and 60 (Ser and Ser ) are O-glycosylated and carry the motifs Xyl(al,3)o-₂Glc and Fuc, respectively; and asparagine residues 145 and 322 (Asn¹⁴⁵ and Asn³²²) are N-glycosylated mainly by structures of the complex type which are primarily bisialylated biantennary in composition. Several recombinant or plasmatic FVII glycoforms have been described (FR2901707; FR2904558; Appa et al. Thrombosis and Haemostasis 2010, 104(2), 243; Klausen et al. Molecular Biotechnology 1998, 9, pp. 195-204, Fenaille et al. Glycoconj. J. 2008, 25(9), pp. 827- 842).

The term "FVII" also includes FVII variants which have the same or greater biological activity in relation to the activity of the wild form, such variants notably include polypeptides that differ from wild FVII, by insertion, deletion, or substitution of one or more amino acids.

The term "factor VII" comprises uncleaved FVII zymogen (FVII) and activated

FVII (FVII a). Factor VII is used in the composition preferably in its activated form.

The term "biological activity of FVIIa" includes the ability to generate thrombin, for example on the surface of activated platelets. The activity of factor VII in the composition can be evaluated in various ways. For example, it can be measured by the ratio between the quantity of FVIIa determined by a coagulation test and the quantity of

FVII determined by immunoreactivity with anti-FVII antibodies.

"Biantennary N-glycan forms linked to Asn¹⁴⁵ and Asn³²², which comprise at least one substituted or unsubstituted GaLVAc(pi,4)GlcNAc antenna" comprise, for example, the forms represented in figure 1.

The N-glycan forms can be sialylated, i.e., comprise in the non-reducing terminal position one or more residues of sialic acid (or N-acetylneuraminic acids, NeuAc). Some of the sialic acid residues, preferably all, can involve a2,6 bonds with the Gal(pi,3/4)GlcNAc antennae. Some of the sialic acids can further involve a2,3 bonds. The sialylated or non- sialylated GaLVAc(pi,4)Gl VAc antennae can further be fucosylated at al,3 on the N-acetylglucosamine (GlcNAc) residue, and/or be sulfated at position 4 on the ^-acetylgalactosamine (GahVAc) residue, and/or be sulfated on the N- acetylglucosamine (GlcNAc) residue. The inventive human factor VII comprises two N-glycosylation sites on Asn¹⁴⁵ and Asn³²². On one N-glycosylation site, the glycan chains are N-linked to an asparagine residue. Each molecule of the inventive factor VII thus comprises two N-linked N- glycan chains. However, factor VII molecules of the composition generally do not have homogeneous glycosylation, i.e., not all of the N-glycan chains are identical; the composition is a heterogeneous mixture of various distinct N-glycan structures.

For the totality of the composition, and thus for the totality of the N-glycan chains of the composition, the quantity of each N-glycan form present in the composition can be determined.

The quantity of each N-glycan form can be determined experimentally by, for example, mass spectrometry or analysis by high-performance liquid chromatography coupled with fluorescence detection (NP-HPLC/FD) after coupling with a fluorophore, or by any other method known to those persons skilled in the art.

In the present invention, the N-glycan forms comprise advantageously at least one substituted or unsubstituted GaLVAc(pi,4)Gl VAc antenna. Indeed, it has been shown that the presence of such antennae reduce the plasma half-life of factor VII.

Thus, a factor VII such as NovoSeven^®, sold by Novo Nordisk, has a half-life approximately 2 hours.

A factor VII comprising an identifiable proportion of motifs such as those claimed in the present invention will have a shorter half-life, possibly as short as a few minutes, for example, less than 10 minutes.

Factor VII is a recombinant human factor VII obtained by genetic engineering from cells, or from transgenic animals such as goats or rabbits. The inventive factor VII can also be obtained by enzymatic modifications of plasma FVII after a series of complex deglycosylation and reglycosylation reactions using exoglycosidases and glycosyltransferases, respectively, in the presence of multiple cofactors.

In a preferred manner, factor VII (preferably in FVIIa form) is produced notably in the milk of a transgenic animal.

According to a preferred embodiment, human factor VII is produced in the milk of nonhuman transgenic mammals that are genetically modified to produce said glycoprotein. Preferably, the milk is of a transgenic rabbit or goat. The secretion of factor VII by the mammary glands, enabling its secretion in the milk of the transgenic mammal, involves the tissue-dependent control of the expression of factor VII. Such methods of control are well known to those skilled in the art. Expression is controlled by sequences that enable expression of the glycoprotein toward a particular tissue of the animal. Such sequences notably include whey acidic protein (WAP), β-casein and β-lactoglobulin promoter sequences and possibly signal peptide sequences. A method for extracting proteins of interest from the milk of transgenic animals is described in patent EP0264166.

The factor VII of the present invention, with a short half-life, is produced in a highly preferred way in the milk of a transgenic goat.

The inventive factor VII, with a short half-life, is useful in applications that do not require prolonged or persistent action.

In the present invention, the term "half-life" refers to the time taken by factor VII to lose half of its pharmacological activity. The concepts of short or long half-lives are in relation to the half- life of the factor VII called NovoSeven^®, sold by Novo Nordisk, which is roughly 2 hours under normal conditions of use.

The invention also relates to a recombinant human factor VII composition, wherein less than 15%, preferably less than 10%, and in particular less than 5% of the molecules of the N-glycan FVIIa composition are forms of the complex type comprising at least one GahVAc(pi,4)GlcNAc antenna substituted or unsubstituted by sialylation, fucosylation or sulfation.

Such a factor VII composition has a long half- life, i.e., of the order of a few hours. In particular, factor VII can be produced in the milk of a transgenic rabbit.

Such a composition may be used as a drug, in situations where the longest possible factor VII half-life is useful and/or necessary, in particular a half-life longer than that of factor VII sold under the name NovoSeven^®. NovoSeven^® is an activated recombinant human factor VII.

Particular pathologies within the scope of the treatments for which the use of such a factor VII composition can be envisaged include hemophilia A or B, acquired hemophilia and congenital factor VII deficiency. Such a composition may also be used as a product for preventing hemorrhages which can occur during surgical procedures.

The below examples illustrate the invention without limiting its scope. Examples

Example 1: MALDI-TOF mass spectrometry of various types of factor VII The objective is to quickly determine the N-glycosylation profile of a glycoprotein by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry. This study notably aims to demonstrate certain glycoforms that are absent or relatively rare in man and are likely to trigger an immune response. To this end, the glycoprotein of interest, desalted and dried, is subjected to total enzymatic N- deglycosylation using peptidyl-N-glycosidase F (PNGase F). After release, the N- glycans are finally purified by porous graphitic carbon solid-phase extraction (PGC- SPE), permethylated and then analyzed by MALDI-TOF mass spectrometry. The glycan species revealed in the mass profiles are identified by their mass/charge ratios (m/z), on the one hand by deducing their monosaccharide composition in terms of hexoses (Hex), N-acetylhexosamines (HexNAc), deoxyhexoses (dHex) and sialic acids, and on the other hand based on the rules of biosynthesis of N-glycans in mammals. Moreover, the relative intensities of the various signals make it possible to determine the coefficients of relative molar abundance of the major structures.

The glycosylation profiles of the following three types of FVII were analyzed by this method:

- the inventive rfiFVII produced in the milk of a transgenic goat,

- the reference rhFVII produced in BHK cells (NovoSeven^®), and

- the rhFVII produced in the milk of a transgenic rabbit. 1. Procedure

1.1 Enzymatic TV-deglycosylation and preparation of released V-glycans

The glycoproteins of interest are enzymatically N-deglycosylated by peptidyl-N- glycosidase F (PNGase F) and the N-glycans thus released are purified by porous graphitic carbon solid-phase extraction (PGC-SPE) and regenerated in acid medium.

1.2 Permethylation of the released oligosaccharides

Permethylation of oligosaccharides in the presence of soda, initially developed by Ciucanu and Kerek (1984) and later modified by Ciucanu and Costello (2003), consists in both blocking by methyl esterification the carboxylic functions of sialic acid residues and blocking by methyl etherification the hydroxyl functions of other saccharide residues. This procedure thus effectively stabilizes sialylated glycans, and more particularly glycans, which easily decompose during analysis by MALDI-TOF mass spectrometry. Permethylated glycans are thus detected with increased sensitivity and the same relative response in positive ionization mode, so that a quantitative glycosylation profile can be obtained.

1.2.1 Preparation of the sample test tubes

The dry residue of the purified oligosaccharides is first placed in solution using a small volume (100-200 μΐ) of a 1 : 1 (v/v) methanol- water mixture, assisted by a homogenization step consisting of several minutes of vigorous vibrating agitation. An aliquot fraction or the totality of the glycan extract is then transferred to a test tube with a 300 μΐ glass insert, equipped with a PTFE-Teflon-PTFE (polytetrafluoroethylene) stopper. If the totality of the glycan extract must be used, it is rinsed twice with 100 μΐ of a 1 : 1 (v/v) methanol- water mixture and the rinsing solutions are then transferred to the glass tube. The glycan sample is dried under a stream of compressed air for half an hour. If moisture is visible at the end of this period, one or more co evaporations of residual water are carried out in the presence of a small volume (-100 μΐ) of absolute acetone under a stream of compressed air at a maximum temperature of 30 °C, until a dry residue is obtained.

1.2.2 Procedure for permethylation of the oligosaccharides

The dry glycan sample is then taken up by adding 100 μΐ of a suspension of soda in dimethyl sulfoxide (DMSO, -200 mg/ml), prepared extemporaneously by means of a mortar and pestle washed beforehand with acetone. The alkaline suspension of glycans is vigorously agitated for 10 minutes, and then 80 μΐ of methyl iodide is added. The alkylation reaction is allowed to proceed under very high agitation for two hours, and 80 μΐ of methyl iodide is supplemented 30 minutes after the start of the reaction. The reaction is quenched by evaporation of the methyl iodide under a stream of compressed air at a maximum temperature of 35 °C for 30 minutes.

The dry alkaline residue is neutralized and placed in solution by adding 100 μΐ of a 50% (v/v) aqueous CH₃CO₂H solution. Excess diiodine, formed during the reaction, at times the cause of the brownish-pink color of the sample solution, is eliminated by adding 50 μΐ of a 100 mg/ml concentrated aqueous Na₂S₂0₃ solution. The permethylated oligosaccharides are then purified by reversed-phase, solid-phase extraction (RP-SPE, CI 8 phase). The C18 SPE cartridge is sequentially wetted and equilibrated by passing 1 ml methanol and two times 1 ml ultrapure water containing 3% (v/v) acetonitrile. The sample, extemporaneously neutralized, and the sample test tube rinsing solutions are loaded on the CI 8 SPE cartridge using ultrapure water containing 3% (v/v) acetonitrile. The sample is then washed by passing two times 1 ml ultrapure water containing 3% (v/v) acetonitrile, followed by 1 ml of a 10% (v/v) aqueous acetonitrile solution. The sample is finally eluted by passing three times 300 μΐ of a 75%) (v/v) aqueous acetonitrile solution. The eluate is dried under a stream of compressed air for 30 minutes, and then under vacuum.

1.3 MALDI-TOF-MS analysis

The mass spectra of molecular ions in MS mode and fragment ions in MS/MS mode are generated on a MALDI-TOF/TOF Autoflex II mass spectrometer (Bruker Daltonics, Bremen, Germany) controlled by the FlexControl 2.4 software. The spectra presented result from the sum of 3000 individual spectra.

Desorption is assured by a nitrogen laser beam operating at 20-50%) and 50-80%> of its maximum power in MS mode and MS/MS mode, respectively. In MS mode, the molecular ions are analyzed in positive reflectron mode with a delayed extraction time of either 120 (m/z<2500), 180 (1500<m/z<3500), or 350 nanoseconds (1500<m/z<8000). The MALDI-TOF mass spectra are externally calibrated using peaks corresponding to the monoisotopic ions of a peptide mass standard mixture including bradykinin fragment 1-7 (m/z=757.400), human angiotensin II (m/z=1046.542), human angiotensin I (m/z=1296.685), substance P (m/z=l 347.735), bombesin (m/z=1619,822), renin (m/z=1758.933) and adrenocorticotropic hormone (ACTH) fragments 1-17 (m/z=2093.087) and 18-39 (m/z=2465.199).

The permethylated oligosaccharide preparation (Q_prot<500 μg) is placed in solution in a 1 : 1 (v/v) methanol- water mixture at two dilutions (10 μΐ and 30 μΐ) and 0.5 μΐ of each preparation is deposited on a smooth stainless steel MALDI target plate and co-crystallized with 0.5 μΐ of a DHB (2,5-dihydroxybenzoic acid) matrix solution concentrated to 10 mg/ml in a 1 : 1 (v/v) methanol- water mixture containing 5 mM NaCH3C0₂. In the case of analysis in negative ionization mode, sodium doping of the matrix solution with NaCHsCC^ is omitted in order to prevent the phenomenon of ionic suppression of the signal of anionic molecular species. The acquisitions of spectra are repeated twice per sample and per dilution.

2. Analysis of results

The mlz ratios of the single-charge (z=l) pseudomolecular [M+Na]⁺ (positive mode) and molecular [M-H]^~ (negative mode) ions present in the mass spectra make it possible to deduce the apparent molecular mass "M" of the various glycan structures present, with a precision in mass lower than 100 ppm. Comparison between theoretical monoisotopic masses and experimental monoisotopic masses makes it possible to identify these structures on the one hand by deducing their monosaccharide composition in terms of hexoses (Hex), N-acetylhexosamines (HexVAc), deoxyhexoses (dHex) and sialic acids, and on the other hand based on the rules of biosynthesis of N-glycans in mammals. The relative intensities of the various signals make it possible to establish the coefficients of relative molar abundance of all of the major glycan species.

3. Discussion

The glycosylation profiles of the three types of factor VII are presented in figures 2 to 4, and the tables summarizing the intensities of the various peaks are presented below (tables 1 to 3).

Table 1: Peak intensities and glycosylation profile analysis (figure 2) of N- glycans, analyzed by MALDI mass spectrometry, of the inventive factor VII produced in the milk of transgenic goats.

Only glycospecies of relative molar abundance greater than 0.5% were taken into account for the calculations.

Key: A=Antenna substituting a GlcNac, F=Fucose, G=Galactose, GaN= V- acetylgalactosamine, GlcNac=N-acetylglucosamine, Man=mannose, NA= V- acetylneuraminic acid, NG=N-glycosylneuraminic acid, N/A=not applicable, N/I=not identified.

Table 2: Peak intensities and glycosylation profile analysis (figure 3) of N- glycans, analyzed by MALDI mass spectrometry, of the reference factor NovoSeven^®.

Key: A=Antenna substituting a GlcNac, F=Fucose, G=Galactose, GalNac/GaN=N-acetylgalactosamine, GlcNac/GlN=N-acetylglucosamine, Man=mannose, NA=N-acetylneuraminic acid, NG=N-glycosylneuraminic acid, P=phosphate group, PGlcNac=GlcNac( l-0)-phosphate group, Man6- PGlN=oligomannose structure with six mannose residues with a 6-O-(Gl VAc(al-0)- phosphoryl)-a-mannose residue, N/A=not applicable, N/I=not identified. Table 3: Peak intensities and glycosylation profile analysis (figure 4) of N- glycans, analyzed by MALDI mass spectrometry, of factor VII produced in the milk of transgenic rabbits.

Key: A=Antenna substituting a GlcNac, F=Fucose, G=Galactose, GalNac/GaN=N-acetylgalactosamine, GlcNac/GlN=N-acetylglucosamine, Man=mannose, NA=N-acetylneuraminic acid, NG=N-glycosylneuraminic acid, P=phosphate group, PGlcNac=GlcNac( l-0)-phosphate group, Man6- PGlN=oligomannose structure with six mannose residues with a 6-O-(Gl VAc(al-0)- phosphoryl)-a-mannose residue, N/A=not applicable, N/I=not identified.

The relative abundance in percentage of glycan forms comprising at least one substituted or unsubstituted N,N'-diacetyllactosamine (GaLVAc(pi,4)Gl VAc) antenna (GaNylated species) is calculated using the following formula by adding the relative intensities of the peaks corresponding to such species.

GaNylation in percentage is calculated with the following formula: n

(number of GaN) * (peak intensity)

j≤

n * ₁₀0

^ (number of A) * (peak intensity)

i=l

wherein

- n represents the number of peaks considered on the MALDI spectrum,

- "number of GaN" represents the number of ^-acetylgalactosamine motifs on the antennae of the glycan form corresponding to the peak, and

- "number of A" corresponds to the number of N-acetylglucosamine antennae of the glycan form corresponding to the peak.

The three types of factor VII considered have, respectively, relative abundances expressed in percentage of glycan forms comprising at least one substituted or unsubstituted N,N'-diacetyllactosamine (GaLVAc(pi ,4)Gl VAc) antenna (GaNylation) of:

- 100% for the inventive factor VII produced in the milk of a transgenic goat,

- 26% for the reference FVII produced in BHK cells (NovoSeven®), and

- 0% for FVII produced in the milk of a transgenic rabbit.

Example 2: Comparison of the pharmacokinetic profile of factor VII produced in the milk of transgenic rabbits and reference factor VII (NovoSeven^®)

The pharmacokinetic profiles of factor VII produced in the milk of transgenic rabbits versus recombinant factor Vila were compared in awake male New Zealand rabbits.

Male rabbits were treated by slow intravenous route with a single 200 μg FVIIa/kg dose of factor VII produced in the milk of transgenic rabbits or recombinant factor Vila NovoSeven^®, manufactured by Novo Nordisk, or an equivalent volume of physiological saline solution.

The blood samples were taken at the following times: during pretreatment (D_₆);

Di at T5 min, Ti5 mi_n, T₃₀ min, T45 mi_n, Ti , T₃ , T5 , Ts h and T₂₄ h-

The circulating plasma level of FVILAg was assayed using an immunoenzymatic method (ELISA).

The parameters measuring exposure (C_max, AUC, MRT), distribution (Vd) and elimination (Ti_/2 and CI) were calculated. Throughout the study (2 days), no mortality or sign of intolerance was observed in the rabbits of the treated groups compared to the controls.

This analysis showed that the pharmacokinetic parameters (Ti/₂, MRT, AUC and Cmax) of the group treated with factor VII produced in the milk of transgenic rabbits are statistically lower than those calculated in the group treated with recombinant FVIIa; these are correlated with the parameters Vd and CI of the group treated with factor VII produced in the milk of transgenic rabbits, higher than those calculated in the group treated with recombinant FVIIa. This suggests faster elimination from the circulating blood of factor VII produced in the milk of transgenic rabbits.

1. Materials and methods

The product to be tested is transgenic coagulation factor Vila from Fl-Ll milk of female transgenic rabbits.

The reference product is recombinant coagulation factor Vila (NovoSeven^®).

The control product is an injectable 0.9% sodium chloride solution.

The excipient used for reconstituting the vial of NovoSeven^® (1/4) was sterile water for injection; the preparation was dissolved for 2 minutes and then administered to the animals within 20 to 22 minutes after reconstitution.

Fifteen male KBL NZW rabbits served as the animal model.

2. Treatment

The products were injected according to the methods described in table 4.

The products tested were injected at a dose of 200 μg/kg FVIIa of factor VII from the milk of transgenic rabbits and of NovoSeven^®.

The reference group received an equivalent volume of 0.9% NaCl (0.35 ml/kg).

Table 4: Rabbit group distribution

¾mpejSmental,_: ;:I¾Miiber& _:of¾iimais . y DOSC -¾^· Volume group per group (Mg kg) injected (ml/kg)

6 rabbits

1 LFB-rFVIIa 200 0.29

(1 to 6)

6 rabbits

2 Recombinant FVIIa 200 0.35

(7 to 12)

3 rabbits

3 0.9% NaCl N/A 0.35

(13 to 15) N/A: not applicable

The product was slowly injected intraveneously into the awake animal, at a rate of 1 ml/min in the marginal ear vein, using a single-use needle/syringe. The intravenous route is on the one hand a classical route of administration in animal pharmacokinetic studies, and on the other hand it is the therapeutic route in man.

Samples of approximately 4 ml of blood were drawn into a Monovette^® tube containing sodium citrate as anticoagulant from the central ear artery of the awake animal at the following days and times:

• D_6 before administration of the product; the sample was compensated by an injection of 0.9% NaCl in an amount of 4 ml/rabbit to avoid hypovolemic stress (this can be carried out in pretreatment but not after treatment with the product under study),

• Di at T5 min, T₁₅ min, T30 min, T45 mi_n, Ti , T3 , T5 , Ts h and T₂₄ h- 3. Assay of factor VII: As

FVILAg is quantified by an immunoenzymatic method (ELISA).

Quantification is carried out on rabbit plasmas and on four aliquot fractions of the products sampled before injection.

i. Assay of human factor VII in the injected products

The product to be tested and the reference product are assayed in the aliquot fractions taken right before injection and frozen.

The fractions are prediluted to 1/20000 then diluted 1/2 in 1/2. They are tested in duplicate, independently, on four dilutions.

The results in U FVII:Ag/ml are the mean of all of the values obtained. ii. Assay of plasmas of rabbits having received human factor VII

All plasmas of animals treated with the test product or the reference product are assayed singly on at least two dilutions.

The results expressed in mU FVII:Ag/ml are the mean of at least two values obtained with two different dilutions with a CV<15%. Hi. Assay of plasmas of rabbits having received physiological saline solution

All plasmas of rabbits having received the control product are tested singly on two dilutions of 1/2 and 1/4. b. Determination of the elimination kinetics of the product

A study of the variation of plasma concentrations as a function of time makes it possible to determine the pharmacokinetic characteristics of a product and to compare them with those of another product.

Analysis of the experimental data (plasma concentration of the product for each sampling time) led to the mathematical expression of the decay curve obtained. i. Representation of the elimination curve over time

Immunological assays of factor Vila (FVILAg) made it possible to construct the elimination curves of the respective products over time.

For each product, mean plasma levels are calculated from the values obtained in six rabbits, as well as the corresponding standard deviations. Then, from the mean plasma levels at various sampling times, a curve is constructed with linear coordinates and semi- logarithmic coordinates. ii. Calculation of pharmacokinetic parameters

From the experimental data and modeling of the individual curves, various pharmacokinetic parameters were calculated: maximum concentration (C_max), elimination half-life (Ti/₂), area under the curve (AUC), mean residence time (MRT), total clearance (CI), volume of distribution (Vd).

Recovery was calculated individually for each rabbit by the formula:

recovery (%)=C_max (mIU/kg)/C_theoreticai (mlU/kg) x 100

with Ctheoreticai (mIU/kg)=dose administered (mIU/kg)/56 ml/kg (=theoretical volume of blood per kg for a rabbit).

4. Results

a. Quantification of human FVII: Ag i. Assay of FVII:Ag in the injected products

The calculation of kinetic parameters (clearance, volume of distribution) requires determination of the injected dose by the analytical method used for the rabbit plasma assays.

Consequently, an immunoenzymatic assay (ELISA) of the level of FVILAg in the two products was carried out from the aliquot fractions sampled before injection. Four fractions of each product were assayed.

The mean concentrations of FVILAg in U/ml in each of the two injected products are presented in table 5 below.

Table 5: Concentration in IUIml of FVILAg (mean±standard deviation; n=4)

ii. Assay of FVILAg in rabbit plasmas

In certain plasmas, the presence of fibrin and/or clots is detected but does not seem to be of consequence on human FVII levels.

The results are the mean of at least two values at two different dilutions with a CV<15%.

The plasmas of rabbits having received the product to be tested were assayed:

- for samples at D_₆, singly on two dilutions of 1/2 and 1/4,

- for samples at Di 5 min, singly on three dilutions from 1/40 to 1/160,

- for samples at Di 15 min, singly on three dilutions from 1/40 to 1/160 for rabbits 1 to 3 and then singly on three dilutions from 1/20 to 1/80 for rabbits 4 to 6 (dilutions adjusted following the results of rabbits 1 to 3), - for samples at Di 30 min, singly on three dilutions from 1/20 to 1/80 except for rabbit 1 assayed on three dilutions from 1/40 to 1/160 (assayed during the first test and then dilutions were adjusted for the other rabbits),

- for samples at Di 45 min, singly on three dilutions from 1/20 to 1/80 for rabbits 1 to 4 (assayed during the first tests) and then singly on three dilutions from 1/10 to 1/40 for rabbits 5 and 6 (dilutions adjusted following the results of rabbits 1 to 4),

- for samples at Di 1 h, singly on three dilutions from 1/10 to 1/40 except for rabbit 1 assayed on three dilutions from 1/20 to 1/80 (assayed during the first test and then dilutions were adjusted for the other rabbits),

- for samples at Di 3 h, singly on three dilutions from 1/10 to 1/40 for rabbit 1 (assayed during the first test) and then for rabbits 2 and 3 on three dilutions from 1/2 to 1/8 (adjustment of dilutions) and then for rabbits 4 to 6 on three dilutions from 1/4 to 1/16 (adjustment of dilutions), - for samples at Di 5 h to Di 24 h, singly on three dilutions from 1/2 to 1/8.

The plasmas of rabbits having received the reference product were assayed:

- for samples at D_₆, singly on two dilutions of 1/2 and 1/4,

- for samples at Di 5 min to Di 45 min, singly on three dilutions from 1/40 to 1/160 except for rabbit 7 at Di 45 min assayed on three dilutions from 1/20 to 1/80 (assayed during the first test and then the dilutions were adjusted for the other tests),

- for samples at Di 1 h, singly on three dilutions from 1/20 to 1/80,

- for samples at Di 3 h, singly on three dilutions from 1/10 to 1/40,

- for samples at Di 5 h, singly on three dilutions from 1/4 to 1/16 except for rabbit 7 assayed on three dilutions from 1/2 to 1/8 (assayed during the first test and then the dilutions were adjusted for the other tests),

- for samples at Di 8 h and Di 24 h, singly on three dilutions from 1/2 to 1/8.

The individual values of levels of human FVII in mU/ml in rabbit plasmas at various sampling times are presented in table 6 below. Table 6: Individual values of human FVII concentrations in mUlml in samples from 15 rabbits

b. Elimination kinetics

i. Comparison of elimination profiles

The elimination profiles of factor VII from the milk of transgenic rabbits and from recombinant FVIIa are presented in figure 5 with a semi- logarithmic scale.

ii. Comparison of pharmacokinetic parameters

The pharmacokinetic parameters of factor VII from the milk of transgenic rabbits and from recombinant FVIIa were calculated using the WinNonlin 5.0 software and are summarized in table 7. The results obtained were compared by the one-way ANOVA method using the Statgraphics plus 5.0 software.

Table 7: Pharmacokinetic parameters during and after single intravenous administration in rabbit (200 μ,ίξ/k , mean±SD, n=6)

N/A: not applicable; N/S: not significant Except for recovery, all of the pharmacokinetic parameters are significantly different between the two types of factor VII studied.

5. Discussion

After administration of a single dose of factor VII from the milk of transgenic rabbits and of recombinant FVIIa by intravenous route in the rabbit, the results are as follows:

• No mortality or clinical sign suggesting potential intolerance are observed the day of administration and throughout the study.

• The graph of the concentrations of factor VII in rabbit plasmas as a function of sampling time shows that the decay follows a bi-exponential model. This indicates that factor VII is distributed in, and eliminated from, rabbit plasma according to a model with two compartments.

• The pharmacokinetic parameters Ti/₂ and MRT of factor VII from the milk of transgenic rabbits are significantly different from those of the reference factor VII. • The C_max of factor VII from the milk of transgenic rabbits is significantly lower than the C_max of the reference factor VII.

• The "recovery" of factor VII from the milk of transgenic rabbits is not significantly different from that of the reference factor VII.

· A highly significant difference (p<0.05) is observed in the pharmacokinetic parameters AUC (F7TG<Novo), CI (F7TG>Novo) and Vd (F7TG>Novo). This suggests a faster elimination from the circulating blood of factor VII from the milk of transgenic rabbits. Example 3: Comparison of the pharmacokinetic profile of factor VII from the milk of transgenic goats and of the reference factor VII (NovoSeven^®)

The pharmacokinetic profiles of factor VII from the milk of goats versus the recombinant factor Vila NovoSeven^® were compared in awake male New Zealand rabbits.

Male rabbits, divided into three experimental groups (n=6 for groups 1 and 2, n=3 for group 3), were treated by slow intravenous route with a single 200 μg FVIIa/kg dose of each factor VII mentioned above or an equivalent volume of physiological saline solution.

The blood samples were taken at the following times: during pretreatment (D_₃);

Di at T5 min, Ti5 mi_n, T₃o min, T45 mi_n, Ti , T₃ , T5 , Ts h and T₂₄ h-

The parameters measuring exposure (C_max, AUC, MRT), distribution (Vd) and elimination (Ti_/2 and CI) were calculated.

Throughout the study (2 days), no mortality or sign of intolerance was observed. The pharmacokinetic parameters calculated, i.e., C_max, Ti_/2, AUC, CI, MRT and Vd of factor VII from the milk of goats are significantly different from those of the recombinant factor Vila (NovoSeven^®).

Under the experimental conditions used, exposure times and exposure levels of

FVII are lower for animals treated with factor VII from the milk of goats than for animals treated with recombinant factor Vila (NovoSeven^®). 1. Materials and methods

The product to be tested is transgenic coagulation factor Vila from the milk of female goats, in liquid form, at a concentration of 440 μg/ml (880 IU/ml), stored at -70 °C.

The reference product is the recombinant coagulation factor Vila NovoSeven^®, a white lyophilized powder in a glass vial, 4.8 mg/8.5 ml, stored at +2/+8 °C.

The control product is an injectable 0.9% sodium chloride solution, stored at room temperature.

Factor VII from the milk of a transgenic goat was thawed for 23 minutes at room temperature; the preparation was administered to the animals within 48 to 50 minutes after thawing began.

Sterile water for injection (8.5 ml) was used to reconstitute the vial of NovoSeven^®; dissolution was immediate and the preparation was administered to the animals within 13 to 15 minutes after reconstitution.

Fifteen KBL NZW rabbits, divided into three groups, served as the animal model.

2. Treatment

The products were injected according to the methods described in table 8.

The products tested were injected at a dose of 200 μg/kg FVIIa of factor VII from the milk of transgenic goats and of NovoSeven^®.

The reference group received an equivalent volume of 0.9% NaCl (0.45 ml/kg).

The volume of the product to be administered was adjusted as a function of the body weight of the animal recorded on the day of administration (Di).

Table 8: Animal group distribution

N/A: not applicable

The product was slowly injected intravenously into the awake animal, at a rate of 1 ml/min in the marginal ear vein, using a single-use needle/syringe.

• D_3 before administration of the product, the sample was compensated by an injection of 0.9% NaCl in an amount of 4 ml/rabbit to avoid hypovolemic stress,

• Di at T5 min, T₁₅ min, T30 min, T45 mi_n, Ti , T3 , T5 , Ts h and T₂₄ h-

After centrifugation (15 min, 3000 rpm, at room temperature 15-20 °C) the collected plasmas were divided into three or four aliquot fractions of approximately 0.3 ml in cryotubes and frozen at -70 °C in the 2 hours following sampling.

3. Assay of factor VII: Ag

FVILAg is quantified by an immunoenzymatic method (ELISA).

Quantification is carried out on rabbit plasmas and on three aliquot fractions of the products sampled before injection.

i. Assay of human factor VII in the injected products

The product to be tested and the reference product are assayed in the aliquot fractions taken right before injection and frozen (see section 7.1.5). The fractions are prediluted to 1/20000 then diluted 1/2 in 1/2. They are tested in duplicate, independently, on four dilutions. The results in U FVII:Ag/ml are the mean of all of the values obtained.

ii. Assay of plasmas of rabbits treated with the product to be tested or the reference product

All plasmas of animals treated with the test product or the reference product are assayed singly on at least two dilutions. The results expressed in mU FVII:Ag/ml are the mean of at least two values obtained with two different dilutions with a CV<15%.

Hi. Assay of plasmas of rabbits treated with physiological saline solution

All plasmas of animals treated with physiological saline solution are tested singly on two dilutions of 1/2 and 1/4. b. Determination of the elimination kinetics of the product

A study of the variation of plasma concentrations as a function of time makes it possible to determine the pharmacokinetic characteristics of a product and to compare them with those of another product. Analysis of the experimental data (plasma concentration of the product for each sampling time) led to the mathematical expression of the decay curve obtained.

i. Representation of the elimination curve over time

For each product, mean plasma levels are calculated from the values obtained in six animals, as well as the corresponding standard deviations. Then, from the mean plasma levels at various sampling times, a curve is constructed with semi-logarithmic coordinates.

ii. Calculation of pharmacokinetic parameters

From the experimental data and modeling of mean curves, the following pharmacokinetic parameters are calculated:

^■ maximum plasma concentration (C_max),

^■ elimination half- life (Ti_/2), ^■ area under the curve (AUC),

^■ mean residence time (MRT),

^■ total clearance (CI),

^■ volume of distribution at equilibrium (Vd),

■ Recovery (%)=C_max (mIU/kg)/C_theoreticai (mlU/kg) x 100,

with Ctheoreticai=dose administered (mIU/kg)/56 ml/kg (=theoretical volume of blood per kg for a rabbit).

4. Results

a. Quantification of human FVII:Ag i. Assay of FVII:Ag in the injected products

Consequently, an immunoenzymatic assay (ELISA) of the level of FVILAg in the two products was carried out from the aliquot fractions sampled before injection in the animals. Three fractions of each product were assayed.

The mean concentrations of FVILAg in IU/ml in each of the two injected products are presented in table 9 below.

Table 9: Concentration in IU/ml of FVILAg (mean±standard deviation; n=3)

ii. Assay of FVII:Ag in rabbit plasmas The plasmas of rabbits having received the product to be tested were assayed:

- for samples at D_₃ and Di 24 h, singly on two dilutions of 1/2 and 1/4,

- for samples at Di 5 min, singly on four dilutions from 1/40 to 1/320 for rabbits 1 and 2 (tested during the first tests) and then singly on four dilutions from 1/2 to 1/16 for all rabbits 1 to 6 (dilutions adjusted following the results of rabbits 1 and 2),

- for samples at Di 15 min, singly on four dilutions from 1/40 to 1/320 for rabbits 1 and 2 (tested during the first tests) and then singly on three dilutions from 1/2 to 1/8 for all rabbits 1 to 6 (dilutions adjusted following the results of rabbits 1 and 2),

- for samples at Di 30 min and Di 45 min, singly on four dilutions from 1/20 to 1/160 for rabbits 1 and 2 (tested during the first tests) and then singly on three dilutions from 1/2 to 1/8 for all rabbits 1 to 6 (dilutions adjusted following the results of rabbits 1 and 2),

- for samples at Di 1 h, singly on four dilutions from 1/10 to 1/80 for rabbits 1 and 2 (tested during the first tests) and then singly on three dilutions from 1/2 to 1/8 for all rabbits 1 to 6 (dilutions adjusted following the results of rabbits 1 and 2),

- for samples at Di 3 h and Di 5 h, singly on four dilutions from 1/2 to 1/16 for rabbits 1 and 2 (assayed during the first tests) and then singly on three dilutions from 1/2 to 1/8 for rabbits 3 and 4, then singly on two dilutions of 1/2 and 1/4 for rabbits 5 and 6 (dilutions adjusted following the results of the preceding rabbits),

- for samples at Di 8 h, singly on four dilutions from 1/2 to 1/16 for rabbits

1 and 2 (assayed during the first tests) and then singly on two dilutions from 1/2 to 1/4 for rabbits 3 to 6 (dilutions adjusted following the results of rabbits 1 and 2).

The plasmas of rabbits having received the reference product were assayed:

- for samples at D_₃ and Di 24 h, singly on two dilutions of 1/2 and 1/4,

- for samples at Di 5 min to Di 45 min, singly on three dilutions from 1/40 to 1/160,

- for samples at Di 1 h, singly on three dilutions from 1/20 to 1/80 for rabbits 7 and 8 (assayed during the first tests) and then singly on three dilutions from 1/40 to 1/160 for rabbits 9 to 12 (dilutions adjusted following the results of rabbits 7 and 8),

- for samples at Di 3 h, singly on three dilutions from 1/10 to 1/40, - for samples at Di 5 h, singly on three dilutions from 1/4 to 1/16,

- for samples at Di 8 h, singly on three dilutions from 1/2 to 1/8.

The individual values of levels of human FVII in mU/ml in rabbit plasmas at various sampling times are presented in table 10 below.

Table 10: Concentration in IU/ml of FVILAg in rabbit plasmas at various sampling times

b. Elimination kinetics

i. Comparison of elimination profiles

The elimination profiles of factor VII from the milk of transgenic goats and from NovoSeven^® are presented in figure 6.

ii. Comparison of pharmacokinetic parameters

The pharmacokinetic parameters of factor VII from the milk of transgenic goats and NovoSeven^® were calculated using the WinNonlin 5.0 software and are summarized in table 1 1. The results obtained were compared by the one-way ANOVA method using the Statgraphics plus 5.0 software. (For a 95% confidence interval, a p value lower than 0.05 indicates than the two populations are statistically different).

Table 11: Pharmacokinetic parameters of factor VII from the milk of transgenic goats and NovoSeven^® after single intravenous administration in rabbit (200 μ,ίξ/kg, mean±SD, n=6))

N/A: not applicable

All of the pharmacokinetic parameters are significantly different between factor VII from the milk of transgenic goats and NovoSeven^®. 5. Discussion

After administration of a single dose of factor VII from the milk of transgenic goats and NovoSeven^® by intravenous route in the rabbit, the results are as follows:

• The graph of the concentrations of factor VII in the plasmas of the animals as a function of sampling time shows that the decay follows a bi-exponential model. This indicates that factor VII is distributed in and eliminated from rabbit plasma according to a model with two compartments.

• The pharmacokinetic parameters Ti_/2 and MRT of factor VII from the milk of transgenic goats are significantly different from those of NovoSeven^®.

• The C_max of factor VII from the milk of transgenic goats is significantly lower than the C_max of NovoSeven^®.

• The "recovery" of factor VII from the milk of transgenic goats is significantly lower than that of NovoSeven^®.

• A highly significant difference (p<0.05) is observed in the pharmacokinetic parameters AUC (F7TG<Novo), CI (F7TG>Novo) and Vd (F7TG>Novo). This suggests a faster elimination from the circulating blood of factor VII from the milk of transgenic goats. According to example 1 , factor VII from the milk of transgenic goats has 100% GaNylated species whereas NovoSeven^® has only 26% GaNylated species.

Thus, the factor VII with a higher percentage of GaNylated species has a shorter half-life.

Claims

1. A recombinant human factor VII composition, wherein at least 50%, preferably at least 70%, and in particular at least 90% of the molecules of the N-glycan factor VII composition are forms of the complex type comprising at least one GahVAc(pi,4)GlcNAc antenna substituted or unsubstituted by a sialic acid, a fucose or a sulfate.

2. The composition according to claim 1, wherein more than 50%, and preferably more than 70% of the N-glycan species have at least one sialylated

GalNAc(pi,4)GlcNAc antenna.

3. The composition according to claim 1 or claim 2, characterized in that at least some of the factor VII sialic acids involve a2,6 bonds.

4. The composition according to claim 3, characterized in that all of the factor VII sialic acids involve a2,6 bonds.

5. The composition according to claim 3, characterized in that it further comprises sialic acids with a2,3 bonds.

6. The composition according to any of the preceding claims, characterized in that the factor VII of the composition is produced in the milk of a transgenic goat.

7. A recombinant human FVII composition wherein less than 15%, preferably less than 10%, and in particular less than 5% of the molecules of the N- glycan factor VII composition are forms of the complex type comprising at least one GahVAc(pi,4)GlcNAc antenna substituted or unsubstituted by a sialic acid, a fucose or a sulfate.

8. The composition according to claim 7, characterized in that the FVII of the composition is produced in the milk of a transgenic rabbit.