WO2011107299A2 - Method for the manufacture of recombinant dspa alpha1 - Google Patents

Abstract

A method for the manufacture of rDSPA alpha1 on a commercial scale is disclosed which comprises various downstream process steps inter alia a cation exchange chromatography for capturing rDSPA alpha1 from the conditioned culture medium. The step of washing the cation exchange resin comprises the change of flow direction. The invention further pertains to rDSPA alpha1 obtainable by this method.

Description

"Method for the manufacture of recombinant DSPA alphal "

FIELD OF THE INVENTION

The present invention relates to a method for the manufacture of recombinant Desmodus rotundus salivary plasminogen activator alphal (rDSPA alphal ) obtainable by this method as well as to compositions containing rDSPA alphal , which is suitable for the pharmaceutical or clinical use.

BACKGROUND OF THE INVENTION

Acute ischemic stroke (AIS) is a leading cause of mortality and the major medical cause of disability in developed countries. Thrombolytic therapy with intravenous (IV) alteplase is the only approved treatment of AIS, but its use is currently restricted to within three hours window after symptom onset. There are indications from published meta-analyses that the IV alteplase treatment-window may extend to 4 ½ hours.

Desmoteplase (DSPA alphal ), a highly fibrin-specific plasminogen activator, is in clinical development for the thrombolytic therapy of AIS after the three hour time window. High fibrin-

specificity, lack of activation of beta-amyloid and the absence of neurotoxicity make DSPA alphal a promising thrombolytic compound. DSPA alphal is known from the European patent EP 0 383 417 B1. Its use as a non-neurotoxic plasminogen activator for the treatment of AIS beyond the three hour time window is known from the European patent EP 1 308 166

B1.

DSPA alphal has been produced in mammalian cell culture by recombinant biotechnology (Kraetzschmer et al., Gene (1991), 105: 229-237) and small scale purification of recombinantly produced DSPA (rDSPA) has been described (Witt et al., Blood (1992), 79: 1213-1217). The isolation and purification of rDSPA alphal on a commercial scale is subject matter of EP 1 731 186 B1. However the method described therein has drawbacks, in particular in view of the yield of the so obtained DSPA alphal and also in terms of the efficacy of virus inactivation. The instant application is concerned with the improvement of the isolation and purification of recombinant DSPA alphal (rDSPA alphal) on a commercial scale.

SUMMARY OF THE INVENTION The present invention provides an improved method for the manufacture of rDSPA alphal on a commercial scale. The method results in a DSPA alphal product as a drug substance, which is suitable for the formulation of a final drug product. The method of the invention comprises the following steps:

(a) applying a conditioned culture medium containing rDSPA alphal in a first flow direction to a cation exchanger under loading conditions which result in a binding of rDSPA alphal ;

(b) washing the cation exchanger, wherein said washing step comprises at least a part having a second flow direction, which is contrary to the first flow direction;

(c) eluting the bound rDSPA alphal from the cation exchanger;

(d) applying the rDSPA alphal containing eluent from step (c) to a hydrophobic interaction chromatography matrix under loading conditions which result in binding of rDSPA alphal ;

(e) optionally, washing the hydrophobic interaction matrix;

(f) eluting the bound rDSPA alphal ; (g) applying the rDSPA alphal containing eluent from step (f) to an affinity chromatography matrix under loading conditions which result in a binding of rDSPA alphal ; (h) optionally, washing the affinity chromatography matrix;

(i) eluting the bound rDSPA alphal from the affinity chromatography matrix.

Furthermore the present invention relates to substantially pure rDSPA alphal obtainable by this method and to composition comprising such rDSPA alphal .

DETAILED DESCRIPTION OF THE INVENTION

As used in the specification and appended claims, unless specified to the contrary, the following terms have the meaning indicated:

The term "manufacture" as used in the present application refers to a process step or a sequence of process steps which result in the provision of the desired product; in the present application this is the provision of a product which comprises a substantially pure target protein, namely rDSPA alphal . The manufacture of a recombinant protein typically comprises upstream and downstream processes. However, the term as applied in the present application is used also for a sequence of downstream processes only. Upstream processes are process steps of providing and growing of a cell line expressing the target protein until the capture of the target protein, e.g. the cultivation of the cell line and the harvest of the conditioned culture medium comprising the target protein. Downstream processes are process steps which relates to the recovery, purification and isolation of the target protein from the cultivation medium.

The term "drug substance" refers to the active pharmaceutical ingredient which is subsequently formulated, usually with at least one excipient, to the final drug product.

The term "substantially pure" as applied to the purity of the rDSPA alphal in a composition or solution means at least 80% of the total proteins in the composition is rDSPA alphal , preferably at least 90%, most preferable at least 98%. Protein content and purity can be measured by reverse HPLC and SDS-page gel analysis.

The term "rDSPA alphal", "rDSPAal" or "DSPA", if not otherwise outlined, as used in the context of the present invention are all synonyms, and synonyms to "desmoteplase". rDSPA alphal is further defined as CAS #: 145137-38-8. The term "DSPA" is used as a short form of rDSPA alphal , if not otherwise outlined. These terms all designate the plasminogen activating factor with the primary amino acid sequence as depicted on Table 2 or a sequence of at least 95% identity, preferably at lest 98% identity thereto, including any microheterogeneous forms thereof, such as proteins with essentially the same primary structure and biological activity but including minor differences as to the N- or C-terminal sequence and glycosylation. These microheterogeneous forms are designated as protein subspecies. Accordingly, unless otherwise explicitly stated, the terms rDSPA alphal and its synonyms as above refer to a mixture of proteins of the primary structure of Table 2, proteins of at least 95 or 98% identity thereto, and microheterogeneous forms thereof. In a particular embodiment this mixture contains only DSPA alphal with the primary structure of Table 2 and microheterogeneous forms thereof. The microheterogeneity can be demonstrated inter alia by cation exchange chromatography, which preferably is accomplished as a high performance liquid chromatography (HPLC), which is known to the skilled person. Cation exchange chromatography performed as HPLC is known as "CIEX-HPLC".

The term "non-rDSPA alphal protein and non-protein contaminants" refers to all material other than rDPSA alphal found within the biological media from which rDSPA alphal is being purified, for example from the cell culture or fermentation process.

The term "drug product" refers to the dosage form in the final primary packaging intended for marketing, i.e. the drug product is the final market product.

The term "drug substance" refers to the active pharmaceutical ingredient (API) of a drug product.

The terms "cation exchanger" or "cation exchange matrix" are used as synonyms and refer to a natural or artificial substance, usually a solid, which is able to exchange bound ions with ions from the surrounding liquid medium. Frequently this material is a resin material. A cation exchanger has negative functional fixed ions and exchanges positive counter-ions. The anchor groups (exchange-active components) in commercially available cation exchange matrixes are usually -C₆H₅0^", -S0₃ ^"-, -COO^", -P0₃ ^"-, or -As0₃ ^". Weaker cation exchange materials/resins are those in which the binding strength of the cation is not high, such as those with carboxyl or carboxyalkyi functionalities. Furthermore, weaker cation exchange resins are usually not fully dissociated at acidic pH. A particular weak cation exchange matrix used in the invention is comprised of a matrix of silica particles covalently bound to polyethyleneimine silane, wherein the amino groups of the polyethyleneimine have been derivatized with carboxyl groups. Such a material/resin is commercially available from J. T. Baker, under the trade name Widepore CBX® chromatography resin. A further possible weak cation exchange matrix comprises a matrix of cross-linked polymethylmetacrylate (PMMA) polymer particles covalently bound to polyethyleneimine (PEI). The PEI can be modified to provide a carboxylic acid group, preferably -CH₂-CH₂-COOH groups. Such a cation exchange matrix is commercially available as Bakerbond XWP500 PolyABx 35. The terms "anion exchanger" or "anion exchange matrix" are synonyms and both refer to a natural or artificial substance which is able to bind anions which can then be exchanged with anions from the surrounding liquid medium. An anion exchanger has positive functional fixed ions and exchanges negative counter-ions. Strong anion exchange matrices can have quaternary ammonium groups from type I:

0 ©

— N(CH₃)₃ X or from type II:

© G

-N(CH₃)₂ X

C₂H₅OH wherein X is an anion selected from the group consisting of hydroxyl, chloride, sulfate, bromide, iodide, fluoride, sulfide, hydrogensulfate, hydrogensulfide, phosphate, diphosphate, monophosphate, carbonate, hydrogencarbonate, citrate, tartrate, or phthalate. A commercially available anion exchange matrix containing diethylaminoethyl (DEAE) groups is Sartorius Sartobind® Q membrane absorber.

The term "membrane adsorber" refers to a device containing membranes, which allow adsorption of ligands such as ions, inorganic compounds, organic compounds like nucleic acids, carbohydrates, lipids, proteins or even viruses or cells. These devices can contain membrane ion-exchangers, ligand membranes and activated membranes. A membrane adsorber device is typically formed of a housing having an inlet and an outlet and one or more layers of an adsorptive membrane located between the inlet and outlet such that all liquid entering the inlet must flow through the one or more membrane layers before leaving the device. Therewith one or more desired constituents of the liquid, such as viruses and/or Host Cell Proteins (HCP) are bound to the membrane surface and are removed from the liquid.

The membrane may be a microporous or macroporous membrane formed of a polymer selected from olefins such as polyethylene, including ultrahigh molecular weight polyethylene, polypropylene, EVA copolymers and alpha olefins, metallocene olefinic polymers, PFA, MFA, PTFE, polycarbonate, vinyl copolymers such as PVC, polyamides such as nylon, polyesters, cellulose, cellulose acetate, regenerated cellulose, cellulose composites, polysulphones, polyethersulphones, polyarylsulphones, polyphenylsulphones, polyacrylonitrile, polyvinylidene fluoride (PVDF), and blends thereof. Additionally, nonwoven and woven fabrics of the same materials, such as Tyvek ®; paper available from E. I. DuPont de Nemours and Company of Wilmington, Delaware, and fibrous media such as a cellulosic pad (e.g. ILLISTAK™ filtration media available from Millipore Corp; Bedford, MA, USA) may be used. The membrane selected can depend upon the Peclet number, the desired filtration characteristics, the particle type and size to be filtered or the flow desired.

The term "tangential flow filtration" (TFF) refers to a pressure driven process that uses a membrane to separate components in a liquid, wherein a liquid (the feed flow) is pumped tangentially along the surface of the membrane and a pressure applied serves to force a part of the liquid through the membrane to the filtrate side of the membrane. TFF materials (e.g., hollow fiber, spiral-wound, flat plate) and methods applying it (e.g. in ultrafiltration (UF), diafiltration (DF), microfiltration) are well known to one of ordinary skill in the art.

The term "hydrophobic interaction matrix" or refers to a natural or artificial substance, usually a solid - frequently a resin - which contains uncharged groups, such as methyl, ethyl, or other alkyl groups. It is applied for hydrophobic interaction chromatography (HIC). These groups form hydrophobic bonds with groups on protein moieties which are passed through the resin and result in separation of proteins based on the strength of interaction between the protein and resin groups. A particular hydrophobic interaction matrix/resin is composed of semi-rigid spherical beads synthesized by a copolymerization of ethylene glycol and methacrylate type polymers derivatized with butyl groups. Such a matrix/resin is commercially available from Toso-Haas, under the trade name Toyo-Pearl® 650M C4.

The term "affinity chromatography resin" refers to a natural or artificial substance, usually a solid, which is used for the purification of proteins. The resin separates proteins based on the affinity which occurs between groups on the protein and groups on the resin. In the instant invention, the resin used as an affinity chromatography resin is usually used as a size exclusion resin to separate proteins based on their size. A particular affinity chromatography resin is a cross-linked co-polymer of allyl dextran and Ν,Ν'-methylene bisacrylamide in the form of beads which are capable of fractionating globular proteins between 20,000 and 8,000,000 kDa. Such a resin is commercially available from Pharmacia, under the trade name of Sephacryl® S-400.

The term "flow direction" refers to the direction of the flow the liquid (e.g. the washing or elution buffer) has when passing through the chromatography devices or any other device in the manufacture process. This can either be downwards, namely top down ("downflow mode"), or bottom-up, i.e. upwardly ("upflow mode"). In the downflow and upflow mode the flow has contrary directions.

The term "virus inactivation" refers to the reduction of the infectivity of a virus containing composition by at least 4 Log

This reduction of virus infectivity is considered by the European Note for Guidance CPMP/BWP/268/95 of February 14, 1996 as indicating a sufficient efficacy of a virus inactivation step in the manufacture of a pharmaceutical substance.

The term "nanofiltration" refers to a filtration based on size exclusion of particles to be removed or separated from a liquid where the pore size of the filter is of nanometer size. In general, the pore size of the nanofiltering unit is less than about 30 nm. However, any membrane having the filter cut-off rating sufficient to reduce or eliminate viruses from a protein-containing sample can be employed for virus removal by nanofiltration. Such a nanofilter is commercially available as PLANOVA 15N from Asahi-Kasei Corp or Viresolve NFP Opticap from Millipore Corp.

The term "ultrafiltration" refers to a process wherein a liquid is placed in contact, typically under some pressure, with a semi permeable membrane containing pores of a specified size, such that molecules or complexes, which are small enough to pass through the pores, permeate the membrane to the opposite side, whereas molecules or complexes which are too large to pass through the pores are retained on the upstream side of the membrane. Ultrafiltration membranes are typically formed from polymers and are specified to have a particular cut-off molecular weight. The term "diafiltration" refers to a variant form of ultrafiltration, which combines the characteristics of dialysis with ultrafiltration. Addition of water or an aqueous buffer to the retentate of the filtration results in a buffer exchange or dilution of the original buffer solution. In this method the diluent, essentially water, is introduced into the buffer tank or between two successive ultrafiltration modules.

The term "depth filtration" refers to a filtration method with a filter consisting of a three dimensional matrix, typically with a thickness of at least 3 mm. In the filtration process the solid ("dirt" or "debris") particles are retained by a combination of adsorption and straining. The term "conditioned culture medium" refers to a culture medium in which cells have been grown. The medium has, therefore, been conditioned by the growth of the cells and contains products excreted or released into the medium during cell growth. These can be both waste products produced during growth or proteins which have been made and secreted into the medium as well as remainders of dead cells.

The term "optional" or "optionally" means that the subsequently described event of circumstances may or may not occur, and that the description includes instances where said event or circumstance occurs and instances in which it does not. DESCRIPTION OF THE PREFERRED EMBODIMENTS

The present invention is directed to a method for the manufacture of rDSPA alphal on a commercial scale and in a form suitable for the use in pharmaceutical formulations, in particular for final drug products. The rDSPA alphal resulting from or obtainable by this invention preferably is substantially pure. The rDSPA alphal is produced by fermenting a mammalian cell line capable of secreting the rDSPA alphal product into the culture media. The media that is obtained from the bioreactor containing the mammalian cells, is harvested, optionally depth filtered, and then the rDSPA alphal is separated from other proteins and contaminants by a series of chromatographic steps, beginning with the use of a cation exchanger, followed by a washing step, which at least partly has a flow direction contrary to the flow direction of the loading step, followed by a selective elution of the rDSPA alphal from the exchange matrix. The rDSPA alphal fraction obtained by selective elution from the cationic exchange matrix is then applied to a hydrophobic interaction resin, where binding and a subsequent selective elution from the matrix provides a second level of purification. The eluted rDSPA alphal fraction is then applied to an affinity chromatography matrix, where again binding and selective elution provides a further level of purification. The purified rDSPA alphal can then be concentrated by conventional techniques, such as ultrafiltration. Preferably the product of the process of the invention constitutes a drug substance (active pharmaceutical ingredient) which can further be processed to a drug product.

In further embodiments of the invention additional downstream steps of purification, such as anion exchange chromatography or nanofiltration, and/or steps of virus inactivation can be implemented between the above mentioned chromatography steps.

The invention can be practiced as exemplified below, without being limited to these examples.

1. Culture Media and Cell Lines

The culture medium comprises a base medium suitable for mammalian cell growth, such as DMEM or Ham's F12. A particular medium is William's E medium (Williams, G. M. and Gunn, J. M, Exp. Cell Res., (1974) 89:139). For the inoculation growth phase, the base medium will usually be supplemented with a serum source, typically bovine serum (BS) or newborn calf serum (CS), present at a concentration in the range from about 0.1 % to 10% by weight, usually being present at about 1% to 5% by weight. Other growth factors or buffers, such as HEPES, may also be added. During the perfusion growth phase, the serum concentration is usually maintained at the same concentration, typically being in the range from about 3% to 8%, usually being about 5%.

Alternatively a serum-free media, in particular CHO-V medium from Irvine Scientific supplemented with glutamine, can be used. Cell lines suitable for use in the present invention include mammalian cell lines capable of non-adherent growth in suspension culture and/or adherent growth on microcarrier beads. Particular cell lines which meet these requirements include Chinese hamster ovary (CHO) cell lines, BHK cells, or the HEK293 cell line (Kraetzschmer et al., Gene, (1991 ) 116: 281 - 284; Petri, T., J. BioTechnology, (1995) 39: 75-83).

A particularly preferred CHO cell line is DXB1 1 , which is described in Urlaub, G. and Chasin, L. A., Proc. Natl. Acad. Sci. USA, (1980) 77: 4216-4220. These cells have been co- transfected with the expression vectors pSVPA11 and pUDHFRI , which contain the coding sequences for DSPA alphal and mouse dihydrofolate reductase, respectively (Petri, T., ibid). The transformed CHO cell line used in the present invention is internally designated CD16.

2. Cation Exchange Chromatography

The conditioned culture medium is applied to a cation exchange matrix (e.g. a resin; usually packed in the form of a column) under conditions selected to provide essentially complete binding of rDSPA alphal to the matrix. Before, the pH can be adjusted to a range of 4 to 6, preferably to 5.5 ± 0.3, with e.g. 0.33% acetic acid by online dilution or direct addition and NaCI (e.g. 110 mM). While other proteins will also be bound, the initial binding stage provides a first level of separation as a number of the undesired or contaminating proteins and other compounds, such as phenol red, in the conditioned media will be unable to bind to the matrix and thus will flow through the matrix. Accordingly this first step can also be referred to as a "capture step". After this loading step, the cation exchange matrix is washed. Compared to the flow direction of the loading step the flow direction of the washing step is reversed, i.e. the flow direction of the washing step is contrary to the flow direction of the loading step. This means that if the loading is performed downwardly the washing is upwards and vice versa. If the washing comprises more than one washing steps, at least one of the steps has a contrary flow direction, i.e. the washing is at least partly accomplished in a contrary flow direction of the loading step. Preferably the washing step which directly precedes the elution of the matrix (i.e. the final wash step) has a flow direction contrary to the flow direction of the loading step. Furthermore it is preferred that the final wash step and the elution step have the same flow direction.

It has been found by the inventors that the change of the flow direction in the capture step impacts on the microhetereogeneity of rDSPA alphal . Accordingly the change of the flow direction in the capture step, which may precede a series of further steps for the purification of a target protein from a conditioned media, can serve as a measure for the determination and/or control of the microheterogeneity of said target protein. This in particular applies if the target protein is rDSPA alphal and if the capture step is a cation exchange chromatography. Suitable cation exchange matrices include a wide variety of resins derivatized with cationic functionalities which are able to bind rDSPA alphal . Preferred are synthetic resins, such as those comprised of silica gel particles, cross-linked agarose, or cross-linked polymethacrylate polymers, derivatized with cationic functionalities such as carboxyl, carboxymethyl, sulfonyl, phosphoryl, and the like. Particularly useful are relatively weak resins, such as those having carboxyl or carboxyalkyl functionalities, such as carboxymethyl or carboxyethyl. A possible resin is comprised of a matrix of silica particles covalently bound to polyethyleneimine silane, with amino groups of the polyethyleneimine silane derivatized with carboxyl groups. Such a resin is Baker Widepore CBX® (45 pm bead size), which is commercially available from J. T. Baker. A particularly preferred weak cation exchange matrix comprises a matrix of cross-linked polymethylmetacrylate (PMMA) polymer particles covalently bound to polyethyleneimine (PEI). The PEI is modified to provide a carboxylic acid group, preferably -CH₂-CH₂-COOH groups. Such a preferred cation exchange matrix is commercially available as Bakerbond XWP500 PolyABx 35 (hereinafter also referred to as "PolyABx resin").

The binding, washing and elution conditions will vary depending on the binding strength of the cationic resin. For weak cationic resins, such as the PolyABx resin, binding may be effected at low ionic strength under slightly acidic conditions, typically pH 4-7, preferably about pH 5.5. The washing preferably is accomplished with 30 to 80 mM, preferably 50 mM acetic or phosphate buffer, about pH 5.

The rDSPA alphal is the eluted from the matrix, where the elution may be accomplished by either a stepwise elution or linear gradient elution by increasing the ionic strength of the buffer. In either case, the rDSPA alphal is collected for further purification as described below. Preferably the elution is accomplished by salt step gradient of 100 to 500 mM NaCI, preferably about 200 mM, in 30 to 80 mM Na₂HP0₄, preferably about 50 mM, about pH 7.5. 3. Hydrophobic Interaction Chromatography (HIC)

The rDSPA alphal fraction collected from the cation exchange matrix is applied to a hydrophobic interaction matrix (in particular a resin; usually in the form of a column) under conditions which allow binding of the rDSPA alphal to the matrix, typically high ionic strength and acidic pH. The rDSPA alphal is then selectively eluted by increasing the organic solvent concentration of a mobile phase applied to the column, using a linear or a step-wise gradient. The rDSPA alphal fraction is collected for further purification. This step reduces DNA contamination by approximately 100 to 1000 fold and inactivates potential vital contaminants.

Suitable hydrophobic interaction matrices include a wide variety of uncharged resins having covalently attached hydrophobic groups, such as propyl, butyl, octyl, phenyl, and the like. The resins may be cross-linked organic polymers, such as styrene-divinylbenzene, silica, agarose, polymethacrylate, or any one of a wide variety of other suitable particulate supports. A particularly preferred resin is comprised of semi-rigid spherical beads synthesized by a copolymerization of ethylene glycol and methacrylate type polymers derivatized with butyl groups. Such a resin is Toyo-Pearl® 650 (40-90 um beads) which is commercially available from Toso-Haas.

Binding to the hydrophobic interaction column is effected under conditions of high ionic strength, usually at an acidic pH from 3-6, preferably pH about 5.5. Substantially all the protein contained in the rDSPA alphal fraction which had been eluted from the cation exchange resin is bound to the hydrophobic interaction column. The various proteins may be selectively eluted based on the differing strengths of hydrophobic interaction with the hydrophobic groups on the matrix, i.e., in order of increasing hydrophobicity of the protein. Elution may be performed with a step-wise or linear gradient, usually with an alcohol eluant, such as ethanol or isopropanol. A particularly preferred alcohol is ethyl alcohol.

Preferably, with the Toyo-Pearl C4 matrix, equilibration may be performed with an equilibration buffer having about 30 to 80 mM (preferably about 50 mM) sodium acetate, 1.0 to 1.5 M (preferably about 1.25 ) sodium chloride, about pH 3 to 6 (preferably of about pH 5.5). After the rDSPA alphal fraction from the ion exchange column is adjusted to pH of 3 to 6 (preferably pH about 5.5), it is applied to the C4 matrix, and the matrix is then re- equilibrated with the equilibration buffer described above.

The column is then washed preferably with a series of different buffers, starting with a washing with the equilibrium as described above, followed by a washing with 10 to 50 mM (preferably about 20 mM) HCI, pH 1-4 (preferably pH 2.5) containing increasing concentrations of an organic solvent. The organic solvent can be ethanol. The concentration of the organic solvent (preferably ethanol) can increase from 10 to 30 % (preferably from 15 to 17 %). The gradient preferably contains about 20 mM HCI, about pH 2.5. The elution of the column can be accomplished with an elution buffer of 10 to 50 mM, preferably about 20 mM, HCI, and 10 to 40 %, preferably about 30 %, in particular 29.5 % ethanol.

4. Affinity chromatography

The rDSPA alphal fraction collected from the hydrophobic interaction resin is diluted with 10 to 50 mM, preferably about 20 mM HCI buffer (two parts eluate and one part HCI) and then applied to an affinity matrix (usually in the form of a column) under conditions which allow binding of the rDSPA alphal to the affinity matrix. Binding of rDSPA alphal to the affinity resin is achieved under conditions of low ionic strength and low pH. Substantially all the rDSPA alphal collected from the hydrophobic interaction resin is bound to the column. While the rDSPA alphal remains bound to the column, contaminants are eluted by washing the column with 10 to 50 mM, preferably about 20 mM, HCI. The rDSPA alphal may then be selectively eluted by raising the pH and/or ionic strength, namely stepwise or linear gradient, preferably with a buffer containing 100 to 500 mM, preferably about 200 mM glycine, 0.1 to 0.5 M, preferably about 0.3 M, NaCI, pH 5 to 7, preferably about pH 6.0.

Accordingly one further purpose of this step can be the buffer exchange. This in particular applies if the preceding step has a low pH and a high concentration of organic solvents.

The matrix used as affinity resins are often ones which are normally used as size exclusion resins. Suitable affinity matrices include resins comprised of a cross-linked polymer of allyl dextran and Ν,Ν'-methylene bisacrylamide in the form of beads with a diameter between 25 and 75 μιη. A particularly preferred resin is Sephacryl 400®, which is commercially available from Pharmacia and is capable of fractionating globular proteins between 20,000 and 8,000,000 kDa.

5. Virus inactivation

This step can optionally be integrated into the downstream processing in order to lower the risk of virus contamination of the final pharmaceutical composition. It can best be applied between the steps of hydrophobic interaction chromatography and affinity chromatography. Its purpose then is to inactivate any virus particle which may be present in the HIC eluent.

In this step the eluent adjusted to a low pH, namely a pH of below 4, preferably below about 2 in combination with a high concentration of an organic solvent, such as ethanol. The organic solvent can have a concentration of 15% or more, in particular 25% or more, more preferred even 29% or more. The incubation of the so adjusted eluent can be at least 3 hours, in particular at least 10 or at least 18 hours. The incubation is preferably accomplished at around room temperature (18 to 24°C).

6. Anion exchange chromatography

For further purification the eluent of the PolyAbx resins can be applied to an anion exchange chromatography even before being the subject matter of a hydrophobic interaction chromatography. This is advantageous not only in view of purification but also in view of further virus inactivation. The anion exchange matrix applicable in the present invention preferably is a strong anion exchange matrix can with quaternary ammonium groups from type I:

0 ©

— N(CH₃)₃ X or from type II: w _{)n se}|_ected from the group consisting of hydroxyl, chloride, sulfate, ⁰¹ loride, sulfide, hydrogensulfate, hydrogensulfide, phosphate, di- pi

sphate, carbonate, hydrogencarbonate, citrate, tartrate, or phthalate. A preferred exchange matrix contains diethylaminoethyl (DEAE) groups (e.g. Sartorius Sartobind® Q membrane absorber).

The anion exchange chromatography can be performed by applying the anion exchange matrix as a membrane absorber with a pore size of 0.1 to 20 μηι, preferably between 3 and 5 pm. The chromatography can be designed as a tangential flow filtration.

Before applying the eluent of the PolyAbx resin to the anion exchange chromatography the eluent can be diluted with WFI (water for injection) to a ration of 1 :1 to 1 :2, preferably to 1 to 1.25. The membrane absorber (in particular Sartorius Sartobind® Q) can be equilibrated to a conductivity of a maximum of about 30 mS/cm, preferably to a conductivity between 3 and 15 mS/cm, most preferred between 8 and approximately 13.5 mS/cm. Fore the equilibration an equilibration buffer containing 10 to 50 mM, preferably about 20 mM, Na phosphate buffer and 50 ro 120 mM, preferably about 80 mM, NaCI, pH 6 to 8, preferably about 7.3, can be employed.

7. Nanofiltration

For the further reduction of the risk of virus contamination in the drug substance or drug product the rDSPA alpha 1 containing composition can be applied to a nanofiltration. Preferably the eluent of step (i), i.e. the eluent from the affinity chromatography is used for this further step before possibly being concentrated. The commercially available nanofilter PLANOVA 15N from Asahi-Kasei Corp or Viresolve NFP Opticap from Millipore can be used.

Before filtration the nanofilter is preferably equilibrated with a buffer with glycine (preferably with 100 to 500 mM glycine, more preferred about 200 mM) and NaCI (e.g. 0.1 to 0.5, preferably about 0.3 M), pH 5 to 7, and preferably about pH 6.0. Before filtration two or three column-eluates from the affinity chromatography can be pooled and homogenized. The eluate or eluate pool is then preferable pre-filtered (e.g. a Millipak 0.1 pm) followed by the filtration over the nanofilter.

8. Concentration and further processing to the drug substance.

After the final purification or decontamination step the composition comprising the rDSPA alphal may be concentrated, typically by filtration. In the present invention an ultrafiltration/diafiltration can be used to concentrate the filtrate from the nanofiltration. This step is at the same time applied for a buffer exchange to the desired final formulation buffer. This combined ultrafiltration/diafiltration and final concentration step can be applied under laminar air flow and room temperature with a TFF system Sartoflow Alpha, which is before equilibrated with the formulation buffer. The composition can be concentrated to a concentration of 3 to 8, preferably about 5.0 mg/ml of rDSPA alphal . This composition can then be stored for the further processing. For the preparation of the drug product the composition can be adjusted to the desired final substance concentration by adding further formulation buffer and can then be aliquoted for the fill and finish steps.

In a specific embodiment of the invention the process of manufacture comprises one or more capture steps with a cation exchanger employing a PolyAbx resin, one or more steps of a membrane absorber, one or more steps of hydrophobic interaction chromatography, one or more virus inactivation steps, one or more steps of anion exchange chromatography, one or more steps of virus filtration and one or more steps of ultra-/ diafiltration and optionally a concentration step. In a particularly preferred embodiment the process of the invention comprises a sequence of steps as outlined in Fig. 1.

9. rDSPA alphal microheterogeneity

The microheterogeneity of rDSPA alphal can be assessed and demonstrated by various analytical methods, such as N-terminal sequencing or gel electrophoresis (e.g. SDS page). A further suitable method of demonstrating microheterogeneity of a protein is the performance of a cation exchange chromatography (CIEX), in particular if it is performed as a high performance liquid chromatography (HPLC). The CIEX-HPLC outcome, i.e the chromatogram of a specific HPLC-CIEX analysis, represents a suitable measure for the description of a protein sample in terms of its microheterogeneity. Accordingly the specific CIEX-HPLC chromatogram (synonym: "CIEX-HPLC profile") of a protein can be used for the identification of a protein sample in the sense of a "finger print". It can also be applied as an in-process control for the manufacture of a drug substance.

According to the present invention a suitable CIEX-HPLC is performed with a strong cation exchanger, preferably with -(CH₂)₃S0₃ ^" as functional groups. The functional groups are preferably surface bound to hydrophilic polymer beads (e.g. hydroxylated methacrylic polymer), which can have a mean particle size of e.g. 10, 13 or 20 μιη, whereas 10 pm is preferred. They can have a mean pore size of about 1000 Angstrom. Suitable commercially available cation exchangers are TSK SP-5PW, Toyopearl SP-650 and TSK-GEL SP-5PW (all available from Toso Haas). The use of TSK SP-5PW is particularly preferred.

According to one embodiment of the invention a TSK SP-5PW column is loaded with rDSPA alphal (or any other rDSPA alphal containing solution), which is dissolved in a mannitol/ glycine buffer. In a further preferred embodiment the buffer contains 2 to 6 % mannitol (preferably about 4%) and 100 to 500 mM glycine (preferably about 200 mM). The column can be pre-equilibrated by successive rinsing using preferably 10 to 50, in particular about 20 mM Na₂HP0₄ buffer, pH 5 to 8, preferred about 7.0. A second buffer with 10 to 50 mM, preferably about 20 mM Na₂HP0₄ and 0.5 to 2, preferably about 1 M, NaCI, pH 5 to 8, preferred about 7.0, can be applied. The loading is then performed using a buffer containing 10 to 50, preferably about 20 mM Na₂HP0_4.. Elution of the captured rDSPA alphal can be done with a linear salt gradient e.g. using a buffer containing 10 to 50 nM, preferably 20 mM, Na₂HP0₄ and 0.5 to 2.0, preferably about 1 M NaCI, pH 5 to 8, preferably about 7.0), resulting in a salt gradient ranging from 0 - 1M NaCI. The elution is controlled by A₂₈₀ measurement. The rDSPA of the invention and using a CIEX-HPLC method as described has a CIEX-HPLC profile with at least two, in particular 3 or more, most preferred exactly six peaks within the retention time (time frame) from 5 min to 25 min, preferred from 7 to 17 min, most preferred from 8 to 16 min. A deviation in the retention time of up to 3% is acceptable. A specifically preferred CIEX-HPLC profile of the rDSPA alphal of the invention is depicted in Fig 10.

Further to the characteristic retention times of the peaks within a CI EX profile, the area under each peak, given as the area percentage related to the area of all peaks within a given time frame, allows the specification of a protein sample, and thus can, in the present invention, serve as a finger print for rDSPA alpha 1 microheterogeneity. The rDSPA of the present invention hence can further be characterized in that the area under at least two peaks within the retention time frame from 5 to 25 min represent at least 40 % of the total area of all peaks within that time frame. In a most preferred embodiment of the invention the rDSPA alphal exhibits a CIEX-HPLC profile wherein one or more of the subsequent conditions are fulfilled, namely peak 1 represents less than 6 % area, peak 2 represents 10 to 22 % area, peak 3 represents 20 to 35 % area, peak 4 represents 27 to 37 % area, peak 5 represents 11 to 21 % area and peak 6 represents less than 9 % area (the peak numbering corresponds to the sequence of appearance of the peaks during elution). For these data a deviation of up to 5% is acceptable. The rDSPA alpha 1 of the invention exhibits the above mentioned CIEX-HPLC profiles in particular in these cases, when the CIEX-HPLC is accomplished with a matrix with -CH₂-CH₂- CH₂-S0₃ ^" as functional groups and a buffer with a step salt gradient of 1 to 3 M NaCI, preferably about 1 M NaCI is used for the elution. It is particularly preferred to perform the CIEX-HPLC for the characterization of the rDSPA alphal of the invention according to example 7. EXAMPLE 1

Production process for rDSPAal

A possible method according to the invention is given in Figure 1 comprising various steps of the isolation and purification. This method is described in more detail below.

1. Overview on the Manufacturing Process

The schematic drawing overleaf (Figure 2) presents an overview of the manufacturing process, including all relevant upstream and downstream steps. Furthermore, a flow-diagram is given in Figures 3A to 3C, which in addition to the schematic drawing explains details of the upstream and downstream manufacturing process, including in-process controls and also critical parameters.

2. Overview on batch pooling strategy

The manufacturing process for desmoteplase uses a flexible batch definition, which is dependent on the amount of DSPA in the PolyABx eluate pool and the maximal loading capacity of the HIC column (max. 0.5 mg DSPAal/ mL resin). The batch pooling strategy is shown in Figure 4.

Overall 15 harvests are collected (one harvest pool consists of two harvest days: E1 , E2, E3...), subjected to a cell clarification step followed by a capture step (PolyABx column). The total fermentation time is 30 days and thus results in 15 capture runs in total. PolyABx- eluates are stored at -20°C until all capture runs are finished and then all eluates are pooled.

The PolyABx-pool is further processed with a membrane adsorption step and is then divided into six aliquots, which are independently subjected to the HIC step. Two HIC eluates are then combined to give the material for one S-400 polishing step resulting in overall three S- 400 eluates. All S-400 eluates are combined and further processed to nano-filtration and ultra/ diafiltration to produce the drug substance. Thus, out of 15 harvests, in total one bulk drug substance is generated.

3. Upstream Processing 3.1 Working Cell Bank

The Working Cell Bank (WCB), is designated "CHO 16.4 desmoteplase", and is stored in the vapour phase of liquid nitrogen. The vials of the WCB contain approximately 4.5 mL of cell suspension. The CHO cells were frozen in CHO-V medium (Irvine Scientific Inc. Santa Ana Inc, CA, USA) with 7 % DMSO. The generation of a seed-train for DSPA manufacturing using the expression strain CHO-CD16.4 is described below. 2011/001884

17

3.2 Pre-culture I

All procedures are carried out under aseptic conditions in a laminar flow cabinet. The cell suspension is thawed in a water bath at 37°C. The thawed suspension is diluted with 45 mL cold CHO-V medium containing 8 mmol/L glutamine to dilute the DMSO in the WCB. After centrifugation, the supernatant from the diluted cell suspension is removed and the cell pellet suspended in 50 mL fresh medium. The cell suspension is diluted with medium to reach a concentration of 500,000 ± 20% viable cells/mL. The cell suspension is transferred into a sterile 500 mL spinner flask which is placed in a C0₂ incubator (T = 37 ± 1 °C and C0₂ cone, of 5% ± 1%) and stirred at 40 rpm for 20 ± 4 h. 3.3 Pre-culture II

A second subcultivation phase is performed to generate sufficient cells for the inoculation of the 10 L fermenter. The calculated cell amount, out of the 500 mL spinner flask, is suspended in sufficient fresh medium (CHO-V medium containing 4 mmol/L glutamine), to reach cell concentration of 300,000 ± 20 % viable cells/mL by using the two day cultivation or 200,000 ± 20 % viable cells/mL by using the three day cultivation. The cell suspension is transferred into 1-8 2000 mL sterile spinner flasks which are then placed in a C0₂ incubator (T = 37 ± 1 °C and C0₂ cone, of 5% ± 1 %) and stirred at 60 rpm. After 5 subcultivations the necessary volume and amount of cells for inoculation of the 20 L fermenter is achieved and the viability should be≥ 70%. 3.4 Pre-Culture III

The 20 L fermenter is assembled and sterilised. Approximately five pre-culture II 2,000 mL spinner flasks are pooled for inoculation of the 20 L fermenter. The cell suspension of the pooled spinner flasks is diluted with CHO-V medium to reach a cell density of 300,000 ± 20% viable cells/mL and is transferred into the 20 L fermenter. The cell culture is further incubated at 37°C, at pH 7.0, p0₂ 50%, 50 mbar overpressure and a stirrer speed of 86 rpm for three to four days.

3.5 Pre-culture IV

The 100 L fermenter is assembled and sterilised. The cell suspension from the pre-culture III 20 L fermentation is diluted with CHO-V medium to the 100 L fermenter to reach a cell density of 300,000 ± 20 % viable cells/mL and used to inoculate the 100 L fermenter. The cell culture is further incubated at 37°C, at pH 7.0, p0₂ 50%, stirrer speed of 48 rpm and an overpressure of 50 mbar for 3-4 days (until cell density of 1.5 x 10⁶ cells/mL).

3.6 500 L Culture

The 500 L fermenter is assembled and sterilized. The cell suspension from the pre-culture IV 100 L fermentation is diluted with CHO-V medium to the 500 L fermenter to reach a cell density of 0.6 - 1.0 x 10⁶ viable cells/mL at a starting working volume of 250 L. The cell culture is further incubated at 37°C, at pH 7.0, p0₂ 50%, stirrer speed of 32 rpm and an overpressure of 50 mbar for overall 30 harvest days. In this step culture volumes of up to 2500 L can be used for fermentation.

3.7 Harvesting

Before harvesting, the 500 L fermenter is filled over two days with CHO-V medium to the working volume of 500 L. The feed pump is set to 125 Uday. When the culture volume in the fermenter reaches approximately 500 L continuous harvesting of the cell suspension is started. The perfusion rate is adjusted to approximately 250-300 LVday (0.5 - 0.6 Wd). The harvest of the first 24 h is discarded. During harvest CHO-V medium containing 5.5 mmol/L glutamine is continuously added. The harvest (medium plus CHO-cells) is pumped into a 750 L harvest bag and stored at 2-8°C. Each harvest bag is connected to the fermenter for two days, resulting in approx. 500 - 600 L harvest per bag. The harvest bag is disconnected and exchanged every 2^nd day.

3.8 Depth Filtration and 0.45 μπι + 0.2 μπι Filtration of Harvest

The purpose of this step is to separate the remaining CHO-cells and cell debris from the medium supernatant that contained the desmoteplase active component. The CHO-cells are retained in the ZETA Plus Maximizer depth filter from CUNO. After depth filtration a 0.45 pm + 0.2 pm filtration with a Sartopore 2 Filter from Sartorius is performed. The harvest suspension is pumped via the filters (at a flowrate of approximately 6-8 Umin) into a collection bag. The pressure over the Zeta Plus Maximizer and the Sartopore 2 filter is kept below 0.5 bar to prevent leakage of cells through the filter. The harvest is stored at 2- 8°C until it is further processed.

4. Downstream Processing (DSP) 4.1 Column A (C10/PolyABx-E) PolyABx Capture Step

PolyABx is a weak cation exchanger (functional group -CH₂CH₂COOH) on polymethacrylate particles. The purpose of this step is to capture the desmoteplase protein from the filtered harvest material. This step retains the desmoteplase protein and removes most of the media components, host cell proteins and DNA.

Before the filtered harvest is loaded on column A, the pH of the solution is adjusted by online dilution from pH 7.0 to pH 5.5 ± 0.3 by adding 0.33% acetic acid + 1 10 mM NaCI. The chromatographic process is performed at room temperature (18 - 24°C). The flow through of the column at the end of the equilibration step should be pH 5.3 - 5.7 and a conductivity of 3.4 - 4.2 mS/cm. The maximum load capacity of the PolyABx resin is 3 mg desmoteplase/mL resin. Elution of the captured desmoteplase is via a step salt gradient (50 mM Na₂HP0 buffer + 200 mM NaCI pH 7.5), which is performed at a flow rate of 200 cm/h ± 10%. The elution pool is collected in a sterile bag. The collected Column A eluate pool is mixed, filtrated (0.2 pm filtration with MilliPak 200 (integrity test performed)) and stored at -20 ± 5°C for further processing.

4.2 PolyABx Eluate Pooling (M20/PolyABx-E-Pool)

To ensure homogeneity of the starting material for further processing, all fifteen PolyABx eluates (stored at -20°C) were pooled. Afterwards, the PolyABx eluate pool was split into aliquots according to the batch definition. The batch definition is determined/defined by the DSPA content in the PolyABx eluate pool and the maximal loading capacity of the HIC column (max. 0.5 mg DSPAal /ml_ resin).

The frozen PolyABx eluates are thawed in two steps to room temperature. First the frozen PolyABx eluates are transferred from -20°C to 5 ± 3°C and are stored for 72 ± 3 h. In the second step the PolyABx eluates are transferred from 5 ± 3°C to room temperature (21 ± 3°C) and are stored for 24 ± 3 h before further processing. The thawed PolyABx eluates are then pooled in a 200 L bag. According to the analytical results and batch definition the PolyABx pool is prepared for the next purification step (Sartobind Q filtration).

4.3 Sartobind Q filtration (120/SQF) DNA, HCP and Virus Reduction

The Sartobind Q filter has positively charged groups within the filter material, acting as an anion exchanger and thus binding negatively charged DNA, host cell proteins and viruses.

The Sartobind Q filtration is performed by using the Amersham BioProcess system. Therefore before use, the BioProcess system is sanitised in place (SIP) with 1 M NaOH for at least 3 h. After incubation time, the BioProcess system is rinsed with 0.01 M NaOH until the conductivity <4 mS/cm is reached.

The autoclaved Sartobind Q Filter is assembled to the BioProcess system, depyrogenated with 1 M NaOH for 60 min and equilibrated with 20 mM Na-Phosphate buffer + 80 mM NaCI pH 7.3 until a conductivity of 8.8 - 13.2 mS/cm and pH 6.6 - 8.0 is reached.

The temperature adjusted PolyABx-eluate pool is diluted with WFI (1 : 2.5) and pumped via a filter (max. load capacity 2 mg DSPAal / cm²; at a linear flow rate of 2.5 cm/ h) into a sterile bag. At the end of the operation, the filter is rinsed with the respective buffer to recover all remaining desmoteplase.

Due to further processing within 24 h, according to the batch definition, the Sartobind Q filtrate is kept at 21 ± 3°C. 4.4 Column B (C20/HIC) Hydrophobic Interaction Purification Step

The functional group butyl (-C₄H₉) can be used in a hydrophobic interaction chromatography when bound to appropriate polymer particles to form chromatography resins. The purpose of this step is to remove host cell DNA and any host cell proteins (HCPs) present. The column is washed with 15 to 25 column volumes (CV) of a 20 mM HCI/ 17% ethanol buffer.

Before loading the solution is diluted proportional 1 :2 with loading buffer (50 m NaAc, 1.25 M NaCI, pH 5.5). The column is sanitised in place (SIP; 3 column volumes (CV) 0.5 M NaOH, total contact time 1 h) and equilibrated (4 CV 50 mM NaAc + 1.25 M NaCI pH 5.5) at 100 cm/h ± 10%. The chromatographic process is performed with the Amersham BioProcess system (6 mm) at room temperature (18-24°C). The maximum load capacity of the butyl resin is 0.5 mg desmoteplase/mL resin. The Sartobind Q filtrate is loaded with a linear velocity of 100 cm/h ± 10% in downflow mode. The column is washed first with 2 CV 50 mM NaAc + 1.25 M NaCI pH 5.5 at 100 cm/h ± 10%, then with 5 CV 20 mM HCI at 100 cm/h ± 10%, then with 10 CV 20 mM HCI/15 % ethanol at 100 cm/h ± 10% and finally with 25 CV 20 mM HCI/17% ethanol at 100 cm/h ± 10%.

Elution of the bound desmoteplase is achieved via decreasing the ionic strength and by increasing the ethanol concentration (5 CV 20 mM HCI / 29.5% ethanol). The eluate is collected in a sterile bag. Directly after sampling the column B eluate-pool is transferred for virus inactivation.

4.5 Virus Inactivation (M30)

The purpose of this step is to inactivate any virus particles that may be present in the HIC eluate pool. The low pH (< 2.0) in combination with a high ethanol concentration (approx. 29.5%) and temperature range at 18 - 24°C has an inactivating effect on virus particles. The inactivation is time-dependent.

The bag containing the HIC Column eluate is stored at 21 ± 3°C for 18 ± 3 h (incubation time) at pH < 2.0 (adjustment with 1 M HCI). The virus inactivated HIC Column eluate pool (intermediate storage step) is stored at 5 ± 3°C for up to 8 days during production.

4.6 Column C (C30/ S-400) Sephacryl S-400 Purification Step

The purpose of this step is to obtain a buffer exchange prior to formulation and it also serves as a polishing step for the removal of residual low molecular weight impurities and salts.

The virus-inactivated HIC Column eluates are combined prior to S-400 loading. The pool is diluted with 20 mM HCI buffer (two parts HIC eluate pool, one part 20 mM HCI). Before loading, the column undergoes SIP (3 CV 0.5 M NaOH, total contact time 1 h), and is equilibrated (3 CV 20 mM HCI / 19% ethanol) at 100 cm/h ± 10% prior to use. The chromatographic process is performed at room temperature (18 - 24°C). The HIC column eluate pool is loaded on the column (the maximum load capacity of the S-400 resin is 2 mg desmoteplase/mL resin) at 60 cm/h ± 10%. The column is then washed with 5 CV 20 m HCI (60 cm/h ± 10%). Elution is with 3 CV 200 mM glycine + 0.3 M NaCI pH 6.0 at 60 cm/h ± 10% and the eluate is collected in a sterile bag. The collection of the eluting product is controlled by A₂₈o measurement. The S-400 column eluate is stored at 21 ± 3°C for≤ 24 h or alternatively at 5 ± 3°C. 4.7 Virus Filtration (130/NF)

The purpose of this step is to remove any virus particles that are present in the product.

The virus filtration process is performed under Laminar Air Flow. A Planova 15 N (Asahi Kasei) virus removing filter (surface area 4 m²) is equilibrated with 200 mM glycine + 0.3 M NaCI pH 6.0. Before filtration two or three S-400 column-eluates are pooled and gently homogenized. The S-400 Column-eluate pool is filtered (constant pressure 0.8 ± 0.1 bar) through a Millipak 0.1 μητι pre-filter followed by the Planova 15 N filter and post wash step (10 ± 0.1 L 200 mM glycine + 0.3 M NaCI pH 6.0) into a sterile bag. The virus filtered S-400 column-eluate pool is stored at 5 ± 3°C for < 24 h. 4.8 Ultrafiltration /Diafiltration and final concentration (I40/UF/DF)

The purpose of this step is to concentrate the virus-filtered S-400 column-eluate to a protein concentration of 5.0 ± 0.5 mg/mL and to exchange the buffer to the formulation buffer (4 % mannitol / 200 mM glycine). The ultrafiltration/diafiltration, final concentration step is performed under Laminar Air Flow at room temperature (21 ± 3°C) with TFF System Sartoflow Alpha, assembled with 6 Sartocon Slice-Cassettes, (Hydrosart; 0.1 m² filtration area, 10 kD cut-off). Before ultrafiltration/ diafiltration the Sartoflow Alpha system (with cassettes) is rinsed with WFI, SIP with 1 M NaOH (for at least 3 h) and equilibrated with 4% mannitol/ 200 mM glycine buffer. The product is concentrated by ultrafiltration to approx. 7 fold under pressure of 1.5 ± 0.5 bar. The diafiltration is done with 8-fold volumes of the ultrafiltration volume against 4% mannitol/ 200 mM glycine buffer at 1.5 ± 0.5 bar. The product is ultrafiltrated to the final concentration is 5.0 ± 0.5 mg/mL. The final concentrated product is stored at 2-8°C for further processing. 884

22

4.9 Preparation of Drug Substance (M40)

The purpose of this step is to adjust the final drug substance concentration of desmoteplase to 3.82 ± 0.38 mg DSPAa1/ml_ by adding formulation buffer (4% mannitol/200 mM glycine) and then to aliquot the drug substance for the Fill and Finish step.

Adjustment of drug substance to the final concentration is performed in a Laminar Air Flow at room temperature (21 ± 3°C). The calculated amount of formulation buffer (calculated by density) is added to the product to achieve a final drug concentration of 3.82 ± 0.38 mg DSPAa1/mL. The drug substance is then sterile filtered (0.2 μιτι MilliPak 40, 200 cm²) and aliquoted into PETG flasks and stored at -70°C until further fill and finish steps.

5. Materials

5.1 Medium

For the serum-free production process the manufacturer use CHO-V medium from Irvine Scientific, which is prepared by mixing WFI with serum-free CHO-V powder, glucose, glutamine and NaHC0₃. As a supplement a 200 mmol/L glutamine stock solution is used. The Irvine Scientific CHO-V medium is a commercially available serum-free medium without human or mammalian-derived components. The medium includes recombinant human insulin.

5.2 The Host Cell Line

5.2.1 Nature and Origin of Cells

The CHO cell line used for production of desmoteplase has been designated as "CD16". The parental CHO dhfr^" (dihydrofolate reductase) cell line originates from the Columbia University, New York and is designated in the scientific literature as DXB1 1 (Urlaub and Chasin, 1980, PNAS, 77: 4216-4220). The CHO dhfr -parental cell line is deficient in the enzyme dihydrofolate reductase. The desmoteplase coding region is inserted adjacent to an SV 40 promoter in the eukaryotic expression plasmid, pSVPA1 1. This plasmid, along with a second plasmid (pUDHFR I) containing the coding region for dhfr, was transfected into the CHO cell line. Clones expressing dhfr were selected by growth in medium lacking nucleosides and screened for production of desmoteplase. The desmoteplase producing clone was then amplified with methotrexate (MTX) and subjected to subsequent subcloning. The original desmoteplase clone, TP8K6, was adapted to grow in 500 nM MTX and then further amplified in 10 μΜ MTX, followed by an independent subcloning process. A subclone of TP8K6, clone 16.4, was chosen for the production of desmoteplase and is now referred to as CD16. The isolation of the desmoteplase coding sequences from salivary cells of Desmodus rotundus and expression of desmoteplase in eukaryotic cells is described by Kraetzschmar et al. (Gene 1991 ; 105: 229-37) and Petri et al. (J Biotechnol. 1995, 39: 75- 83). The cloning and selection of the CD16 clone is outlined in Figure 5.

5.2.2 cDNA Library Construction

Salivary glands from the Mexican vampire bat, Desmodus rotundus, were the starting material for cloning of the desmoteplase gene. Total RNA was isolated from salivary glands using a guanidinium isothiocyanate lysis procedure followed by ultracentrifugation through a CsCI layer. Poly (A)^* RNA was obtained through double affinity chromatography using oligo (dT) cellulose spin columns. A first oligo (dT)-primed cDNA library was constructed in the lambda gt10 vector. A second library was made in the Uni-ZAP™ vector after synthesis of the first strand with MMLV reverse transcriptase in the presence of 5-methyl dCTP and after size selection for cDNAs greater than approximately 500 base pairs.

5.2.3 Library Screening and In Vivo Excision of Positive Clones

The Desmodus rotundus salivary cDNA library was screened with a probe derived from human t-PA (tissue plasminogen activator) cDNA. About 50,000 primary clones from the first cDNA library were transferred onto nylon membranes and screened with a nick-translated human t-PA cDNA. Hybridisation was carried out for 14 hours at 42°C in 6 x SSC, 1% SDS. The membranes were washed under low-stringency conditions (4 x SSC, 0.1% SDS, 42°C), and 11 positive clones were detected after two rounds of screening. As no full-length clone was obtained, a nick-translated 76 bp Alul-BamHI fragment (positions "269" to "344") derived from the 5' end of the longest positive clone was used to screen 120,000 independent clones plated from the second unamplified library.

Hybridisation was carried out for a total of 16 hours at 50°C in 6 x SSC, 1 % SDS; in 4 x SSC, 0.1 % SDS; and eventually in 2 x SSC. Out of 200 positive clones which were thus found, 35 were subjected to in vivo excision after superinfection with the R408 helper phage, giving rise to pBluescript(R) SK phagemids containing the cDNA inserts, following the supplier's instructions (Stratagene). Individual clones were partially sequenced and assigned according to one of four distinct categories (alpha 1 , alpha 2, beta, or gamma) dependent on the structural domains predicted to be encoded by each clone.

5.2.4 Nucleotide Sequencing and Gene Structure for Desmoteplase

The nucleotide sequence of the longest insert for desmoteplase has been determined by Sanger^'s dideoxy sequencing methods. An open reading frame (ORF), starting with an ATG codon, was found in a position favourable for translation initiation. The ORF should encode a polypeptide of 477 amino acids. A polyadenylation signal is found in the 3' untranslated region of the inserts which all terminate with a poly (A) stretch. The predicted N-terminal sequence appears to contain a eukaryotic signal peptide sequence which is expected for the extracellular export and secretion of the protein. N-terminal sequencing of desmoteplase collected from bat saliva showed that the mature form of the enzyme begins with the sequence Ala-Tyr-Gly. This Ala residue corresponds to residue 37 of the predicted sequence. Thus, the first 36 amino acids of the primary translation product of the desmoteplase gene are removed during the maturation process, resulting in a 441 amino acid secreted protein.

5.2.5 Construction of Desmoteplase Expression Vectors

CHO cells were chosen for production of recombinant desmoteplase because they have been successfully used to express a variety of other recombinant proteins including t-PA and erythropoietin (EPO). An expression vector was constructed using pSVL from Pharmacia. The desmoteplase gene fragment, including the entire 5' untranslated region preceded by an EcoRI adapter fragment, TCTAGAATTC, the open reading frame, and part of the 3' untranslated region ending the internal EcoRI site was inserted into the Xbal site of pSVL. Plasmid pSVPA11 (Kraetzschmar et al., 1991), see Figure 6A, is a Klenow fill-in, blunt-end fusion of an EcoRI fragment containing the desmoteplase cDNA and the Xbal digested vector pSVL-EcoRI(-), a derivative of the mammalian expression vector pSVL (Pharmacia) in which the single EcoRI site had been removed by filling in and religation. The correct orientation of the inserted cDNAs with respect to the promoter region of pSVL was verified through restriction analysis. pSVPA1 1 has successfully been used for transient desmoteplase expression in COS-7 cell transfection experiments (Kraetzschmar et al., 1991). The complete nucleotide sequence of the vector pSVPA1 1 is given in Figure 6B to 6K.

5.2.6 Construction of DHFR and Co-transfection Plasmid

The selection vector, pUDHFR I contains a 2627 bp Clal fragment cloned into the Accl cleaved and dephosphorylated vector pUC19. The Clal fragment contains the entire SV40 late promoter based mouse dihydrofolate reductase expression cassette isolated from the plasmid pPA207 (see Figure 7A). The complete nucleotide sequence of the vector pSVPA1 1 is given in Figure 7B to 7F.

5.2.7 Construction of rCHO Cells Expressing Desmoteplase For the construction of the CHO cell line the CHO line known as DXB1 1 , deficient in DHFR was used and CHO cells were cultivated in alpha MEM with nucleosides, containing 2.5% foetal bovine serum (FBS) (Gibco BRL) and 0.01 % Serextend. The cells were transfected using the calcium phosphate method with 5 mg Sail linearised pSVPA 1 1 (see Figure 6A) and 0.5 mg EcoRI linearised pUDHFR I (see Figure 7A) and dhfr* positive cells were selected in alpha MEM without nucleosides, supplemented with 2.5% dialysed foetal bovine 884

25

serum (dFBS) and 0.01 % Serextend. Positive cell clones were selected, picked with sterile cotton swabs, and treated individually with increasing doses of MTX from 5 to 500 nM. The desmoteplase protein production of these cell clones before and during treatment with MTX was monitored by fibrin plate assays. One clone, termed TP8K6, responded especially well to this MTX treatment with an increasing desmoteplase production (Petri et al., 1995).

5.2.8 Subcloning of the Desmoteplase Cell Line

The CHO cells were grown in alpha MEM, supplemented with 10% dialysed foetal bovine serum (dFBS). They were amplified in increasing levels of MTX to increase the production of desmoteplase. Cells that had been adapted to 5 and 10 μΜ MTX were subsequently subcloned by limited dilution in 96-well plates to isolate single-cell, high-producing clones. Clones that were positive for desmoteplase production were then trypsinised and subcultured in six-well plates. After the cells had become confluent, a 24-hour measurement of desmoteplase production was performed by S-2288 hydrolysis and cells were counted so that specific production could be calculated. Clones that produced over 20 pg of desmoteplase protein per cell, per day (pg/c/d) were then expanded. The desmoteplase subclone, CD16, was then chosen for its superior growth and high production of desmoteplase. Clone CD16 was expanded, and vials of the so called "CD16.4 cells" were frozen and stored in liquid nitrogen until they were used to prepare a first cell bank.

5.2.9 Generation of MCB (serum free medium)

After adaptation to serum-free conditions and starting from two vials, the cells were expanded in a spinner flask up to a volume of 76 mL in CHO-V (Irvine Scientific) medium with 4 mmol/L glutamine. The cells were transferred to a second spinner flask and expanded to a volume of 118 mL under the same culture conditions. The cells were then distributed into two spinner flasks (59 mL each) and expanded to a culture volume of 198 mL in each spinner. The pooled cell suspension from both spinners was used to inoculate a 2.5 L bioreactor to produce a sufficient cell mass for the MCB. The cells were harvested by batch centrifugation at a cell density of 12.9 x 10⁵ and a cell viability of 95%. The recovered cells were mixed with the freezing medium (culture medium + 7% DMSO) and aliquoted from a single batch into 170 ampoules (1.5 mL). The ampoules were frozen down to -150°C at - 1 °C/min and subsequently stored at -130°C or lower in the gas phase of liquid nitrogen. Each vial contains 1.3 x 10⁷ cells with a viability of 94%. After freezing, an appropriate number of vials were thawed and taken into culture for internal release testing in respect to viability, mycoplasma and microbial contaminants.

5.2.10 Generation of a Working Cell Bank (WCB)

The cells from two vials of master cell bank were expanded in a 125 mL spinner flask up to a cell concentration of 1.0 x 10⁶ cells/mL in CHO-V medium with 4 rtiM glutamine. The cells were transferred to a second spinner flask (500 mL) and expanded to a cell concentration of

1.0 x 10⁶ cells/mL The cells were then distributed into two spinner flasks (500 mL each) and expanded to a cell concentration of 1.0 x 10⁶ cells/mL in each spinner. From these two spinners the cells were distributed into four new spinners and expanded to a total cell concentration of 1.20 x 10⁹ cells/mL (four spinners) and a viability of > 85%. The pooled cell suspension from all four spinners was used to inoculate a 6.0 L bioreactor to produce a sufficient cell mass for the WCB. The cells were harvested at a cell concentration of > 1.3 x 10⁶ cells/mL and < 1.8 x 10⁶ cells/ mL and a cell viability of > 85% by centrifugation using four centrifuge beakers. The cells were resuspended in 640 mL freezing medium (culture medium + 7% DMSO) and aliquoted from a single batch into ampoules. The ampoules were frozen down to -150°C with a temperature controlled freezing machine and subsequently stored at - 130°C or lower in the gas phase of liquid nitrogen.

6. Description of the Drug Substance

A summary of the analytical data for the drug substance as result of the above described production process is given in Tables 1A and 1 B. The specific methods are described below.

6.1 DSPA alphal -related properties

6.1.1 Nomenclature

INN: desmoteplase

Other names: DSPAcrt , desmoteplase (recombinant Desmodus rotundus salivary plasminogen activator alpha 1), ZK 152 387, DSPA

CAS # : 145137-38-8

6.1.2 DSPA alphal structure

6.1.2.1 Primary amino acid sequence Desmoteplase has a molecular mass of approximately 49.5 kD, based on the amino acid sequence as derived from the DNA sequence. The theoretical molecular mass has been verified by peptide mapping, SDS-PAGE and mass spectrometry analysis. The predicted primary amino acid sequence of the secreted active desmoteplase is presented in Table 2. The sequence beginning with Ala-Tyr-Gly-... matches that found for the first 15 residues of desmoteplase isolated from vampire bat salivary glands. The amino acid composition is given in Table 3.

6.1.2.2 Secondary and tertiary structure of DSPA alphal

Desmoteplase appears mainly as a monomer in non-reduced SDS-PAGE analysis and size- exclusion chromatography (SEC). A theoretical prediction and structure-based estimate of the secondary structural features of the desmoteplase molecule has been conducted. The theoretical prediction is based on the full length sequence for salivary plasminogen activator alpha 1 from Desmodus rotundus (Swiss-Prot Entry P98119) and was performed by using PredictProtein (Secondary Prediction Interface). The modelling resulted in 2.27% a-helix, 28.34% β-sheet and 69.9% coil structures.

A similar result was obtained when applying a structure-based estimate of the specific domains of desmoteplase or homologous proteins of known 3D-structure for the calculation; the results are given in Table 4. In this calculation a total length of 441 amino acids (without signal peptide) was assumed and resulted in 5.2 % a-helix, 28.1 % β-sheet and 66.7 % coil as structural features.

Furthermore, modelling of the desmoteplase molecule revealed the structure shown in Figure 8 as ribbon and pearl-model. The domains of the DSPA molecule were modelled using the "Swiss-Model" modelling server (www.expasy.org) based on the sequence P98119 (Swiss-Prot entry).

The structure of the catalytic domain (PDB code 1A5I) was taken without modifications. The template structure of the finger domain given by "Swiss-Model" was based on the structures with the PDB code 1TPM (t-PA finger domain, X-ray structure) and 1TPN (t-PA finger domain, NMR structure). The template for the EGF-like domain was based on the structure with the PDB code 1TPG (t-PA finger and EFG-like domain, NMR structure). A common structure for both domains was delivered by "Swiss-model". The structure of the kringle domain was build by "Swiss-Model, using the templates 1 PK2 (t-PA kringle 2, NMR structure), 1TPK (t-PA kringle 2, X-ray structure) and 1PML (t-PA kringle 2, X-ray structure). The domain structures were connected using modelling program "Deep-View" (www.expasy.org). In a non-automated procedure, missing peptide pieces were inserted manually, the relative orientation of the domains was determined in a way to obtain a maximum overlap, and a number of adaptations of the side chains positions were performed. The final structure was energy minimised and this structure served as input for the GROMOS 96 simulation program. The computer simulation of the DSPA molecule was performed using a rectangular box and explicit water conditions. The simulation to equilibrate the structure and to explore the conformational space was started with an initial temperature of 300 Kelvin (K) and the simulation was continued over 3.6 ns. For graphical representation a snapshot of the structure at 1 ns was taken and a model carbohydrate side chain was attached at the position of amino acids Asn117 and Asn 362.

The general structure of desmoteplase has also been analysed using FT-IR spectroscopy. The FT-IR spectra clearly indicate a high degree of ^β-sheets, since marker bands for β- sheets are visible at 1636 cm^"1 (parallel) and 1678 cm^"1 (antiparallel), turns result in a band at 1660 cm^"1

Figure 9 shows the structural changes of desmoteplase at different pH and temperature values. Desmoteplase appears to be stable over a wide pH range but shows some denaturation at temperatures above 55°C (appearance of random coils signal at 1651 cm^"1). Also, at higher temperatures the band at 1636 cm^"1 is shifted to lower and the band at 1678 cm^"1 is shifted towards higher wave-numbers, indicating the formation of intermolecular C- sheets and aggregation.

In conclusion the molecule is temperature-sensitive above 55°C but very stable over a wide pH range from 2.8 to 10.4.

The analysis of disulphide bridges of desmoteplase is a further important parameter to understand the structure of the protein. The S-S bridges are important for the proper folding of desmoteplase and consequently for its correct function. According to the primary sequence and as also highlighted in the pearl-model in Figure 8, desmoteplase has 28 cysteines which potentially result in 14 S-S bridges. Based on homology searches in public structure databases, proposed S-S bridges are listed in Table 5.

Using mass spectrometry analysis of peptides generated with a non-reduced trypsin digest of desmoteplase it was possible to identify peptides and peptide groups ("S-S clusters"), which reflect the theoretical masses of S-S linked peptide clusters (Table 6). All of the six expected S-S peptide clusters (Peptide Clusters 2-6) were identified and confirmed by MS analysis (not shown).

Some of the peptide clusters contain more than one S-S bridge, which made it impossible to exactly determine the S-S linkage even though the correct peptide cluster was identified. However, additional proteolytic digests besides trypsin were used (see Table 7) and all subfragments are also in alignment with the theoretical S-S bridges. In summary it is concluded that the identification of all expected proteolytic fragments strongly supports the theoretical S-S linkage structure. 6.1.3 Microheterogeneity of desmoteplase

Characterisation work has been performed with regard to glycosylation and N-terminal microheterogeneity of desmoteplase. The desmoteplase preparation resulting from the method as disclosed herein comprises a mixture of different glycospecies and N-terminal subspecies, which is characterised by the profile in the CIEX-HPLC analysis. Each peak in the CIEX-HPLC can be assigned to a certain mixture of desmoteplase subspecies with different glycosylation and N-terminus but possessing the CIEX-elution characteristics. All these DSPAalphal subspecies have been found to be active as shown by S-2288 activity assay. Figure 10 shows a typical CIEX profile from DSPA drug substance. Separation of the individual peaks and MS analysis of the subspecies of desmoteplase in each peak results in a desmoteplase subspecies distribution as listed in Table 8.

The separation of the DSPA subspecies is generally based on a quadruple pattern of glycostructures that is overlapped by the N-terminal heterogeneity. In peak 1 to 4 desmoteplase with the original N-terminus is found and a sequence of glycostructures 16, 11 and 12, 4 and no glycosylation at Asn1 17 can be detected (see Table 9 for description of the main different glycoforms). The first peak of the quartet can be assigned to fully processed desmoteplase glycoforms with doubly sialylated complex biantennary type sugars with and without core fucosylation. Peak two of the quartet contains proteins with mainly singly sialylated complex biantennary and hybrid type sugars. Peak three contains mainly the glycoform with a high mannose type sugar. Finally, peak four can mainly be assigned to non- glycosylated desmoteplase at Asn117. A similar sequence of glycostructures but with a N- terminus of N+3 is then found in peaks 2 to 5. Peak 5 thus contains mainly non-glycosylated protein at Asn117 of incompletely processed desmoteplase (N+3). Further, incompletely processed protein (N+7) can also be found in this fraction. Cleavage of the N-terminal pro- sequence was also incomplete for the main protein form detected in fraction "Pool 6", in which the protein form N+12 was dominant. In addition, isoelectric focussing (IEF) was applied to separate the differently charged subspecies of the desmoteplase molecule. The banding pattern of all desmoteplase subspecies is between the pi markers 9.3 and 7.35 and resembles the distribution of the CIEX peak pattern. A gel section with the results of the IEF analysis is shown in Figure 11. 6.1.4 Site occupancy of the glycosylation sites

The glycosylation sites of desmoteplase have been analysed to determine their occupancy. The results of this work are given in Table 10.

The data show that the Asn362 and the fucosylation sites Thr61 are nearly completely occupied in desmoteplase. For Asn1 17 the occupation is on average about 65%.

6.2 Process-related impurities

Process-related impurities such as host cell proteins (HCP), host cell DNA, Bioburden and endotoxins are analysed for desmoteplase drug substance and drug product. The results are summarised in Table 1B. 7. Analytical methods for characterization of the drug substance

7.1 N-terminal Sequencing

The procedure is used to determine the identity of desmoteplase and is based on automated N-terminal sequence analysis, using Edman degradation chemistry. Prior to Edman degradation samples are concentrated and desalted on RP-HPLC and lyophilised. Prepared samples are then put through fifteen Edman degradation cycles, respectively, which include coupling, cleavage and conversion, the identity is confirmed by on-line RP-HPLC.

There are several identifiable sequences in the DSPA alphal molecule; two major sequences originate from two distinct cellular processing events. Due to different processing events, desmoteplase contains sequences corresponding to "N" (authentic, expected processing) and "N+3" (processed at a site three amino acids upstream from the authentic processing site). Low levels of additional sequences are typically seen which appear to correspond to proteolytic cleavages which have occurred in the desmoteplase molecule following its original biosynthesis. To identify the product as desmoteplase, not less than 95% of N-terminus sequences determined for the product must be derived from DSPAal species.

7.2 Gel Electrophoresis (SDS-PAGE)

The purpose of this protocol is to determine the identity and purity of desmoteplase with respect to other contaminating proteins present in the sample. SDS-PAGE is performed under reducing and non-reducing conditions. Under reduced conditions disulphide bonds are reduced with dithiothreitol in SDS-containing sample buffer and alkylated with iodacetamide. The samples are loaded onto polyacrylamide gels and proteins separated on the basis of molecular weight. The resulting gels are then stained colloidal blue and/or silver stain. Densitometry scanning of the stained SDS-PAGE gels may be performed. The assay standard consists of desmoteplase reference standard SR-444PI2 with a concentration of 1 Mg/pL. Standards are prepared for electrophoresis either with disulphide bonds reduced and alkylated or non-reduced. The identity of desmoteplase is determined by comparing the banding pattern and relative position of bands obtained for the sample to be analysed with the banding pattern and relative position of bands of the reference standard. The identity is confirmed if the banding pattern complies. The purity of desmoteplase is determined from the densitometry scan as the area of the desmoteplase peak divided by the total area of all peaks. Additionally, a 10% limit calibrator and the LOQ concentration are run in parallel to quantify impurities between those limits. The silver stained gel is observed visually to determine if any impurity proteins are observed that are not detected in colloidal blue-stained gels.

7.3 Peptide Mapping

The peptide map profile is obtained by Lys-C digestion of the protein, reduction, denaturation and subsequent RP-HPLC separation of the fragments. The characteristic Lys-C digest chromatograms at 214nm are qualitatively compared to those of reference standard run in parallel and allow verifying the identity of rDSPAal . Altogether, 19 Lys-C cleavage products can be identified in the chromatogram referring to 24 predicted peptides for desmoteplase.

7.4 CIEX-HPLC

Cation-exchange High Performance Liquid Chromatography (CIEX-HPLC) is used to test the identity of desmoteplase in a given sample. The standard used is an aliquot of desmoteplase reference material. The analysis results in a peak profile with 6 distinct peaks indicating the microheterogeneity of the molecule with regard to glycosylation including sialylation and various N-termini.

The relative abundance of each peak is calculated with regard to the total peak area of the whole sample. The relative ratio of the peaks is summarised in Table 1A.

7.5 Glycosylation/Fingerprint

To determine the antennarity of desmoteplase carbohydrate structures are enzymatically cleaved with N-glycosidase F from the protein. Prior to chromatography any interfering proteins and salts are removed. For the determination of neutral oligosaccharide structures (antennarity status), the oligosaccharides have to be desialylated.

Neutral oligosaccharides are separated by using HPAEC (high performance anion-exchange chromatography) and detected by pulsed amperometric detection (PAD). The antennarity fingerprint has to comply qualitatively with that of the current rDSPAal reference standard.

7.6 Sialylation

For the determination of sialic acid content in the desmoteplase molecule the glycosidically attached sialic acid is cleaved by the enzyme sialidase to obtain free N-acetyl-neuraminic acid (NeuAc). Released free sialic acids are separated by HPAEC (high performance anion- exchange chromatography) and detected by pulsed amperometric detection (PAD). Sialic acid (NeuAc) is quantified by external calibration with an appropriate standard and expressed as mol sialic acid per mol rDSPAal . 7.7 RP-HPLC

C4-reversed Phase-High Performance Liquid Chromatography (RP-HPLC) is used to determine the concentration of desmoteplase in a given sample. Samples are chromatographed using a linear gradient of trifluoroacetic acid diluted 1/1000 in purified water (Eluent A) and trifluoroacetic acid diluted 1/1000 in acetonitrile (Eluent B). The standard used is desmoteplase reference material at a known concentration.

The concentration of desmoteplase is proportional to the area obtained at 214nm and is evaluated by using the calibration curve of the reference standard. The result is expressed as mg/mL.

7.8 A₂₈₀/UV

This analysis is used to determine the concentration of desmoteplase in a given sample. Therefore, the absorbance is measured at 280 nm and the amount of desmoteplase expressed as mg/mL is calculated using the absorbency index of 1.71 ml_/(mg x cm) for desmoteplase.

7.9 Aggregates (SEC)

Size exclusion-HPLC separates monomers from potentially occurring dimers or oligomers (aggregates) depending on the size of the molecule. The eluate is monitored at 214 nm and the purity is determined as the relative percentage in area of monomer to that of total area of all protein peaks detected.

7.10 HCP

The content of residual host cell proteins (HCPs) is determined using a commercial ELISA kit from Cygnus Technologies (CM015) specially intended for products manufactured from recombinant expression in Chinese Hamster Ovary (CHO) cells. The polyclonal antibodies in the kit are raised against HCPs typically found in protein free growth media of recombinant

CHO cells.

In this sandwich ELISA HCPs present in samples are bound via an affinity purified capture antibody to microtiter plate. After an incubation step a second horseradish peroxidase (HRP) labelled anti-CHO-HCP antibody is added, resulting in the formation of a sandwich complex (solid phase antibody - CHO-HCPs - HRP labelled antibody). Unbound reactants are removed by a washing step and subsequently the substrate tetramethyl benzidine (TMB) is added. The amount of oxidised substrate is read on a microtitre plate reader at 450 nm and is directly proportional to the concentration of CHO-HCPs present. Results are reported in ng HCPs per mg desmoteplase. 7.11 DNA (Threshold)

Residual DNA is detected using the commercial Threshold Total DNA Assay Kit (Molecular Devices). Diluted samples are first treated with proteinase K in the presence of SDS to remove interfering protein. Following the proteinase K pre-treatment, samples are heat- denatured to form single stranded DNA (ssDNA). A ssDNA-Protein complex is then formed by incubating the sample with "labelling reagent" (component of the kit) containing biotinylated ssDNA binding protein, streptavidin and a monoclonal anti-ssDNA-antibody conjugated with urease. The quantification is independent of the DNA-sequence and is based on high affinity interaction of biotinylated E. coli single-strand binding protein (SSB) with ssDNA. The biotinylated ssDNA-protein complex is then captured on a biotinylated nitrocellulose membrane by filtration utilising the high affinity of streptavidin and biotin. The enzymatic activity of the immobilised urease causes a pH change over time, which is proportional to the amount of DNA present in the test article. Results are reported in ng per dose. 7.12 Bioburden

The bioburden is determined according to the method described in Ph.Eur. 2.6.12 and expressed as colony-forming units (CFU) per ml_.

7.13 Endotoxins

Bacterial endotoxins are determined according to the turbidimetric kinetic method (method C), which is described in Ph.Eur. 2.6.14. Results are reported in endotoxin units (EU) per mg desmoteplase.

7.14 S-2288

S-2288 [D-Ile-Pro-Arg-NH-Phenyl-N0₂] is a chromogenic substrate for a broad range of serine proteases. Desmoteplase cleaves the p-Nitroaniline group from the substrate, and the p-Nitroaniline absorbs light at a wavelength of 405 nm. The change in absorbance value obtained over time is proportional to the amount of desmoteplase enzymatic activity present. The result is expressed in mg/mL of standard showing the same S-2288 activity. Additionally, a positive control is used for this assay. 7.15 Plate Clot Lysis

This procedure is used to measure the fibrinolytic activity of desmoteplase by microtitre plate clot lysis assay using an in vitro generated fibrin clot as a substrate. A fibrin clot forms when fibrinogen is converted to fibrin by thrombin-induced specific proteolytic cleavages by thrombin. Fibrin aggregates with plasminogen to form a clot, which causes the solution to become turbid. The turbidity is quantified by measuring the absorbance at 405 nm. Concurrently, plasminogen is proteolytically cleaved to active plasmin by desmoteplase, a plasminogen activator. The plasmin cleaves the fibrin polypeptides, which dissolves the clot. This reduces the turbidity and thus the absorbance. The decrease in absorbance at 405 nm over time is proportional to the fibrinolytic activity of desmoteplase. The quantitative parameter is the time in seconds required for the absorbance at 405 nm to decrease from its maximal value to the 50% value of initial maximum absorbance (T50). The result is expressed in KU/mL of standard showing the same fibrinolytic activity. A positive control is used for this assay.

7.16 PCL/RP-HPLC

The quotient between the fibrinolytic activity of desmoteplase towards fibrin via its specific substrate plasminogen divided by the protein concentration as determined with RP-HPLC was calculated to give a value for the specific activity of desmoteplase.

7.17 Mannitol/Glycine

Content of glycine and mannitol in desmoteplase formulations and placebo is determined by HPLC with refractive index (Rl) detection. Glycine and mannitol are separated on a NH₂- phase column at 40°C using an isocratic gradient of sodium diphosphate (8.3 mmol/L)/acetonitrile (22:78) at a flow rate of 1.5 mUmin. The standards used for the reference solutions are USP reference standard materials.

For quantification a one-point calibration with the respective reference solutions (100% level) in water/acetonitrile (1 :1) is performed. The 100%-concentration is defined with 5.3 mg/mL (mannitol) and 2.0 mg/mL (glycine), respectively.

8. Container Closure System

The drug substance is initially stored in 20 L or 50 L Flexboy bags (Stedim) directly after the dia-ultrafiltration step. Drug substance is then transferred into either 1 L and 30 mL PETG containers (NNI) for long-term storage at -70°C.

9. Stability of the drug substance

The stability of the drug substance has been assessed in a stability study at -70 °C. The parameters, methods and acceptance criteria are given in Table (back up). Based on this stability study the drug substance demonstrated stability at -70°C for at least up to 9 months. EXAMPLE 2: Manufacturing of the drug product

1. Description and Composition of the drug product

The drug product is a lyophilised material containing 9 mg desmoteplase in a 10ml glass vial (e.g. a 10R vial). The material is reconstituted with 10 mL Water for Injection (WFI) immediately prior to administration. The complete formulation is shown in Table 11.

2. Description of manufacturing process

A flow-chart of the manufacture of the drug product is given in Figure 12. A detailed description is given below.

2.1 Formulation/sterile filtration

The formulation to the drug product comprises thawing and mixing the drug substance, following final sterile filtration through two in-line Hydrophilic 0.22 μιτι Gamma Gold Millipak filters.

The formulation process is performed in a LAF (Laminar Air Flow) hood (Class C). Frozen drug substance solution is carefully thawed from -70°C to -20°C to 2-8°C within 2 days for each step. After thawing drug substance from all containers these are pooled and carefully mixed for 10 min. The solution is then filtered through two in-line Hydrophilic 0.22 m Gamma Gold Millipak filters into a sterile bulk filling bottle.

2.2 Aseptic filling

Aseptic filling is performed with a filling machine under LAF (Class A) in a Class B background. The sterile filtered desmoteplase solution is automatically pumped into depyrogenated and sterilised DIN vials. The filling volume is checked periodically by weight (3.40 g/vial; ± 4%). Directly after the filling point a sterilised stopper is automatically placed in the vial neck in lifted position. The vials are transferred to the freeze dryer.

2.3 Lyophilisation

The vials with aseptic filled product are lyophilised. During the lyophilisation process the vacuum and the temperature are controlled. When the lyophilisation is complete, the freeze dryer is pre-aerated with sterile filtered nitrogen. The vials are automatically stoppered inside the freeze dryer. When the vials were fully stoppered, the freeze dryer is aerated to atmospheric pressure.

2.4 Capping and Visual Inspection

When the lyophilisation process is complete, the freeze dryer is opened and the vials are capped with 20 mm aluminium flip-off caps with a plastic disc. After capping the vials are visually inspected for defects. After visual inspection the vials are transferred to and stored at the warehouse at 2-8°C. 3. Description of the Drug Product

A summary of the analytical data for the drug product as result of the above described production process is given in Tables 12A and 12B. The specific methods are described in Example 1 chapters 6 and 7.

4. Stability of the Drug Product

The stability of the drug product has been assessed in a stability study at 2-8°C. The parameters, methods and acceptance criteria are given in Table 13. Based on this stability study the Drug Product demonstrated stability at 2-8°C for at least up to 48 months.

EXAMPLE 3 General strategy of the virus validation study.

In order to test the feasibility of the different purification steps for virus removal or inactivation six different model viruses were used with in the study: Murine leukaemia virus (MuLV), Pseudorabies virus (PRV), Encephalomyocarditis virus (EMCV), Porcine Parvovirus (PPV), Herpes simplex virus (HSV) and Bovine papilloma virus (BPV) (for further description see also Table 14). MuLV is a relevant virus since CHO cells reveal endogenous C-type retroviruses. PPV is a model for human parvoviruses and BPV is a model for the human papilloma viruses. The latter two viruses constitute a severe test of the capacity of the downstream process. PRV, HCV and EMCV complete the spectrum of physico-chemical properties and therefore contribute to the general virus clearance capacity of the process.

The DSPA alpha 1 containing sample was spiked separately with the six model viruses and purified using four different downscaled purification methods of the original DSPA alpha 1 downstream process. After the purification steps the virus load was analysed and a decrease in the virus load calculated as log₁₀ values. For each production step three runs were conducted to demonstrate that the step is reproducible at downscale and the model is comparable to the production scale.

In addition the yield and activity of rDSPA alphal was determined after each purification step in order to show that the respective virus removal/inactivation step has no relevant impact on the yield or the activity of rDSPA alphal .

Description of the virus inactivation by acidic incubation in the presence of ethanol

The purpose of this step is the inactivation of theoretically occurring acid-labile enveloped viruses and acid-resistant non-enveloped viruses. The enveloped virus species MuLV, PRV and the EMCV as non-enveloped virus were used for the virus validation of this process step. The inactivating effect results from the low, acidic pH value of less than 2.5 and the high ethanol concentration of at least 25%. The product (=eluate from the HIC chromatography) containing 29.5% of ethanol will be adjusted to a pH value of 2 by addition of 20 mM HCI in a ratio of 3 to 1. After incubation for 18 hours at room temperature, the product can be stored at 2-8°C for up to 8 days until the product is processed further

Methods

In the virus validation study the samples from the low pH treatment batches were handled as follows:

20 ml of starting material was adjusted to pH 2.15 - 2.20 and incubated in a water bath at

21 °C ± 3°C. The pH-adjusted and tempered starting material was spiked with 2 mL virus stock solution. The pH-value was controlled and re-adjusted to pH 2.15 - 2.20 if necessary and the time period was started after the target pH was reached. The sample "load" was taken from a batch designated as "medium control"; 10 mL of medium was spiked with 1 mL virus spike. Samples were withdrawn after 0, 10, 10, 30, 60, and 180 minutes of incubation, for the EMCV batches, additional samples were drawn after 360 and 900 minutes. Each sample was diluted with cell culture medium directly after receipt and titrated immediately.

Results

The results for the batches are given in Tables 15 to 17.

MuLV: As displayed in table 15, the inactivation of MuLV was highly effective even after 0 minutes with a virus reduction of at least 5.95± 0.29 log₁₀. The inactivation was therefore almost completed immediately after spiking. Due to the lower detection limit for samples "60 min" and "180 min", the complete inactivation resulted in a virus reduction of at least 6.25 ± 0.29 log₁₀.

PRV: As displayed in table 16, the inactivation of MuLV was effective even after 0 minutes. The inactivation was therefore completed immediately after spiking. Due to the lower detection limit for samples "60 min" and "180 min", the complete inactivation resulted in a virus reduction of at least 6.43 ± 0.27 log₁₀.

EMCV: Since EMCV is a non-enveloped virus species and thus more resistant to physico- chemical treatment the incubation time for inactivation was extended to 900 min. In both batches the inactivation of EMCV was complete after 360 min (Table 17). Due to the low detection limit for the samples "360 min" and "900 min" the complete inactivation resulted in a virus reduction of at least 6.25 ± 0.29 log₁₀. The inactivation kinetics for the EMCV-spiked demonstrates very nicely the very rapid time dependent virus inactivation with effective virus inactivation at 60 min (4.02 ± 0.38 log₁₀) and an almost complete inactivation at 180 minutes of acidic incubation (5.95 ± 0.34 log₁₀) (see Figure 13). 1884

38

Conclusion

The low pH treatment showed complete inactivation for the runs with MuLV and PRV (both enveloped) after 60 minutes. The incubation time was extended for the non-enveloped virus EMCV. Inactivation of EMCV was almost completed at 180 minutes with a maximum inactivation reached at 360 minutes.

EXAMPLE 4 Virus removal using a Sartobind Q membrane adsorber (MAQ)

For this purification step the Sartobind Q membrane adsorber was chosen. Due to the positively charged ammonium groups this filter is capable to bind negatively charged contaminants like DNA, endotoxins and viruses. Methods

In order to improve the MAQ binding capacity for the above listed impurities the capture eluate from the Poly ABX step (see Example 3) was diluted 2.5-fold with Water for injection so that the conductivity of the sample is approx. 11 mS/cm at a pH of 7.0 to 7.5. As an equilibration/wash buffer 20 mM phosphate pH 7.3 + 80 mM NaCI was used.

For each run 150 ml_ starting material were spiked with 7.5 ml_ of virus stock of the particular virus type. After mixing the spiked starting material was filtered over the appropriate pre-filter (MuLV/PRV: 0.45 pm; EMCV/PPV; 0.1 m). The sample "load" was withdrawn from the pre- filtered material and titrated immediately. The sample "hold" was withdrawn as well, but kept at room temperature until reception of the final fraction, in order to normalize for a potential reduction in virus load due to incubation in the respective buffer.

After pre-equilibration with 200 mM Na-Phosphate pH 7.2 the column was equilibrated with 20 mM Na-Phosphate pH 7.3 + 80 mM NaCI and 144 ml of the diluted pre-filtered material was loaded and four fractions FT1 to FT4 of approx. 36 ml were collected. Afterwards the column was flushed with 18 mL of 20 mM Na-Phosphate pH 7.3 + 80 mM NaCI. The samples FT1 to FT4, the flush fraction and the "hold" sample were titrated immediately and collectively. The flow rate for all MAQ process steps amounts to 1.44 ml/min. Results

The results for the batches are given in Tables 18 to 22.

MuLV: The titer of MuLV was below the limit of detection throughout the fractions collected. A reduction factor of at least 5.54 log₁₀ was demonstrated (see Table 18). No infectivity was found in the flash fraction. Therefore, the MAQ filtration can be rated as effective concerning the removal of this virus type. PRV: The titer of PRV was below the limit of detection throughout the fractions collected. A reduction factor of at least 5.30 log₁₀ was demonstrated (see Table 19). No infectivity was found in the flash fraction. Therefore, the MAQ filtration can be rated as effective concerning the removal of this virus type.

EMCV: The fact that EMCV was not depleted from the product pool points out that there are substantial differences in properties of virus capsids throughout the wide range of virus types (see Table 20). In this case, EMCV seemed to carry a positive net charge and hence did not bind to the matrix under the conditions tested. This effect was already known from Polio virus, but was not predicted for EMCV.

PPV: PPV was effectively removed by MAQ. A reduction factor of at least 4.46 log₁₀ was demonstrated in fractions 2 to 4 and the flush fraction (see Table 21). Minimum activity was found in fraction 1 which was analyzed at a higher sensitivity compared to the other fractions. The viral titer remained stable in the hold sample.

DNA and HCP: In addition to the virus titration the samples were analysed for the DNA and host cell proteins (HCP) The results are summarized in Table 22. The DNA content was reduced by MAQ by a factor of 2.54 log₁₀. The HCP were removed very effectively by MAQ with a reduction factor of 12.35 log₁₀. The optimised process resulted in a remarkable high yield of approx. 95% of DSPA alpha 1.

Conclusion

The MAQ filtration as implemented in the purification process of DSPA alpha 1 proved to be very effective concerning the reduction of MuLV, PRV and PPV. The virus titers for these virus types were pushed below the limit of detection. However, EMCV was not depleted by MAQ, which points out that there are substantial differences in properties of virus capsids throughout the wide range of virus types. In this case, EMCV seemed to carry a positive net charge on its surface and hence did not bind to the matrix under the conditions tested. In this regard it is important to note that the EMCV virus can be inactivated by the pH inactivation step and removed by the nanofiltration step.

EXAMPLE 5 Virus validation study of the cation exchange chromatography using Poly ABx-35 resin

In this capture step a majority of the impurities will be removed and the product will be concentrated. It could be shown that this capture step leads also to a removal of both model viruses tested in the study. For the BPV virus the removal with a log value of above 2 can be rated as effective and with regard to this virus the capture chromatography step contributes to the overall virus safety of the downstream process. Methods

The bioreactor harvest containing the recombinant DSPA alpha 1 is stored at ambient temperature (15-25°C). Purification of the harvest begins with a column chromatography step using the Bakerbond™ XWP 500 PolyABX-35 resin made by J. T. Baker.

During the loading process the pH value of the harvest will be adjusted to 5.5 by addition of 0.33% acetic acid containing 1 10mM NaCI in an online dilution procedure. After adjustment of the pH value the harvest is filtered through a 0.45pm and a 0.2pm filter prior to loading onto the cation exchange column in order to separate occurring precipitates. The column is first washed with 50 mM acetate buffer pH 5.5 followed by a 50 mM phosphate buffer pH 7.1 The product is eluted from the column with 50 mM phosphate pH 7.5 as a starting buffer and step-gradient up to 0.2 M NaCI whereby the conductivity is increased to 25mS/cm. Results

It was found out that the pH increase of the second wash buffer from pH 5.5 to pH 7.1 led to a significant desorption of HCP and isoforms of DSPA alpha 1. The optimised process resulted in a yield of approx. 85% of DSPA alpha 1. HSV: The herpes simplex virus was reduced by a factor of 1.2 log₁₀ (see Table 23). No infectivity was found in the flash fraction. Therefore, the MAQ filtration can not be rated as effective concerning the removal of this virus type.

BPV: For the bovine papilloma virus a reduction factor of at least 2.3 log was demonstrated (see Table 23). Therefore, the Poly ABx step can be rated as effective concerning the removal of this virus type.

Conclusion

The results demonstrated that the cation exchange chromatography using Poly Abx-35 was able to reduce the virus titers of both viruses: the HSV and the BPV.

EXAMPLE 6

Virus removal using a nanofiltration

Nanofiltration is an established method for removal of viruses and other particulate impurities. The membrane used in this step is arranged as hollow fibers with a pore size of 15±2 nm and is made of cuprammonium regenerated cellulose, providing minimum nonspecific interactions with proteins. This nanofiltration membrane is designed to provide process safety concerning virus safety issues. Due to its narrow pore size it is capable of clearing even the smallest known viruses (18 - 24 nm) effectively. Methods

For this purification step the PLANOVA 15N filter was chosen which contains cuprammonium regenerated cellulose as hollow fibre set-up. The membrane area of 0.001 m² represents a 1 :1000 scale-down factor of the production scale nanofiltration equipment with a membrane area of 1.0 m². However, the maximum load volume per membrane surface area of 25.0 L/m² and the operational pressure of 0.9 to 0.95 bar as relevant process parameters are nearly identical to the production scale. The virus-spiked nanofiltration runs followed the same procedure as given throughout the validation of the downscale model with exception that the loading material was spiked with the four test viruses.

35 ml of the starting material was pre-filtered over a 0.1 pm-syringe filter to avoid product aggregates, and 30 ml of this material was spiked with 3.0 ml of virus stock solution. The samples "load" and hold" were withdrawn after spiking and mixing. The sample "load" was immediately titrated for virus titer. The virus-spiked virus material was pre-filtered again with a 0.45 m filter (MuLV/PRV) or a 0.1 pm filter (EMCV/PPV) resulting in the sample "F1 filtrate" which was also immediately titrated. At least 25 g (equivalent to 25 ml) of the starting material was processed through the PLANOVA 15N filter plus at least 5 g (equivalent to 5 ml) equilibration buffer as post wash. "F2 filtrate" was collected within two fractions F21 and F22; fraction. F21 comprised the F2 filtrate from 0 to 15g and F22 comprised 15 to 25g (plus "post-wash"). Samples F21 and F22 was titrated immediately after receipt; the sample hold was titrated together with sample F22. After completion of the virus-spiked nanofiltration the post-integrity test was performed.

Results

The results for the batches are given in Tables 24 to 27.

MuLV: As no virus was found in the filtrate fractions, the reduction of MuLV from the product pool by PLANOVA 15N filtration can be rated as effective (see Table 24). Due to the detection limit the complete lack of viral infectivity in the nanofiltrate fractions resulted in a virus reduction of > 5.48±0.22 log₁₀.

PRV: As no virus was found in the filtrate fractions, the reduction of PRLV from the product pool by PLANOVA 15N filtration can be rated as effective (see Table 25). Due to the detection limit the complete lack of viral infectivity in the nanofiltrate fractions resulted in a virus reduction of > 6.31 ± 0.28 log₁₀.

EMCV: As no virus was found in the filtrate fractions, the reduction of EMCV from the product pool by PLANOVA 15N filtration can be rated as effective (see Table 26). Due to the detection limit the complete lack of viral infectivity in the nanofiltrate fractions resulted in a virus reduction of > 6.07 ± 0.29 log₁₀. T EP2011/001884

42

PPV: As no virus was found in the filtrate fractions, the reduction of PPV from the product pool by PLANOVA 15N filtration can be rated as effective (see Table 27). The virus reduction is at least > 5.08 ± 0.42 log₁₀. Conclusion

The results demonstrated that the eight virus validation runs were highly comparable. The samples drawn from the nanofiltration runs with MuLV/PRV/EMCV showed no residual viral activity. These three virus species were cleared below the limit of detection. For PPV nanofiltrations residual viral infectivity was found in both nanofiltrate fractions. The virus reduction of at least 5.08±0.42 log₁₀ can be rated as effective.

The log reduction factors observed for the MAQ, low pH treatment and nanofiltration are compiled in Table 28.

EXAMPLE 7

Analysis of rOSPA alphal microheterogenity

The purpose of this chromatography is to evaluate the rDSPA alphal microheterogeneity. A TSK SP-5PW column is packed with porous hydroxyiated polymethacryiate beads (mean particle size: 10pm, mean pore size: 1000A) which are surface modified with a strong cation exchanger (functional group -CH₂-CH₂-CH₂-SO₃ ^").

The loading sample of the TSK SP-5PW column is DPSA dissolved in 4% mannitol/ 200 mM glycine buffer. The column is pre-equilibrated by successive rinsing using 20 mM Na₂HP0 buffer pH 7.0 for 15 minutes and 20 mM Na₂HP0₄ buffer + 1 M NaCI pH 7.0 for 15 minutes (each step until a constant base line is reached). The loading step is performed using 20 ml of a buffer containing 20 mM Na₂HP0₄ at a flow rate of ImUmin. Elution of the captured desmoteplase is via a linear salt gradient (20 mM Na₂HP0₄ buffer + 1 M NaCI pH 7.0), which is performed at a flow rate of 1 ml_/min. Collection of the eluting product is controlled by A₂₈₀ measurement.

The DSPA eluate shows six main peaks that are denominated with increasing retention times as peak 1 to 6. This profile is exemplified in Figure 10. The analysis of the area under the peak curves (given as area percentage related to the area of the combined six peaks) represents a characteristic finger print defining the DSPA microheterogeneity (as given in Table 1A). 001884

43

List of Abbreviations

A₂8o - Absorption at 280nm

aa / AA - amino acid

Accl - Restriction Endonuclease Accl

Alpha MEM - alpha Minimum Essential Medium

Alul - Restriction Endonuclease Alul

BamHI - Restriction Endonuclease BamHI

CD16 - desmoteplase producing CHO cell line

cDNA - complementary DNA

CFU - Colony Forming Units

CHO - Chinese Hamster Ovary

CHO-V - CHO-V™ medium (Irvine Scientific)

CHO K1 - Chinese Hamster Ovary cells

CIEX - Cation Exchange Chromatography

Clal - Restriction Endonuclease Clal

CUNO - CUNO® filter ZETA PLUS® 90 LP

CV - Column Volume

DAD Diode Array Detector

dFBS - dialysed Foetal Bovine Serum

DHFR - Dihydrofolate Reductase

DIAS - Desmoteplase In Acute ischemic Stroke

DMSO - Dimethyl Sulphoxide

DNA - Desoxyribonucleic Acid

DSPA - Desmodus rotundus Salivary Plasminogen Activator

rDSPAal - recombinant Desmodus rotundus

Salivary Plasminogen Activator alpha 1

p0₂ - partial pressure Oxygen

DXB11 - CHO cell line DXB11

E. coli - Escherichia coli EcoRI - Restriction Endonuclease EcoRI

ELISA - Enzyme Linked Immunosorbant Assay

EMCV - Encephaolomyocarditis virus

E-MuLV - Ecotropic Murine leukaemia virus

EPC - End of Production Cells

EPO - Erythropoetin

ESI-MS - Electrospray lonisation Mass Spectroscopy

EU - Endotoxin Units

FBS - Foetal Bovine Serum

FCS - Foetal Calf Serum

FID Flame ionisation detection

FT-IR - Fourier Transformation - Infrared Spectroscopy

HA Haemagglutination

HCP - Host Cell Proteins

HEPES - 4-(2-hydroxyethyl)-1 -piperazineethanesulfonic acid

HIC - Hydrophobic Interaction Chromatography

HPAEC High performance anion-exchange chromatography

HPLC - High Performance Liquid Chromatography

HRP Horseradish peroxidase

HSE - House Method

ICH - International Conference on Harmonisation

ID - Identification

INN - International Non-proprietary Name

IPC - In Process Control

kD - kilo Dalton

LAF - Laminar Air Flow

LC/MS - Liquid Chromatography / Mass Spectrometry

LCMV - Lymphocytic Choriomeningitis Virus

LOD - Limit of Detection

Log₁₀ decadic logarithm LOQ Limit of Quantification

LysC - Endoproteinase LysC

m - month(s)

MALDI-reTOF Matrix Assisted Laser Desorption lonisation - reflection Time

Flight

MAP Mouse Antibody Production

MCB Master Cell Bank

MEM Minimal Essential Medium

min Minute(s)

MMLV MMLV reverse transcriptase

MRS Matches Reference Standard

MTX Methotrexate

MuLV Murine Leukaemia Virus

NaAc Sodium Acetate

NaCI Sodium Chloride

N/D not done

NeuAc N-acetyl-neuraminic acid

ORF Open Reading Frame

ORI Origin of Replication

PAGE Polyacrylamide Gel Electrophoresis

PAI-1 plasminogen-activator inhibitor 1

pBluescript(R) SK- pBluescript(R) SK phagemid

PCL Plate Clot Lysis

PCR Polymerase Chain Reaction

Ph.Eur. European Pharmacopoeia

PPC Post Production Cells

PPV Porcine Parvo Virus

PRV Pseudo Rabies Virus

pSVPA1 1 eukaryotic expression plasmid pSVPA1 1

pUDHFR I dhfr coding region containing plasmid pUDHFR pSVL mammalian expression vector pSVL

QA Quality Assurance

QC Quality Control

RH Relative Humidity

Rl Refractive Index

RNA Ribonucleic Acid

RP-HPLC Reversed Phase High Pressure Liquid Chromatography

RT Reverse Transcriptase / Room temperature

S-2288 Desmoteplase S-2288 potency assay

Sail Restriction Endonuclease Sail

SDS Sodium Dodecyl Sulfate

SDS-PAGE - Sodium Dodecyl Sulfate-PolyAcrylamide Gel Electrophoresis

SIP sanitised in place

SSC Saline-Sodium Citrate buffer

SOP Standard Operating Procedure

SV40 Simian Virus 40

TEM Transmission Electron Microscopy

TFF Tangential flow filtration

TMP Transmembrane pressure

TOF Time Of Flight

TP8K6 desmoteplase producing CHO cell line

t-PA tissue plasminogen activator

UF Ultrafiltration

USP United States Pharmacopoeia

UV Ultraviolet

WCB Working Cell Bank

WFI Water for Injection

w/o without

Xbal Restriction Endonuclease Xbal

XC continuous rat cell line 4

47

Table legends

Table 1A&B: Summary of the DSPA drug substance characterization

Table 2: Primary amino acid sequence for secreted active desmoteplase

Table 3: Amino acid composition for desmoteplase

Table 4: Prediction and structure-based estimate of the secondary structure of desmoteplase

Table 5: Proposed S-S Bridges in desmoteplase

Table 6: Peptide Clusters of desmoteplase by S-S linkage (non-reduced trypsin digest)

Table 7: S-S linked peptide clusters in "Third" Optimised material (Trypsin (grey) and alternative proteolytic digests (white))

Table 8: Heterogeneity of Desmoteplase

Table 9: Main Glycoforms of Desmoteplase

Table 10: Ratio of glycosylated/non-glycosylated desmoteplase

Table 11: Steps of Lyophilisation process

Table 12A&B: Summary of the DSPA drug product characterization

Table 13: Methods for DSPA drug substance stability analysis

Table 14: Model viruses used in the virus validation studies

Table 15: Log₁₀ reduction factors for MuLV in low pH treatment step

Table 16: Log₁₀ reduction factors for PRV in low pH treatment step

Table 17: Log₁₀ reduction factors for EMCV in low pH treatment step

Table 18: Log₁₀ reduction factors for MuLV spiked MAQ filtration

Table 19: Log₁₀ reduction factors for PRV spiked MAQ filtration Table 20: Log₁₀ reduction factors for EMCV spiked MAQ filtration

Table 21 : Log₁₀ reduction factors for PPV spiked MAQ filtration

Table 22: Log₁₀ reduction factors for DNA and HCP in MAQ filtration

Table 23: Log₁₀ reduction factors of virus load with PolyABx column

Table 24: Log₁₀ reduction factors for MuLV spiked nanofiltration

Table 25: Log₁₀ reduction factors for PRV spiked nanofiltration

Table 26: Log₁₀ reduction factors for EMCV spiked nanofiltration

Table 27: Log₁₀ reduction factors for PPV spiked nanofiltration

Table 28: Compilation of the virus validation methods with MAQ filtration, low pH incubation and nanofiltration

Figure legends

Fig. 1 : Flow scheme for the purification of DSPA alpha 1

Fig. 2: Pictogram of desmoteplase manufacturing process (Upstream and

downstream).

Fig. 3 A - C: Flow Diagram of Upstream process and Downstream process

Fig. 4: DSPA batch definition

Fig. 5: Flowchart for isolation of CD16

Fig. 6 A: Schematic Diagram of Desmoteplase Expression Plasmid pSVPA11

Fig 6 B-K: Complete nucleotide sequence of the plasmid pSVPA1 1

Fig. 7 A: Schematic Diagram of dhfr + co-transfection Plasmid pUDHFR

Fig. 7 B-F: Complete nucleotide sequence of the plasmid pUDHFR

Fig. 8: Domain Structure for Desmoteplase in Schematic Form (left: pearl-model;

right: ribbon.

Fig. 9: FT-IR spectra at different pH (left panel) and temperatures (right panel)

Fig. 10: CIEX profile of desmoteplase (Optimised drug substance lot DS 30309012)

Fig. 11 : IEF of Desmoteplase

Fig. 12: Flow Diagram for the DSPA drug product manufacturing process

Fig. 13: Inactivation of EMCV by acidic ethanolic incubation as function of the

incubation time. Table 1 A

used for calculation of mass based values Table 2 B

Category Method Results Reference

> 90% DSPAocl

SDS-PAGE

% individual impurities to SDS-PAGE Colloidal Blue Colloidal Blue

be determined, report stain

reduced

Purity results above LOQ (%)

> 90% DSPAocl

SDS-PAGE

% indivual impurities to be SDS-PAGE Colloidal Blue Colloidal Blue non- determined, report results stain

reduced

above LOQ (%)

Product-related ≤ 5% aggregates; (report Size exclusion

Aggregates (SEC)

impurities % monomer) chromatograhy

≤ 100 ng/mg ELISA kit Cygnus

HCP

desmoteplase Technologies

Process-related DNA (Threshold) < 10 ng/dosis (Dosis 9 mg) Threshold total DNA kit Impurities

Bioburden ≤ 1 CFU/ ml. Ph. Eur. 2.6.12

Endotoxins ≤ 3.0 EU/mg Ph. Eur. 2.6.14, Method C

Desmoteplase cleaves the p-Nitroaniline from a

S-2288 3.82 mg/mL ± 0.57 mg/mL chromogenic substrate, which is detected at 405

Activity / Potency nm

PCL/RP-HPLC 40 - 120 kU/mg -

Fibrinolytic activity and

PCL 200 - 410 kU/mL quantitation by

absorbance at 405nm

Mannitol 40.0 mg/mL ± 4.0 mg/mL HPLC

Glycine 15.0 mg/mL ± 1.5 mg/mL HPLC

General tests I

PH 5.0 - 7.0 Ph. Eur. 2.2.3

435 mOsmol/kg ± 45

Osmolality Ph. Eur. 2.2.35

mOsmol/kg

Table 2

I N+3 I 1

After secretion and processing --> (GlySerArg) -AlaTyrGlyVal aa# 5 AlaCysLysAspGluIleThrGlnMetThrTyrArgArgGlnGluSerTrpLeuArgPro aa# 25 GluValArgSerLysArgValGluHisCysGlnCysAspArgGlyGlnAlaArgCysHis aa# 45 ThrValProValAsnSerCysSerGluProArgCysPheAsnGlyGlyThrCysTrpGln aa# 65 AlaValTyrPheSerAspPheValCysGlnCysProAlaGlyTyrThrGlyLysArgCys aa# 8 5 GluValAspThrArgAlaThrCysTyrGluGlyGlnGlyValThrTyrArgGlyThrTrp aa# 105 SerThrAlaGluSerArgValGluCysIleAsnTrpAsnSerSerLeuLeuthrArgArg aa# 125 ThrTyrAsnGlyArgMetProAspAlaPheAsnLeuGlyLeuGlyAsnHisAsnTyrCys aa# 145 ArgAsnProAsnGlyAlaProLysProTrpCysTyrVallleLysAlaGlyLysPheThr aa# 165 SerGluSerCysSerValProValCysSerLysAlaThrCysGlyLeuArgLysTyrLys aa# 185 GluProGlnLeuHisSerThrGlyGlyLeuPheThrAspIleThrSerHisProTrpGln aa# 205 AlaAlallePheAlaGlnAsnArgArgSerSerGlyGluArgPheLeuCysGlyGlylle aa# 225 LeuIleSerSerCysTrpValLeuThrAlaAlaHisCysPheGlnGluSerTyrLeuPro aa# 245 AspGlnLeuLysValValLeuGlyArgThrTyrArgValLysProGlyGluGluGluGln aa# 2 65 ThrPheLysValLysLysTyrlleValHisLysGluPheAspAspAspThrTyrAsnAsn aa# 285 AspIleAlaLeuLeuGlnLeuLysSerAspSerProGlnCysAlaGlnGluSerAspSer aa# 305 ValArgAlalleCysLeuProGluAlaAsnLeuGlnLeuProAspTrpThrGluCysGlu aa# 325 LeuSerGlyTyrGlyLysHisLysSerSerSerProPheTyrSerGluGlnLeuLysGlu aa# 345 GlyHisValArgLeuTyrProSerSerArgCysAlaProLysPheLeuPheAsnLysThr aa# 365 ValThrAsnAsnMetLeuCysAlaGlyAspThrArgSerGlyGluIleTyrProAsnVal aa# 385 HisAspAlaCysGlnGlyAspSerGlyGlyProLeuValCysMetAsnAspAsnHisMet aal 05 ThrLeuLeuGlyllelleSerTrpGlyValGlyCysGlyGluLysAspValProGlyVal aa# 425 TyrThrLysValThrAsnTyrLeuGlyTrpIleArgAspAsnMetHisLeu⁽⁴⁴¹⁾

Secreted protein of 441 amino acids calculated Mol. Weight = 49543.91

Table 3

Protein Name = desmoteplase

Amino Acid Molecular Weight Calculator

Amino Acid MW Count SumMW

Alanine Ala A 89 23 2047

Valine Val V 117 28 3276

Leucine Leu L 131 30 3930

Isoleucine He I 131 14 1834

Proline Pro P 115 22 2530

Phenylalanine Phe F 165 14 2310

Tryptophan Trp W 204 10 2040

Methionine Met M 149 6 894

Glycine Gly G 75 36 2700

Serine Ser S 105 34 3570

Threonine Thr T 1 19 28 3332

Cysteine Cys c 121 28 3388

Tyrosine Tyr Y 181 20 3620

Asparagine Asn N 132 22 2904

Glutamine Gin Q 146 19 2774

Aspartic Acid Asp D 133 20 2660 w/ASN

Glutamic Acid Glu E 147 25 3675 w/GLN

Lysine Lys 146 23 3358

Arginine Arg R 174 27 4698

Histidine His H 155 12 1860

Total AA = 441 Total w/o H₂0 = 49480

2011/001884

54

Table 4

PDB- Length ot-helix β-sheet

Domain

Id. [aa] [aa] [aa]

Fibronectin-I 1 FBR 43 24 0

EGF-like 1 IXA 39 8 0

Kringle 1TPK 82 14 0

Peptidase S1 1A5I 251 78 17

Table 5

1 2 3 4 5 6 7

Cys-aa# 6-36 34-43 51-62 56-73 75-84 92-173 1 13-155

8 9 10 11 12 13 14

Cys-aa# 144-168 178-309 221-237 229-298 323-398 355-371 388-416

Table 6

Number

Number

Peptide of S-S W MW of S-S S-S Linked Peptides (aa numbers)

Cluster linked theor. experim *

Bridges

peptides

1 5 5 1 -7 31-38 43-55 56-82 84-89 8198.1 8199.4

130- 163-

2 2 3 90-101 4539.1 4539.8

145 175

111- 146-

3 1 2 31 19.6 3120.9

123 159

176- 307- 377-

4 3 3 7740.8 7741.5

181 330 419

219- 293-

5 2 2 4848.5 4849.2

248 306

355- 364-

6 1 2 181 1.1 1810.6

358 376

Table 7

Table 8

N-Glycosylation

N-terminal heterogeneity

(main forms only)

N N+3 N+7 N+12 Asn117 Asn362

Peak! ++ (+) none none 16 26, 30

Peak 2 ++ ++ none none 11 , 12 26, 30

Peak 3 ++ ++ none none 4 26, 30 non-

Peak 4 ++ ++ (+) none 26, 30 glycosylated

non-

Peak 5 none +++ ++ (+) 26, 30 glycosylated

Peak 6 none (+) ++ +++ non-

26, 30 glycosylated

Table 9

Table 10 t Number glycosylation at Asn117 [%] non-glycosylated Asn-117 [%] j

2017376 66.3 33.7

2018656 62.5 37.5 t Number glycosylation at Asn-362 [%] non-glycosylated Asn-362 [%]

2017376 98.1 1.9

2018656 98.1 1.9 t Number fucosylation at Thr-61 [%] non-fucosylated Thr-61 [%]

2017376 100 None Detected

2018656 100 None Detected

Table 11

9 mg Vial

Ingredients Quantity Final Concentration

(per vial) (reconstituted in 10 mL)

Active Ingredient

9 mg 0.9 mg/mL desmoteplase

Inactive Ingredients

Glycine 36 mg 3.6 mg/mL Mannitol 96 mg 9.6 mg/mL

(Residual Water) <4.5 mg

Total Cake Weight <150 mg

Table 12 A

Table 12 B

Category Method Result Reference

> 90% DSPAal

SDS-PAGE Colloidal SDS-PAGE colloidal

% indivual impurities to be determined,

Blue reduced Blue stain report results above LOQ (%)

Purity

> 90% DSPAal

SDS-PAGE Colloidal SDS-PAGE colloidal

% indivual impurities to be determined,

Blue non-reduced Blue stain report results above LOQ (%)

Product- Size exclusion related Aggregates (SEC) ≤5% aggregates: (report % monomer) chromatography Impurities

Sterility Sterile Ph. Eur. 2.6.1

Ph. Eur. 2.6.14, Method

Endotoxins ≤ 3.0 EU/mg C

Process- related ≥ 10 μιτ) : < 6000/vial Ph. Eur. 2.9.19, Method

Sub-visible particles

Impurities ≥ 25 μιτι : < 600/vial I

Visible particles Free from visible particles Ph. Eur. 2.9.20

Residual moisture < 3.0% H₂0 Ph.Eur. 2.5.32

Desmoteplase cleaves the p- Nitroaniline from a

S-2288 2.5 mg/mL ± 0.37 mg/mL

chromogenic

Activity/ substrate, which is Potency detected at 405 nm

PCIJRP-HPLC 40-120 kU/mg -

Fibrinolytic activity

PCL 185 kU/mL ± 65 kU/mL and quantitation by absorbance at 405nm

Mannitol 96 mg/vial ± 9.6 mg/vial HPLC

Glycine 36 mg/vial ± 3.6 mg/vial HPLC

General tests pH 5.0 - 7.0 Ph.Eur. 2.2.3

Osmolality 290 mOsmol/kg ± 30 mOsmol/kg Ph.Eur. 2.2.35

Uniformity of dosage

Corresponds to Ph. Eur./ USP Ph.Eur. 2.9.40 units

1 001884

64

Table 13

Parameter Method Acceptance criterion

Appearance Clarity Clear liquid solution: Not more opalescent than reference suspension I

Appearance Colour Colourless liquid solution: Not more coloured than

reference solution B9

General pH value 5.0 - 7.0

General Glycine 40.0 mg/mL ± 4.0 mg/mL. Trend analysis

General Mannitol 15.0 mg/mL ± 1.5 mg/mL. Trend analysis

Content Protein content (A280) 3.82 mg/mL ± 0.38 mg mL. Trend analysis

Activity S-2288 3.82 mg/mL ± 0.57 mg/mL. Trend analysis

Activity Plate Clot Lysis 200 kU/mL -410 kU/mL. Trend analysis

Identity Western Blot To be determined. Evaluation of changes relative to to

> 90% DSPAal ; % individual impurities to be determined,

Purity/Identity SDS-PAGE, red. (Colloidal) Report results above LOQ (%); Main band corresponding to

Reference Standard; Evaluation of changes relative to to

> 90% DSPAal ; % individual impurities to be determined,

Purity Identity SDS-PAGE, non-red. (Colloidal) Report results above LOQ (%); Main band corresponding to

Reference Standard,; Evaluation of changes relative to to

Main band corresponding to reference standard, minor

Identity SDS-PAGE, red. (Silver)

bands below 55kD; Evaluation of changes relative to to

Content Protein content (RP-HPLC) 3.82 mg/mL ± 0.38 mg/mL. Trend analysis

Purity SE-HPLC < 5% aggregates; (report % monomer)

Corresponding to Reference Standard,

percentage distribution of peaks:

Peak no. Area-%

Peak 1 < 6

Peak 2 10- 22

Identity CIEX-HPLC

Peak 3 20-35

Peak 4 27-37

Peak 5 1 1- 21

Peak 6 < 9

Trend analysis

Identity Glycosylation / Fingerprint Corresponding to Reference Standard

Content Sialylation 1.5 - 3.4 mol/mol. Evaluation of changes relative to to

Identity Peptide Mapping Corresponding to reference standard

> 95% of N-terminus sequences derived from DSPAal ;

Identity N-Terminus

evaluation of changes relative to to

Process related

Bacterial Endotoxins < 3.0 EU/mg

impuritites

Process related

Bioburden < 1 CFU/mL

impuritites Table 14

Table 15

Sample Description TCIDgo titer Detection limit Reduction factor Code [log₁₀/mL] [logio] [log ]

M02 Virusstock 7.45 ± 0.23 4.57 —

1M21 Load 6.98 ± 026 3.57 —

1M22 0 min. ≤ 0.97 ± 0.00 0.97 >6.01 ±0.26

1M23 10 min. < 0.97 ±0.00 0.97 ≥ 6.01 +0.26

1M24 30 min. < 0.97 ±0.00 0.97 > 6.01 ±0.26

1M25 60 min. < 0.67 ± 0.00 0.67 ≥ 6.31 ± 0.26

1M26 180 min. ≤ 0.67 ± 0.00 0.67 ≥ 6.31 ± 0.26

1M27 Hold 6.80 ± 0.24 3.57 0.18 ±0.36

M02 Virusstock 7.45 ± 0.23 4.57 -

2M21 Load 6.92 ±0.29 3.57 -

2M22 0 min. < 0.97 + 0.00 0.97 > 5.95 + 0.29

2M23 10 min. < 0.97 ±0.00 0.97 > 5.95 ±0.29

2 24 30 min. 0.97 ± 0.00 0.97 ≥ 5.95 ±0.29

Sample Description TCID₅₀ titer Detection limit Reduction factor Code [log₁₀/mL] [logio] [logio]

2M25 60 min. < 0.67 ± 0.00 0.67 ≥ 6.25 ±0.29

2 26 80 min. ≤ 0.67 ± 0.00 0.67 > 6.25 ±0.29

2 27 Hold 6.62 ±0.25 3.57 0.30 ± 0.38

Table 16

Sample Description TCIDso titer Detection limit Reduction factor Code [log₁₀/mL] [logio] [log₁₀]

P02 Virusstock 7.62 ±0.17 4.57 —

1P21 Load 6.68 ±0.25 3.57 —

1P22 0 min. < 0.49 ± 0.00 0.49 > 6.19 ± 0.25

1P23 10min. < 0.49 ± 0.00 0.49 > 6.19 ±0.25

1P24 30 min. < 0.49 ± 0.00 0.49 ≥ 6.19 ±0.25

1P25 60 min. < 0.19 ±0.00 0.19 ≥ 6.49 ± 0.25

1P26 180 min. ≤ 0.19 ±0.00 0.19 ≥ 6.49 ± 0.25

1P27 Hold 6.74 ± 0.30 3.57 -0.06 ± 0.39

P02 Virusstock 7.62 ±0.17 4.57 —

2P21 Load 6.62 ± 0.27 3.57 —

2P22 0 min. ≤ 0.49 ± 0.00 0.49 ≥ 6.13 ±0.27

2P23 10 min. < 0.49 ± 0.00 0.49 ≥ 6.13 ±0.27

2P24 30 min. ≤ 0.49 ± 0.00 0.49 > 6.13 ±0.27

2P25 60 min. < 0.19 ±0.00 0.19 ≥ 6.43 ± 0.27

2P26 180 min. < 0.19 ±0.00 0.19 > 6.43 ±0.27

2P27 Hold 6.68 ± 0.25 3.57 -0.06 ± 0.37

Table 17

: Actual titer, which is calculated by Spearman-Karber is lower than detection limit, which the result of a probability assessment given by the Poisson distribution.

TEP2011/001884

69

Table 18

Reduction

TCIDso titer Detection limit Total volume

Sample code Description factor

[logiomL] llogiol [mLJ

[logio]

1M11 Load sample 6.21 3.57 145.29 -

1M12 Fraction 01 < 0.49 ± 0.00 0.49 36.18 ≥ 5.71 ±0.28

1MI3 Fraction 02 < 0.49 ± 0.00 0.49 36.28 ≥ 5.71 ±0.28

1M14 Fraction 03 < 0.49 ±0.00 0.49 36.20 > 5.71 ±0.28

1M15 Fraction 04 < 0.49 ± 0.00 0.49 36.63 > 5.71 ± 0.28

1M16 Flush fraction < 2.57 ± 0.00 2.57 17.61 > 3.63 ±0.28

1 17 Hold 6.10 ±0.17 3.57 145.29 0.11 ±0.33

2M11 Load sample 6.03 ± 0.33 3.57 142.74 -

2M12 Fraction 01 < 0.49 ± 0.00 0.49 35.69 > 5.54 ± 0.33

2M13 Fraction 02 < 0.49 ± 0.00 0.49 35.84 > 5.54 ±0.33

2M14 Fraction 03 < 0.49 ±0.00 0.49 34.87 > 5.54 ±0.33

2M15 Fraction 04 < 0.49 ± 0.00 0.49 36.34 > 5.54 ±0.33

2M16 Flush fraction < 2.57 ± 0.00 2.57 17.66 > 3.46 ±0.33

2M17 Hold 5.86 ±0.31 3.57 142.74 0.17 ±0.45

P T/EP2011/001884

70

Table 19

Reduction

TCIDso titer Detection limit Total volume

Sample code Description factor

[logio mL] [logiol [mL]

[logiol

1P11 Load sample 5.85 ± 0.32 3.57 145.49 -

1P12 Fraction 01 < 0.49 ± 0.00 0.49 36.25 > 5.36 ± 0.32

1P13 Fraction 02 < 0.49 ± 0.00 0.49 36.26 ≥ 5.36 ± 0.32

1P14 Fraction 03 < 0.49 ± 0.00 0.49 36.25 > 5.36 ± 0.32

1P15 Fraction 04 < 0.49 ± 0.00 0.49 36.73 > 5.36 ± 0.32

1P16 Flush fraction < 2.57 ± 0.00 2.57 17.68 > 3.28 + 0.32

1P17 Hold 5.91 ± 0.21 3.57 145.49 -0.06 ± 0.39

2P11 Load sample 5.79 ± 0.23 3.57 145.18 -

2P12 Fraction 01 < 0.49 ± 0.00 0.49 36.18 > 5.30 ± 0.23

2P13 Fraction 02 < 0.49 ± 0.00 0.49 36.19 > 5.30 ± 0.23

2P14 Fraction 03 < 0.49 ± 0.00 0.49 36.18 > 5.30 ± 0.23

2P15 Fraction 04 0.49 ± 0.00 0.49 36.63 > 5.30 ± 0.23

2P16 Flush fraction < 2.57 ± 0.00 2.57 17.67 > 3.22 ± 0.23

2P17 Hold 5.73 ± 0.27 3.57 145.18 0.06 ± 0.36

1884

71

Table 20

Reduction

TCn>5<j titer Detection limit Total volume

Sample code Description factor

[logjo mL] [logio] [mL]

[logio]

1E11 Load sample 6.62 ± 023 3.57 144.11 --

1E12 Fraction 01 6.36 ± 0.25 0.49 35.90 0.26 ± 0.34

1E13 Fraction 02 6.66 ± 0.25 0.49 36.00 -0.03 ± 0.34

1E14 Fraction 03 6.48 ± 0.24 0.49 36.00 0.15 ± 0.33

1E15 Fraction 04 6.60 ± 0.25 0.49 36.21 0.03 ± 0.34

1E16 Flush fraction 6.16 ± 0.25 2.57 17.48 0.46 ± 0.34

IE17 Hold 6.68 ± 0.30 3.57 144.11 -0.06 ± 0.38

2E1 1 Load sample 6.57 ± 0.32 3.57 146.09 --

2E12 Fraction 01 6.24 ± 0.33 0.49 36.41 0.32 ± 0.46

2E13 Fraction 02 6.42 ± 0.33 0.49 36.42 0.15 ± 0.46

2E14 Fraction 03 6.30 + 0.21 0.49 36.42 0.26 ± 0.39

2E15 Fraction 04 6.60 ± 0.26 0.49 36.48 -0.03 ± 0.41

2E16 Flush fraction 6.04 ± 0.30 2.57 17.80 0.52 ± 0.44

2E17 Hold 6.62 ± 0.25 3.57 146.09 -0.06 ± 0.41

Table 21

Reduction

TCIDjo titer Detection limit Total volume

Sample code Description factor

[log,o mLJ [logiol [mL]

[logiol

1PP1 1 Load sample 7.03 ± 0.26 4.57 140.78 -

1PP12 Fraction 01 0.02 ± 0.00 0.02 36.16 7.01 ± 0.26

1PP13 Fraction 02 < 2.57 ± 0.00 2.57 36.23 > 4.46 ± 0.26

1PP14 Fraction 03 < 2.57 ± 0.00 2.57 36.26 > 4.46 ± 0.26

1PP15 Fraction 04 < 2.57 ± 0.00 2.57 32.13 > 4.46 ± 0.26

1PP16 Flush fraction < 2.57 ± 0.00 2.57 17.29 > 4.46 ± 0.26

1PP17 Hold 7.33 ± 0.32 4.57 140.78 -0.30 ± 0.41

2PP11 Load sample 7.09 + 0.20 4.57 144.80 --

2PP12 Fraction 01 0.02 ± 0.00 0.02 36.10 7.07 ± 0.20

2PP13 Fraction 02 < 2.57 ± 0.00 2.57 36.10 > 4.52 ± 0.20

2PP14 Fraction 03 < 2.57 ± 0.00 2.57 36.10 > 4.52 ± 0.20

2PP15 Fraction 04 < 2.57 ± 0.00 2.57 36.50 > 4.52 ± 0.20

2PP16 Flush fraction < 2.57 ± 0.00 2.57 17.60 > 4.52 ± 0.20

2PP17 Hold 7.21 ± 0.27 4.57 144.80 -0.12 ± 0.34

Table 22

Description DNA-Exp. 1 DNA-Exp. 2 HCP-Exp. 1 HCP-Exp. 2

[pg/mL] [pg/mL] [ng/mL] [ng/mL]

MAQ-Load 11400 20200 2320 1753

MAQ-filtrate < 6200 < 6200 231 98,7

Overall reduction 2,54 12,35

[lOQiol

Table 23

HSV BPV

Virus reduction Virus reduction

Total virus load Total virus load

Fraction factor factor

[log] [log]

[log] [log]

Load 7.9 n a 8.0 n/a

Flowthrough /

7.3 n/a 8.1 n/a Wash 1 / Wash 2

Eluate 6.7 1.2 5.7 2.3

Table 24

Sample Description TCID₅₀ titer Detection Total Total Reduction Code [log₁₀/mL] limit volume virus load factor

[log ] [mL] [log™] [logio]

M03 Virusstock 7.86 ± 0.21 4.57 ~ —

1M31 Load 6.86 ± 0.26 3.57 33.00 8.38 ± 0.26

1 M32 Pre-filtrate 6.27 ± 0.28 3.57 31.30 7.76 ± 0.28 -

Nanofiltrate < 0.67 ± < 1.85 ±

1 M33 0.67 15.08 ≥ 5.60 ±

Fraction 1 0.00 0.00 0.28

Nanofiltrate < 0.67 ± < 1.85 ±

0.67 15.14 > 5.60 ±

1 M34

Fraction 2 0.00 0.00 0.28

1M35 Hold 6.86 ± 0.25 3.57 33.00 8.38 + 0.25 0.00 ± 0.36

M03 Virusstock 7.86 ± 0.21 4.57 - - ~

2M31 Load 6.51 ± 0.21 3.57 33.00 8.02 ± 0.21

2M32 Pre-filtrate 6.15 ± 0.22 3.57 31.30 7.64 + 0.22 —

Nanofiltrate < 0.67 ±

2M33 0.67 15.05 < 1.85 ± ≥ 5.48 ±

Fraction 1 0.00 0.00 0.22

Nanofiltrate ≤ 0.67 ±

2M34 ≤ 1.85 ±

0.67 15.09 > 5.48 ± Fraction 2 0.00 0.00 0.22

2 35 Hold 6.03 ± 0.25 3.57 33.00 7.55 ± 0.25 0.10 ± 0.33

Table 25

Sample Description TCIDso titer Detection Total Total Reduction Code [log₁₀/mL] limit volume virus load factor

[logic] [mL] [logio] [logio]

P03 Virusstock 7.68 ± 0.24 4.57 -

1 P31 Load 6.57 ± 0.36 3.57 33.00 8.08 + 0.36 -

1 P32 Pre-filtrate 6.57 ± 0.29 3.57 31.30 8.06 ± 0.29

Nanofiitrate ≤0.19 ± ≤ 1.37 ± ≥6.37 +

1 P33 0.19 15.04

Fraction 1 0.00 0.00 0.29

Nanofiitrate < 0.19 ± < 1.37 ± > 6.37 ±

1 34 0.19 15.07

Fraction 2 0.00 0.00 0.29

1 P35 Hold 6.57 ± 0.27 3.57 33.00 8.08 ± 0.27 0.00 ± 0.45

P03 Virusstock 7.68 ± 0.24 4.57 - -

2P31 Load 6.51 ± 0.30 3.57 33.00 8.02 ± 0.30 -

2P32 Pre-filtrate 6.51 ± 0.28 3.57 31.30 8.00 ± 0.28

Nanofiitrate ≤0.19 ±

2P33 0.19 ≤ 1.37 ±

15.05 > 6.31 ± Fraction 1 0.00 0.00 0.28

Nanofiitrate ≤0.19 ±

2P34 0.19 ≤ 1.37 ±

15.10 > 6.31 ± Fraction 2 0.00 0.00 0.28

2P35 Hold 6.80 ± 0.32 3.57 33.00 8.32 ± 0.32 -0.30 ± 0.43

Table 26

Sample Description TCID₅₀ titer Detection Total Total Reduction Code [log_1c mL] limit volume virus load factor

[•og₁₀l tmL] [log ₀] [log ]

E03 Virusstock 7.21 ± 0.29 4.57 — ~ —

1 E31 Load 6.33 ± 0.27 3.57 33.00 7.85 ± 0.27 —

1 E32 Pre-filtrate 6.45 ± 0.25 3.57 31.30 7.94 ± 0.25 —

Nanofiltrate

1E33 ≤0.19 ± 0.19 ≤ 1.37 ±

15.08 > 6.25 ± Fraction 1 0.00 0.00 0.25

Nanofiltrate < 0.19 ± 0.19 ≤ 1.37 ±

1E34 15.03 ≥ 6.25 ±

Fraction 2 0.00 0.00 0.25

1 E35 Hold 6.33 ± 0.30 3.57 33.00 7.85 ± 0.30 0.00 ± 0.40

E03 Virusstock 7.21 ± 0.29 4.57 ~ - -

2E31 Load 6.45 ± 0.29 3.57 33.00 7.96 ± 0.29 ~

2E32 Pre-filtrate 6.27 ± 0.29 3.57 31.30 7.76 ± 0.29 -

Nanofiltrate < 0.19 ±

0.19 ≤ 1.37 ±

2E33 15.03 > 6.07 ±

Fraction 1 0.00 0.00 0.29

Nanofiltrate ≤0.19 ±

2E34 0.19 ≤ 1.37 ±

15.10 ≥ 6.07 ± Fraction 2 0.00 0.00 0.29

2E35 Hold 6.51 ± 0.34 3.57 33.00 8.02 + 0.34 -0.06 ± 0.44

Table 27

Sample Description TCID₅₀ titer Detection Total Total Reduction Code [log₁₀/mL] limit volume virus load factor

[logio] [mL] [logio] [logio]

PP03 Virusstock 9.22 ± 0.28 5.57 — — -

1 PP31 Load 8.40 ± 0.30 4.57 33.00 9.92 ± 0.30 —

1 PP32 P re-filtrate 8.40 ± 0.23 4.57 30.80 9.89 ± 0.23

Nanofiltrate

1 PP33 2.54 ± 0.20 2.17 15.08 3.72 ± 0.20 5.86 ± 0.31

Fraction 1

Nanofiltrate

1 PP34 2.78 ± 0.27 2.17 15.06 3.96 ± 0.27 5.62 ± 0.35

Fraction 2

1PP35 Hold 8.22 ± 0.29 4.57 33.00 9.74 ± 0.29 0.18 ± 0.42

PP03 Virusstock 9.22 ± 0.28 5.57 — —

2PP31 Load 8.16 ± 0.30 4.57 33.00 9.68 ± 0.30 -

2PP32 Pre-filtrate 8.16 ± 0.26 4.57 31.30 9.66 ± 0.26 -

Nanofiltrate

2PP33 2.66 ± 0.23 2.17 15.08 3.84 ± 0.23 5.50 ± 0.35

Fraction 1

Nanofiltrate

2PP34 3.08 ± 0.33 2.17 15.12 4.26 ± 0.33 5.08 ± 0.42

Fraction 2

2PP35 Hold 8.04 ± 0.26 4.57 33.00 9.56 ± 0.26 0.12 ± 0.40

Table 28

MuLV PRV EMCV PPV study K2/R01/07 (2007)

120

MAQ filtration

Fraction 1 > 5.54 ± 0.33 ≥ 5.30 ±0.23 7.01 ±0.26 Fraction 2 > 5.54 ±0.33 > 5.30 ± 0.23 > 4.46 ± 0.26 Fraction 3 > 5.54 ±0.33 ≥ 5.30 ± 0.23 no reduction of 4.46 ± 0.26 virus obtained

Fraction 4 5.54 ±0.33 5.30 ± 0.23 4.46 ±0.26 Post-flush 3.46 ± 0.33 3.22 ± 0.23 > 4.46 ± 0.26

Overall MAQ*' > 4.34 ±0.33 4.10 ±0.23 ≥ 4.S2 ± 0.26

study 2/05.136 (2008)

M30

low pH treatment

60 min. ≥ 6-25 + 0.29 > 6.43 ± 0.27 4.02 ± 0.36 n.t

180 min. 6.25 ± 0.29 > 6.43 ± 0.27 > 5.95 ±0.34 n.t. 360 min. n.t n.t > 5.96 ±0.23 n.t. 900 min. n.t. n.t 5.96 ±0.23 n.t.

130

nanofiltration

Fraction 1 5.48 ±0.22 > 6.31 ±0.28 6.07 ±0.29 5.50 ±0.35 Fraction 2 5.48 ± 0.22 ≥ 6.31 ±0.28 > 6.07 ± 0.29 5.08 ± 0.42

Overall reduction*⁵ 15.23 > 16.06 ll.Sl*³ >8.92

Theoretical viral load

8.24 n.a. n.a. n.a. per dose

Safety margin >6.99 n.a. n.a. n.a.

Claims

Patent claims

A method for the manufacture of rDSPA alphal from a medium containing DSPA alphal the method comprises the following steps:

(a) applying the medium in a first flow direction to a cation exchanger under loading conditions which result in a binding of rDSPA alphal ;

(b) washing the cation exchanger, wherein said washing step comprises at least a part having a second flow direction, which is contrary to the first flow direction;

(c) eluting the bound rDSPA alphal from the cation exchanger;

(d) applying the rDSPA alphal containing eluent from step (c) to a hydrophobic interaction chromatography matrix under loading conditions which result in binding of rDSPA alphal ;

(e) optionally, washing the hydrophobic interaction matrix;

(f) eluting the bound rDSPA alphal ;

(g) applying the rDSPA alphal containing eluent from step (f) to an affinity chromatography matrix under loading conditions which result in a binding of rDSPA alphal ;

(h) optionally, washing the affinity chromatography matrix;

(i) eluting the bound rDSPA alphal from the affinity chromatography matrix.

The method of claim 1 , wherein the first flow direction is downwards and the second flow direction is upwards.

The method of claim 1 or 2, wherein the elution of step (i) results in substantially pure rDSPA alphal .

The method of one of the above claims, wherein the cation exchange matrix of step (a) is selected from silica gel particles, cross-linked agarose, cross-linked polymethacrylate polymers, derivatized with carboxyl or carboxyalkyl groups, in particular cross linked polymethylmethacrylate (P MA), covalently bound to polyethyleneimine.

5. The method according to one of the above claims, wherein said method further comprises one or more of the steps:

- adjusting the medium before loading according to step (a) to a pH of 4 to 6, preferably to about 5.5, preferably with a buffer containing acetic acid and NaCI;

- washing the matrix according to step (b) with a buffer of pH 4 to 7, preferably with a buffer containing 30 to 80 mM acetic or phosphate buffer, pH of about

5.0;

- eluting the rDSPA alphal according to step (c) with a salt step gradient of 100 to 500 mM NaCI, preferably about 200 mM, in 30 to 80 mM Na₂HP0₄, preferably about 50 mM, at pH of about 7.5; or a salt of equivalent ionic strength.

6. The method of one of the above claims, wherein said method further comprises one or more of the following steps:

- diluting the eluent from step (c) with a loading buffer with about 30 to 80 mM, preferably about 50 mM, sodium acetate, 1.00 to 1.5 M, preferably about 1.25

M, NaCI, about pH 3 to 6, preferably about pH 5.5, and loading the hydrophobic interaction matrix;

- washing the hydrophobic interaction matrix according to step (e) with a series of different washing buffers, preferably starting with a buffer with 30 to 80 mM, preferably about 50 mM, sodium acetate, 1.00 to 1.5 M, preferably about 1.25

M, sodium chloride, about pH 3 to 6, preferably about pH 5.5, followed by a washing step with 10 to 50 mM, preferably about 20 mM, HCI, pH 1-4, preferably pH 2,5, containing an organic solvent, preferably ethanol;

- eluting the hydrophobic interaction matrix according to step (f) with an elution buffer containing 10 to 50 mM, preferably about 20 mM, HCI, and 10 to 40 %, preferably about 30 %, in particular 29.5 %, ethanol.

7. The method according to claim 6, wherein said washing is performed with an increased concentration of organic solvents and said washing buffers contain 10 to 30 %, preferably from 15 to 17%, organic solvent.

8. The method according to one of the above claims, wherein said method further comprises one or more of the following steps: - diluting the eluent from step (f) before loading to the affinity chromatography matrix according to step (g) with a loading buffer with 10 to 50 mM, preferably about 20 mM HCI;

- washing the affinity chromatography matrix according to step (h) with 10 to 50 mM, preferably about 20 mM HCI;

- eluting the affinity chromatography matrix to step (i) with a buffer containing 100 to 500 mM, preferably 200 mM, glycine and 0.1 to 0.5 M, preferably 0.3, M, NaCI, pH 5 to 7, preferably about 6. 9. The method of one of the above claims, wherein the eluent of step (f) is incubated in an organic solvent with a pH of 3 or below for at least 3 hours.

10. The method of claim 9 wherein the eluent of step (f) is incubated for at least 10, preferably for at least 18 hours.

1 1. The method of claim 9 or 10, wherein the incubation of the eluent of step (f) is performed at a temperature between 18 and 24°C.

12. The method according to one of the above claims wherein the eluent of step (c) before being applied in step (d) is applied to an anion exchange matrix under loading conditions that result in a binding of virus, host cell proteins and/or DNA to the anion exchange matrix.

13. The method according to claim 12 wherein the anion exchange matrix contains quaternary ammonium groups from type I or type II

0

(type I)

© 0

— (CH₃)₂ X (type II)

C₂H₅OH and X is an anion selected from the group consisting of hydroxyl, chloride, sulfate, bromide, iodide, fluoride, sulfide, hydrogensulfate, hydrogensulfide, phosphate, diphosphate, monophosphate, carbonate, hydrogencarbonate, citrate, tartrate, or phthalate.

14. The method according to claim 12 or 13 wherein the anion exchange matrix contains diethylaminoethyl groups (DEAE) groups.

15. The method according to any of the claims 12 to 14 wherein the anion exchange matrix is applied as a membrane adsorber.

16. The method according to claim 15, wherein the membrane adsorber step is performed as a tangential flow filter (TFF) process. 17. The method according to any of the claims 12 to 16, wherein said method further comprises one or more of the following steps:

- diluting the eluent from step (c) before loading to the anion exchange chromatography matrix with WFI (water for injection), preferably in a ratio of eluent to WFI from 1 :1— 1 :2, preferably about 1 :1.25;

- equilibrating the anion chromatography matrix before loading to a conductivity of a maximum of about 30 mS/cm, preferably to a conductivity between 3 and

15 mS/cm, most preferred between 8 and 13.5 mS/cm;

equilibrating the anion exchange chromatography matrix with an equilibration buffer containing 10 to 50 mM, preferably about 20 mM, Na phosphate buffer and 50 to 120 mM, preferably about 80 mM, NaCI, pH 6 to 8, preferably about

7.3.

18. The method according to any of the above claims wherein the eluent of step (i) is applied to a step of nanofiltration.

19. The method according to claim 18, wherein the matrix for virus filtration is selected from the group consisting of cross-linked organic polymers, ceramics or inorganic compounds, in particular polysulfone (PS), sulfonated polysulfone (SPS), polyimide (PI) or polyamide.

20. The rDSPA alphal obtainable by any of the claims 1 to 19.

21. The rDSPA alphal of claim 20, which is substantially pure. 22. The rDSPA alphal exhibiting a CIEX-HPLC profile with at least 2, preferred at least 3, most preferred 6 peaks, whereas each of said peaks exhibits a retention time frame from approximately 5 to 25 min, preferably approximately 7 to 17 min, most preferred 8 to 16 min, and the CIEX-HPLC is performed with a strong cation exchanger, preferably with a cation exchanger with the functional group -(CH₂)₃S0₃ ^" .

23. The rDSPA alphal according to claim 22, wherein the area under the at least two peaks represent at least about 40 % of the total area under all peaks within the retention time window.

24. The rDSPA alphal according to claim 22 or 23 exhibiting a CIEX-HPLC profile as depicted in Fig. 10.

25. The pharmaceutical composition comprising rDSPA alphal according to one of the claims 20 to 24.

26. The pharmaceutical composition according to claim 25 further comprising a pharmaceutically acceptable excipient.