WO2011131787A2

WO2011131787A2 - Antibody preparations

Info

Publication number: WO2011131787A2
Application number: PCT/EP2011/056487
Authority: WO
Inventors: Wolfgang Möller; Dieter Rudnick; Oliver Maneg; Michael Rodemer; Matthias Germer; Veit Braun
Original assignee: Biotest Ag
Priority date: 2010-04-22
Filing date: 2011-04-21
Publication date: 2011-10-27
Also published as: US20180319871A1; ES2553385T3; IL222373A; RU2017111820A; JP6283517B2; JP2016222674A; HK1180601A1; AU2011244241A1; CA2796263C; IL222374A0; WO2011131786A3; JP6118248B2; RU2016140455A; US20150064170A1; EP2560691A2; ZA201208540B; EP2560682A2; JP2013530950A; MX2012012263A; ZA201208541B

Abstract

Provided is an antibody preparation suitable for intravenous administration in humans comprising IgG, IgA and at least 5% IgM antibodies by weight of the total amount of antibodies, wherein the preparation is prepared from human plasma, wherein the antibody preparation has specific complement activating activity and wherein in an in vitro assay with human serum suitable to determine the ability of the antibody preparation to activate complement unspecifically the antibody preparation generates substantially no C5a and/or substantially no C3a. Further provided are medical uses of the antibody preparation.

Description

ANTIBODY PREPARATIONS

FIELD OF THE INVENTION

The present invention relates to an antibody (immunoglobulin) preparation comprising IgM which has specific complement activating activity but low unspecific complement activation capacity. The present invention also relates to the use of the antibody preparation in medicine.

BACKGROUND OF THE INVENTION

Immunoglobulin compositions prepared from human plasma and suitable for intravenous administration are known in the art and for several decades have played an important role in the treatment of a wide range of diseases. Immunoglobulins are used, for example, for the treatment of infections in humans and can be assigned to various classes with various biochemical and physiological properties. Immunoglobulin G participates in defending against viral antigens, whereas IgM is predominantly active in antibacterial and antitoxin immune responses.

The immunoglobulin solutions comprise IgG, IgA and IgM in various percentages, with different preparations having different treatment applications, e.g. preparations with a higher percentage of IgM are used in the prophylaxis or treatment of bacterial infections.

The immunoglobulin solutions are usually prepared from fractions of blood plasma or serum, e.g. Cohn fractions. These fractions are then subjected to a number of purification steps to remove contaminants such as viruses, denatured proteins, proteases and lipids.

Human plasma for fractionation is collected from thousands of donors and may contain pathogen viruses despite thorough testing of the source plasma. Therefore process steps to inactivate or remove viruses are essential in order to achieve safe products for use in medicine. Several techniques for virus inactivation/removal are known in the art, e.g. chemical treatments, irradiation with UVC light or nanometer filtration, which are performed in order to ensure overall virus safety. The virus removal or inactivation capacity of the process steps is validated using laboratory scale models of the production process and for each step a removal or inactivation factor is determined. An increase of the inactivation / removal factor adds additional viral safety to the pharmaceutical product. Today guidelines from regulatory authorities require at least two effective steps for enveloped and non-enveloped viruses in the manufacture of plasma-derived pharmaceuticals. Although several methods, such as solvent/detergent treatment, octanoic acid treatment, nanometer filtration and heat treatment, are effective to inactivate or remove enveloped viruses there are only a few methods known to inactivate or remove non-enveloped viruses, for example Parvo viruses. These non-enveloped viruses are mostly very small, usually passing through nanometer filters with pore sizes above 20 nm. This pore size is too small for IgM molecules having a diameter up to 30 nm. Non enveloped viruses are effectively inactivated by chemicals like β-propiolactone which, however, also leads to a modified immunoglobulin with impaired functions. Another effective treatment is UVC-irradiation (EP 1842561 , CAF-DCF). However, known solvent/detergent treatments, octanoic acid treatment and mild heat treatment have no substantial effect on non- enveloped viruses.

As mentioned above, in addition to viruses which are potentially present it is also necessary to remove other contaminants like lipids, proteases, protein aggregates, and denatured immunoglobulins. The removal of all these contaminants is essential (1) to ensure the product complies with bio-safety guidelines regarding viral contamination, (2) in order for the product to be tolerated by the patient after intravenous administration, (3) to allow the product to be stable during long-term storage (any residual proteolytic activity might lead to degradation of the product over long-term storage, e.g. 2 years), and (4) to generate the desired compound mixture / pharmaceutical composition.

At the same time, however, it is essential that the purification steps to remove the contaminants do not interfere with the immunoglobulin molecules, so that as far as possible these retain their normal biological activity and are retained at high yield in solution. This balance is difficult to achieve since many known purification steps can also have a negative impact on the activity of the immunoglobulins, and in particular on IgM; for example extended irradiation times with UVC can reduce the yield of native and active IgM obtained in the final immunoglobulin solution. Not only does this lead to a reduction in efficacy of the final immunoglobulin solution but it can also cause the solution to be less well tolerated in vivo.

Aggregates and denatured immunoglobulins, the amount of which can be increased by certain purification steps, especially are a potential risk for the patients because they have a high capacity to activate complement unspecifically, leading to severe side effects in patients receiving these denatured immunoglobulins. Unspecific complement activation refers to the initiation of the complement cascade in the absence of specific antibody-antigen complexes. Unspecific complement activation is strictly to be avoided since it may cause undesirable side effects such as hypotension, flushing, headache, fever, chills, nausea, vomiting, muscle pain, dyspnoea and tachycardia. Specific complement activation, on the other hand, is desirable and it occurs only after the immunoglobulins have bound to their specific antigens.

Unspecific complement activation is measured as the so called anticomplementary activity (ACA) by a standardized test described in the European Pharmacopoeia.

The role of the complement system in the immune defense of pathogens is well known. The complement system consists of about 20 proteins, which are activated sequentially. The classical complement pathway typically requires a specific antigen antibody complex for activation, whereas the alternative pathway can be activated by antigens without the presence of antibodies. The classical and the alternative pathway of complement activation all generate a protease C3- convertase. The C3-convertase cleaves and activates component C3, creating C3a and C3b, and causing a cascade of further cleavage and activation events over C5 convertase to C5a and C5b. C5b initiates the membrane attack pathway, which results in the membrane attack complex, consisting of C5b, C6, C7, C8, and polymeric C9. This is the cytolytic endproduct of the complement cascade which forms a transmembrane channel, which causes osmotic lysis of the target cells like bacteria.

Complement activation additionally results in the formation of anaphylatoxins, including the biologically active protein C5a. This anaphylatoxin is a potent chemotactic agent for immune and inflammatory cells and induces cell activation and causing the release of histamine from mast cells. In situations of excessive or uncontrolled and/or unspecific complement activation, the overproduction of C5a can cause deleterious effects to patients. C5a is an effective leukocyte chcmoattractanl, causing the accumulation of white blood cells, especially neutrophil granulocytes, at sites of complement activation. C5a activates white blood cells and is a powerful inflammatory mediator. Whereas these functions are beneficial during specific antibody-antigen complex reactions all unspecific generation of C5a has to be avoided due to the potential side effects.

Unspecific complement activation is a particular issue for IgM immunoglobulin preparations (i.e. those comprising at least 5% IgM) as in contrast to IgG preparations IgM antibodies easily aggregate in solution. IgM preparations are difficult to stabilize especially if they are enriched compared to plasma concentrations and stored in liquid solution. It is also known that IgM is a vigorous activator of complement; a single molecule bound to an antigen can activate complement. This is in contrast to IgG, where two or more molecules of IgG must be bound to an antigen in close association with each other to activate complement.

Still further, the main indications treated by IgM containing immunoglobulin preparations are bacterial infections and sepsis. As these patients are already suffering from hypotension an additional unwanted generation of unspecific complement activation and C5a would lead to a clinical worsening of the patient^'s condition. Accordingly, IgM preparations have been described as being difficult to prepare for intravenous application.

There are several methods described in the art for the production of IgM containing immunoglobulin preparations from human plasma.

The initial purification of human IgM solutions has been carried out by classical Cohn plasma fractionation methods or its well known modifications (e.g. Cohn/Oncley, Kistler/Nitschmann). Using cold ethanol precipitation processes the IgM fraction is recovered in fraction III or fraction I/I 11 (also called B or B+I). Starting from fraction III o I/I II methods have been described for purification of protein solutions enriched in IgM. EPOO 13901 describes a purification method starting from fraction III including steps using octanoic acid, β-Propiolactone treatment and an adsor^ption ste using an anionic exchange resin. This method is used to produce Pentaglobin^® - to date the only commercially available intravenous IgM product, β-propiolactone is a well known chemical used in sterilization steps in order to inactivate viruses which are potentially present. As β- propiolactone is a very reactive substance which causes the chemical modification of proteins there is also substantial loss of the anti-viral and anti-bacterial activities of the immunoglobulins. On the other hand this chemical modification results in an reduced anticomplementary activity compared to an chemically unmodified immunoglobulin. EP0352500 describes the preparation of an IgM concentrate for intravenous application with a reduced anti-complementary activity by using anionic exchange chromatography, β-Propiolactone, UVC light irradiation and an incubation step at increased temperature (40 °C to 60 °C). The preparation produced by this method was stable in liquid solution for a limited time due to the chemical modification. The IgM concentration was above 50 % from the total immunoglobulin content.

The preparation of protein solutions enriched in IgM without chemical modification by β- propiolactone has been described in EP0413187 (Biotest) and EP0413188 (Biotest). These methods involve subjecting a suitable protein solution to octanoic acid treatment and anionic exchange chromatography, starting from Cohn fraction III or Will. In patent EP0413187 (Biotest) the octanoic acid treatment is carried out by stirring for 15 min, in order to remove lipids being present in Cohn fraction III.

The preparation according to EP0413187 had a low anticomplementary activity, between 0,6 and 0,8 CH50/mg protein, but had to be stabilized and virus inactivated by β-propiolactone. Low anticomplementary activity is considered to be < 1 CH50/mg protein according to EP monograph for immunoglobulins.

EP0413188B 1 (Biotest) describes the preparation of an IgM-enriched preparation for intravenous administration by using an anion exchange chromatography in order to reduce the anticomplementary activity. Additionally a heat treatment at pH 4 - 4.5 at 40 to 60°C, preferably between 50 and 54°C, was described to reduce the anticomplementary activity. This preparation had to be lyophilized to ensure stability of the preparation for several months. Long term stability as a liquid solution could not be shown. M.Wickerhauser et al. "Large Scale Preparation of Macroglobulin", Vox Sang 23, 1 19-125 (1972) showed that IgM preparations isolated by PEG precipitation had high anticomplementary activity (AC A) by a standard complement fixation test and this AC A activity was reduced 10 fold by incubating the IgM preparation at pH 4.0 at 37°C for 8 hours followed by readjustment to neutral pH. It was not demonstrated if this 10 fold reduction is sufficient to ensure intranenous tolerability. The authors did not assess the specific complement activating potential of their IgM concentrate, nor did they assess safety in any animal or human model.

Another method describes the use of mild-heat treatment of IgM preparations at 40 to 62°C, preferably 45 to 55°C, at pH 4.0 to 5.0 (EP 0450412, Miles) to reduce the unspecific complement activation. In this patent application octanoic acid is added to a Cohn fraction III suspension in order to remove prekallikrein activator and lipoproteins by centrifugation. Nevertheless this mild heat treatment led to partial loss of antigenic determinants of IgM. This may increase the risk of generating neo-antigens leading to a increased immunogenicity in humans or the loss of activity.

The preparation of an IgM containing protein solution for intravenous application by using a protease treatment (e.g. with pepsin) after an octanoic acid precipitation step has been described in EP0835880 (US 6136312, ZLB). Protease treatment leads to partial fragmentation of the immunoglobulin molecule impairing the full functional activity of the Fab and Fc parts. Therefore protease-treated immunoglobulins cannot be regarded as unmodified. Also this preparation method leads to about 5 % fragments with a molecular weight of <100kD.

The described methods to carry out the octanoic acid treatment (EP0413 187 and EP0835880) have the drawback that the octanoic acid treatment is not effective with respect to removal and inactivation of non-enveloped viruses, and does not remove substantially all proteolytic activity. In EP 0345543 (Bayer, Miles) a highly concentrated IgM preparation with at least 33% IgM for therapeutic use is disclosed, the preparation being substantially free of isoagglutinin titres. In this patent application an octanoic acid precipitation is carried out by adding the octanoic acid and the isoagglutinins are removed by Synsorb affinity chromatography. The final preparation had to be freeze dried. Altogether the production of an IgM containing preparation with low anticomplementary activity is possible if the immunoglobulins are chemically or cnzymatically modified and/or further purified by chromatography and/or subjected to a mild heat treatment. However, these methods have their drawbacks in the lack of virus removal/virus inactivation (and therefore virus safety), reduction in the amount of immunoglobulin molecules in native form and/or residual anticomplementary activity. As such, there is still a need to provide improved IgM containing immunoglobulin preparations suitable for intravenous administration in humans.

SUMMARY OF THE INVENTION

In a first aspect the present invention provides an antibody preparation suitable for intravenous administration in humans comprising IgG, IgA and at least 5% IgM antibodies by weight of the total amount of antibodies, wherein the preparation is prepared from human plasma, wherein the antibody preparation has specific complement activating activity and wherein in an in vitro assay with human serum suitable to determine the ability of the antibody preparation to activate complement unspecifically the antibody preparation generates substantially no C5a and/or substantially no C3a.

The present applicants have surprisingly found that the production of an IgM antibody preparation from human serum is possible which has specific complement activating activity and substantially no unspecific complement activity. This product is advantageous as it maintains product efficacy while reducing unwanted side-effects such as hypotension, associated with unspecific complement activation after intravenous administration.

A further aspect of the present invention provides a method of producing the antibody preparation of the present invention from human plasma comprising the steps of:

(a) preparing from the human plasma a plasma fraction as a solution containing immunoglobulins;

(b) mixing a C₇ to CQ carboxylic acid with the solution and treating the mixed solution with a vibrating agitator to precipitate contaminating proteins:

(c) separating the precipitated proteins from the solution to yield the IgM containing immunoglobulin composition; (d) incubating the IgM containing immunoglobulin composition at between pH3.5 and pH 4.5 to form an incubated solution;

(e) irradiating the incubated solution with UVC to form a UVC irradiated solution; and

(f) filtering the UVC irradiated solution under sterile conditions to form the antibody preparation suitable for intravenous administration in humans.

The present applicants have surprisingly found that the use of a vibrating agitator during the step where the immunoglobulin solution is mixed with the carboxylic acid is extremely advantageous. This method step provides a more efficient removal of unwanted proteins (including proteases) and produces an intermediate product which is better suited to downstream processing steps utilised to produce an immunoglobulin medicament; the intermediate product allows these downstream processing steps to be more efficient. Accordingly, the downstream processing steps can be less harsh, helping to achieve the antibody preparation of the present invention which is capable of specific complement activation and substantially no unspecific complement activation.

In particular, the IgM immunoglobulin containing composition obtained from step (c) can be combined with further treatment steps, such as treatment with mild acid conditions and treatment with UVC irradiation, to produce an IgM containing immunoglobulin product or antibody preparation which is suitable for intravenous administration and which has the following advantageous properties: having low anti-complementary activity; retaining a high level of native and active IgM; and being virus safe and thus suitable for intravenous administration in humans. The level of virus safety achieved with the methods described herein has not previously been obtaininable. Additional advantages are having low proteolytic activity (and therefore being stable during long term storage) and being chemically unmodified.

Still further, the present invention provides an antibody preparation of the present invention for use in medicine. In one embodiment the antibody preparation is for use in the treatment of immunological disorders and bacterial infections.

A further aspect of the present invention provides a method of treatment comprising administering the antibody preparation of the present invention to a patient. The present invention will now be described in further detail by way of example only, with reference to the accompanying figures.

FIGURE 1 provides an overview of the steps that can be utilised to form an antibody preparation suitable for intravenous admininstration according to the present invention. The octanoic acid treatment step employing a vibromixer device, the pH4 treatment and the UVC treatment are highlighted The starting material is generated from a standard cold ethanol precipitation process of human plasma.

FIGURE 2 provides a graph showing time dependent mean C5a concentrations found in human serum after incubation with IgM preparations.

FIGURE 3 provides a graph showing time dependent mean C3a concentrations found in human serum after incubation with IgM preparations.

DETAILED DESCRIPTION OF THE INVENTION

Antibody preparation

As described above, the present invention provides an antibody preparation suitable for intravenous administration in humans comprising IgG, IgA and at least 5% IgM antibodies by weight of the total amount of antibodies, wherein the preparation is prepared from human plasma, wherein the antibody preparation has specific complement activating activity, and wherein in an in vitro assay with human serum suitable to determine the ability of the antibody preparation to activate complement unspecifically the antibody preparation generates substantially no C5a and/or substantially no C3a.

The antibody preparation of the present invention comprises human plasma proteins of which at least 90%, preferably at least 95% is made up of immunoglobulins (polyclonal antibodies). In particular the preparation comprises the immunoglobulins IgG, IgA and IgM wherein at least 5% of the immunoglobulins are IgM. The amount of IgG, IgA and IgM immunoglobulins can be determined by nephelometry or by immunoprecipitation according to Ph. Eur. 2.7. 1 . More preferably the antibody preparation comprises at least 10% Ig and most preferably at least 15% IgM. In relation to IgG and IgA, preferably the antibody preparation comprises more than 5% IgA and/or more than 40% IgG. All percentages are percentage of total amount of antibodies (for example, g of IgM / (g of IgG + g of IgA + g of IgM) x 100).

Methods for determining that the antibody preparation has specific complement activating activity (i.e. the ability to activate the complement cascade in the presence of antigen) through assessing the functional activity of the Fc part of the immunoglobulin molecule are known in the art. In particular, a suitable method is described by the current Eur. Ph. method according to the European Guidelines ICH S6 (CPMP/ICH/302/95) which utilizes Rubella antigen. Further details regarding specific complement activation are provided below in reference to biological activity.

The antibody preparation causes substantially no unspecific complement activation (i.e. activation of the complement cascade by immunoglobulins in the absence of antigen) in in vitro assays suitable to determine unspecific complement activation in normal human serum (i.e. serum from healthy humans). In particular, the assay can determine the amount of C5a and/or C3a generated in the assay in the absence of antigen. As noted above, complement activation results in the production of C5a and C3a. Since both of these proteins are involved in the terminal pathway of the complement system (rather than in either the classical/lectin pathway or the alternative pathway) they are particularly useful to determine complement activation.

The antibody preparation generates substantially no C5a and/or substantially no C3a when used in an appropriate in vitro assay with human serum in the absence of antigen. In a preferred embodiment the antibody preparation adjusted to an IgM concentration of 1.72 mg/ml generates less than 200 ng/ml C5a after 60 minutes of the assay, and/or the antibody preparation adjusted to an IgM concentration of 1.72 mg/ml generates less than 6000 ng/ml C3a after 60 minutes of the assay.

Alternatively, or in addition, the amount of C5a and/or C3a generated by the antibody preparation in the assay is the same as the amount of C5a and/or C3a generated in the same assay by human serum alone ± 70%. Preferably this is after 60 minutes of the assay.

Suitable assays are known in the art. In a preferred embodiment the assay comprises the steps of: (a) adding an amount of the antibody preparation to 100 μΐ human serum to create a reaction mixture containing 1.72mg/ml IgM and incubating the reaction mixture for 60 minutes at 37 °C with constant agitation;

(b) preparing a set of dilutions of the reaction mixture suitable for an ELISA;

(c) performing a sandwich ELISA on the set of dilutions of the reaction mixture utilizing a primary and a secondary antibody to C5a or C3a and a chromogenic substance, wherein the secondary antibody is conjugated to an enzyme and the chromogenic substance is the substrate of the enzyme; and

(d) determining the amount of C5a or C3a in the reaction mixture based on a colour change obtained as a result of contacting the chromogenic substance with the enzyme bound to C5a or C3a via the secondary antibody.

In the ELISA the set of dilutions are contacted with wells of an assay plate coated with the primary antibody. After incubation the wells are washed to remove the dilution sample, The second antibody is then incubated and binds to any C3a/C5a bound to the primary antibody in the wells, since it has a different epitope on the C3a/C5a to the primary antibody. After further washing to remove unbound secondary antibody the chromogen is incubated and reacts with the enzyme conjugated to the second antibody. The resulting colour change can be measured via optical density determinations with a photometer, being proportional to the concentration of C5a/C3a present in the set of dilutions.

In particular, the amount of the antibody preparation added in step (a) is that appropriate to create a concentration of 1.72mg/ml IgM in the reaction mixture. Steps (c) and (d) can comprise: (i) applying the set of dilutions of the reaction mixture to the wells of an assay plate which are coated with a primary antibody to C3a/C5a (i.e. "the capture antibody"); (ii) incubating the plate to allow any C3a/C5a to bind the primary antibody; (iii) washing the plate to remove any material in the dilutions not bound to the primary antibody; (iv) applying a secondary enzyme linked antibody (the detection antibody) that also binds to C3a/C5a; (v) incubating the plate to allow any secondary antibody to bind to C3a/C5a; (vi) washing the plate to remove unbound secondary antibody; (vii) applying a chemical that is converted by the enzyme into a colour signal; and (viii) measuring the absorbency of the plate wells to determine the presence and quantity of C3a/C5a. The sandwich ELISA is performed according to methods known in the art, and/or with commercially available kits according to manufacturer's instructions. Suitable, and particularly preferred, commercially available enzyme linked immunosorbent assay (ELISA) kits are Quidel Micro Vue C5a Plus EIA Kit; A025, and Quidel Micro Vue C3a Plus EIA Kit; A032.

In a further embodiment of the present invention the antibody preparation comprises less than 2% aggregates of 1200 kDa or above, preferably less than 1.5%. This refers to the % of the immunoglobulin content. The amount of aggregates can be determined by high performance size exclusion chromatography (HPSEC). This can be performed by methods known in the art.

Alternatively, or in addition the ability of the antibody preparation to generate substantially no unspecific complement activation can be defined as the anti-complementary activity of the preparation being less than 1 .0 CH50/mg protein, more preferably less than 0.75 CH50/mg protein. The assay to determine the anti-complementary activity on this scale can be carried out according to the method described in the European Pharmacopoeia (method 2.6.17, Ph. Eur. 6. Edition, 2008). Further details of this assay are provided in the assay section below.

In a preferred embodiment the antibody preparation has been prepared from human serum in the absence of a step involving chemical or enzymatic modification of the antibodies, i.e. the process of production of the antibody preparation from human serum does not comprise a step of contacting the antibodies with a reagent which would cause their enzymatic or chemical modification. In particular, the process does not comprise contacting the antibodies with β-propiolactone, which causes chemical modification of the antibodies, or comprise contacting the antibodies with pepsin, which would cause enzymatic cleavage of the antibodies.

Alternatively, or in addition, the antibody preparation has been prepared from human serum in the absence of a step involving heating of the antibodies to a temperature of 40°C or more for 10 minutes or more. In particular, it is known that heating steps can denature the immunoglobulins and causes immunoglobulin aggregation.

Further preferably the antibody preparation is prepared by a process which is capable of a more than 31ogio , preferably more than 4 logio, and most preferably by more than 5 login, removal of non-

Ι Δ enveloped viruses, thus making the antibody preparation virus safe. The antibody preparation is therefore safer than the antibody preparations of the prior art, particularly with respect to active non enveloped viruses like, for example, parvoviruses. This results in an antibody preparation that is substantially free of virus, and in particular substantially free of non-enveloped virus. Still further, the method of the present invention is able to achieve this level of viral particle removal/inactivation without a significant impact on the amount of active IgM or on the anticomplementary activity of the antibody preparation.

In particular, the antibody preparation can be prepared from human plasma by a process comprising the steps of:

(b) mixing a C₇ to C₉ carboxylic acid with the solution and treating the mixed solution with a vibrating agitator to precipitate contaminating proteins;

(c) separating the precipitated proteins from the solution to yield the IgM containing immunoglobulin composition;

(d) incubating the IgM containing immunoglobulin composition at between pH 3.5 and pH 4.5 to form an incubated solution;

It is preferred that the process further comprises subjecting the incubated solution obtained from step (d) to nanofiltration prior irradiation in step (e). Further details and preferred aspects of the method of production are described in the section below.

In a further preferred embodiment of the invention the antibody preparation is capable of administration to cynomolgus monkeys at 1 15 mg IgM/kg body weight/hour in the absence of a 10% or greater reduction in arterial pressure from pretreatment level. As noted above, unspecific complement activation causes hypotension and therefore the lack of a significant change in arterial pressure indicates that substantially no unspecific activation of complement is occurring in healthy

1 J monkeys in vivo. Arterial pressure can be measured by inserting a pressure catheter into the lower abdominal aorta via the right lcrmoral artery.

In a preferred embodiment the antibody preparation also comprises antibodies against one or more of Pneumococcus saccharide, Escherichia coli, Enterococcus faecalis, Candida albicans, and Chlamydia.

In a further preferred embodiment at least 90% of the antibodies in the antibody preparation are biologically active. The term biologically active means that the antibodies in the preparation are in native form and in particular are capable of activation of the complement cascade as a result of specific binding to an antigen. The biological activity of an antibody preparation can be assessed based on assays to determine antibody titre/binding activity and Fc integrity/function known in the art. In particular, in an in vitro Rubella antigen based assay suitable to determine Fc function the activity of the Fc part of the antibodies of the antibody preparation is the same as that of a biological reference preparation ± 10%, more preferably ± 5%.

Biological reference preparations are utilised by the international medical and healthcare community and help to ensure consistency in medical products. As such, suitable biological reference preparations for the assay are known and are available in the art (e.g. Immunoglobulin Biological Reference Preparation (Batch No. 3). In particular, the assay can be performed according to the Internationally recognised Eur. Ph. 2.7.9 Test for Fc Function of Immunoglobulin (current edition April 201 1) which utilises Human Immunoglobulin Biological Reference Preparation (Batch No. 3) as a control, against which the % activity of the antibody preparation is determined. This test comprises the steps of (i) loading tanned group O human red blood cells with rubella virus antigen to create antigen coated blood cells; (ii) incubating an amount of the antibody preparation with the blood cells; adding guinea pig complement to start complement initiated lyses of blood cells; (iii) measuring the kinetic of haemolysis via time-dependent changes of absorbance at 541 nm; (iv) evaluating the function of the antibodies of the antibody preparation using the maximal change of absorbance per time.

The antibody preparation preferably also has a lower proteolytic activity than the antibody preparations described in the prior art. In particular, no proteolytic activity is detectable in the preparation when it is stored at between 2 to 8°C. The proteolytic activity can be measured by standardized test methods known in the art, such as that using the chromogenic substrate which is described in the assay section below, and in Example 6.

The antibody preparation of the present invention may further comprise a stabilizing agent, such as glycine.

As with preparations known in the art the antibody preparation of the present invention can be stored at 5±3°C. However, due to the efficient purification with the method of the present invention the stability of the antibody preparation is extremely good. The final product is stable in liquid form for at least 3 months, preferably at least 6 months and most preferably at least two years at 2 to 8°C, which means that there is no fragmentation or polymerization of IgM above 1.5% measured in HPSEC, no increase of proteolytic activity, no decrease of IgM antibody activity against Escherichia coli and IgM antibody activity against Pneumococcus saccharide of more than 25 % and no increase in anticomplementary activity of more than 25 %, staying below 1 CH50/mg protein. Still further, the final product produced by the method of the present invention is stable in liquid form for at least 3 months, preferably at least 6 months, and most preferably at least one year at room temperature (between 23 and 27°C) as assessed by the same criteria.

Method of production of antibody preparation

As described above, the present invention provides a preparation of an IgM containing antibody preparation from a plasma fraction comprising immunoglobulins. In particular, the present invention provides a method of producing the antibody preparation described herein from human plasma comprising the steps of:

(b) mixing a C₇ to C9 carboxylic acid with the solution and treating the mixed solution with a vibrating agitator to precipitate contaminating proteins;

Plasma fractions suitable for the preparation of pharmaceutical immunoglobulin compositions, and methods for their production are well known in the art. The plasma fraction is preferably a precipitated plasma fraction and most preferably a precipitated plasma fraction obtained by the process of Cohn fractionation or its well known modifications (e.g. Kistler-Nitschmann). Most preferably the fraction is fraction I/III or fraction III (also known as fraction B+I or fraction B) out of cold ethanol fractionation. It is preferred that the immunoglobulins of the plasma fraction comprise at least 5% IgM.

Step (a) comprises providing a plasma fraction as a solution containing the immunoglobulins. In many cases the plasma fraction containing the immunoglobulins will be in solid or semi-solid form. Thus the aim of this step is to ensure or to bring the protein of the plasma fraction into solution such that it is in a suitable state for mixing with the carboxylic acid in step (b). This step may comprise mixing the plasma fraction with a suitable buffer. Preferably the buffer is of low molarity (i.e. less than 1M) and has a pH between 4.5 and 5.5 e.g. 0.1 M sodium acetate buffer pH 5.05±0.1. Mixing can be completed using a blade mixer or a vibrating agitator.

In step (b) the solution formed in step (a) is mixed using a vibrating agitator with a C₇ to C₉ carboxylic acid to precitipate contaminating proteins (e.g. proteases, viruses etc). The carboxylic acid may be branched and/or may include substituents which do not substantially alter the effect of step (b). The carboxylic acid is preferably octanoic acid. The carboxylic acid is preferably added at a concentration of at least 0.075 kg/kg of plasma fraction, up to a concentration of 0.2 kg/kg. More preferably the carboxylic acid is added at 0.8 to 0.15 kg/kg of plasma fraction, and most preferably between 0.09 kg/kg and 0.13 kg/kg. Acid of any convenient molarity may be used to provide the correct concentration. Any type of commercially available vibrating agitator, suitable for use in the chemical/pharmaceutical industry, may be used. Examples of suitable vibrating agitators are available from Graber + Pfenninger GmbH. In particular, the "Labormodell Typ 1 " vibromixer can be used for lab scale experiments, and the "Industriemixer Typ 4" can be used for production scale preparations. The vibrating mixers can be used according to manufacturer's instructions, and in particular at settings which are described by the manufacturers as suitable for mixing solutions containing proteins. For example the vibrating mixers can usually be operated at less than 100 Hz with an amplitude less than 10 mm, e.g. the vibration mixing using the "Labormodell Typ 1" at lab scale was carried out by the present inventors at 50 Hz, when 230 V power supply is used. The vibration amplitude of the mixing process was varied between 0 and 3 mm, and for the IgM preparation preferably 3 mm was used. Stirrer plates with a diameter between 23 mm and 65 mm were used for lab scale experiments. For production scale a stirrer plate diameter of 395 mm was used (hole diameters of 13.5 mm and 16 mm).

In step (b) the pH of the mixed solution is preferably between 4.5 to 5.5, and more preferably between pH 4.8 and pH 5.3. The step can be carried out in sodium acetate buffer, and, for example, with approximately 0. 1 M sodium acetate buffer. The temperature at which step (b) is conducted is preferably between 10°C and 35°C, and more preferably 14 to 30 °C.

The mixing time using the vibrating agitator is not particularly limited but is preferably at least 30 minutes and not more than 3 hours, and more preferably from 40 - 1 10 minutes. Incubation times of less than 30 minutes can reduce the level of virus inactivation.

In one embodiment of step (b) tri-calcium phosphate is mixed with the solution in step (b). Preferably this is added at 0.01 to 0.02 kg/kg plasma fraction (as it is in solid or semi-solid form). The tri-calcium phosphate can be added simultaneously, separately or sequentially to the carboxylic acid. In a preferred embodiment the tri-calcium phosphate is added at least 20 minutes after the carboxylic acid.

In step (c) the contaminating proteins precipitated in step (b) are separated off from the solution to yield the IgM containing immunoglobulin composition (i.e. an immunoglobulin containing solution). This step of separation is not particularly limited by can be performed by any suitable method known in the art. However, the separating step is preferably performed using filtration, and more preferably ultrafiltration, and the result of step (c) is therefore a filtered solution.

As described above, the method of the present invention is advantageous in manufacturing terms since it appears to cause a more efficient precipitation of contaminating proteins, and, as a result, step (c) is easier to perform. When the mixture resulting from step (b) is separated, a transparently clear solution, i.e. the IgM containing immunoglobulin composition, is achieved. Filtration is is therefore quicker and easier.

Further process steps (d) to (f) are required to convert the IgM containing immunoglobulin composition obtained from step (c) into an antibody preparation suitable for intravenous administration.

Step (d) comprises treating the IgM containing immunoglobulin composition obtained from step (c) with mild acid conditions, step (e) comprises subjecting the acid treated composition to UVC irradiation to form a UVC irradiated solution, and step (f) comprises filtering the UVC irradiated solution under sterile conditions to form the antibody preparation suitable for intravenous administration in humans.

For the treatment with mild acid conditions the IgM containing immunoglobulin composition obtained from step (c) is incubated at between pH 3.5 to pH 4.5, and preferably between pH 3.8 and pH 4.2, to form an incubated solution. The mild acid conditions can be created by adding a suitable acid to the IgM containing immunoglobulin composition, for example the pH can be adjusted by adding 0.2M HCl.

This incubation step is preferably carried out at between 32 and 42°C, and more preferably at between 35 and 39°C. The incubation time is preferably at least 2 hours and not greater than 24 hours, and more preferably at least 9 hours but not greater than 16 hours.

In the irradiation step the incubated solution obtained from the mild acid treatment described above is treated with UVC light to form a UVC irradiated solution. This step can be performed using devices which are commercially available, such as the UVivatec® device (Bayer Technology Services). It is preferred that the incubated solution is treated at 254 ± 10 nm between 200 and 500 J/m², more particularly between 200 and 300 J/m², in order to further inactivate viruses and proteases which are potentially present. It is noted that UVC treatment under gentle conditions than would normally be required is only possible with the water-clear filtrate which is obtained by the present invention after the octanoic acid treatment with vibromixing. More opalescent or opaque solutions normally received by standard stirring techniques would necessitate longer irradiation times which would lead to more denaturation of the IgM activity and lower virus inactivation rates.

In step (f) the irradiated solution is filtering under sterile conditions to form the antibody preparation suitable for intravenous administration in humans. Preferably the filtration is nanofiltration, more preferably through a filter having a 40 to 50 nm pore size.

In addition to the mild acid treatment the UVC irradiation and the filtration step, additional steps to achieve an immunoglobulin preparation for intravenous administration can optionally also comprise one or more further filtration steps. In one embodiment the protein solution being processed can be adsorbed onto DEAE-Sephadex and then separated from the Sephadex by depth filtration. For example, it may further be subjected to a batch adsorption with 75 mg per kg protein DEAE Sephadex at pH 5.8, in order to remove the unwanted accompanying protein Ceruloplasmin.

In a particularly preferred embodiment the incubated solution obtained from the mild acid treatment is subjected to adsorption onto DEAE-Sephadex and then separated from the Sephadex by depth filtration, before being treated to UVC irradiation.

In another embodiment the immunoglobulin solution being processed may be filtered through a nanometer filter. Filters of 75 ± 5 nm to 35 ± 5 nm pore size, or filters having a nominal pore size of 75 to 35 nm (for example Pall Ultipor DV50), can be used at various stages during the process. (A nominal pore size of e.g. 50 nm means retention rate of > 4 logl O for virus with size of 50 nm or larger). In a preferred embodiment the solution obtained from the DEAE-Sephadex step described in the above paragraph is filtered through a 0.2μηι filter prior to UVC irradiation.

The final antibody preparation (i.e. the processed IgM containing immunoglobulin solution) obtained from the process defined above may be directly filled into a container under sterile conditions. Alternatively, the antibody preparation may be formulated in a glycine-containing buffer at a pH between 4 and 5.5, and preferably between 4. 1 to 4.5. The antibody preparation may also be diluted to a protein concentration between 40 and 80 g/L and preferably between 55 and 70 g/L. It is noted that it is also possible to enrich the IgM content of the antibody preparation by well known methods like e.g. anion exchange chromatography.

As previously indicated above, the method described above leads to a higher inactivation and removal of virus particles, especially very resistant, non-enveloped viruses such as Parvo viruses, which are usually not very susceptible to octanoic acid treatment. Furthermore, an improved removal of proteolytic activity is achieved in comparison to conventional stirring. These features are achieved while keeping a high yield of IgM that is chemically unmodified. This finding contrasts with the conventional view that the treatment with octanoic acid is not an effective step against non- enveloped viruses and improved viral safety must be achieved through inactivation of virus through harsher methods such as β-Propiolactone treatment. Also it was well known that increasing e.g. octanoic acid concentration to completely remove proteolytic activity results in a massive loss of IgM.

The results of the method are achieved through the use of mixing devices using a vibrating mode in combination with the octanoic acid treatment. This is particularly surprising since it is known that IgM is very susceptible to shear stress, which may lead to an undesired high anticomplementary activity. Accordingly, one would not consider using a vibrating mixer to prepare an Ig composition and would not expect such a favorable impact when using a vibrating mixing during processing o an IgM containing solution.

Furthermore with the method the separation achieved by step (c), such as clarification by filtration of the octanoic acid treated solution resulting from step (b), is enhanced when a vibrating mixing device is used. Separation is more easily achieved, reducing processing time and manufacturing costs, and step (c) leads to a limpid solution which creates advantages for downstream processing. Conventional solutions, achieved by filtering the results of octanoic acid treated IgM containing solutions which have been stirred, are opalescent or opaque. The resulting IgM containing composition obtained from step (c) is preferably subjected to treatment with mild acid conditions (e.g. pH 4) and an UVC-irradiation step to further improve virus safety and stabilize the final product. Due to the enhanced clarification o the IgM containing immunoglobulin composition obtained from step (c) it is possible to lower the necessary irradiation time with UVC to achieve a virus inactivation of non enveloped viruses of more than 3 or 4 logio. This results in a higher yield of native and active IgM during UVC treatment.

Surprisingly these steps lead to a chemically and enzymatically unmodified IgM containing solution which has higher yields of native and active IgM, having low anticomplementary activity and low proteolytic activity and having high antibacterial and antiviral activity, with an outstanding virus safety concerning enveloped and non enveloped viruses; a key feature for pharmaceuticals which are for intravenous administration. Moreover, a treated IgM containing solution has improved long term stability being very stable in liquid solution for more than 12 months at 2 - 8°C.

Medical Use

The antibody preparation of the present invention is suitable for use in medicine and can be used in the treatment of immunological disorders and infections, particularly IgM deficiency disorder and bacterial infections. Human IgM-enriched polyvalent immunoglobulin preparation for intravenous administration contain higher antibody titres against clinically relevant Gram-negative and Gram- positive bacteria as well as higher antibody titres against endotoxins of Gram-negative bacteria and exotoxins of Gram-negative and Gram-positive bacteria compared with polyvalent immunoglobulin G preparations.

In particular, the antibody preparations of the present invention are suitable for intravenous administration to patients.

The invention also provides a method of treatment of a patient comprising a step of administering the antibody preparation of the present invention to the patient. In particular, the patient can be suffering from an immunological disorder or a bacterial infection. In a preferred embodiment the antibody preparations are administered intravenously. The present invention will now be described further by way of example only.

EXAMPLES

Assay methods

Distribution of molecular size by I I PLC for IgM Concentrate

The below method can be utilized to determine the % aggregates in an antibody preparation (as used in Example 8).

Test solution: Samples are injected undiluted at approx. 50 g/L with an injection volume of 10 μΐ (approx. 500 μg protein load).

Reference solution: human immunoglobulin (e.g. Intratect, Biotest AG)

Standard solution: Bio-Rad gel filtration standard (Art.-No. 151-1901)

Column:

— size: 1 = 30 mm, 0 = 7.8 mm,

— stationary phase: Tosoh Bioscience TSK-Gel G4000 SWXL, suitable for fractionation of globular proteins with relative molecular masses in the range 20 000 to 7 x 10⁶ Da.

Mobile phase: dissolve 4.873 g of disodium hydrogen phosphate dihydrate, 1.741 g of sodium dihydrogen phosphate monohydrate, 1 1.688 g of sodium chloride and 50 mg of sodium azide in 1 litre of water.

Flow rate: 0.5 ml/min

Detection: spectrophotometer at 280 nm. In the chromatogram obtained with the reference solution, The chromatogram is integrated according to the following scheme and the peaks are identified:

• Polymer (> 1200 kD), 10 - 13 min

• IgM (1200 - 750 kD), 13 - 19 min

• Dimer / IgA (750 - 350 kD), 19 -20 min

• IgG (350 - 100 kD), 20 - 26 min

• Fragments (< 100 kD), 26 - 40 min

• Fragments (< 100 kD), 26 - 40 min Determination of unspecific complement activation

Hemolysin pre-treated sheep erythrocytes are hemolysed by complement. By complement-binding antibodies in the sample the hemolysis is suppressed. The amount of complement is determined, which is bound (inactivated) by 1 mg immunoglobulin.

A certain amount of immunoglobulin (10 mg) is mixed with complement of guinea pig and the free complement is titrated. The anti-complementary activity is expressed a used complement in respect to the used complement of a reference solution. The hemolytic unit of complement activity (CH₅₀) is the amount of complement leading to hemolysis of 2.5 x 10⁸ optimally prepared erythrocytes of a total amount of 5 x 10⁸ erythrocytes in optimal buffer conditions.

Optimally prepared erythrocytes (8ml stabilized erythrocytes from sheep, washed three times with gelatine-barbital-buffer, finally 1 ml erythrocyte sediment are suspended in 24 ml gelatine-barbital- buffer) are prepared by mixing 20 ml erythrocytes suspension with 20 ml hemolysine (adjusted to 2 MHE/ml - minimal hemolytic unit) and incubation for 15 min at 37°C.

An equivalent of 10 mg immunoglobulin is diluted in gelatine-barbital-buffer (1 g gelatine in 1 L barbital buffer pH 7.3, 5-fold barbital-buffer solution: 83 g sodium chloride, 10.192 g sodium barbital in 2 liters water, pH 7.3). To a final volume of 1 ml 200 μΐ complement 100 CH₅o / ml are added. The test tubes are incubated under shaking for 1 h at 37 °C. The samples are diluted and titrated against optimally prepared erythrocytes. After an incubation for 1 h at 37°C the samples are centrifuged and the optical density is determined by using a spectrophotometer at a wavelength of 541 nm.

Determination of proteolytic activity

Proteolytic activity can be assessed by mixing a chromogenic substrate (in particular those sensitive to at least one serine protease) and a sample of the antibody preparation (usually diluted in buffer to meet the linear range of the assay) at 37°C and monitoring the absorption kinetics using a spectrophotometer. The proteolytic activity of the sample is calculated from the initial absorption difference (AAbs/min) by using the equation C (U/L) = 3 S x AAbs/min x F (C = proteolytic activity; S = conversion factor relating to specific adsorption change o the chromogenic substrate; and F = dilution factor). Use of the substrate is according to manufacturer's instructions.

The proteolytic activity can in particular be assessed via the following steps:

(a) 25 mg of the substrate S-2288 (Chromogenix) is dissolved in 7.2 ml of water- for-injection;

(b) a sample of the antibody preparation is diluted into buffer (lOOmM Tris.HCl pH 8.4, 106 mM NaCl) to meet the linear range of the assay and temperature is adjusted to 37°C;

(c) equal amounts (e.g. 200μ1) of the diluted antibody preparation and the dissolved substrate are mixed;

(d) the absorption kinetics are measured at 405 nm for 1 to 3 minutes at 37°C using a spectrophotometer;

(e) the proteolytic activity of the sample is calculated from the initial absorption difference (AAbs/min) by using the equation C (U/L) = 313 x AAbs/min x F (C = proteolytic activity, F = dilution factor)

The limit of quantitation of this method is 8 U/l, and using a sample of the antibody preparation of the present invention proteolytic activity is undetectable. As such the level of the proteolytic activity in the final product of the present invention is below 8 U/l.

Example 1 -Preparation of an I gM enriched preparation from fraction 1/111

180 kg Cohn Fraction I/I II, originating from cold ethanol fractionation of human plasma are suspended in 720 L 0.1 M sodium acetate buffer pH 5.05 and mixed for 15 - 30 minutes after the suspension temperature is reached (22 ± 4 °C).

The solution is treated by addition of a 19.8 kg octanoic acid (0.1 10 kg per kg paste I/I I I used) at room temperature and the protein solution is further mixed for 80 minutes, using a vibrating mixer (Vibromixer^®, Size 4, Graber+Pfenniger GmbH, Vibromixer adjusted to level 2 - 3). The octanoic acid is added slowly over 30 min.

Approx. 3 kg tri-calcium phosphate (Ca₃(P0₄)₂) are added and the protein solution is further mixed for at least 15 min. The precipitate is removed by clarifying filtration using a filter press. An additional 0.2 μηι filtration is carried out and the protein solution is subjected to ultrafiltration with 10 Id⁾ membranes. The protein solution is diafiltered against 0.04 M NaCl solution and afterwards adjusted to a protein concentration of 40 g/L.

The protein solution is treated at pH 4.0 ± 0.1 after dilution 1+1 with water for injection. pH adjustment is carried out by using 1 M HC1 and the protein solution is incubated for 9 h at 37 °C ± 2°C. After the pH 4 incubation the protein solution is adjusted to pH 5.8, using 1 M NaOH. The resulting protein solution is further purified by adding DEAE Sephadex in a batch mode (75 g DEAE Sephadex per kg protein). The protein solution is incubated under stirring for > 60 min at room temperature. The DEAE Sephadex is removed by clarifying filtration. The protein solution is subjected to a 0.2 μηι filtration.

The protein solution is filtered through a 0.1 μιη filter and a Pall, Ultipor VF DV50, 20" filter. The filtrate is further processed by UVC light treatment at 254 nm, using a flow-through UVivatec® process device (Bayer Technology Services / Sartorius Stedim) at a UVC dose of 240 J/m². The flow velocity through the UVC reactor is calculated using the manufactures instructions. The irradiated protein solution is concentrated to a protein concentration of 50 - 70 g/1 by ultrafiltration and is subjected to diafiltration (10 kD membrane, using 0.32 M glycine buffer pH 4.3). The final product is filtered through a 0.2 μη filter and is stored at 2 to 8°C.

Example 2 - Investigation of Conditions in Octanoic Acid Treatment Step

For the octanoic acid treatment the following experimental ranges were tested, also in combination with each other using the method described in Example 1 (results not shown).

Octanoic acid amount: 0.09 kg/kg to 0.13 kg/kg (Amount octanoic acid per kg used fraction

I/I 11) (120 to 180 mM octanoic acid)

pH of the octanoic acid treatment between pH 4.8 and 5.3

Temperature range of the reaction: 14°C to 30°C

Incubation time: 40 to 110 min

All conditions tested lead to intermediates being easy to clarify for further processing and with a extensive reduction of proteolytic activity from several thousand U/L in suspended Cohn fraction I/III). These intermediates result in a final product with a proteolytic activity below 8 U/l (calculated as described below in Example 6) which is the limit of quantitation.

Example 3 - Virus Reduction Through Use of Vibromixer - Determination of virus removal factors for the octanoic acid treatment with and without use of a vibromixer.

250 ml of suspended faction I/III were homogenised for 30 min at pH 5.05 and 22°C. The suspension was spiked with 2.6 ml of the virus stock solution. Octanoic acid was added (1 10 g/kg) and homogenised for 60 min using a vibromixer. In a parallel experiment the same mixture was homogenised with standard stirring. After 60 min tri-calcium phosphate (0.15 g/kg octanoic acid) was added and the suspension stirred for 15 min. The suspension was cleared by depth filtration using a filter disc. The filter disc was pre-rinsed with 70 - 80 ml of buffer. After filtration, the filter was rinsed with 80 ml of buffer. Filtrate and wash were pooled and a sample was drawn for virus titration.

Virus titres from samples taken prior to addition of octanoic acid and after filtration were determined on appropriate indicator cells for SV40, Reo and PPV (CV-1 , CCL.7.1 and PK13). Finally, the removal factor was calculated in compliance with the current guidelines for virus validation studies.

In virus validations studies, non-enveloped viruses such as SV40 and Reo were effectively removed in the order of more than 4 logio and more than 5 log_{] 0}, respectively. Moreover, PPV was removed by more than 3 logui. These values are more than 10 times and up to 1000 times higher than with the same octanoic acid treatment under standard stirring conditions without vibromixing.

Table 1: Comparison of the virus reduction factors (Iog_{] 0}) for the octanoic acid treatment with and without the use of a vibromixer.

Example 4 - Evaluation of UVC treatment

The optimal range for the dosage of UVC radiation has been evaluated. There is a balance between the minimal necessary dosage to achieve at least 4 logio inactivation for non enveloped viruses and the maximum tolerable dosage to avoid denaturation of the IgM molecules leading to an impaired Fab function to bind antigens and impaired Fc function influencing complement activation. In the range of 200 to 400 J/m² one could observe only a slight increase of immunoglobulin aggregates and no significant impact on fragment content.

For the experiments the optical density (OD) of the original protein solution is used to calculate the flow rate in the UVivatec lab system with the vendor provided Excel-Sheet from BTS (customer Master Calculation Sheet UVivatec Lab I I Version 3.0). The flow rate is calculated by taking into account the lamp performance, the set point of the UV signal lamp sensor and the desired UVC irradiation dose.

IgM containing solution with a protein content of about 55 g/1 (Batch 86GB005BE07) was pumped at a flow rate of 5.8 l/h through the UVivatec system in order to achieve a dose of 200 J/m² for a single flow-through. A dose of 300 J/m² was achieved by pumping the protein solution at a flow rate of 3.9 L/nr through the system. 400 J/m² were achieved by pumping the protein solution at a flow rate of 2.9 L/m² through the system. Table 2: Analytical results for the activity and titre determinations before and after UVC treatment in the concentrated final product

No significant difference could be observed for immunoglobulin content, proteolytic activity or ACA in the range of 200 to 400 J/m² . The preferred range for the dosage was set between 200 and 300 J/m² because 200 J/m² are well enough to inactivate non enveloped viruses and at 300 J/m² no significant impact could be seen on aggregate formation and antibody titres. The preferred dosage is 225 J/m2

Diluted IgM containing solution with a protein content of 8 to 12 g/1 (Batch 86BB059BE07) was pumped at a flow rate of 5.8 1/h through the UVivatech system in order to achieve a doses between 200 and 300 J/m² for a single flow-through. Table 3: Analytical results for IgM solutions before and after UVC irradiation at different

UVC doses

Batch fraction 1/111 before UVC: UVC: UVC: UVC:

86BB059BE07 UVC 200 J/in² 225 J/m² 250 J/m² 300 J/m²

Anti-Enterococcus

faecalis - IgM U/mg 17.2 14. 1 16.7 14.0 13.9

Anti-Pneumococcus

Saccharic! -IgG U/mg 23.2 24.1 24.7 24.0 25.7

Anti-Pneumococcus

Saccharid -IgA U/mg 13.3 12.1 18.0 16.5 14.8

Anti-Pneumococcus

Saccharid -IgM U/mg 17.5 15.1 18.0 16.4 16.6

The distribution between the immunoglobulin classes remains unaffected by the UV irradiation within this dosage range procedure. The molecular weight distribution pattern analyzed by HPSEC is also not changed. The level of purity analyzed by CZE remains unchanged. Proteolytic activity (PA), prekallikrein activator (PKA) and anti-complementary activity (ACA) are unchanged. Also the anti bacterial activity measured by an Elisa method are not significantly altered for all immunoglobulin classes.

The aliquots - irradiated with increasing UV intensities - were further processed until final product and subjected to the same panel of analytical tests. There was also no significant difference observable in the final products. All tested antibody titres are always in the range of 100 + 10 % of the control preparation not UVC treated.

Example 5 - Overall Virus Reduction Through Use of Vibromixer/pH4 treatment and UVC treatment - Determination of virus removal factors

The validation of virus removal/inactivation of the three steps octanoic acid treatment with vibromixing, pH4 treatment and UVC treatment (215 J/m²) was performed using the following model viruses : Bovine Viral Diarrhea Virus (BVDV) as model virus for Hepatitis C Virus,

Pseudorabies Virus (PRV) as model virus for Human I lerpes Viruses, Human Immunodeficiency virus (HIV- 1 ), Equine Arteritis Virus (EAV) as model virus for corona viruses, Sindbis Virus

(SinV) as model virus for Flavi viruses, Murine Encephalomyelitis Virus (MEV) as model virus for

Hepatitis A Virus, Reovirus (Reo) as model virus for other non enveloped viruses, Porcine

Parvovirus (PPV) as model virus for human Parvovirus B 19. The results of these studies with the three steps octanoic acid treatment, pi 14 treatment and UVC treatment are listed in the following Tabic 2.

a Reduction factor without data for validation of the UVC irradiation step

The optional nano filtration with filters with a nominal pore size of about 50 nm adds additional safety by increasing the total reduction up to more than 17 log 10 depending on the size of the virus. E.g > 17.5 logio are then reached for HIV-1 whereas PPV was not further removed by nano filtration.

Therefore the purification procedure according to the invention leads to an outstanding virus safe IgM preparation with up to now for such an IgM containing preparation unreached virus inactivation / reduction rates of more than 8 logio. This is especially important for the non enveloped viruses like MEV, Reo and PPV which are generally more resistant against virus inactivation and removal procedures due to their small size and the lack of a lipid envelope.

Example 6: Determination of residual proteolytic activity for the octanoic acid treatment with and without use of a vibromixer.

The octanoic acid treatment was performed like in example 1 and in a parallel experiment without a vibromixer but with vigorous standard stirring with a blade stirrer. The proteolytic activity in samples after octanoic acid / tricalcium phosphate treatment and ultra-/diafiltration were determined using the chromogenic substrate S-2288 (Chromogenix), following the manufacturers instructions. 25 mg of the substrate S-2288 (Chromogenix) are dissolved in 7,2 ml water-for-injection. Samples are diluted into buffer (100 raM Tris/HCI pH 8.4, 106 niM NaCl) to meet the linear range of the assay, e.g. 200 μΐ buffer are mixed with 200 μΐ sample (mixing and temperature adjustment to 37 °C) and 200 μΐ chromogenic substrate solution. The absorption kinetics are measured at 405 nm (1-3 min) at 37 °C, using a spectrophotometer. The proteolytic activity of the sample is calculated from the initial absorption difference (AAbs/min) by using the equation C (U/L) = 3 13 * AAbs/min * F (C = proteolytic activity. F = dilution factor)

Table 5: Reduction of roteol tic activit b octanoic acid treatment

The filtrate after octanoic acid treatment was limpid when vibromixing was employed. In the comparison experiment the filtrate after octanoic acid treatment with blade stirrer was very opaque and difficult to filtrate.

Example 7: Anti-bacterial titres in an IgM preparation according to the invention

For comparison with the only commercially available intravenous tolerable IgM containing preparation Pentaglobin, the anti -bacterial activities were analyzed in three batches of this well established drug and compared to a preparation according to the invention. The determination of antibodies of the IgA or IgM class in the IgM preparation versus antibacterial or antifungal antigens was carried out by ELISA. Microtitre plates were coated with a corresponding antigen and incubated with a standard or the IgM preparation. Antibodies bound to the antigen were detected with an anti- human-IgA or anti-human-IgM conjugate. The detection was carried out by using an enzyme substrate. The resulting colour change is corresponding to the amount of antibodies present in IgM preparation.

Table 6 Comparison of anti bacterial binding activity of IgM in an preparation according to the invention and commercially available Pentaglobin

Table 7 Comparison of anti bacterial binding activity of IgA in an preparation according to the invention and commercially available Pentaglobin

The IgM and IgA mediated activities in the new preparation were typically at least 1.5 times as high as in Pentaglobin which can be explained by the fact that IgM and IgA in Pentaglobin is chemically modified with β-Propiolactone This step is replaced by the more gentle procedures according to this invention.

Overall these data demonstrate that the binding region of the IgM molecules in the final preparation is functionally full active. Example 8 Storage stability studies with liquid IgM product

Product according Example 1 without UVC treatment was stored in 10 or 100 ml glass vials (filling volume 5 ml or 50 ml) at 2 - 8°C and analyzed for all parameter according to specification. The results are shown in Table 8. Parameter which are relevant to show stability are the aggregate and fragment content measured with high performance size exclusion chromatography (HPSEC), proteolytic activity (PA) and anticomplementary activity (ACA). These parameter are critical for intravenous tolerability and likely to change during long term storage. At 2 - 8 °C there was no significant change of these parameters. Even at storage at room temperature (23 - 27°C) these values remained within specification, although there is a slight increase of fragments after 24 months at room temperature. Other parameter like coloration, opalescence, pH value were also determined and stayed unchanged over the whole study period. IgM and IgA titres against various bacteria remain stable over 2 years at 2 - 8°C.

Product according example 1 with UVC treatment was also stored in 10 or 100 ml glass vials (filling volume 5 ml or 50 ml) at 2 - 8°C and room temperature and analyzed for all parameter according to specification. The results are shown in Table 9. In this ongoing stability study the currently available 12 month date show the same stability profile of the product as without UVC treatment which allows the extrapolation to a 24 month stability.

Table 8 Stability of the batch A586067 tested at 2-8 °C lying position Filling size: 5 ml

(%)

IgM content > 20 % 21 .6 22.1 22.1 n.t. 22.3 n.t. 20.9 20.5 anticomplementary

activity < 1.0 0.48 0.56 0.48 0.66 0.70 0.64 0.54 0.38

(CH50/mg protein)

n.t. = not tested

Table 9 Stability of the batch A586057 tested at 2-8 °C lying position Filling size: 50 ml

Example 9 In vitro unspecific complement activation with the IgM product

Example 9A - Determination of C5a levels

Analysis of the potential of the IgM preparation to activate complement in vitro unspecifically was performed using factor C5a as a marker for activation of the terminal complement pathway. For this purpose human serum was incubated with immunoglobulin products or buffer for 120 min. Samples were taken after 0, 5, 15, 60 and 120 minutes of incubation. In order to demonstrate appropriate function of the in vitro system complete inhibition as well as full activation of the complement system was shown. The complement factor concentration was measured via optical density determinations with a photometer using a commercial available enzyme linked immunosorbent assay (ELISA) kit (Quidel MicroVue C5a Plus EIA Kit; A025).

Human serum (Quidel NHS; Al 13) was thawed quickly at 37°C and immediately put on ice. Every single sample consisted of a reaction batch containing serum (100 μΐ). Additives were first pipetted followed by the addition of human serum to start the reaction in every reaction batch.

Human serum without any additives served as a blank and showed baseline complement activation due to the experimental setup. Addition of heat aggregated IgG (HAAG Quidel; A1 14; 1.3 μΐ) served as a strong activator of human serum complement to demonstrate the responsiveness of the in vitro system. EDTA (final concentration 10 mM) was added to human serum in order to completely inhibit complement activation over the entire reaction time and experimental treatment. IgM preparation, Pentaglobin (according to EP0013901) and an IgM preparation according to EP0413187 were adjusted to an IgM-concentration of 1.72 mg/ml in each reaction. As negative controls the respective volume of formulation buffer was used.

All reactions were stopped after incubation for 0, 5, 15, 60 and 120 minutes at 37°C under constant agitation by addition of stabilizing solution (Quidel sample stabilizer A9576; 140 μΐ). Subsequent sample dilution and ELISA analysis was performed following the manufacturer's protocol. The experiment was performed in two independent replicates and mean values were calculated. The results are shown in Table 10 and Figure 2.

Addition of activator (heat aggregated IgG) lead to a strong increase of C5a within 15 minutes indicating a sensitive response of the in vitro system to detect complement activation.The addition of EDTA as inhibitor resulted in unchanged values over the entire incubation time showing that complement activation is specific and not an artefact due to sample handling or preparation. Incubation of human serum at 37°C and exposure to artificial surfaces induced a slight complement activation ducumented as blank values. The IgM reference preparation according to EP0413187 resulted in complement activation up to more than 1000 ng/ml after 60 Minutes (table 10). The commercially available chemically modified reference preparation Pentaglobin (EPOO 13901 ) still showed half of the complement activating potential compared to the EP0413187 product.

The concentration of C5a in serum treated with the IgM preparation according to this invention is comparable to the C5a concentration measured in serum without additives (blank) or in serum treated with the formulation buffers (300 mM Glycin, pH 4.3 or 0.45 % NaCl/2.5% Glucose, pH 6.8). Thus the immunoglobulins in the IgM preparation according to this invention substantially do not unspecifically activate complement in human serum in the in vitro test system.

Table 10 Mean C5a concentration detected in human serum treated with IgM containing immunoglobulins

Example 9B Determination of C3a levels

Analysis of the potential of the IgM preparation to activate complement in vitro unspecifically was performed using factor C3a as a marker for activation of the complement pathway. For this purpose human serum was incubated with immunoglobulin products or buffer for 120 min. Samples were taken after 0, 5, 15, 60 and 120 minutes of incubation. In order to demonstrate appropriate function of the in vitro system complete inhibition as well as full activation of the complement system was shown. The complement factor concentration was measured via optical density determinations with a photometer using a commercial available enzyme linked immunosorbent assay (ELISA) kit (Quidel Micro Vue C3a Plus EIA Kit; A032).

Human serum (Quidel NHS; A113) was thawed quickly at 37°C and immediately put on ice. Every single sample consisted of a reaction batch containing serum (100 μΐ). Additives were first pipetted followed by the addition of human serum to start the reaction in every reaction batch.

Human serum without any additives served as a blank and showed baseline complement activation due to the experimental setup. Addition of cobra venom factor (CVF Quidel; A600; 20 U/ml) served as a strong activator of human serum complement to demonstrate the responsiveness of the in vitro system. EDTA (final concentration 10 mM) was added to human serum in order to completely inhibit complement activation over the entire reaction time and experimental treatment. IgM preparation and Pentaglobin (according to EP0013901) were adjusted to an IgM-concentration of 1.72 mg/ml in each reaction. As negative controls the respective volume of formulation buffer was used.

All reactions were stopped after incubation for 0, 5, 15, 60 and 120 minutes at 37°C under constant agitation by addition of stabilizing solution (Quidel sample stabilizer A9576; 140 μΐ). Subsequent sample dilution and ELISA analysis was performed following the manufacturers protocol. The experiment was performed in two independent replicates and mean values were calculated. The results are shown in Table 1 1 and Figure 3.

Addition of activator (CVF) lead to a strong increase of C3a within 15 minutes indicating a sensitive response of the in vitro system to detect complement activation. The addition of EDTA as inhibitor resulted in unchanged values over the entire incubation time showing that complement activation is not an artefact due to sample handling or preparation. Incubation of human serum at 37°C and exposure to artificial surfaces induced a slight complement activation documented as blank values. The commercially available chemically modified reference preparation Pentaglobin (EP0013901) showed a three times higher C3 activating potential compared to the blank.

The concentration of C3a in serum treated with the IgM preparation according to this invention is comparable to the C3a concentration measured in serum without additives (blank) or in serum treated with the formulation buffers (300 mM Glycin, pH 4.3 or 0.45 % NaCl/2.5% Glucose, pH 6.8). Thus the immunoglobulines in the IgM preparation according to this invention substantially do not unspecifically activate complement in remarkable amounts in human serum in the in vitro test system.

Table 11 Mean C3a concentration detected in human serum treated with IgM containing immuno lobulins

Example 10 In vivo experiments with IgM product

To confirm safety and tolerability the effects of the IgM preparation on arterial blood pressure following repeated intravenous infusions over 5 days were studied in 8 conscious cynomolgus monkeys. A dose of 190 mg/IgM/kg/day of the IgM preparation prepared according to the methods described herein was administered. Pentaglobin, the commercially available intravenous tolerable IgM containing preparation was administered to some monkeys as a comparison substance. Pentaglobin was administered in such a way that the same IgM dose was administered. Blood pressure was determined following injection to determine whether administration was associated with an intolerable level of unspecific complement activation. A control dose of 0.9% NaCl was administered to the animals several hours prior to the administration of the immunoglobulin preparations. Blood pressure was determined by inserting a pressure catheter into the lower abdominal aorta via the right femoral artery. Results were transmitted by telemetry.

The administration of the IgM preparation (15 ml/kg/day) had only minor effects on arterial blood pressure (mean, systolic and diastolic). The differences up to 4 hours after every infusion compared to pretest values did not exceed 4 mmHg. These differences can be considered not biologically relevant.

Table 12a C3a levels [ng/mlj after the administration of the IgM preparation

C3a levels were determined in plasma samples taken after injection as a marker for unspecific activation of the complement pathway. C3a levels [ng/ml] were only slightly increased by the administration of the IgM preparation (15 mL/kgBW) and were even lower than with the commercially available reference preparation Pentaglobin at equal amounts of IgM. Blood samplings were performed approximately 6 hours after treatment.

No substantial toxicological findings could be attributed to the IgM preparation, and there were no relevant alterations that have not been observed with Pentaglobin. As the safety of Pentaglobin is well established in the clinical practice of many years it is reasonable to conclude that these alterations do not have any clinical relevance.

The good tolerability and safety of the IgM preparation was also verified in a human Phase I study in 24 healthy male and female volunteers. Systolic blood pressure in the first 4 hours after administration in the mean decreased only about 9 % (1 1.9 mmHg) after infusion of 91 to 274 mg of the IgM preparation per kg BW/d at 0.5 ml/min.

This was in the same range like the placebo 0.9 % NaCl-solution (9.4%, 1 1.7 mmHg).

No serious advsere events were recorded and all non-serious adverse events were self limiting. Further, there was no evidence for the transmission of an infectious agent, as shown by PCT determinations.

It is noted that the usefulness of efficacy studies in animal models of relevant diseases is limited due to the immunogenicity and preformed Gal-antibodies in IgM preparations obtained from human plasma. However, given the prior art knowledge regarding the use of Pentaglobin in the treatment of disease and the anti-bacterial antibody titres of the IgM preparation prepared by the method of the present invention (as demonstrated in Example 7) it can be concluded that the IgM preparation have clinical efficacy. Example 1 1 Functional integrity of the Fc part of the antibody preparation

Functional integrity of the Fc part of the antibodies in the antibody preparation prepared in accordance with the method described herein was analysed using the current Ph. Eur. method (2.7.9 Test for Fc Function of Immunoglobulins Eur. Ph. Edition current April 201 1) according to the European Guidelines ICH S6 (CPMP/ICl 1/302/95) (Note for Guidance on preclinical safety evaluation of biotechnology-derived pharmaceuticals) for IgG preparations. The European Pharmacopoeia's monograph for immunoglobulins (01/2005:20709) proposes a Rubella antigen- based test for Fc function of immunoglobulins.

In particular, tanned group O human red blood cells were loaded with rubella virus antigen. Specific volumes of the antibody preparations were incubated with antigen coated blood cells. The complement-initiated lysis of the blood cells was started by adding guinea pig complement. The kinetics of subsequent haemolysis was measured via time-dependent changes of absorbance at 541 nm. The evaluation was carried out using the maximal change of absorbance per time. Human Immunoglobulin Biological Reference Preparation; BRP Batch no. 3 was used as the comparison.

The activity of the Fc part of the antibody molecule was determined in 7 batches of the IgM containing antibody preparation and was in all batches between 96.5 and 103.3 % compared to the biological reference preparation (BRP), therefore proving the functionality of the IgM containing antibody preparation.

Claims

1 . An antibody preparation suitable for intravenous administration in humans comprising IgG, IgA and at least 5% IgM antibodies by weight of the total amount of antibodies, wherein the preparation is prepared from human plasma, wherein the antibody preparation has specific complement activating activity, and wherein in an in vitro assay with human serum suitable to determine the ability of the antibody preparation to activate complement unspecifically the antibody preparation generates substantially no C5a and/or substantially no C3a.

2. An antibody preparation according to claim 1 further comprising more than 5% IgA and more than 40% IgG by weight of the total amount of antibodies.

3. An antibody preparation according to claims 1 or 2 comprising at least 10% IgM by weight of the total amount of antibodies.

4. An antibody preparation according to claim 3 comprising at least 15% IgM by weight of the total amount of antibodies.

5. An antibody preparation according to any preceding claim wherein the antibody preparation adjusted to an IgM concentration of 1.72 mg/ml generates less than 200 ng/ml C5a after 60 minutes of the assay.

6. An antibody preparation according to any preceding claim wherein the antibody preparation adjusted to an IgM concentration of 1.72 mg/ml generates less than 6000 ng/ml C3a after 60 minutes of the assay.

7. An antibody preparation according to any preceding claim wherein in the in vitro serum assay the antibody preparation with human serum generates the same amount of C5a and/or C3a as human serum alone ± 70%. .y

to determine that the antibody preparation generates substantially no C5a comprises the steps of: (a) adding an amount of the antibody preparation to 100 μΐ human serum to create a reaction mixture containing 1.72mg/ml IgM and incubating the reaction mixture for 60 minutes at 37 °C with constant agitation;

(b) preparing a set of dilutions of the reaction mixture suitable for an ELISA;

(c) performing a sandwich ELISA on the set of dilutions of the reaction mixture utilizing a primary and a secondary antibody to C5a and a chromogenic substance, wherein the secondary antibody is conjugated to an enzyme and the chromogenic substance is the substrate of the enzyme; and

(d) determining the amount of C5a in the reaction mixture based on a colour change obtained as a result of contacting the chromogenic substance with the enzyme bound to C5a via the secondary antibody.

9. An antibody preparation according to any preceding claim wherein the in vitro serum assay to determine that the antibody preparation generates substantially no C3a comprises the steps of:

(a) adding an amount of the antibody preparation to 100 μΐ human serum to create a reaction mixture containing 1.72mg/ml IgM and incubating the reaction mixture for 60 minutes at 37 °C with constant agitation;

(b) preparing a set of dilutions of the reaction mixture suitable for an ELISA;

(c) performing a sandwich ELISA on the set of dilutions of the reaction mixture utilizing a primary and a secondary antibody to C3a and a chromogenic substance, wherein the secondary antibody is conjugated to an enzyme and the chromogenic substance is the substrate of the enzyme; and

(d) determining the amount of C3a in the reaction mixture based on a colour change obtained as a result of contacting the chromogenic substance with the enzyme bound to C3a via the secondary antibody.

10. An antibody preparation according to any preceding claim comprising less than 2% aggregates of 1200 kDa or above.

1 1. An antibody preparation according to claim 10 comprising less than 1.5% aggregates of 1200 kDa or above.

12. An antibody preparation according to any preceding claim wherein the anti-complementary activity of the preparation is less than 1.0 CH50/mg protein.

13. An antibody preparation according to claim 12 wherein the anti-complementary activity is less than 0.75 CH50/mg protein.

14. An antibody preparation according to any preceding claim comprising an immunoglobulin content of greater than 95% of the total protein content.

15. An antibody preparation according to any preceding claim prepared from human serum in the absence of a step of heat treatment at a temperature 40°C or above for more than 10 minutes.

16. An antibody preparation according to any preceding claim prepared from human serum in the absence of a step involving chemical or enzymatic modification of the antibodies.

17. An antibody preparation according to claim 16 wherein the step of chemical modification is a step of contacting the antibodies with β-propiolactone.

18. An antibody preparation according to any preceding claim prepared by a process which is capable of a more than 31ogio removal of non-enveloped viruses.

19. An antibody preparation according to any preceding claim prepared from human plasma by a process comprising the steps of:

(d) incubating the IgM containing immunoglobulin composition at between pH3.5 and pH 4.5 to form an incubated solution;

20. An antibody preparation according to claim 19 wherein the process further comprises subjecting the incubated solution obtained from step (d) to nanofiltration prior irradiation in step (e).

21. An antibody preparation according to any preceding claim wherein the antibody preparation is capable of administration to cynomolgus monkeys at 1 15 mg IgM/kgBW/hr in the absence of a greater than 10% reduction in arterial pressure from pretreatment level.

22. An antibody preparation according to any preceding claim wherein at least 90% of the antibodies in the preparation are biologically active.

23. An antibody preparation according to any preceding claim wherein in an in vitro Rubella antigen based assay suitable to determine Fc function the activity of the Fc part of the antibodies of the antibody preparation is the same as that of a biological reference preparation ±10%.

24. A method of producing an antibody preparation according to any preceding claim from human plasma comprising the steps of:

25. An antibody preparation according to any one of claims 1 to 23 for use in medicine.

26. An antibody preparation according to claim 25 for use in the treatment of an immunological disorders or a bacterial infection.

27. An antibody preparation according to claim 26 wherein the immunological disorder is an

IgM deficiency disorder.

28. Use of an antibody preparation according to any one of claims 1 to 23 for the manufacture of a medicament for treatment of an immunological disorder or a bacterial infections.

29. A method of treatment of a patient comprising administering an antibody preparation according to any one of claims 1 to 23.

30. A method of treatment according to claim 29 wherein the patient is suffering from an immunological disorder or a bacterial infection.

31. A method of treatment according to claim 29 and 30 wherein the antibody preparation is administered to the patient intravenously.