A method of removing bacterial endotoxin from a pharmaceutical process solution is disclosed. In one embodiment, the method comprises treating the solution with a surfactant effective to dissociate the endotoxin from a pharmaceutical drug or vaccine substance in the solution, and then filtering the solution through a molecular cut-off filter having a pore size effective to retain the pharmaceutical drug or vaccine substance but allow the dissociated bacterial endotoxin to pass therethrough.

FIELD OF THE INVENTION

This invention relates to a method of removing bacterial endotoxins from pharmaceutical compositions, e.g. vaccine compositions.

BACKGROUND OF THE INVENTION

Endotoxins are lipopolysaccharides which are typically derived from the cell walls of bacteria, and in particular Gram negative bacteria. Contamination with endotoxin is a common problem in the pharmaceutical industry, particularly in non-sterile production processes.

Vaccines, for example viral vaccines, can be produced by inoculating a suitable viral substrate with a virus, incubating the substrate to allow viral replication and then harvesting the virus from the substrate. The virus is then inactivated using a suitable inactivating agent and the viral solution is further processed and purified to give the vaccine.

In one method of producing vaccines, and in particular viral vaccines such as influenza vaccine, eggs are used as the substrate for the virus. Thus the eggs are inoculated with seed virus, incubated to allow viral replication, and the allantoic fluid containing the virus harvested from the eggs. The allantoic fluid is then subjected to a sequence of purification steps to give a purified viral fraction which may then be lysed with a detergent such as Triton to disrupt the virus and release the desired viral antigens. The viral antigens are further purified and optionally blended, where desired, with other antigens to give a multivalent vaccine composition. Preparation of the vaccines is typically conducted in sterile rooms, and preservatives such as thiomersal are added at various stages during the process to minimise or prevent bacterial growth.

When eggs are used as the viral substrate, although they go through a cleaning process prior to inoculation to reduce the bioburden, the cleaning process does not guarantee complete removal of microorganisms. Moreover, it is possible that the eggs may not be pathogen free and may contain bacteria. Consequently, during the period of viral growth in the eggs, any contaminating bacteria will also multiply. Although the bacteria may be killed by preservative, the endotoxin-containing bacterial cell walls remain. During the disruption of the viral particles with detergent, the bacterial cell walls are also ruptured leading to release of endotoxin which may copurify into the final vaccine composition. If high levels of bacterial endotoxin are detected in the vaccine, current practice is to discard the contaminated vaccine rather than attempt to remove the endotoxin. As a consequence, there can be significant and expensive wastage during vaccine production as a result of endotoxin contamination.

Endotoxin contamination represents a particular problem in the manufacture of influenza vaccines. The primary constituents of influenza vaccines are the haemagglutinin (HA) surface antigens, along with smaller amounts of neuraminidase surface antigens (NA). Both the HA and NA can form characteristic rosette structures, either alone or as mixed rosettes containing both types of antigens. This behaviour is typical of membrane proteins where the hydrophobic stalks which traverse the membrane are forced to self-associate to produce a stable hydrophobic microenvironment.

The haemagglutinins generally exist in the form of trimers whilst the neuraminidase antigens tend to exist in tetrameric forms. Once formed, the rosettes are very stable and, for example, are not easily disrupted with detergents (Sian Renfrey PhD Thesis, University of Oxford 1994).

Bacterial endotoxins are lipopolysaccharides which possess a large hydrophilic polysaccharide chain, and a hydrophobic fatty acid-containing tail. Thus they have an amphiphilic structure. When in an aqueous environment, there will be a tendency for them to form aggregates.

It has been found that influenza surface antigens and endotoxin are difficult to separate from one another, and it is believed that this may be due to the fact that both influenza surface antigens, such as HA, and endotoxin have an amphiphilic structure and may become strongly associated under aqueous conditions. It also appears that endotoxin is incorporated into the HA/NA rosettes.

Any method of removing endotoxin from vaccine compositions must fulfil a number of criteria. Firstly, the method must not result in excessive loss of the product antigen. Secondly, it must be capable of removing relatively high concentrations of endotoxin (for example it should be capable of reducing the levels of endotoxin to less than 200 EU/ml, e.g. to less than 100 EU/ml). The method must not introduce potentially toxic chemicals into the product and any chemicals which are used must not adversely affect the antigens. Any such method must also be suitable for scaling up for use under production conditions. A method which works on only a small scale, but cannot be scaled up efficiently, is of no use in a manufacturing context. Finally, it is desirable that the method of removal of endotoxin should be capable of being carried out using existing process and production equipment.

A number of attempts have been made to remove endotoxin from vaccines but these have hitherto been generally unsuccessful.

Sucrose density gradient centrifugation is a technique which relies on the separation of substances on the basis of their specific molecular densities, assuming that there is no interaction between them. This technique has been applied by the present applicants to the problem of endotoxin removal from influenza antigen solutions; it was found that although 30% of the endotoxin eluted separately from the haemagglutinin on the sucrose density gradient, the remaining 70% remained associated with the surface antigen. Addition of detergent to the sucrose density gradients only improved endotoxin resolution from haemagglutinin by a relatively small extent.

It has been reported that triethylamine interacts with endotoxin causing an increase in endotoxin hydrophobicity and hence it was postulated that triethylamine may induce separation of the HA/NA/endotoxin complex believed to be formed. Triethylamine was therefore added to the sucrose density gradient, but this was unsuccessful in improving the resolution of endotoxin from haemagglutinins.

Non specific absorption of endotoxin onto activated charcoal, glass, anion exchange media (DEAE) and polystyrene have also been attempted but such attempts were not successful.

A number of affinity chromatography media are commercially available for the purpose of removing endotoxin from pharmaceuticals. The proprietary detoxifying media "Acticlean Etox, Prosep-Remtox and CUNO Zeta Plus ZA" were tried with varying degrees of success. Thus "Acticlean Etox" was successful in separating 99% of the endotoxin from the haemagglutinins in a small scale experiment, but the small scale separation could not be reproduced adequately and the method did not work effectively on a larger scale. "Prosep-Remtox" did not adsorb either endotoxin or haemagglutinin, whereas CUNO Zeta Plus ZA" adsorbed 100% endotoxin and 100% haemagglutinin.

Thus, there remains a pressing need for a method of removing endotoxin from pharmaceutical compositions, and in particular vaccine compositions.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide a method of removing endotoxin which is effective on a manufacturing scale, does not leave toxic or objectionable chemical residues in the pharmaceutical product, is compatible with conventional process equipment and, particularly in the case of vaccine, does not result in the excessive loss or denaturing of antigen.

It has now been found that influenza surface antigens, such as HA and NA, and endotoxin can be separated by addition of a detergent and then filtering the resulting mixture through a molecular weight cut-off (MWCO) filter having a pore size such that the antigen remains in the filter retentate whilst the endotoxin passes through the filter with the filtrate. It is envisaged that this technique will also be applicable to other pharmaceutical products and in particular drugs and vaccines which contain polypeptide chains or have an amphiphilic structure.

Accordingly, the invention provides a method of removing bacterial endotoxin from a pharmaceutical process solution which method comprises treating the solution with a surfactant effective to dissociate the endotoxin from a pharmaceutical drug or vaccine substance in the solution, and then filtering the solution through a molecular weight cut-off filter having a pore size effective to retain the pharmaceutical drug or vaccine substance but allow the dissociated bacterial endotoxin to pass therethrough.

The term "pharmaceutical process solution" as used herein refers to any pharmaceutical drug substance-containing or vaccine antigen-containing solution that may be produced as part of the pharmaceutical manufacturing process, up to and including the final product. The pharmaceutical process solution may, for example, be a solution of partially purified vaccine antigens from which cell debris or unwanted materials from a microbial particle (e.g. a viral particle) or cell have been removed.

The pharmaceutical substance can be any pharmaceutical drug or vaccine substance which is susceptible or potentially susceptible to endotoxin contamination during its manufacturing process. For example, the substance can be a vaccine antigen or a therapeutic polypeptide product such as a peptide hormone, or a blood product. In particular, the therapeutic peptide product can be a recombinant peptide product. The terms "peptide" and "polypeptide" as used herein embraces substances which consist solely of peptides, as well as substances such as glycoproteins which comprise peptides in association with or bonded to other moieties such as saccharide groups.

The substance will typically be one which associates with the endotoxin, for example forming a complex. Non-peptide drugs which have amphiphilic structures, and which consequently may form complexes with endotoxins, may also benefit from the method of the present invention.

The surfactant can be a cationic, non-ionic, zwitterionic, amphoteric or anionic surfactant, provided that it is effective to dissociate the endotoxin from the antigen whilst not adversely affecting the properties of the antigen and in particular its ability to be retained on the MWCO filter. Preferably, however, it is an ionic surfactant and in particular an anionic surfactant.

Particular surfactants are anionic surfactants having a steroidal structure, and in particular the surfactants corresponding to or analogous to bile salt acids. Thus, for example, the surfactant can be a suitable salt of deoxycholate, cholate, glycocholate, taurodeoxycholate or taurocholate for example. Salts of deoxycholate (DOC) are presently preferred.

Other examples of anionic surfactants include long chain alkyl and alkylaryl sulphonates such as alkyl-benzenesulphonates and dodecylsulphonate.

The concentration of surfactant will be one which is effective to dissociate the endotoxin from the antigen whilst not having an adverse effect on the antigen or its ability to be retained on the MWCO membrane. Preferably the concentration of the surfactant is at least equal to its critical micelle concentration (CMC), and more preferably is greater than its critical micelle concentration. It is presently most preferred to use concentrations of surfactant of one and a half to five times, for example two to four times, the critical micelle concentration.

The surfactant is most preferably one which can readily be removed from the product, for example by dialysis techniques.

The molecular weight cut-off (MWCO) filtration membrane is selected so as to have a molecular weight cut-off which enables retention on the membrane of the desired product antigens, but which allows the endotoxin and surfactant to pass through the membrane.

For example, with influenza antigens, a 100 kD MWCO membrane can be employed. This will retain the majority of the haemagglutinin antigens which typically exist in a trimeric form having a molecular weight of approximately 230 kD.

It is possible that the endotoxin forms micelles or aggregates with the surfactant and, in such circumstances, the surfactant should be such that the total molecular weight of any micelles formed with the endotoxin is less than the molecular weight cut-off of the filter membrane.

The membrane can be, for example, a regenerated cellulose acetate membrane, or a polysulphone membrane. A particular example of a membrane suitable or use in the method of the present invention is the 100 kD MWCO cellulose acetate Millipore membrane.

The method of the present invention has been found to be useful for separating bacterial endotoxin from the influenza viral surface antigens, but it is envisaged that the method will also find application with other viral, bacterial, protozoal and parasitic vaccine antigens, and in particular other haemagglutinin antigens, as well as other polypeptide substances and other drug substances in which endotoxin contamination is a problem.
 

Claim 1 of 12 Claims

1. A method of removing bacterial endotoxin from a pharmaceutical process solution containing an amphiphilic pharmaceutical drug or vaccine, wherein said method comprises: a) treating the pharmaceutical process solution with a concentration of sodium deoxycholate that is effective to dissociate the endotoxin from the amphiphilic pharmaceutical drug or vaccine in the pharmaceutical process solution without affecting the ability of the drug or vaccine to be retained by a molecular cut-off filter having a pore size effective to retain the amphiphilic pharmaceutical drug or vaccine substance but allow the disassociated bacterial endotoxin to pass through; b) directly thereafter filtering the treated pharmaceutical process solution though a 30 kDa molecular weight cut-off filter; and c) thereafter, subjecting the filtered pharmaceutical process solution to a further process step in which the sodium deoxycholate is removed, wherein after this process step the amount of sodium deoxycholate remaining in the pharmaceutical process solution is less than 0.002%.
 

____________________________________________
If you want to learn more about this patent, please go directly to the U.S. Patent and Trademark Office Web site to access the full patent.