Purity tests

The purity of biotechnology products is affected by protein contaminants and degradation products that co-purify with the protein or form during processing and by impurities introduced with the starting materials or during the purification process. The starting materials, including the manufacturer's working cell bank and raw materials used for cell culture or fermentation, can be the source of contaminants such as mycoplasma, bacteria, mold and yeast, bacterial endotoxins, adventitious virus, host cell DNA, host cell proteins, and added chemicals, including antifoams, antibiotics, and inducing agents.

Purification processes for biotechnology products are usually centered around column separation procedures and often include affinity chromatography steps that either provide primary purification of the protein or are included to remove specific contaminants. Since the chemical linkage between an affinity ligand and the solid support is potentially labile, affinity ligands can leach from the chromatography columns and become contaminants in the purified protein. Columns that are used repeatedly for protein purification are a potential source of both mioobial contamination and bacterial endotoxins. In addition, chemicals can be added throughout the purification process, and the absence of these added chemicals in the final product must be demonstrated. These added chemicals can include agents used for cleaning processing equipment, agents and preservatives used for the cleaning and storage of separation columns, and ingredients contained m buffers used m the purification process but that are not desirable in the final product.

The absence of process contaminants and degradation products in the final product is demonstrated by a combination of final product testing and process validation studies. A partial list of potential impurities and the analytical technique used for their detection and quantitation is presented in Table 6.1. Each category is described in more detail in the following sections.

Protein contaminants

The detection and quantitation of protein contaminants and degradation products that may co-purify with the protein of interest require the development of sensitive and specific assays. SDS-PAGE run under both reducing and nonreducing conditions provides a very effective method of separating unknown protein impurities and degradation products from inact protein. Following electrophoresis, resolved impurity bands can be visualized by staining with Coomassie blue. The visualized bands can then be quantitated by densitometry and area normalization or by densitometry and interpolation from a standard curve. Since different proteins have different Coomassie blue staining affinities, the accuracy of the quantitation is dependent on a knowledge of the staining intensity of each impurity and degradation product. Efforts should be made to isolate and identify contaminants and degradation products. Once isolated and characterized, these impurities and degradation products can be used as reference standards during assay validation studies.

In general, SDS-PAGE with Coomassie blue staining has a sensitivity of 0.2-1 , depending on the protein, and provided that this level of impurity is represented in a single resolved band. The quantitative power of this analytical method is limited when multiple diffuse bands are present Sensitivity can be increased by visualizing the resolved bands with silver nitrate. Silver binds to the protems, producmg a black band, but the binding is not stoichiometric. Silver staining is usually 10-100 times more sensitive than Coomassie blue staining, although the increase in sensitivity varies from protein to protein. Quantitation of silver-stained gels is usually not possible because the band intensity response is not linear, especially when carbohydrate is present, and because high background staining is frequently found.

Care must be exercised in quantitating degradation products from SDS-PAGE gels to avoid overestimating the total degradation. An understanding of the degradation pathway is essential so that the origin of multiple species can be established. For example, a single cleavage may give rise to two or more fragments; therefore, the extent of degradation cannot be determined as the arithmetic sum of the amount of each fragment.

The detection of protein impurities at levels below 0.1-0.5 typically requires an enzyme immunoassay or a radioimmunoassay that has been developed to detect a specific contaminant The assay and the assay reagents are customized to the product's manufacturing process. The ability of the test to provide reproducible results is dependent on the lot-to-lot uniformity of the reagents. These immunoassay reagents typically have sensitivities in the range of 1-100 ppm.

ELISAs are frequently employed for quantitating impurities at the ppm level. In these procedures, either an antigen or an antibody against the impurity of interest is immobilized onto a microtiter plate. The test article is added to the wells of the plate so that the potential contaminant can bind to the immobilized antigen or antibody. Following this reaction, a second antibody reagent, consisting of a detector (i.e., an enzyme or a fluorescent tag bound to an anti-impurity antibody)is added to th eplate. The second antibody binds to the immobilized impurity, forming a sandwich with the impurity bound between the two test reagents. Quantitation of the amount of bound impurity is made by reacting me second antibody reagent with a specific substrate to produce either fluorescence or a colored solution whose absorbance can be measured spectrophotometrically. Quantitation is then made by interpolation from a standard curve. For these assays to be truly effective in accurately quantitating impurities, the impurities must be isolated and characterized so that they can be used in reagent preparation and as reference standards in the assays.

Added chemicals

Fermentation, cell culture, and protein purification processes commonly require the addition of various chemicals to aid in fermentation or cell culture, to perform specific separation functions, or to protect against destruction of the protein during purification. The absence of these chemicals in the final product can be demonstrated by end-product testing and by validation studies that quantitate that the process is effective in eliminating these added substances.

Since these added substances are primarily chemicals, a wide variety of sensitive analytical methods, including gas chromatography, atomic absorption spectroscopy, NMR, HPLC, immunoassay, and bioassay, can be employed for final product testing. Expectations are that these added chemicals will be removed to levels near or below the detection limit of the assay. Validation studies employing radio-labeled chemicals are conducted on a small scale to measure removal of each added substance by individual steps in the process. Studies also are conducted during manufacturing runs to quantitate removal at each processing step to confirm by mass balance calculations the complete clearance of these added substances.

Affinity ligands

Purification procedures for recombinant proteins and monoclonal antibodies often include passing the material over an affinity column to achieve a high (>90%) degree of purification in a single step or to remove a defined contaminant The affinity media usually consists of a monoclonal or polyclonal IgG immunoglobulin or a large protein antigen bound to a solid support through a chemical linkage.

Since these linkages may be labile, the protein or antibody may leach from the support and become a contaminant in the protein being purified. Since these contaminating proteins are highly undesirable in the purified product, end-product testing must be able to detect the affinity ligand at the ppm level, and immunoassays are typically employed. Depending on the nature of the affinity ligand, reagents for these assays can either be obtained from commercial sources or produced and quality controlled in-house. In addition to end-product testing, validation studies snowing that purification steps downstream of the affinity chromatography step remove the affinity ligand provide strong supporting data that the final protein product is free of the potential contaminant.

Residual DNA

Concern continues to exist regarding the presence of DNA in purified biotechnology products, since these proteins are derived from transformed cell lines or from genetically altered host organisms. Host cell DNA is found in the starting material (i.e., the fermentation paste, the fermentation supernatant, or the cell culture fluid) and arises from dead and lysed cells. The purification process must be designed to reduce host cell DNA to very low levels, and the extent of the removal should be confirmed by both validation studies and end product testing.

Residual host cell DNA in purified proteins can be quantitated by a DNA hybridization assay using a radio-labeled probe prepared by nick translation of DNA extracted from the host cell or from the transformed cell line. Sample DNA is separated from the protein by extracting the DNA with phenol and then precipitating any remaining protein with alcohol. The DNA is denatured with heat, and the single-stranded DNA is immobilized on nitrocellulose membranes. Following hybridization and autoradiography, residual DNA is quantitated by visual comparison to a series of standards. Although DNA content in biological drug products is evaluated on a case-by-case basis, PDA (CBER) guidelines require the DNA assay to have a sensitivity on the order of 10 picograms DNA per dose (1-3), and the World Health Organization (WHO) guidelines require less than 100 picograms per dose [8].

Bioburden and bacterial endotoxins

Since most therapeutic biotechnology products are administered parenterally, the purified protein must be free of bacterial endotoxins and other pyrogenic substances. Bacterial endotoxins in protein products can be quantitated using the limulus amebocyte lysate (LAL) assay. Lysates with sensitivitiesof0.03EU/mLarerecommendedforuse with protein products [9].The rabbit pyrogen test as described in 21 CFR 610.13(b) [10] and in USP, Chapter <151>, "Pyrogen Test", is used for end-product testing to ensure freedom from pyrogenic substances.

The validation of the removal of bacterial endotoxins from the starting materials by the purification process can be demonstrated on a small scale by spiking column loads and process intermediates with endotoxin and quantitating clearance, or on a production scale by monitoring endotoxin clearance at each process step and verifying overall clearance by mass balance calculations the complete clearance of these added substances.

Cultivable and noncultivable mycoplasma

Since both recombinant proteins and monoclonal antibodies are produced by cell culture methods, it is possible that the cell culture fluid can become contaminated with mycoplasma. The absence of mycoplasmain me purified protein must be demonstrated by end-product testing, and the purification procedure must be validated to demonstrate the clearance of mycoplasma. The absence of cultivable mycoplasma in the purified protein can be demonstrated by broth and agar culture under both aerobic and anaerobic conditions using the procedure described by CBER[2]. Freedom from noncultivable mycoplasma can be demonstrated by bisbenzimidazole fluorochrome staining following incubation of the test sample with Vero cells [2]. A preference for testing unfiltered and quick-frozen/single-thaw samples has been recently expressed by CBER.

Validation studies can be used to support periodic rather than lot-by-lot testing for mycoplasma. These studies involve the spiking of starting materials and/or column load materials with high titers of mycoplasma, simulating the purification steps on a small scale, and calculating mycoplasma removal/inactivation factors. Purification processes should show the clearance (inactivation plus removal) of at least six logs of mycoplasma.

Adventitious agents

An area of great concern at the present time is demonstrating that the purified protein is free of adventitious agents, especially viruses. This must be demonstrated by both end-product testing and by validation data that show the removal of viruses to levels that no longer present a risk factor. The assessment of risk is evaluated on a case-by-case basis and is based in part on me patient population, on the dose level and regimen, and on the health status of the patients.

The identification and quantitation of viral particles in the master cell bank (MCB) and the manufacturer's working cell bank (MWCB), and the identification of infectious virus produced by the MCB and the MWCB are of primary importance. Viral particle enumeration in the MCB/ MWCB can be accomplished by thin section electron microscopy, although the sensitivity of this method is Id^-lO8 particles/mL. Infectious retrovinis can be detected by reverse transcriptase testing, direct assay, or co-cultivation of MCB/MWCB cells with detector cell lines known to support the replication of viruses.

If infectious virus can be detected, then end-product testing for those virus(es) can be performed using an appropriate in vitro cell-based plaque or focus assay. Viral assays that first employ amplification of the virus and then detection in a cell culture system are more sensitive than direct cell culture assays and thereby provide greater assurance of the absence of infectious virus.

Validation studies are needed to demonstrate the removal and inactivation of virus by column purification steps and by processes and/or formulations that are designed specifically to inactivate virus. If infectious viruses have been identified, they can be used in small-scale studies to measure clearance and inactivation. Similarly, radio-labeled viral particles prepared by culturing the MWCB in media containing radiolabeled nutrients can be used to measure and define particle removal. Typically, logs of viral clearance (inactivation plus removal) are expected for processes used for the purification of a biotechnology protein product.