BACKGROUND:CAMPATH-1 Abs have been used for T-cell depletion in stem-cell transplantation since the early 1980s. During that time there has been substantial progress in manufacturing techniques and quality control procedures. This article summarizes the methods used to produce the Abs for clinical use and describes results of quality control tests on representative batches.METHODS:Rat hybridoma and recombinant CHO cells were cultured in hollow-fiber fermentors. Antibodies were purified from the culture supernatant by fractionation with ammonium sulphate, or by column chromatography. Additional steps were added to assure the removal of DNA and viruses. A range of analytical methods was used to characterize the antibodies. Samples were stored frozen at -70 degrees C and re-analyzed many years later to assess the long-term stability.RESULTS:Hollow-fiber fermentors provided a simple and reliable means for antibody production, with yields between 3-10 mg/h and a convenient concentration for further processing (0.6-2.0 mg/mL). All of the CAMPATH-1 Abs (rat IgM, rat IgG2b and human IgG1) could be purified by affinity chromatography on Protein A, but the low pH required for elution caused unacceptable aggregation of the IgM. CAMPATH-1H contained approx. 20% dimeric IgG, which could be removed by size exclusion chromatography. Antibodies were stable for at least 6 years at -70 degrees C, but there was unacceptable aggregation of CAMPATH-1M in one batch stored for 9 years.DISCUSSION:Pilot-scale production of MAbs for clinical studies is feasible in a small academic center, but regulatory requirements now demand that great attention is paid to all aspects of manufacturing and quality assurance. Although the underlying principles of cell culture and protein chemistry remain the same, the level of documentation, validation and quality control has increased greatly over the last 20 years.

Bacterial endotoxin is probably the most common significant contaminant that might be found in antibody preparations. It is found ubiquitously in normal environments but its pyrogenic effects in vivo can be lethal. It is absolutely vital to control the level of endotoxin in therapeutic products, but its significance for experimental work must not be underestimated since it can have numerous confounding effects both in vivo and in vitro. Methods for removing endotoxin from products have been described (1-4), but in our experience, it is much better to avoid it from the beginning by scrupulous control of raw materials and use of good aseptic technique. To grow, bacteria need water; therefore, the most likely sources of endotoxin are water and any process equipment that has been wet. Water must be obtained from a controlled source that is low in endotoxin. If you do not have a suitable dedicated water purification system, then it is best to purchase purified water. Water for irrigation (from a pharmaceutical supplier) is possibly the most economic. Equipment in contact with cells or products should preferably be sterile disposable plastic. Standard tissue culture ware is usually very reliable. Plastic bags for media and process intermediates are now widely available in all sizes (e.g., Stedim, Aubagne, France) and should be used in preference to glass bottles. Silicone tubing (food or medical grade) is suitable for all fluid transfers and should not be reused. If reusable equipment is essential, it can be soaked in 0.5 M NaOH and/or baked in an oven at high temperature (steam sterilization is not sufficient) (1).

Ideally, injectable drugs are sterilized in their final containers by a foolproof method like autoclaving. This is not possible for biologicals like monoclonal antibodies (mAbs), so they must be manufactured aseptically, sterilized by filtration and then filled into sterile vials or ampoules. The final filling procedure is the most critical aseptic process and should be done in a very clean environment. Automatic machines are used for large production processes and eliminate the risk of contamination associated with manual processes. However, preparing material for early clinical trials can be problematic because the batch size is normally too small for a filling machine (e.g., 500-1000 vials). Normal practice is to fill this number of vials by hand, but the vials and closures have to be washed, depyrogenated (by baking in an oven), and sterilized, and the filling has to be carried out in a very strictly controlled environment, because the vials are open throughout the process and are only stoppered and sealed in a second step.

During every stage of development and production of diagnostic or therapeutic antibodies, it is necessary to have an assay to measure antibody concentration. Several techniques are routine in virtually all antibody laboratories. High-performance liquid chromatograpy (HPLC) using an affinity matrix (protein A or protein G) is rapid and quantitative and measures most types of antibody, though the equipment is quite costly. Enzyme-linked immunosorbent assay (ELISA) is widely used and extremely versatile. By judicious choice of anti-Ig reagents, specific for heavy chain, light chains, or particular domains, it is possible to screen for almost any desired Ig molecule or fragment. This is specially useful when analyzing a complex mixture (1). Native gel electrophoresis is a useful technique for screening relatively concentrated samples (e.g., from fermentors) since different antibodies can be readily distinguished by their characteristic mobilities. However, none of these methods are really suited to the rapid semiquantitative testing of huge numbers of samples that is often necessary early in a project (when screening for a rare hybridoma or transfectant, or for somatic mutants) and for routine analysis of process samples during cell culture. Instead we have found that red cell agglutination as originally developed by Coombs (2,3), is convenient, quick, and very cheap.

Protein A attached to an affinity adsorbent was shown to be remarkably resistant to 0.5 M NaOH. Even repeated treatments gave only a small decrease in functional capacity and no adverse effect on leakage of the protein A into the eluate. This simple cleaning procedure should be useful in applications where antibodies need to be purified free from the risk of contamination with endotoxins or micro-organisms, e.g., for in vivo therapy, either in clinical trials or experimental animals. It can also prevent cross-contamination when the same protein A adsorbent is used for different batches of antibody.