APPROPRIATE MAMMALIAN EXPRESSION SYSTEMS FOR BIOPHARMACEUTICALS

Process development for biopharmaceuticals is dictated by product qual ity, drug safety and economy of the manufacturing process. Not surpris ingly, these factors also play a key role in the evaluation of mammali an cell expression systems to be used in the production of pharmacolog ically active glycoproteins. To date, the most prominent candidates fo r efficient expression of glycoproteins are mammalian cell lines such as mouse fibroblast cells (C 127-BPV), Chinese hamster ovary cells (CH O-DHFR, CHO-NEOSPLA, CHO-GS), mouse myeloma cells (NSO-GS) as well as transgenic animals carrying c-DNA or genomic DNA which codes for the p rotein of interest. The expression titer in the case of glycoproteins is mainly determined by the promoter construct, the site of integratio n into the chromosome, the copy number and the type of protein in ques tion. Based on expression titer, CHO-NEOSPLA and NSO-GS expression sys tems are most effective in the production of monoclonal antibodies and , to a lesser extent, of recombinant DNA derived proteins. However, ba sed on overall product yield, expression of recombinant DNA derived pr oteins in transgenic animals is by far the most promising system. Ther efore, for proteins required in large quantities, transgenic expressio n systems offer an attractive choice. However cost of goods for produc ts for which the dosage or the overall annual quantities are low is do minated by downstream processing, filling, lyophilization and packagin g and not by the fermentation process. Such proteins are preferentiall y produced by classical mammalian eel culture systems. Concerns which have to be addressed with respect to drug safety in the transgenic ani mal approach are the size of the herd, genetic stability from animal t o animal, variation in productivity and in impurity profiles during la ctation periods, microbial, viral, mycoplasma and prion contaminants, the dependence on health status and the life span of the animal. In a number of cases glycosylation of the protein is relevant for the preve ntion of immunogenicity of the protein, the pharmacological activity, the pharmacokinetic profile, solubility and stability against proteoly sis. The glycosylation pattern, depending on protein structure, is inf luenced by the enzymatic system of the host cell as well as by ferment ation conditions. Therefore, selection of host cells and culture condi tions must take into account the requirement for a specific and stable glycosylation pattern. For the assessment of glycovariants, a number of protein analytical methods such as peptide mapping, isoelectric foc using, oligosaccharide mapping, MALDI-TOF (matrix assisted laser desor ption mass spectrometry-time of flight), capillary electrophoresis and specific potency assays an available. In our experiments, glycosylati on of proteins expressed in CHO cells was demonstrated to be very stab le. Only extreme process times, cultivation methods and ammonium ion c oncentrations had an influence on the glycosylation profile. Among the three products investigated tissue plasminogen activator (t-PA), inte rferon omega and soluble intercellular adhesion molecule (s-ICAM) - t- PA expressed the most stable glycosylation pattern. Only at extreme am monium concentrations an increase of mannose-5 structures was observed , whereas biantennary complex structures were reduced. On the other ha nd, interferon omega and s-ICAM showed greater susceptibility to incre ased ammonium concentrations and to adherent cultivation. Such conditi ons induced quantitative changes to the glycosylation pattern favoring the appearance of higher branched structures. Short cultivation times resulted in more heterogenous oligosaccharide structures. Since the g lycosylation of the three proteins is different in the same host cell, the amino acid sequence of the protein apparently influences the glyc osylation pattern and its sensitivity to culture conditions. In NSO-mo use myeloma cells, production of s-ICAM is two times as high as in CHO cells with titers reaching up to 410 mu g/ml. However, the glycosylat ion is totally different. Oligosaccharide content is lower and more he terogenous. N-Acetylneuraminic acid content is low and the amount of N -glycolyneuraminic acid content comparatively high. Based on the diffe rent oligosaccharide composition it can be expected that glycoproteins produced in NSO cells will have a higher clearance rate due to bindin g to asialo receptors. Since glycoproteins with N-glycolylneuraminic a cid are only expressed at the embryonic stage in humans, this monosacc haride can be expected to be potentially immunogenic in adults. Consid ering all these relevant aspects, CHO cell expression systems combined with an efficient metabolic design for the culture media seems to be the first choice for glycoproteins required in kilogram quantities. Fo r proteins where cost of goods are on the critical path, transgenic ex pression systems might offer an advantage, provided that drug safety i ssues are properly addressed. If no proprietary expression system of a dequate economy is available, down payments and royalties for the lice nsing of the technology have to be included in the overall economics o f the manufacturing process.