PRODUCTION OF RECOMBINANT H-AT-III WITH MAMMALIAN-CELL CULTURES USINGCAPILLARY ELECTROPHORESIS FOR PRODUCT MONITORING

The importance of mammalian cell cultures for biotechnological product ion processes is steadily increasing, despite the high demands of thes e organisms on their culture conditions. Efforts towards a more effici ent bioprocess generally concentrate on maximizing the culture's life time, the cell number, and the product concentration. Here recombinant BHK 21 c13 cells are used to produce rh-AT III, an anticoagulant of h igh therapeutic value. The influence of the process mode (batch, repea ted batch, continuous perfusion) and the process temperature (30 degre es C vs. 37 degrees C) on the above mentioned parameters is investigat ed. It is possible to increase the length of the culture from 140 h (b atch) to more than 500 h (continuous perfusion culture), while concomi tantly increasing the cell density from 0.72 10(6)/ml (batch) to 2.27 10(6)/ml (repeated batch) and 2.87 10(6)/ml (continuous perfusion cult ure). The accumulation of toxic metabolites, such as lactate, can be c urtailed by reducing the bioreactor temperature from 37 degrees C to 3 0 degrees C during the later part of the exponential growth phase. Fas t and reliable product monitoring became essential during process opti mization. Capillary zone electrophoresis (CZE) in uncoated fused silic a capillaries was studied for that purpose and compared to the standar d ELISA. Under optimized conditions an AT III quantification could be done within 2 min with CZE. The detection limit was 5 mu g/ml. A relat ive standard deviation of less than 0.9% was calculated. The detection limit could be lowered by one order of magnitude by using a two dimen sional system, where an liquid chromatographic (LC) system is coupled to the CZE. Concomitantly the resolution is improved. The two-dimensio nal analysis required 5 min. Membrane adsorbers (MA) were used as stat ionary phase in the LC-system, to allow the application of high flow r ates (5-10 ml/min). The correlation between the LC-CZE analysis and th e standard AT III-ELISA was excellent, with r(2): 0.965. Using the ass ay for at line product monitoring, it is shown, that the process tempe rature is of no consequence for the productivity whereas the process m ode strongly influences this parameter.