| Several alternate flowsheets for the production of tissue plasminogen activator (t-PA) in Chinese hamster ovary (CHO) cells were simulated using "BioProcess SimulatorTM" (BPS) of Aspen Technology Inc. The flowsheet developed for typical conditions reported for laboratory-scale production and purification of t-PA was defined as the base case. The alternate flowsheets differed in the fermentation media, type and mode of operation of the bioreactor and scale-up strategy. A trial and error method was used to design the downstream processing train so that the required purity and specified overall recovery are met within a predetermined batch time. The simulator was also used to carry out economic evaluation of process alternatives. Overall economic performance was measured in terms of return on investment (ROI) and gross margin (GM). The flowsheet that used a fed-batch mode of operation for the bioreactor was found to result in the largest ROI. This was followed by a multiple bioreactor flowsheet, where the downstream train was much smaller compared to the base case. A two-stage cell culture process increased the ROI of the process by 44% whereas, the ROI increased only marginally by replacing the stirred tank bioreactor with an airlift bioreactor.; Simulation of the base case was also carried out with "SuperPro DesignerRTM" of Intelligen Inc. The simulation was found to be simpler and faster using the short cut unit operation models. However, due to flexibility and thoroughness of the simulation, as well as analytical ability, BPS was found to be more suitable for a detailed process design. On the other hand, contrary to experimental observations, the rigorous models of BPS for some units (e.g. chromatography) generated results which were found to be insensitive to variations in process parameters.; To backup the simulation results, a sensitivity study of the process performance to uncertain parameters and variables was carried out using BioProcess SimulatorTM. The membrane separation units were found to be most sensitive to flow velocity and exponent 'b', that accounts for dependence of gel resistance on Reynolds number. Both of these quantities affect the thickness of the boundary layer, thereby reducing gel resistance. For the chromatography columns, solute capacity was found to be the critical parameter. The performance of the affinity chromatography was sensitive to the choice of the Langmuir isotherm constant, whereas the ion exchange column was sensitive to charge on the t-PA molecule. Resolution in the gel filtration column was most significantly affected by column length and to a lesser extent, by load volume and flow rate within the range studied. Among the economic parameters, ROI was found to be most sensitive to changes in equipment cost, followed by raw materials cost. A large increase in ROI was observed with improvements in overall recovery and could serve as an incentive to optimize the operation of the processes. |
