(583h) Automated Bioreactor Feedback Control Via Mast-Nova-Deltav Integration | AIChE

(583h) Automated Bioreactor Feedback Control Via Mast-Nova-Deltav Integration

Authors 

Lowe, C. - Presenter, University of Massachusetts-Amherst
Houle, Z., Takeda Pharmaceuticals
Karmiy, L., Takeda Pharmaceuticals
Dhamal, M., Takeda Pharmaceuticals
Harmon, R., Takeda Pharmaceuticals
Process analytical technologies (PAT) have introduced the ability to obtain rapid, detailed, on-line information about processes and products throughout manufacturing and development. While on-line process and analytical readouts greatly reduce traditional turnaround times, an ultimate goal would be utilizing these outputs to control and optimize in-process conditions; positively influencing productivity or product quality profiles. Additionally, implementation of feedback control mechanisms between at line analytical equipment and bioreactor control systems can offer substantial reductions in operational burden through automated sampling and feeding, allowing for increased sample/feed frequency, or expanding the number of bioreactors that operators can manage at once.

Herein we report the implementation of closed loop feedback control for automated feeding, gassing adjustments, and pH adjustment utilizing the existing bioreactor control system (Delta V), cell culture analyzer (Nova BioProfile® Flex 2), and automated MAST® sampling system present in our process development laboratories. Automated delivery of cell culture samples from the bioreactor to the Nova Flex2 was made possible by installation of the MAST system. Typically, following manual sampling and use of an offline feed calculator, the operator would manually input feed volume into the DeltaV bioreactor control software for each feed. Additionally, offline pH and pCO2 values would be compared to action criteria and the operator would perform manual adjustment if necessary. These manual operations have been eliminated by integration between the Flex2 and DeltaV, allowing DeltaV to identify automated sample results from the Flex2. With these results, DeltaV can calculate and automatically deliver the required volume of feed to the bioreactor, induce a gassing adjustment to control pCO2 levels, or automatically align online pH reading to the offline analyzer. This automated workflow has been put into practice for both fed batch and perfusion bioreactor studies in our process development laboratories across multiple programs under development. Consistent bioreactor performance has been observed for both traditional manual bioreactor operations and these new automated workflows. A high success rate has regularly been observed for automated operations allowing for >90% reduction in hands on time required for an operator to perform routine sampling and feed activities.

While ongoing research and new developments in the PAT space will continue to expand the ability for direct feedback control to be implemented more widely within the industry, this work demonstrates that closed loop feedback control can be implemented with existing laboratory infrastructure. Based on this initial work with automated glucose feeding, additional automated strategies are being investigated with the current closed loop feedback mechanism. Additionally, the finer glucose control afforded by this strategy via higher frequency automated sampling and feeding may allow cell culture performance improvements by reducing required residual glucose. Further, similar integration with other on-line analytical equipment will allow for additional metrics and targets for feedback control, particularly as correlations are established between product quality attributes and in-process conditions. Ultimately the implementation and expansion of these new capabilities offer the potential to greatly expand in-process controls while simultaneously reducing operational burden and increasing throughput of small-scale bioreactors in the development space.