Implementing True PAT in Bioreactor Operations: Gaining Optimized Cell-Level “Observability” and Data-Driven Process “Guidance” by Leveraging On-Line technologies such as Dielectric Spectroscopy and TM At-Line Tools such as the Modular Automated Sampling Technology (MAST ) Platform
Authors:
C.
a Pepper,
L.
a Graham,
B.
a Downey,
Abstract
J.
b Germon,
“On-Line” Technologies
In implementing process analytical technology (PAT), biopharmaceutical companies are continually striving to gain a fundamental understanding of what is happening to the cells within their bioreactors and the impact that the process has on the cell’s ability to deliver the target product quality. While implementation of on-line tools like dielectric spectroscopy and Raman are helping to provide insight, there is still a gap integrating the data produced by these techniques with off-line measurements such as cell density, viability, metabolite levels, and titer and presenting that data to the end user that gives them real insight into their process. Bend Research Inc., in collaboration with Pfizer Inc. and other major biopharmaceutical companies, is working to advance the Modular Automated Sampling Technology (MAST) platform, the goal of which is to provide a complete data management solution to the end user. The MAST system prototype covers all aspects of data management from providing aseptically collected bioreactor samples to analytical devices, maintaining the raw analytical data in a database and then providing the user with a Graphical User Interface (GUI) to rapidly modulate the data, providing instantaneous guidance. Early studies with this modular sampling platform have demonstrated that automated at-line measurements are representative of parallel manual samples. During these studies, maintenance of sterility has not been an issue.
D.
a Newbold,
and D.
b Sullivan
(
a Bend
b
Research Inc., Pfizer Inc.)
Enabling Cell & Protein Level “Observability” “Guidance”
“At-Line” Technologies “Process-Level”
“Cell-Level” Dielectric Spectroscopy (DS) Data Correlated to Off-Line Caspase Activation Measurements Enable On-Line Monitoring of Cell Population State As cell viability drops in response to staurosporine addition, transformations in the cell lead to variations in capacitance. By scanning over a frequency range, it was discovered that distinct cell populations could be distinguished during apoptosis. These observations in the capacitance spectra were related to off-line measurements of Caspase 3 activation to provide an on-line measure of cell health in the bioreactor.
This poster illustrates how the MAST platform is incorporated with on-line analytical technologies to deliver overall cell-level “observability.” For example, as a result of this integration, dielectric spectroscopy measurements can be corrected to estimate total and viable cell density more accurately.
Delivering Biotherapeutics: Our Mission Deliver Innovative and Transformative Technologies & Services … To Achieve Target Quality Proteins … through Science & Engineering
Bend Research
Laboratory SP200
Cell Culture
2
30
69
No contamination. Results consistent with manual sampling
Bend Research
Laboratory SP200
Media Challenge
3
98
>1141
No contamination
Pfizer Pilot Plant
30-L Reactor SP100
131
No contamination. Results consistent with manual sampling
113
No contamination. Results consistent with manual sampling
64
No contamination. Results consistent with manual sampling
160
No contamination. Results consistent with manual sampling
244
Overnight and repetitive sampling with no clogging or fouling
403
No contamination. Results consistent with manual sampling
10
No contamination. Results consistent with manual sampling
Situation
Cell Culture
Pfizer Pilot 130-L Reactor Cell Culture Plant SP100 Pfizer Pilot 500-L Reactor Cell Culture Plant SP100
FACSaria instrument (Fluorescence activated cell sorter)
From this new level of insight into the operation of the system, scientists can obtain enhanced data-driven “guidance” for key activities like cell-line selection and optimized process operation. With this new technology, the bioprocess industry can make major advances toward advanced real-time testing, predictive control, and overall enhanced bioprocess design and operation.
Location
Scale / Sample Pilot Model
Number of Runs
4
2
3
Total Accumulated Duration (days)
47
69
36
Total Samples Taken
Outcome
Company A
Pilot-Scale SP100
Cell Culture
Company B
Pilot-Scale SP200
High Cell Density Challenge
Company B
Pilot-Scale SP200
Company C
Pilot-Scale SP100
Company D
Laboratory SP200
Cell Culture
1
16
16
No contamination. Results consistent with manual sampling
Company E
Laboratory SP200
Cell Culture
1
7
4
No contamination
Cell Culture
Microbial
3
6
4
2
40
6
60
10
Good comparison shown between manual and MAST samples taken from the 500-L reg tox pilot-scale runs
Sample Pilot SP100
Sample Pilot SP200
Fixed Reactor Installation - Steam or Liquid Sanitization
Bench-Scale and SUB Reactors - Liquid Sanitization
Pilot bioreactor with two Raman probes installed
Raman Data Correlated to Nova Results Collected At-Line Using the MAST Sample Pilot (SP100)
Conclusions Recent innovations in on-line measurement (dielectric spectroscopy and Raman) and modular, automated aseptic sampling (MAST) have created the conditions needed for true bioreactor PAT implementation. The MAST system is ideal for use in bioreactors. It demonstrates reliable contamination-free sampling with greater sample consistency and reproducibility when compared to manual samples. This scale-independent, lowcost sampling system is capable of frequent sampling to enable more intensive process-control schemes. Savings in labor and process optimization/efficiency can be expected. The highly efficient Sample Pilot modules also have potential use in disposable systems and downstream applications. The combination of advanced on-line measurement with at-line analysis (e.g., dielectric spectroscopy and Raman) creates a synergistic environment for accelerated process understanding. Additional PAT, coupled with cell-based bioreactor models, have also been developed to enhance guidance. This emerging process-development methodology holds promise to shorten development timelines and deliver a higher-quality process that significantly reduces the cost of goods.
Cell Separation Time-series plots of Raman models and NovaFlex data collected through the MAST Sample Pilot. By enabling samples to be collected at frequent, regular intervals without an operator present, we were able to increase our off-line data without any additional labor. Those data were used to build robust process monitoring models and develop the Raman system for laboratory use. The Sample Pilot is an enabling existing technology for implementation of PAT applications.
Acknowledgments: The authors thank J. Bouressa, N. Jenkins, E. Yu, S. Casnocha, J. Weber, T. Barreira, P. Jeffers, and C. Crowley of Pfizer, and T. Wigle, D. Millard, A. Carroll, N. Glutting, P. Glover, B. Russell, L. Davis, and R. Ray of Bend Research for their contributions to this work.
and / or
Example schematic showing potential for sampling from multiple bioreactors to enable “at-line” analyses (in conjunction with “in-line” technologies (e.g., Dielectric Spectroscopy and Raman); Potential application: Process Control
Bend Research Inc. 64550 Research Road, Bend, OR 97701 USA Phone: 1-800-706-8655 © Bend Research Inc. 2014
MAST Sample Pilot (SP100) on a 30-L Bioreactor
In-house testing shows excellent consistency between manual and MAST samples taken from benchscale cell culture runs