June 15, 2005 (Vol. 25, No. 12)
Achieving Better Process Development Outcomes
Improved process development technology can positively impact process development outcomes in three ways. First, process development technology with increased throughput allows the evaluation of more process conditions, increasing the probability of finding superior culture processes.
Second, process development technology with increased capabilities may allow for an improved understanding of a culture process, or enable new, improved culture processes. Third, it may consume less resources (e.g., time, money, space), increasing the resources available for process development. Therefore, scientists should use optimal process development technology when possible.
Bench-scale (less than 20 L) stirred tank bioreactors (STBRs) are a work-horse process development platform. Bench-scale STBRs are frequently used due to their ability to be operated in a manner analogous to the production scale (e.g., controlled dissolved oxygen and pH), thus minimizing unforeseen problems during scale up.
In addition, instrumented STBRs produce more data, thus enabling greater process understanding, compared to non-instrumented systems (e.g., shakeflask).
Traditional vs. CellFerm Pro STBR Design and Capability
Traditional bench-scale STBRs are generally operated analogous to production-scale vessels, meaning independently (Figure 1). Control actions for each reactor are determined via a programmable logic controller (PLC) executing simple PID algorithms.
Ancillary equipment integration, or the development of a supervisory control and data acquisition system (SCADA) for multiple SBTRs, frequently requires custom third party hardware and software. Without a SCADA program, data from parallel cultures can only be compared after manual compilation.
In contrast, the CellFerm Pro STBR system, from Dasgip (Jlich, Germany), was designed specifically for parallel operation (Figure 2). The CellFerm Pro system uses a Windows XP-based SCADA program. This program allows simultaneous set-up, calibration, operation, and shut down of up to eight STBRs. Due to the presence of a personal computer, advanced control schemes are possible.
For example, the program can calculate oxygen uptake rate (OUR) using predicted influent gas composition and culture volume and dissolved oxygen (DO). Feed profiles or OUR-based feeding can be performed. Process events, such as pH or DO values or trends, can trigger control actions, such as feeds with variable time delay.
The SCADA program also simplifies data analysis. Data from all reactors are plotted in real time. It is also possible to export spreadsheets containing all online information, including process variables, set points, control parameters, calibration data, and process events.
Finally, batches are executed via guided work flows, and most procedures have step-by-step instructions, which can significantly reduce the learning curve for inexperienced operators or those unfamiliar with process controls terminology.
The hardware capabilities of a CellFerm Pro and traditional STBR system are similar. Pressurized reference gel pH, polarographic DO, and platinum RTD temperature probes are used. Thermal mass flow controllers are used for air, O2, CO2, and N2 addition. Per vessel agitation and temperature control is possible. Liquid addition is achieved via peristaltic pumps.
However, the CellFerm Pro system is distinguished by allowing up to four, precisely controlled, continuously variable peristaltic pumps per vessel, as opposed to traditional STBRs, which generally use fixed speed, on/off, peristaltic pumps. Precise pump estimates are used to estimate culture volume for OUR calculations and to prevent vessel overflow.
Finally, the CellFerm Pro is distinguished from traditional STBR systems in terms of size and cost. A traditional 3-L bioreactor system, using jacked vessels equipped with stainless steel headplates and mechanically coupled impellers, can cost three to four times as much, and require two to three times the bench space, per vessel, compared to a CellFerm Pro system equipped with 1-L magnetically coupled paddle agitated spinners, when installation costs and ancillary equipment are considered.
materials and methods The CellFerm Pro was evaluated along side a traditional 3-L STBR system by running identical experimental arms simultaneously on both systems. An E1A-complementing suspension cell line was infected with an E1A-deficient recombinant adenovirus at a specified cell concentration.
Cell concentration and viability were determined offline via CEDEX (Innovatis; Bielefeld, Germany) automated trypan blue dye exclusion cell counting. Adenovirus was harvested by detergent-based cell lysis, and virus concentration was determined by anion exchange chromatograph with UV/Vis detection.
The CellFerm Pro system used four modified 1-L paddle agitated spinner flasks with electric heating blankets. The traditional system used 3-L jacketed vessels with stainless steel headplates and mechanically coupled impellers. Both systems had similar per vessel hardware capabilities: DO, pH, and temperature probes, O2, N2, CO2, and air mass flow controllers, heating, agitation, and liquid addition control.
results Experiment 1 consisted of two arms: A and B. One traditional 3-L bioreactor and one CellFerm Pro bioreactor were used for each arm (Figures 3a, 3b).
Experiment 2 consisted of two arms: A and B. Duplicate traditional 3-L bioreactor and duplicate CellFerm Pro bioreactors were used for each arm (Figure 4a, 4b).
discussion The minor experimental discrepancies between the traditional 3-L bioreactor and the CellFerm Pro system were unexpected. Offline pH, DO, and temperature values were compared, but revealed no significant differences. Known differences included material of construction, surface area to volume ratio, and agitation rate.
The exact reasons for differences in peak viable cell concentration and adenovirus titer remain unknown. However, the magnitude of the experimental differences were small, and both culture systems selected the same optimal culture process for both experiments, indicating the CellFerm Pro should be useful.
The CellFerm Pro software increases throughput by allowing simultaneous operation of up to eight vessels. For experiments using more than eight vessels, multiple CellFerm Pro systems must be used, which diminishes the benefits of parallel operation.
The reuse of probes and vessels significantly contributes to experiment turn around time and workload for both of the STBR systems. Both systems require probe maintenance and calibration, decontamination, disassembly, cleaning, reassembly, sterilization, and set up.
The development of disposable vessels with reliable sensors is required to further increase STBR throughput and significantly reduce workload. For both systems, a double set of vessels is recommended to maximize throughput; one set of vessels is set up while the second set is in operation. Technicians should be used for equipment operation to allow scientists to focus on process development.
A quantitative comparison of bioreactor system throughput (e.g., vessel turnover time) was not possible; a refined vessel turnover strategy had been developed for the traditional STBR system, but not for the CellFerm Pro evaluation system.
The CellFerm Pro system has advantages over traditional, independently operated STBRs, including parallel operation (increased throughput), increased capabilities (e.g., advanced control schemes), reduced cost and space requirements, and increased ease of use, which may improve process development outcomes.
A CellFerm Pro system will most greatly benefit labs having limited resources, or those looking to acquire STBR capabilities. Resource-rich labs using large numbers of traditional STBRs and customized, sophisticated SCADA programs with equipment integration will benefit less.
The choice of optimal process development technology is ultimately determined by process development strategy. However, knowledge of available process development equipment is required if the best system is to be chosen.