Running a fermentation process is a bit like walking a tightrope. It is a balancing act involving monitoring and control of physical and analytical parameters, and in the case of fed-batch mode operations, also regulating substrate feeding. If you maintain your poise throughout the run, the results are high fermentation yield in the quickest possible time. Wobble, and yield reduces and run-time increases. Stumble, and the batch can be lost. For some APIs loss of a 15,000-litre batch can amount to US$1m.
Monitoring and controlling analytical parameters is central to successful fermentation. Analytical sensor failure during the run demands labour-intensive, off-line lab tests of grab samples, and the possibility of corrective actions that may come too late to save the batch.
On top of that, the risk of human error in equipment operation is ever present. Up to three analytical parameters (pH, dissolved oxygen and dissolved CO2) in combination with redundant sensors must be handled properly. If each sensor is connected to a single transmitter the installation becomes very complex and needs a significant amount of panel space. Such set ups lead easily to human error if operating procedures are different from transmitter to transmitter.
Importance of in-line process analytics during fermentation
Part of the FDA’s role is to ensure that the pharmaceutical industry improves and modernises pharmaceutical manufacturing processes. This is the purpose of the Process Analytical Technology (PAT) initiative. The more recently launched International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) initiative, Quality by Design (QbD), is also aligned with the PAT concept. PAT plays a key role in creation of a robust control strategy for production processes that reduces the possibility of operator error.
Relevance of key process parameters
APIs produced via biofermentation are extremely sensitive to the manufacturing process. Lot-to-lot variability in product quality is commonly observed even when fermentation has been performed using exactly the same process; and variability has also been known to affect final product quality. In order to minimise batch inconsistency, analytical parameters are measured and controlled. The goal is to keep these parameters at a constant setpoint or on a predefined trajectory. However, as a run progresses, cell metabolism and accumulation of metabolic side products may alter the fermentation medium in an unpredictable way, necessitating corrective action.
pH, oxygen, and carbon dioxide
Of the three most commonly controlled analytical parameters (pH, dissolved oxygen and dissolved CO2) pH is the most critical because enzymatic activities and cellular metabolism are sensitive to pH changes. Therefore, optimal cell growth conditions are heavily dependent on maintaining the correct pH. A reliable pH measurement allows not only monitoring but also closed-loop control to be implemented.
In-line sensors for continuous, real-time pH measurement have been quite common for many years but are not without their risks. Should a pH electrode fail during a batch, pH must be measured with off-line samples, which carries risk of contamination, and is labour and time intensive.
Cell cultures require dissolved oxygen for the production of energy from organic carbon sources, e.g. glucose. Given oxygen’s poor solubility in water, the control of air flow is carefully regulated to ensure it does not become a rate-limiting factor in the process. Overdosing dissolved oxygen is not harmful to the micro-organisms, but is a waste of energy used to run air compressors. The signal from an oxygen sensor can be used to build-up closed-loop control for the air supply.
Dissolved carbon dioxide levels can be indicative of the quality of cellular metabolism and are thus routinely monitored as an indicator of culture performance. High CO2 levels can inhibit growth and metabolism and can affect product quality characteristics, such as glycosylation of the protein product. In fed-batch mode the CO2 signal allows control of medium fed to maintain the CO2 level in an optimal range. Yield increases of up to 40% are possible from implementing a dissolved CO2 control strategy.
Multi-parameter transmitters
As mentioned above, monitoring three parameters on three separate transmitters with different interfaces can be a challenge. In addition, sensor failure during a run is inconvenient at best, and leads to batch loss at worst.
Multi-parameter transmitters go a long way to reducing human error. The ability to configure and calibrate sensors for three different parameters on a single transmitter not only simplifies measurement system installation (and reduces costs), it also cuts the training requirement for operators and hence the possibility of mistakes being made.
However, use of a standard multi-parameter transmitter by itself does nothing to lessen cumbersome, risky, and time-consuming calibration procedures. Nor does the transmitter provide information as to when sensor calibration or replacement will be required. Operators therefore have to install a sensor that they hope will survive the next batch run or, for the sake of process safety, discard a sensor that may in fact be perfectly useable. METTLER TOLEDO has found a solution to these problems.
Digital analytical technology for higher performance and safety
METTLER TOLEDO’s Intelligent Sensor Management (ISM) is a technology for analytical sensors that is able to predict when a sensor will next require calibration and when it may fail. ISM is a digital sensor handling and maintenance concept designed to simplify sensor operations.
In the head of an ISM sensor a microprocessor runs algorithms that use data from the sensor (past and present process measurement data, as well as information on the sensor’s physical condition) to generate predictive diagnostics tools. When an ISM sensor is connected to an ISM-equipped transmitter, or a computer running METTLER TOLEDO’s iSense Asset Suite software, the diagnostics tools show when sensor maintenance or replacement will be due. In the case of a pH electrode that is close to end-of-life, the tools will alert the operator to this fact, preventing the electrode being used for the next batch.
A further useful feature of ISM is the ability to calibrate sensors accurately in a convenient location. Whereas conventional sensors require calibration at the measurement point, involving bringing buffer solutions into a clean zone, ISM sensors can be calibrated in a lab or maintenance shop using iSense software and a USB connection. Calibrated ISM sensors can be stored until required, and quickly swapped at the fermentation reactor for fast measurement point start up. Sensor pre-calibration and Plug and Measure functionality reduces the risk of human error, further improving process safety.
ISM sensors also record their exposure to temperature, and SIP and CIP cycles. This data, plus the calibration data, is stored on iSense for traceability and compliance purposes.
The M800 is an ISM-equipped, multi-parameter, multi-channel transmitter. Its large, touchscreen display shows data from up to four sensors. It also includes the iMonitor, a view of ISM’s predictive diagnostics using traffic light colour coding to warn operators of sensor problems, and information on how to rectify an issue before measurement performance and the fermentation are affected. With its highly organised menu structure, navigation on the transmitter is easy and intuitive.
Conclusion
Efficient fermentation operations involve continuous monitoring and control of various process parameters. Complex reactor set-ups carry an increased risk of operator error. METTLER TOLEDO’s Intelligent Sensor Management sensors with an M800 transmitter significantly reduce this risk by providing predictive sensor diagnostic tools and a unified user interface.