Achieving consistently high quality data for upstream development play an essential role in bioprocess science and engineering. Typically, the product and process related engineering strategies are implemented rather late in the pharmaceutical and biopharmaceutical industry.
As a research institution we are indented to provide an appropriate analytical platform applicable early in the process development, generating an improved, in-depth process understanding and knowledge. According its relevance, the concept was established and further developed for a commercial IgG expressed in CHO cells.
With the aim to realize advanced process monitoring by linking critical process parameters with critical quality attributes for improved process understanding, process monitoring concepts and mathematical modelling are provided.
In translational research, we are intended to establish an analytical platform contributing to the regulatory concepts within the PAT - GMP - Initiative (process analytical technology) to push the implementation of Quality by Design (QbD) framework.
The relevant building blocks within this initiative are:
Quality planning
Quality control
Quality improvement
Through these criteria, knowledge is used achieving an optimized risk-benefit ratio for the product quality (quality planning). Quality is controlled by multitude of analyzers (quality control), which should improve the process quality and stability (quality improvement). It is aimed that the manufactures switch from more inflexible towards more flexible processes with enlarged operation space. However, this flexibility is only tolerated since the effect on the product quality is known. The major expected advantages should be the reduction of batch failures, faster product release and the reduction of costs.
Although the QbD approach is meanwhile well established the expected consistent product quality is not comprehensively reached because variation in material for instance, raw material, has not been adequately recognized. The challenge to overcome this concept weakness is the implementation of an advanced process control regime (APC). By the Quality by Control (QbC) strategy it is able to fill this gap, but is still challenging in respect to variable/parameter identification, process setup, data generation and sophisticated process software.
Quality Control Strategy
Due to the long lasting collaboration with the upstream research group of the Institute and various industrial cooperation, we were able to continuously evolve a platform of advanced process analytics. This platform is convenient to provide in-depth information about the process dynamics, is appropriate to measure the analyte and attributes in complex varying matrices. This is achieved by optimized and minimized sample preparation and enables analysis of increasing number of samples that can be reliably analyzed in a reasonable time. Thus, the balance of information needed and analytical effort is found.
For the QbC concept several different information about cell related attributes, media and material characteristics, as incoming criteria as well as in-process controls, but also importantly, product-related attributes is required.
In our research group, we focus an media/material analysis as well as product-related attributes in complex, varying matrices (Figure 1). For this purpose, different methods were established and their reliability confirmed. For methods, which are applied for many years also quality control charts are implemented to control the life cycle quality.
Figure 1
Opportunities & Challenges
Within the process-related advanced analytical platform we see the opportunity to implement new and adapted methodologies for batch and continuous operations with a balanced effort.
The ultimate goal for the future will be online measurements of the critical process- and product attributes, full real time release application, model based quality prediction and continuous processes with advanced integrated concepts.
However, the way to reach this goal, definitely requires optimized analytical platforms. Those must provide high quality data in complex matrices, provide reliable data during process development but also support validation studies of online-measuring-devices. Without these data process models, predictive strategies, process monitoring and automatization initiatives will hamper.
However, for the concept development, extremely high work load and sophisticated analytical investigations are demanded. Automation is partially available but not in sufficient manner. Therefore, we decided to work on sample preparation to make it as simple but also robust as possible. This concept is so far very successful and may in future attribute to sufficient automation.
The second strategic point is to reduce the general workload by reducing the number of samples. With statistical tools, retrospectively the number of required analysis is calculable. For repeated and comprehensive studies, the work load is significantly reduced without affecting the degree of information. As backup, retaining samples can the drawn.
