Christoph Clemens

Process analytical technology (PAT) for biopharmaceuticals

Process analytical technology (PAT), the regulatory initiative for building in quality to pharmaceutical manufacturing, has a great potential for improving biopharmaceutical production. The recommended analytical tools for building in quality, multivariate data analysis, mechanistic modeling, novel models for interpretation of systems biology data and new sensor technologies for cellular states, are instrumental in exploiting this potential. Industrial biopharmaceutical production has gradually become dependent on large-scale processes using sensitive mammalian cell cultures. This further emphasizes the need for improved PAT solutions. We summarize recent progress in this area based on an expert workshop held at the 8(th) European Symposium on Biochemical Engineering Sciences (Bologna, 2010), and highlight new opportunities for exploiting PAT when applied in biopharmaceutical production. We conclude with recommendations for advancing PAT applications in the biopharmaceutical industry.

Into the unknown: expression profiling without genome sequence information in CHO by next generation sequencing

The arrival of next-generation sequencing (NGS) technologies has led to novel opportunities for expression profiling and genome analysis by utilizing vast amounts of short read sequence data. Here, we demonstrate that expression profiling in organisms lacking any genome or transcriptome sequence information is feasible by combining Illumina’s mRNA-seq technology with a novel bioinformatics pipeline that integrates assembled and annotated Chinese hamster ovary (CHO) sequences with information derived from related organisms. We applied this pipeline to the analysis of CHO cells which were chosen as a model system owing to its relevance in the production of therapeutic proteins. Specifically, we analysed CHO cells undergoing butyrate treatment which is known to affect cell cycle regulation and to increase the specific productivity of recombinant proteins. By this means, we identified sequences for >13 000 CHO genes which added sequence information of ∼5000 novel genes to the CHO model. More than 6000 transcript sequences are predicted to be complete, as they covered >95% of the corresponding mouse orthologs. Detailed analysis of selected biological functions such as DNA replication and cell cycle control, demonstrated the potential of NGS expression profiling in organisms without extended genome sequence to improve both data quantity and quality.

CHO gene expression profiling in biopharmaceutical process analysis and design

Increase in both productivity and product yields in biopharmaceutical process development with recombinant protein producing mammalian cells can be mainly attributed to the advancements in cell line development, media, and process optimization. Only recently, genome-scale technologies enable a system-level analysis to elucidate the complex biomolecular basis of protein production in mammalian cells promising an increased process understanding and the deduction of knowledge-based approaches for further process optimization. Here, the use of gene expression profiling for the analysis of a low titer (LT) and high titer (HT) fed batch process using the same IgG producing CHO cell line was investigated. We found that gene expression (i) significantly differed in HT versus LT process conditions due to differences in applied chemically defined, serum-free media, (ii) changed over the time course of the fed batch processes, and that (iii) both metabolic pathways and 14 biological functions such as cellular growth or cell death were affected. Furthermore, detailed analysis of metabolism in a standard process format revealed the potential use of transcriptomics for rational media design as is shown for the case of lipid metabolism where the product titer could be increased by about 20% based on a lipid modified basal medium. The results demonstrate that gene expression profiling can be an important tool for mammalian biopharmaceutical process analysis and optimization.

Quality-by-Design for biotechnology-related pharmaceuticals

The following article is a report from a workshop on Quality-by-Design (QbD) held at the 7th European Symposium on Biochemical Engineering Science (7 September 2008, Faro, Portugal).The aim of the workshop was to provide an update on the present status of using QbD in biotechnology-related applications in the pharmaceutical industry. The report summarizes the essential parts of the presentations and covers the industrial, academic, and regulatory aspects of QbD. It concludes with recommendations for further work and development.

Metabolic Control in Mammalian Fed-Batch Cell Cultures for Reduced Lactic Acid Accumulation and Improved Process Robustness

Biomass and cell-specific metabolic rates usually change dynamically over time, making the “feed according to need” strategy difficult to realize in a commercial fed-batch process. We here demonstrate a novel feeding strategy which is designed to hold a particular metabolic state in a fed-batch process by adaptive feeding in real time. The feed rate is calculated with a transferable biomass model based on capacitance, which changes the nutrient flow stoichiometrically in real time. A limited glucose environment was used to confine the cell in a particular metabolic state. In order to cope with uncertainty, two strategies were tested to change the adaptive feed rate and prevent starvation while in limitation: (i) inline pH and online glucose concentration measurement or (ii) inline pH alone, which was shown to be sufficient for the problem statement. In this contribution, we achieved metabolic control within a defined target range. The direct benefit was two-fold: the lactic acid profile was improved and pH could be kept stable. Multivariate Data Analysis (MVDA) has shown that pH influenced lactic acid production or consumption in historical data sets. We demonstrate that a low pH (around 6.8) is not required for our strategy, as glucose availability is already limiting the flux. On the contrary, we boosted glycolytic flux in glucose limitation by setting the pH to 7.4. This new approach led to a yield of lactic acid/glucose (Y L/G) around zero for the whole process time and high titers in our labs. We hypothesize that a higher carbon flux, resulting from a higher pH, may lead to more cells which produce more product. The relevance of this work aims at feeding mammalian cell cultures safely in limitation with a desired metabolic flux range. This resulted in extremely stable, low glucose levels, very robust pH profiles without acid/base interventions and a metabolic state in which lactic acid was consumed instead of being produced from day 1. With this contribution, we wish to extend the basic repertoire of available process control strategies, which will open up new avenues in automation technology and radically improve process robustness in both process development and manufacturing.

Universal Capacitance Model for Real-Time Biomass in Cell Culture

Capacitance probes have the potential to revolutionize bioprocess control due to their safe and robust use and ability to detect even the smallest capacitors in the form of biological cells. Several techniques have evolved to model biomass statistically, however, there are problems with model transfer between cell lines and process conditions. Errors of transferred models in the declining phase of the culture range for linear models around +100% or worse, causing unnecessary delays with test runs during bioprocess development. The goal of this work was to develop one single universal model which can be adapted by considering a potentially mechanistic factor to estimate biomass in yet untested clones and scales. The novelty of this work is a methodology to select sensitive frequencies to build a statistical model which can be shared among fermentations with an error between 9% and 38% (mean error around 20%) for the whole process, including the declining phase. A simple linear factor was found to be responsible for the transferability of biomass models between cell lines, indicating a link to their phenotype or physiology.

A robust feeding strategy to maintain set‐point glucose in mammalian fed‐batch cultures when input parameters have a large error

Industrial CHO cell cultures run under fed‐batch conditions are required to be controlled in particular ranges of glucose, while glucose is constantly consumed and must be replenished by a feed. The most appropriate feeding rate is ideally stoichiometric and adaptive in nature to balance the dynamically changing rate of glucose consumption. However, high errors in biomass and glucose estimation as well as limited knowledge of the true metabolic state challenge the control strategy. In this contribution, we take these errors into account and simulate the output with uncertainty trajectories in silico in order to control glucose concentration. Other than many control strategies, which require parameter estimation, our assumptions are founded on two pillars: (i) first principles and (ii) prior knowledge about the variability of fed‐batch CHO cell culture. The algorithm was exposed to an in‐silico Design of Experiments (DoE), in which variations of parameters were changed simultaneously, such as clone‐specific behavior, precision of equipment and desired control range used. The results demonstrate that our method achieved the target of holding the glucose concentration within an acceptable range. A robust and sufficient level of control could be demonstrated even with high errors for biomass or metabolic state estimation. In a time where blockbuster drugs are queuing up for time slots of their production, this transferable control strategy that is independent of tedious establishment runs may be a decisive advantage for rapid implementation during technology transfer and scale up and decrease in campaign change over time.

Workflow for target-oriented parametrization of an enhanced mechanistic cell culture model†

The goal of this study is to develop a macroscopic mechanistic model describing growth and production within fed-batch cultivations of CHO cells. The model should be used for process characterization as well as for process monitoring including real-time parameter adaptations. The model proved to be able to describe a data-set of 40 processes differing in clones, scales, and process conditions with a normalized root mean square error of approximately 10%. However, due to limited parameter identifiability and limited knowledge about physiologically meaningful parameter values, a broad range of parameters could describe the data with similar quality. This hampered comparison of the model parameters as well as their real-time estimation. Therefore an iterative workflow combining techniques like sensitivity and identifiability analysis, analysis of the specific rates as well as structural adaptations of the parameter space is developed. By applying it the parameter variability could be reduced by 80% with similar predictive power as the original parameters. Summing up, based on a mechanistic CHO model, a generic and transferrable workflow is created for target-oriented parameter estimation in case of limited parameter identifiability. Finally, we suggest a methodology, which fits ideally into the frame of Process Analytical Technology aiming to increase process understanding.