OR WAIT null SECS
Using a systematic approach and achieving run-to-run consistency are essential.
The recent resurgence in microbial fermentation for biologics production can be attributed to many factors. Advances in biochemistry and a growing understanding of when glycosylation is not needed, improved capabilities for modification of naked proteins, the emergence of new protein formats and scaffolds well-suited to this production technology, and its applicability to DNA plasmid manufacturing are among them. Successful optimization efforts result in scalable fermentation processes that not only afford high titers and productivity, but also exhibit a high level of robustness and reliability.
The cost-effectiveness of microbial fermentation compared to cell culture makes it attractive when suitable, but there are numerous challenges that must be overcome. Bacteria and yeast typically do not have the cellular machinery capable of coping with complex biological structures, so it is of utmost importance to spend some initial time and effort to identify the best production system for each molecule of interest, according to Jonas Müller, team manager for microbial upstream processing at Lonza.
Another fundamental challenge has been the lack of a good understanding of expression systems and processes, observes Christopher Lennon, director of microbial upstream processing with FUJIFILM Diosynth Biotechnologies, Billingham, United Kingdom. “Without this understanding, it can be difficult to produce the right amount of protein at the required product quality,” he remarks. He does point out, however, that increased access to more sensitive technologies, particularly mass spectrometry, has provided greater insight into product quality at a much earlier stage in development than was previously available. This requires more iterative process development to ensure product quality attributes are met and at earlier stages in the process.
Process control can be an issue when scaling to large volumes, particularly with respect to the oxygen transfer rate and heat management. “For fast-growing bacterial cultures, it is necessary to ensure that sufficient oxygen transfer occurs throughout the entire culture volume to maximize growth, which in turn requires sufficient mixing and airflow,” says Cameron Graham, senior process engineer with BIOVECTRA.
Heat management is important because microbial fermentations typically have tight temperature ranges of +/- 1 or 2 °C of setpoint, according to Graham. “Every millimole of oxygen consumed produces heat, and agitation produces heat, all of which must be removed by the equipment to ensure there are no uncontrolled temperature spikes. This issue becomes a major concern when scaling up production, particularly beyond the 1000-liter scale,” he explains.
Options include replacing chilled water-cooling systems with ultra-low temperature glycol in the vessel jacket, but Graham observes that above 1000 liters, increases to the internal heat transfer area may be required. Surface area enhancements such as baffles that have internal cooling coils have been used to address this challenge.
The use of antibiotics to select for bacteria that maintain high expression levels of the desired product is an issue specific to bacterial fermentation. During downstream processing, the antibiotic must be removed, and regulatory agencies expect demonstration of removal, according to Gregor Awang, a specialist in biologics development and innovation at BIOVECTRA. Typically, this goal is achieved by using very low (10s of ug/mL) starting concentrations of specific, preferred antibiotics and validating the removal of the antibiotic using specific assays and the multiplicative dilution effect of several purification steps. “The development of cell lines that do not require antibiotics but maintain stable expression has consequently become a major goal in recombinant protein and plasmid production,” he notes.
Increasing plasmid expression is another challenge, Awang says. “Producing more plasmid is not as straightforward as simply increasing the number of cells. Retaining plasmids can impact cell growth, thus it is essential to find a balance that optimizes plasmid production and growth rates, which can be tricky to achieve,” he explains.
Another feature of microbial fermentation that dictates its applicability for biologics production is the glycosylation capability of the specific microbes involved. Escherichia coli (E. coli) can perform very few post-translational modifications, while yeast do glycosylate proteins, but not in the same manner as mammalian cells. For these reasons, E. coli and Pichia pastoris (P. pastoris) are perfect and complementary production systems for biomolecules that do not need human glycosylation, Müller states.
While the glycans made by yeast are simpler than those generated during mammalian cell culture and typically not an issue for product consistency, they are mannose-rich, and thus protein products generated in yeast may potentially be immunogenic in humans, according to Awang. He notes that a few companies have worked on humanizing the glycosylation process by inserting specific enzymes into the yeast genome, making yeast capable of making a truncated human glycosylation pattern, but he is not aware of any of these cell lines being used for commercial products.
The types of post-translational modifications required for a product will therefore determine the optimum host organism, according to Müller. While some modifications might be done ex vivo, he emphasizes that the most elegant approach is to leverage the host organism’s cellular machinery. It may be possible to improve results through process parameter optimization, but the existing pathways cannot be replaced completely.
Post-translation modifications, says Lennon, therefore first need to be identified. How they will be controlled then becomes specific to the types of modifications involved. “Some can be controlled by using different expression strains of the same organism, while others may require a more wholesale change to the process, host or cell line,” he concludes.
An important development in commercial-scale microbial fermentation is the development of single-use technologies (SUTs) designed to address the specific processing needs of fermentation reactions. “Single-use systems have historically struggled with heat exchange and have been of limited use for more intensive recombinant protein processes; however, new systems are coming out that close this gap,” Lennon says.
SUTs are advantageous because they eliminate the need for cleaning between batches and simplify the setup and qualification process. BIOVECTRA has shown that single-use units are suitable at the 100- and 1000-liter scales when using high-growth bacteria and yeast. “We’ve used our 100-liter vessel successfully in a bacterial process with an oxygen transfer rate over 400 millimoles per liter per hour, supplying the necessary oxygen transfer rate and cooling capacity. We’re currently using our 1000-liter single-use unit in a similarly demanding process with high-growth yeast and seeing performance comparable to what is expected from a stainless-steel unit,” states Graham.
Currently, single-use microbial fermentation technology is generally limited to a capacity of around 1000 liters, mainly due to cooling capacity constraints. However, as technology continues to advance, Graham thinks it likely that there will be growing demand within the industry to see larger single-use volumes—at least up to 10,000 liters—that take advantage of single-use technology.
For larger stainless-steel vessels, meanwhile, Graham notes that easily cleanable modifications that increase internal heat transfer surface area can simplify the process and reduce downtime between runs. BIOVECTRA has, for instance, modified a 17,000-liter stainless steel vessel with internal cooling baffles that are cleanable to ASME BPE standards and improve consistency and reproducibility between runs.
It is also worth noting, according to Awang, that SUTs do create the opportunity to more easily produce smaller product quantities, such as those required for DNA-based personalized medicines. “By running a small single-use fermentation from 10 to 100 liters, enough DNA can be obtained to create the therapeutic. When larger amounts are needed, large stainless-steel vessels can produce kilograms or tens of kilograms of DNA; having both options available gives the best of both worlds,” he comments. He goes on to observe that because SUTs make bacterial fermentation much simpler, it may be possible to allow installation of a fermentation plant on site or nearby a hospital for faster delivery of personalized medicines for which one-batch-per-patient targeted, small-scale production of highly potent autologous therapeutics (e.g., DNA and mRNA vaccines and gene-modified cell therapies) is involved.
Real-time monitoring of biologics manufacturing to enable improved process control has become an important goal in the biopharmaceutical industry. Production via microbial fermentation is no exception. “Use of PAT [process analytical technologies] during a fermentation run enables detection of potentially problematic variations and allows for manual or automatic corrections to bring the process back to the center of the validated operating space,” Awang says. “By using PAT to ensure success and consistency, you can save time and resources while maintaining the highest quality standards,” he adds.
Optimization of microbial fermentation processes for protein production requires consideration of the type of organism being used. Yeast effectively secretes proteins into the fermentation medium, making product recovery relatively easy using tangential flow filtration (TFF) and/or centrifugation, according to Awang. E. coli, however, produces proteins in the cytoplasm either as soluble proteins or insoluble inclusion bodies, although some may be secreted into the periplasmic space between the inner and the outer membranes.
“These different means of production require different harvest strategies depending on the specific product being made,” Awang says. Products secreted into the periplasm can be recovered by removing the outer membrane via osmotic shock, while proteins secreted into the periplasm will require full lysis of the cells, which is a more complicated process.
Some cytoplasm-located proteins are produced as insoluble inclusion bodies, which require recovery from the lysed culture supernatant by high-speed centrifugation, denaturation, and renaturation to produce soluble, properly folded product, according to Awang. The notes that the additional steps may introduce more impurities or potentially lead to damage to the product.
In fact, while inclusion bodies historically served as a means for pre-purifying protein products, the subsequent refolding process requires dedicated unit operations with large buffer volume capacity and handling of a large amount of chaotropes (typically urea), according to Müller. “These steps are usually the bottleneck of many current facilities and are quite cost-intensive, especially if we consider the waste management aspect. A promising approach to address this challenge is cell engineering, which drives the formation of correctly folded molecules in the cell,” he observes.
Cell line engineering is just one aspect to consider when planning a strategy for microbial fermentation of biologics. The diversity of processes and potential biomolecules that can be made using this technology necessitates the development of a specific strategy and tailor-made process for each product, according to Müller. “We recommend investing time and effort in host selection and early process development,” he stresses. “To achieve the highest possible titer/yield, it is essential to ensure right from the project start that the most suiting organism, expression system, sequence of unit operations and operating conditions are selected,” Müller adds.
The diversity of fermentation processes and molecules produced by them makes establishing a universal microbial fermentation platform process nearly impossible. “Developing a platform or library of cell lines and plasmids that can be effectively utilized depending on the type of protein or product would be a great solution to a big problem in fermentation,” Awang notes. Moving toward a platform-like approach for the downstream processing steps could, meanwhile, provide one means for standardizing some aspects of these manufacturing processes, according to Müller.
Rather than seek to establish one overall platform process, FUJIFILM Diosynth Biotechnologies has developed individual platform solutions optimized for different product accumulation routes (soluble/insoluble, cytoplasmic, or secreted). The company’s pAVEway expression workflow starts with identifying the best expression host/expression vector/gene sequence using automated bioreactor and downstream analysis processes. “This enhanced workflow enables optimal strain selection based on both product titer and product quality. Once the starting conditions are selected, we can improve yield by evaluating the effects of factors such as media and feed composition, induction temperature, and pH,” Lennon comments.
It is also possible to establish a platform process for a particular type of product. Platform solutions for the manufacture of plasmid DNA are increasingly common, but even these approaches require flexibility to accommodate different plasmids and formats.
While process optimization often focuses on product yield for microbial fermentation, consistency from run to run is most important. “Regulatory agencies have provided clear guidelines for developing a fermentation process that meets quality requirements. This involves understanding the system and its capabilities and identifying the design space in which the fermentation can operate, especially with E. coli,” explains Awang.
“It is best,” Awang contends, “to use a systematic approach to development and optimization that includes design-of-experiment studies to understand how the cell line and the culture conditions interact to influence growth and product yield.” He adds that most of the process development and characterization work can be performed at lab scale using small single-use fermenters for model building.
Consistency must then be achieved between lab- and plant-scale models, which is realized by fine-tuning the process until agreement in performance at both scales is reached. Maintaining consistency in multiple parameters to optimize comparability can be a challenge, though, particularly for certain parameters, according to Graham. He points to the concentration of dissolved oxygen as being particularly problematic and notes that it is crucial to carefully consider the type of vessel being used to optimize agreement between the models.
Having consistent small and large-scale models makes it possible to investigate deviations from the norm in plant runs at lab scale, which accelerates the analysis of process consistency and deviations while reducing the cost, Awang observes. In addition, lab-scale models are used to more rapidly qualify raw materials and assess the risk that certain raw materials may be potential sources of variability in plant-scale manufacturing runs.
Optimization of fermentation reactions should not focus on one aspect such as yield. It should include process robustness and consistency, among other attributes such as the oxygen demand and oxygen transfer rate. In addition, optimization is influenced by the ultimate goal for the fermentation process, according to Awang.
“If the goal is to supply hundreds of tons of material to the global market, that expectation must be considered when creating the cell line and fermentation process. Jumping into process development without thinking about the end goal often leads to problems with process scalability later on,” Awang comments. He also notes that careful documentation of cell lines (where they came from, how they were stored, confirmation of identity) is essential to ensure both product consistency and regulatory compliance.
Cynthia A. Challener, PhD, is a contributing editor to BioPharm International®.
Vol. 36, No. 10
When referring to this article, please cite it as Challener, C. A. Optimizing Fermentation Processes for Biologics Manufacturing. BioPharm International 2023 36 (10).