Increasing cost pressures on the biotech industry, in particular, mean that scaleup doggedly remains a major stumbling block in bioprocess manufacturing.
"Because of the high risks of clinical failure, large-scale process trials are often done late when there is little time left for iteration," says Nigel Titchener-Hooker Ph.D., from the EPSRC-funded Innovative Manufacturing Research Center in the Advanced Center for Biochemical Engineering at University College London in the U.K.
Dr. Titchener-Hooker, who was speaking at the recent "Sartorius Downstream Technology Forum" in London, proposed a variety of cost-effective ultra scale-down models that mimic pilot-scale situations which, he believes, could really improve the scale-up process. For example, he discussed an ultra scale-down 10-mL rotating disk device that can be used to predict shearing effects during pilot-scale centrifugal recovery of mammalian cells.
According to Dr. Titchener-Hooker, shearing causes problems because the sub-micron particles that result from it are difficult to remove and can adversely affect the purification of biologicals. Therefore, being able to predict when such shearing may occur and then find design solutions that minimize shearing is highly desirable.
"The information gained from this approach has been used to confirm shearing effects in a centrifuge operating at 20,000-L scale at Lonza Biologics (www.
lonzabiologics.com) and has provided useful information for process scale-up," said Dr. Titchener-Hooker.
He also discussed another model his group developed using a 1-mL chromatography column to mimic process-scale behavior. This model uses flow rates and the amount loaded onto the column to determine the compression and fouling a column would experience. This is then used to produce a "window of operation" graph in which the optimum flow rate and load can be predicted to produce maximum yield.
"When you consider that manufacturing a biological drug can provide one million dollars in revenue each day, it is imperative to get the process up and running rapidly. We believe using these ultra scale-down models will contribute to a more informed scale-up," concluded Dr. Titchener-Hooker.
Tanja Oswald, Ph.D., head of process development at Bayer Healthcare (www.bayerhealthcare.com), a company producing animal vaccines, highlighted the issue of cleaning.
To Clean or Not to Clean
"The regulatory authorities are getting stronger and stronger and are demanding higher GMP standards than ever before. Even though we are making animal vaccines, we still have to manufacture them in the same way as drugs for human consumption," explained Dr. Oswald.
Since biological vaccines cannot be autoclaved for sterility, as this would destroy their activity, pharmas are left with the choice of either maintaining sterility throughout the downstream purification process or filter-sterilizing at the final step.
Bayer has chosen the route of maintaining sterility throughout the process, which presents several challenges. "With the closure of Chiron's (Emeryville, CA) flu vaccine manufacturing unit, cleaning and validation has become a priority. Therefore, we have to have a method of cleaning our columns and of validating that this is clean to present to the regulatory authorities.
"Since we have had difficulties with getting cleaning guidelines and even sometimes the recipes that manufacturers use for cleaning buffers, we have now switched to using disposable filtration systems for purification as this means we can save time and effort cleaning and validating," continued Dr. Oswald.
As an additional justification for using membrane filtration, Dr. Oswald presented an example of how, using a Sartorius Sartobind disposable filtration unit, they were able to recover 69% of a foot-and-mouth disease virus in polyethylene glycol compared with 12.9% from a conventional ion exchange chromatography column.
"Due to the diffusion limits of the pores in the Sartobind it captures more virus than ion exchange media, therefore it can both purify and concentrate in one step, which for us at Bayer is ideal," commented Dr. Oswald.
Andreas Immelmann, Ph.D., managing director of Analysis (www.analysis-gmbh.de), a firm providing validation of viral clearance in the downstream processing of biologicals, presented information on some of the problems being faced when designing viral clearance studies.
"Validation studies can take up to four months, so manufacturers need to plan studies well ahead. Studies should be planned and conducted not only by production bioengineers but also by virologists, as you cannot expect them to be familiar with the manufacturing process. Therefore, virologists need to be included or there will be problems when studies are taken from the laboratory to process scale.
"One of the greatest technical challenges we are facing in bioprocessing is removing the non-envelope viruses, such as the model porcine parvovirus. Ethanol has been the classical method of virus removal, but lately nanofiltration is being considered as a viable alternative.
This is because filters that have 20-nm pore sizes are becoming more widely available and membrane filtration can be easily integrated into the purification process," said Dr. Immelmann.
To prove that ion adsorption filtration offers a good technique for removing non-envelope viruses, Ranil Wickramasinghe, Ph.D., associate professor at Colorado State University, presented the results of his research.
Dr. Wickramasinghe's team challenged four Sartorius anion and cation exchange membranes with the parvovirus Aedes aegypti desonucleosis virus (AeDNV) and tested for the presence of viruses within the filtrate by PCR.
"Viruses are generally negatively charged, and since the pores within an anionic membrane filter are positively charged, theoretically this means that this type of membrane should remove viruses. With column chromatography, you have complicating effects due to size exclusion.
"Further, diffusion through the media particles is slow and makes removing viruses and viral vectors difficult, whereas this is not the case with membrane filters," explained Dr. Wickramasinghe.
The results of the research show anion exchange membranes remove AeDNV, whereas the cation exchange membranes do not. Therefore, these types of membranes have potential uses for viral clearance in biopharmaceutical manufacture or for purifying viral vectors in gene therapy and viral vaccine production, both of which are difficult to do with column chromatography."
A Long Way to Go?
All those attending the "Downstream Technology Forum" agreed that issues such as viral clearance, scale-up, and cleaning make manufacturing biologicals a difficult process.
"There always has to be a reasonable compromise between cost, time, and yield, and achieving the maximum yield is currently impossible with the constraints we have in bioprocess today," said Dr. Oswald.
Dr. Titchener-Hooker agreed. "We have to accept that cost pressures might dictate more generic processes that can be applied quickly as a viable route forward rather than always developing each process de novo.
"All of the challenges we face mean we are often looking for new ways to remove a small amount of contaminants, which have a significant effect on the properties of the process stream. In this respect, we are at a disadvantage compared to the chemical sector where the materials are often more straightforward and simpler to define.
"However, with enough good research in bioprocessing the situation will improve dramatically in the next decade, making biological manufacture a more productive and cost-effective prospect."