January 1, 2006 (Vol. 26, No. 1)
Susan Aldridge, Ph.D.
Biomanufacturing Trends and Opportunities Are Meeting Highlights
Process development, facility management, and large-scale manufacturing were among the key issues discussed at the “Sixth European Biotechnology Symposium,” which was sponsored by Genetic Engineering News and held in Copenhagen.
According to Uwe Gottschalk, Ph.D., vp of the new purification technologies unit at Sartorius (www.sartorius.com), downstream processing (DSP) continues to be a challenge. He spoke about several issues involving DSP as demand for Mabs increases and more manufacturing capacity comes onstream.
“These developments will put purification in the spotlight,” he said. “Matching upstream and downstream processing has become a serious issue. We have seen a revolution in upstream processing but not downstream.”
The former, driven by biology attracted far more investment than the latter, driven by engineering, but the imbalance must now be tackled.
Dr. Gottschalk went on to describe some of the current challenges in DSP. These are complex, technically demanding processes involving high initial volumes and, perhaps, low-target protein titer. There is a need for robust and scalable processes for the entire DSP.
Meanwhile, there are also increasing regulatory demands to be dealt with. DSP accounts for 75% of manufacturing costs which, in turn, account for 1525% of cost of goods.
For Mab DSP, the capture step with Protein A is one important bottleneck because the process has a physical limit, although there is a new version of Protein A from Millipore (www.millipore.com) with a higher capacity.
“Protein A is the standard for the near future at least,” said Dr. Gottschalk. “There is a lot of activity looking for alternatives and there are plenty of questions. But there are no answers as yet.”
Polishing and viral inactivation also represent bottlenecks in DSP, and the use of membranes rather than columns might be useful here.
Meanwhile, Sartorius has a global distribution agreement with LevTech (www.levtech.net) for the LevTech system that can be used for more efficient viral inactivation, noted Dr. Gottschalk.
The product uses superconductor technology to drive a single-use impeller inside a sterile, disposable bag, eliminating cross contaminations, particle generation, and friction.
The last few years saw a focus on high-tech solutions, but there could be a place for traditional methods in DSP, such as precipitation and crystallization.
“I’m in favor of revisiting some of the robust aspects of older technologies and keeping things simple,” said Dr. Gottschalk.
CHO Cells Rule
Florian Wurm, Ph.D., of the Swiss Federal Institute of Technology, Lausanne and founder and CSO of spin-out ExcellGene (www. excellgene.com), provided the conference attendees with some observations on mammalian cell culture and recombinant protein production.
“CHO cells are still dominant and account for 80 to 90% of production. I am absolutely convinced CHO will continue as the leading production system,” he said. “Through media development, we have learned what to feed the cells and when.”
Dr. Wurm, however, feels there may also be a place for transgenic production systems in the future.
“Products with a limited lifespan will never be made in transgenics, because it is so much faster to use CHO cells. But products needed in 10’s to hundreds-of-ton volumes and high-dose products like herceptin could perhaps usefully be made in animals or potatoes.”
From 1986 to 2004, recombinant protein production has gone up dramatically from 50 mg/L to 4.7 g/L with the main contributor being a massive increase in the number of cells, thanks to media improvement.
Process control, process design, and host cell engineering have also been significant factors in yield improvement, while promoters, enhancers, the nature of the overall plasmid construct, and better cloning techniques have not been.
No Generic Process
There is still no generic process for producing proteins in mammalian cells, said Dr. Wurm, and nor will there be in the foreseeable future. A huge body of knowledge has to be built up for production of one particular protein and it is different for each product.
Batch and extended batch culture are the main methods of production today. Process development, however, is still vital and Dr. Wurm’s team has been looking at a shaking technology, TubeSpin, which involves ventilated 50 mL tubes.
They hope to discover why there is (non-productive) lag phase in culture growth and, ideally, eliminate this and get more productivity through creating conditions for continuous log phase. The team would also like to develop media for higher density cell culture, because currently only 12% of the culture volume is actually occupied by cells.
Finally, better control is required over the way DNA gets into host cells. In the absence of any project on the CHO genome, mouse and human data will need to be used to try to identify regions that accept foreign genes.
Simon Curvers, Ph.D., CEO of AC Biotec (www.acbiotech.com), talked about how his team uses miniaturization and parallelization in process development.
“Screening and process development are now coming closer together,” he noted. AC Biotec has a focus on shaken rather than stirred bioreactors in parallel, which is a simple and reliable technology with no moving parts and low investment costs. Three specific tools are used to help in media optimization.
RAMOS, invented by the co-founder of AC Biotech, Tibor Anderlei, Ph.D., is a device that measures the respiratory activity of microbial, plant, and cell cultures online. COSBIOS is a continuous fermentation system with six parallel bioreactors that gives results more rapidly than a single bioreactor and is applicable to substrate inhibited systems, foamy systems, and metabolic flux analysis.
Finally, the FeedBead is a polymer particle that allows slow-release of glucose in miniaturized fed batch processes and proved useful in selecting optimal production strains during screening.
Of course, biomanufacturing now encompasses many more entities other than recombinant proteins and Mabs. Peter Steiner of ESBATech (www.estbatech.com) discussed the advantages of these compared to conventional antibodies, such as increased stability and solubility.
“We have addressed the critical aspects of single-chain antibody fragments. We expect them to be most useful in compartments where there are no systemic applications like the eye, lung, joint, and CNS,” he said.
As far as production is concerned, the stable antibody fragments are easier than conventional antibodies because they do not require glycosylation and can be produced in high-yield in E. coli.
Meanwhile, Troels Koch, Ph.D., vp of chemistry and manufacturing at Santaris Pharma (www.santaris.com) described some of the challenges involved in synthesizing locked nucleic acids (LNAs) that are chemically-modified oligonucleotides for application in therapeutics based on RNA regulation.
“These are the most powerful nucleic acid analogs ever made,” he claimed. “They can be put into any oligo or DNA sequence.”
The LNAs Santaris is developing are 16-mers, where the synthesis of the monomers involves 32 steps. “We pursued an aggressive cost-cutting strategy, so that the LNAs are now cheaper than the corresponding RNA technologies,” he added.
Improvement of yields, decrease of cost of goods, and assurance of supply involved developing good relationships with three different CMOs. The first, from the EU, had a good track record in nucleoside chemistry and was able to produce monomers in the 10 to 20 g scale that represented a yield improvement from the 0.510% range to the 2540% range, according to Dr. Koch.
In the next phase, up to 300 g of monomer were being synthesized. The collaboration with the first CMO continued, and a second was engaged as a back-up and for validation.
Then, for the final scale-up to kg scale and beyond, a collaboration was set up with a CMO in India that offered an attractive combination of low-cost manufacturing, bulk facilities, and nucleic acid experience.
“Finding the right supplier can be difficult and time consuming,” admitted Dr. Koch. “However, we have a relationship with our suppliers that’s on a personal basis. Our people are working in India. I believe that working as a partnership rather than as a customer is important, and we will continue that way.”
Meanwhile, pre-clinical testing of lead compounds in the Santaris LNA program show high-affinity, high-specificity, and good biostability, suggesting they have the looked-for edge over antisense and RNAi approaches.
Santaris Pharma’s most advanced program on SPC2996 is in Phase I/II testing for chronic lymphocytic leukemia. Data from this program is expected in the first half of 2006.
Process Scale Chromatography
Ian Sellick, marketing director of Pall Life Sciences (www.pall.com), discussed the relationship between capacity, resolution, and flow rate in process-scale chromatography and how this can be optimized.
Recent research involved the use of the Ciphergen ProteinChip technology to look at how the chromatographic surface interacts with the protein product to discover the best elution conditions.
This has been exemplified by work done by Fons Bosman, Ph.D., of Innogenetics (www.innogenetics.com), on a specific Mab. Good transferability of on-chip conditions to the full sized column were shown.
Process development and optimization on-chip take only two to three days compared to several weeks by more conventional methods. In another example, Pall scientists used 96-well plate screening for optimizing the polishing step in a recombinant protein.
“These are powerful techniques for scouting conditions for chromatography and are especially applicable for molecules where we don’t know what the purification procedure should be,”explained Sellick.
Meanwhile, membrane separation shows promise as a complement to column chromatography. The Pall Mustang Q Z XT5000 membrane capsules can be configured in series or in parallel to increase the effective membrane volume.
With this kind of set-up, as much as 1.5 kg of protein can be purified and, indeed, it has now been approved by the FDA for a polishing application for DNA and viral clearance.
Concluding the conference session, Friedrich Nachtmann, Ph.D., head of biotech co-operations at Sandoz (www.
sandoz.com), discussed how a manufacturing facility can be managed for maximum utilization.
Sandoz offers contract manufacture of biopharmaceuticals in microbial and mammalian cell culture as well as developing its own program of biogenerics. The manufacturing cost element of facility, personnel, and energy is where most efficiency gains are possible. Only limited savings can be made in the other two elements of material and service costs.
“A decent facility requires an investment of over $100 million,” said Dr. Nachtmann, drawing on cost data from recently built mammalian cell culture facilities.
CMOs can reduce these costs because they share the risk, he noted, adding that CMOs also have multipurpose facilities, and if your product does not succeed, the capacity can be used for something else.
“Ninety percent of all biotech plants are not being used for the product originally intended,” he pointed out.
There are, however, some specific challenges associated with the multipurpose plant. Regulation will demand special changeover procedures, and staff must be more highly trained than for a plant dedicated to one product.
Continuous process development is less feasible too, because the multipurpose facility operates on a campaign basis.