Scale-up and manufacturing strategy for insect cell culture (ICC) should be assessed at the earliest stages of process development. Your final goals will shape the process and determine at what stage to implement steps that can make the process more reliable and robust at manufacturing scale. Only a few vendors are experienced enough to guide you through them. On 7 December 2016, Sharyn Farnsworth (associate principal scientist at Fujifilm Diosynth Biotechnologies) discussed some upstream challenges BEVS/ICC in a BPI webinar.
The process we develop is an ICC using Sf21, Sf9, and High-Five cell lines to produce recombinant proteins. The BEVS delivers a gene of interest into cells. More than 500 baculoviruses can infect and produce recombinant proteins in insect cells. This system can make very large molecules at high concentrations — e.g., virus-like proteins (VLPs) for vaccines. It’s also capable of certain posttranslational modifications, but not to the extent that Chinese hamster ovary (CHO) cells can manage.
Although this is a relatively inexpensive system, difficulty with robust and reproducible statistical testing of both virus titers and protein titers can lead to significant process variation. This is even more exaggerated when viruses and/or virus stocks are unstable.
Seed Train Scale-Up: ICC growth is relatively standard, but thawing recovery can take more time than for mammalian cells (to reach consistent doubling times). Both minimum and maximum passage numbers need to be evaluated because they affect how infective the baculovirus will be. The results can help determine a cell-cycling strategy and how long a GMP cell bank will last.
Infection cell density (ICD) is the cell density at which an ICC will be infected with virus to begin protein production. Multiplicity of infection (MoI) is the amount of virus that will be added at infection. Together, they have the most influence on robust high yielding culture. Additionally, these two parameters will significantly affect the size of your virus bank and direct your virus stock strategy.
Harvest: Time to harvest typically is determined by the point at which protein production tapers off. Productivity begins to decline substantially after peak cell size. It is also influenced by whether production occurs through a primary or secondary infection. For scheduling consistency in GMP manufacturing, a time-based harvest is the best method ultimately.
Culture Media: We continually evaluate new media modifications as they become available from vendors. Nothing fully chemically defined has worked consistently for seed train and infection because the two processes have different nutrient needs. Currently available media typically include some type of yeastolate to supply the needs of both cell growth and production postinfection. In some cases, supplementation at infection can increase productivity.
Questions and Answers
Without chemically defined media, what do you do to minimize variability? We perform growth promotion and titer evaluations on new media lots and unknown components.
What tools or process analytical technologies (PATs) can be used to determine peak cell size? How do you determine when it is declining? Cell counting is the easiest way. We’ve used a Vi-CELL counter (Beckman Coulter), which measures viable cell density and average cell size. Once the virus infects insect cells, they start to swell. Usually about 24–48 hours later is when productivity starts to decline — when most of the virus has escaped and the protein is in solution.
When VLPs assemble inside the Sf9 nucleus, what harvest procedure do you recommend? Usually a centrifugation step removes most of the supernatant, followed by homogenization or chemical cell disruption, then another centrifugation and filtration after that. It depends on how sticky the protein is to the cell debris once you burst the cells open. There could be several steps after homogenization.
How long does the low-temp/low-pH hold last? That depends on the hardiness of both the virus and the protein of interest. Typically, it’s 30 minutes to an hour for low pH, but longer for low-temperature holds.
For ICC processes, do you use single-use bioreactors — and at what scale? We have processes in both stainless steel and single-use reactors and have been able to run them interchangeably. For GMP manufacturing, we have 50-L to 2,000-L scale; for process development, 2-L to 200-L scales.
What is the yield range? That depends on the protein. ICC yields are typically lower than for mammalian cell culture: from 20% to 70%.
The full presentation of this “Ask the Expert” webcast — with data — can be found on the BioProcess International website.