Baculovirus infected insect cell systems maintain an ongoing interest for producing biopesticides, recombinant proteins, vaccines and vectors for gene delivery. It is critical to enhance virus yields for economic feasibility of this production system, especially for the application of low cost biopesticides and veterinary vaccines. Two important approaches are the use of low cost serum free medium and infections at high cell densities in order to maximize volumetric production. For the first approach, the current low cost medium, VPM3, available to this project, needs to be optimized to reduce the variability in the performance of the final medium. The second approach relates to the so called “cell density effect”, in which the cell specific yield declines with an increase in the infection cell density (ICD). This PhD thesis focuses on these two limiting issues for enhancing baculovirus yields.
The variability of the VPM3 medium was addressed through a medium optimization process with the Helicoverpa zea (HzAM1) cell line and a wild-type Helicoverpa armigera single nucleopolyhedrovirus (HearNPV) virus isolate, using a high throughput means involving 50-mL tube bioreactors and a statistical approach. Through a series of screening and validation experiments and a response surface method, two of the most reliable hydrolysates out of five that were previously used were selected and optimized in terms of levels added. The result showed that the simplified low cost medium was as good as the default medium andthe commercially available media in terms of cell growth and virus production. The medium optimization process also generated an animal components free version of the low cost medium. Ideally a chemically defined medium or at least a low cost medium containing a single hydrolysate would have resulted from this work, but it was felt beyond the scope of this PhD program to persist with these studies. Commercially available media were subsequently adopted for the cell density studies as such media display lower variability in performance and because, as far as is known, they contain a single hydrolysate, most likely a yeast extract.
The cell density effect has been documented extensively in the literature. However, previous investigators restricted their studies to the later stage of the virus infection process, the protein production phase. Therefore, a comprehensive study of the cell density effect has been conducted, not only assessing the late stage of protein production, but also the upstream processes of mRNA expression of the virus polyhedral gene or of a recombinant gene under control of the polyhedral promoter, and of viral DNA (vDNA) replication. The cell density effect was investigated for different cell line/virus systems including Sf9/rAcMNPV, Hi5/rAcMNPV and HzAM1/HearNPV propagated in either Sf900™III or Express Five®. The results showed that the decline in cell specific yield with increasing ICDs was not only significant for protein production but also applied to mRNA expression under control of the polyhedrin promoter and vDNA replication. These declines were significantly reduced but not completely arrested with fresh medium replacement at the time of infection.
The decline in protein production may result entirely because of the impairment of the early stage of virus replication and/or transcription at high cell densities (6-36 hpi), but other limitations during the protein expression phase, (24-72hpi), cannot be ruled out. The data indicates that the timing of transcription, replication and protein expression are similar for all three cell line/virus systems and while the increase in the cell density at the time of infection affects the yield dramatically, the timing of events do not appear to be affected by the ICD. It is of interest that the Hi5 cells produce a peak recombinant protein yield per unit volume (pg/µm3) that is 3.7 times higher than that produced by the Sf9 cells and 2.8 times that of the HzAM1 cells used in these studies. The reason for the productivity variations between the cell lines is not clear.
The hydrolysates have potent impacts on cell growth and virus infection in serum free medium. They likely play a role in the cell density effect, in which a higher hydrolysate content in the medium may prolong the stable range of the cell specific yield with increasing ICD. As reported in the literature a yeast additive is critical to cell growth, but in addition it was shown that in the IPL-41 basal medium plus yeast and lipid additives (IPLYL), the specific b-Gal yield of infected Sf9 cells was stable for longer, up to the peak cell density (PCD) of 2.5×106 cell/mL compared to only 1.3×106 cell/mL in Sf900™III, which has a lower hydrolysate content than was used in the IPLYL formulation, which we speculate is the reason for the longer period of stable specific yield. Additionally, yeast was shown to be important for protein production during the first 24 hpi. It is particularly interesting to note that the IPLYL medium used gave similar peak volumetric yields for the Sf9/rAcMNPV system compared to that obtained using the commercially available SF900™III medium in batch cultures, despite containing lower amino acid and glucose levels.
Finally, in contrast to some reports in the literature, the IPL-41 based studies demonstrated an important role for lipids not only for cell growth and production of BV prior to infection but also during the infection period itself. If the lipid emulsion is withdrawn post-infection, recombinant protein yields drop dramatically. Even when lipids are withdrawn as late as 24 hpi, yields drop compared to appropriate controls.