The cell density effect is a well-established constraint in the baculovirus-insect cell expression platform, in which cell-specific productivity declines with increasing cell density, hence limiting the maximum achievable volumetric yield of protein product. A deeper elucidation of this phenomenon is sought in this study, by tracking the peak production of viral DNA (vDNA), recombinant LacZ mRNA, and β-galactosidase (β-gal) protein, over a wide range of cell densities. Sf9 suspension cell cultures were propagated in Sf-900 III serum-free medium and synchronously infected with rAcMNPV at multiple infection cell densities (ICDs) of between 0.5 and 8 × 10(6) cells/mL. There was a strong negative linear correlation between the specific β-gal yield and the peak cell density (PCD) post-infection, but contrary to previous reports, the yield decline started at a lower PCD of around 1 × 10(6) cells/mL. Most interestingly, there also was a corresponding strong negative linear correlation between the specific vDNA or LacZ mRNA yield, and the PCD. Comparing the infections at the highest and lowest PCDs tested, the yield decline was most dramatic for β-gal protein (95 %) and LacZ mRNA (90 %), while it was more moderate for vDNA (50 %). These declines were significantly reduced but not completely arrested, when spent medium was replaced with fresh at the ICD. These findings suggest that protein yield deterioration with increasing cell density originated from limitations during upstream events such as virus gene replication or transcription, rather than during the translational phase. Such limitations may be largely nutritional, but a more complex mechanism may be implicated.

The phenomenon of the cell density effect is not readily explained by an obvious nutrient limitation, and a recent study has suggested that for recombinant Autographa californica multiple nucleopolyhedrovirus (rAcMNPV)-infected Sf9 cells, a drop in messenger RNA (mRNA) levels may be sufficient to explain the cell density effect for this system. The current study aims to investigate the response in cell-specific yields (viral DNA (vDNA), LacZ mRNA and β-galactosidase (β-Gal) protein) with increasing infection cell density (ICD) for rAcMNPV-infected Hi5 cells, where the rAcMNPV expresses the β-Gal gene under control of the polyhedral promoter. Hi5 cells in suspension culture of Express Five® medium were synchronously infected with a rAcMNPV at multiple ICDs between 0.5 and 6 × 10(6) cells/mL and a multiplicity of infection of 10 plaque-forming units (PFU)/cell either in the original or fresh medium conditions. There were negative correlations between the three key virus infection indicators (vDNA, mRNA and β-Gal) and the peak cell density (PCD). However, unlike infected Sf9 cells, the yield decline started at the lowest PCD investigated (0.6 × 10(6) cells/mL). Generally, the yield decline with increasing PCD was most pronounced for β-Gal followed by mRNA and was more moderate for vDNA. The decline was significantly reduced but not totally arrested when fresh medium replacement was used. The results suggest that the reduction in recombinant protein-specific yields at high PCDs is associated with limitations during the up-stream processes of replication and transcription rather than entirely caused by limitations during translation. In addition, low production rates at late infection stages of moderate to high ICDs are a probable cause of the cell density effect.

No abstract

The phenomenon of the reduction in the cell-specific yield with increasing infection cell density (ICD), the cell density effect, is one of the main hurdles for improving virus yields in vitro. In the current study, the reduction in cell-specific yields (viral DNA [vDNA], polyhedrin mRNA and occlusion body [OB]) with increasing ICD for Helicoverpa armigera nucleopolyhedrovirus (HearNPV)-infected HzAM1 (Helicoverpa zea) insect cells has been investigated. HzAM1 cells were propagated in Sf900™ III serum-free medium and synchronously infected with wild-type HearNPV at various ICDs of 0.5-5 × 10(6) cells/mL at an MOI of 5 PFU/cell. Infection was conducted either in the original medium or in fresh medium. As found previously for Sf9 and High Five cells, there were negative correlations between the three key virus infection indicators (vDNA, mRNA and OB) and the peak cell density (PCD). Generally, the yield decline with increasing PCD was most pronounced for OB, followed by mRNA, and was more moderate for vDNA. The decline was significantly reduced, but not totally arrested, when fresh medium was used. There were also strong correlations between OB and mRNA, mRNA and vDNA, and OB and vDNA levels. These results suggest that the reduction in baculovirus yield (OB) at high PCDs is associated with limitations during the upstream processes of replication and transcription together with limitations during protein translation. Furthermore, the peak protein productivity per unit of cell volume in the HzAM1/HearNPV system was shown to be higher than that of the Sf9/rAcMNPV system, but lower than that of the High Five/rAcMNPV system.

A critical component of an in vitro production process for baculovirus biopesticides is a growth medium that is efficacious, robust, and inexpensive. An in-house low-cost serum-free medium, VPM3, has been shown to be very promising in supporting Helicoverpa armigera nucleopolyhedrovirus (HaSNPV) production in H. zea insect cell suspension cultures, for use as a biopesticide against the Heliothine pest complex. However, VPM3 is composed of a significant number of undefined components, including five different protein hydrolysates, which introduce a challenging lot-to-lot variability to the production process. In this study, an intensive statistical optimization routine was employed to reduce the number of protein hydrolysates in VPM3 medium. Nearly 300 runs (including replicates) were conducted with great efficiency by using 50 mL TubeSpin® bioreactors to propagate insect cell suspension cultures. Fractional factorial experiments were first used to determine the most important of the five default protein hydrolysates, and to screen for seven potential substitutes for the default meat peptone, Primatone RL. Validation studies informed by the screening tests showed that promising alternative media could be formulated based on just two protein hydrolysates, in particular the YST-AMP (Yeast Extract and Amyl Meat Peptone) and YST-POT (Yeast Extract and Lucratone Potato Peptone) combinations. The YST-AMP (meat-based) and YST-POT (meat-free) variants of VPM3 were optimized using response surface methodology, and were shown to be just as good as the default VPM3 and the commercial Sf-900 II media in supporting baculovirus yields, hence providing a means toward a more reproducible and scalable production process for HaSNPV biopesticides.