RESULTS
Effect of plasmid DNA availability on AAV production at low cell densities
In the present study, we investigated the improvement of AAV6 production via transient transfection of suspension-adapted HEK293SF cells by increasing cell density. The standard AAV production in our laboratory is conducted via transfection at low cell density with 1 µg of plasmid DNA per mL. First, we assessed if plasmid DNA availability played an important role during production in an increased cell density. We compared the AAV6 yield from productions at 1 and 2 million cells per mL. During transfection, plasmid DNA was delivered either on a volumetric basis (1 µg/mL) or a cell basis (1 µg/106 cells). PEI concentration for all experiments was set at 2 µg per µg of plasmid DNA. Viral vectors were harvested 48 hours post-transfection (hpt) to determine viral yield in terms of genome-containing particles (viral genomes, VG). The cell density and viability profiles during AAV production are shown in Figure 1A. Lower transduction efficiencies were observed in both productions at 2 million cells per mL compared to the control (Figure 1B). A 2.4-fold increase in viral titer was observed in the transfection of 2 million cells per mL when DNA was supplied on a cell basis (5.4 × 109 VG/mL), resulting in a better-than-linear increase of viral titer when compared to the standard production (2.2 × 109 VG/mL) (Figure 1C). Productions at 2 million cells per mL showed the highest cell-specific viral yield (CSVY) only when DNA was supplied on a cell basis (approximately 3100 VG/cell). The delivery of DNA as 1 µg/mL in the transfection of 2 million cells per mL caused a reduction in the CSVY to only 900 VG/cell, resulting in a volumetric yield of only 1.8 × 109 VG/mL.
Cell-specific production maintained at a medium cell density
The next step was to evaluate the viral vector production at a medium cell density (MCD, 4 × 106 cells/mL at the time of transfection). Again, the cells were transfected with plasmids on either a volumetric or cell basis and the viral titer VG/mL was assessed every 24 hours. Figure 2A shows the cell density and viability profiles during AAV production at low and medium cell densities. A drop in the transfection efficiency was observed even when a higher amount of DNA was supplied (Figure 2C). However, as previously observed, this lower transduction efficiency did not impact viral titer (Figure 2B, DNA delivered on a cell basis). Similar to what was observed in the transfection of 2 million cells per mL, the titer of the MCD production with DNA delivered on a cell basis was 3.8 times the control, reaching 8.6 × 109VG/mL (Figure 2B). DNA delivery on a volumetric basis resulted in 4.9 × 109 VG/mL (a 2.2-fold increase from the control). When DNA is supplied on a cell basis, the CSVY was maintained at around 2,000 VG/cell. On the other hand, volumetric delivery of DNA resulted in a decrease to 1,200 VG/cell (Figure 2B). The functional viral titers (enhanced transduction units (ETU)/mL) are summarized in Figure 2D. An increase in cell density resulted in a rise of functional titer by 2.5-3.9 times, depending on the harvest time. At 48 hpt, no statistically significant difference (P = 0.4261) was observed in the ratio VG/ETU between low- and medium-cell-density production with 1 µg of DNA per million cells (Figure 2E).
The cell density effect is observed at higher densities
To study if cell-based delivery of DNA would be enough to prevent the cell density effect at a high-cell-density (HCD) production, we evaluated AAV6 production at 10 million cells per mL. A perfusion-like mode of operation was devised in 50-mL TubeSpin bioreactors to achieve the desired cell density. A working volume of 10 mL and medium exchange of 1 vessel volume per day (VVD) were set. The medium exchange started 24 hours after the seeding of the cells and was maintained past the transfection to keep providing fresh nutrients to the cells. Cells were transfected at a density of 10 × 106 cells/mL, with plasmid DNA delivered on a cell basis, and viral titer was assessed every 24 hours. A low-cell-density (LCD) control was also prepared in similar vessels. As expected, the transduction of the high cell-density production was lower than the control (Figure 3A). Cell densities and volumetric titers during AAV6 productions are shown in Figure 3B. AAV6 yield peaked at 72 hpt, and the HCD production resulted in 9.3 × 109 VG/mL, a 4.6-fold increase compared to the LCD control. Even though the yield was higher than the control, this result shows that the titer does not increase linearly with the increase in cell density at the time of transfection. This difference becomes even more evident at the cell-specific level. At 72 hpt, the LCD batch control yielded an average of 1,600 VG/cell, while the HCD perfusion yielded less than 1,000 VG/cell (Figure 5B).
Medium supplementation improves cell-specific production of genome-containing particles
Nutrients and cell metabolites were analyzed during the production of AAV6 at low and high cell densities in TubeSpin bioreactors. Even though the levels of metabolites (lactate and ammonium) would be maintained low, and levels of glutamine and glutamic acid were kept constant by the medium exchange, a significant drop in glucose concentration was observed (Figures 4A and B). At 48 hpt, glucose concentration was 3.2 g/L for LCD, while for HCD, as low as 1.8 g/L. Two other HCD productions were performed to evaluate the effect of medium supplementation. This time, the medium was supplemented with either glucose at a final concentration of 7.2 g/L or 15% Cell Boost 5, which resulted in a similar concentration of glucose between LCD and HCD production at their lowest points (Figure 4B).
Figure 5A shows the cell density and viability profiles during AAV production at high cell density with basal and supplemented media, and Figure 5B shows the transduction efficiency 24 hours post-transfection. Medium supplementation with Cell Boost 5 resulted in a higher growth rate after transfection, with cell density reaching 25.5 × 106 cells/mL 96 hpt, while cells in basal medium reached up to 20.4 × 106 cells/mL. Cells cultured in the glucose-supplemented medium showed a slower growth rate, resulting in the transfection being done one day later compared to the medium supplemented with Cell Boost 5. At 72 hpt, when compared to the use of the basal medium, the use of supplementation resulted in increased volumetric titer (1.4- and 1.6-fold increase for glucose and Cell Boost 5, respectively) and a significant improvement of cell-specific titer (1.8-fold increase for Cell Boost 5) (Figure 5C).
Medium supplementation alleviated the observed cell density effect in the HCD production, resulting in CSVY similar to those in the LCD control, leading to an almost 10-fold increase in viral titer. There was an increase in functional titer when HCD production was supplemented. Cell-specific functional titer also increases depending on the supplementation, but it was 2 times inferior compared to the LCD (Figure 6A). The VG/ETU ratio for LCD production was 34.4, while for HCD, it reached 79.9 when the medium was supplemented with Cell Boost 5 (Figure 6B).