Optimization of YF17D-Vectored Zika Vaccine Production by Employing Small-Molecule Viral Sensitizers to Enhance Yields

Abstract

Background: Modern viral vector production needs to consider process intensification for higher yields from smaller production volumes. However, innate antiviral immunity triggered in the producer cell may limit virus replication. While commonly used cell lines (e.g., Vero or E1A-immortalised cells) are already compromised in antiviral pathways, the redundancy of innate signaling complicates host cell optimization by genetic engineering. Small molecules that are hypothesized to target antiviral pathways (Viral Sensitizers, VSEs) added to the culture media offer a versatile alternative to genetic modifications to increase permissiveness and, thus, viral yields across multiple cell lines. Methods: To explore how the yield for a chimeric Zika vaccine candidate (YF-ZIK) could be further be increased in an intensified bioprocess, we used spin tubes or an Ambr15 high-throughput microbioreactor system as scale-down models to optimize the dosing for eight VSEs in three host cell lines (AGE1.CR.pIX, BHK-21, and HEK293-F) based on their tolerability. Results: Addition of VSEs to an already optimized infection process significantly increased infectious titers by up to sevenfold for all three cell lines tested. The development of multi-component VSE formulations using a design of experiments approach allowed further synergistic titer increases in AGE1.CR.pIX cells. Scale-up to 1 L stirred-tank bioreactors and 3D-printed mimics of 200 or 2000 L reactors resulted in up to threefold and eightfold increases, respectively. Conclusions: Addition of single VSEs or combinations thereof allowed a further increase in YF-ZIK titers beyond the yield of an already optimized, highly intensified process. The described approach validates the use of VSEs and can be instructive for optimizing other virus production processes.

Keywords: antiviral defense; bioprocess engineering; design of experiments; small molecules; vectored live-attenuated vaccines; viral sensitizers.

Figure 1

Initial screening of eight VSEs to increase infectious YF-ZIK titers in three cell lines. Low (blue), medium (gray), and high (red) concentrations of each VSE were used. (A) Maximum infectious virus titers (PFU/mL) are shown as fold-changes relative to YF-ZIK-producing cells without VSEs. (B) Fold-changes in maximum total viral genomes measured using qRT-PCR. Promising VSEs are highlighted with dashed boxes. Values represent single experiments carried out in spin tubes (n = 1).

Figure 2

Dose-dependent effect of selected VSEs in AGE1.CR.pIX (A) and HEK293-F (B) cells to increase YF-ZIK titers. Maximum infectious virus titers (PFU/mL) are shown as fold-changes relative to YF-ZIK-producing cells without VSEs. Error bars represent the mean ± STD of biological replicates (n = 1–6 for VSEs and n = 6 for the control) carried out in spin tubes. Normality was assessed using the Shapiro–Wilk test. Data (n ≥ 3) were analyzed using one-way ANOVA, followed by Dennett’s comparison test, for comparison with the control. Significant differences are indicated by asterisks (* p < 0.05; ** p < 0.01; **** p < 0.0001).

Figure 3

Evaluation of selected VSE concentrations for YF-ZIK production in AGE1.CR.pIX cells using an Ambr15 system. Cell growth and fold-changes in infectious virus titers (PFU/mL) using an Ambr15 system. (A) VCC (full symbols) and cell viability (empty symbols) are shown for AGE1.CR.pIX cells grouped by VSE. Control infections (sterile water instead of VSEs) are shown as red circles. (B) Maximum infectious virus titers are shown as fold-changes relative to the control infections. Error bars represent the mean ± STD of biological replicates (n = 3 for VSEs and n = 11 for the control). Normality was assessed using the Shapiro–Wilk test. Data (n ≥ 3) were analyzed using one-way ANOVA, followed by Dennett’s comparison test, for comparison with the control. Significant differences are indicated with asterisks (** p < 0.01; **** p < 0.0001).

Figure 4

Evaluation of synergistic and antagonistic effects of combinations of two VSEs in AGE1.CR.pIX cells using a full-factorial (FF) DoE design. (A) Maximum infectious virus titers (PFU/mL) are shown as fold-changes relative to the control infections (n = 13). Error bars represent the mean ± STD of biological duplicates. Significant differences are indicated with asterisks (* p < 0.05; **** p < 0.0001). (B) Effect size of factors and factor interactions measured using the partial Eta-square from a 2 × 2 FF ANOVA. The statistical significance was set at 0.05. (C) Synergy plot showing a combination index (CI) analysis according to the response additivity approach. CI values >1 indicate antagonism, =1 indicate additivity, and <1 indicate synergy. All experiments were carried out in the Ambr15 system.

Figure 5

Four-dimensional response contour plot of a CCF DoE design investigating triple-VSE combinations in AGE1.CR.pIX cells using an Ambr15 system. Maximum infectious virus titers (PFU/mL) are shown as fold-changes relative to the control infections (n = 15). The influence of the respective VSE concentration on the fold-increase responses is shown.

Figure 6

Maximum infectious YF-ZIK titers across various cell lines and production formats with and without VSE addition. BHK-21 and HEK293-F cells were cultivated in spin tubes, and AGE1.CR.pIX cells were cultivated in the Ambr15 system. For BHK-21 cells, the addition of 25 µM VS-E (n = 1) was compared to the control infection (n = 2). For AGE1.CR.pIX cells, the addition of 1 µM VS-F and 15 µM VS-G (n = 3) was compared to the control infection (n = 15). For HEK293-F cells, the addition of 45 µM VS-E (n = 3) was compared to the control infection (n = 6). Error bars represent the mean ± STD of replicates.

Figure 7

YF-ZIK production in benchtop 1 L STRs in batch mode using AGE1.CR.pIX (left, green) and HEK293-F cells (right, red) with or without VSE addition. VSEs (colored) or PBS (black) were added to the cells prior to infection. Replicates are represented by different symbols (controls: triangle, diamond, square; VSE addition: circle and downwards triangle); 1 µM VS-F and 15 µM VS-G were added to AGE1.CR.pIX cells, and 25 µM VS-C was added to HEK293-F cells. At the time of infection, cells were diluted with fresh medium at a ratio of 1:2, and the temperature was reduced to 33 °C. (A) VCC (full symbols) and culture viability (empty symbols). Dashed lines indicate the time of infection. (B) Maximum infectious virus titers determined with a plaque assay. Error bars represent the mean ± STD of biological replicates (n = 2 for VSEs and n = 3 for the control). Significant differences are indicated with asterisks (** p < 0.01; *** p < 0.001).

Figure 8

Simulated mixing curves for 200 L (A) and 2000 L (B) with the small-scale plain vessel without a structure (red), the target reactor (either 200 or 2000 L, blue), and the corresponding small-scale model including a structure (green). The dotted line shows the 95% threshold of the relative standard deviation of the scalar over the entire domain (RSD), representing homogenous mixing. (C) The structure of the 2000 L surrogate pDS reactor (right) compared to the small-scale plain vessel without a structure (left) and the tracer distribution after 10 s (red) and 8.2 s (blue), respectively. Depicted in red are the shaft and stirrer blade of the used stirrer at the respective stirrer height.

Figure 9

Comparison between plain glass (red), polymer 3D-printed (black), and 3D-printed pDS reactor A (blue) or pDS reactor B (gray) for YF-ZIK production using AGE1.CR.pIX cells with or without the addition of VSEs. Prior to infection, cells were diluted at a ratio of 1:2 with fresh medium, the temperature was reduced to 33 °C, and 1 µM VS-F and 15 µM VS-G or PBS were added to the cells. (A) VCC (full symbols) and culture viability (empty symbols). Dashed lines indicate the time of infection. (DO control failure in reactor A without VSE at 4 dpi.) (B) Glucose (black) and lactate (gray) uptake/production rates (left) and cell-specific growth rates (middle) prior to infection and/or VSE addition. Simulated mixing times are shown on the right. Error bars represent the mean ± STD of biological replicates (n = 2 for reactors A and B and polymer reactors; n = 5 for glass reactors). Significant differences are indicated with asterisks (*** p < 0.001; **** p < 0.0001). (C) Maximum infectious virus titers determined with a plaque assay. Black patterns indicate VSE addition. Due to DO control failure in reactor A without VSE at 4 dpi, no fold change is indicated for reactor A.