As previously explained, modifying the number of retransfections, the amount of

DNA used in each re-transfection and the combination of different medium ex-

change time points resulted in variations of VLP titers and overall yield and effi-

ciency of the process. Therefore, a systematic methodology for optimization was

used to combine these parameters in a design of experiment approach. Following a

Box-Behnken design, a mathematical model could be rendered using the time of re-

transfection (measured in hours post-transfection after the first transfection), the

amount of DNA used for this retransfection and the medium exchange rate (mea-

sured in pL·cell⁻¹·day⁻¹, cell-specific perfusion rate) as variables to find an optimal

combination of these parameters for the first re-transfection time point. A similar

method could be used, once the parameters for the first re-transfection have been

optimized, to further optimize a second or even a third re-transfection point if these

prove to be significant to the final obtained VLP titer.

For the optimization of the first re-transfection, the limits for the analyzed

variables were set 24 and 72 hpt as lower and upper limit for the time of re-

transfection, 0.5 and 2 μg/mL of DNA as lower and upper limit for the amount of

DNA and 30 and 1,000 pL·cell⁻¹·day⁻¹ as lower and upper limit for the cell-specific

perfusion rate (CSPR). Following the Box-Behnken method for optimization, a

mathematical model was obtained and further optimized to find the optimal solution

of 24 hpt, 1.7 μg/mL of DNA and 30 pL·cell⁻¹·day⁻¹ as the optimal parameters

for the first re-transfection time point. The method of optimizing the next re-

transfection point using a design of experiment approach did not offer any

significant result due to the cytotoxic effect of PEI together with the use of several

re-transfection points close in time to the point that the cell culture did not have

enough time to recover. Therefore, performing one single re-transfection with the

given parameters already improved VLP titers 7.5-fold [107].

These operational conditions were transferred to bioreactor scale where a process

based on continuous perfusion was implemented. This time, the perfusion approach

was not a conventional perfusion-based process where the aim is to achieve high

cell density and maintain a steady-state through bleeding. In TGE bioprocesses, the

aim of perfusion is to achieve the appropriate continuous medium exchange rate, or

CSPR, while subsequent retransfections are performed. This makes the viable cell

density to enter a plateau due to the effects of PEI and transfection itself on cell

growth [108]. To implement this continuous perfusion approach and the previously

optimized parameters at a bioreactor scale, a cell retention device is needed to retain

and recirculate the cells back to bioreactor while the culture medium is constantly

being circulated through it. In order to achieve this, a very suitable device is the

alternating tangential flow (ATF) cell retention device [109]. For the final bioreactor

setup, the vessel and the ATF hollow fiber module (HFM) are placed on a scale

which constantly monitor their weight. A harvest flow rate from the HFM is con-

trolled using a pump that constantly extracts the filtrated medium. This mass dis-

placement is monitored by the scale, to which a mass setpoint has been set. The

scale is connected to another pump triggering the addition of fresh culture medium

to balance the mass displacement driven by the harvest flow. Like this, the weight is

maintained constant in the bioreactor and the culture medium is renew at a given

CSPR.

254

Bioprocessing of Viral Vaccines