Optimization of Viral Vector Formulation Buffers Using a Microscale Filtration Process

Oncolytic viruses (OVs) are a class of cancer therapies that are currently undergoing clinical trials and on their way to full regulatory approval. The viral manufacturing process is generally divided into two major steps: upstream and downstream processing. Downstream processing (DSP) generally aims at removal of process-related (e.g. cell debris, host cell proteins, and nucleic acids) and product-related impurities (e.g. aggregated and empty virus particles) using a combination of chromatography and membrane-based processes. One of the primary DSP challenges is membrane fouling as indicated by a dramatic transmembrane pressure (TMP) increase and consequent poor virus recovery due to reduced transmission of the virus particles through the membrane. A leading cause of membrane fouling is the aggregation of virus particles, mediated by various biophysical interactions in solution. Aggregated viral particles also affect bio-distribution following in vivo administration, resulting in unwanted inflammatory immune responses, and may even cause reduced vector transduction efficiency. While additives can be used to prevent aggregation, this phenomenon is often poorly understood. Further compounding the problem, additives do not work equally among different viruses to prevent aggregation; thus, methods are required to rapidly screen additives to identify the ones efficiently preventing aggregation.

In our previous work, we have shown how fouling behavior of a membrane could be controlled by selecting the right filter characteristics including membrane material, number of layers, and layer pore size [1]. In this study, we aim to reduce utilized virus solution volume per experiment to overcome the most important shortcoming in our previous process design. Our main objective is to develop a simple and effective procedure utilizing the lowest virus solution volumes to rapidly assessing microscale filtration performance and gaining a comprehensive understanding of virus particle-virus particle, virus particle-impurity, and virus particle-membrane interactions. A Rhabdoviral vector with a unique bullet shape was selected as a vector of interest (provided by the Virus Manufacturing Facility at the Ottawa Hospital Research Institute). A custom-built dead-end small-scale filtration setup was utilized to filter the viral vector solution at a constant flux while simultaneously monitoring changes in the transmembrane pressure (TMP). Using this setup, effects of different excipients including various biomolecules, emulsifiers and polymers on microfiltration of the Rhabdoviral vector and consequently the TMP profile can be screened in a rapid and simple manner. Due to the analogies between the test setup and actual DSP unit operations, results from this work could be easily and more accurately applied to process optimization, leading to reduced virus aggregation and increased process yield.

[1] S. Shoaebargh, I. Gough, M. Fe Medina, A. Smith, J. van der Heijden, B.D. Lichty, J.C. Bell, D.R. Latulippe, Sterile filtration of oncolytic viruses: An analysis of effects of membrane morphology on fouling and product recovery, J. Memb. Sci. 548 (2018) 239–246.