Effective Stabilization of Viral Vectors in Liquid Using an Algorithm-Based Development Approach

Background

In recent years approvals of both CAR-T cell therapies and promising clinical data of gene therapy trials such as in hemophilia or sickle cell disease, prove that significant progress of gene and cell therapy in the field of hematological malignancies as well as rare hematologic genetic disorders has been made. Replication-deficient recombinant viral vectors such as adenovirus serotype 5 (Ad5), adeno-associated virus (AAV) and members of the poxvirus family represent a rapidly growing field of vaccine development and gene therapy.

Viral vectors are known as complex supra-molecular ensembles of macromolecules produced by living organisms (nucleic acids, proteins, polysaccharides and in the case of lipid enveloped viruses of phospholipids) which are prone to a variety of complex chemical and physical degradation pathways, in particular due to stress induced by manufacturing, storage and distribution; hence representing a significant hurdle for the development and administration of stable vector-based pharmaceuticals such as vaccines or gene therapeutics.

AIM

To effectively develop stabilizing formulations for replication-deficient recombinant viral vectors, we applied a database and algorithm-based development strategy including Design of Experiment (DoE).

Methods

As a representative example of viral vectors we used Ad5. DoE was applied to select the most effectively stabilizing formulations. Ad5 was stored at 37 °C and 25 °C as accelerated aging temperature to identify the most effective stabilizing excipients and combinations thereof. Based on these results, the best selected formulations were further iteratively optimized and used for real-time storage at 5 °C in both liquid and lyophilized formulation. Infectivity was measured by titration assay and immunostaining of the Ad5 specific Hexon protein in HEK293 cell culture.

ResultS

By analysis of the infectious virus titers and mathematical combination of these results with the DoE matrix, the linear influence of each amino acid used in the DoE matrix was determined. The accelerated aging conditions were shown to be predictive for real-time aging. Several of the excipients indicated a neutral influence, whereby a well-balanced combination of the components was important regarding concentration and stabilizing interactions. The predictive power of this approach was confirmed by by long-term storage of two iteratively improved formulations tailored for Ad5.

Conclusion

The highly efficient pre-selection strategy of effectively stabilizing excipients by an algorithm-based development approach and an accelerated aging model enables the generation of best-in-class stable liquid formulations for viral vectors in short development time. Stable titres in liquid formulations can avoid lyophilization, loss of infectivity via freeze/thaw procedures or high overage and thus significantly reduce manufacturing costs. This approach is transferred to AAV and could have beneficial impact when applied early in downstream processing. A transfer to the frequently applied gene therapy vector system AAV is currently being implemented and an update will be presented at the conference.