How to accelerate vaccine scale-up

The Problem

A vaccine which was under significant commercial pressure to deliver large quantities of clinical material was encountering scale-up issues due to high shear rates limiting the final product yield.

The Breakthrough

Rather than relying on traditional, time consuming methods to investigate the process hydrodynamic environment (mixing, shear rates, turbulent eddy dissipation, etc…) , an approach using Computational Fluid Dynamics (CFD) was used to simulate the current mixing environment and identify process conditions to replicate the process across different scales.

The Impact

The CFD approach enabled the avoidance of scaling up the process from lab-scale to production-scale using traditional methods. These methods can be very time consuming and laborious, often requiring scaling in a stepwise manner between discreet scales e.g., 1L to 100L to 1000L etc. This approach enabled an order of magnitude reduction in development time.

Understanding scale-up parameters can be problematic when one considers the different reactor configurations and geometries involved. Factors such as, type of impeller and number and type of baffles need to be considered.

Traditionally scale-up relies on engineering correlations. However, multiple assumptions such as maintaining impeller type and geometric ratios are associated with these correlations. These assumptions are often broken when transferring to a different vessel, particularly with the recent industry movement towards bespoke, single-use systems.

To solve the scale-up challenges APC scientists utilised CFD to create 3D model of the bespoke reactor configurations at each scale. The exact agitation rates to mimic the velocity profiles and mixing environment between scales were iteratively determined via CFD simulations. The resulting shear profiles were then compared to ensure the shear stress at larger scales was equal or less than the shear the product was exposed to at lab scale.

This CFD workflow enabled scale-up of a cell production operation with a minimal number of upstream experiments. Once the working process had been optimized at lab scale CFD was used to replicate the mixing regimes and shear profiles for the different scales. Cell growth and viability profiles were held constant at each reactor scale and/or configuration. This limited the number of experiments to right first-time validation batches for scale-up, reducing the time required by approximately 2/3rds.