Application of mechanistic modeling to speed up purification process development of novel formats

A virus-like particle case study

Virus VLP in blue setting

VLPs are a promising complex molecular format for the development of vaccines, such as vaccines targeting emerging viruses and therapeutic cancer vaccines. Their size and shape enable an easy uptake by antigen-presenting cells (APCs) resulting in an effective induction of an adaptive immune response.

VLPs are formed by the recombinant expression of viral structural proteins. A method for the large-scale production of VLPs is production in suspension cell culture. Within such a VLP production project, a very time and labor consuming part is the development of a purification process to extract the VLPs from cell culture supernatant. The development of a purification strategy is a mainly experimental approach. Conventionally, the operational parameters are optimized via time and material consuming Design-of-Experiment (DoE) studies. Dozens of experiments are needed to screen a process design space that was manually chosen.


In a cooperative project, GoSilico and Cevec pursued an innovative approach for the purification of VLPs. As chromatography media, a CIMmultus QA monolithic adsorber (BIA Separations) was selected and two initial chromatograms were generated: a gradient and a step-wise elution.

Based on this preliminary model, robust elution conditions were developed and the model uncertainty was quantified. Most of the host cell proteins could be removed and >95% of the VLPs could be recovered. Furthermore, the elution conditions achieved a base-line separation of two VLP variants that were co-eluting in the initial calibration experiments.


Afterwards, a third chromatogram was used to reduce the model uncertainty and enable final process optimization with respect to flow rate and switch from gradient to step elutions. Based on the finalized model, an understanding of product quality attributes and their relation to process parameters could be shown, which is a requirement of the Quality-by-Design guideline set forth by the EMA and FDA. What is especially valuable for this case is the mechanistic model’s ability to predict the outcome of a scaled-up process, as the employed annulus shaped monolith is only available in a limited set of sizes.

In summary, the combination of a limited set of experiments and a mechanistic model enabled very efficient process optimization and design space characterization for an innovative molecular scaffold. This approach might constitute a future technology platform for the purification of VLPs and whenever a downstream platform process doesn’t yet exist.

See full poster “Application of mechanistic modeling to speed up purification process development of novel formats: A virus-like particle case study