To make cell therapies more widely available, the manufacturing costs must be reduced. When asked what makes cell therapies so costly to produce, Esmond Lee—a doctoral student at the Institute for Stem Cell Biology and Regenerative Medicine at Stanford University School of Medicine—points to a multi-center study in Europe.
The study found that the cost of raw materials makes up 26–82% of the total cost of goods of producing advanced therapy medicinal products under hospital exemption, compassionate use programs, or in clinical trials.
“Hence, lowering the cost of raw materials required to produce a cell therapy will make a significant impact on the total cost of goods for cell and gene therapy,” Lee says.
In 2015–2016, Lee worked in Steve Oh’s lab at the Bioprocessing Technology Institute in Singapore. There, he used Design of Experiments (DoE) to improve media formulations to scale up the production of immortalized erythroblasts for transfusion therapy.
“I wondered if these statistical principles, originally used by RA Fisher for industrial agriculture, could also be applied to the field of CRISPR/Cas9 based gene therapy,” Lee explains.
Working on a PhD in Rosa Bacchetta’s lab at Stanford, Lee is part of a team that is developing novel therapies to treat a primary immune regulatory disorder called IPEX syndrome.
“I wanted to understand the key variables that drive the frequency of targeted integration for knock-in gene editing at the FOXP3 locus,” continues Lee. “We initiated a collaboration with the Umetrics team at Sartorius, and they provided access to their MODDE software, as well as technical expertise on applying DoE to our pre-clinical gene-editing construct.”
In addition to finding the crucial variables, DoE can identify their optimal ranges of operation. “The output response from our DoE experiment was used as inputs for a model estimating the cost of producing 100 million FOXP3 gene-edited cells, which could be a therapeutically relevant dose,” Lee says.
That work showed that the most significant factors impacting gene editing are single guide RNA (sgRNA) and recombinant adeno-associated virus (rAAV). “Surprisingly, while using the highest amount of sgRNA and rAAV led to the best frequency of targeted integration, this was not the most cost-optimal solution,” Lee explains. “Increasing rAAV multiplicity of infection beyond a certain point led to increasing costs but diminishing returns in terms of editing frequency.”
These findings might be used to reduce the cost of manufacturing cell-based therapies.
“While our work is broadly applicable in establishing drivers of raw material costs in gene editing, cost-optimal conditions are context dependent and would need to be experimentally determined,” Lee notes. “Within the broader field of cell therapy, other gene-delivery systems may also benefit from this approach.”