Gene-based therapies are relatively new but the buzz surrounding them is deafening. Stakeholders are already thinking about cost of goods (COG), and bioprocessing is always a good place to look for COG improvements.
But despite their 22nd-century price tags, traditional viral vector production processes still rely on 20th-century technologies—essentially laboratory production methods which are subsequently industrialized as product volume requirements dictate.
A group led by Suzanne S. Farid, PhD, professor in the department of biochemical engineering, University College London, has recently completed a study of COG in the manufacture of lentiviral vectors. Farid compared viral vector production methods based on adherent cultures with suspension cells cultured in a single-use bioreactor. Her cost estimation tool was based on a whole-process economic model which was linked to optimization algorithms associating COG with various process conditions.
Farid concluded that suspension cultures in single-use bioreactors achieve an approximately 90% improvement in COG per dose, compared with more traditional methods based on adherent HEK293 culture in stackable, multi-layer vessels.
Her analysis also uncovered major contributors to COG according to production volume. For up to around 100 doses per year the most significant factors were labor and indirect costs. Above around 500 doses the relative contribution from indirect costs diminishes, and raw materials become the leading COG contributor.
For example, at the 100-dose level labor and indirect costs accounted for 35% and 40%, respectively, followed by raw materials and quality-related activities. Above 500 doses raw material costs begin to dominate. By the time the 1000-dose scale is reached these costs account for 40% of the total, and QC costs become almost negligible. Raw material costs dominated since they increase, more or less linearly, with scale while facility overheads are spread over more doses.
Trends in economies of scale
“The trends in economies of scale for viral vector processes are similar to those for mAbs with the dominant costs shifting from fixed costs at small scales to material costs at large scales,” Farid tells GEN. “But of course the specific raw material cost drivers differ given that most viral vector processes still rely on transient transfection that is dependent on the supply of plasmid DNA, and many use lab-based methods. For example, for processes relying on multi-layer vessels, single-use components dominate given the large numbers of units required, while for processes relying on more scalable technologies, such as single-use bioreactors, the key material influencer shifts to the costly plasmid DNA required in transient transfection processes.”
Yet one gets the feeling that COG considerations are different for gene therapies than for monoclonal antibody production, which for years seemed to be disconnected from normal economic forces.
“COG is a hot topic in cell and gene therapy given several notable failures attributed to manufacturing concerns, including high COG,” Farid adds. “Given the relative infancy of the sector, the cost of manufacturing processes can represent a significant proportion of the selling price, and in that regard is not as mature as the mAb sector. Currently, viral vector costs represent a major component of the material manufacturing costs for gene-modified cell therapies such as CAR T-cell therapies, and these costs are even more pronounced for higher-dose products such as hematopoietic stem cell (HSC) therapies.
Given the pressures to reduce the price of these therapies (e.g., approx. $400k for CAR T therapies and $1.8M for HSC therapies), there is a lot of interest to drive down viral vector cost contributions to these gene-modified cell therapy costs.”
With currently available production technology viral vectors represent anywhere from 15% to more than 50% of the COG of a gene-modified cell therapy. “For this reason this sector must shift away from lab-scale methods used nowadays to scalable alternatives. This will not only reduce costs but lead to more robust industrialized processes.”
Farid’s analysis illustrates how production scale and facility footprint change with different flowsheet options, degrees of process optimization, and market capture assumptions. “This helps determine whether processes will lead to practical facility sizes, and based on this the analysis can help prioritize R&D targets that will help with cost or space reduction.”