Therefore, it is best if the CMO and client assess the quality of material produced at the CMO in a collegial and collaborative
manner. If there is a significant variance from expectations, there are several possible responses: 1) delayed progression
into GMP production while troubleshooting efforts are undertaken to determine the cause of the variance; 2) delayed progression
into GMP production to generate more pilot runs or engineering runs at scale that can define the variance, so that new specifications
can be established to release CMO-produced lots; or 3) live with the variance and proceed as planned, which is probably acceptable
only for early-stage projects.
Figure 3 presents one such example evaluation, where Millennium and a CMO statistically compared the peaks from an oligosaccharide
analysis of product from the pilot plants at both organizations. The analysis of both samples should be analyzed by the same
assay from the same laboratory, especially if the assays have not yet been qualified or validated, as is often the case during
Process transfer criteria also should be established for all downstream purification steps. Table 4 presents an example illustrating
the benefits. This table represents the process transfer of a polishing chromatography step to a CMO (a different CMO from
the one used in all the previous examples).
Table 4. Transfer of early-development process to a CMO. Data represents the purification yield and percentage of high molecular
weight (HMW) species in the load and eluate of a polishing chromatography step for lots from the Millennium (MPI) pilot facility,
the CMO pilot, and the full-scale manufacturing suite at the CMO.
In this case, because of time constraints and the early nature of the project, a full-scale engineering run was deemed prudent
because of the limited process history generated at that point. Millennium and the CMO decided to generate more process history
as part of the transfer by executing pilot lots in parallel at both sites. Data from these seven pilot lots are presented
in the first seven columns of Table 4. The percentage of high molecular weight species (% HMW) in both the load to and eluate
from this step was similar at both sites and the two organizations agreed that this step was transferred successfully.
However, after executing the first cycle of the full-scale engineering run, it was clear that the HMW percentage in both the
load and eluate were significantly higher than in the transferred process. Because the process was transferred successfully,
it was neither a process issue nor a transfer issue. Instead, focus was trained on the scale-up parameters and the full-scale
equipment. It was quickly determined to be an equipment issue with the full-scale chromatography skid (imprecise valve timing
resulted in brief product contact with sanitizing solution). After correction, the second cycle of the engineering run and
the next two GMP runs achieved a HMW percentages of 0.6%, 0.3%, and 0.3%.
Good transfer criteria based on solid process history can foster and enable effective client–CMO interactions. It is vital
to clearly demonstrate process reliability before a transfer and to set guidelines for evaluating a transfer based on client
process history. If these guidelines are followed, standards will be met and client–CMO interactions will be smooth.
Michael Glacken is the director of biologics process development at Millennium: The Takeda Oncology Company, Cambridge, MA, 617.679.7000,