There is an obvious commercial imperative for small-batch cell therapies, but process changes cannot be made without careful consideration. Biological products can be sensitive to very minor manufacturing process changes. Even relatively small process differences can significantly affect the nature of a finished product — most importantly, how it functions in a patient's body. “When ‘the process is the product’ — as would be the case with industrialized production of stem cells — variability is the enemy and must be reduced and controlled as completely as possible” (4).
Successful biological manufacturing therefore addresses scale-up, process comparability, and process characterization. Because cell therapy products cannot be fully characterized in a laboratory, manufacturers must ensure product consistency, quality, and purity by ensuring that their manufacturing processes remain substantially the same over time. So companies must tightly control the source and nature of their starting materials and establish appropriate process controls for each unique product and/or manufacturing process.
Further complicating these matters is the fact that during research and early stage clinical trials, most cell therapies and associated process controls are developed and produced for a limited number of patients. They use technologies that are readily available rather than what will ultimately be needed for commercial-scale production. Typically those are manual, laboratory-scale technologies that are not optimized for specific cells or process needs.
Authors from Aastrom Biosciences, Inc. (www.aastrom.com) highlighted in 2008 the criticality of batch failure for autologous therapies: “The risk that a process will fail to yield a finished cell dose meeting specifications must be extremely low, whatever the reason. For PSCT, manufacturing failure is not merely an operational inefficiency issue; instead, it directly translates into a failed patient treatment” (5). By adopting a quality by design (QbD) approach to developing cell therapy processes, we can identify and address critical failure modes and process parameters.
For example, one useful measure of process risk is the number of sterile connections that must be made throughout cell therapy manufacturing. In a typical manual process, operators make multiple sterile connections in a biological safety cabinet (BSC) over several days to add reagents, collect quality control (QC) samples, and package a final product. In one relatively simple process we analyzed recently, we counted 76 sterile connections, all of which were performed manually. Each connection carries an inherent risk to the product. When patient numbers are small, such risk is theoretically manageable by skilled operators using good standard operating procedures (SOPs). But risk to products is considerable in commercial-scale production.
A natural consequence of the QbD approach to redesigning such a process would be to minimize the number of manual sterile connections. In fact, when this concept of minimizing sterile connections is taken to the extreme, it implies a process without any manual sterile connections. It is therefore hardly surprising that much innovation in small-batch manufacturing for cell therapies has revolved around using presterilized closed consumables (single-use technology).
Process analytical technology (PAT) has been defined as “a system for designing, analyzing, and controlling manufacturing through timely measurements (that is, during processing) of critical quality and performance attributes of raw and in-process materials and processes, with the goal of ensuring final product quality” (7). This philosophy for monitoring and controlling product quality is independent of batch size, and its implications can be profound as those batch sizes reduce.
For example, consider QC for an autologous therapy. Regulators are likely to require that the same tests be performed regardless of batch size. Each QC sample must be a specific volume to suit a given analytical method. Due to economies of scale, the volume of cells used for a large batch and the subsequent cost of performing QC tests are relatively minor. However, for an autologous therapy, such identical QC samples are likely to represent a significant proportion of the total cells available — and the total cost per patient.
Final-product release testing must satisfy a difficult combination of constraints. Such tests must be rapid (for products with short shelf-lives), relatively inexpensive (because they must be performed for every dose), limited in necessary sample volume (to prevent excessive loss of final product), and able to test a complex product composition. (5)
So for small batches, it is even more important that we invest efforts in characterizing processes so that we understand precisely which parameters need to be monitored, controlled, and tested — including operating limits within which a predictable outcome can be assured. If we combine the need to take samples for PAT and QC with the objective to use closed consumable processing, then we have a specific small-batch requirement for which a solution must be found. How do we efficiently collect samples from the right place, at the right time, without affecting product sterility (Photo 2)?