Addressing The Particulate Challenge In Cell Therapy Products

A conversation between Rafi Mohammad, Bayer, and Life Science Connect's Jon O'Connell

Cell therapies upend most traditional particle detection and mitigation strategies used for legacy biologics. That's because cell therapy products are suspensions of therapeutic cells, which are themselves particles. This makes it difficult to detect extrinsic particles in the turbid cell suspensions and to distinguish them from inherent product-related particles.

The conundrum obviates use of sterile filtration and light obscuration methods, and it puts mitigation wholly on the control strategy.

Rafi Mohammad, Ph.D., scientific director at Bayer, has spent most of his career working on formulation and CMC strategy for advanced therapy products. He offered to answer some of our questions about detection and prevention. Our interview has been edited for clarity.

What challenges do you face in detecting visible and subvisible particles in cell therapy drug products?

Mohammad: Cell therapy product consists of cells in suspension. The presence of turbid cell suspension makes the product’s appearance very cloudy, and it's challenging to detect visible foreign particles. In addition, it's challenging to differentiate the visible foreign particles from the inherent particulates. You have inherent particulates — product related particles, which may be acceptable when their presence is measured, characterized and determined to be part of the clinical profile (USP <1790>)— whereas the foreign, intrinsic to the manufacturing process particulates are not acceptable. These are called extrinsic and intrinsic particulates.

Another challenge is that, in a typical antibody manufacturing process the product is sterile filtered, so you have better control. For cell therapy products, there is no sterile filtration.

Finally, processing time can affect the overall stability of the product. When we are working through visual inspections, we cannot spend too much time inspecting or reinspecting.

In terms of the subvisible particulates, it's really challenging to differentiate those in cell therapy products. We will not be able to use traditional techniques like light obscuration to test for subvisible particles because the cells themselves are in the subvisible particulate range.

Recent evidence shows single-use components shed particles we can't filter out. How do you address this class of particles? Do you envision hybridization where high-shear zones use steel and fixed infrastructure instead of single-use technology?

Mohammad: We have to independently evaluate how cell therapy products behave under different manufacturing conditions and assess any overall impact. The control starts even before manufacturing; this includes, for example, understanding what kinds of controls the vendor has in place.

Single-use technologies have advanced considerably and are increasingly used in the manufacturing of cell therapy products. From a manufacturing perspective, single-use systems offer important advantages over stainless steel systems, including greater ease of use, faster changeover, and reduced cleaning requirements. However, it is important to characterize the particulate profile of single-use assemblies before implementation and to perform additional testing, as needed, to confirm that particulate levels are adequately controlled.

Can you talk about the work to build image libraries or other solutions to help operators distinguish between the product of interest and contaminants?

Mohammad: These are turbid cell suspensions, and understanding product behavior is important. Some of the cells can aggregate, and we need to demonstrate that they are product-related and won't have an impact on product performance. Next, we need a good defect library that can simulate all the visible particulates we can expect in the manufacturing environment.

Once we have a detailed characterization, we can then evaluate their impact. We should ask what kind of particulates we have in the suspension, and do they contain any extrinsic particulates? That's why visual inspection plays a very important role.

In addition, we want to have controls at each processing step, not only at the end. Then, when you're looking into the drug product vials, you have more confidence.

Beyond a good defect library and visual inspection qualification, the operator's training is important. They must be able to differentiate between the product versus non-product and understand the classification of these particles overall.

On that note, defect libraries are often proprietary. Do we need more robust public databases so that operator training doesn't have to start from scratch or developing them doesn't have to start from scratch per product per company?

Mohammad: The intent of defect libraries is to simulate your specific product and where defects are generated, such as, for example, in steps like manufacturing or filling.

We can leverage publicly available resources from literature and guidance documents as a useful starting point in early product development. Prior knowledge from these sources can help build initial defect libraries and support process-based simulations to predict the types of defects that may arise for a given product. However, because cell therapy products and manufacturing processes are highly product-specific, these libraries should be refined and tailored using product-specific data.

Thawed cryoprotectants often get mistaken for particulates. Could better formulation help prevent this one source of a false positive signal more than improving optics?

Mohammad: It's not just cryoprotectants. More broadly, it's helpful to consider all inherent or product-related particulates. We should explore approaches that we can leverage beforehand to control them.

But using cryoprotectants as an example, if we do more up-front cryoformulation development, really looking at optimization, we can minimize formation of these product-specific particulates.

That said, up-front optimization would help to minimize product-specific particulate contribution, but on its own, it's not going to fix everything. It must be part of the larger proactive control strategy.

What are your thoughts on phase-appropriate filtration? Is there a world where filtration with a marginal loss in viable cells becomes acceptable during manufacturing?

Mohammad: For cell therapy products, terminal sterile filtration — if somebody wants to use a 0.2-micron filter, for example — is not practically feasible. However, we can ask: Can we use any of the large and appropriate porosity filters/cell strainers to filter the product? It must be evaluated based on the product. If it's going to cause a lot of loss, then it may not be practical to implement.

On the other hand, sterile filtration is useful in other areas. For example, we can filter media and buffers to minimize particulates in later manufacturing steps.

What role, if any, does automated visual inspection have in particulate detection?

Mohammad: For large size allogeneic cell therapy batches, automated visual inspection is helpful to speed up the visual inspection process during manufacturing. Automated visual inspection is very challenging due to the turbid cell suspensions and there's a lot more up-front work is required to develop and validate the method. The AI/machine learning tools are being used for traditional biologics where the large data set becomes available and accordingly large data set is needed for cell therapy products to implement.

Robust libraries and control strategies can be all for nothing if particulates enter the product during administration. From a packaging or training perspective, what's the latest thinking on controlling contamination at the bedside?

Mohammad: We want to ensure that, even at the bedside, proper controls are in place. Dose preparation should be contained in a sterile environment, such as biosafety cabinets.

Even at the dose preparation step, closed and automated processing steps would help reduce the risk of introduction of particles. In addition, hospital staff must be fully trained to maintain aseptic conditions to ensure no other particulates are being introduced at the time of administration.

And the key considerations for patient administration are the route of administration, frequency of administration, compositions of particulates and the patient population which influences how we think about particulates. If something is being administered intravenously, then we must be more careful. The particulate controls are very stringent. Similarly, exceptional stringency applies when injecting into the brain, because any particulate introduction can have major adverse events.

Control starts before manufacturing, before you receive the raw materials including the packaging material, while you're manufacturing, after clinical sites receive it, and while preparing the doses and administering them to the patient.

About The Expert:

Rafi Mohammad is a scientific director at Bayer led cell and gene therapy drug product formulation development and CMC. He has extensive experience in developing and advancing formulation strategies that enable robust, scalable manufacturing complex biologics and cell therapies — helping bridge development into reliable clinical and commercial supply. He received his Ph.D. in internal medicine (drug delivery) from the University of Tokyo and performed postdoctoral research at University of California, Berkeley.