The regenerative medicine field is still quite young, and companies developing commercial processes for the production of cell and gene therapies face several hurdles. These products are highly complex and often involve production methods that, while in some ways are similar to those used in biologics and vaccine manufacturing, in many ways are quite different.
Often the level of understanding of the product and process requirements is less than desired. This results in part due to the complexity of the product, but also due to a lack of analytical tools that can provide relevant information with sufficient sensitivity. There is also a need for appropriately designed equipment and technologies. These factors make process development and validation for cell and gene therapies a challenging and highly critical task.
Depending on the application (i.e., autologous vs. allogeneic cell therapies or gene therapy vs. cell therapy applications), the focus of scaling of the manufacturing process may be different.
Scale up and mass production of the final product starting from a small seed bank (for instance, a master cell bank or working cell bank established from a healthy donor) is the main focus of the allogeneic applications, according to Behnam Ahmadian Baghbaderani, global head of cell and gene therapies for Lonza Pharma & Biotech. In comparison, for autologous applications, scale out and parallel production of multiple manufacturing processes (each starting from a patient sample) is the focus of scaling strategies.
Looking at cell therapies, Baghbaderani notes that the main differences in scaling allogeneic applications lies in the cellular characteristics, cell culture systems, and unit operations utilized in these processes, including the downstream methods employed to produce the final product. “The cellular characteristics (i.e., anchorage-dependent versus anchorage-independent cell types) or the biological mechanism through which the final cell therapy product is generated (e.g., tissue-specific stem cells undergoing expansion in the culture versus directed differentiation of pluripotent stem cells into specialized cells) can be the determining factors for the choice of unit operations, production strategy, and scale up considerations for the allogeneic cell therapies.”
For instance, the use of microcarrier systems versus aggregate cultures in suspension systems or the use of static two-dimensional (2D) large-scale cell-culture systems require different scale-up considerations, including the seed train strategy, the choice of bioreactor, bioreactor configurations, etc. In addition, the downstream processing steps can vary significantly depending on the product specifications and available downstream technologies, such as tangential flow filtration, counter-flow centrifugation, and continuous centrifugation methods.
Importantly, the process control strategy could be different for different cell therapy applications. “While production of large numbers of actively proliferating cells would be the primary focus for tissue-specific stems cells (e.g., mesenchymal stem cells), the control strategy for directed differentiation of pluripotent stem cells into specialized cells requires tight control of change in cell phenotype from pluripotent into multi-potent stem cells or progenitor cells and eventually fully committed cell types,” Baghbaderani explains.
The fact that cell therapies are living products is a key difference when looking at cell versus gene therapies, according to Alain Lamproye, CEO of Yposkesi. “It is essential to have a sufficient amount of starting material cells that will provide a sufficient quantity of the final product. In addition, end sterilization is not possible, so the scaled expansion and production processes must guarantee sterility. For in-vivo gene therapy applications, on the other hand, the final product is a viral vector containing the gene of interest. The expansion process step is followed by clarification, purification, and final sterilization steps. The additional steps aim to increase the quality of the product, but may also reduce the quantity,” he observes.
Similar process development challenges
Early processes developed for cell therapies involve many manual operations and have been—according to Lamproye—created using art more than science, and lack necessary control. “To develop industrial processes,” he says, “a great deal of understanding of the production process is required, starting with the demonstration of repeatability.” In addition, companies must not only drastically increase manufacturing capacity, but at the same time reduce raw material (media, culture ware/bioreactors, etc.) costs to make their products accessible to more than the few people that participate in clinical trials.
“To develop a process that is repeatable and transferable to large scale is not trivial. The challenge to increasing capacity for a cell therapy process is the biological variation of cellular starting materials. Statistically relevant measurements must be repeated many times to improve process yields and reduce manufacturing cost, which adds cost and time to the development process,” Lamproye observes. He points in particular to the avoidance of the use of plasmids and the development of packaging/stable cell lines as current key issues.
Reducing manufacturing costs, as well as development of scalable, automated manufacturing processes under closed systems that can reach the target quantities while maintaining the critical quality attributes (CQA) of the final products are main challenges for allogeneic cell therapies, Baghbaderani agrees. In this context, he identifies multiple important challenges:
- Modification of 2D cell-culture systems into a scalable, three-dimensional (3D) suspension-culturesystems with adequate control strategies
- Completely closing processes from the starting materials to generation of final products (in particular when starting from cell-isolation steps)
- The ability to control the quality of intermediate materials during directed differentiation processes for iPSCs
- Technological limitations associated with automation of manufacturing processes
- The pace of development of automated downstream processing technologies, which have not caught up with the scale required for some of these therapies
- The fact that research-grade raw materials are often used in baseline processes for which equivalent GMP-grade materials or high-quality materials generated by qualified vendors are not available
- A lack of adequate process characterization tools and technologies.
Many of these challenges apply to the development of viral-vector manufacturing processes as well. The biggest challenge, according to Baghbaderani, is the development of large-scale manufacturing processes using 3D computer-controlled suspension bioreactors with appropriate downstream processes. “Gene therapies require high productivity to meet the high dose demands. The productivity largely depends on the harvest titer and downstream process yield. Moreover, the product design at the molecular level, the producer cell line, and the production medium can all be limiting factors for harvest titer. Most importantly, appropriate analytical methods and assays are needed for implementation of proper in-process control strategies needed to achieve the desired quality and quantity of the final product,” he comments.
On a big-picture level, the increased demand for viral-vector manufacturing services to support clinical trials for recombinant adeno-associated virus (rAAV)-based gene therapies has led to a shortage of production capacity. Wait times for production slots can be 18 months or more (1). “As rAAV demand goes unmet, companies face lost opportunities, patient access to existing treatments is reduced, and development plans for new gene therapies stall,” Lamproye asserts.
Common validation hurdles
One of the primary requirements for process validation is to demonstrate documented evidence that the scaled process can consistently produce the final product, meeting the CQAs based on the specified critical process parameters and critical material attributes. To achieve this goal, according to Baghbaderani, it is crucial to perform characterization and process limit evaluations of the scaled manufacturing process. “A big challenge with allogeneic applications is making sure the identity and the quality of the final product is controlled within the design parameters of the process. Lack of adequate process characterization and appropriate in-process controls due to the myriad number of process variables along with poorly defined CQAs poses significant challenges for characterization and validation of allogeneic cell therapies,” he states.
Demonstrating repeatable processes when there is biological variation due to the starting material—a key challenge mentioned for process development—also creates difficulties for process validation, agrees Lamproye. “Production engineering principles require a controlled process to utilize well-understood and consistent methods for production. The ability to generate and manufacture well-characterized batches of cells for therapeutic use requires significantly different approaches to typically implemented laboratory-based cell-culture techniques, protocols that are often labor-intensive and limited to small scale,” he says. In many cases, it is necessary to create a new process and then perform sufficient runs to generate enough data to demonstrate consistency, an effort that comes at a high cost and consumes significant resources.
For gene-therapy processes, similar to allogeneic processes, poorly defined CQAs for the final product and inadequate process characterization (potentially due to the numerous process variables and lack of appropriate in-process monitoring and controls) are the main challenges for process validation, according to Baghbaderani. Precise in-process monitoring of complex parameters (metabolite concentrations, pH, oxygen and carbon dioxide, number and viability of the cells) within the bioreactor is necessary to gather the data needed to enable effective scale up of upstream processes, Lamproye adds. In many cases, he notes that the need to develop, validate, and implement novel analytical techniques is an added challenge.
Some new tech is helping
Overall, the pace of technology development is not in harmony with the growth in the number of cell and gene therapies moving through the clinic, according to Baghbaderani. There are, however, some new technologies that have been helpful with respect to the development of robust and reproducible manufacturing processes.
For cell-therapy applications, Baghbaderani points to the Cocoon Platform from Lonza, an automated GMP-in-a-box concept for patient-scale cell therapy manufacturing, other iPSC platform technologies that allow the generation and characterization of high-quality pluripotent stem cells, and new downstream technologies specific for processing cell therapy products, such as Acoustic Cell Processing (FloDesign Sonics), the Sepax Cell Separation System (GE Healthcare), and the LOVO system (Fresenius Kabi).
For Lamproye, a key advance for cell-therapy manufacturing has been the development of bioreactors that enable 3D production using systems that were originally designed for 2D flatware. “The ability to transfer production systems for adherent cells to bioreactors aids in the scale up of these processes. Bioreactors can be used as closed systems and provide the opportunity to scale up to larger volumes. And importantly, the science of bioreactor scaleup is better understood and generally requires few process changes,” he explains.
For gene therapies, advances in analytical techniques designed to address the needs of these complex products, such as digital droplet polymerase chain reaction, flow cytometry, and high-performance liquid chromatography, have enabled more consistent and accurate measurement of quality attributes, observes Baghbaderani.
Scale out of autologous therapies
With autologous therapies, scaling studies actually focus on scale-out, with the implementation of multiple parallel processes for manufacturing of final products, each of which uses material from one patient. “To produce products for large numbers of patients would require running of hundreds to thousands of parallel processes in a cleanroom environment, which is not feasible. It will be necessary to develop an automated closed system with a minimal footprint that can be placed in a cleanroom space yet still maintain material traceability (including chain of custody). Such a strategy would help reduce the overall cost of goods, making the process more viable for commercial manufacturing and ultimately offering more affordable treatments,” Baghbaderani asserts.
Any production system that is used must also ensure that processes are robust enough to enable the use of raw/starting materials with significant variability, adds Lamproye. “For instance, the variation of transplanted CD34+ cells/kg can be up to four orders of magnitude around the median for autologous purposes. Strategies for autologous cell and gene therapy production must therefore be sufficiently robust to enable the development and implementation of effective in-process controls,” he states.
The process is the product
Although there are common unit operations and similarities between manufacturing processes for cell and gene therapies, every manufacturing process is fundamentally different and unique due to the number of process variables involved in each process, the definition of the relevant CQAs, and the fact that characterization tools may be lacking, according Baghbaderani. “It can therefore be argued that the process is the product to a large extent, and it is crucial to follow best practices during early development studies and carry out proper risk assessments for each process,” he remarks. One general approach to minimize variabilities and manufacturing risks is to minimize suboptimal, open, manual unit operations and replace 2D cell-culture systems with computer-controlled, 3D suspension bioreactors.
The focus for many manufacturers, whether innovators or contract development and manufacturing organizations (CDMOs), is on automation. “Many CDMOs are now integrating sufficiently flexible production workflows by utilizing automated systems. This approach can potentially enhance process and product standardization, provide effective tracking of processes and in-process controls, and lead to higher throughput and simpler processes,” Lamproye observes. Automation and implementation of in-process controls and in-process monitoring technologies during early development, adds Baghbaderani, can significantly prevent future delays and manufacturing or commercial failures for cell and gene therapy products.
1. C. O’Donnell and T. Mathias, “Pfizer, Novartis Lead $2 Billion Spending Spree on Gene Therapy Production,” Reuters.com, Nov. 27, 2019.