Kiji Therapeutics is developing gene-engineered, iPSC-derived MSCs, an approach that Miguel Forte, M.D., Ph.D., CEO and Co-Founder of Kiji Therapeutics and immediate past president of ISCT, described as a way to establish more consistent starting material. Its manufacturing strategy illustrates a broader question for developers: how can production expand while preserving the cell characteristics associated with the benefit seen in clinical studies?
Manufacturing Changes Can Reshape the Clinical Question
For MSC therapies, early evidence of benefit gives developers a reason to expand production. It also creates an obligation to understand whether the material produced through a revised process remains comparable to the material evaluated in patients.
Michael Schuster, Co-Founder and Head of Business Development at Mesoblast, stressed the need to watch the manufacturing process closely to preserve the product as production expands. His point has a clinical consequence: evidence generated under an earlier process cannot be assumed to apply to cells made after that process changes.
Automation, a move to a different production platform, or a revised formulation may all have sound operational reasons behind them. Each can also raise questions about cell function or potency. Forte stressed that characterization and comparability must preserve the relationship between the therapy’s biological activity and its documented benefit.
This makes manufacturing development part of the interpretation of clinical evidence. A program needs to explain both what its cells do and how it knows those properties have been retained as production evolves. Greater capacity alone does not answer either question.
A Consistent Cell Source Still Needs Functional Evidence
How developers approach this work depends in part on their starting material. With bone marrow donor-derived and umbilical cord-derived MSCs, they need to account for donor differences and cell expansion. An iPSC-derived approach offers another route to controlling the starting material.
Forte described establishing a master cell bank as the basis for reducing starting-point variability in Kiji Therapeutics’ approach. That is a useful manufacturing advantage, but it should be interpreted within its scope. Consistency at the beginning of production still needs to be connected to the function of the resulting therapy.
Schuster emphasized the role of screening, testing, and assays in checking that the intended cells have been selected and determining whether their properties have changed. A stable source alone does not resolve the biological question of which cell characteristics contribute to therapeutic activity.
For clinical development teams, the useful question is whether characterization can support a clear explanation of potency across production changes. That explanation provides the bridge between a controlled manufacturing process and the evidence used to judge treatment benefit.
Build Comparability Into Decisions Before Pivotal Development
An early development process may differ from the one needed for commercial production. Developers therefore face choices about when to optimize and how to assess the consequences of those changes.
Ed Ahn, Ph.D., CEO and Director of MEDIPOST, described these decisions as specific to each development team. The objective might involve time, cost, a process parameter, or product performance. He emphasized understanding the links among production conditions, cell biology, and clinical function early in development.
A practical decision framework would ask:
- Which operational problem would the proposed change address?
- Which biological characteristics could it affect?
- What characterization and potency evidence would support comparability?
- Can the change be evaluated before pivotal development adds pressure to finalize production?
These questions give teams a way to weigh trade-offs without directing them toward a single manufacturing strategy. Greater efficiency may justify a change. Teams still need to account for the work involved in showing that the therapy retains the properties underpinning its clinical evidence.
AI Depends on the Data Developers Can Provide
AI offers possible support for process development, automation, and analysis of complex manufacturing datasets. The discussion among the executives remained cautious about treating it as a general solution.
Ahn placed the larger expense in generating suitable training data rather than in the model itself. Forte also stressed the investment needed to adapt AI tools to a company’s own processes and datasets.
Developers therefore need to focus on building a data foundation that helps explain the relationship between manufacturing conditions and product behavior. AI may help them explore that relationship, but its usefulness depends on the information available and how specifically the task is defiThe core implications ultimately returnreturn to comparability. As an MSC program moves toward wider use, its clinical case should continue to reflect the product being manufactured. Teams have a clearer basis for expanding supply when they plan characterization, potency assessment, and process changes before pivotal development. This planning helps them avoid leaving unresolved questions about whether the resulting therapy still reflects the evidence generated in patients.
