Human cell-based models are rapidly becoming central components of modern drug discovery, as researchers study disease directly in patient-derived cells rather than relying on indirect animal or simplified culture systems. This shift has been accompanied by the rise of 3D organoids, improved automation, and a growing emphasis on reproducibility and regulatory alignment.
Steve Smith
Chief Executive Officer
iXCells Biotechnologies
In this Innovation Spotlight, Steve Smith, the chief executive officer of iXCells Biotechnologies, highlights this evolving landscape and discusses how innovations are helping to close the long-standing gap between preclinical findings and clinical outcomes. Advances in automation and strategic collaborations, such as the partnership between iXCells Biotechnologies and Rosebud Biosciences, are further enabling scalable, high-quality model development.
Over the past decade, how has the field of human cell-based disease modeling evolved?
The shift has been profound and continues to accelerate at exciting speeds given recent announcements from regulatory bodies. A decade ago, the field was largely built on animal models and simple cell cultures that were practical, but often poor predictors of what would happen in humans. Since then, we have seen a fundamental rethinking of where human biology needs to enter the drug discovery process, and the answer is much earlier.
Induced pluripotent stem cells (iPSCs) were the catalyst. Once you could take cells from a patient and reprogram them into virtually any disease-relevant cell type, the whole premise of preclinical modeling changed. As opposed to approximating human disease, researchers started studying it directly. Critically, specialized cell services and technologies such as reprogramming, genome editing, and directed differentiation have become much more accessible. Today, they are within reach of most research programs, which has helped to democratize the field in a way that's genuinely exciting.
Now, that foundation is maturing. We're seeing models evolve across multiple dimensions simultaneously; greater structural complexity with 3D organoids, more sophisticated co-culture systems that capture how different cell types interact, automation that makes high-throughput work feasible, and an increasing emphasis on personalization through patient-derived lines. The US Food and Drug Administration’s recent moves to prioritize human-relevant models have added regulatory momentum to what was already a scientific inevitability. The direction of travel is clear, and iXCells is well-positioned to support the needs of the scientific community.
What are the advantages of moving from traditional 2D cell cultures to more complex 3D models?
2D models have earned their place. They are accessible, scalable, and still valuable for a wide range of questions. We use them every day and will continue doing so without question. But there are limits to what a flat monolayer can tell you about how a disease behaves in the body.
In three dimensions, cells are organized the same way they are in tissues, simulating cell-to-cell and cell-to-matrix interactions, the biochemical gradients, and the spatial architecture that drive so much of human biology. Some disease phenotypes simply don't manifest in 2D. They only appear when you give cells the structural context that they need. This has direct implications on whether a drug candidate looks promising or gets correctly flagged as problematic before it reaches a patient.
The translational gap in drug development is a real and costly problem. A significant proportion of successful preclinical programs fail in the clinic, backed by a large amount of funding, with part of that failure tracing back to models that didn't accurately reflect human biology. More physiologically relevant models won't eliminate that gap entirely, but they can meaningfully reduce it by generating data that is more predictive from the start. We believe that this is where 3D systems, and organoids in particular, are starting to demonstrate their value.
3D organoid systems derived from patient cells recreate the structural and functional complexity of human tissues, offering more predictive insights into disease biology and drug response than traditional 2D cultures.
©iStock, RollingCat
As organoids and patient-derived models become more widely used, what are the biggest technical and regulatory challenges in ensuring reproducibility, scalability, and quality control?
This is the question the field must answer if these technologies are going to fulfill their promise, and it's one we think about constantly at iXCells.
On the technical side, controlling process variability is the central challenge in differentiation efficiency. Batch variation in the raw materials can reduce model reproducibility and undermine confidence in the data if not actively managed.
Early in the organoid field, differentiation protocols were highly heterogeneous across labs, which made it difficult to compare results or build on each other's work. This has improved considerably, in large part due to automation. Standardized, automated workflows reduce the human variability that inevitably creeps into complex biological processes, making it possible to generate data at a scale and consistency that translational research demands. Moreover, eliminating process-induced variables allows donor-to-donor differences (patient cohorts versus healthy controls) to be discovered, investigated, and ultimately understood, advancing our understanding of human biology.
In addition, regulatory expectations are also evolving. The US Food and Drug Administration and other agencies are moving in the right direction, but the field still needs to develop consensus around identity, potency, stability, and comparability standards for these models. That work is underway, and we see ourselves as an active participant in helping to define positive growth.
What does your recent partnership with Rosebud Biosciences entail, and why did you decide to work together?
Our partnership with Rosebud is designed to bridge this gap. Rosebud brings capabilities in organoid modeling at scale, enabled by their advanced automation infrastructure.
Our core strength is in provision of cells to build models with a focus in neuroscience, although we have completed projects across many specialties. Combined with Rosebud's automation depth and cardiac and renal modeling capabilities, we create something more valuable than either organization could offer independently. Researchers working on diseases that involve the brain, heart, and kidney can access integrated, high-quality models across those systems through a single relationship. The underlying need we're addressing is the gap between biologically complex models and the industrial-scale consistency required for serious drug discovery programs.
Sporadic amyotrophic lateral sclerosis (ALS) research lacks robust preclinical models. How is iXCells involved in changing that landscape?
Sporadic ALS is one of the hardest modeling challenges in the field. Unlike familial ALS, it doesn't have a single defining mutation you can engineer into a model. The majority of ALS patients fall into the sporadic category, yet most preclinical work has historically been done in transgenic systems that represent a small minority of cases. Models and patients have been misaligned from the start.
Our work through Project Mosaic directly targets that problem. By generating patient-specific iPSC-derived neuronal and glial models from sporadic ALS patients, we begin to capture the biological heterogeneity that defines this disease rather than flattening it out. The current focus is on deep characterization of 2D models, establishing the transcriptomic signatures that reflect different patient subtypes. The goal is to use those signatures to inform precision medicine strategies, matching patients with therapies based on their underlying biology rather than treating sporadic ALS as a single, uniform condition. Expanding into 3D systems is a natural next step as this work matures.
How do you see human-relevant models such as those developed at iXCells reshaping the future of research and drug development?
We are now at the beginning of a genuine inflection point. The core problem drug development has faced for decades is that models used to make early decisions have not been reliable predictors of what happens in humans. The result has been high attrition rates, enormous cost, and most importantly, patients waiting far too long for therapies that work.
Human-relevant models change that calculus. When you can study disease in patient-derived cells, generate consistent and functionally validated data sets at scale, and do it in systems that accurately reflect the biology of the patient cohort you are trying to help, you make better decisions earlier, advancing the right programs and stopping the wrong ones before the costs compound. For complex and rare diseases, that clarity can be the difference between a therapy reaching patients and a program that never gets there.
What we are building at iXCells is infrastructure at scale for that future: industrially consistent, deeply characterized, functionally validated models that meet the pace and rigor that modern drug development demands. The science is ready. The regulatory environment is moving in the right direction. The next opportunity is execution, which is what we at iXCells Biotechnologies are investing in and focused on.
