Researchers in the life sciences industry have long been committed to finding preclinical models that more accurately reflect human biology than animal models can. But it wasn’t until December of 2022, when Congress passed the FDA Modernization Act, that the industry saw a clear marker in the sand declaring that it’s time to move forward and find new ways to replace animal testing with in vitro models. The new law allows developers of novel drugs to study safety and efficacy in the preclinical setting with animal alternatives such as cell-based and digital models.

As a result, there is increasing interest in using organoids—3D, human-derived cellular models—for disease modeling and drug discovery. But to capitalize on that interest and drive the adoption of organoids, the industry needs to address several challenges. Developing and scaling up organoids in a reproducible way is difficult and labor-intensive. And getting regulators on board with data generated from organoid research is a hurdle, because the relevance and reliability of data gained using organoids are still not as well understood as animal model data are.

The latest technology advances, particularly in artificial intelligence, are helping to advance organoid scalability and reproducibility, driving wider applications and regulatory acceptance. To make the most of these advances, the industry should embrace a three-pillar strategy:

1. Mapping organoids to complex human biology

Adult stem cells and induced pluripotent stem cells (iPSCs) have traditionally been the foundation of organoids, and incremental improvements in protocols have allowed organoid developers to model different tissue types, disease types, and patient scenarios in a dish. In recent years, advances in bioprinting have greatly improved the assembly of 3D models.

One key limitation has been that it’s difficult to build organoids composed of more than one embryonic germ layer lineage. Stem cell-derived organoids of the epithelium, for example, may lack stroma, vasculature, and other cell types. Researchers are able to add those separately, but it’s far from simple. Now there’s a growing demand for models that reflect the intrinsic complexity of the native tissue or disease status. Examples such as the tumor microenvironment, organ barrier functionalities like the blood brain barrier, or the gut/microbiome interactions are often utilized to understand interaction between an experimental drug and a diseased organ.

Bioprinting helps assemble these organoids, but it comes with challenges: The more complex the organoid, the more difficult it is to scale up and analyze in a reproducible way.

Here’s where AI can play an important role. AI-driven research tools can image organoids as they’re being built and determine whether the spatial orientation and morphology is correct, if the organoids are growing as expected, and when they will need media changes, growth factor additions, or passaging. Then, as the organoids are scaled up, the tools can predict those needs and automatically address them.

2. Improving scalability and reproducibility

The complexity of organoids raises the risk of variability, which in turn complicates scalability and reproducibility. Some variability is good—humans are intrinsically variable, and organoids should reflect that—but anything that impedes scalability and reproducibility should be addressed. For example, each lab follows its own protocols for growing organoids and achieving research endpoints, resulting in small variations in models obtained from different sources. The industry should work toward standardizing protocols to minimize this variability. Incorporating AI-driven decision-making during the culture process could be valuable in standardizing protocols.

Another major source of variability that the industry should work to eliminate is the reagents that are used for organoid culture. Right now, most organoids are cultured in animal-derived basement membrane extract gels, which are inherently variable. Researchers use a host of different growth factors and cytokines that can be variable, as well. One way to address this issue is to develop protocols for large batches of organoids that are all grown under the same conditions, with the same reagents, by the same operator. This will improve consistency.

Improving the analysis of organoids is another challenge. Researchers often want to use assays that they know, such as biochemical assays, which tend to be too simplistic because they reduce the complex biology reflected in an organoid down to a single data point. Better options are emerging now, including high-content, high-throughput imaging systems with AI-enabled image analysis software built in, which researchers can use to quickly analyze the morphology and the morphometry of their organoids so they can determine whether their assays are working properly. Other AI-powered instruments can accumulate data from multiple instruments and then parse out the details that are relevant to the experiment at hand.

3. Building acceptance for organoids among regulators

Improving scalability and reproducibility will help build acceptance for organoid research among regulators, but that’s just the first step in addressing the regulatory challenge. Drug developers need to develop protocols for managing large datasets and storing them in ways that will meet regulatory requirements for traceability. And they’ll need to standardize the creation of model systems that satisfy the FDA’s “context of use” provision, which requires developers to provide a clear description of the appropriate use and application of preclinical models in drug development and regulatory review, along with measures of quality control and assurance.

The industry can expedite the development of standards for organoid context of use by sharing information and experience through groups such as the International Consortium for Innovation and Quality in Pharmaceutical Development (IQ Consortium). The consortium, which brings together biopharma companies to develop solutions that will benefit the R&D community, has an affiliate that's focused on sharing experiences and best practices in building 3D model systems.

The more the industry gathers and shares data around the development of organoids, the more comfortable the FDA and other regulatory agencies will be moving away from animals in preclinical research and toward 3D models.

No doubt this evolution of preclinical research will ultimately be better for patients. One of the biggest challenges the industry faces is that only a small percentage of drugs entering clinical trials actually make it to market. That massive attrition is often the result of challenges with safety and translational efficacy—problems that could be more readily identified and solved with models that more accurately reflect human biology.

Nikki Carter is the Director of Commercial Organoid Innovation and Victoria Marsh Durban, Ph.D., is the Director of Custom Organoid Services at Molecular Devices.

Recent posts