A human induced pluripotent stem cell-based high-throughput screening platform for teratogenicity hazard identification
This paper presents a high-throughput, automated human induced pluripotent stem cell (hiPSC) platform that utilizes four germ layer models to accurately identify teratogens early in drug discovery, offering a scalable and efficient alternative to traditional animal testing.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Every new medicine carries a hidden question: is it safe for a developing baby? For decades, the answer has come too late. Drug developers typically test for birth defects only after a candidate has already spent years in the lab and clinical trials, often using animal models that are slow, expensive, and not always perfect mirrors of human biology. This delay creates a difficult reality for women of childbearing age. Because the risk is unknown until very late in the process, many promising drugs are abandoned entirely, or they reach the market with strict warnings that prevent women from taking them while pregnant or trying to conceive. The result is a gap in treatment options and a system that often fails to account for how drugs affect female biology. Scientists have long sought a way to peek at this risk much earlier, using human cells to see if a chemical might harm a developing embryo before it ever enters a living body.
A team of researchers at Evotec has now built a new tool designed to answer that question quickly and accurately. They created a high-speed screening platform that uses human stem cells—specifically, cells that have been reprogrammed to act like the earliest building blocks of a human embryo. These cells are coaxed into becoming four different types of tissue that represent the three main layers of the early human body: the lateral plate mesoderm, which helps form blood and bone; the definitive endoderm, which becomes the gut and lungs; the neural crest, which gives rise to nerves and skin; and neural rosettes, which are clusters of cells that will become the brain and spinal cord. By watching how these four distinct human tissues react to a chemical, the system can spot signs of trouble that might be missed by looking at just one type of cell.
The researchers put this system to the test with a large collection of 146 known substances. This group included 90 chemicals already proven to cause birth defects and 56 that are known to be safe. The goal was to see if the new platform could correctly sort them. The cells were grown in tiny wells, treated with the chemicals, and then examined with advanced microscopes that can count cells and measure their shapes in great detail. The system looks for two main things: whether the cells die or stop growing, and whether they fail to turn into the specific tissue they were supposed to become. If a chemical stops the cells from developing correctly, but does not simply kill them, the system flags it as a potential hazard.
The results were striking. When the researchers combined the data from all four tissue models, the platform correctly identified 88 percent of the known harmful chemicals. It also correctly recognized 75 percent of the safe chemicals as safe. In total, the system got the right answer 83 percent of the time. This level of accuracy is comparable to other lab methods, but the speed is what sets this new approach apart. Because the process is automated and runs on small plates, the team can test hundreds of chemicals in a single month, a task that would take far longer with older methods. This speed means drug developers can now check for birth defect risks right at the start of their work, allowing them to discard dangerous candidates early and focus their efforts on the safest possible molecules.
The study also revealed why some chemicals are harder to detect than others. For instance, the system found that a common vitamin-like drug called retinoic acid affects the gut tissue model much less strongly than the other three tissue models. By looking at the genetic activity inside the cells, the researchers discovered that the gut cells produce a specific enzyme that breaks down the drug, effectively protecting themselves. This kind of insight helps scientists understand not just that a drug is risky, but how it works, which is crucial for designing safer medicines. The researchers noted that the system did miss a few known harmful drugs, such as phenytoin, likely because those drugs cause damage through complex interactions between different organs or require specific genetic traits that a single lab dish cannot fully replicate. However, the platform successfully identified the vast majority of risks and provided a clear, human-relevant way to assess safety.
This work does not replace the need for final safety checks in animals or humans, which are still required by regulators before a drug can be sold. Instead, it offers a powerful new filter for the early stages of drug discovery. By catching potential dangers early, the platform could help bring more effective treatments to market faster and ensure that women have access to safer medicines without unnecessary restrictions. It represents a shift toward a more efficient and humane way of developing drugs, using human biology to protect human health from the very beginning of the process.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.