The Drugs in Your Medicine Cabinet Were First Tested on Mice. A NASA Experiment May Change That.

The drugs you take were probably first tested on mice. A technology meant to change that — testing drugs on human organ tissue instead of animal models — has been building toward this moment for over a year, and the regulatory pieces are finally aligning to make it happen. This is where the science stands as of early 2026: the FDA has issued a draft validation framework for organ chips, a NASA experiment has returned from deep space with the first in-vivo comparison data, and the pharmaceutical industry is beginning to treat human-relevant testing as something worth building around. Whether it becomes routine depends on questions the data hasn't answered yet.

The shift has been building for more than a year. In April 2025, the FDA released a strategic roadmap to make animal testing "the exception rather than the rule" for preclinical safety and toxicity studies within three to five years, and the NIH reinforced that same month by requiring all new funding notices involving animal models to incorporate human-focused approaches like organ chips. That policy foundation matters: it means the regulatory infrastructure is now actively encouraging adoption, not just permitting it. GEN News

The technology has been building toward this moment. Emulate — which commercializes organ-chip technology developed at the Wyss Institute for Biologically Inspired Engineering at Harvard — launched its AVA Emulation System in June 2025. The instrument is a self-contained benchtop unit that cultures, incubates, and images up to 96 independent organ-chip samples in a single run. Compared to the previous generation, it cuts consumable costs fourfold, requires 50% fewer cells and media per sample, and reduces hands-on lab time by more than half. A typical seven-day experiment generates more than 30,000 time-stamped data points from daily imaging and effluent assays, with post-experiment omics pushing the total into the millions. Emulate

Emulate's Liver-Chip S1 has already produced the kind of data that makes regulators pay attention. In a 2022 study published in Communications Medicine, the liver chip showed 87% sensitivity and 100% specificity for drug-induced liver injury — outperforming animal models on the metric that matters most for patient safety. It was the first organ chip admitted to the FDA's Innovative Science and Technology Approaches for New Drugs (ISTAND) program, which supports tools that fall outside existing qualification programs but may still be useful for drug development. Emulate Lorna Ewart, Emulate's chief scientific officer, described 2025 as a pivotal year — both for the external regulatory momentum and for the company's internal product development. "Moving towards reduction and in some cases replacement of animal models demands both biological fidelity and throughput," she said. "AVA meets those dual requirements."

CN Bio, a U.K.-based organ-chip company built on technology from MIT's Linda Griffith lab, has taken a parallel path. Its PhysioMimix platform supports single- and multi-organ models, with applications in toxicology, drug-induced liver injury, and disease modeling for metabolic liver disease, COPD, and brain drug delivery. The company is currently working to add immune cell integration to its liver model — a technically demanding step that would allow customers to study the toxicity of monoclonal antibodies in a human-relevant context. GEN News CN Bio's CSO, Tomasz Kostrzewski, is direct about the misconception the field still battles: the idea that organ chips can fully replace animal models today. "The focus should be on using these tools to answer the right question and in the right context of use at the right time alongside all those other approaches that are out there."

The bone marrow experiment on NASA's Artemis II mission illustrates both the ambition and the challenge. In April 2026, the four Artemis II astronauts launched carrying an unusual piece of cargo: organ chips containing cells from their own bone marrow. The experiment, called AVATAR (A Virtual Astronaut Tissue Analog Response), placed bone marrow chips alongside the crew inside the Orion spacecraft while an identical set remained on Earth. The astronauts have now returned. Researchers are comparing both sets — the first in-vivo human comparison data for organ-chip technology generated under conditions no earthbound lab can replicate. Deep space radiation and microgravity are hazards that cannot be simulated on the ground. The bone marrow was chosen because it is particularly vulnerable to radiation and because it produces the circulating blood cells that are measurably altered in some astronauts after spaceflight. Wyss Institute

That is the honest version of where organ chips are. Not a replacement for animal testing — not yet — but a set of tools that are increasingly generating the kind of data that regulators and drug developers cannot ignore. The pharmaceutical industry has billions of dollars invested in animal-model infrastructure, standardized protocols, historical databases of what animal results predicted about human outcomes, and a workforce trained to work with animal models. The organ-chip sector is a fraction of that size. Adoption is slow not because the technology fails but because the ecosystem around it — the validation standards, the regulatory comfort, the established workflows — is still being built. Morgan Lewis

In March 2026, the FDA issued a draft guidance establishing a validation framework for new approach methodologies (NAMs) — including organ chips — built around four principles: context of use, human biological relevance, technical characterization, and fit-for-purpose. The guidance also includes a clarification that matters for adoption: a fit-for-purpose NAM may be used in a regulatory submission even without full validation. That is not a small carve-out. It means the path forward is incremental — use by use, context by context — rather than a years-long comprehensive validation that precedes any deployment. FDA

"Technological advances are allowing us to move beyond animal testing in drug development, which has a poor track record of predicting safety and efficacy in humans," said FDA Commissioner Marty Makary when the draft guidance released. The framing was direct. The biomedical establishment has known for years that animal models fail in ways that matter. What it lacked was an alternative that could operate at the scale and cost structure the industry requires — and a regulatory framework that would accept the results.

The deeper implication is one that the industry is only beginning to articulate. The entire biomedical research system was built around a constraint: direct testing on human biology was too slow, too expensive, and too risky for early-stage discovery. Animal models were a proxy — imperfect but workable. The constraint is now lifting. Organ chips, combined with advances in computational modeling and AI-assisted analysis, mean that for some testing scenarios, the proxy is no longer necessary. What happens when you can test directly on human-relevant platforms at scale?

The answer is not simply faster drug development. It is different drugs. There are classes of targets — in rare diseases, in pediatric conditions, in human-specific biology where animal models structurally cannot capture the relevant physiology — that have been economically and scientifically intractable because the testing platform could never deliver a clean answer. If human-relevant platforms become standard, those targets become tractable. That is the unlock that the FDA's roadmap is quietly engineering toward. It is not a guarantee. Adoption remains early, validation standards are still being defined, and the pharmaceutical industry's inertia is real. But the direction is clear, and the clock is now regulatory, not technological.

The AVATAR comparison data will take time to fully analyze. That will not settle whether organ chips can replace animal models in drug development. But it adds the most convincing evidence yet that the replacement is no longer a question of if — only of when. NASA Science