“Replacing animal models means reshaping the scientific system”

As part of the European Commission’s Joint Research Centre, EURL ECVAM is tasked with some of the most foundational work in the field: developing, validating, and promoting alternative methods to reduce, replace or refine animal testing for scientific purposes. As such, EURL ECVAM shapes the very systems by which new methods become accepted for chemical assessment and, one day, an internationally accepted standard. “In some areas, we’re very close to phasing out animals,” says Dr. Ingrid Langezaal, toxicologist and scientific officer at EURL ECVAM. “In others, there’s still a long road ahead. But data helps us move forward.”

Mapping the road to regulatory acceptance

At the heart of EURL ECVAM’s infrastructure is TSAR: the Tracking System for Alternative Methods towards Regulatory Acceptance. Developed more than a decade ago, TSAR does what its name promises: it follows the journey of alternative methods submitted for safety testing through every milestone like submission, validation, peer review, international standardization, and final regulatory adoption. “It’s not so much a tool for researchers as it is a tool for dissemination and transparency,” Langezaal explains. “Regulatory agencies, validation bodies, and intergovernmental organizations like the OECD rely on TSAR to coordinate efforts, share progress, and avoid duplicating work.” Pierre Deceuninck, data scientist at the Joint Research Centre: “If you want to know how many non-animal methods have been accepted, or how long each phase takes, TSAR shows you. That’s crucial for building trust, in the methods and the process.” Still, TSAR has its limitations. “It’s mainly focused on regulatory toxicology,” says Deceuninck. “It doesn’t include early-stage research or exploratory methods that haven’t reached the validation phase.” For researchers working outside regulatory frameworks, especially in biomedical science, this creates a gap.

BimmoH

Set to launch later in 2025, the BioMedical Model Hub (BimmoH) is EURL ECVAM’s newest move yet to capture the explosion of non-animal models emerging from biomedical research. Powered by artificial intelligence, BimmoH is designed to extract, filter, and categorize the scientific publications that use human-biology based models allowing the use of non-animal approaches.

“This is not a database you fill out manually,” says Deceuninck. “It’s been trained on PubMed. It scans hundreds of thousands of papers and only includes those describing non-animal or hybrid models relying on human-biology. If a paper only reports on animal model use, it’s excluded.”

The AI sorts the models according to three main categories:

• Anatomy, histology, and cell types (e.g., liver, brain, skin)
• Clinical conditions and disease areas (e.g., cancer, respiratory diseases)
• Model type (e.g., organoids, in silico models, organ-on-a-chip)

“It’s the kind of tool we wish had existed  years ago when we launched our reviews of advanced non-animal models in biomedical research,” says Deceuninck. “Researchers will be able to search by body system, disease, or model type. That saves time, narrows searches, and makes the latest insights visible, something current tools like PubMed can’t do well.”

From data to decisions

Even with tools like TSAR and BimmoH, the road from innovation to implementation is long. “Validation is essential,” says Langezaal. “You can’t get regulatory acceptance without validation. That means not just proving a method works, but proving it’s reliable, reproducible, and fit for purpose across labs and sectors.” That’s why TSAR remains important. It captures the validation process, the peer-reviewed studies, up to OECD test guideline adoptions. “And it also provides open access to protocols,” adds Langezaal. “When a method reaches final validation, we share the actual procedures used to generate the data. That helps others build on the work.” Still, the process can be slow. “Validation is resource-intensive,” says Deceuninck. “And many cutting-edge models, like multi-organ systems or AI-driven predictions, are still too new or too diverse to standardize easily. That’s why tools like BimmoH are so important. They show what’s out there, even if it’s not yet validated.”

Debating ‘animal-free’

In conversations about non-animal methods, terminology matters. Langezaal points out that even some “animal-free” methods may include animal-derived reagents like fetal bovine serum or Matrigel. [1] “We recently mapped all the reagents used in a method validation study,” she says. “Almost all of them contained some animal-derived components.” Does that make them less valid? No, but it underscores the importance of clear definitions, especially when setting funding priorities or writing policy. “There’s a difference between reducing animal use, replacing whole-animal models, and achieving fully animal-free methods,” Langezaal explains. “We need to be honest about where we are, and where we’re going.” That’s another reason why EURL ECVAM makes documentation available. “Researchers can learn a lot from the methods in TSAR,” she adds. “Not just the protocols, but the entire validation journey, what kind of data was collected, how it was reviewed. That transparency helps move the field forward.”

_{[1] Bartnicka Joanna, Langezaal Ingrid and Coecke Sandra, Towards animal-free in vitro methods in the Thyroid Validation Study, EUR 30818 EN, Publications Office of the European Union, Luxembourg, 2021, ISBN 978-92-76-41164-2, doi:10.2760/544332, JRC125904.}

An ecosystem of tools

With the archiving of DB-ALM in 2019, a now-retired database of (validated) method protocols, EURL ECVAM recognized the need to evolve. “DB-ALM was useful in its time,” says Deceuninck, “but it only captured a fraction of the landscape of methods available, and maintaining it wasn’t sustainable. But it is still useful to collect or check certain information. That is why we archived it and provide the link on our website, so that method protocols can still be consulted.” The team shifted its focus to the other tools: TSAR for tracking, BimmoH for discovery, and comprehensive peer-reviewed reviews on specific disease areas, from cardiovascular to neurodegenerative conditions. “These reviews formed the foundation for BimmoH,” says Deceuninck. “We systematically analyzed thousands of models. Now, with AI, we’re upscaling that effort to keep up with the literature almost in real time.”

International harmonization

One of the biggest challenges in the adoption of alternative methods is international harmonization. “To replace animal testing at scale, we need more international agreement,” says Langezaal. “That’s why the preferred route is through the OECD, or the ICH in case of human and veterinary medicines. Their test guidelines are essential for regulatory trust. If you can prove a method is reproducible and meets a regulatory need, you can open doors internationally.” Meanwhile, tools like BimmoH support researchers, funders, and policymakers by offering a clearer picture of what’s already working. Deceuninck: “For example, we’ll even be able to identify models with the most citations, models that are already gaining traction. That’s valuable for everyone.”

What the future holds

Asked about their long-term vision, both Langezaal and Deceuninck are realistic. “Ten years won’t be enough to replace all animal testing,” says Deceuninck. “But in certain fields, like cosmetics, we’ve shown it’s possible. And in others, like oncology or cardiovascular diseases research, we’re seeing slow, but still, some movement.” It is still necessary to increase the level of trust in non-animal methods, and clarify how they can be used for regulatory decision making, Langezaal says. Deceuninck: “Our own analyses show that since around 2020, more papers are being published on non-animal models than on animal models. Tools like BimmoH can help speed that up. Science is diverse. There’s no one-size-fits-all solution. But with the right tools and mindset, we can rethink how we study disease, develop treatments, and protect human health, without relying on animal models.”

TSAR

DB-ALM

Interview: Hannah van der Vegt