Each year, enormous investment and scientific effort advance drug candidates through rigorous animal studies, yet many still encounter unexpected challenges when evaluated in humans. These outcomes highlight an opportunity to strengthen how preclinical evidence is generated and interpreted. By leveraging human-based NAMs, the biomedical community can better anticipate human-specific toxicity, refine candidate selection earlier, and make smarter, more confident decisions. This approach aligns scientific models more closely with the ultimate goal of biomedical research: protecting human health while accelerating the delivery of safe, effective therapies.
The Limits of Traditional Toxicology
For decades, animal-based toxicology has served as the cornerstone of safety assessment. The approach is comprehensive, standardized, and required by regulators worldwide.
Yet, traditional toxicology has clear disadvantages. The issue lies in species differences, like variations in metabolism, receptor biology, and immune responses, that don’t always translate well to human outcomes. For example, thalidomide, found to be safe in rats, can cause devastating birth defects in humans. TGN1412, an antibody, cleared primate safety studies but triggered life-threatening immune responses in human volunteers.
In addition, while there are ethical constraints on how traditional toxicology is conducted, the rationale to find better ways to reduce, refine, and replace animal testing is as strong as the scientific one.
New Approach Methodologies: More Than Just Alternatives
New Approach Methodologies (NAMs) are more than "alternatives" as they represent a shift in how we assess toxicity. This new toolkit is expansive, including human cell cultures, organ-on-chip devices, computational models, and high-throughput screening platforms. Their unifying principle is simple yet powerful: complement, and when viable replace, animal models with systems that are more relevant to human biology.
The initial excitement around NAMs was about efficiency, but their potential runs far deeper. We can now use human-derived cell systems to capture genetic diversity. Organ-on-chip platforms faithfully maintain the complex tissue architecture and cell-to-cell signaling that traditional lab cultures destroy. Computational models can integrate data streams to predict outcomes no single assay could ever reveal alone.
The Systemic Toxicity Problem
So, why isn't the transition complete? Because toxicity is rarely a simple, one-step event. A chemical might damage the liver directly, or it might damage the kidney, which then releases inflammatory signals that harm the liver secondarily. It’s a systemic biological network, and traditional toxicology captured this complexity by the brute force of testing the entire organism. NAMs, in their focus on precision and reproducibility, isolate specific systems. That isolation is both their greatest strength and their biggest vulnerability.
For example, testing a compound on human heart cells in a dish can confirm direct cardiotoxicity. But it completely misses how that compound might affect the nervous system's control over heart rate, or how kidney impairment might slow the drug's clearance, leading to dangerous accumulation in the heart. Capturing these interconnected, systemic effects requires linking multiple NAM platforms in ways that accurately reflect physiological reality.
Combinatorial Approaches: Building Predictive Power
The cutting edge of NAM development lies in combinations. Researchers are developing integrated testing strategies that link multiple assays to capture biological complexity:
- Organ interaction platforms that connect liver chips to kidney chips via flowing medium, mimicking blood circulation
- Multi-endpoint panels that assess multiple toxicity pathways simultaneously in the same sample
- Computational models that integrate data from diverse NAMs to predict in vivo outcomes
These combinations do more than just add information; they create novel predictions. A liver model might show a compound’s metabolic pathway. A kidney model might show its excretion rate. But only by linking them can we see that toxic metabolites accumulate to dangerous levels that neither model predicts in isolation.
What This Means in Practice
Validation standards developed for animal studies don't always map cleanly to human-relevant systems. Risk assessment approaches built on animal dose-response data need recalibration for mechanistic human data.
For researchers and companies, NAMs offer tantalizing possibilities like faster iteration in drug discovery, better prediction of human-relevant toxicity, and reduced costs for safety assessment. However, implementation requires significant investment in new technologies, training, and validation studies.
For innovators, the challenge is clear: develop NAM platforms and testing strategies rigorous enough to satisfy regulatory requirements, human-relevant enough to improve prediction, and practical enough for widespread adoption. The gap between proof-of-concept and regulatory acceptance remains large, but is narrowing steadily.
The transformation of toxicology from purely animal-based to human-based testing isn't just scientifically sound, it's increasingly necessary. As precision medicine demands more nuanced safety data from biological complex systems, NAMs represent not merely an alternative path forward but the most promising one.
