FDA roadmap to reducing animal testing: a regulatory and scientific paradigm shift in nonclinical safety assessment

Abstract

On April 20, 2026, the U.S. Food and Drug Administration FDA published a 1-year progress report of its initiative to reduce animal testing requirement in drug development. The FDA roadmap, announced on April 10, 2025, represented a pivotal paradigm shift toward human-centric New Approach Methodologies (NAMs), including stem cell-derived organoids, organs-on-chips, in silico modeling, and AI-enabled tools. The aim is to address the longstanding challenges of conventional preclinical safety assessment using animal models, including poor translational predictability, high cost, ethical concerns, and inherent interspecies biological differences, resulting in high clinical attrition and delayed access to effective therapies. In this commentary, we critically evaluate the scientific rationale, first-year implementation progress, and global regulatory impact of the FDA initiative. We highlight landmark advances including the permanent Innovative Science and Technology Approaches for New Drugs (ISTAND) program, human-centric validation principles, streamlined nonclinical frameworks for biologics, and alignment with global agencies and regulators. We also discuss existing, persistent challenges, such as uneven validation across toxicological endpoints, incomplete global data sharing, and cultural inertia, and propose actionable strategies to accelerate the safe, systematic adoption of NAMs in regenerative medicine and drug development.

Introduction and rationale

The 2026 progress report released in April by the US Food and Drug Agency (FDA) on reducing animal use in nonclinical safety testing, entitled “Reducing Animal Testing in Nonclinical Studies: Year One Progress and the Path Forward”, represents a measurable and significant departure from long-standing regulatory reliance on animal models, advancing a human‑centric, technology‑integrated framework that reshapes the standards of evidence, efficiency, and ethical conduct in drug development [1]. For decades, preclinical safety evaluation has centered on rodent and non‑human primate studies as the primary gateway to human clinical trials, a practice rooted in tradition rather than consistent predictive success. Mounting evidence of translational failure, economic unsustainability, and ethical strain has rendered this model increasingly untenable, creating an urgent mandate for regulatory modernization [2]. The FDA’s one‑year progress report does not merely document incremental policy change; it establishes and institutionalizes a new scientific and regulatory logic in which human‑relevant New Approach Methodologies (NAMs), as presented in the FDA’s April 10, 2025 press release, “FDA Announces Plan to Phase Out Animal Testing Requirement for Monoclonal Antibodies and Other Drugs” [3], are formally positioned as the default and preferred foundation of safety assessment, with animal studies reserved as a supplementary or exceptional recourse rather than a default requirement.

The limitations of animal‑centric preclinical testing extend far beyond ethical concerns to fundamental, irredeemable scientific inadequacy. Interspecies divergence in disease pathogenesis, immune function, metabolic pathways, pharmacokinetic properties, and cellular and molecular responses creates a persistent translational gap that no procedural refinement or model optimization can fully overcome [4]. More than 90% of drug candidates that demonstrate favorable safety and efficacy profiles in animal studies fail during human clinical development, most commonly due to unanticipated toxicities or insufficient therapeutic activity that animal models failed to reveal [4]. The TGN1412 incident is a defining illustration of this risk [5]. A humanized monoclonal antibody, TGN1412 (Theralizumab), designed as a CD28 superagonist to treat B-cell chronic lymphocytic leukemia, rheumatoid arthritis, and other immune-related diseases by expanding regulatory T-cells, which was well tolerated in non‑human primates at doses hundreds of times greater than the intended human exposure, triggered life‑threatening cytokine storm in human volunteers, driven by structural and functional differences in CD28 signaling that animal testing could not detect. Such discrepancies are not anomalous but characteristic of broader limitations in animal models for cancer therapies, Alzheimer’s disease interventions, and inflammatory conditions, where genomic and physiological responses in laboratory species correlate poorly with human biology [6–8].

These scientific weaknesses are compounded by economic and logistical burdens: monoclonal antibody development frequently requires more than 140 non‑human primates per program, with individual animals costing up to US$50,000, extending development timelines to nearly a decade, and driving total program costs above $650 million [9, 10]. Beyond terrestrial laboratory species, the annual harvest of more than one million horseshoe crabs for Limulus amoebocyte lysate reagents used in endotoxin testing introduces ecological strain and conservation risk [11], while the physiological stress of laboratory housing alters immune, metabolic, and inflammatory phenotypes in research animals, introducing confounding variability that distorts the reliability of study endpoints [11]. Together, these factors form a system that is slow, costly, ethically fraught, and misaligned with the core objective of reliably forecasting human drug responses.

First-year implementation and progress

Against this backdrop, the FDA’s first‑year implementation of its animal testing reduction roadmap has translated strategic vision into durable institutional and regulatory change. A central achievement has been the restructuring of internal governance to support cross‑disciplinary evaluation of NAMs, which inherently transcend the therapeutic‑area silos that structure traditional regulatory review. The agency has established agency‑wide coordination committees, specialized technical workgroups, and a planned NAM Integrated Review Team to centralize expertise, ensure consistent evaluative standards, and streamline integration of novel tools across drug classes and safety endpoints.

Concurrently, the elevation of the FDA Innovative Science and Technology Approaches for New Drugs (ISTAND) program from pilot to permanent status has created a predictable, transparent pathway for qualifying novel drug development tools, attracting more than 16 active industry submissions and resolving longstanding uncertainty that discouraged investment in alternative methods. Formalized collaboration with the National Institutes of Health (NIH) through the Complement‑Animal Research in Experimentation (ARIE) initiative further aligns translational research capabilities with regulatory priorities, ensuring that emerging technologies are developed with regulatory applicability in mind rather than advancing in scientific isolation. This integrated institutional architecture addresses a longstanding barrier to NAM adoption: the fragmentation of regulatory oversight across disease‑specific divisions, which historically created inconsistent expectations and delayed acceptance of cross‑cutting technologies such as computational toxicology and organ‑on‑chip systems, the latter often derived from the application of stem cell-based tissue engineering.

Regulatory transparency has emerged as an equally powerful driver of behavioral change across the drug development ecosystem. The launch of a searchable, regularly updated database documenting acceptable NAM applications has replaced implicit reviewer discretion with explicit, accessible guidance, eliminating the guesswork that previously deterred sponsors from pursuing animal‑sparing study designs. A retrospective 15‑year analysis of NAM submissions further refined strategic focus by revealing that 93% of NAMs used in drug applications submitted to FDA fall into in silico or in vitro categories, enabling targeted resource allocation toward the most widely adopted and technically mature platforms [12]. For example, technical publications refining protocols for liver‑on‑chip systems have directly addressed reproducibility barriers in evaluating drug‑induced liver injury [13], a leading cause of clinical attrition and post‑market drug withdrawal [14]. The most conceptually transformative step came in March 2026, when the FDA released a draft guidance establishing four core validation principles—Context of Use, Human Biological Relevance, Technical Characterization, and Fit‑for‑Purpose—and reversed the long-standing burden of proof [15]. Rather than requiring sponsors to demonstrate that NAMs perform equivalently to animal studies, the agency now requires adherence to human‑centric validation criteria, establishing NAMs as the conceptual default in regulatory decision‑making. This subtle but profound shift legitimizes integrated, multi‑modal evidence packages as scientifically superior to single animal studies, rather than treating them as discretionary concessions to animal welfare.

Immediate reductions in animal use underscore the roadmap’s tangible real‑world impact. The December 2025 draft guidance on monoclonal antibodies eliminated routine six‑month non‑human primate studies, replacing rigid testing mandates with weight‑of‑evidence frameworks that combine shorter‑duration studies, in vitro data, computational modeling, and other human‑relevant assays [16]. This single regulatory adjustment is expected to spare hundreds of non‑human primates annually while preserving stringent patient safety standards, and in select cases may eliminate primate testing entirely. Parallel progress has been made in reducing reliance on horseshoe crabs, with updated guidance supporting the transition from animal‑derived Limulus amoebocyte lysate reagents to recombinant alternatives that perform equivalently without ecological harm [17]. The qualification of AIM‑NASH (Artificial Intelligence-Based Histologic Measurement of Nonalcoholic Steatohepatitis) by FDA and EMA (European Medicines Association) as the first AI‑enabled drug development tool marks a critical technological milestone, validating machine learning as a legitimate component of regulatory assessment and opening avenues for broader integration of computational approaches in toxicity screening, histological scoring, and clinical endpoint evaluation [18]. These advances are not isolated victories but interconnected steps in a systemic shift toward efficiency, human relevance, and ethical stewardship.

International regulatory alignment has amplified these domestic achievements by transforming national policy into global precedent. The 2025 FDA‑NIH workshop brought together leading global regulators, including the EMA, Japan’s Pharmaceuticals and Medical Devices Agency (PMDA), Germany’s Federal Institute for Risk Assessment (BfR), and Australia’s Therapeutic Goods Administration (TGA), to harmonize strategies for NAM adoption and weight‑of‑evidence evaluation [19]. Such communication and coordination are essential, as divergent national requirements historically forced sponsors to retain animal testing to satisfy the most conservative jurisdictions, undermining progress even in regions with progressive regulatory frameworks. By aligning validation standards, regulatory expectations, and implementation roadmaps, the FDA has reduced cross‑border frictions and created a supportive global environment for animal‑sparing development practices. Engagement with the Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM) and the International Council for Harmonisation (ICH) further institutionalizes this alignment, ensuring that scientific progress scales into a cohesive global regulatory framework.

Current challenges and barriers

Despite strong early momentum, there are scientific, structural, and cultural challenges to the long‑term success of FDA’s initiative. Validated NAMs remain unevenly distributed across toxicological endpoints; for example, while there are robust alternatives available for skin sensitization and mutagenicity [20, 21], there is limited coverage for high‑stakes domains, including developmental and reproductive toxicity, chronic carcinogenicity, and long‑term organ damage [22, 23]. This fragmented validation landscape forces continued animal use for critical safety assessments, constraining the overall scope of reduction. The vision of a comprehensive, open‑access global toxicity data repository remains incomplete, hindered by inconsistent data formats, unresolved governance structures, and incomplete international data‑sharing agreements. Without a unified, interoperable database, redundant testing continues and the full evidentiary value of existing studies remains underutilized, perpetuating inefficiency and unnecessary animal use.

Cultural and technical barriers persist across the drug development ecosystem. Decades of institutional tradition have entrenched reliance on animal models among researchers, sponsors, and regulators, many of whom lack specialized training in NAM design, implementation, and interpretation. Weight‑of‑evidence frameworks, while scientifically flexible and powerful, demand integrated evaluation of diverse data streams that requires specialized expertise not uniformly distributed across regulatory review teams. Small biotechnology firms and academic laboratories face additional financial barriers, as upfront investment in NAM validation and standardization may outweigh short‑term cost savings, even with clear regulatory acceptance. These challenges are not failures of policy implementation but reflections of the systemic nature of the transition; replacing a century‑old testing paradigm requires more than regulatory updates—it demands sustained investment in workforce expertise, technical infrastructure, and targeted incentives to ensure equitable participation across the innovation landscape.

Strategic directions and future actions

To realize its long‑term vision, the FDA must build on first‑year successes with deliberate expansion and institutional deepening. The streamlined testing frameworks validated for monoclonal antibodies provide a replicable model that should be extended systematically to other biologics, small molecules, gene therapies, cell therapies, and medical countermeasures, supported by rigorous scientific oversight to maintain uncompromised safety standards. Validation efforts should include priority for under-covered endpoints, particularly developmental toxicity and chronic carcinogenicity, using the 15‑year NAM landscape analysis to focus resources on in silico and in vitro tools with demonstrated industry uptake. Implementation of quantitative animal‑use tracking systems will strengthen accountability, enabling data‑driven priority adjustment and transparent public reporting of progress. Completion of the global toxicity data repository, supported by standardized data formats and binding international agreements, will eliminate redundant testing and strengthen the collective evidence base that supports regulatory decisions.

Sustained progress also requires intensified international harmonization as NAMs grow more technically diverse and context‑specific. Continued collaboration among global regulatory bodies will prevent divergent standards from eroding gains and ensure that animal reduction delivers consistent benefits across global markets. Broader cultural transformation, supported by expanded training programs, integration of NAMs into academic and professional curricula, and widespread dissemination of real‑world case studies, will solidify human‑centric testing as the default mindset rather than a specialized alternative. Targeted incentives, including research grants, priority review vouchers, and public‑private partnerships, will democratize access to NAM technologies, ensuring that small developers and academic researchers can contribute to and benefit from the transition without being priced out by upfront costs.

Conclusion

The FDA’s one‑year progress in reducing animal testing represents far more than a regulatory update; it embodies a fundamental reorientation of nonclinical safety assessment toward human relevance, scientific rigor, and ethical responsibility. By moving beyond incremental refinement of animal use to systematic adoption of human‑centric tools, the agency addresses the core limitations of traditional preclinical research while delivering concurrent benefits to patient safety, development efficiency, and animal welfare. The institutional infrastructure, regulatory frameworks, and global partnerships established in the first year provide a resilient foundation for continued advancement, reducing uncertainty for industry, standardizing evaluation for regulators, and accelerating the translation of innovative technologies into routine practice.

As the roadmap advances into its second and third years, success will be measured not only by the number of animals spared—hundreds of non‑human primates and more than one million horseshoe crabs each year—but by reductions in clinical trial attrition, accelerated patient access to life‑saving therapies, and the gradual establishment of a global regulatory system in which animal studies are truly the exception rather than the rule. This transition represents regulatory science at its most adaptive and forward‑thinking: responsive to technological progress, grounded in translational evidence, and committed to maximizing public health benefit. In the years ahead, the FDA’s framework is poised to serve as a global model, demonstrating that scientific advancement, regulatory rigor, and ethical stewardship are not competing priorities but complementary pillars of a modern, effective drug development ecosystem (Fig. 1).

Fig. 1.

FDA roadmap to reducing animal testing and advancing human‑centric nonclinical safety assessment. FDA, US Food and Drug Administration; NAMs, New Approach Methodologies; ISTAND, Innovative Science and Technology Approaches for New Drugs; mABs, monoclonal antibodies; NIH, National Institutes of Health; NHPs, Non-human primates; AIM-NASH, Artificial Intelligence-Based Histologic Measurement of Nonalcoholic Steatohepatitis; EMA, European Medicines Agency; PMDA, Pharmaceuticals and Medical Devices Agency; TGA, Therapeutic Goods Administration. (Created with BioRender.com)

Author contributions

YL: review and analyze literature, wrote the manuscript; ZAL: develop concept, wrote the manuscript, provide research funding; RST: initiate and develop concept, wrote the manuscript, provide research funding.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

RST is aco-Editor-Chief of Stem Cell Research & Therapy.