Validating NAMs in Ecotoxicology: A Strategic Roadmap for Replacing Animal Testing

Abstract

This article provides a comprehensive guide for researchers and drug development professionals on the validation and application of New Approach Methodologies (NAMs) in ecotoxicology. It explores the scientific and regulatory drivers behind the global shift away from animal testing, details a range of established and emerging non-animal methods, and addresses key challenges in standardization and implementation. A central focus is placed on practical validation frameworks and comparative analyses that demonstrate the reliability and human relevance of NAMs, offering a strategic overview for integrating these approaches into modern safety assessment pipelines.

The Paradigm Shift: Scientific and Regulatory Drivers for NAMs in Ecotoxicology

Analyzing the Limitations of Traditional Animal Test Reproducibility and Human Relevance

For decades, traditional animal-based toxicological testing has served as the cornerstone of safety assessments for chemicals, pharmaceuticals, and consumer products. However, this paradigm faces growing criticism on multiple fronts—scientific, ethical, and practical. The field is now undergoing a fundamental shift toward New Approach Methodologies (NAMs) that offer more human-relevant safety assessments while reducing reliance on animal models [1]. This transition is driven by significant limitations in the reproducibility and human relevance of animal testing, which have profound implications for drug development, chemical safety assessment, and environmental protection.

The scientific community increasingly recognizes that what is toxic in a rat will not necessarily be toxic in a human, and vice versa [1]. Species variation in physiology, metabolism, and genetics presents a fundamental challenge to extrapolating animal data to human health outcomes. These limitations are not merely theoretical; they have real-world consequences manifested in high drug failure rates and limited predictive capacity for human toxicity. This analysis examines these limitations through a critical lens while exploring the promising alternative methodologies that are reshaping 21st-century toxicology.

Critical Analysis of Traditional Animal Testing Limitations

Scientific Limitations in Human Relevance and Predictive Power

The scientific case against overreliance on animal models stems from fundamental biological differences between species and concerning performance metrics in predicting human outcomes.

• Species Variation: Rodents and other common laboratory animals possess unique metabolic processes, organ sensitivities, and disease tendencies compared to humans [1]. These differences mean that animal models often fail as reliable surrogates for predicting human responses to chemicals and drugs. The infamous case of thalidomide exemplifies this translational gap—while the drug exhibited limited developmental toxicity in several animal species during initial screening, it caused serious birth defects in humans [1].

• Poor Predictive Performance: Comprehensive analyses reveal that rodents have a true positive human toxicity predictivity rate of only 40%-65% [2]. This performance gap has direct consequences for drug development, where approximately 95% of drugs developed for brain diseases fail in clinical trials despite promising animal studies [3]. Similarly, one out of every four new medicines fails due to brain side effects that did not manifest in animal testing [3].

• False Positives and Negatives: The limitations of animal models generate both false positives (substances flagged as toxic in animals but safe for humans, leading to abandonment of promising compounds) and more dangerously, false negatives (substances deemed safe in animals but toxic to humans) [1].

Challenges in Experimental Reproducibility

Beyond human relevance, traditional animal testing faces significant reproducibility challenges that undermine its reliability as a scientific tool.

• Inter-species Variability: Differences in physiology, metabolism, and genetics between species make it difficult to reproduce findings across animal models and translate them to humans [1]. For instance, laboratory mice do not develop asthma like children, and their bodies do not mount the same immune defenses [4].

• Standardization Difficulties: While animal tests follow standardized protocols, biological variability between individual animals and differences in laboratory practices can introduce significant inconsistencies. This contrasts with NAMs, which can offer more controlled and standardized experimental conditions using human-derived cells and tissues [4] [3].

• High Variable Costs: Animal research is incredibly costly, with expenses for animal purchase, housing, feeding, veterinary care, and specialized facilities creating substantial financial burdens on industry and research institutions [1]. These high costs can limit sample sizes and statistical power, further impacting reproducibility.

Quantitative Assessment of Animal Testing Limitations

Table 1: Quantitative Limitations of Traditional Animal Testing

Performance Metric	Data	Impact/Consequence
Human Toxicity Predictivity Rate	40%-65% for rodents [2]	High rate of false positives/negatives in human risk assessment
Drug Development Success Rate	95% failure rate for brain disease drugs [3]	High attrition despite promising animal data
Neurological Side Effect Prediction	25% of new medicines fail due to unpredicted brain side effects [3]	Significant patient risk and drug development costs
Translation to Human Approvals	5% success rate in translating preclinical results to human approvals [4]	Inefficient resource allocation in drug development
Time Requirements	Months to years for endpoints like carcinogenicity or chronic toxicity [1]	Delayed development of new drugs and chemicals

New Approach Methodologies (NAMs): A Human-Relevant Framework

New Approach Methodologies represent a diverse range of innovative tools and approaches that enable more human-relevant assessment of chemical hazards and risks. NAMs include any in vitro, in chemico, or computational methods that, when used alone or in combination, provide improved chemical safety assessment through more protective and relevant models while reducing reliance on animal testing [5] [2]. These approaches leverage advances in molecular biology, genomics, bioinformatics, and computational science to forecast potential adverse effects based on perturbations of key biological pathways in human-derived systems, rather than relying on observed effects in animals [1].

The conceptual foundation of NAMs represents a shift from empirical, high-dose animal testing toward a more mechanistic, human-relevant, and efficient hazard and risk assessment paradigm. At the heart of 21st-century toxicology is understanding how chemicals interact with human biological systems, utilizing human-relevant models to provide insights into toxicity mechanisms [1]. The NIH has established the Complement-ARIE program to speed the development, standardization, validation, and use of human-based NAMs, recognizing their potential to significantly advance understanding of human health and disease [5].

Categories of NAMs and Their Applications

NAMs encompass three primary categories of methodologies, each offering distinct advantages for toxicological assessment:

• In Chemico Methods: Experiments performed on biological molecules outside of cells, studying molecular interactions. These include non-mammalian models like zebrafish, nematodes, and planaria that enable rapid, cost-effective assessment of chemical effects on biology and behavior [5].

• In Vitro Methods: Experiments performed on cells outside of the body, including various cell, organoid, and tissue culture techniques [5]. These range from simple 2D cell cultures to complex 3D models such as organoids, spheroids, and microphysiological systems (organ-on-chip devices) that better replicate human physiology [1]. For example, vascularized liver cancer-on-a-chip models can evaluate vessel remodeling and cell death in response to embolic agents, providing human-relevant alternatives for testing cancer treatments [4].

• In Silico Methods: Experiments performed using computing platforms, encompassing mathematical modeling, simulation, machine learning, and other computational techniques [5]. These include quantitative structure-activity relationships (QSARs), physiologically based pharmacokinetic (PBPK) modeling, and artificial intelligence/machine learning (AI/ML) approaches for predicting toxicity [1]. Advanced computer simulations can model biological processes and predict effects of previously untested chemicals and drugs [5].

Key Technological Frameworks in NAMs Implementation

The effective implementation of NAMs relies on several conceptual frameworks that enable integration of diverse data streams:

• Adverse Outcome Pathways (AOP): The AOP framework serves as a central conceptual tool that links sequences of molecular and cellular events leading to adverse health outcomes [1] [6]. This framework provides structure and mechanistic foundation for evaluating safety, helping researchers understand the key events between initial chemical exposure and eventual adverse outcome [6].

• Integrated Approaches to Testing and Assessment (IATA): IATA represent systematic approaches that combine information from multiple NAMs, existing knowledge, and exposure data to reach well-founded safety conclusions [1]. These frameworks enable strategic integration of various data streams to build comprehensive risk assessments without relying solely on animal data.

• Next Generation Risk Assessment (NGRA): NGRA is defined as an exposure-led, hypothesis-driven approach that integrates in silico, in chemico and in vitro approaches to risk assessment [2]. In this framework, NAMs are the tools used to achieve the overall objective of more human-relevant, mechanistically informed safety assessments.

The following diagram illustrates the conceptual relationship between these frameworks and their application in modern toxicology:

Comparative Performance Data: Animal Models vs. NAMs

Quantitative Comparison of Predictive Performance

Table 2: Performance Comparison - Animal Models vs. NAMs

Performance Characteristic	Traditional Animal Models	New Approach Methodologies (NAMs)
Human Physiological Accuracy	30% for animal models [4]	80% for organ-on-a-chip systems [4]
Testing Timeline	Months to years for chronic endpoints [1]	Days to weeks for many endpoints [5]
Cost Considerations	High (animal purchase, housing, veterinary care) [1]	Lower (especially for high-throughput screening) [4]
Species Relevance	Limited by interspecies differences [1]	High (uses human cells/tissues) [3]
Mechanistic Insight	Limited to observable phenotypes	High (molecular initiating events, pathway analysis) [1]
Personalized Assessment	Not feasible	Possible (patient-specific cells/organoids) [3]

Specific NAMs Applications and Validation Case Studies

The practical application of NAMs across various toxicological endpoints demonstrates their growing utility and reliability:

• Skin Sensitization Assessment: Defined Approaches (DAs) combining in silico, in chemico, and in vitro data sources with fixed data interpretation procedures have been formally adopted in OECD Test Guidelines (TG 497) for skin sensitization assessment [2]. These approaches demonstrate that combination of human-based in vitro approaches can outperform the traditional Local Lymph Node Assay (LLNA) performed in mice, particularly in terms of specificity [2].

• Crop Protection Product Evaluation: A multiple NAM testing strategy was performed for crop protection products Captan and Folpet using 18 different in vitro studies, including eye and skin irritation and skin sensitization assays compliant with OECD guidelines [2]. The NAM package appropriately identified both chemicals as contact irritants, demonstrating that suitable risk assessments could be performed with available NAM tests, aligning with risk assessments conducted using traditional mammalian test data [2].

• Transcriptomic Points of Departure (tPODs): Developments in in vitro transcriptomics enable establishment of transcriptomic points of departure by mapping gene transcripts to known Adverse Outcome Pathways [7]. This approach has shown that results from the tPOD method closely replicate existing points of departure for numerous chemicals that had been set using combinations of animal and in vitro data [7].

Experimental Protocols and Methodologies in NAMs

Protocol for Transcriptomic Point of Departure (tPOD) Assessment

The transcriptomic point of departure approach represents a powerful methodology for deriving protective exposure limits without animal testing:

• Cell Culture Preparation: Human primary cells or induced pluripotent stem cell (iPSC)-derived cells are cultured under defined conditions appropriate for the target tissue type. For liver toxicity assessment, this might involve hepatocytes or liver spheroids; for neurotoxicity, brain organoids or neuronal cultures would be utilized [7].

• Chemical Exposure: Cells are exposed to a range of chemical concentrations, typically spanning several orders of magnitude, with appropriate vehicle controls. Exposure duration varies based on the endpoint but often ranges from 24 hours to 14 days depending on the biological pathway being evaluated [7].

• RNA Extraction and Sequencing: Following exposure, total RNA is extracted using standardized kits (e.g., Qiagen RNeasy). RNA quality is verified (RIN >8.0) before library preparation and sequencing using platforms such as Illumina. Minimum sample size is n=3 per concentration group to ensure statistical power [7].

• Bioinformatic Analysis: Sequencing reads are aligned to the human reference genome, and differential gene expression analysis is performed using established pipelines (e.g., DESeq2, EdgeR). Significantly altered genes (p<0.05 with false discovery rate correction) are identified for each concentration [7].

• Pathway Mapping and Benchmark Dose Modeling: Differentially expressed genes are mapped to known Adverse Outcome Pathways using databases such as the AOP-Wiki. The benchmark dose (BMD) method is then applied to identify the concentration at which significant pathway perturbation occurs, establishing the tPOD [7].

• Validation and Uncertainty Assessment: The tPOD is compared to existing in vivo data when available, and uncertainty factors are applied based on the robustness of the pathway evidence and cross-species extrapolation considerations [7].

Protocol for Organ-on-Chip Toxicity Assessment

Microphysiological systems offer sophisticated platforms for evaluating organ-specific toxicity:

• Chip Preparation and Seeding: Organ-on-chip devices (e.g., Emulate, Mimetas) are prepared according to manufacturer specifications. Appropriate human cell types are seeded into the device chambers—for liver chips, this typically involves primary hepatocytes or HepaRG cells combined with endothelial and stellate cells in relevant ratios [4].

• Maturation and Conditioning: Chips are maintained in specialized instrumentation that provides physiological cues (flow, stretch, electrical stimulation) for 7-28 days to allow tissue maturation. Media is changed regularly, and tissue formation is monitored via transepithelial electrical resistance (TEER), microscopy, or biomarker secretion [4] [3].

• Compound Dosing and Exposure: Test compounds are introduced into the chip system at physiologically relevant concentrations, typically through the perfusion system to mimic blood flow. Multiple concentrations are tested in separate chips, with n≥3 chips per concentration [4].

• Real-time Monitoring and Endpoint Assessment: During exposure, chips are monitored using integrated sensors where available (e.g., oxygen, pH, TEER). At designated timepoints, chips are harvested for specific endpoint assessments which may include:

  • Immunofluorescence staining for marker expression and tissue integrity
  • ELISA/Luminex for cytokine and biomarker secretion
  • Metabolite analysis via LC-MS
  • RNA/protein extraction for transcriptomic or proteomic analysis [4]

• Data Integration and Cross-Talk Analysis: For multi-organ systems, communication between organ compartments is assessed through transfer of metabolites or signaling molecules. Physiologically based pharmacokinetic (PBPK) modeling may be applied to translate in vitro concentrations to human relevant doses [1].

The following workflow diagram illustrates the application of human relevance assessment for toxicological pathways and associated NAMs:

Essential Research Reagent Solutions for NAMs Implementation

Successful implementation of NAMs requires specific research reagents and technological platforms that enable human-relevant toxicological assessment:

Table 3: Essential Research Reagent Solutions for NAMs

Reagent/Platform Category	Specific Examples	Function in NAMs Implementation
Stem Cell Sources	Human induced pluripotent stem cells (iPSCs), Primary tissue-derived stem cells	Foundation for developing human-relevant organoids and tissue models for disease modeling and toxicity testing [3]
Organoid Culture Systems	Intestinal organoid kits, Brain organoid differentiation kits, Liver organoid platforms	Create 3D tissue-like structures that replicate complexity and function of human organs for physiological toxicity assessment [5] [4]
Organ-on-Chip Devices	Emulate organs-on-chips, Mimetas Phaseguides, TissUse HUMIMIC	Microphysiological systems that mimic key physiological aspects of tissues/organs with microenvironments that align with human biology [5] [4]
Extracellular Matrix Products	Matrigel, Collagen I, Fibrin-based hydrogels, Synthetic PEG hydrogels	Provide scaffolding and biochemical cues to support 3D tissue formation and maintenance in organoid and tissue culture systems [4]
High-Content Screening Tools	Transcriptomic assays (RNA-seq kits), Proteomic arrays, Metabolomic platforms	Enable comprehensive molecular profiling for mechanism-based toxicity assessment and adverse outcome pathway identification [1] [7]
Computational Toxicology Resources	OECD QSAR Toolbox, EPA ToxCast database, ICE (Integrated Chemical Environment)	In silico platforms that integrate diverse data streams for predictive toxicology and chemical prioritization [5] [7]

The limitations of traditional animal testing in reproducibility and human relevance are no longer speculative concerns but empirically demonstrated realities with significant implications for public health and chemical safety assessment. The scientific evidence compellingly shows that animal models frequently fail as reliable predictors of human toxicity, with poor translational success rates across multiple domains, particularly in drug development and complex disease modeling.

New Approach Methodologies represent not merely alternative testing strategies but a fundamental paradigm shift toward human-relevant, mechanism-based safety assessment. The integration of advanced in vitro models such as organoids and organ-on-chip systems with powerful in silico approaches and comprehensive molecular profiling enables a more predictive, efficient, and ethically advanced approach to toxicology. The growing validation of these approaches across regulatory agencies worldwide—including recent landmark policies from the NIH and FDA that no longer consider animal testing as the sole standard—signals an irreversible transition in toxicological science [3] [8].

While challenges remain in standardization, validation, and implementation of NAMs, the scientific foundation, economic incentives, and ethical imperatives for this transition are firmly established. The continuing evolution of these technologies—including advances in bioprinting, single-cell technologies, and explainable AI—promises to further accelerate the move toward a truly human-centered toxicological framework that better protects both human health and the environment [1].

The regulatory landscape for drug development is undergoing a profound transformation, marked by a global transition away from traditional animal testing toward more human-relevant New Approach Methodologies (NAMs). This paradigm shift is driven by scientific evidence demonstrating that human-relevant models often provide more accurate safety and efficacy data than animal models, alongside growing ethical concerns and technological advancements. In the United States, the FDA Modernization Act 2.0 (December 2022) served as the pivotal catalyst, removing the longstanding statutory requirement for animal testing and explicitly allowing non-animal alternatives in drug development [9] [10]. This legislative change has been followed by substantial policy implementation, including the FDA's April 2025 detailed roadmap to phase out animal testing requirements, beginning with monoclonal antibodies and other biologics [11] [12].

Parallel to U.S. developments, the United Kingdom has launched its own ambitious strategy. In November 2025, the UK government published a comprehensive roadmap to phase out animal testing in science, backed by £75 million in funding [13]. This strategy outlines six key objectives to accelerate the replacement of animals, drive private investment in alternative methods, and position the UK as a global leader in this field. Together, these regulatory milestones in two major biomedical research hubs represent a coordinated global effort to modernize drug development through advanced, human-relevant science.

Analysis of Key U.S. Regulatory Milestones

The Foundation: FDA Modernization Act 2.0

The FDA Modernization Act 2.0, signed into law on December 29, 2022, represents the cornerstone of U.S. regulatory change [10]. This legislation fundamentally updated the Federal Food, Drug, and Cosmetic Act of 1938, which had effectively mandated animal data before human trials could begin. The Act achieved this by replacing the term "preclinical tests (including tests on animals)" with "nonclinical tests," and explicitly defining these to include cell-based assays, microphysiological systems, and computer models [9] [10]. This change empowered sponsors to use NAMs and instructed FDA reviewers to consider them on their scientific merits, though it did not ban animal testing outright.

Table: Key Provisions of FDA Modernization Act 2.0

Provision	Description	Impact
Terminology Update	Replaced "preclinical tests (including tests on animals)" with "nonclinical tests"	Removed statutory animal-test mandate from original 1938 law
Definition of Nonclinical Tests	Defined to include cell-based assays, microphysiological systems, bioprinted tissues, and computer models	Provided legal basis for using human-relevant testing methods
Regulatory Flexibility	Authorized FDA to consider these alternatives for investigational new drug applications	Enabled more flexible approach to drug safety and efficacy assessment

Implementation and Expansion: Recent U.S. Regulatory Actions

Following the foundational legislation, U.S. regulatory agencies have implemented a series of decisive actions to accelerate the adoption of NAMs throughout 2024-2025. These actions represent a coordinated effort across multiple government bodies to operationalize the principles established in the FDA Modernization Act 2.0.

Table: Recent U.S. Regulatory Actions Supporting NAMs (2024-2025)

Date	Agency	Action	Significance
Sep 2024	FDA	First Organ-on-a-Chip accepted into ISTAND Program	Created precedent for regulatory qualification of microphysiological systems [10]
Apr 2025	FDA	Announced phased elimination of routine animal testing & released Roadmap	Stated animal use should become "the exception rather than the rule" [11] [10]
Apr 2025	NIH	Shifted funding priorities toward human-based technologies	Incentivized research using Organ-Chips, organoids, or computational models [10]
Jul 2025	NIH	Barred funding for animal-only studies	Required integration of at least one validated human-relevant method [10]

The FDA's April 2025 announcement specifically outlined a plan where "animal testing requirement will be reduced, refined, or potentially replaced using a range of approaches, including AI-based computational models of toxicity and cell lines and organoid toxicity testing in a laboratory setting" [11]. This initiative immediately encouraged the inclusion of NAMs data in investigational new drug (IND) applications, beginning with monoclonal antibodies and expanding to other biologics and eventually new chemical entities [12]. The agency also committed to using pre-existing, real-world safety data from other countries with comparable regulatory standards where drugs have already been studied in humans [11].

Analysis of UK Regulatory Strategy

The UK's Comprehensive Phase-Out Roadmap

In November 2025, the UK government unveiled a detailed strategy to phase out animal testing in science, supported by £75 million in funding [13]. This strategy delivers on a manifesto commitment to improve animal welfare while advancing scientific research. The roadmap acknowledges that phasing out animal use can only occur where reliable and effective alternative methods can replace them, and includes specific commitments with timelines extending to 2030. The strategy will be overseen by a committee chaired by science minister Lord Vallance, with key performance indicators to be published the following year [13].

The UK strategy establishes six key objectives designed to create a comprehensive ecosystem for transitioning to NAMs:

• Accelerate the replacement of animals in science to phase out their use
• Achieve equal or better research and testing outcomes using alternative methods
• Drive private investment in alternative methods to boost innovation and growth
• Improve regulatory confidence and acceptance of alternative methods
• Create infrastructure and partnerships to unlock value from UK data
• Position the UK as a global leader in alternative methods [13]

This multifaceted approach addresses not only the scientific and technological challenges but also the economic, regulatory, and infrastructure requirements necessary for a successful transition to animal-free research methodologies.

Parallel Developments in UK Medical Product Regulation

Concurrent with the animal testing phase-out strategy, the UK's Medicines and Healthcare products Regulatory Agency (MHRA) has undertaken significant regulatory modernization across multiple domains. In June 2025, the MHRA implemented a major overhaul of medical device regulation with new Post-Market Surveillance (PMS) requirements that strengthen oversight of devices once they're in use [14]. These regulations require manufacturers to proactively monitor product safety and performance, collect comprehensive real-world data, and report serious incidents within shortened timelines.

Additionally, in November 2025, the MHRA announced a sweeping reform of rare disease therapy regulation [13] [15]. The proposed framework includes innovative licensing models that could allow "an early, single approval [to] be issued for both a clinical trial application and marketing authorisation based on compelling but limited evidence" [15]. This patient-centered approach emphasizes the use of real-world evidence, adaptive trial designs, and flexible benefit-risk thresholds to address the unique challenges of rare disease drug development where patient populations are small.

Comparative Analysis of US and UK Approaches

Regulatory Pathway Comparison

The US and UK regulatory milestones share common objectives but demonstrate distinct implementation approaches and strategic priorities, reflecting their different regulatory traditions and research ecosystems.

Table: Comparative Analysis of US and UK Regulatory Pathways for NAMs

Aspect	United States Approach	United Kingdom Approach
Primary Driver	Legislative action (FDA Modernization Act 2.0) followed by agency implementation	Government strategy with significant funding (£75 million) and committee oversight
Key Mechanism	ISTAND Program for qualifying novel Drug Development Tools [10]	Comprehensive roadmap with six strategic objectives and specific KPIs [13]
Timeline	Immediate for INDs; animal studies to become "the exception" within 3-5 years [11] [12]	Phased approach with commitments extending to 2030 [13]
Funding Strategy	NIH prioritization of grants incorporating human-based technologies [10]	Direct government funding combined with focus on driving private investment [13]
Initial Focus	Monoclonal antibodies and other biologics [11]	Broader scientific applications, with specific therapeutic areas emerging

Both regulatory systems recognize the critical importance of regulatory confidence in alternative methods. The FDA has emphasized building confidence through case studies and collaborative validation [16], while the UK strategy explicitly includes "improve regulatory confidence and acceptance of alternative methods" as one of its six key objectives [13]. This focus on establishing scientific confidence in NAMs is essential for their widespread adoption in regulatory decision-making.

Scientific Validation Pathways

The validation of New Approach Methodologies follows logically structured pathways to establish scientific confidence and regulatory acceptance. The diagram below illustrates this process from development to full implementation.

This validation pathway is exemplified by specific milestones, such as the September 2024 acceptance of the first Organ-on-a-Chip into the FDA's ISTAND Pilot Program [10]. The Liver-Chip submission for predicting drug-induced liver injury (DILI) established an important evidentiary precedent, showing 87% sensitivity and 100% specificity for a set of hepatotoxic drugs that animal models had deemed safe [10]. This head-to-head validation against traditional methods provides the compelling evidence needed for regulatory acceptance.

Essential Research Reagent Solutions for NAMs Implementation

The successful implementation of NAMs in regulatory science requires specific research reagents and technological platforms. These tools form the foundation for generating robust, human-relevant safety and efficacy data.

Table: Key Research Reagent Solutions for NAMs Implementation

Reagent/Platform	Function	Application in Regulatory Science
Organ-on-a-Chip Systems	Microfluidic devices containing human cells that emulate organ-level physiology	Replaces animal models for drug safety testing (e.g., DILI prediction) [10] [17]
Human Induced Pluripotent Stem Cells (iPSCs)	Patient-derived cells reprogrammed to pluripotency, capable of differentiation into various cell types	Creates disease models that reflect human genetic diversity for efficacy testing [17]
High-Throughput Screening Assays	Automated in vitro tests that rapidly evaluate thousands of compounds	Generates toxicity and efficacy data at scale for chemical prioritization [18] [16]
Computational Toxicology Models	AI and machine learning algorithms that predict chemical hazards	Predicts drug toxicity, metabolism, and off-target effects using structural biology and multiomics data [17]
3D Organoids	Self-organizing 3D tissue structures derived from stem cells that mimic organ architecture	Provides more physiologically relevant models for disease modeling and drug testing [17]

These research tools enable the generation of human-relevant data that addresses the pharmacogenomic differences between animals and humans that have traditionally contributed to high clinical trial failure rates [17]. For example, enzymes such as cytochrome P450 that are involved in drug metabolism vary significantly between species, leading to differences in how drugs are broken down and cleared from the body [17]. Human-based systems provide more accurate prediction of these processes.

Experimental Protocols and Methodologies

Organ-on-a-Chip Validation Protocol

The validation of Organ-on-a-Chip technology for regulatory applications follows rigorous experimental protocols designed to demonstrate reliability and predictive capacity. The workflow below illustrates the key stages in this validation process.

The validation of the Liver-Chip for drug-induced liver injury prediction exemplifies this protocol. This study demonstrated the chip's superior performance compared to animal models, showing 87% sensitivity and 100% specificity for detecting hepatotoxic drugs that had previously been deemed safe based on animal studies [10]. The experimental methodology involved:

• Chip Fabrication: Creation of microfluidic devices containing primary human hepatocytes, hepatic stellate cells, and endothelial cells in a physiologically relevant architecture [10]
• Compound Testing: Exposure of the Liver-Chip to 27 known drugs (16 with known DILI risk in humans, 11 without) at clinically relevant concentrations [10]
• Endpoint Analysis: Measurement of multiple parameters including albumin secretion, urea production, ATP content, and reactive oxygen species to create a comprehensive safety profile [10]
• Benchmarking: Comparison of results against historical animal data and human clinical outcomes to establish predictive capacity [10]

This rigorous validation approach provides the evidence base needed for regulatory acceptance, as demonstrated by the technology's inclusion in the FDA's ISTAND program and specific reference in the FDA's NAMs roadmap [10].

Computational Toxicology Workflow

Computational approaches represent another essential methodology in the NAMs toolkit. The integration of artificial intelligence and machine learning for toxicity prediction follows a structured workflow:

• Data Curation: Compilation of high-quality in vitro and historical in vivo toxicity data from diverse sources
• Feature Selection: Identification of relevant molecular descriptors and structural features correlated with toxicity endpoints
• Model Training: Development of machine learning algorithms using training datasets with known toxicity outcomes
• Validation: Testing model performance against independent compound sets not used in training
• Interpretation: Application of explainable AI techniques to understand the basis of predictions

The MHRA has endorsed this approach through funding of AI projects, including one that will "use artificial intelligence and NHS data to predict side effects from drug combinations before they reach patients" [13]. This project aims to spot interactions by analyzing patterns in anonymized NHS data showing how different medicines behave when used together, focusing initially on cardiovascular medicines.

The regulatory milestones achieved from the FDA Modernization Act 2.0 in 2022 to the UK's replacement strategy in 2025 represent a decisive global shift toward human-relevant safety and efficacy testing. While implementation approaches differ between regions, the fundamental objective remains consistent: to create more predictive, efficient, and humane drug development processes through the adoption of New Approach Methodologies.

This regulatory evolution is underpinned by growing scientific evidence that NAMs can provide equal or better protection of human health compared to traditional animal models. The continuing alignment of scientific capability, regulatory policy, and funding priorities across major research nations suggests that this transition will accelerate in the coming years, fundamentally reshaping global drug development paradigms. As these regulatory frameworks mature, they promise to deliver safer, more effective therapeutics to patients while reducing reliance on animal testing.

The Growing Body of Evidence Supporting NAMs for Improved Human Toxicity Prediction

New Approach Methodologies (NAMs) are transforming toxicology by providing human-relevant, mechanistic data that often outperforms traditional animal testing in predicting human outcomes. This shift is driven by advances in in vitro (cell-based), in silico (computer-based), and in chemico (biochemical) methods, which offer more accurate, efficient, and ethical tools for safety assessment. The following comparison guide objectively evaluates the performance of these NAMs against traditional animal models, providing supporting experimental data and detailed methodologies for researchers and drug development professionals.

The table below summarizes key performance metrics of validated NAMs compared to traditional animal tests, demonstrating their superior predictive accuracy for human toxicity in several critical areas.

Table 1: Performance Comparison of NAMs vs. Animal Tests for Human Toxicity Prediction

Toxicity Endpoint	Traditional Animal Test	Performance (Accuracy)	New Approach Methodologies (NAMs)	Performance (Accuracy)	Supporting Evidence
Skin Sensitization	Guinea Pig Tests (e.g., GPMT)	72%-74% [19]	Defined Approaches (e.g., DPRA, KeratinoSens, h-CLAT)	Up to 85% [19]	OECD TG 442 series [20]
Skin Irritation	Draize Rabbit Test	~60% [19]	Reconstituted Human Skin Models (e.g., EpiDerm, EpiSkin)	Up to 86% [19]	OECD TG 439 [19]
Developmental Toxicity	Animal Tests (e.g., rodents, rabbits)	~60% Sensitivity [19]	Human Stem Cell-Based Test (devTOX quickPredict)	93% Sensitivity [19]	FDA Biomarker Qualification Program [21]
Shellfish Toxin Detection	Mouse Bioassay	Less effective for human protection [19]	Analytical Chemistry Method (e.g., LC-MS)	Superior for human protection [19]	Fully replaced in regulatory testing [19]

Detailed Experimental Protocols and Methodologies

Defined Approaches for Skin Sensitization Assessment

Skin sensitization is a key event in the development of allergic contact dermatitis. The Adverse Outcome Pathway (AOP) for skin sensitization has been well-defined, allowing for the development of NAMs that target specific molecular and cellular events.

Experimental Workflow for Integrated Skin Sensitization Testing

Detailed Protocols:

• Direct Peptide Reactivity Assay (DPRA) [20]: This in chemico method evaluates the molecular initiating event—the covalent binding of a chemical to skin proteins.

  • Procedure: Incubate the test chemical with a synthetic peptide containing either cysteine or lysine for 24 hours at 25°C.
  • Measurement: Use high-performance liquid chromatography (HPLC) to quantify the percentage of peptide depletion.
  • Prediction Model: Chemicals with peptide depletion above a specific threshold (e.g., >6.38% for cysteine) are classified as sensitizers.

• KeratinoSens [20]: This in vitro assay measures the second key event—keratinocyte activation.

  • Cell Line: Use a genetically engineered human keratinocyte cell line (HaCaT) containing a luciferase gene under the control of the Antioxidant Response Element (ARE).
  • Procedure: Expose cells to the test chemical for 48 hours.
  • Measurement: Quantify luciferase activity; an induction above a threshold (e.g., 1.5-fold) indicates sensitization potential.

• human Cell Line Activation Test (h-CLAT) [20]: This in vitro assay measures the third key event—dendritic cell activation.

  • Cell Line: Use a human monocytic leukemia cell line (THP-1 or U937).
  • Procedure: Expose cells to the test chemical for 24 hours.
  • Measurement: Use flow cytometry to measure the increased expression of surface markers CD86 and CD54.
  • Prediction Model: Chemicals causing an increase in expression above predefined thresholds are classified as sensitizers.

Stem Cell-Based Assay for Developmental Toxicity

The devTOX quickPredict platform is a prominent example of a NAM validated for developmental toxicity testing.

Protocol for devTOX quickPredict Assay [21]:

• Cell Culture: Utilize human induced pluripotent stem cells (iPSCs).
• Differentiation: Guide the iPSCs to differentiate into metabolically active progenitor cells over 72 hours.
• Chemical Exposure: Expose the differentiating cells to a range of concentrations of the test compound.
• Endpoint Measurement: Assess changes in cell viability and metabolic activity using high-content imaging. The key measurement is the compound's impact on metabolic pathways, which serves as a biomarker for developmental toxicity.
• Data Analysis: A proprietary algorithm converts the metabolic data into a prediction of the compound's potential to cause human developmental toxicity, with a reported 93% sensitivity [19].

In Silico Prediction of Developmental Toxicity

Computational tools like the DeTox database leverage existing data to predict toxicity of new chemical structures.

Protocol for Using the DeTox Database [21]:

• Input: Chemical structure of the compound of interest (e.g., in SMILES or SDF format).
• Processing: The tool employs Quantitative Structure-Activity Relationship (QSAR) models that have been built and validated using data from the US FDA, the Teratogen Information System (TERIS), and other datasets.
• Output: The model provides a probability score for whether the chemical is likely to cause developmental toxicity. It can also predict risk specifically for the first, second, or third trimester of pregnancy.
• Limitation: The model may encounter "activity cliffs" where two structurally similar chemicals have different toxicities, and it currently does not provide mechanistic insight [21].

The Scientist's Toolkit: Essential Research Reagents and Solutions

Table 2: Key Reagents and Platforms for Implementing NAMs

Item	Function/Description	Example Use Cases
Human iPSCs	Pluripotent cells capable of differentiating into any cell type; provide a human-relevant biological system.	devTOX quickPredict assay; generating organoids for toxicity screening [21].
Reconstituted Human Tissues	3D models of human skin (EpiDerm, EpiSkin) or other organs; mimic in vivo architecture and response.	Replacement of Draize rabbit test for skin irritation/corrosion [19].
THP-1 or U937 Cell Lines	Human monocyte-derived cell lines used to model dendritic cell activation.	h-CLAT for assessing the skin sensitization potential of chemicals [20].
ARE-Luciferase Reporter Cells	Genetically engineered cells (e.g., KeratinoSens) that activate luciferase upon oxidative stress.	Identifying chemicals that trigger the Keap1-Nrf2-ARE pathway, a key event in skin sensitization [20].
Synthetic Peptides (Cysteine/Lysine)	Used in the DPRA to measure a chemical's direct protein reactivity.	Quantifying the molecular initiating event in the skin sensitization AOP [20].
Microfluidic Organ-on-a-Chip Devices	Devices containing human cells that simulate organ-level physiology and dynamic flow.	Creating liver, kidney, or multi-organ (body-on-a-chip) models for systemic toxicity assessment [5] [22].
QSAR Software & Databases	Computational tools that predict toxicity based on chemical structure and existing data.	DeTox database for developmental toxicity screening; ChemMaps for exploring environmental chemicals [5] [21].

Integrated Testing Strategies and Regulatory Adoption

The true power of NAMs is realized when they are combined in integrated testing strategies, such as Integrated Approaches to Testing and Assessment (IATA). The tiered Next Generation Risk Assessment (NGRA) framework exemplifies this: in a case study evaluating 37 compounds with known DART outcomes, an NGRA using in silico and in vitro NAMs in its first tier correctly identified 16 out of 17 high-risk compounds, demonstrating its utility as a protective screening tool [21].

Regulatory bodies are actively supporting this transition. The FDA's recent roadmap to phase out animal testing for monoclonal antibodies and the FDA Modernization Act 2.0 encourage the submission of NAMs data [23]. Internationally, the Organisation for Economic Co-operation and Development (OECD) has adopted numerous test guidelines for validated NAMs, facilitating their global use in regulatory decision-making [5] [20].

New Approach Methodologies (NAMs) are defined as any technology, methodology, approach, or combination thereof that can provide information on chemical hazard and risk assessment while avoiding the use of animals. This broad definition encompasses in silico (computational), in chemico, in vitro (cell-based), and ex vivo approaches [24]. The fundamental goal of NAMs is to provide more human-relevant, efficient, and reliable approaches to risk assessment and regulatory decision-making compared to traditional animal test methods [24]. The transition toward NAMs is driven by multiple factors: ethical concerns regarding animal testing, scientific limitations in translating animal data to human outcomes, and the potential for significant cost and time reductions in safety assessment [25] [26].

The term "NAM" does not necessarily imply that the methodologies themselves are new; rather, it is their application to regulatory decision-making or their use as replacements for traditional testing requirements that represents the innovation [24]. This distinction is crucial for understanding the regulatory landscape, where established non-animal methods may be considered "new" in the context of specific regulatory guidelines. The push for NAMs has gained substantial momentum through international regulatory initiatives. For instance, the U.S. Food and Drug Administration (FDA) has enforced a plan to phase out animal testing for monoclonal antibodies (mAbs) in the preclinical stage, while the European Commission has updated its roadmap (Directive 2010/63/EU) for animal testing phase-out [27]. Health Canada's adoption of NAMs under the Canadian Environmental Protection Act (CEPA) further demonstrates the global shift toward these methodologies [26].

Scope and Definition of NAMs

Comprehensive Scope of NAM Technologies

The technological scope of NAMs is extensive and continuously evolving, encompassing a wide array of platforms and approaches designed to address different aspects of toxicological assessment. The table below categorizes the primary types of NAMs and their specific applications in safety assessment.

Table 1: Categories and Applications of New Approach Methodologies (NAMs)

Category	Description	Example Technologies	Primary Applications
In Vitro Models	Uses cells, tissues, or organs maintained in a controlled laboratory environment.	2D/3D cell cultures, reconstructed human epidermis (RHE), organoids, spheroids, microphysiological systems (organ-on-a-chip) [28] [27] [29].	Cytotoxicity, skin irritation/corrosion, genotoxicity, organ-specific toxicity screening.
Ex Vivo Models	Uses tissues or organs obtained from living organisms and maintained in a laboratory setting.	Human skin models from surgical donations (e.g., Genoskin) [29].	Dermal absorption and toxicity, studies requiring intact human tissue architecture.
In Silico Models	Computational approaches to predict toxicity.	Quantitative Structure-Activity Relationship (QSAR) models, read-across, molecular docking, pharmacokinetic/pharmacodynamic (PK/PD) modeling [26] [30].	Hazard prediction, prioritization of chemicals for testing, bioaccumulation potential.
Omics Technologies	High-throughput analysis of cellular molecules.	Toxicogenomics, transcriptomics, proteomics, metabolomics [28] [27].	Mechanistic toxicology, biomarker discovery, understanding signal transduction pathways.
High-Throughput Screening (HTS)	Automated assays that rapidly test thousands of compounds.	Cell-based assays, high-content imaging, biochemical assays [28] [26].	Rapid toxicity screening, prioritization of large chemical libraries.

Core Principles: The 3Rs and Human Relevance

The development and implementation of NAMs are guided by two foundational sets of principles: the ethical framework of the 3Rs and the scientific pursuit of human biological relevance.

The 3Rs Principle—Replace, Reduce, and Refine animal use—forms the ethical cornerstone of the NAM paradigm [25] [26]:

• Replacement: Using technologies that avoid or replace the use of animals, such as human volunteers, human tissues and cells, computer models, or established cell lines. Replacement can be absolute (full) or relative (partial), such as using invertebrates like fruit flies or worms [25].
• Reduction: Employing methods that minimize the number of animals needed per study while maintaining scientific and statistical rigor. This includes techniques like longitudinal imaging, blood microsampling, and data/resource sharing between research groups [25].
• Refinement: Improving animal welfare by using procedures that minimize potential distress, pain, and harm when animal studies are still necessary. This encompasses comfortable housing, improved husbandry, appropriate analgesia, and training animals to cooperate during procedures [25].

Scientifically, NAMs aim to enhance human biological relevance in toxicological assessment. Traditional animal models, while the long-standing gold standard, often have questionable or limited biological relevance to human effects [24]. NAMs address this by directly utilizing human cells and tissues, thereby offering mechanistic insights that may be more useful for regulatory decision-making than animal data [24] [26]. For example, organ-on-chip technologies and human-relevant cell lines are designed to better mimic human physiological responses, thereby improving predictive accuracy for human outcomes [28].

The Critical Need for a Unified Validation Framework

Limitations of Current Validation Processes

Despite significant technological advancements, the validation, acceptance, and implementation of NAMs within regulatory decision-making have not kept pace with their development [24]. The traditional process for validating new test methods, as outlined in documents like the OECD Guidance Document No. 34 (2005), has become increasingly viewed as rigid, cumbersome, and ill-suited for the rapid evolution of NAMs [24]. A primary limitation has been the heavy reliance on inter-laboratory ring trials, which are lengthy, expensive, and require extensive coordination across multiple organizations [24].

Furthermore, the regulatory acceptance of NAMs has often been contingent on demonstrating predictive capacity by comparing NAM results with data from traditional animal tests. This creates a fundamental paradox, as animal test results themselves can be variable and of limited human relevance [24]. Many existing regulatory statutes were written with specific animal test data in mind, making it difficult to accept NAMs that provide different, yet potentially more human-relevant, information [24]. As noted by Steve Bulera of Charles River Laboratories, a significant hurdle is that "regulators are also going to have to figure out how to use this information to make a decision on a drug’s development," and a lack of standardization where "company A may validate its assay completely differently from company B" creates major implementation barriers [27].

Proposed Frameworks for Scientific Confidence and Validation

In response to these challenges, several frameworks have been proposed to modernize the process of establishing scientific confidence in NAMs. A key proposal, outlined by Clippinger et al., suggests a framework built on five essential elements [24]:

• Fitness for Purpose: The NAM must be suitable for its intended regulatory or safety application.
• Human Biological Relevance: The method should be based on human biology and provide mechanistically informative data for health-protective decisions.
• Technical Characterization: The assay must be reliable and reproducible.
• Data Integrity and Transparency: The data generated and its methodological basis must be transparent and robust.
• Independent Review: The method and data should undergo objective, independent scientific assessment.

This framework emphasizes that a NAM's results need not directly align with traditional animal test data to be valuable; instead, it should provide biologically relevant information and mechanistic insights that are more useful for regulatory decision-making [24].

Another multi-stakeholder effort produced a 3-step evaluation framework intended to provide a consistent set of universal criteria for evaluating a NAM's fit-for-purpose [31]. The goal of this framework is to support greater consistency across initiatives, accelerate the development of new NAMs, and systematically determine their suitability for regulatory application [31]. These frameworks collectively represent a call to action for a unified, cross-industry approach grounded in measurable quality standards and standardization to accelerate the integration of NAMs into regulatory decision-making [32].

The following diagram illustrates the logical relationships and workflow within a modern, hypothesis-driven framework for validating and utilizing NAMs in risk assessment.

Implementation in Research and Regulatory Contexts

Current Applications and Experimental Workflows

NAMs are being implemented across various industries and for a range of toxicological endpoints. The following table summarizes key application areas and the NAMs commonly employed.

Table 2: Implementation of NAMs Across Industries and Endpoints

Industry/Area	Key Applications	Commonly Used NAMs	Regulatory Status & Examples
Pharmaceuticals	Early toxicity screening, safety assessment of new drugs, hepatotoxicity, cardiotoxicity.	High-throughput screening, organ-on-chip, toxicogenomics, cellular assays [28] [27].	~70% of Investigational New Drug (IND) applications rely on non-animal methods in early screening [28]. FDA phasing out animal testing for specific mAbs [27].
Cosmetics & Household Products	Skin irritation, skin corrosion, eye irritation, dermal absorption.	Reconstructed Human Epidermis (RHE), ex vivo human skin models [29] [33].	EU Cosmetics Regulation (EC 1223/2009) bans animal testing. OECD guidelines exist for RHE models for corrosion/irritation [29] [33].
Industrial Chemicals	Hazard identification, prioritization of data-poor substances, ecotoxicology.	In silico (QSAR, read-across), high-throughput in vitro assays, omics technologies [26] [30].	Used in EPA's ToxCast, Health Canada's Existing Substances Program for grouping, read-across, and risk-based prioritization [26].
Ecotoxicology	Assessment of toxicity to aquatic and terrestrial organisms.	In vitro assays, QSAR, tests with alternative species (e.g., zebrafish embryos) [26] [30].	OECD guidelines for fish embryo tests. Frameworks use GHS for classification and tools like P2OASys for comparative hazard assessment [30].

A typical workflow for integrating NAMs into a chemical safety assessment, particularly for data-poor substances, might involve the following steps [26] [30]:

• Compile Existing Data: Gather all available physicochemical, in silico, and in vitro data.
• Identify Data Gaps: Determine key missing information required for hazard characterization.
• Apply Hypothesis-Driven Testing: Use a targeted battery of in vitro assays (e.g., for cytotoxicity or specific mechanistic targets) to generate bioactivity data.
• Conduct In Silico Modeling: Use computational tools to predict properties like bioaccumulation or persistence.
• Integrate Evidence: Use a weight-of-evidence approach to combine all data streams (existing, in silico, in vitro) to characterize hazard and inform risk assessment.

The Scientist's Toolkit: Key Reagents and Platforms

The successful execution of NAMs relies on a suite of specialized reagents, tools, and platforms. The following table details essential components of the modern toxicologist's toolkit.

Table 3: Essential Research Reagent Solutions for NAMs

Tool/Reagent	Function	Key Characteristics & Examples
Human Cell Lines	Provide a human-relevant biological system for toxicity testing.	Includes primary cells, immortalized lines, and induced pluripotent stem cells (iPSCs). Critical for organ-specific models [28].
3D Culture Matrices	Support the formation of complex 3D tissue structures like spheroids and organoids.	Hydrogels and scaffolds that mimic the extracellular matrix, improving physiological relevance over 2D cultures [27] [33].
Ex Vivo Human Skin Models	Used for dermal toxicity testing as a highly physiologically relevant alternative.	Comprises real, surgically donated human skin kept alive ex vivo (e.g., Genoskin). Contains mature stratum corneum, hair follicles, and immune cells missing from reconstructed models [29].
Organ-on-a-Chip Platforms	Microfluidic devices that emulate the structure and function of human organs.	Contain living human cells and tissues that mimic key physiological functions. Used to study complex organ-level responses and inter-organ interactions [28] [27].
Biosensors & HTS Reagents	Enable high-throughput, automated screening of chemical effects.	Fluorescent or luminescent probes, antibodies, and other reagents designed for high-content imaging and high-throughput screening platforms [28] [33].
OMICS Reagents	Facilitate large-scale analysis of molecular changes induced by chemicals.	Kits and platforms for toxicogenomics, transcriptomics, proteomics, and metabolomics to uncover mechanisms of toxicity [28] [27].

The transition to a toxicology and ecotoxicology paradigm centered on New Approach Methodologies is well underway, propelled by ethical imperatives, scientific advancements, and growing regulatory support. The global in vitro toxicology testing market, valued at USD 18.23 billion in 2024 and projected to reach USD 32.88 billion by 2030, is a testament to this rapid evolution [28]. The core scope of NAMs encompasses a wide array of human biology-based tools, from sophisticated in vitro models to powerful in silico predictions, all guided by the principles of the 3Rs.

The primary challenge is no longer the development of the technologies themselves, but the establishment of a unified, streamlined, and internationally harmonized framework for their validation and regulatory acceptance. The success of this transition hinges on continued collaboration among industry, academia, and regulators to build scientific confidence, standardize protocols, and transparently share data. As envisioned by leaders in the field, the future will likely involve a period of hybrid approaches, combining NAMs with targeted animal testing, ultimately moving toward a more human-relevant, efficient, and ethical system for safety assessment [27] [24] [32].

The NAMs Toolbox: From In Silico Models to Microphysiological Systems

The field of ecotoxicology is undergoing a transformative shift, driven by the ethical, financial, and methodological limitations of traditional animal testing. Regulatory agencies worldwide are increasingly advocating for New Approach Methodologies (NAMs) that can reduce, refine, and eventually replace animal studies [34]. Among these NAMs, in silico (computational) approaches have emerged as powerful tools for predicting the ecotoxicological effects of chemicals. These methods leverage the power of computational chemistry and machine learning to construct mathematical models that correlate the structural and chemical properties of molecules with their biological activity and toxicity profiles [35]. The core premise is that the biological activity of a substance is a function of its physicochemical properties, a principle that underpins Structure-Activity Relationship (SAR) and its quantitative counterpart, Quantitative Structure-Activity Relationship (QSAR) [36]. The adoption of these methods is not merely an ethical imperative but also an economic and scientific one, offering the potential to screen thousands of chemicals rapidly and cost-effectively, thereby providing a more scalable solution for assessing the vast number of substances in commerce and the environment [37] [38].

Core Computational Methodologies

The predictive power of computational toxicology stems from a suite of interrelated methodologies. Understanding these core approaches is essential for evaluating their applications and limitations.

• Quantitative Structure-Activity Relationship (QSAR): QSAR is a methodology that develops mathematical models to correlate quantitative descriptions of molecular structure (descriptors) with a biological or toxicological endpoint [36]. The development of a robust QSAR model requires a curated dataset of molecules with known activity, calculation of molecular descriptors, and the application of statistical methods to establish a reliable correlation [36]. These models evolve in complexity, from one-dimensional (correlating simple properties like pKa) to three-dimensional (accounting for steric and electronic interactions), and even four-dimensional models that include multiple ligand conformations [36].

• Read-Across: Read-across is a technique used to fill data gaps for a "target" chemical by using data from similar "source" chemicals. It relies on the principle that structurally similar compounds are likely to have similar biological properties and toxicological profiles [37] [39]. While powerful, its acceptance depends on the transparent justification of the chemical similarity and the biological plausibility of the transfer.

• Machine Learning (ML) and Deep Learning (DL): Modern computational toxicology heavily utilizes ML and DL, subsets of artificial intelligence. These methods can analyze complex, high-dimensional data to identify patterns that may not be apparent through traditional statistical methods [35]. Support Vector Machines (SVM), Random Forests (RF), and Deep Neural Networks (DNN) are among the most commonly used algorithms. For instance, the ToxinPredictor tool uses an SVM model that achieved state-of-the-art results with an AUROC of 91.7% in predicting molecule toxicity [38].

• Cross-Structure-Activity Relationship (C-SAR): A more recent innovation, C-SAR, accelerates structural development by analyzing pharmacophoric substitution patterns across diverse chemical classes, rather than being limited to a single parent structure. This approach can identify transformative solutions to convert an inactive compound into an active one and is particularly useful for SAR expansion [40].

The following diagram illustrates the logical workflow for developing and applying these computational models, from data collection to regulatory decision-making.

Performance Comparison of Key In Silico Tools

The landscape of available in silico tools is diverse, with numerous software and web services designed for toxicity prediction. These tools vary in their computational methods, predictive endpoints, and performance metrics. The table below summarizes a selection of representative tools, highlighting their key features and applications.

Table 1: Overview of Representative In Silico Tools for Toxicity Prediction

Tool Name	Type/Method	Key Features	Reported Performance (Examples)
ToxinPredictor	Machine Learning (SVM)	Predicts toxicity of small molecules using structural properties; includes a user-friendly webserver.	AUROC: 91.7%, F1-Score: 84.9%, Accuracy: 85.4% [38]
C-SAR	Cross-Structure-Activity Relationship	Accelerates structure development by identifying transformative pharmacophoric substitutions across diverse chemotypes.	Applied to HDAC6 inhibitors; provides strategic options for molecular transformation [40]
QSARPro	Quantitative Structure-Activity Relationship	Performs group-based QSAR, correlating chemical group variation at different molecular sites with biological activity [35].	N/A
McQSAR	Quantitative Structure-Activity Relationship	Free program to generate QSAR equations using the genetic function approximation paradigm [35].	N/A
PADEL	Descriptor Calculation	Open-source software to calculate molecular descriptors and fingerprints for model development [38] [35].	N/A
MedChem Studio	Cheminformatics	Supports lead identification, de novo design, scaffold hopping, and lead optimization [35].	N/A

A critical application of these tools is their integration into a Weight of Evidence (WoE) approach for regulatory hazard assessment. The reliability of a QSAR model for such purposes is often evaluated against specific criteria, including the definition of its Applicability Domain (AD), its scientific validity, and the possibility for a mechanistic interpretation [39].

Table 2: Key Criteria for Evaluating QSAR Models in a Regulatory Context

Evaluation Criteria	Description	Importance for Ecotoxicity Assessment
Defined Endpoint	The specific toxicological effect the model is designed to predict (e.g., fish acute toxicity, skin sensitization).	Ensures the model is fit-for-purpose and its predictions are unambiguous [36].
Unambiguous Algorithm	A transparent and well-defined mathematical model for making predictions.	Promotes reproducibility and allows for critical scientific evaluation [36].
Applicability Domain (AD)	The physicochemical, structural, or biological space on which the model was trained and for which it is applicable.	Crucial for determining the reliability of a prediction for a new chemical; predictions outside the AD are unreliable [39] [36].
Measure of Goodness-of-Fit	Statistical parameters (e.g., R², accuracy) that describe how well the model fits the training data.	Provides an initial indication of the model's predictive capability [36].
Model Validation	The process of assessing the model's performance using data not used in training (e.g., test set, external validation).	Ensures the model is robust and not over-fitted to the training data [38].

Experimental Protocols and Methodologies

Protocol for Developing a QSAR Model

The development of a robust QSAR model is a multi-step process that requires careful execution at each stage. The following protocol outlines the key steps, as derived from standard practices in the field [36] [35].

• Dataset Curation: A set of molecules with known, reliable biological activity (e.g., IC50, EC50) for the desired endpoint is assembled. The dataset should be structurally similar enough to allow for modeling but diverse enough to cover a relevant chemical space.
• Descriptor Calculation: Molecular descriptors, which are numerical representations of the molecules' structural and physicochemical properties, are calculated using software like PADEL or RDKit [38] [35]. These can range from simple properties (e.g., molecular weight, log P) to complex quantum mechanical indices.
• Data Splitting: The dataset is randomly divided into a training set (typically 70-80%) used to build the model and a test set (20-30%) used to evaluate its predictive performance on unseen data.
• Model Building: Statistical or machine learning methods (e.g., Partial Least Squares, Random Forest, Support Vector Machine) are applied to the training set to establish a mathematical relationship between the descriptors and the biological activity.
• Model Validation: The model's predictive ability is assessed using the test set. Common metrics include the coefficient of determination (R²), root mean square error (RMSE) for regression, and accuracy, sensitivity, and AUROC for classification [38].
• Defining Applicability Domain: The AD of the model is defined to identify the region of chemical space where the model can make reliable predictions. This is often based on the leverage and residual of the compounds in the training set.

Protocol for a C-SAR Analysis

The C-SAR approach offers a complementary strategy focused on molecular transformation [40].

• Library Construction: A chemical library of diverse chemotypes targeting a specific biological entity (e.g., HDAC6) is compiled from databases like ChEMBL, with consistent biological activity data.
• Matched Molecular Pairs (MMPs) Generation: The library is analyzed to identify MMPs—pairs of molecules that differ only by a single, specific structural transformation at a particular site.
• Activity Landscape Analysis: The activity difference (e.g., high vs. low potency) for each MMP is calculated. Pairs with large activity differences for small structural changes are known as "activity cliffs" and are of high interest.
• Identification of C-SAR Highlights: The analysis searches for repetitive pharmacophoric substitution patterns across different MMP chemotypes that consistently lead to activity cliffs.
• Application: The identified C-SAR highlights provide strategic guidance for converting an inactive compound into an active one, applicable even to novel chemotypes not present in the original dataset.

The Scientist's Toolkit: Essential Research Reagents and Software

Successful implementation of in silico ecotoxicity prediction relies on a suite of software tools and computational resources. The table below details key solutions used by researchers in the field.

Table 3: Key Research Reagent Solutions for In Silico Ecotoxicology

Tool / Resource	Type	Primary Function
DataWarrior	Software	An open-source program for data visualization and analysis, used for cheminformatics tasks, dataset curation, and initial SAR exploration [40] [35].
RDKit	Open-Source Library	A collection of cheminformatics and machine learning software written in C++ and Python, used for calculating molecular descriptors, fingerprinting, and integrating with ML workflows [38] [35].
PADEL	Open-Source Software	Used specifically for calculating molecular descriptors and fingerprints, which are essential inputs for QSAR and machine learning models [38] [35].
KNIME	Software Platform	An open-source platform for data analytics that integrates various cheminformatics and machine learning nodes (e.g., RDKit, WEKA) to create complex workflow for virtual screening and model building [35].
ChEMBL	Database	A manually curated database of bioactive molecules with drug-like properties, providing high-quality SAR data for model training and validation [40].
SHAP	Analysis Framework	(SHapley Additive exPlanations) A game-theoretic approach used to interpret the output of machine learning models, identifying the most important molecular descriptors contributing to a toxicity prediction [38].

The integration of these tools into a coherent workflow is fundamental for modern computational toxicology. The following diagram maps the tools to a generalized predictive modeling workflow, showing how they contribute to different stages of the process.

In silico approaches, particularly SAR and computational toxicology models, have firmly established themselves as indispensable components of the ecotoxicologist's toolkit. Framed within the broader thesis of validating NAMs, these methods offer a compelling alternative to animal testing that is not only more ethical but also faster, less expensive, and capable of handling the vast scale of chemical assessment required today [37] [34]. While challenges remain—particularly regarding standardization, transparency, and regulatory acceptance—the continuous improvement of machine learning algorithms, the expansion of high-quality datasets, and the development of innovative methodologies like C-SAR are rapidly advancing the field [40] [39] [35]. The future of ecotoxicity prediction lies in the intelligent integration of these in silico tools within a Weight of Evidence framework, potentially combined with other NAMs such as organ-on-a-chip technology, to build a more predictive, mechanistic, and humane safety science for the 21st century [37] [41].

The field of toxicology and drug development is undergoing a fundamental paradigm shift, moving from traditional animal-based testing toward more human-relevant, efficient, and ethical New Approach Methodologies (NAMs) [1]. This transition is driven by significant limitations of animal models, including species variation that can lead to inaccurate human response predictions, high costs, time-consuming procedures, and ethical concerns [42] [1]. The passing of the FDA Modernization Act 2.0 in 2022 marked a pivotal regulatory milestone by eliminating the mandatory requirement for animal testing before human drug trials, accelerating the adoption of advanced in vitro models [43] [44]. Among the most promising NAMs are organoids, organ-on-a-chip (OoC) systems, and high-throughput assays, which offer more predictive platforms for disease modeling, drug screening, and safety assessment [43] [1]. This guide provides a comparative analysis of these technologies, focusing on their applications in validating animal testing alternatives for ecotoxicology and pharmaceutical research.

Organoids and organs-on-chips represent two advanced types of three-dimensional (3D) culture systems designed to mimic critical structural and functional features of human tissues or organs in vitro [43]. While both aim to overcome the limitations of traditional 2D cell cultures, they differ significantly in their design principles, capabilities, and applications.

Organoids: Self-Organizing 3D Microtissues

Organoids are complex 3D structures derived from the self-assembly of stem cells (either adult stem cells or induced pluripotent stem cells) under specific culture conditions or from tissue explants [43] [45]. They typically range from several hundred micrometers to millimeters in size and closely recapitulate the cell types, cellular organization, and certain functions of their in vivo counterparts [43]. The development of organoids relies on the innate self-organizing capacity of stem cells grown in a gel-like extracellular matrix (such as Matrigel) with specific media formulations containing growth factors and signaling inhibitors [45].

Table 1: Key Characteristics, Advantages, and Limitations of Organoids

Feature	Organoids	Organs-on-Chips (OoCs)
Fundamental Principle	Self-assembly of stem cells [43] [45]	Engineered microfluidic culture [43] [45]
Architecture & Control	Self-organized; Limited external control [45]	Precisely controlled microenvironment [43] [45]
Physiological Relevance	Recapitulates cellular complexity and organization [43]	Mimics tissue-tissue interfaces, mechanical forces, flow [43] [45]
Throughput & Scalability	Moderate; suitable for medium-throughput screening [45]	Evolving toward high-throughput systems (HT-OoC) [46]
Key Advantages	Patient-specific modeling, genetic disease studies, cancer research [43] [45]	Incorporation of physiological flow, mechanical stimuli, real-time monitoring [43] [45]
Primary Limitations	Batch-to-batch variability, limited maturation, absence of vascularization and systemic cues [43] [45]	Engineering complexity, higher cost, expertise requirements [45]

Organs-on-Chips: Engineered Microphysiological Systems

Organs-on-chips (OoCs) are microfluidic devices that contain continuously perfused chambers inhabited by living cells arranged to simulate tissue- and organ-level physiology [43] [45]. Unlike organoids, OoCs represent a more engineered approach that leverages microfluidic technology to create a dynamic, controlled microenvironment. These systems are typically fabricated from optically transparent materials (like polydimethylsiloxane, or PDMS) and feature microchannels lined with living cells, often separated by semipermeable membranes or embedded in extracellular matrix gels [45]. This setup enables the replication of physiological fluid flow, mechanical forces (such as cyclic strain in lung chips), and chemical gradients found in human organs [43] [45].

Experimental Protocols and Methodologies

Establishing Organoid Cultures

Protocol for Patient-Derived Organoid Development [45]:

• Cell Source Isolation: Obtain adult stem cells (ASCs) or induced pluripotent stem cells (iPSCs) from patient tissue samples (e.g., tumor biopsies, blood). ASCs do not require reprogramming, preserving donor-specific disease characteristics [46].
• 3D Matrix Embedding: Suspend cells in a gel-like extracellular matrix (e.g., Matrigel or optimized collagen I) to provide structural support that mimics the in vivo environment [45] [46].
• Directed Differentiation: Culture cells in specific media formulations tailored to the target organ type, supplemented with precise combinations of growth factors, and signaling inhibitors to guide differentiation and self-organization [45].
• Culture Maintenance: Maintain organoids in specialized media, with regular passaging every 1-4 weeks depending on the organ type. Organoids can be cryopreserved for biobanking [45].

Operating Organ-on-Chip Systems

Generalized Workflow for OoC Experiments [45] [47]:

• Device Preparation: Acquire or fabricate microfluidic chips (e.g., via soft lithography or 3D printing). Some commercial plates come pre-seeded with extracellular matrix like collagen I [45] [46].
• Cell Seeding: Introduce relevant cell types (e.g., primary cells, iPSC-derived cells) into specific compartments of the chip. For example, a gut-on-a-chip model uses a mixture of epithelial and goblet cells [47].
• System Perfusion: Connect the chip to external reservoirs, pumps, and tubing to create a dynamic flow system. This perfusion supplies fresh media, removes waste, and can apply physiological mechanical stimuli [45].
• Compound Exposure and Monitoring: Introduce test compounds into the flow system. Utilize the chip's transparency for real-time, in situ imaging of cellular responses and collect effluent for analysis of metabolites and biomarkers [45] [47].

High-Throughput Screening Applications

High-throughput organ-on-chip (HT-OoC) platforms combine physiological relevance with scalability by using parallelization to increase the number of replicates per chip [46]. These systems are designed for automation, facilitating the screening of growing libraries of drug compounds [46].

Protocol for High-Throughput Screening using Multi-well OoC Platforms (e.g., OrganoPlate) [46]:

• Platform Selection: Utilize SBS-compliant plates (e.g., 40-, 64-, or 96-independent tissue culture chips) compatible with automated liquid handling systems.
• Automated Cell Seeding: Seed cells into the microfluidic channels using robotic liquid handling. The OrganoPlate's design, for instance, features three adjacent channels per chip—one for ECM gel and two for perfusion—allowing direct access to both apical and basolateral sides of the culture.
• Parallelized Compound Testing: Apply different drug candidates or concentrations to individual chips in parallel via automated systems.
• High-Content Analysis: Employ integrated analytical techniques such as automated, in situ microscopy, and online sensors to monitor barrier integrity, cell viability, and functional endpoints across all chips simultaneously.

Diagram 1: A decision workflow to guide researchers in selecting the most appropriate in vitro model based on their specific research objectives and constraints.

Performance Data and Validation

Robust experimental data demonstrates the predictive capacity of these advanced models. The following table summarizes key performance metrics from validation studies.

Table 2: Experimental Performance Data of Advanced In Vitro Models

Model Type	Application / Assay	Key Experimental Finding	Predictive Performance	Reference
Liver-Chip	Predictive Toxicology	Evaluation of 27 known hepatotoxic and non-toxic drugs	Sensitivity: 87%, Specificity: 100% [44]
Proximal Tubule OoC	Nephrotoxicity Screening	Tested antisense oligonucleotide SPC-5001	Successfully predicted clinical nephrotoxicity not detected in mice/non-human primates [44]
Vessel-Chip	Thrombosis Assessment	Evaluated prothrombotic effect of anti-CD40L mAb (Hu5c8)	Detected thrombotic complications that occurred in clinical trials but not in preclinical animal tests [44]
In Silico Trials (Human Cardiomyocytes)	Pro-Arrhythmic Cardiotoxicity	Population-based drug trials using computer models	89% accuracy in predicting clinical arrhythmia (vs. 75% for animal models) [44]
Multi-Organ OoC (Gut/Liver)	ADME (Absorption, Distribution, Metabolism, Excretion)	Interconnected gut and liver models to study organ crosstalk	Allows sampling for concentration-time profiles; compares oral/IV dosing [47]

Essential Research Reagents and Materials

Successful implementation of organoid and OoC technologies requires specific reagents and materials. The following table details key components used in these experimental workflows.

Table 3: Essential Research Reagent Solutions for Advanced In Vitro Models

Reagent / Material	Function / Application	Examples / Specifications
Extracellular Matrix (ECM)	Provides 3D structural support for cell growth and organization	Matrigel, optimized collagen I (e.g., rat-tail collagen I) [45] [46]
Specialized Culture Media	Supports cell survival, proliferation, and directed differentiation	Organ-specific formulations with growth factors (e.g., EGF, Wnt-3A) and signaling inhibitors [45]
Cell Sources	Forms the basis of the 3D tissue model	Adult stem cells (ASCs), induced pluripotent stem cells (iPSCs), primary cells (e.g., hepatocytes, endothelial cells) [45] [47]
Microfluidic Devices	Serves as the platform for housing tissues and enabling perfusion	AIM Biotech (idenTx, organiX), Emulate, MIMETAS (OrganoPlate), CN Bio (PhysioMimix) [46] [48] [47]
Analysis Kits & Assays	Enables assessment of cell viability, function, and response	Metabolite detection kits, cell viability assays (e.g., ATP-based), barrier integrity measurements (TEER) [47]

Organoids and organs-on-chips represent complementary, rather than competing, technologies in the New Approach Methodologies landscape [45]. Organoids excel in modeling patient-specific diseases and cancer biology due to their cellular complexity and genetic fidelity [43] [45]. In contrast, OoCs provide superior capabilities for investigating drug transport, pharmacokinetics, and toxicity by replicating dynamic physiological microenvironments [43] [45] [47]. The emerging integration of these systems into "organoids-on-a-chip" combines the biological fidelity of organoids with the controlled perfusion and analytical capabilities of microfluidic platforms, offering a promising path toward more predictive human-relevant models [45]. For researchers in ecotoxicology and drug development, the selection between these technologies should be guided by the specific research question, with organoids ideal for genetic and personalized medicine studies and OoCs better suited for ADME and toxicological applications requiring physiological relevance. The continued development and standardization of these NAMs are crucial for advancing toward a future with significantly reduced reliance on animal testing.

Leveraging Omics Technologies for Mechanistic Insights into Toxicological Pathways

In the face of a global push to reduce animal testing, toxicology is undergoing a fundamental transformation toward predictive, mechanism-based approaches. New Approach Methodologies (NAMs) that leverage omics technologies are central to this shift, providing the deep mechanistic insights needed to support quicker, more human-relevant risk assessments [49]. This guide compares the performance of key omics technologies—transcriptomics, metabolomics, and proteomics—in illuminating toxicological pathways, providing objective data to help researchers select the right tools for de-risking drug development.

Core Omics Technologies in Toxicology

Omics technologies enable the systematic study of large sets of biological molecules to uncover the molecular and cellular changes that occur in response to chemical exposures [49]. The table below compares the three most prominent omics disciplines in modern toxicology.

Table 1: Comparison of Core Omics Technologies in Toxicological Research

Technology	Primary Focus	Key Advantage in Toxicology	Commonly Used For
Transcriptomics [49] [50]	Profiling of all RNA molecules (the transcriptome)	Most mature omics technology; highly sensitive for detecting early biological perturbations [49] [50]	Deriving transcriptomic Points of Departure (tPODs); mechanism of action identification; potency ranking [49]
Metabolomics [50] [51]	Analysis of small-molecule metabolites (the metabolome)	Closest to phenotype; can reveal toxicity signatures at lower doses than classical methods [51]	Identifying sensitive biomarkers of toxicity; understanding metabolic pathway disruptions [50] [51]
Proteomics [50] [52]	Study of the full set of proteins (the proteome)	Directly reflects functional cellular activity and signaling events [52]	Uncovering protein-level signaling perturbations and post-translational modifications [50] [52]

Experimental Protocols for Omics-Informed Risk Assessment

Integrating omics into toxicology studies requires specific experimental designs to generate regulatory-grade data. The following workflow is characteristic of studies used to support Next-Generation Risk Assessment (NGRA).

Protocol 1: Short-Term In Vivo Study with Omics Endpoints

This protocol is designed as a bridge between traditional animal testing and fully in vitro-based hazard assessment [49] [50].

• 1. Study Design: A typical design involves a 5- to 28-day repeated oral dose study in rodents. The study should include 8 or more dose groups to ensure robust benchmark dose (BMD) modeling, along with a control group [49] [50].
• 2. Tissue Sampling and Preparation: At study termination, potential target organs (e.g., liver, kidney) are collected. Tissues are flash-frozen in liquid nitrogen or placed in stabilizers like RNAlater to preserve molecular integrity. Consistent cryopreservation strategies, including strict documentation of time and temperature, are critical for reproducibility [51].
• 3. Omics Analysis: RNA, metabolites, or proteins are extracted from the tissues.
  • For Transcriptomics: Use robust, high-throughput RNA-seq methods (e.g., DRUG-seq, BRB-seq) for cost-effective, transcriptome-wide analysis [50]. Process samples for next-generation sequencing.
  • For Metabolomics: Use mass spectrometry (MS) or nuclear magnetic resonance (NMR) spectroscopy to profile small molecules in biological samples [51].
• 4. Data Processing and Bioinformatics: This is a critical step for ensuring reliable and reproducible results.
  • Transcriptomics Data: Process raw sequencing data through a standardized bioinformatics pipeline. This typically includes quality control, alignment to a reference genome, and normalization. Differentially expressed genes are then identified [50].
  • Use tools like BMDExpress for benchmark dose (BMD) modeling. The software fits dose-response models to gene expression data to identify the lowest dose at which a significant molecular change occurs [50]. Frameworks like the Regulatory Omics Data Analysis Framework (R-ODAF) can help standardize the selection of differentially expressed genes and pathway-level interpretation [50].
• 5. Derivation of a Molecular Point of Departure (POD): The final step is to calculate a transcriptomic Point of Departure (tPOD). This is defined as the lower 95% confidence limit of the lowest median benchmark dose (BMD) for a gene ontology biological process set that shows consistent changes and contains at least 3 genes [49]. This tPOD can then be used to derive a Transcriptomic Reference Value (TRV) for human health risk assessment [49].

Figure 1: A generalized workflow for generating omics data to support a Next-Generation Risk Assessment (NGRA). The process moves from in vivo study design to the application of computational methods for deriving a protective Point of Departure (POD).

Performance Comparison: Omics vs. Traditional Apical Endpoints

A key measure of an omics technology's performance is how its results align with those from traditional, longer-term studies. The following quantitative data summarizes this comparison.

Table 2: Concordance Between Omics-Derived Points of Departure and Traditional Apical Endpoints

Study Focus / Chemical	Omics Technology Used	Molecular POD	Traditional Apical POD	Fold Difference	Experimental Summary
Data-poor PFAS (MOPA) [49] [50]	Transcriptomics (targeted RNA-seq)	0.09 µg/kg-day (tPOD)	Not available (data-poor)	N/A	5-day rat study; tPOD derived from BMD modeling of gene ontology processes in multiple tissues.
Developmental/Reproductive Toxicity (Dicyclohexyl phthalate) [49]	Transcriptomics (fetal testis)	Specific value not stated	Apical developmental/reproductive POD	Within 2.5-fold	In utero exposure pilot study; tPODs from fetal tissue showed strong concordance with apical endpoints.
Developmental PFAS & Co-exposure [49]	Metabolomics & Transcriptomics	Metabolomic & tPODs	Lowest concurrent apical data	Within 3- to 8-fold	Short-term study comparing pup and maternal responses; enabled ranking of individual chemicals and mixture by potency.

The data shows that omics-derived PODs from short-term studies are consistently within a narrow margin of traditional PODs, demonstrating their utility as conservative and protective alternatives for regulatory decision-making [49].

The Scientist's Toolkit: Essential Research Reagents & Platforms

Successful implementation of omics in toxicology relies on a suite of wet-lab and computational tools.

Table 3: Key Research Reagents and Solutions for Omics Toxicology

Item / Solution	Category	Primary Function in Omics Toxicology
RNAlater / TRIzol	Sample Stabilization	Preserves RNA integrity in tissues and cells immediately after collection, preventing degradation.
High-Throughput RNA-seq Kits (e.g., DRUG-seq, BRB-seq)	Genomics Reagent	Enables scalable, cost-effective, transcriptome-wide gene expression profiling from low-input samples [50].
Mass Spectrometry (MS) Platforms	Metabolomics/Proteomics Instrument	Identifies and quantifies small-molecule metabolites or proteins for functional phenotyping.
BMDExpress	Bioinformatics Software	Performs benchmark dose (BMD) modeling on transcriptomic or other omics data to derive points of departure [50].
CompTox Chemicals Dashboard (EPA)	Data Resource	Provides access to chemistry, toxicity, and exposure data for chemical interpretation and prioritization [16].
Adverse Outcome Pathway (AOP) Knowledgebase	Conceptual Framework	Organizes mechanistic knowledge from a Molecular Initiating Event (MIE) to an adverse outcome at the organism level [16].

Standardization and Bioinformatics: The Path to Regulatory Acceptance

The regulatory application of omics depends on consistent, reproducible bioinformatics pipelines, as variability in workflows can significantly influence results [49] [50]. The diagram below outlines a standardized bioinformatics process for transcriptomic data.

Figure 2: A standardized bioinformatics pipeline for processing transcriptomic data, from raw sequencing reads to the derivation of a transcriptomic Point of Departure (tPOD).

Collaborative efforts, such as those led by the Health and Environmental Sciences Institute (HESI), are actively working to harmonize these bioinformatics pipelines across sectors to build regulatory confidence [49] [50]. The use of tools like BMDExpress and frameworks like R-ODAF is crucial for filtering non-biologically relevant signals and ensuring robust, interpretable results [50].

The integration of omics technologies into toxicological research is no longer a theoretical future but an ongoing reality. As a core component of NAMs, omics provide the mechanistic depth, sensitivity, and human relevance required to accelerate safety evaluations, reduce reliance on animal testing, and build a more predictive toxicology paradigm. For researchers and drug development professionals, the strategic adoption of these tools—coupled with standardized protocols and a clear understanding of their performance—offers a competitive advantage in building safer products and navigating the evolving regulatory landscape.

The increasing synthesis and application of multicomponent nanomaterials (MCNMs)—such as bimetallic nanoparticles, doped metal oxides, and surface-functionalized materials—has heightened concerns regarding their potential environmental impact [53] [54]. Traditional in vivo ecotoxicity testing, while reliable, is often too resource-intensive and ethically demanding to keep pace with the rapid development of novel nanomaterials [55]. Consequently, U.S. federal agencies and the international scientific community are prioritizing the development and regulatory acceptance of New Approach Methodologies (NAMs) to improve efficiency and reduce animal use in chemical safety evaluations [55] [56] [5].

This case study examines a specific in silico NAM: a classification Structure-Activity Relationship (SAR) computational framework for predicting the ecotoxicity of metal and metal oxide MCNMs [53] [54]. Such models are crucial for enabling the early hazard assessment of a vast and growing number of nanomaterials, supporting the development of safe-by-design nanomaterials, and addressing the mandates of initiatives like the U.S. Strategic Roadmap for establishing new safety evaluation methods [55] [56].

Experimental Protocol: Building the Classification SAR Model

Data Curation and Toxicity Classification

The model was constructed using a substantial and heterogeneous dataset, representing the largest compiled collection of MCNM ecotoxicity data to date [54].

• Data Source: The dataset was curated from 102 scientific publications, comprising 652 individual ecotoxicity measurements for 214 distinct metal and metal oxide MCNMs [54].
• Toxicity Endpoints: The data included half-maximal effective concentration (EC50), inhibitory concentration (IC50), and lethal concentration (LC50) values, collectively referred to as EC50 for clarity [54].
• Biological Targets: The model encompassed ecotoxicity towards a wide range of organisms, including bacteria (e.g., E. coli, S. aureus), eukaryotes (e.g., C. albicans), fish (D. rerio), crustaceans (D. magna), and plants [54].
• Classification Scheme: Each measurement was classified as "toxic" or "non-toxic" based on a predefined threshold of 100 mg/L, a standard established in prior nanotoxicology research [54].

Descriptor Calculation and Model Development

The core of the SAR approach lies in identifying physicochemical descriptors that correlate strongly with toxicological activity.

• Descriptor Selection: The study found that just two fundamental descriptors could effectively classify MCNM ecotoxicity across the entire heterogeneous dataset [53] [54]:
  • Hydration Enthalpy (ΔHhyd) of the Metal Ion: This property influences the stability and dissolution behavior of the nanomaterial in an aqueous environment, which is critical for its biological reactivity [54].
  • Energy Difference (ΔE) between the MCNM Conduction Band and the Redox Potential of Biological Media: This descriptor reflects the potential of the nanomaterial to induce electron-transfer-driven oxidative stress, a primary mechanism of nanomaterial toxicity [53] [54].
• Modeling Approach: A computational framework was developed to classify MCNMs as toxic or non-toxic based on the computed values of these two key descriptors for a given MCNM's composition and structure [54].

The experimental workflow is summarized in the diagram below.

Results & Discussion: Performance and Mechanistic Insight

Key Findings and Model Interpretation

The primary achievement of this study was the demonstration that a simple model based on two intuitive descriptors can provide a holistic understanding of MCNM ecotoxicity.

• Predictive Power: The model achieved adequate classification of MCNM ecotoxicity across the entire dataset, which included diverse organisms and MCNM types [54]. This indicates the model's robustness and generalizability.
• Mechanistic Insight: The selection of hydration enthalpy and conduction band energy difference is not arbitrary; it directly points to dominant toxicity pathways. ΔHhyd is linked to ion release and dissolution toxicity, while ΔE is a key factor in oxidative stress via the generation of reactive oxygen species (ROS) [54]. This aligns with a catalytic perspective on nanotoxicity, where surface reactivity is the primary driver of biological effects [57].
• Comparison with Other Modeling Approaches The field of nano-QSAR is evolving rapidly. The table below compares this classification SAR model with another recent, advanced model, highlighting differences in scope and methodology.

Table 1: Comparison of Recent In Silico Models for Nanomaterial Toxicity

Feature	Classification SAR for MCNM Ecotoxicity [53] [54]	NanoToxRadar (Multitarget Nano-QSAR) [58]
Model Type	Classification SAR	Quantitative (Regression) Nano-QSAR
Target Application	Ecotoxicity (Environmental Organisms)	Cytotoxicity (110 Mammalian Cell Lines)
Nanomaterial Scope	Metal/Metal Oxide MCNMs	Multicomponent Nanoparticles (MC-NPs)
Key Descriptors	Hydration Enthalpy, Conduction Band Energy	Size-Dependent Electron-Configuration Fingerprint
Primary Output	Toxic/Non-Toxic Classification	pIC50 value (Quantitative Potency)
Accessibility	Described in Research Paper	Web Platform (https://www.kitox.re.kr/nanotoxradar)

Implications for Mixture and Multicomponent Nanomaterial Toxicity

Understanding MCNM toxicity is complex, as the components can interact in ways that are not simply additive. Research on mixtures of individual ENPs shows that joint effects can be additive, synergistic, or antagonistic [59]. For instance, a review found that 53% of studied nano-mixture interactions were antagonistic, 25% synergistic, and 22% additive [59]. The combination of nCuO and nZnO was one of the most frequently studied and exhibited strong interactions across all three types [59]. The mechanistic insights from the classification SAR model, particularly the role of dissolution (linked to ΔHhyd) and oxidative stress (linked to ΔE), help explain these interactions. For example, one ENP in a mixture might adsorb to another, reducing its dissolution or blocking reactive surfaces, leading to an antagonistic effect [59].

The diagram below illustrates the core toxicity mechanisms identified by the model's descriptors.

The Researcher's Toolkit for MCNM Ecotoxicity Assessment

For scientists embarking on the development or application of similar models, the following reagents, data sources, and computational tools are essential.

Table 2: Essential Research Toolkit for MCNM Ecotoxicity Modeling

Tool/Resource	Function & Application	Example / Source
ECOTOX Knowledgebase	Curated database of single-chemical ecotoxicity data; useful for benchmarking or additional data mining.	U.S. EPA (https://www.epa.gov/ecotox) [60]
ICCVAM/NICEATM Resources	Provide guidance, validation, and support for the development and evaluation of New Approach Methodologies (NAMs).	National Toxicology Program [55] [5]
Computational Descriptors	Fundamental material properties used as inputs in SAR models to predict biological activity.	Hydration Enthalpy, Conduction Band Energy [54]
Machine Learning Algorithms	Algorithms used to build predictive regression or classification models from descriptor and toxicity data.	CatBoost, Random Forest, Support Vector Machine (SVM) [58] [61]
FAIR Data Principles	A guiding framework for ensuring data is Findable, Accessible, Interoperable, and Reusable.	Critical for building robust, shared datasets for model development [60] [58]

This case study demonstrates that a classification SAR model based on a large, heterogeneous dataset can effectively predict the ecotoxicity of multicomponent nanomaterials while also providing mechanistic understanding of the underlying toxicological pathways. The identification of hydration enthalpy and conduction band energy as key descriptors underscores the importance of dissolution and oxidative stress in MCNM toxicity. This in silico approach represents a powerful NAM that aligns with the strategic goals of regulatory agencies to replace, reduce, and refine (3Rs) animal testing [5]. By enabling early-tier hazard assessment, such models facilitate the design of safer nanomaterials and contribute to the sustainable and responsible development of nanotechnology [53] [54] [56].

The field of toxicology is undergoing a fundamental paradigm shift, moving away from traditional animal models toward more human-relevant, efficient, and ethical testing approaches. This transition is driven by the recognition that traditional animal models often poorly predict human outcomes, with over 90% of oncology drugs that succeed in animal studies failing in human clinical trials due primarily to efficacy and toxicity concerns [62]. New Approach Methodologies (NAMs) represent a broad category of innovative tools—including in vitro models, computational approaches, and high-throughput screening methods—that offer more human-relevant solutions for safety assessment [16] [18].

Integrated Testing Strategies (ITS) represent the logical evolution of NAMs application, moving beyond single-method approaches to combine multiple NAMs in a strategic framework. These integrations address the fundamental limitation of any individual NAM: while excel at answering specific questions, they rarely capture the complexity of whole-organism biology. As noted by Health Canada, "currently available cell-based NAMs may lack complete biological coverage" for assessing systemic toxicity [26]. The ITS framework overcomes this by combining complementary methodologies to create a more complete safety assessment picture than any single approach could provide alone, ultimately delivering improved chemical safety assessment through more protective and relevant models that reduce reliance on animals [2].

Key NAM Platforms for Integration

Core Technological Platforms

The foundation of any Integrated Testing Strategy lies in understanding the available NAM platforms and their respective strengths and limitations. Several core technologies have reached sufficient maturity for regulatory consideration:

• Organoids: These three-dimensional cell aggregates derived from patient tumors or pluripotent stem cells retain most of the histological, genetic, and phenotypic characteristics of the source tissue. Their key strength lies in preserving intratumoral heterogeneity, making them particularly valuable for patient-specific drug sensitivity profiling and biomarker discovery [62].

• Organ-on-a-Chip (OoC) Systems: These microphysiological systems incorporate microfluidic channels to recreate the dynamic microenvironment of human organs, including fluid flow, mechanical forces, and multicellular interactions. OoC platforms provide physiologically relevant context for evaluating drug delivery, immune cell trafficking, and barrier function under controlled conditions that mimic human physiology [62] [63].

• Advanced In Vitro Assays: This category includes functional assays such as multielectrode arrays (MEA) for measuring neuronal and cardiac electrical activity, impedance-based systems for monitoring cell viability and barrier integrity, and high-content imaging platforms. These tools provide functional readouts of cellular responses to compounds in real-time without labels or dyes [63].

• Computational and AI/ML Models: In silico approaches range from quantitative structure-activity relationship (QSAR) models and physiologically based pharmacokinetic (PBPK) modeling to artificial intelligence and machine learning algorithms that can identify patterns in complex datasets. These methods excel at data integration and prediction, particularly when anchored to known agents within the same therapeutic class [62] [64].

Comparative Analysis of NAM Platforms

Table 1: Comparison of Key NAM Platforms for Integrated Testing Strategies

Platform	Key Strengths	Technical Limitations	Primary Applications in ITS	Regulatory Readiness
Organoids	Retains patient-specific heterogeneity; Self-organizing 3D structure	Limited immune/vascular components; Standardization challenges	Patient-specific drug profiling; Biomarker discovery; Mechanistic studies	Intermediate (Case-by-case acceptance)
Organ-on-a-Chip	Physiologically relevant flow and shear stress; Multicellular interactions	Technical complexity; Limited throughput; High development cost	Barrier function studies; ADME evaluation; Immune cell trafficking	Emerging (Pilot programs)
In Silico & AI/ML	High throughput; Low cost; Mechanism-independent	Dependent on quality training data; Limited biological resolution	Chemical prioritization; Read-across; Risk-based screening	Advanced for specific endpoints
High-Throughput Screening	Rapid data generation; Cost-effective; Standardized protocols	Limited biological complexity; Often reductionist	Early hazard identification; Toxicity screening; Prioritization	Advanced for specific endpoints

Framework for Integrating Multiple NAMs

Foundational Principles of Integration

Successful integration of multiple NAMs requires more than simply running different assays in parallel. It demands a strategic framework guided by several key principles:

• Defined Context of Use: The most critical factor for regulatory acceptance is establishing a clear "context of use" for each NAM within the integrated strategy [64]. This involves precisely defining what biological question each method addresses and how the combined data will inform specific safety decisions. Regulatory agencies are more likely to accept NAM data when sponsors provide robust scientific justification for their application to specific endpoints [63] [64].

• Fit-for-Purpose Validation: Unlike traditional validation against animal data, NAM validation should focus on demonstrating human biological relevance and reliability for the specific context of use. As noted by researchers, "NAMs do not aim to recapitulate the animal test without the animal, but to provide more relevant information on a chemical to allow exposure-based safety assessment" [2].

• Tiered Testing Approach: Effective ITS implement a tiered strategy that begins with higher-throughput, lower-complexity assays for prioritization, followed by progressively more complex and physiologically relevant models for compounds of interest. This approach efficiently allocates resources while gathering data of appropriate complexity for decision-making [26] [2].

Implementation Workflows

The specific configuration of an ITS varies depending on the assessment goals, but several proven workflows have emerged:

Diagram 1: Tiered testing strategy workflow for NAM integration.

For complex endpoints like systemic toxicity, a hypothesis-driven workflow has proven effective:

Diagram 2: Hypothesis-driven workflow based on adverse outcome pathways.

Validation and Regulatory Considerations

Building Confidence in Integrated Approaches

The regulatory acceptance of Integrated Testing Strategies depends on demonstrating scientific validity and reliability through several key approaches:

• Case Studies and Cross-Method Validation: Regulatory confidence is built through accumulating evidence from well-documented case studies. For example, a comprehensive NAM testing strategy for crop protection products Captan and Folpet employed 18 different in vitro studies and successfully identified these compounds as contact irritants, demonstrating that appropriate risk assessment could be performed with available NAM tests [2].

• Mechanistic Validation: Rather than focusing solely on correlating NAM results with animal data, successful validation frameworks emphasize biological plausibility and understanding of mechanisms. The U.S. EPA actively develops Adverse Outcome Pathways (AOPs) to establish scientific rationale supporting the use of NAMs in evaluating potential chemical impacts [16].

• Interlaboratory Reproducibility: Standardized protocols and demonstration of reproducibility across laboratories are essential for regulatory acceptance. Recent initiatives like the Complement-ARIE public-private partnership aim to accelerate the development and evaluation of NAMs through collaborative validation studies [65].

Regulatory Landscape and Pathways

Significant regulatory developments have created pathways for incorporating ITS into safety assessment:

• FDA Modernization Act 2.0: This legislative change permits the use of "scientifically justified, robust, and fit-for-purpose non-animal methods" to support regulatory submissions, effectively opening the door for integrated NAM strategies [62] [63].

• Pilot Programs for Biologics: The FDA has initiated pilot programs focusing on monoclonal antibodies and other biologics, allowing select developers to use NAMs-based testing strategies that will inform broader policy changes [63].

• International Harmonization Efforts: Organizations like the OECD are developing test guidelines for Defined Approaches (specific combinations of NAMs with fixed data interpretation procedures), which have been formalized in guidelines for skin sensitization and serious eye damage/eye irritation [2].

Essential Research Tools and Reagents

Research Reagent Solutions for NAM Implementation

Table 2: Key Research Reagents and Platforms for NAM Integration

Tool Category	Specific Examples	Primary Function in ITS	Key Features
Stem Cell Models	iPSC-derived cardiomyocytes/neurons	Provide human-relevant cells for functional toxicity assessment	Human biology; Differentiate into multiple cell types
Microphysiological Systems	Organ-on-chip platforms	Recreate tissue-tissue interfaces and physiological flow	Fluid flow; Mechanical stimulation; Multicellular interactions
Functional Assessment Platforms	Multielectrode array (MEA) systems	Measure real-time electrical activity of excitable cells	Label-free; Non-invasive; High-content functional data
Biomarker Detection	Omics technologies (transcriptomics, proteomics)	Identify mechanistic signatures of toxicity	Pathway analysis; Mechanism identification
Barrier Integrity Models	Transwell systems with TEER measurement	Assess barrier function in gut, lung, blood-brain barrier models	Quantitative integrity measurement; Permeability assessment
Computational Tools	PBPK modeling platforms; AI/ML algorithms	Extrapolate in vitro data to human exposure contexts	Data integration; Human exposure prediction

Integrated Testing Strategies represent the future of safety assessment, moving beyond one-to-one replacement of animal tests toward a more holistic, human-relevant framework that combines multiple NAMs in a strategic manner. The success of these integrated approaches depends on several critical factors: establishing clear contexts of use, demonstrating human biological relevance rather than simply correlating with animal data, and engaging early with regulatory agencies to align on validation strategies.

The field continues to evolve rapidly, with emerging areas including the development of virtual tissue models that use computer simulations to predict chemical effects on human development [16], and the integration of AI/ML approaches to interpret complex multimodal data from different NAM platforms [64]. Furthermore, international collaborations and public-private partnerships are accelerating the validation and qualification of integrated approaches [65].

As the scientific community builds confidence in these strategies through case studies, cross-laboratory reproducibility testing, and continued technological refinement, ITS are poised to transform safety assessment across multiple sectors—delivering more human-relevant, predictive, and efficient evaluation while reducing reliance on traditional animal models.

Navigating Implementation Hurdles: Standardization, Data Quality, and Regulatory Acceptance

Overcoming Challenges in Physico-Chemical Characterization of Complex Materials like Nanomaterials

The field of nanotechnology has revolutionized medicine, energy, and materials science, with nanomaterials serving as transformative mediators across diverse scientific disciplines [66] [67]. However, the unique physicochemical properties that make nanomaterials so promising—including their high surface area-to-volume ratio, quantum effects, and size-dependent behavior—also present significant characterization challenges [68] [69]. Thorough characterization is critically important not only to provide a complete picture of the nanomaterial itself and to establish structure-property relationships, but also to provide feedback in nanomaterial design, as their physiochemical characteristics directly affect their performance, utility, and safety [70].

The complexity of nanomaterials is magnified in regulatory and safety contexts, where the particle-chemical duality of nanomaterials complicates several aspects of their safety assessment [69]. Established test methods for identifying potential hazards often require adjustment or confirmation for nanomaterials, as their behavior in biological and environmental systems differs markedly from conventional chemicals [69]. This characterization challenge becomes particularly acute within the framework of New Approach Methodologies (NAMs), which aim to replace, reduce, and refine animal testing in ecotoxicology research [71] [72]. Without comprehensive and reliable characterization data, the development and validation of these alternative methods face significant hurdles, potentially slowing the transition toward more human-relevant and ethical testing paradigms [72] [69].

Critical Characterization Parameters and Technical Challenges

Key Physico-Chemical Properties

For complex nanomaterials, several physico-chemical parameters must be thoroughly characterized to understand their behavior, potential applications, and toxicological profiles. These properties govern the interactions of nanomaterials with biological systems and the environment, ultimately dictating their utilities, performance, and fate [70]. The table below summarizes the essential characterization parameters and the specific challenges they present in the context of safety assessment and NAMs development.

Table 1: Essential Characterization Parameters for Nanomaterials and Associated Challenges

Parameter Category	Specific Properties	Characterization Challenges	Relevance to NAMs and Ecotoxicology
Size & Morphology	Size, size distribution, shape, surface area, aggregation/agglomeration state	Dynamic nature in biological media; interference from complex matrices [69]	Determines cellular uptake, biodistribution, and toxicity endpoints [69]
Surface Properties	Surface chemistry, ligand structure, surface charge (zeta potential), hydrophobicity, binding affinity	Low ligand concentration, heterogeneity, surface curvature effects [70]	Governs protein corona formation, cellular interactions, and biocorona evolution [70] [69]
Structural Properties	Crystal structure, elemental composition, purity, chemical composition	Nanoscale effects on crystallinity; trace element detection	Influences dissolution rates, catalytic activity, and generation of reactive oxygen species [73]
Application-Specific Properties	Radiolabeling efficiency (for theranostics), magnetic properties, drug loading and release	Maintaining functionality while ensuring stability; quantifying in complex systems [66]	Critical for validating NAMs for specific applications like targeted drug delivery [66] [74]

Technical Obstacles in Characterization

Characterizing these parameters presents multiple technical obstacles that complicate safety assessment and NAMs validation. A primary challenge is the lack of reference nanomaterials with well-defined properties, which are essential for method calibration and comparison across laboratories [68] [69]. The dynamic transformation of nanomaterials in different media further complicates characterization; nanoparticles in biological or environmental matrices are surrounded by a corona of biomolecules that continually evolves, altering their identity and behavior from the originally characterized material [69].

Additionally, analytical limitations persist in detecting and quantifying nanomaterials in complex matrices like tissue homogenates or environmental samples [69]. The dose metrics for nanomaterials also present conceptual challenges, as traditional concentration-based approaches (e.g., mg/L) may be less relevant than particle number or surface area for certain toxicological endpoints [69]. These technical obstacles highlight the need for complementary characterization approaches and careful interpretation of data within the context of NAMs development.

Comparative Analysis of Characterization Techniques

No single technique can fully characterize the diverse properties of nanomaterials, necessitating a combinatorial approach that leverages the strengths of multiple methodologies [73]. The table below provides a comparative analysis of major characterization techniques, their applications, and limitations, with particular emphasis on their utility for generating reliable data for NAMs validation.

Table 2: Comparison of Techniques for Nanomaterial Characterization

Technique	Principles	Information Obtained	Strengths	Limitations	Utility for NAMs Development
Electron Microscopy (TEM/SEM)	Electron beam interaction with sample	Size, shape, morphology, composition, crystal structure	High resolution; direct visualization	Expensive; vacuum conditions; sample preparation artifacts; statistical representation	Provides crucial baseline characterization for test material identity [73] [67]
Nuclear Magnetic Resonance (NMR) Spectroscopy	Magnetic properties of atomic nuclei	Ligand structure, conformation, density, binding mode, dynamics	Comprehensive molecular information; non-destructive	Signal broadening for bound ligands; requires significant sample amount	Elucidates surface-biomolecule interactions critical for corona formation [70]
Dynamic Light Scattering (DLS)	Brownian motion of particles in suspension	Hydrodynamic size, size distribution, aggregation state	Rapid measurement; minimal sample preparation	Limited resolution; biased by large particles/aggregates	Monitoring stability in exposure media for ecotoxicity tests [73]
Fourier-Transform Infrared (FTIR) Spectroscopy	Molecular vibrations	Surface chemistry, functional groups, ligand identity	Accessible; minimal sample preparation	Limited quantitative analysis; interpretation complexity in complex systems	Verification of surface modifications in engineered nanomaterials [70]
X-ray Photoelectron Spectroscopy (XPS)	Photoelectric effect	Elemental composition, chemical state, electronic state	Surface-sensitive (top 10 nm); quantitative	High vacuum; limited spatial resolution	Assessing surface purity and oxidation states relevant to toxicity [73]

Advanced Workflows and Integrated Methodologies

Complementary Characterization Workflow

Overcoming characterization challenges requires integrated workflows that combine multiple techniques to build a comprehensive understanding of nanomaterial properties. The following diagram illustrates a logical workflow for complementary characterization of nanomaterials, emphasizing how different techniques address various aspects of nanomaterial properties in a coordinated approach:

Experimental Protocols for Key Characterization Methods

NMR Analysis of Surface Ligands

Objective: To characterize the structure, conformation, and dynamics of surface ligands on nanomaterials using solution-phase NMR spectroscopy [70].

Methodology:

• Sample Preparation: Prepare a concentrated dispersion of functionalized nanoparticles in deuterated solvent (e.g., D₂O, CDCl₃). For quantitative analysis, use a known concentration of internal standard (e.g., 1,3,5-trioxane for ¹H NMR) [70].
• Data Acquisition: Acquire ¹H NMR spectrum using standard pulse sequences. For advanced structural analysis, implement 2D techniques:
  • DOSY (Diffusion Ordered Spectroscopy): To differentiate bound and unbound ligands based on diffusion coefficients [70].
  • TOCSY (Total Correlation Spectroscopy): To determine through-bond correlations between protons within the ligand structure [70].
  • ROESY (Rotating Frame Overhauser Effect Spectroscopy): To obtain through-space correlations and connectivity information between neighboring ligands [70].
• Data Analysis:
  • Compare chemical shifts of free ligands versus bound ligands to confirm surface attachment.
  • Analyze peak broadening and relaxation times (T₂) to understand ligand dynamics and packing density.
  • For quantitative analysis, integrate characteristic proton peaks relative to internal standard to determine ligand density.

Applications in NAMs: This protocol enables precise characterization of surface modifications that dictate nanomaterial-biomolecule interactions, providing critical data for understanding test article identity in in vitro systems [70].

Radiolabeling Characterization for Theranostic Nanomaterials

Objective: To determine radiolabeling efficiency and stability of radiolabeled nano-formulations for theranostic applications [66].

Methodology:

• Radiolabeling Procedure:
  • For ¹⁷⁷Lu-labeled magnetic nano-formulations: Incubate the nanomaterial with ¹⁷⁷Lu-chloride in the presence of a chelator (e.g., DOTA, DTPA) at optimized pH, temperature, and reaction time [66].
  • Purify the radiolabeled product using size exclusion chromatography or membrane filtration.
• Quality Control:
  • Determine radiolabeling efficiency using instant thin-layer chromatography (iTLC) or paper chromatography.
  • Assess in vitro stability by incubating in physiological buffers (e.g., PBS) and human serum at 37°C, with periodic sampling over 24-48 hours.
  • Determine radiochemical purity using high-performance liquid chromatography (HPLC) with radiodetection.
• Characterization of Physicochemical Properties:
  • Confirm that radiolabeling does not alter size, surface charge, or morphology using DLS, zeta potential measurements, and TEM.
  • Evaluate magnetic properties using vibrating sample magnetometry (VSM).

Applications in NAMs: This protocol enables tracking of nanomaterial distribution and uptake in in vitro systems, providing quantitative data for pharmacokinetic modeling without animal use [66].

The Scientist's Toolkit: Essential Research Reagents and Materials

Successful characterization of complex nanomaterials requires specialized reagents and reference materials. The table below details key research solutions essential for overcoming characterization challenges in nanomaterials research, particularly in the context of NAMs development.

Table 3: Essential Research Reagents and Materials for Nanomaterial Characterization

Reagent/Material	Function	Specific Application Examples	Considerations for NAMs
Stable Isotope-labeled Compounds	Tracing and quantification in complex matrices	¹³C, ¹⁵N-labeled precursors for NMR tracking; stable isotope-labeled biomarkers	Enables precise tracking without radiological concerns; supports read-across approaches [70] [69]
Reference Nanomaterials	Method calibration, interlaboratory comparison	Certified particle size standards; well-characterized nanogold for TEM calibration	Essential for assay validation and regulatory acceptance of NAMs [68] [69]
Surface Modification Reagents	Controlled surface functionalization	Thiol-terminated PEG for gold NPs; silane coupling agents for metal oxides	Enables systematic study of structure-activity relationships critical for safe-by-design approaches [70] [74]
Advanced Chelators and Radiolabels	Theranostic nanomaterial development	DOTA, NOTA, and other macrocyclic chelators for radiolabeling with ¹⁷⁷Lu, ⁶⁴Cu	Facilitates development of theranostic NAMs with integrated imaging capabilities [66]
Biomolecular Corona Standards	Standardized assessment of nano-bio interactions	Defined protein mixtures simulating plasma or cellular environments	Supports consistent evaluation of bio-nano interactions across testing platforms [69]

The challenges in physico-chemical characterization of complex nanomaterials are substantial but not insurmountable. A combinatorial approach that leverages multiple complementary techniques, coupled with advanced data integration strategies, provides a path forward for comprehensive nanomaterial assessment [73]. The ongoing development of standardized protocols, reference materials, and data reporting standards is essential for building regulatory confidence in NAMs and accelerating the transition away from animal testing [71] [72] [68].

Future progress will depend on closer integration between characterization experts and toxicologists, the development of increasingly sophisticated analytical tools for complex matrices, and the adoption of FAIR (Findable, Accessible, Interoperable, and Re-usable) data principles to maximize the value of characterization data [69]. By addressing these characterization challenges systematically, the scientific community can unlock the full potential of nanomaterials while ensuring their safe and sustainable development through human-relevant testing approaches.

Establishing Standardized Protocols and Ensuring Data Reproducibility Across Labs

The adoption of New Approach Methodologies (NAMs) in ecotoxicology and regulatory toxicology represents a paradigm shift from traditional animal testing toward innovative in vitro, in silico, and in chemico methods. While these approaches offer tremendous potential for human-relevant toxicity assessment and reduced animal use, their regulatory acceptance and scientific credibility hinge on resolving critical challenges in protocol standardization and data reproducibility across laboratories. The fundamental premise of NAMs lies in their ability to provide reliable, human-relevant data for chemical safety assessment, but this potential can only be realized through standardized frameworks that ensure consistent implementation and interpretation of results [2]. Without such standardization, the promise of NAMs to accelerate chemical risk assessment while reducing animal testing remains limited.

The validation of NAMs faces a unique challenge: the historical process often involves direct comparison to animal tests that were themselves never fully validated for their reproducibility or relevance to humans [75]. This creates a circular problem where new, potentially more human-relevant methods must benchmark against traditional methods with known limitations. Growing consensus within the scientific community indicates that the validation paradigm needs to be viewed through a new lens to identify and implement superior methods in a timely manner [75]. This article examines the current state of protocol standardization and reproducibility efforts for NAMs, providing comparative analysis of emerging frameworks and practical guidance for implementation across research laboratories.

Current Challenges in NAMs Standardization

Technical and Scientific Barriers

The transition from traditional animal testing to human-relevant NAMs faces multiple technical and scientific hurdles that impact reproducibility. A primary challenge is the inadequate benchmarking of NAMs against animal data, despite well-documented limitations of rodent models, which have a poor true positive human toxicity predictivity rate of only 40-65% [2]. This benchmarking paradox creates a situation where NAMs are expected to replicate results from methods with known translational limitations, rather than being validated based on their human relevance and protective capability.

Additional technical barriers include:

• Model complexity: NAMs encompass a broad spectrum of technical approaches with varying regulatory readiness, including computational modeling, high-throughput screening, omics-based systems, organ-on-a-chip technologies, and complex in vitro models [2].
• Validation inconsistencies: The maturity of these methods differs across scientific and regulatory sectors, leading to patchy and limited adoption for discovery research and regulatory use [72].
• Inter-laboratory variability: Differences in technical expertise, reagent sources, equipment, and environmental conditions introduce variability that challenges reproducibility, particularly for complex models like organoids and microphysiological systems [76].

Regulatory and Cultural Hurdles

Beyond technical challenges, significant regulatory and cultural barriers impede standardization efforts. There remains a pervasive concern that data derived from NAMs will not find acceptance by regulatory agencies, sponsors, or the wider scientific community [2]. This perception creates inertia, with researchers maintaining familiarity and comfort with established animal methods despite their limitations. The regulatory landscape itself presents obstacles, as many current classification and labeling regulations rely on internationally harmonized guideline methods that predominantly use animal data [2]. Transitioning to risk-based approaches that embrace NAMs requires robust context-specific exposure assessment and significant advances in exposure science [2].

Emerging Standardization Frameworks and Validation Approaches

Defined Approaches and OECD Guidelines

Significant progress in standardization has been achieved through the development of Defined Approaches (DAs) – specific combinations of data sources with fixed data interpretation procedures [2]. These approaches have facilitated the use of NAMs within regulatory contexts for specific endpoints:

• Skin sensitization: OECD Test Guideline 497 outlines defined approaches for skin sensitization that combine in silico, in chemico, and/or in vitro data [2].
• Eye irritation: OECD Test Guideline 467 provides a framework for defined approaches addressing serious eye damage and eye irritation [2].

These DAs represent a crucial advancement because they package NAMs into standardized testing strategies with predetermined data interpretation procedures, reducing variability in implementation and interpretation across laboratories. The success of these approaches for specific toxicity endpoints demonstrates the feasibility of standardizing NAMs for regulatory decision-making.

Proposed Unified Validation Frameworks

A growing consensus supports the need for a unified, cross-industry approach to NAMs validation grounded in measurable quality standards and standardization [32]. Proposed frameworks emphasize:

• Clearly defined standards based on human biological relevance rather than correlation to animal data [32] [75].
• Standardized protocols with detailed methodological specifications [32].
• Transparent data sharing to build confidence in NAM performance [32] [2].
• Context-of-use determinations that align method validation with specific regulatory applications [2].

The U.S. FDA's biomarker qualification program provides one example of a detailed validation process, as demonstrated by the qualification of Stemina's devTOX quickpredict assay, which uses stem cells to predict toxicity based on metabolism [21]. Similarly, the UK government has committed to establishing a new UK Centre for the Validation of Alternative Methods (UKCVAM) that will coordinate a cross-sector network of public and private laboratories to accelerate validation and regulatory acceptance [72].

Table 1: Comparative Performance of Standardized NAMs for Specific Toxicity Endpoints

Toxicity Endpoint	Standardized Method	Validation Framework	Inter-laboratory Reproducibility	Regulatory Acceptance
Skin Sensitization	OECD TG 497 Defined Approaches	Combination of in chemico and in vitro assays with fixed interpretation procedure	High (validated across multiple laboratories)	Accepted in multiple regions
Eye Irritation	OECD TG 467 Defined Approaches	In vitro tissue models with standardized prediction models	High	Accepted in multiple regions
Developmental Toxicity	devTOX quickPredict	FDA biomarker qualification program	Moderate (limited inter-lab validation)	Qualified by FDA for specific contexts
Acute Aquatic Toxicity	ADORE database computational models	Benchmark dataset for machine learning model comparison	Variable (depends on model implementation)	Emerging acceptance

Case Studies: Reproducibility Assessments Across Laboratories

Developmental and Reproductive Toxicity (DART) Assessment

The 2025 SOT Symposium Session "Qualifying NAMs for Developmental and Reproductive Toxicity: Advancements and Pitfalls" showcased several approaches addressing standardization for complex endpoints [21]. Key case studies included:

• Stemina's devTOX quickPredict: This assay underwent the FDA's rigorous biomarker qualification program, demonstrating a structured pathway to regulatory acceptance. The process highlighted both the potential for standardization and challenges such as long wait times and communication gaps between agencies [21].
• DeTox Database: Developed at the University of North Carolina, this in silico tool uses quantitative structure-activity relationship (QSAR) models to predict developmental toxicity. While promising, the database faces standardization challenges with "activity cliffs" where structurally similar chemicals have different toxicities [21].
• Zebrafish Female Reproductive Toxicity Model: Research from Brown University demonstrated the utility of zebrafish for DART assessment but identified inconsistent methodological reporting in the literature as a significant barrier to reproducibility [21].

Next Generation Risk Assessment (NGRA) Framework

Unilever researchers presented a tiered NGRA framework for DART safety assessment that correctly identified 16/17 compounds with high DART risk using in vitro and in silico assays as the first tier [21]. This approach demonstrates how standardized testing strategies can provide protective risk assessment without replicating animal studies. The framework acknowledges that while the approach may not determine mechanisms of DART, it serves effectively for protection and risk flagging [21].

ADORE Benchmark Dataset for Ecotoxicology

The ADORE (Aquatic Toxicity Benchmark Dataset) provides an extensive, well-described dataset on acute aquatic toxicity in three taxonomic groups (fish, crustaceans, and algae) [77]. This initiative addresses standardization challenges by providing:

• A curated core dataset from the EPA ECOTOX database with standardized toxicity values and experimental conditions [77].
• Expanded features including phylogenetic, species-specific, and chemical properties [77].
• Defined data splits based on chemical occurrence and molecular scaffolds to enable standardized comparison of machine learning models [77].
• Specific challenges to the research community to test model performance on standardized tasks [77].

This benchmark dataset enables objective comparison of computational NAMs across studies by ensuring training and testing occur on the same data with identical splitting strategies, directly addressing reproducibility challenges in ecotoxicological modeling.

Experimental Protocols for Key NAMs

Defined Approaches for Skin Sensitization (OECD TG 497)

The OECD Test Guideline 497 represents a standardized protocol for assessing skin sensitization using defined approaches that integrate data from multiple NAMs [2]:

Purpose: To identify chemicals that have the potential to induce skin sensitization and characterize their potency without animal testing.

Methodology Overview:

• Data Generation: Obtain results from a minimum of three information sources selected from:
  • In chemico assays measuring peptide reactivity (e.g., DPRA)
  • In vitro assays assessing keratinocyte activation (e.g., KeratinoSens)
  • In vitro assays evaluating dendritic cell activation (e.g., h-CLAT)
• Data Interpretation: Apply a fixed prediction procedure that mathematically integrates results from the selected information sources
• Classification: Categorize chemicals as sensitisers (subcategorised by potency) or non-sensitisers based on the prediction procedure

Critical Standardization Parameters:

• Strict adherence to OECD Test Guidelines for each component assay
• Validation against known human sensitization data where available
• Qualification of laboratory proficiency through testing of reference chemicals
• Controlled environmental conditions (temperature, humidity) for cell-based assays

devTOX quickPredict Protocol for Developmental Toxicity

The devTOX quickPredict assay provides an example of a standardized in vitro protocol that has undergone regulatory qualification [21]:

Purpose: To predict compound-specific potential for developmental toxicity based on human stem cell metabolism.

Methodology Overview:

• Cell Culture: Maintain human pluripotent stem cells under defined culture conditions
• Compound Exposure: Expose cells to test compounds across a range of concentrations
• Metabolite Analysis: Measure changes in biomarker metabolites in the culture medium
• Prediction Model: Apply a qualified algorithm to convert metabolite changes into developmental toxicity potential

Standardization Requirements:

• Use of predetermined cell lines with verified characteristics
• Strict quality control for culture media and supplements
• Calibration of analytical instrumentation using reference standards
• Implementation of standardized data analysis pipelines

Research Reagent Solutions and Essential Materials

Successful implementation of standardized NAMs requires careful selection and quality control of research reagents. The following table details essential materials and their functions in standardized NAM workflows:

Table 2: Essential Research Reagents and Materials for Standardized NAM Implementation

Reagent/Material	Function	Standardization Considerations	Quality Control Metrics
Defined Cell Lines	Provide biological substrate for in vitro assays	Authentication, passage number limits, mycoplasma testing	STR profiling, viability assessments, functional competence
Specialty Culture Media	Support cell growth and maintenance	Lot-to-lot consistency, component concentration verification	pH, osmolarity, growth promotion testing
Reference Chemicals	Method calibration and proficiency assessment	Purity verification, stability monitoring	Certificate of analysis, independent verification
Antibody Panels	Cell phenotyping and endpoint measurement	Clone specificity, cross-reactivity profiling	Titration optimization, staining index quantification
Microphysiological Systems	Organ-on-chip models for complex biology	Chip lot consistency, membrane integrity	Barrier function testing, metabolic competence
Computational Platforms	In silico prediction and data integration	Algorithm version control, database currency	Prediction accuracy on reference sets, processing speed

Automation systems like the Curiox C-FREE Pluto platform address standardization challenges in sample preparation by eliminating manual centrifugation steps, which reduces variability in critical workflows such as antibody cocktailing, cell washing, and fixation [76]. Implementation of such systems has demonstrated 95% retention of CD45+ leukocytes post-lysis and consistent staining indices across immune markers, significantly improving inter-laboratory reproducibility [76].

Visualization of Standardization Workflows

NAM Validation and Qualification Pathway

The following diagram illustrates the standardized pathway for validation and qualification of new approach methodologies:

Integrated Testing Strategy Workflow

This diagram illustrates how standardized NAMs integrate into a complete testing strategy for chemical safety assessment:

Establishing standardized protocols and ensuring data reproducibility across laboratories remains both a critical challenge and essential prerequisite for widespread adoption of NAMs in ecotoxicology and regulatory decision-making. Significant progress has been made through defined approaches, benchmark datasets, and qualified methods, but considerable work remains. The path forward requires continued collaboration between researchers, regulatory agencies, and standards organizations to develop and implement robust frameworks that prioritize human relevance while ensuring reliable, reproducible results.

The scientific community must embrace a cultural shift toward transparent method reporting, data sharing, and proficiency testing to build confidence in NAMs. As highlighted throughout this analysis, initiatives like the ADORE benchmark dataset, OECD defined approaches, and the emerging UKCVAM represent concrete steps toward harmonization. By adopting these standardized frameworks and contributing to their refinement, researchers can accelerate the transition to a more human-relevant, efficient, and ethical approach to chemical safety assessment.

The field of ecotoxicology is at a pivotal juncture. With increasing regulatory pressure to reduce and eventually phase out animal testing—a commitment enshrined in EU policies like REACH and championed by the U.S. EPA—New Approach Methodologies (NAMs) have emerged as the scientific frontier for next-generation chemical safety assessment [78] [16]. These methodologies, which include in silico (computational) models, in chemico (biomolecular) assays, and in vitro (cell-based) systems, promise more human-relevant, efficient, and ethical testing paradigms [5] [16]. However, their widespread adoption faces a significant hurdle: a crisis of confidence. For regulators and scientists to trust NAMs enough to replace traditional animal studies, the data generated must be robust, reproducible, and transparent. This is where the FAIR Data Principles become not just beneficial, but critical.

The FAIR principles—making data Findable, Accessible, Interoperable, and Reusable—provide a structured framework to overcome these validation challenges [79] [80]. Originally conceived in 2016, these principles were designed to enhance the reusability of digital assets by both humans and computational systems [80] [81]. In the context of NAMs, FAIR compliance ensures that complex datasets from organ-on-a-chip models, high-throughput screening assays, and computational toxicology models are managed in a way that allows for independent verification, meta-analysis, and seamless integration—the very processes required to build the necessary scientific confidence for regulatory acceptance [79] [5].

The FAIR Principles: A Framework for Trustworthy Science

The FAIR principles establish a systematic foundation for managing scientific data. Their power lies in their focus on machine-actionability, ensuring that data can be automatically discovered and used by computational systems, which is essential for handling the volume and complexity of modern NAMs data [80] [82].

• Findable: The first step toward reuse is discoverability. Data must be easy to find for both humans and computers. This is achieved by assigning globally unique and persistent identifiers (such as Digital Object Identifiers or DOIs) and describing the data with rich, machine-readable metadata. The data must then be registered or indexed in a searchable resource [79] [80] [81].
• Accessible: Once found, data should be retrievable. The "Accessible" principle stipulates that data and metadata are obtainable using a standardized, open, and universally implementable communication protocol. Importantly, accessibility is not synonymous with being "open." Data can be restricted and protected behind authentication and authorization layers while still being FAIR, a crucial aspect for proprietary research data in the pharmaceutical industry [79] [80] [82].
• Interoperable: For data to be integrated with other datasets and analyzed across different tools and workflows, it must be interoperable. This requires using shared and broadly applicable languages for knowledge representation, such as standardized vocabularies, formal ontologies, and structured data formats. This ensures that data from a liver-on-a-chip model, for instance, can be meaningfully combined with genomic data or historical toxicological records [79] [80].
• Reusable: The ultimate goal of FAIR is to optimize the reuse of data. This demands a strong foundation of transparency and provenance. Reusability is achieved by providing rich, accurate metadata, clear usage licenses, and detailed provenance information describing the origin and processing history of the data. This allows other researchers to replicate, validate, and build upon the original findings [79] [80] [81].

Comparative Analysis: FAIR vs. Open Data in NAMs Research

A common misconception in life sciences is equating FAIR data with Open Data. Understanding their distinction is vital for effective data strategy in ecotoxicology and drug development.

Table 1: FAIR Data vs. Open Data

Aspect	FAIR Data	Open Data
Primary Goal	Optimizes data for reuse by humans and machines [80] [82]	Makes data freely available to all without restrictions [82]
Accessibility	Can be open or restricted; accessible under defined conditions [79] [82]	Always open and freely accessible [82]
Key Focus	Machine-actionability, interoperability, and rich metadata [80] [82]	Unrestricted sharing and transparency [82]
Ideal Use Case	Structured data integration in R&D; protecting sensitive or proprietary data while enabling reuse [79] [82]	Democratizing access to large public datasets; accelerating collaborative research [82]

As shown in Table 1, FAIR data is not necessarily open. A company's internal high-throughput screening data, governed by intellectual property, can be made highly FAIR for internal and collaborative research without being made public [82]. Conversely, open data might be publicly available but lack the structured metadata and standardized formats required for machine-driven integration and analysis, thus limiting its utility for building large-scale, predictive toxicological models [80].

The Validation Pipeline: How FAIR Data Principles Build Confidence in NAMs

The path from developing a NAM to its regulatory acceptance relies on a robust validation process. FAIR data principles underpin every stage of this pipeline, creating a virtuous cycle of verification and trust.

Validation Pipeline

Enhanced Reproducibility and Provenance Tracking

The "Reusable" and "Interoperable" pillars of FAIR require detailed data provenance and the use of community standards [79] [80]. For a NAM like a vascularized liver-cancer-on-a-chip model—used to test embolic agents for cancer therapy—this means that every aspect of the experiment is meticulously documented [4]. The FAIRified dataset would include the origin and passage number of the cell lines, the precise composition of the hydrogels, the flow rates of the microfluidic system, and the raw and processed analytical readouts, all described using standardized ontologies [5] [4]. This level of detail allows other laboratories to replicate the system exactly and confirm the results, a cornerstone of scientific validation.

Enabling Meta-Analysis and Cross-Domain Integration

A single NAMs dataset rarely provides sufficient evidence to replace an animal test. Confidence is built by integrating evidence from multiple sources, such as combining data from Quantitative Systems Pharmacology (QSP) models, transcriptomics, and high-throughput in vitro assays [4]. FAIR principles make this integration possible. When datasets are Findable (indexed in searchable resources) and Interoperable (using common vocabularies), they can be aggregated for powerful meta-analyses. For example, the U.S. EPA's Integrated Chemical Environment (ICE) resource leverages this approach, providing a platform where diverse NAMs data can be accessed and analyzed to support chemical safety decisions [5]. This integrated evidence base is far more compelling for regulators than a collection of isolated studies.

Table 2: Quantitative Impact of FAIR Data on Research Efficiency

Metric	Non-FAIR Data Challenge	FAIR Data Benefit	Supporting Evidence
Data Discovery	56% of scientists cite lack of standardization as a major barrier [79].	Datasets are easily discoverable by humans and machines via rich metadata and persistent identifiers [79] [80].
Time-to-Insight	Significant time lost locating, understanding, and formatting data [80].	Accelerates analysis; one study reduced gene evaluation time from weeks to days [80].	Oxford Drug Discovery Institute [80]
Research Reproducibility	High failure rate in replicating published results, especially in biomedical fields [79].	Rich metadata and provenance ensure data can be validated and replicated [79] [80].	BeginNGS coalition reduced false positive DNA differences to <1 in 50 subjects [80]
Model Accuracy	Animal models have a 5% success rate in translating to human approvals [4].	Human-relevant NAMs like organ-on-chip can achieve 80% accuracy in replicating human physiology [4].	AInvest Market Analysis [4]

Supporting Computational Modeling and AI

The future of ecotoxicology lies in Adverse Outcome Pathway (AOP)-driven, AI-enhanced risk assessment [16]. These approaches are entirely dependent on large, high-quality, machine-readable datasets. The "Interoperable" and "Reusable" nature of FAIR data makes it the perfect fuel for these advanced applications. For instance, the EPA is developing virtual tissue models and AI-driven toxicology predictions that rely on integrated FAIR data to simulate the effects of chemicals on human development and health, thereby reducing reliance on animal tests [16]. Eli Lilly's AI-powered drug discovery platform, TuneLab, trained on over $1 billion worth of proprietary data, is a prime industry example of leveraging curated data to accelerate discovery and reduce animal testing [4].

Experimental Protocols: Generating FAIR Data for NAMs

To illustrate the practical application of FAIR principles, consider a typical experiment evaluating chemical toxicity using a human liver organoid model. The following protocol ensures the resulting data is FAIR-compliant from the outset.

Detailed Methodology: Toxicity Screening Using a Human Liver Organoid Model

1. Experimental Design and Metadata Planning (Findable, Reusable)

• Objective: To assess the hepatotoxic effects of Chemical X at 3 dose levels and 3 time points, with 6 biological replicates per condition.
• Persistent Identifier: Prior to data generation, reserve a Digital Object Identifier (DOI) for the future dataset from a repository like Zenodo or Figshare.
• Metadata Schema: Define the metadata structure using a community standard such as the ISA (Investigation, Study, Assay) framework. This includes:
  • Investigation: Principal investigator, project title, funding source.
  • Study: Detailed protocol for organoid differentiation from human induced pluripotent stem cells (iPSCs), including cell line identifier and culture conditions.
  • Assay: Description of the treatment protocol, analytical methods (e.g., RNA-Seq, ATP-based viability assay), and data processing workflows.

2. Sample Preparation and Data Generation (Interoperable, Reusable)

• Biological Model: Differentiate human iPSCs (line: XYZ-001) into 3D liver organoids using a defined cytokine protocol. Document any deviations.
• Chemical Exposure: Treat organoids with Chemical X (CAS RN: XXX-XX-X) at 1, 10, and 100 µM. Include vehicle (DMSO 0.1%) and positive control (100 µM Acetaminophen) groups.
• Readouts:
  • Viability: Measure ATP content using CellTiter-Glo 3D assay.
  • Transcriptomics: Perform bulk RNA-Seq on all samples. Use controlled vocabulary (e.g., Gene Ontology terms) to annotate findings.

3. Data Processing and Curation (Interoperable, Reusable)

• Bioinformatics: Process RNA-Seq raw FASTQ files through a standardized pipeline (e.g., STAR aligner + DESeq2). Log all software versions and parameters in a workflow language like Nextflow or Snakemake.
• Data Annotation: Annotate the resulting gene expression matrix with the relevant ontology terms (e.g., Cell Ontology for cell type, UBERON for anatomy, ChEBI for the chemical).

4. Data Deposition and Publication (Findable, Accessible)

• Repository Selection: Deposit the dataset in a public, domain-specific repository such as ArrayExpress or GEO for transcriptomics data, which assigns and manages persistent identifiers.
• Data Files: Upload raw data (FASTQ), processed data (normalized counts), and the complete ISA-structured metadata.
• Access License: Apply a clear usage license (e.g., CC-BY 4.0) to the dataset to define terms of reuse.

The Scientist's Toolkit: Essential Reagents and Solutions

Table 3: Research Reagent Solutions for Liver Organoid Toxicity Assay

Item	Function in Protocol	Key Consideration for FAIRness
Human iPSC Line	Source of biologically relevant human tissue for organoid generation.	Record the specific cell line identifier and donor characteristics. Use Cell Line Ontology (CLO) terms.
Defined Differentiation Kit	Ensures consistent and reproducible generation of liver organoids.	Document the commercial catalog number, lot number, and exact protocol.
Chemical X (Reference Standard)	The test agent of unknown toxicity.	Use a certified standard with a known Chemical Abstracts Service (CAS) Registry Number.
CellTiter-Glo 3D Assay	Measures cell viability as a functional endpoint of toxicity.	Report the exact kit lot and protocol modifications for 3D cultures.
RNA-Seq Library Prep Kit	Prepares genetic material for high-throughput sequencing.	Document the kit version and all quality control metrics (e.g., RIN scores).
Standardized Bioinformatics Pipeline	Processes raw sequencing data into analyzable gene counts.	Use version-controlled, containerized software (e.g., Docker, Singularity) for full reproducibility.

Visualizing the Workflow: From FAIR Data to Regulatory Confidence

The following diagram maps the logical flow of how individual FAIR-compliant NAMs experiments contribute to a larger, integrative framework like the Adverse Outcome Pathway (AOP), ultimately leading to regulatory acceptance.

FAIR Data to AOP Workflow

The transition to an animal-free future for ecotoxicology and drug development is no longer a question of scientific possibility, but of scientific confidence. New Approach Methodologies (NAMs) are rapidly advancing, demonstrated by innovations in organ-on-chip technology, AI-driven toxicology, and complex organoid systems [4]. However, their potential cannot be fully realized within a framework of disconnected, poorly documented data. The FAIR principles provide the essential scaffolding to systematically build the required trust. By ensuring that data from these novel methods is Findable, Accessible, Interoperable, and Reusable, the scientific community can create an irrefutable, integrated, and transparent evidence base. This will empower regulators to make decisions based on human-relevant biology and accelerate the development of a safer, more ethical, and more efficient paradigm for chemical and pharmaceutical risk assessment.

The validation and regulatory acceptance of New Approach Methodologies (NAMs) in ecotoxicology and drug development represent a critical frontier in modern science. For researchers and drug development professionals, navigating the evolving regulatory landscape is paramount. Strategic engagement through official pilot programs and pre-submission meetings has become essential for successfully integrating non-animal methods into regulatory submissions. Recent legislative and policy changes have fundamentally transformed this landscape, with the FDA Modernization Act 2.0 (December 2022) removing the longstanding mandate for animal testing and explicitly authorizing cell-based assays, microphysiological systems, and sophisticated computer models as valid evidence for investigational new drug applications [10]. This foundational shift empowers sponsors to use NAMs and instructs FDA reviewers to consider them on their scientific merits, making strategic regulatory engagement both a scientific and business imperative for modern research organizations.

Available Regulatory Pilot Programs

FDA's ISTAND Program

The Innovative Science and Technology Approaches for New Drugs (ISTAND) pilot program, launched in December 2020, represents a groundbreaking pathway for establishing novel Drug Development Tools (DDTs) that fall outside existing qualification programs [10]. This program explicitly includes microphysiological systems such as Organ-Chips as qualifying technologies [10]. The first Organ-on-a-Chip was accepted into ISTAND in September 2024—a liver-chip system designed to predict drug-induced liver injury (DILI) [10]. This acceptance established a critical procedural precedent for all future microphysiological systems seeking qualification. The program's significance lies in its outcome: technologies approved through ISTAND can be included in IND and NDA applications "without needing FDA to reconsider and reconfirm its suitability" for the qualified context of use [10].

FDA's Monoclonal Antibody (mAb) Pilot Program

In April 2025, the FDA announced a targeted pilot program allowing developers of monoclonal antibody products to use primarily non-animal-based testing strategies under close FDA consultation [11] [83]. This program emerged from scientific recognition that animal testing of mAbs has proven particularly challenging due to species-specific variations in target biology and immune responses [83]. The program encourages developers to leverage computer modeling, artificial intelligence, organoids, and organ-on-a-chip systems to evaluate safety [11] [83]. For participating sponsors, the FDA aims to provide regulatory incentives such as streamlined reviews, potentially accelerating development timelines and reducing costs associated with traditional animal studies, particularly those using non-human primates [83].

Cross-Agency Coordination Initiatives

The Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM), established with congressional authorization in 2000, coordinates technical reviews of alternative test methods across federal agencies [84] [10]. ICCVAM's charter explicitly focuses on accelerating regulatory acceptance of test methods that replace, reduce, or refine animal use without compromising safety or efficacy [10]. Additionally, the NIH announced in April 2025 its intention to establish the Office of Research, Innovation, and Application (ORIVA) to coordinate NIH-wide efforts to develop, validate, and scale the use of non-animal approaches across the agency's biomedical research portfolio [84]. These coordinated initiatives provide multiple touchpoints for researchers seeking validation of their NAMs approaches.

Table 1: Key Regulatory Pilot Programs for NAMs Validation

Program Name	Lead Agency	Focus Area	Key Benefits	Current Status
ISTAND Program	FDA	Novel Drug Development Tools including Organ-Chips	Qualified context of use without need for re-review	First Organ-Chip accepted September 2024 [10]
mAb Testing Pilot	FDA	Monoclonal antibody safety assessment	Streamlined review for non-animal approaches	Launch planned within coming year [11]
ICCVAM Coordination	Multiple Agencies	Validation of alternative test methods	Cross-agency acceptance of validated methods	Ongoing since 2000, recently expanded [84] [10]
ORIVA Initiative	NIH	Biomedical research methods	Funding priority for human-based approaches	Established April 2025 [84]

Experimental Design and Methodologies for NAMs

Toxicokinetic-Toxicodynamic (TK-TD) Modeling

The General Unified Threshold models of Survival (GUTS) framework represents a robust computational approach for predicting effects under time-variable exposure scenarios in ecotoxicology [85]. The calibration and validation process for these models requires careful experimental design. Models must be calibrated on existing survival data from both acute and chronic tests under static exposure regimes, then validated against time-variable exposure profiles [85]. For example, in a study focusing on neonicotinoid insecticides and aquatic macroinvertebrates, researchers calibrated models on standard laboratory toxicity data, then performed validation experiments using two distinct approaches: one testing multiple species sensitivity to a single compound (imidacloprid), and another testing multiple compounds (imidacloprid, thiacloprid, and thiamethoxam) on a single species (Cloeon dipterum mayfly) [85]. This rigorous validation framework demonstrated acceptable prediction accuracy for four of five tested species in the multiple species dataset [85].

Organ-on-a-Chip Validation Protocols

The validation of organ-on-a-chip systems for regulatory use requires comprehensive experimental protocols that demonstrate predictive capacity. For the Liver-Chip S1 system accepted into the FDA's ISTAND program, the validation approach involved a peer-reviewed study demonstrating 87% sensitivity and 100% specificity for predicting drug-induced liver injury for a set of hepatotoxic drugs that animal models had deemed safe [10]. This head-to-head comparison against traditional animal models provided the evidentiary basis for regulatory qualification. The experimental workflow typically involves establishing human cell cultures under microfluidic conditions that replicate organ-level functions, followed by controlled exposure studies with reference compounds with known human toxicity profiles [10].

Computational Ecotoxicology and nano-QSAR

For nanomaterials and other novel chemical entities, quantitative structure-activity relationship (QSAR) models adapted for ecotoxicological applications provide powerful screening tools. A perturbation model for nano-QSAR problems was developed to simultaneously predict ecotoxicity of different nanoparticles against several assay organisms, while accounting for chemical compositions, sizes, measurement conditions, shapes, and exposure times [86]. This model, derived from a database containing 5520 cases (nanoparticle-nanoparticle pairs), demonstrated accuracies of approximately 99% in both training and prediction sets [86]. The experimental protocol involves characterizing nanoparticle properties, assembling toxicity data across multiple species and endpoints, and applying advanced machine learning algorithms to identify predictive relationships that transcend individual experimental conditions.

Table 2: Key Experimental Protocols for NAMs Validation

Methodology	Key Components	Data Requirements	Validation Metrics	Applicable Domains
TK-TD Modeling (GUTS) [85]	Calibration on acute/chronic data, time-variable exposure validation	Survival data under static and pulsed exposures	Prediction accuracy for survival under variable exposure	Aquatic ecotoxicology, insecticide risk assessment
Organ-on-a-Chip [10]	Microfluidic culture, human primary cells, functional endpoints	Reference compounds with known human toxicity	Sensitivity/specificity vs human response	Drug-induced liver injury, organ-specific toxicity
Computational nano-QSAR [86]	Nanopperty characterization, multi-species toxicity data	5520+ nanoparticle pairs, multiple experimental conditions	~99% prediction accuracy	Nanomaterial risk assessment, priority setting
High-Throughput Screening [87]	Assay development, transferability assessment, IVIVE	Concentration-response data for diverse chemicals	Reproducibility across platforms	Chemical prioritization, respiratory toxicity

Strategic Regulatory Engagement

Pre-submission Meeting Preparation

Early and frequent engagement with regulatory agencies through pre-submission meetings significantly enhances the likelihood of successful NAMs validation. Sponsors should engage FDA early (e.g., in pre-IND meetings) to discuss strategies for developing and incorporating NAMs [83]. Effective meeting packages should include comprehensive data comparing NAMs performance against conventional in vivo tests to demonstrate validity, especially when fully validated methods are not yet established [83]. Preparation should focus on clearly articulating the proposed context of use, presenting robust scientific evidence, and directly addressing potential limitations of the alternative methods. The FDA has emphasized that clear communication between the agency and sponsors alleviates uncertainty and drives more widespread utilization of these methods, leading to more rapid validation of NAMs [83].

Data Package Requirements

Regulatory submissions for NAMs require comprehensive data packages that establish scientific validity and reliability. For computational models, this includes detailed documentation of model development, training data, performance characteristics, and uncertainty quantification [85] [86]. For organ-on-a-chip systems, submissions should include evidence of system characterization, functional benchmarks, and validation against reference compounds with known human effects [10]. The FDA's roadmap encourages developers to leverage computer modeling and artificial intelligence to predict drug behavior, noting that software models could simulate how a monoclonal antibody distributes through the human body and reliably predict side effects based on distribution and molecular composition [11]. Sponsors should generate data using selected NAMs alongside conventional in vivo tests to demonstrate comparative validity during this transition period [83].

Global Regulatory Alignment

Strategic regulatory engagement must consider international regulatory landscapes, as alignment with global standards facilitates broader acceptance. The European Medicines Agency (EMA) has established a network of member states to define and implement a strategy for the reduction, refinement and replacement of animal use in drug development [83]. Similarly, the International Organization for Standardization (ISO) has revised standards for biological evaluation of medical devices (ISO 10993) to reduce animal testing by giving preference to in vitro models where these methods yield equally relevant information [84]. Understanding these global frameworks enables sponsors to develop testing strategies that meet multiple regulatory requirements simultaneously, optimizing resource allocation and accelerating international product development.

Several publicly available databases provide critical support for NAMs development and validation. The ECOTOX Knowledgebase from the U.S. Environmental Protection Agency is a comprehensive system providing chemical environmental toxicity data on aquatic and terrestrial species, compiled from over 53,000 references and including over one million test records covering more than 13,000 species and 12,000 chemicals [88]. This resource enables efficient data mining while reducing the need for animal tests [88]. The SeqAPASS (Sequence Alignment to Predict Across-Species Susceptibility) database allows researchers and regulators to extrapolate toxicity information across species, supporting read-across approaches that minimize animal testing [89]. Additionally, the FDA is collaborating with ICCVAM to create a shared central database for validated NAMs, with a beta version expected by mid-2025 [83].

Technical Implementation Tools

Successful implementation of NAMs requires specialized technical resources and platforms. High-Throughput Screening (HTS) systems form a cornerstone of this approach, enabling rapid testing of multiple substances using different cell types [87]. The Integrated Chemical Environment (ICE) provides tools and data used to predict toxicity, supporting computational toxicology approaches [89]. For specialized applications like intestinal toxicity assessment, human-relevant models incorporating advanced cell cultures enable precise evaluation of how exposures impact intestinal function [87]. Similarly, hepatic metabolism and clearance studies employ advanced in vitro models to evaluate metabolic pathways and elimination processes, utilizing proprietary and commercially available assays to deliver crucial insights into pharmacokinetics [87].

Table 3: Research Reagent Solutions for NAMs Implementation

Resource Category	Specific Tools/Platforms	Key Applications	Accessibility	Regulatory Recognition
Knowledgebases [88] [89]	ECOTOX Knowledgebase, SeqAPASS	Toxicity data mining, cross-species extrapolation	Publicly available	EPA-approved, used in risk assessment
Microphysiological Systems [10]	Organ-on-a-Chip (Liver-Chip S1)	Drug-induced liver injury prediction	Commercial platforms	FDA ISTAND acceptance [10]
Computational Modeling [85] [86]	GUTS framework, nano-QSAR models	Time-variable exposure prediction, nanomaterial risk	Open and proprietary	Peer-reviewed validation
Assay Development [87]	High-Throughput Screening, IVIVE	Chemical prioritization, in vitro to in vivo extrapolation	Commercial and custom	Used in regulatory decision-making

The strategic integration of pilot programs and pre-submission meetings represents a critical pathway for advancing New Approach Methodologies in ecotoxicology research and drug development. The rapidly evolving regulatory landscape, shaped by recent legislative changes and agency initiatives, offers unprecedented opportunities for researchers to implement human-relevant, computationally-driven approaches that reduce reliance on animal testing. Success in this new paradigm requires proactive regulatory engagement, robust experimental validation, and strategic utilization of available resources and pilot programs. As Commissioner Makary noted, this approach "marks a paradigm shift in drug evaluation and holds promise to accelerate cures and meaningful treatments for Americans while reducing animal use" [11]. For research organizations and drug developers, mastering these regulatory engagement strategies is no longer optional but essential for leadership in 21st-century toxicological science and drug development.

The transition to New Approach Methodologies (NAMs) in regulatory toxicology and ecotoxicology research represents a paradigm shift from traditional animal testing toward more human-relevant, mechanistic approaches. NAMs encompass a diverse suite of tools and technologies, including in vitro models (cell- and tissue-based systems), in silico models (computational tools and AI), microphysiological systems (organ-on-a-chip), omics technologies, and adverse outcome pathways (AOPs) [22]. Despite significant advancements, the widespread adoption of NAMs for regulatory decision-making faces substantial technical hurdles, particularly in the areas of dosimetry, sample preparation, and reference materials [2] [32].

The core challenge lies in establishing scientific confidence that these new methods provide information of equivalent or better quality and relevance for regulatory decision-making compared to traditional animal tests [90]. This is especially complex for systemic toxicities resulting from chronic exposure or involving multiple mechanisms, where the relationship between exposure concentration and biological effect is not straightforward [2]. Successfully addressing these technical gaps requires a unified framework for validation grounded in measurable quality standards and standardization to ensure reliable, reproducible results across different laboratories and platforms [32].

Critical Technical Gaps in NAMs Implementation

The Dosimetry Challenge in ComplexIn VitroSystems

Dosimetry in NAMs extends beyond simple concentration measurements to encompass the biologically effective dose that reaches the molecular target within complex in vitro systems. This is particularly challenging for repeated-dose and chronic toxicity assessments, where ensuring consistent, relevant exposure concentrations over time is critical for generating meaningful data [2]. The limitations of current approaches become especially apparent when modeling systemic toxicities, as NAMs may not fully mimic every aspect of human-relevant acute or chronic exposure, even in sophisticated physiologically-based models [2].

Microphysiological systems (MPS), such as organ-on-chip platforms, introduce additional dosimetry complexities through fluid flow, tissue-tissue interfaces, and mechanical forces that influence compound distribution and metabolism [22]. These systems must accurately reflect human pharmacokinetic and pharmacodynamic processes, including species-relevant metabolite formation, which may differ from those observed in traditional animal models or simple static cell cultures [90]. Furthermore, accounting for inter-individual variability in human populations requires sophisticated study designs that incorporate genetic diversity into cell-based test systems, adding another layer of complexity to dosimetry considerations [90].

Sample Preparation and Standardization Hurdles

Sample preparation for NAMs lacks standardized protocols across different testing platforms, leading to variability in results that complicates inter-laboratory comparisons and method validation. The absence of standardized procedures is particularly problematic for emerging technologies like organ-on-chip systems and complex 3D models, where differences in cell sourcing, culture conditions, and media composition can significantly impact system performance and experimental outcomes [91]. This standardization gap presents a substantial barrier to regulatory acceptance, as regulators require assurance that NAM-derived data are robust, reproducible, and reliable for safety decision-making [32] [91].

The validation process itself faces a fundamental tension: while NAMs are not intended to simply recapitulate animal tests without animals, they are often benchmarked against animal data that may have poor human toxicity predictivity rates of only 40-65% [2]. This creates a circular problem where human-relevant methods are evaluated against potentially flawed animal data, complicating the establishment of human-relevant positive and negative controls. Additionally, the sourcing and characterization of biological samples—including primary cells, stem cell-derived tissues, and engineered constructs—introduce further variability that must be controlled through standardized sample preparation protocols and comprehensive system characterization [90].

Reference Material Deficiencies

The development and characterization of qualified reference materials for NAMs lag behind method development, creating a critical gap in validation workflows. Unlike traditional animal studies, which have established historical control databases and standardized animal strains, many NAMs lack well-characterized reference compounds and control materials with demonstrated performance characteristics across different laboratories [90] [91]. This deficiency is particularly acute for complex endpoints such as developmental and reproductive toxicity (DART), where multiple alternative assays within a tiered or battery approach must provide a level of confidence for human safety assurance at least equivalent to current testing paradigms [92].

The problem extends to technical reproducibility as well. Current organ-on-chip technologies face significant standardization challenges, with different laboratories potentially using slightly different cell lines, flow parameters, or structural configurations that affect standardized readouts [91]. Without universal reference materials and standardized protocols, it becomes difficult to determine whether variability in results stems from true biological differences or technical artifacts. This challenge is recognized by initiatives like the Foundation for the National Institutes of Health (FNIH) Validation Qualification Network, which aims to define shared data elements and unify reporting practices across preclinical, clinical, and safety testing [91].

Table 1: Technical Gaps and Current Limitations in NAMs Implementation

Technical Area	Specific Challenges	Impact on NAMs Validation
Dosimetry	- Determining biologically effective dose in complex systems- Accounting for metabolite formation- Modeling chronic exposure scenarios- Incorporating population variability	Affects accuracy of point-of-departure estimation and extrapolation to human exposure scenarios
Sample Preparation	- Lack of standardized protocols across platforms- Variability in cell sourcing and culture conditions- Inconsistent media composition and supplement use- Differences in system assembly and maintenance	Leads to inter-laboratory variability that undermines reproducibility and regulatory confidence
Reference Materials	- Limited qualified reference compounds- Lack of universal controls for complex endpoints- Insufficient characterization of technical reproducibility- Absence of standardized positive/negative controls	Hinders method benchmarking and validation across different testing platforms and laboratories

Comparative Performance of NAMs Versus Traditional Methods

Established Successes for Defined Endpoints

NAMs have demonstrated significant success for specific, well-defined toxicity endpoints where mechanisms are relatively straightforward and human relevance can be clearly established. For skin sensitization, a combination of human-based in vitro approaches has shown similar performance to the traditionally used Local Lymph Node Assay (LLNA) in mice, with some defined approaches (specific combinations of data sources with fixed data interpretation procedures) even outperforming the LLNA in terms of specificity [2]. These successes have been formally recognized through the development of OECD test guidelines for defined approaches for serious eye damage/eye irritation (OECD TG 467) and skin sensitization (OECD TG 497), which are now widely used in regulations worldwide [2].

Further validation comes from case studies with specific compounds. For crop protection products Captan and Folpet, a multiple NAM testing strategy incorporating 18 in vitro studies—including eye and skin irritation and skin sensitization assays compliant with OECD test guidelines—appropriately identified these chemicals as contact irritants, demonstrating that suitable risk assessments could be performed with available NAM tests broadly aligned with risk assessments conducted using existing mammalian test data [2]. These examples illustrate that for local toxicity endpoints with direct mechanistic links to human biology, NAMs can provide equivalent or superior protection compared to traditional animal methods.

Performance Gaps for Complex Endpoints

While NAMs show promise for complex toxicity endpoints such as carcinogenicity and developmental and reproductive toxicity (DART), significant performance gaps remain compared to traditional methods for these endpoints. The predictive power of NAMs has greatly improved in the last five to ten years, but animal models are still considered the gold standard for many drug-induced adverse events, particularly for small molecules with complex metabolism and distribution patterns [27]. This performance gap is most evident for systemic toxicities resulting from repeated exposure or involving multiple organ systems, where the complexity of organism-level responses is difficult to capture in current in vitro systems.

The table below summarizes comparative performance data for NAMs versus traditional animal methods across different toxicity endpoints:

Table 2: Comparative Performance of NAMs vs. Traditional Animal Methods

Toxicity Endpoint	Traditional Method	NAM Alternatives	Comparative Performance
Skin Sensitization	Local Lymph Node Assay (LLNA) in mice	Defined Approaches (DPRA, KeratinoSens, h-CLAT)	Combination of human-based in vitro approaches outperforms LLNA in specificity [2]
Eye Irritation	Draize rabbit eye test	OECD TG 467 (Defined Approaches)	Validated for regulatory use with equivalent protection [2]
Cardiovascular Toxicity	In vivo cardiovascular safety pharmacology	In silico proarrhythmia risk prediction models	Appropriately qualified models can assess torsades de pointes risk according to context of use [92]
Developmental Toxicity	In vivo DART studies in two species	Alternative assays (in vitro, ex vivo, nonmammalian)	Multiple alternative assays in tiered approach can provide equivalent confidence for human safety [92]
Systemic Toxicity (Repeat Dose)	90-day rodent toxicity study	Microphysiological systems (organ-on-chip)	Not yet capable of fully replacing animal studies for complex, multi-organ effects [2] [27]
Carcinogenicity	2-year rodent bioassay	Weight-of-evidence approaches, transgenic mouse models	Product-specific assessment recommended for biologics; WoE may inform need for rat study [92]

Hybrid Approaches for Transitional Implementation

Given the current performance gaps for complex endpoints, a hybrid approach combining alternative and animal-based testing models is emerging as a practical strategy during this transitional period [27]. This method involves running in vitro tests in parallel with standard animal tests to demonstrate that NAMs are at least as good as—if not better than—traditional methods at predicting adverse outcomes in humans. Such side-by-side comparisons are crucial for building confidence in NAMs among researchers, regulators, and other stakeholders, particularly for novel drug targets or mechanisms where historical data is limited [27].

The hybrid approach also facilitates the stepwise implementation recommended in the FDA's roadmap for reducing animal testing, which suggests starting with monoclonal antibodies as a promising area for reducing animal use in preclinical safety testing before expanding to other biological molecules and eventually new chemical entities [91]. This cautious, evidence-based transition acknowledges both the potential of NAMs and their current limitations, particularly for small molecule development where comprehensive accounting of metabolites and multi-organ interactions presents significant technical challenges [27].

Experimental Protocols for Method Validation

Protocol for Dosimetry Assessment in Microphysiological Systems

Objective: To quantify nominal versus free concentration relationships and metabolic stability in microphysiological systems (MPS) for accurate dosimetry assessment.

Materials:

• Physiologically relevant cell culture models (e.g., primary cells, iPSC-derived cells, or appropriate cell lines)
• Microphysiological system platform (organ-on-chip or spheroid culture)
• Test articles including positive control compounds and vehicle controls
• Analytical instrumentation for concentration verification (LC-MS/MS, HPLC)
• Biomarkers for functional assessment (e.g., albumin for hepatocytes, TEER for barrier models)

Methodology:

• System Characterization: Pre-condition MPS for 24-48 hours to establish stable phenotypes and functionality. Confirm tissue-specific functions using relevant biomarkers.
• Dosing Strategy: Apply test articles across a physiologically relevant concentration range, including vehicle controls. Use both single and repeated dosing regimens as appropriate for the exposure scenario.
• Concentration Verification: Collect media samples at multiple timepoints (e.g., 0, 1, 6, 24 hours) for analytical quantification of parent compound and major metabolites.
• Mass Balance Assessment: Determine recovery of administered dose in media, cells/tissues, and system components to account for non-specific binding.
• Temporal Profile Analysis: Model concentration-time relationships to determine key pharmacokinetic parameters (e.g., C~max~, T~max~, AUC, half-life).
• Functional Assessment: Correlate compound exposure with functional responses using appropriate endpoint measurements.
• Data Normalization: Express results as both nominal concentrations and free concentrations, noting relationship between the two.

Validation Criteria: System demonstrates <20% coefficient of variation in control compound responses across three independent experiments. Free concentration remains >80% of nominal concentration at all measured timepoints, or relationship is well-characterized.

Protocol for Cross-Laboratory Sample Preparation Standardization

Objective: To establish standardized sample preparation protocols that ensure reproducibility of NAMs across different laboratories and platforms.

Materials:

• Standardized reference materials with certificate of analysis
• Qualified cell sources with documented characterization
• Defined culture media and supplements from single qualified vendor
• Standard operating procedures for all critical processes
• Quality control assays for system performance assessment

Methodology:

• Cell Sourcing and Qualification: Source cells from qualified repositories with comprehensive characterization data. Perform additional qualification assays to confirm phenotype and functionality.
• Culture Protocol Harmonization: Establish standardized protocols for cell expansion, passage procedures, and seeding densities with defined acceptance criteria.
• Differentiation Standardization: For stem cell-derived systems, implement standardized differentiation protocols with quality control checkpoints.
• Sample Processing Uniformity: Define fixed procedures for compound dosing, media changes, and sample collection timepoints.
• Reference Compound Testing: Include standardized reference compounds in each experiment to control for system performance.
• Blinded Testing: Implement cross-laboratory testing with blinded samples to assess reproducibility.
• Data Collection Standardization: Use harmonized data collection templates and metadata documentation.

Validation Criteria: <30% coefficient of variation in reference compound responses across participating laboratories. Successful blinded sample classification with >80% accuracy across all testing sites.

Essential Research Reagent Solutions for NAMs

Successful implementation of NAMs in ecotoxicology research and regulatory testing requires access to well-characterized research reagents and materials. The following table details key solutions necessary for addressing technical gaps in dosimetry, sample preparation, and reference materials:

Table 3: Essential Research Reagent Solutions for NAMs Implementation

Reagent Category	Specific Examples	Function in NAMs	Technical Considerations
Qualified Cell Sources	Primary hepatocytes, iPSC-derived cells, tissue-specific progenitors	Provide biologically relevant models for toxicity assessment	Must demonstrate tissue-specific functionality; require comprehensive characterization [90] [22]
Reference Compounds	Prototypical toxicants with known mechanisms (e.g., acetaminophen for hepatotoxicity)	Method calibration and cross-laboratory comparison	Require certificate of analysis with purity confirmation; should span multiple mechanism classes [90]
Extracellular Matrix Components	Defined hydrogels, basement membrane extracts, synthetic scaffolds	Provide physiological context for 3D culture systems	Batch-to-batch variability must be minimized; composition should be well-characterized [22] [27]
Specialized Media Formulations	Serum-free defined media, tissue-specific supplements	Support phenotypic maintenance in complex systems	Require exact formulation documentation; essential for reproducibility [91]
Metabolite Standards	Stable isotope-labeled metabolites, synthetic metabolite analogs	Dosimetry assessment and metabolic capability verification	Critical for confirming metabolic competence; enable accurate quantification [2] [90]
Performance Assay Kits	Functional assessment kits (albumin, urea, CYP activity)	System qualification and functionality assessment	Must be validated for specific cell types; provide quantitative readouts [90]
Quality Control Biomarkers	Transcriptomic signatures, proteomic markers, functional endpoints	Batch-to-batch consistency assessment	Should represent key biological processes; enable system qualification [90] [22]

Visualizing NAMs Validation Workflows and Relationships

NAMs Validation Pathway Diagram

Dosimetry Considerations in Microphysiological Systems

Addressing the technical gaps in dosimetry, sample preparation, and reference materials represents the critical path toward robust validation and regulatory acceptance of New Approach Methodologies. The current state of NAMs reflects a transitional period where these methods show significant promise—particularly for defined endpoints with clear mechanistic foundations—but still face substantial challenges for complex, systemic toxicity assessments [2] [27]. Successfully navigating this transition will require coordinated efforts across multiple stakeholders, including researchers, regulatory agencies, standards organizations, and material suppliers.

The establishment of a unified framework for validation and regulatory acceptance, grounded in measurable quality standards and standardization, represents the most pressing need in the field [32]. Initiatives like the FDA's roadmap for reducing animal testing, the FNIH's Validation Qualification Network, and international harmonization efforts provide the structural foundation for this framework [91]. By systematically addressing the technical gaps in dosimetry, sample preparation, and reference materials through collaborative research and standardization, the scientific community can accelerate the adoption of NAMs that offer more human-relevant, mechanistic, and predictive approaches to safety assessment—ultimately benefiting human health, environmental protection, and scientific progress.

Building Confidence: Validation Frameworks and Comparative Performance of NAMs

Blueprint for a Unified Cross-Industry Validation Framework for NAMs

The transition from traditional animal testing to human-relevant New Approach Methodologies (NAMs) represents a paradigm shift in toxicology and ecotoxicology. This guide provides a comparative analysis of the evolving validation frameworks designed to establish scientific confidence in these non-animal methods. As regulatory agencies worldwide increasingly call for NAMs to streamline chemical hazard assessment [93], the development of a unified, cross-industry validation framework has become critical. We examine the core components of modern validation, present quantitative comparisons with traditional approaches, and provide detailed experimental protocols to guide researchers and drug development professionals in navigating this rapidly advancing field. The move toward a fit-for-purpose validation paradigm promises to accelerate the integration of human-relevant data into safety assessments while supporting the 3Rs principles (Replace, Reduce, Refine) in toxicological research [5].

New Approach Methodologies encompass any technology, methodology, approach, or combination that can be used to replace, reduce, or refine animal toxicity testing while enabling more rapid or effective prioritization and assessment of chemicals [93]. These include in silico (computational), in chemico (abiotic measures of chemical reactivity), and in vitro (cell-based) assays, as well as non-animal testing methods employing omics technologies or non-protected taxonomic groups [93] [5]. The fundamental driver for NAMs adoption extends beyond ethical considerations to their potential for providing more human-relevant, efficient, and mechanistic toxicity data compared to traditional animal models [32] [24].

The validation of NAMs presents a significant challenge, as existing regulatory frameworks were predominantly designed for animal test methods. A 2024 analysis highlights the pressing need for a unified, cross-industry approach to NAMs validation, grounded in measurable quality standards and standardization [32]. This blueprint examines the emerging frameworks aimed at addressing this need, comparing their components against traditional validation paradigms and providing experimental guidance for implementation.

Comparative Framework Analysis: Traditional vs. Modern Approaches

Evolution of Validation Paradigms

The validation of toxicological test methods has traditionally been governed by principles established in the Organisation for Economic Co-operation and Development (OECD) Guidance Document 34 (2005), which emphasized reliability and relevance for a defined purpose [24]. However, this framework has increasingly been recognized as inflexible for assessing NAMs, often requiring lengthy, expensive inter-laboratory ring trials and favoring comparison to animal data over human biological relevance [24].

Modern frameworks propose a more flexible, fit-for-purpose approach that recognizes NAMs may provide different but more human-relevant information than traditional animal tests [24]. The emerging consensus, articulated by the Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM), focuses on establishing scientific confidence through multiple elements tailored to the specific context of use (CoU) [94] [24].

Table 1: Comparative Analysis of Traditional vs. Modern Validation Paradigms

Validation Component	Traditional Paradigm	Modern NAMs Framework
Primary Validation Standard	Concordance with animal test data	Human biological relevance and fitness for purpose [24]
Validation Process	Rigid, sequential process often requiring ring trials	Flexible, modular approach based on context of use [24]
Biological Relevance Assessment	Assumed for animal models	Focused on human biology and mechanistic understanding [24]
Reference Standards	Animal response data	Human-relevant mechanistic data; historical animal variability used for benchmarking [24]
Implementation Timeline	Often years to decades	Accelerated through early regulator engagement [94]
Data Requirements	Fixed based on regulatory statutes	Adaptable based on defined context of use [94]

The Core Components of Modern NAMs Validation

Current scientific consensus identifies five to six essential elements for establishing scientific confidence in NAMs. These elements provide a structured approach to validation that accommodates the diversity of NAMs technologies while maintaining scientific rigor.

Table 2: Core Components of a Modern NAMs Validation Framework

Framework Component	Description	Application in Regulatory Context
Defined Context of Use (CoU)	Clear specification of the purpose and limitations of the NAM application [94] [24]	Determines validation stringency; pre-regulatory uses may require less validation than quantitative risk assessment [94]
Biological Relevance	Assessment of alignment with human biology and mechanistic understanding [24]	Focuses on human relevance rather than concordance with animal data [24]
Technical Characterization	Evaluation of reliability, reproducibility, and robustness [24]	Includes intra- and inter-laboratory reproducibility assessments [24]
Data Integrity	Implementation of quality assurance measures for data generation and handling [94]	Ensures transparency and traceability of data and methods [94]
Information Transparency	Complete reporting of methods, data, and limitations [94]	Enables independent assessment and builds regulatory confidence [94]
Independent Review	Critical assessment by external experts and regulatory bodies [94] [24]	Provides objective evaluation of fitness for defined purpose [24]

Experimental Protocols and Methodologies

Establishing Technical Characterization

Protocol Title: Assessing Intra-laboratory and Inter-laboratory Reproducibility for NAMs

Purpose: To quantitatively measure the reliability and reproducibility of a New Approach Methodology across multiple testing conditions and laboratories.

Materials and Equipment:

• Standardized test method protocol
• Reference chemicals with known responses (covering full range of expected results)
• Qualified personnel across multiple laboratories
• Appropriate cell culture systems, analytical instruments, or computational platforms specific to the NAM

Procedure:

• Pre-validation Phase: Develop standardized protocol with detailed specifications for all critical procedures and acceptance criteria.
• Reference Chemical Selection: Curate a set of 10-15 reference chemicals representing strong, weak, and negative responses for the endpoint of interest [24].
• Intra-laboratory Testing: Have at least two qualified operators within the same laboratory test the reference chemical set following the standardized protocol at different times (e.g., different days or weeks).
• Inter-laboratory Testing: Select a minimum of 3-4 independent laboratories to test the same reference chemical set using the standardized protocol.
• Data Collection: Collect quantitative results using predetermined metrics and scoring systems.
• Statistical Analysis: Calculate intra-laboratory repeatability and inter-laboratory reproducibility using appropriate statistical measures (e.g., coefficient of variation, concordance correlation coefficient).

Validation Criteria: A NAM demonstrates sufficient reliability when intra-laboratory results show >80% repeatability and inter-laboratory results show >70% reproducibility for quantitative endpoints, or appropriate statistical equivalence for qualitative assessments [24].

Assessing Biological Relevance

Protocol Title: Establishing Human Biological Relevance for NAMs

Purpose: To demonstrate the alignment of the NAM with human biology and its relevance for predicting human toxicological responses.

Materials and Equipment:

• Human-derived cells, tissues, or biochemical systems
• Relevant positive and negative control substances
• Appropriate analytical platforms for measuring mechanistic endpoints
• Comparative animal data (where available for context, not as gold standard)

Procedure:

• Mechanistic Basis Definition: Identify and document the key biological events and mechanisms the NAM is designed to measure.
• Human Biological Alignment: Demonstrate that the NAM captures aspects of human biology through:
  • Use of human cells, tissues, or proteins
  • Expression of relevant human biomolecules (receptors, enzymes)
  • Appropriate metabolic capacity where relevant
• Pathway Activation Testing: Expose the NAM to reference chemicals with known mechanisms of action and measure expected pathway responses.
• Performance Assessment: Compare NAM responses to available human data (e.g., clinical observations, epidemiological data, human in vitro studies) where possible.
• Context Establishment: Document how the NAM provides information useful for health-protective decisions, even without direct comparison to animal data.

Validation Criteria: Biological relevance is established when the NAM demonstrates: (1) alignment with known human biological processes; (2) appropriate responses to mechanistically-acted chemicals; and (3) capability to provide information useful for human health risk assessment [24].

Visualization of the Validation Framework

The following diagram illustrates the integrated relationship between the core components of the modern NAMs validation framework and the pathway to regulatory acceptance:

Framework Workflow Relationship

This workflow visualization demonstrates how the validation components interact sequentially, with each element building upon the previous to establish comprehensive scientific confidence.

Research Reagent Solutions Toolkit

Successful implementation of NAMs requires specific research tools and platforms. The following table details essential reagents and resources for developing and validating NAMs:

Table 3: Essential Research Reagent Solutions for NAMs Development

Tool/Reagent Category	Specific Examples	Function in NAMs Development
In Vitro Model Systems	Organ-on-chip devices, 3D organoids, primary human cells [5] [91]	Provide human-relevant tissue models that mimic organ-level responses for toxicity assessment
Computational Toxicology Tools	CompTox Chemicals Dashboard, ToxCast, httk R package [95]	Enable chemical prioritization, exposure prediction, and high-throughput toxicity screening
Ecotoxicology Resources	SeqAPASS, ECOTOX Knowledgebase, Web-ICE [95]	Support species extrapolation and ecological risk assessment without animal testing
Data Integration Platforms	Integrated Chemical Environment (ICE), invitroDB [5] [95]	Centralize NAMs data for comparative analysis and method evaluation
Reference Chemical Sets	Curated chemical libraries with known toxicity profiles [24]	Provide benchmark substances for validating NAMs performance and reproducibility
Quality Control Materials	Standard operating procedures, positive/negative controls, proficiency substances [24]	Ensure technical robustness and inter-laboratory reproducibility of NAMs

Case Studies and Regulatory Progress

Successful Implementations Across Industries

Several NAMs have successfully navigated the validation process and gained regulatory acceptance for specific applications:

• Skin Sensitization: Multiple non-animal testing strategies incorporating in vitro, in chemico, and in silico inputs have demonstrated equivalent or superior performance to the in vivo model when compared to both animal and human data [94]. The OECD Guideline 497 for Defined Approaches for Skin Sensitization represents a successful example of regulatory adoption [24].

• Endocrine Disruption: Certain NAMs have been validated and accepted by the EPA as alternatives for certain Tier 1 endocrine disruptor assays, while others serve for prioritization and as scientifically relevant information in weight-of-evidence evaluations [94].

• Inhalation Toxicology: Human lung complex 3D models combined with computational modeling have been used to assess hazard for human occupational exposure by the inhalation route, as demonstrated in an OECD IATA Case Study [94].

Emerging Regulatory Frameworks

Recent developments indicate significant momentum toward regulatory acceptance of NAMs:

• The FDA Modernization Act 2.0 (December 2022) replaced the word "animal" with "non-clinical" in relevant text, officially recognizing in vitro, in silico, and in chemico approaches as valid for regulatory submissions [94] [91].

• The FDA Roadmap to Reducing Animal Testing (April 2024) outlines a strategic, stepwise approach for implementing validated NAMs in preclinical safety studies [91].

• The Complement-ARIE program by NIH aims to speed the development, standardization, validation, and use of human-based NAMs through a consortium of researchers and dedicated funding [5] [91].

The following diagram illustrates the evolving pathway from traditional validation to modern NAMs acceptance:

Validation Paradigm Shift

The blueprint for a unified cross-industry validation framework for NAMs represents a fundamental shift in how toxicological methods are evaluated and accepted. By moving from rigid, animal-centric validation to flexible, fit-for-purpose approaches focused on human biological relevance, the scientific community can accelerate the adoption of more predictive, efficient safety assessment tools.

Critical success factors for widespread implementation include:

• Early engagement between NAM developers and regulatory agencies to align on context of use and validation requirements [94]
• International harmonization of standards and acceptance criteria to facilitate global chemical assessment [32] [91]
• Transparent data sharing through resources like the NAMs Data Hub to build collective evidence and confidence [91]
• Education and training for researchers and assessors on appropriate development and application of NAMs [95]

As the validation framework continues to evolve, the focus must remain on establishing scientific confidence through rigorous, relevant assessment of NAMs' capabilities rather than demanding alignment with historical animal data. This approach promises to transform chemical safety assessment while advancing both human health protection and the principles of ethical science.

The paradigm for safety and efficacy testing in biomedical and ecotoxicological research is undergoing a fundamental transformation. This shift from traditional animal models to New Approach Methodologies (NAMs) is driven by persistent concerns over the predictive accuracy of animal studies for human outcomes and evolving ethical standards. This comparative analysis examines the scientific evidence supporting NAMs—encompassing advanced in vitro systems, microphysiological models, and computational approaches—against conventional animal testing, with a specific focus on their performance in predicting human responses. The integration of these human biology-based approaches is reshaping preclinical research and regulatory frameworks worldwide, supported by legislative changes such as the FDA Modernization Act 2.0 [96] [97].

Quantitative Comparison of Predictive Accuracy

Extensive validation studies have quantified the performance differences between NAMs and traditional animal models across multiple testing domains. The data consistently demonstrate the enhanced predictive capability of human biology-based systems.

Table 1: Predictive Accuracy Comparison Across Testing Domains

Testing Domain	Traditional Animal Models	New Approach Methodologies (NAMs)	Performance Gap
Skin Allergy	72-74% accuracy (guinea pigs, mice) [19]	85% accuracy (combined chemistry- and cell-based methods) [19]	+11-13%
Skin Irritation	60% accuracy (Draize rabbit test) [19]	86% accuracy (reconstituted human skin models) [19]	+26%
Developmental Toxicity	60% sensitivity (animal tests) [19]	93% sensitivity (human stem cell tests) [19]	+33%
Drug-Induced Liver Injury (DILI)	Frequently fails to detect human-specific hepatotoxicity [96]	87% identification of hepatotoxic drugs (Liver-Chip model) [97]	Significant improvement
Cardiotoxicity	Limited prediction of human arrhythmia responses [97]	Human stem cell-derived cardiomyocytes successfully flag arrhythmia risks [97]	Enhanced mechanistic prediction

The economic implications of these predictive failures are substantial. The chronically high attrition rate of new drug candidates traces back to two major factors: lack of efficacy in human biology and human-specific safety issues, both reflecting the poor predictability of traditional preclinical models [97]. With the likelihood of approval for compounds entering Phase 1 trials at just 6.7% as of early 2025—down from 10% a decade prior—the limitations of current testing paradigms have become economically unsustainable [96].

Fundamental Limitations of Traditional Animal Models

Interspecies Disconnect

The fundamental challenge with animal models stems from interspecies differences in drug metabolism, immune response, and transporter expression that often lead to poor translatability of safety findings [96]. These biological differences manifest in concrete failures:

• TGN1412 Case Study: A monoclonal antibody caused life-threatening cytokine release syndrome in human volunteers despite appearing safe in preclinical monkey studies [97].
• Smoking and Cancer: Human population studies revealed that smoking causes cancer, while mice and rats do not develop cancer from smoking [19].
• Species-Specific Toxic Responses: Animal studies cannot adequately model patient-specific variables such as genetic polymorphisms, comorbidities, or environmental exposures that drive human responses [96].

Economic and Temporal Inefficiencies

Traditional animal testing imposes significant economic and temporal burdens on the drug development process:

• Monoclonal Antibody Programs: A typical program consumes approximately 144 monkeys at a cost of up to $50,000 each, totaling $7 million in animal expenses before clearing early safety gates [97].
• Development Timeline: Conventional toxicology studies represent multi-month processes involving drug administration to animals followed by tissue dissection to understand effects [91].

New Approach Methodologies: Scientific Basis and Experimental Protocols

Microphysiological Systems (Organs-on-Chips)

Microphysiological systems (MPS) replicate organ-level functionality in vitro through perfused, physiologically relevant architectures that maintain human-specific biological responses [98].

Table 2: Experimental Protocol for Organ-on-Chip Validation

Protocol Component	Technical Specifications	Validation Metrics
Cell Sourcing	Human induced pluripotent stem cells (iPSCs) or primary tissue-derived cells [98]	Donor-to-donor reproducibility, differentiation markers
Platform Architecture	2D, 2D+, or 3D configurations with physiological flow rates [98]	Barrier function integrity, tissue organization
Functional Assessment	Contractility, membrane action potential, calcium handling (cardiac); albumin production, urea synthesis (liver) [97]	Comparison to clinical data from human tissues
Exposure Conditions	Physiological relevant dosing, including metabolite exposure [96]	Pharmacokinetic modeling integration
Endpoint Analysis	High-content imaging, multi-omics analyses, functional readouts [98]	Correlation with known human toxicants

Experimental Workflow: The Emulate Liver-Chip was validated through a rigorous process comparing its predictions to known human outcomes for hepatotoxic drugs. The model correctly identified 87% of hepatotoxic drugs that caused liver injury in patients, leading to its acceptance into the FDA's Innovative Science and Technology Approaches for New Drugs (ISTAND) pilot program [97].

Stem Cell-Based Disease Modeling

Human induced pluripotent stem cells (iPSCs) enable patient-specific toxicology assessment through differentiation into target tissues while retaining donor genetic information [98].

Diagram: iPSC-based Predictive Toxicology Workflow

Experimental Protocol - Chemotherapy Cardiotoxicity Modeling:

• iPSC Generation: Blood samples from patients experiencing doxorubicin-induced cardiotoxicity and matched controls are reprogrammed to iPSCs using chemically defined, cost-effective methodologies (~$500 per line) [98].
• Cardiac Differentiation: iPSCs are differentiated into cardiomyocytes using established protocols, generating tissues with physiological cell composition (cardiomyocytes and cardiac fibroblasts in 3:1 ratio) [97].
• Compound Exposure: Differentiated cardiomyocytes are exposed to doxorubicin at clinically relevant concentrations.
• Endpoint Assessment: Functional measurements include contractility, reactive oxygen species production, DNA damage, calcium handling, and mitochondrial dysfunction.
• Genetic Validation: Genome-wide association studies identify genetic variants (e.g., RARG variant), followed by SNP correction in iPSCs to reverse hypersensitivity, confirming mechanistic basis [98].

Computational and In Silico Approaches

Computational models leverage existing biological data to predict toxicity without additional animal testing:

• Quantitative Structure-Activity Relationships (QSAR): Mathematical models linking chemical structure to biological activity [99].
• Physiologically Based Pharmacokinetic (PBPK) Modeling: Simulates absorption, distribution, metabolism, and excretion of compounds in humans [97].
• Data Mining Tools: Predict hazard of substances based on existing data from similar compounds [19].

The Scientist's Toolkit: Essential Research Reagent Solutions

Table 3: Key Research Reagents for Implementing NAMs

Reagent Category	Specific Examples	Research Application
Stem Cell Resources	Human induced pluripotent stem cells (iPSCs) [98]	Patient-specific disease modeling, toxicology screening
Differentiation Kits	Cardiomyocyte differentiation kits, hepatocyte differentiation media [98]	Generating target tissues for organ-specific toxicity assessment
Extracellular Matrix	Matrigel, collagen, fibrin-based hydrogels [98]	3D tissue scaffolding that mimics native extracellular environment
Microphysiological Systems	Organ-on-chip platforms (Liver-Chip, Heart-on-Chip) [97]	Emulating human organ level functionality and inter-tissue crosstalk
Biosensing Platforms	TEER electrodes, multiparametric metabolic sensors, microelectrode arrays [100]	Real-time functional monitoring of tissue responses
Specialized Media	Serum-free organotypic media, metabolic induction supplements [98]	Maintaining tissue-specific phenotypes and functions
Viability/Cytotoxicity Assays	MTT, resazurin, ATP-based luminescence assays [100]	High-throughput screening of compound toxicity
Functional Assay Kits	Calcium handling dyes, contractility sensors, albumin ELISA [97]	Quantifying tissue-specific functional endpoints

Regulatory Evolution and Implementation Framework

Global regulatory agencies are establishing structured pathways for integrating NAMs into safety assessment:

• FDA Modernization Act 2.0: Explicitly authorized use of non-animal alternatives to support investigational new drug applications [97].
• FDA Pilot Programs: Encouraging sponsors to submit NAM data parallel with animal data to build experience repositories [97].
• Fit-for-Purpose Validation: Assessing NAMs based on utility in specific decision contexts rather than requiring wholesale replacement of animal models [96].
• International Harmonization: Collaboration between U.S. FDA, European Commission, Japanese regulators, and Indian ministries to align validation standards [91].

The validation framework requires multi-site reproducibility studies, standardized protocols, and demonstrated predictive capacity for human outcomes [97]. The Foundation for the National Institutes of Health (FNIH) is establishing a Validation Qualification Network (VQN) to define shared data elements and unify reporting practices across preclinical, clinical, and safety testing [91].

The comparative evidence demonstrates that human biology-based NAMs consistently outperform traditional animal models in predicting human responses across multiple testing domains. The scientific and economic case for transitioning to these approaches is compelling, with improvements in predictive accuracy ranging from 11% to 33% across key endpoints [19].

The future of predictive toxicology lies in integrated testing strategies that combine human-based cell systems, computational modeling, and patient-specific approaches. As noted by Dr. Eckhard von Keutz, former SVP at Bayer, "By adopting human-relevant models early, companies can make more informed go/no-go decisions, ultimately saving both time and capital while advancing safer, more effective therapeutics" [97]. This transition represents not merely a technical improvement but a fundamental transformation toward more predictive, ethical, and efficient safety assessment paradigms.

The field of toxicology is undergoing a paradigm shift, moving away from traditional animal models toward more human-relevant, efficient, and ethical New Approach Methodologies (NAMs). This transition is particularly evident in the regulatory assessment of skin sensitization and ocular irritation, two key endpoints for chemical and product safety. Driven by scientific advancement, ethical concerns, and legislative action, regulatory agencies worldwide are now formally accepting non-animal testing approaches for these endpoints. This guide documents the success stories of this transition, providing researchers and drug development professionals with a clear comparison of alternative methods, their experimental protocols, and their established regulatory status. The adoption of these methods represents a critical achievement within the broader thesis on validating NAMs, demonstrating that a future free of animal testing is not only possible but is already being realized in specific, well-defined areas [101] [11].

Skin Sensitization

The Adverse Outcome Pathway (AOP) and Defined Approaches

Skin sensitization is a complex biological process that has been meticulously delineated into an Adverse Outcome Pathway (AOP). The AOP describes a sequence of measurable key events (KEs) beginning with the molecular initiating event and progressing to the adverse outcome in an organism [101]. This mechanistic understanding has been foundational for developing non-animal test methods that target each key event.

The following diagram illustrates the Key Events in the Skin Sensitization Adverse Outcome Pathway and the associated alternative testing methods.

Rather than relying on a single test, regulatory acceptance has been achieved through Defined Approaches (DAs). DAs are fixed data interpretation procedures that integrate results from multiple non-ananimal methods, often covering different KEs, to predict a skin sensitization hazard or potency [102] [103]. The OECD Guideline No. 497, adopted in 2021 and updated in 2025, is the first internationally harmonized guideline to describe DAs that can officially replace the need for an animal test [103].

Regulatory Acceptance and Key Test Methods

Major regulatory agencies have moved from research to implementation, embedding these alternative methods into official policy.

U.S. Environmental Protection Agency (EPA): In 2018, the EPA released a draft interim science policy stating it would accept specific defined approaches for identifying skin sensitizers under the conditions described in the policy. This applies to pesticides and other chemicals regulated under statutes like TSCA and FIFRA. The EPA now accepts data from in chemico and in vitro tests that feed into DAs, effectively replacing the former animal test requirements for many applications [102] [103].

U.S. Food and Drug Administration (FDA): In June 2023, the FDA finalized guidance stating that it will consider skin sensitization data generated using a battery of in silico, in chemico, and in vitro studies that have been shown to predict human sensitization with accuracy comparable to in vivo methods. This applies to the safety assessment of products under its purview [103].

The table below summarizes the key alternative test methods for skin sensitization and their regulatory standing.

Table 1: Key Non-Animal Test Methods for Skin Sensitization Assessment

Test Method (OECD Guideline)	Key Event Targeted	Brief Principle	Regulatory Acceptance Status
Direct Peptide Reactivity Assay (DPRA) (TG 442C)	KE1: Molecular Initiating Event	Measures the ability of a chemical to bind to synthetic peptides, modeling the covalent binding to skin proteins.	Accepted within defined approaches under EPA draft policy and FDA guidance; part of OECD TG 497 [102] [103].
KeratinoSens (TG 442D)	KE2: Keratinocyte Response	Uses a genetically modified keratinocyte cell line to measure the activation of the Nrf2 antioxidant response pathway, a key cellular stress response.	Accepted within defined approaches under EPA draft policy and FDA guidance; part of OECD TG 497 [103].
Human Cell Line Activation Test (h-CLAT) (TG 442E)	KE3: Dendritic Cell Activation	Measures changes in surface marker expression (CD86 and CD54) on a human monocytic cell line (THP-1) to simulate dendritic cell activation.	Accepted within defined approaches under EPA draft policy and FDA guidance; part of OECD TG 497 [103].
EpiSensA	KE2: Keratinocyte Response	A more recent assay measuring gene expression in a 3D reconstituted human epidermis model; a proposal for an OECD test guideline is under consideration [103].	Under evaluation; peer review of validation study completed. Represents a newer, potentially more refined method [103].

Case Study: The OECD TG 497 Defined Approaches

A prime success story is the adoption of OECD TG 497: Defined Approaches for Skin Sensitisation. This guideline provides regulators with validated, non-animal testing strategies. For example, one DA within TG 497, the "2o3" rule, integrates data from the DPRA, KeratinoSens, and h-CLAT assays. A positive prediction in any two of these three assays leads to an overall classification of the substance as a skin sensitizer. The performance of these DAs has been extensively evaluated, with analyses showing that most perform better than standard animal methods in predicting human skin sensitization hazard and potency [103].

The ICCVAM Skin Sensitization Workgroup and its successor, the Skin Sensitization Expert Group, have been instrumental in this progress, coordinating implementation efforts across U.S. federal agencies [103].

Ocular Irritation

Moving Beyond the Draize Test

For decades, the Draize rabbit eye test was the standard for assessing ocular irritation. The development of human-relevant tissue models and physicochemical understanding has now provided a suite of validated alternatives. The regulatory strategy for ocular irritation often involves a testing framework that uses a combination of in vitro and ex vivo assays to categorize materials without using live animals [102].

Accepted Testing Frameworks and Assays

The EPA has been a leader in accepting alternative approaches for ocular irritation testing, particularly for antimicrobial cleaning products. The agency's updated guidance from 2015 describes a tiered testing framework using three core assays [102].

The workflow for a typical tiered testing strategy for ocular irritation assessment is shown below.

U.S. Environmental Protection Agency (EPA): The EPA's framework for antimicrobial cleaning products is a landmark in regulatory acceptance. It allows for the use of the Bovine Corneal Opacity and Permeability (BCOP), EpiOcular, and Cytosensor Microphysiometer assays in a strategic sequence to classify the eye irritation potential of formulations. This approach is also considered on a case-by-case basis for other pesticide classes [102].

To expand the use of NAMs, the EPA collaborated with NICEATM and the PETA Science Consortium to evaluate the performance of two proposed Defined Approaches for classifying agrochemical formulations into EPA hazard categories. A retrospective analysis of 29 agrochemical formulations demonstrated the utility of these approaches, further building confidence in non-animal methods [102].

Table 2: Key Non-Animal Test Methods for Ocular Irritation Assessment

Test Method	Model Type	Brief Principle	Regulatory Acceptance Status
Bovine Corneal Opacity and Permeability (BCOP)	Ex vivo (isolated tissue)	Uses bovine corneas to measure changes in opacity and permeability as indicators of damage, effectively identifying severe irritants and corrosives.	Accepted by EPA in a tiered testing strategy for antimicrobial cleaning products and other pesticides [102].
EpiOcular	In vitro (3D reconstructed human tissue)	Uses a 3D model of human corneal epithelium. Cell viability is measured after exposure; a significant reduction indicates irritation potential.	Accepted by EPA in a tiered testing strategy; core method for distinguishing mild and non-irritants [102].
Cytosensor Microphysiometer	In vitro (cell-based)	Measures changes in the metabolic rate of cultured cells (e.g., L929 mouse fibroblasts) upon exposure to a test substance.	Accepted by EPA as part of its tiered testing framework for ocular irritation [102].

The Scientist's Toolkit: Essential Reagents and Models

Implementing these alternative methods requires specific research tools. The following table details key reagents and model systems essential for conducting these assays.

Table 3: Essential Research Reagent Solutions for Skin Sensitization and Ocular Irritation Testing

Reagent / Model Solution	Function in Testing	Specific Example Assays
Synthetic Peptides	Serve as nucleophilic targets to measure a chemical's covalent binding potential (reactivity), which is the Molecular Initiating Event of skin sensitization.	Direct Peptide Reactivity Assay (DPRA) [103].
Engineered Keratinocyte Cell Lines	Reporter gene cell lines used to detect the activation of specific cellular stress response pathways relevant to the skin sensitization process.	KeratinoSens, LuSens [103].
Human Monocyte Cell Line (THP-1)	Used to measure the activation of dendritic cells, a key event in the immune response of skin sensitization, by tracking surface marker expression.	Human Cell Line Activation Test (h-CLAT) [103].
Reconstructed Human Epidermis (RhE) Models	3D human cell-derived tissues that mimic the structure and biology of the outer layer of human skin. Used for both sensitization (e.g., EpiSensA) and irritation testing.	EpiSensA (sensitization), EpiOcular (irritation) [102] [103].
Isolated Corneal Tissues	Ex vivo tissues, typically from livestock, used to model damage to the human cornea by measuring changes in physical and optical properties.	Bovine Corneal Opacity and Permeability (BCOP) Assay [102].

The regulatory acceptance of alternative methods for skin sensitization and ocular irritation stands as a testament to the power of collaborative science. Through the elucidation of adverse outcome pathways, the development of sophisticated in vitro, in chemico, and in silico methods, and the creation of validated Defined Approaches, the toxicology community has successfully built a new, human-relevant paradigm for safety assessment. The official adoption of these methods by agencies like the U.S. EPA and FDA under specific policies and international guidelines like OECD TG 497 provides a clear roadmap for researchers and industry. These success stories not only spare animals but also provide data that is often more predictive of human responses, ultimately enhancing public health protection. They serve as a powerful model and inspiration for ongoing efforts to replace animal testing for other more complex toxicological endpoints.

The Role of Initiatives like ICCVAM and OECD in Test Guideline Development and Harmonization

The global scientific community faces a critical challenge in transitioning from traditional animal testing to more human-relevant New Approach Methodologies (NAMs) while maintaining scientific rigor and regulatory acceptance. This paradigm shift is orchestrated primarily through two pivotal organizations: the Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM) and the Organisation for Economic Co-operation and Development (OECD). These organizations establish the foundational test guidelines that define internationally recognized methods for chemical safety assessment, thereby creating a standardized framework for regulatory toxicology and ecotoxicology research. The development of these guidelines represents a concerted effort to balance scientific advancement with regulatory needs, focusing on methods that reduce, refine, or replace animal use (the 3Rs) while ensuring human and environmental protection [104] [105].

The validation process for these methodologies is defined as "the process by which the reliability and relevance of a particular method or approach is established for a defined purpose" [24]. For researchers and drug development professionals, understanding the roles and interactions of ICCVAM and OECD is essential for designing studies that will meet international regulatory standards and advance the adoption of NAMs in toxicological research and chemical safety assessment.

Organizational Profiles and Mandates

ICCVAM: United States Framework

ICCVAM is a permanent committee of the National Institute of Environmental Health Sciences (NIEHS) that was formally established in 2000 by the ICCVAM Authorization Act (42 U.S.C. 285l-3) [106]. The committee comprises representatives from 18 U.S. federal regulatory and research agencies that require, use, generate, or disseminate toxicological and safety testing information.

The statutory purposes of ICCVAM include increasing the efficiency and effectiveness of federal agency test method review, eliminating unnecessary duplication of effort, optimizing utilization of external scientific expertise, ensuring new and revised test methods are validated to meet federal agency needs, and reducing, refining, or replacing animal use in testing where feasible [106]. This mandate positions ICCVAM as a central coordinating body for alternative method validation within the U.S. regulatory landscape.

Table 1: Key U.S. Agencies Represented on ICCVAM

Agency Type	Examples	Primary Responsibilities
Regulatory Agencies	U.S. Environmental Protection Agency (EPA), U.S. Food and Drug Administration (FDA)	Chemical and product safety regulation, pesticide registration, tolerance setting
Research Agencies	National Institute of Environmental Health Sciences (NIEHS), National Center for Advancing Translational Sciences (NCATS)	Method development, validation research, scientific advancement
Health Agencies	Department of Veterans Affairs Office of Research and Development	Health impact assessment, veteran healthcare research

OECD: International Framework

The OECD Test Guidelines Programme provides a mechanism for international harmonization of chemical safety assessment methods across its 38 member countries. The OECD guidelines are developed through a consensus-based approach with input from member countries, including subject matter experts from ICCVAM agencies [104] [107]. These guidelines form the basis for chemical safety assessments across most industrialized nations and fall under the OECD Mutual Acceptance of Data (MAD) decision, which stipulates that data generated in accordance with OECD Test Guidelines in one member country must be accepted by all other member countries for regulatory purposes [104].

This mutual acceptance system eliminates significant duplicative testing, conserves scientific resources (including minimizing laboratory test animal use), and forms a basis for work sharing and cooperation among all OECD countries. The harmonization of test guidelines through OECD thereby reduces non-tariff barriers to trade while maintaining high standards of environmental and human health protection.

Collaborative Mechanisms and Harmonization Processes

Guideline Development Workflows

The development and revision of test guidelines follow structured processes that incorporate scientific peer review, international stakeholder input, and regulatory consideration. The collaboration between ICCVAM and OECD creates an integrated framework from initial method development through international regulatory acceptance.

Diagram 1: Test Guideline Development Workflow

ICCVAM Workgroups and Their Functions

ICCVAM establishes specialized workgroups to address specific testing challenges and advance the development of alternative methods. These workgroups comprise representatives from multiple agencies with relevant expertise and are instrumental in driving the scientific agenda for NAM development and validation.

Table 2: Active ICCVAM Workgroups and Focus Areas (2022-2023)

Workgroup	Participating Agencies	Key Activities and Focus Areas
Acute Toxicity Workgroup	7 ICCVAM agencies	Evaluating in silico models for acute oral and inhalation toxicity; assessing variability of in vivo benchmark data [71]
Ecotoxicology Workgroup	7 ICCVAM agencies	Identifying in vitro and in silico methods for ecological hazards; evaluating alternatives to acute fish toxicity tests [71]
PFAS Workgroup	9 ICCVAM agencies	Developing NAMs to assess per- and polyfluoroalkyl substances toxicity; addressing scientific and regulatory challenges [71]
Common Data Elements Workgroup	7 ICCVAM agencies	Standardizing terms and formats for data sharing; supporting NIH Complement-ARIE program repository [71]

The outputs from these workgroups frequently form the scientific basis for U.S. positions in OECD test guideline development activities. For instance, the Acute Toxicity Workgroup's analysis of in vivo data variability informed performance benchmarks for evaluating alternative methods [71], which subsequently contributed to OECD discussions on validation criteria.

International Cooperation through ICATM

The International Cooperation on Alternative Test Methods (ICATM) provides a formal mechanism for coordination among validation organizations worldwide. Established in 2009 through an agreement between ICCVAM, the European Union Reference Laboratory for alternatives to animal testing (EURL ECVAM), the Japanese Center for the Validation of Alternative Methods (JaCVAM), and Health Canada, ICATM has since expanded to include the Korean Center for the Validation of Alternative Methods (KoCVAM), with additional participation from Brazil, the United Kingdom, Singapore, and Taiwan [108].

ICATM's primary goals are to establish international cooperation in validation studies, ensure independent peer review, develop harmonized recommendations, and promote worldwide acceptance of alternative methods and strategies. This cooperation helps avoid duplication of effort, leverages limited resources, and supports timely international adoption of alternative methods [108]. The collaboration is particularly evident in specific validation studies, such as the JaCVAM-coordinated validation of the EpiSensA skin sensitization test method, in which NICEATM participated in the peer review [108].

Advancements in Key Methodological Areas

Skin Sensitization Assessment

The development and validation of defined approaches for skin sensitization assessment represent a landmark achievement in the implementation of NAMs. This effort demonstrates the successful collaboration between ICCVAM and OECD in translating scientific advances into internationally accepted test guidelines.

The OECD Guideline No. 497: Defined Approaches for Skin Sensitisation, adopted in 2021 and updated in 2023, is the first internationally harmonized guideline to describe a non-animal defined approach that can replace animal tests for identifying skin sensitizers [107] [108]. The development of this guideline was facilitated by a 2016 ICATM workshop that reviewed international regulatory requirements and identified steps needed to support regulatory acceptance of non-animal approaches [108].

The defined approaches for skin sensitization integrate multiple information sources, including in chemico and in vitro methods that measure key events in the adverse outcome pathway for skin sensitization. These key events include covalent binding to proteins (measured in assays like DPRA and Direct Peptide Reactivity Assay) and keratinocyte activation (measured in assays like KeratinoSens and LuSens) [108]. The validation process for these approaches considered their alignment with human biology rather than solely comparing results to historical animal test data [24].

Acute Systemic Toxicity Testing

The ICCVAM Acute Toxicity Workgroup has made significant progress in advancing alternatives to traditional animal tests for acute oral, dermal, and inhalation toxicity. Their work includes publishing a scoping document that identifies U.S. agency information requirements and decision contexts, analyzing variability in in vivo data used as benchmarks, and organizing global projects to develop in silico models for specific regulatory endpoints [71].

A key achievement is the organization of a global project to develop in silico models of acute oral systemic toxicity that predict five specific endpoints needed by regulatory agencies [71]. This collaborative effort exemplifies how ICCVAM facilitates the development of tools that can potentially replace animal testing while meeting regulatory needs. The workgroup is currently applying similar approaches to acute inhalation toxicity, addressing one of the more challenging areas for alternative method development.

Ecotoxicology and Environmental Safety

While initial ICCVAM activities focused primarily on human health toxicology, the establishment of the Ecotoxicology Workgroup reflects growing attention to implementing non-animal approaches for ecotoxicity testing. This workgroup has compiled a comprehensive summary of agency needs for ecotoxicity testing and emerging technologies for evaluating ecological and environmental hazards [71].

The workgroup is currently focusing on evaluating alternatives to the acute fish toxicity test, a widely used animal test in environmental hazard assessment. This effort aligns with broader international activities at OECD, where ICCVAM agency representatives contribute to the development and revision of ecotoxicology test guidelines that emphasize reduction, refinement, or replacement of animal testing [104].

Validation Frameworks and Scientific Confidence

Evolution of Validation Principles

The validation principles for toxicological test methods were initially formalized in the OECD Guidance Document 34, "Guidance Document on the Validation and International Acceptance of New or Updated Test Methods for Hazard Assessment," published in 2005 [24]. While these fundamental principles remain relevant, there is widespread recognition that validation processes need updating to accommodate advances in NAMs and encourage their timely regulatory uptake.

Traditional validation approaches have emphasized multi-laboratory ring trials to assess transferability and performance, with predictive capacity typically determined through comparison to historical animal test results [24]. However, participants in a 2023 ICATM coordination meeting acknowledged that this approach may not be practical for many new technologies underlying NAMs [108]. There is growing consensus that validation should focus more on biological relevance and mechanistic understanding, with well-defined protocols, diverse reference chemicals, clear acceptance criteria, and coordinated peer review [108].

Modern Framework for Establishing Scientific Confidence

A contemporary framework for establishing scientific confidence in NAMs emphasizes five essential elements [24]:

• Fitness for Purpose: The NAM must be technically adequate for its intended application and decision context.
• Human Biological Relevance: Assessment should focus on alignment with human biology and mechanistic understanding, rather than solely comparing results to animal tests.
• Technical Characterization: The method must demonstrate reliability through measures of intra- and inter-laboratory reproducibility.
• Data Integrity and Transparency: Complete and transparent reporting of methods, data, and analyses is essential.
• Independent Review: External expert evaluation provides critical assessment of the method's suitability.

This framework recognizes that NAMs need not produce identical information to traditional animal tests and may instead provide biologically relevant information and mechanistic insights more useful for regulatory decision-making [24]. This shift in perspective is crucial for advancing the integration of NAMs into regulatory toxicology.

Research Tools and Reagent Solutions

The implementation of validated test methods requires specific research tools and reagents that enable standardized and reproducible testing across laboratories. The following table highlights key research reagent solutions used in NAMs that have been incorporated into ICCVAM and OECD test guidelines.

Table 3: Essential Research Reagents for Alternative Test Methods

Reagent/Assay System	Application	Function in Testing Paradigm	Regulatory Status
Reconstructed Human Epidermis (RHE) Models	Skin corrosion/irritation, sensitization	Models human skin barrier function and tissue response; used in EpiSensA test method [108]	OECD TG 439, 431, 442D
Human Hepatoma Cell Line (HepaRG)	Metabolic competence, hepatotoxicity	Provides in vitro model of human liver metabolism and CYP enzyme induction [108]	Under consideration for OECD guideline
LuSens Keratinocyte Assay	Skin sensitization	Measures gene expression changes associated with keratinocyte activation, a key event in skin sensitization [108]	Part of defined approaches in OECD TG 497
SENS-IS Test Method	Skin sensitization	Evaluates skin sensitization potential using a proprietary reconstructed epidermis model [108]	Undergoing validation peer review
Defined Approach (DA) for Skin Sensitization	Skin sensitization hazard identification and potency categorization	Integrates results from multiple in chemico and in vitro methods to classify sensitization potential without animal testing [107]	OECD TG 497 (adopted 2021, updated 2023)

The collaborative efforts of ICCVAM and OECD in test guideline development and harmonization represent a critical framework for advancing the scientific and regulatory acceptance of New Approach Methodologies. Through structured validation processes, specialized workgroups, and international cooperation, these organizations facilitate the transition to human-relevant toxicological testing while maintaining scientific rigor and regulatory oversight.

The ongoing revision of validation principles to focus on human biological relevance rather than correlation with historical animal data marks a significant evolution in toxicological science. For researchers and drug development professionals, understanding these frameworks and the associated test guidelines is essential for designing studies that will meet international regulatory standards and contribute to the continued advancement of NAMs in toxicology and ecotoxicology research.

As noted in the ICCVAM Biennial Report, future efforts will continue to focus on developing and implementing alternative approaches across a broadening range of endpoints and chemical classes, including particularly challenging areas such as PFAS assessment and developmental neurotoxicity [107] [71]. This ongoing work ensures that test guideline development remains responsive to both emerging scientific advances and evolving regulatory needs in chemical safety assessment.

The field of ecotoxicology is undergoing a significant transformation with the emergence of New Approach Methodologies (NAMs), defined as any technology, methodology, approach, or combination thereof that can refine, reduce, or replace reliance on traditional animal toxicity testing [93]. This shift is driven by regulatory agencies worldwide that are calling for these methods to streamline chemical hazard assessment while adhering to the 3Rs principles (Replacement, Reduction, and Refinement) [16] [93]. The validation of NAMs requires rigorous benchmarking against established toxicological outcomes to ensure they provide information of "equivalent or better" scientific quality and relevance for regulatory decision-making [16].

For researchers, scientists, and drug development professionals, establishing confidence in NAMs requires a systematic framework for evaluating their scientific validity and reliability. This involves demonstrating that these methods consistently produce accurate, reproducible results that are predictive of human and ecological health outcomes [27]. The U.S. Environmental Protection Agency (EPA) has developed a strategic approach that includes characterizing the scientific quality of existing tests, developing recommended reporting requirements, and demonstrating NAM applications through case studies [16]. This comprehensive guide examines the key metrics, experimental protocols, and benchmarking frameworks essential for establishing NAMs as trustworthy tools in ecotoxicological research and chemical safety assessment.

Core Performance Metrics for NAMs

Benchmarking NAM performance requires assessing multiple dimensions of method validity and reliability, drawing from established principles of measurement instruments used in scientific research [109] [110]. The key indicators include reliability estimates that evaluate the stability of measures, internal consistency of measurement instruments, and interrater reliability of instrument scores [109]. Meanwhile, validity represents the extent to which the interpretations of test results are warranted for their intended use [109].

Reliability Metrics

Reliability refers to the consistency and stability of a measurement method across different conditions, timepoints, and operators [110]. For NAMs, this encompasses several distinct types of reliability, summarized in the table below.

Table 1: Reliability Metrics for NAM Performance Assessment

Metric Type	Definition	Assessment Method	Target Threshold
Test-Retest Reliability	Consistency of measures when the same test is administered at different times to the same biological system	Correlation between scores from two timepoints (Pearson's r)	≥ +0.80 [110]
Internal Consistency	Consistency of responses across items in a multiple-measure assay	Split-half correlation or Cronbach's α	≥ +0.80 [110]
Inter-rater Reliability	Consistency of judgments between different observers or raters	Cohen's κ (categorical) or Cronbach's α (quantitative)	Varies by context [110]

Validity Metrics

Validity represents the extent to which a measurement method actually assesses the theoretical construct it purports to measure [110]. For NAMs in ecotoxicology, this involves demonstrating that the method accurately predicts adverse outcomes relevant to human health or ecological systems.

Table 2: Validity Metrics for NAM Performance Assessment

Validity Type	Assessment Approach	Evidence Examples for NAMs
Face Validity	Extent to which a method appears to measure the target construct	Neuronal network inhibition measured via Microelectrode Array (MEA) for neuroactive substances [111]
Content Validity	Extent to which a measure "covers" the construct of interest	Coverage of multiple neurotoxic pathways in the ENRICH list of 250 prioritized chemicals [112]
Criterion Validity	Correlation with established variables (criteria)	Correlation of in vitro bioactivity with known in vivo toxicity outcomes [112] [111]

Additional Performance Metrics

Beyond traditional reliability and validity measures, NAMs require assessment of additional parameters relevant to their application in regulatory and research contexts. Responsiveness of the measure to change is of particular interest in applications where improvement in outcomes as a result of treatment is a primary goal [109]. The predictive capacity of NAMs is frequently evaluated through statistical measures including sensitivity (true positive rate), specificity (true negative rate), and overall accuracy compared to traditional animal test results [27].

For high-throughput screening methods, additional practical metrics include throughput capacity (number of compounds screened per unit time), cost-efficiency compared to traditional methods, and technical reproducibility across laboratories [16] [93]. The EPA's CompTox Chemicals Dashboard represents one such tool that enables data interpretation, translation, and chemical prioritization at scale [16].

Experimental Protocols for Benchmarking NAMs

Microelectrode Array (MEA) Recordings for Neurotoxicity

Microelectrode Array (MEA) recordings provide a functional measure of neuronal network activity and have emerged as a key NAM for assessing neurotoxicity [111]. The following protocol details the methodology for benchmarking MEA performance against traditional neurotoxicity assessments.

Experimental Workflow:

• Primary Cortical Culture Preparation: Dissociate cortical tissue from Sprague-Dawley rat embryos (day 18), plate cells onto MEA plates pre-coated with poly-D-lysine and laminin, and maintain cultures in neurobasal medium with B-27 supplement for 21-28 days until mature network activity develops [111].
• Test Substance Preparation: Prepare botanical extracts or isolated constituents in dimethyl sulfoxide (DMSO) followed by dilution in culture medium, ensuring final DMSO concentration does not exceed 0.5% [111].
• Baseline Recording: Record spontaneous electrical activity for 10 minutes before compound addition to establish baseline network parameters including mean firing rate, burst frequency, and network synchronization [111].
• Compound Exposure & Recording: Expose cultures to test compounds across a concentration range (typically 0.1-100 µg/mL) with continuous MEA recording for 30-60 minutes post-exposure [111].
• Data Analysis: Calculate dose-response curves for each activity parameter, determine No Observed Effect Levels (NOELs), and classify compounds as inhibitory (e.g., kava, kratom) or excitatory (e.g., aconite, oleander) based on their activity phenotypes [111].

Benchmarking Parameters:

• Potency Comparison: Compare NOELs between botanical extracts and their known bioactive constituents (e.g., oleandrin for oleander) [111].
• Phenotypic Concordance: Assess whether constituent compounds recapitulate the neuronal activity phenotypes of whole extracts [111].
• Predictive Validation: Correlate in vitro MEA results with known in vivo human neurotoxicity for benchmark compounds [111].

Figure 1: MEA Experimental Workflow for Neurotoxicity Assessment

High-Throughput Toxicokinetics for Biomonitoring

The ENRICH (Environmental NeuRoactIve CHemicals) list development exemplifies a computational NAM approach for prioritizing chemicals for neurotoxicity testing and biomonitoring [112]. This methodology combines database mining with high-throughput toxicokinetic modeling to predict chemical exposure and biological detection.

Experimental Workflow:

• Database Mining: Compile chemicals from multiple sources including EPA's CompTox Chemicals Dashboard, identify substances with potential neuroactivity through in vitro high-throughput screening, animal testing, and human epidemiological data [112] [16].
• Exposure Assessment: Evaluate likelihood of human exposure through presence in consumer products, environmental monitoring data, and chemical properties suggesting biological persistence [112].
• Chemical Prioritization Index Calculation: Rank chemicals using a scoring system that incorporates neuroactivity evidence, exposure potential, and detectability in biological samples [112].
• Toxicokinetic Modeling: Apply high-throughput toxicokinetic models to predict parent compounds and metabolites in urine as a practical biological matrix for biomonitoring [112].
• Manual Curation: Apply inclusion/exclusion criteria to ensure environmental relevance and analytical feasibility (e.g., mass spectrometry compatibility) [112].

Validation Approach:

• Cross-Source Verification: Compare chemical priorities across multiple regulatory and scientific sources [112] [16].
• Analytical Feasibility Assessment: Verify that prioritized chemicals can be detected using current biomonitoring methods [112].
• Retrospective Validation: Compare predictions against existing human biomonitoring data where available [112].

Signaling Pathways and Neurotoxicity Mechanisms

Understanding the molecular pathways affected by neurotoxic substances is essential for developing mechanistically relevant NAMs. Research has identified several key pathways that can be monitored using in vitro and in silico approaches.

Figure 2: Key Neurotoxicity Pathways for NAM Assessment

The Adverse Outcome Pathway (AOP) framework provides a structured approach for linking molecular initiating events to adverse outcomes at organismal and population levels [16]. The EPA is actively developing AOPs for high-priority pathways and chemicals, which helps establish the scientific rationale supporting the use of NAMs in evaluating potential chemical impacts [16]. For neuroactive substances, common molecular initiating events include ion channel modulation (e.g., aconitine from aconite) [111], receptor activation/blockade (e.g., yohimbine from yohimbe) [111], and disruption of calcium homeostasis [111]. These initiating events lead to cellular responses such as altered neuronal firing patterns, network synchronization defects, and ultimately neuronal cell death, which can be quantified using MEA recordings and high-content imaging [111].

Research Reagent Solutions for NAM Implementation

Successful implementation of NAMs requires specific research tools and reagents tailored to neurotoxicity assessment and high-throughput screening. The following table details essential materials and their applications in benchmarked experimental protocols.

Table 3: Essential Research Reagents for NAM Implementation

Reagent/Category	Specific Examples	Research Application	Protocol Reference
Cell Culture Systems	Primary rat cortical cultures; Human induced pluripotent stem cell (iPSC)-derived neurons	Provide biologically relevant models for neuroactivity screening	[111]
Bioactivity Detection Platforms	Microelectrode Array (MEA) platforms; High-content imaging systems	Functional assessment of neuronal network activity	[111]
Chemical Libraries	ENRICH list of 250 prioritized neuroactive chemicals; EPA's CompTox Chemicals Dashboard	Defined screening libraries for standardized assessment	[112] [16]
Toxicokinetic Modeling Tools	High-throughput toxicokinetic models; Physiologically based kinetic (PBK) models	Prediction of biological exposure and metabolite formation	[112]
Analytical Standards	Certified reference materials for prioritized chemicals and metabolites	Quality control and method validation for biomonitoring	[112]

Benchmarking NAM performance requires a multifaceted approach that assesses reliability, validity, and predictive capacity across multiple experimental contexts. While significant progress has been made in developing methods such as MEA recordings for neurotoxicity [111] and computational approaches for chemical prioritization [112], the field continues to face challenges in standardization and regulatory adoption [27]. The implementation of NAMs on a wide scale will require time, research investment, and focused validation efforts to build confidence in their predictive power [27].

A hybrid approach that combines alternative methods with traditional animal testing during the transition period allows for comparative data generation while supporting the gradual phase-out of animal studies where scientifically justified [27]. This strategy is particularly relevant for complex toxicological endpoints such as neurodevelopmental effects, where the EPA is developing virtual tissue models to evaluate chemical exposure impacts during human development [16]. As the scientific community works toward standardized validation frameworks and increased mechanistic understanding, NAMs are poised to transform chemical safety assessment while reducing reliance on traditional animal testing approaches.

Conclusion

The validation and integration of New Approach Methodologies mark a pivotal advancement in ecotoxicology, moving the field toward more human-relevant, efficient, and ethical safety assessments. The convergence of robust scientific evidence, evolving regulatory frameworks, and innovative technologies has created an irreversible momentum. For researchers and drug development professionals, the path forward requires active participation in standardizing protocols, contributing to open data repositories, and engaging with regulatory pilots. Future success hinges on global collaboration to harmonize validation standards, continuously refine computational and organ-on-chip models, and ultimately build an ecosystem where NAMs are the default, not the exception, for protecting human health and the environment.

References

• 1. The dawn of animal-free safety: A comprehensive review ... [https://www.sciencedirect.com/science/article/pii/S305047592500853X]
• 2. New approach methodologies (NAMs): identifying and ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10964841/]
• 3. The end of animals-only testing brings both opportunities and ... [https://hub.jhu.edu/2025/07/16/3-questions-organoids-animal-testing-alternatives-thomas-hartung/]
• 4. Non-Animal & Human-Relevant Research News [https://riseforanimals.org/news/sci-news-sept-2025/]
• 5. Alternatives to Animal Testing [https://www.niehs.nih.gov/health/topics/science/sya-iccvam]
• 6. A pragmatic workflow for human relevance assessment of ... [https://pubmed.ncbi.nlm.nih.gov/40228575/]
• 7. Advancing NAMs for Risk Assessment: Perspectives from SOT ... [https://www.jhutoxicologypolicyresearch.org/tox-blog/blog43]
• 8. The FDA's historic move beyond animal testing mandates [https://www.sciopen.com/article/10.26599/CO.2025.9410013]
• 9. FDA Modernization Act 2.0 – what, if anything, has ... [https://www.eara.eu/post/fda-modernization-act-2-0-what-if-anything-has-changed-in-us-regulations]
• 10. A Recent Timeline of U.S. Actions to Reduce and Replace ... [https://emulatebio.com/alternatives-to-animal-testing-in-drug-development/]
• 11. FDA Announces Plan to Phase Out Animal Testing ... [https://www.fda.gov/news-events/press-announcements/fda-announces-plan-phase-out-animal-testing-requirement-monoclonal-antibodies-and-other-drugs]
• 12. FDA Announces Plan to Phase Out Animal Testing ... [https://barragan.house.gov/2025/04/16/jdsupra-fda-announces-plan-to-phase-out-animal-testing-requirement/]
• 13. Products | UK Regulatory Outlook November 2025 [https://www.osborneclarke.com/insights/regulatory-outlook-november-2025-products]
• 14. First major overhaul of medical device regulation comes ... [https://www.gov.uk/government/news/first-major-overhaul-of-medical-device-regulation-comes-into-force-across-great-britain]
• 15. Major Overhaul of UK Rare Disease Regulation: MHRA ... [https://www.morganlewis.com/blogs/asprescribed/2025/11/major-overhaul-of-uk-rare-disease-regulation-mhra-proposes-enhanced-framework]
• 16. Efforts to Reduce Use of Vertebrate Animals in Chemical ... [https://www.epa.gov/chemical-research/epa-new-approach-methods-efforts-reduce-use-vertebrate-animals-chemical-testing]
• 17. FDA Modernization Act 2.0: transitioning beyond animal ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10617761/]
• 18. New Approach Methodologies [https://www.setac.org/page/sciencehts/]
• 19. Alternatives to animal testing [https://crueltyfreeinternational.org/about-animal-testing/alternatives-animal-testing]
• 20. Antibody-based alternatives to animal testing for toxin ... [https://www.frontiersin.org/journals/toxicology/articles/10.3389/ftox.2025.1636246/full]
• 21. 2025 Annual Meeting Report: NAMs Hit the Bullseye in ... [https://toxchange.toxicology.org/blogs/rebecca-somach/2025/05/15/2025-annual-meeting-report-nams-hit-the-bullseye-i]
• 22. New Approach Methodologies (NAMs): The Future of ... [https://emulatebio.com/new-approach-methodologies-nams-the-future-of-toxicology-and-safety-testing/]
• 23. Validating NAMs for FDA's Plan to Phase Out Animal Testing [https://www.jhutoxicologypolicyresearch.org/tox-blog/blog45]
• 24. A framework for establishing scientific confidence in new ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9525335/]
• 25. Alternatives to Animal Research | Harvard Medical School [https://hms.harvard.edu/research/animal-research/alternatives-animal-research]
• 26. Use of new approach methods (NAMs) in risk assessment [https://www.canada.ca/en/health-canada/services/chemical-substances/fact-sheets/use-new-approach-methods-risk-assessment.html]
• 27. Animal testing alternatives to dominate, but not overnight [https://www.clinicaltrialsarena.com/interviews/animal-testing-alternatives-to-dominate-but-not-overnight/]
• 28. $32.88 Bn In-vitro Toxicology Testing Market Trends and ... [https://finance.yahoo.com/news/32-88-bn-vitro-toxicology-092400584.html]
• 29. New alternatives to animal testing [https://genoskin.com/animal-testing/alternatives-to-animal-testing/]
• 30. 7 Assessment of Ecotoxicity | A Framework to Guide Selection ... [https://nap.nationalacademies.org/read/18872/chapter/9]
• 31. An evaluation framework for new approach methodologies ... [https://www.sciencedirect.com/science/article/pii/S0273230020300180]
• 32. A call to action: Advancing new approach methodologies ... [https://www.sciencedirect.com/science/article/pii/S0273230025001345]
• 33. In-Vitro Toxicology Testing Market Size 2025 to 2034 [https://www.precedenceresearch.com/in-vitro-toxicology-testing-market]
• 34. New Approach Methods (NAMs) [https://www.fda.gov/food/toxicology-research/new-approach-methods-nams]
• 35. Overview of Computational Toxicology Methods Applied in ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11677645/]
• 36. Quantitative structure–activity relationship-based ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9300454/]
• 37. A Review of In Silico Tools as Alternatives to Animal Testing [https://pubmed.ncbi.nlm.nih.gov/33119417/]
• 38. Computational models to predict the toxicity of molecules [https://www.sciencedirect.com/science/article/pii/S0045653524028078]
• 39. Overview of In Silico Tools to Evaluate Human Health ... [https://pubmed.ncbi.nlm.nih.gov/40925860/]
• 40. A new computational cross-structure-activity relationship (C ... [https://www.sciencedirect.com/science/article/pii/S0010482525005207]
• 41. In silico trials can reduce animal testing [https://www.novainsilico.ai/in-silico-trials-can-reduce-animal-testing/]
• 42. Replacing Animal Testing with Stem Cell-Organoids [https://pmc.ncbi.nlm.nih.gov/articles/PMC11319430/]
• 43. Organoids and organs-on-chips: Recent advances ... [https://www.sciencedirect.com/science/article/abs/pii/S2666634025000947]
• 44. Beyond Animal Testing: The Rise of Organs-on-Chips ... [https://clarivate.com/life-sciences-healthcare/blog/beyond-animal-testing-the-rise-of-organs-on-chips-technology/]
• 45. Organoids VS Organ-on-a-chip: A Comparative Review ... [https://elveflow.com/microfluidic-reviews/organoids-vs-organ-on-a-chip/]
• 46. State-of-the-art in high throughput organ-on-chip for ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12149869/]
• 47. Organ on a chip Models | Single & Multi Organ Chip [https://cn-bio.com/organ-models/]
• 48. Organ-on-Chip in 2025: In the Spotlight After the NIH, FDA, ... [https://www.hmbioconsulting.com/post/organ-on-chip-in-2025-in-the-spotlight-after-the-nih-fda-and-gao-push]
• 49. Navigating complexity in modern toxicology: the role of omics ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12599865/]
• 50. Omics in Short-Term Toxicology Studies: Are We Ready to ... [https://alitheagenomics.com/blog/omics-in-short-term-toxicology-studies-are-we-ready-to-replace-traditional-approaches]
• 51. How Omics Revolutionize Toxicology: ETS2025 ... [https://www.linkedin.com/posts/shermaine-mitchell-ryan_ets2025-ets2025-omics-activity-7377465836179197952-MoOe]
• 52. Omics technologies to understand drug toxicity mechanisms [https://www.sciencedirect.com/science/article/pii/S1359644622003415]
• 53. Assessing the ecotoxicity of multicomponent nanomaterials ... [https://pubs.rsc.org/en/content/articlelanding/2025/en/d4en01183j]
• 54. Assessing the ecotoxicity of multicomponent nanomaterials ... [https://pubs.rsc.org/en/content/articlehtml/2025/en/d4en01183j]
• 55. Current Ecotoxicity Testing Needs Among Selected U.S. ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9623878/]
• 56. Use of new approach methodologies (NAMs) to meet ... [https://www.frontiersin.org/journals/toxicology/articles/10.3389/ftox.2022.964553/full]
• 57. A catalytic perspective to nanomaterials reactivity-based ... [https://www.sciencedirect.com/science/article/pii/S2452074825000023]
• 58. NanoToxRadar: A Multitarget Nano-QSAR Model for ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12532064/]
• 59. Review and Prospects on the Ecotoxicity of Mixtures ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9671040/]
• 60. A Curated Database of Ecologically Relevant Toxicity Tests to ... [https://academic.oup.com/etc/article/41/6/1520/7730623]
• 61. Biological Safety and Sustainability of Metal Nanomaterials ... [https://www.mdpi.com/2072-666X/16/1/15]
• 62. Integrating New Approach Methodologies (NAMs) into ... [https://www.mdpi.com/2313-7673/10/12/796]
• 63. New Approach Methodologies (NAMs): Advances in ... [https://www.axionbiosystems.com/new-approach-methodologies-nams-advances-preclinical-safety-studies]
• 64. New Approach Methodologies: What Clinical Pharmacologists ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12404678/]
• 65. ACC Research Program Joins Partnership for Validation of ... [https://www.americanchemistry.com/chemistry-in-america/news-trends/press-release/2025/acc-research-program-joins-partnership-for-validation-of-new-approach-methodologies-nams]
• 66. 177Lu-Labeled Magnetic Nano-Formulations: Synthesis, Radio- and... [https://www.mdpi.com/1420-3049/30/21/4290]
• 67. Progress in nanomaterials: Synthesis, characterization, ... [https://www.sciencedirect.com/science/article/pii/S2949829525001329]
• 68. 25 years of research and regulation: Is nanotechnology safe to... [https://pubmed.ncbi.nlm.nih.gov/40708971/]
• 69. Frontiers | 25 years of research and regulation: Is nanotechnology safe... [https://www.frontiersin.org/journals/toxicology/articles/10.3389/ftox.2025.1629813/full]
• 70. Analytical Methods for Characterization of Nanomaterial ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC7941215/]
• 71. ICCVAM Workgroups - National Toxicology Program - NIH [https://ntp.niehs.nih.gov/whatwestudy/niceatm/iccvam/wg]
• 72. Replacing animals in science: A strategy to support the ... [https://www.gov.uk/government/publications/replacing-animals-in-science-strategy/replacing-animals-in-science-a-strategy-to-support-the-development-validation-and-uptake-of-alternative-methods]
• 73. Characterization techniques for nanoparticles: comparison ... [https://pubs.rsc.org/en/content/articlelanding/2018/nr/c8nr02278j]
• 74. Polymeric Nanoparticles: Production, Characterization, Toxicology and... [https://pmc.ncbi.nlm.nih.gov/articles/PMC7464532/]
• 75. Chemical Testing Using New Approach Methodologies ... [https://www.frontiersin.org/research-topics/19075/chemical-testing-using-new-approach-methodologies-nams/magazine]
• 76. How Merging NAMs and Traditional Models Look in ... [https://www.criver.com/eureka/how-merging-nams-and-traditional-models-look-discovery]
• 77. A benchmark dataset for machine learning in ecotoxicology [https://www.nature.com/articles/s41597-023-02612-2]
• 78. Alternative approaches to animal testing [https://cefic.org/policy/alternative-approaches-to-animal-testing/]
• 79. FAIR Data Principles: Benefits, Challenges & FAQs [https://zontal.io/fair-data-principles/]
• 80. FAIR data principles: What you need to know [https://tiledb.com/blog/fair-data-principles-explained]
• 81. A Guide to the FAIR Principles [https://www.openscience.eu/article/infrastructure/guide-fair-principles]
• 82. Understanding FAIR vs Open Data in Life Sciences [https://ranchobiosciences.com/2025/01/unlocking-data-potential-understanding-fair-vs-open-data-in-life-sciences/]
• 83. Advancing the transition from animal models to alternative ... [https://www.parexel.com/insights/blog/advancing-the-transition-from-animal-models-to-alternative-methods-fdas-initiative-and-sponsors-opportunities]
• 84. FDA and NIH Announce Initiatives to Reduce Animal ... [https://www.lw.com/en/insights/fda-and-nih-announce-initiatives-to-reduce-animal-testing-and-encourage-alternative-methodologies]
• 85. Calibration and validation of toxicokinetic-toxicodynamic models for... [https://pubmed.ncbi.nlm.nih.gov/29717389/]
• 86. (PDF) Computational ecotoxicology : Simultaneous prediction of... [https://www.academia.edu/13622300/Computational_ecotoxicology_Simultaneous_prediction_of_ecotoxic_effects_of_nanoparticles_under_different_experimental_conditions]
• 87. Predictive and Exploratory Toxicology Solutions [https://www.rti.org/solution/predictive-toxicology-exploratory-biology-nams]
• 88. (ECOTOX) Knowledgebase Resource Hub | US EPA Ecotoxicology [https://www.epa.gov/comptox-tools/ecotoxicology-ecotox-knowledgebase-resource-hub]
• 89. 3Rs in Ecotoxicity Testing | National Agricultural Library [https://www.nal.usda.gov/animal-health-and-welfare/animal-use-alternatives/3Rs-ecotoxicity-testing]
• 90. Applying the Key Concepts to Build Confidence in NAMs - NCBI [https://www.ncbi.nlm.nih.gov/books/NBK614927/]
• 91. Are Animals Out? How NAMs Are Changing the Rules of ... [https://www.biospace.com/fda/are-animals-out-how-nams-are-changing-the-rules-of-the-game]
• 92. Replacing and Reducing Animal Testing at CDER [https://www.fda.gov/about-fda/center-drug-evaluation-and-research-cder/replacing-and-reducing-animal-testing-cder]
• 93. Progress on new approach methods (NAMs) in ecotoxicology [https://academic.oup.com/etc/article-abstract/44/9/2389/8111641]
• 94. New Methods in Human Safety Assessment: Right Here ... [https://www.criver.com/eureka/new-methods-human-safety-assessment-right-here-right-now]
• 95. New Approach Methods (NAMs) Training [https://www.epa.gov/chemical-research/new-approach-methods-nams-training]
• 96. Commentary on FDA's shift from animal testing and ... [https://www.sciencedirect.com/science/article/pii/S0273230025001266]
• 97. FDA NAMs Roadmap: Economic and Scientific Impact ... [https://www.4dcell.com/post/fdanamsroadmap]
• 98. In Vitro Alternatives to Animal Models - NCBI [https://www.ncbi.nlm.nih.gov/books/NBK507212/]
• 99. In vitro or not in vitro: a short journey through a long history [https://enveurope.springeropen.com/articles/10.1186/s12302-018-0151-3]
• 100. Editorial: New Approaches in Toxicity Testing of ... [https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2022.922551/full]
• 101. Alternative Methods for Skin-Sensitization Assessment - PubMed [https://pubmed.ncbi.nlm.nih.gov/36548573/]
• 102. Strategic Vision for Adopting New Approach Methodologies - Replacement Strategies | US EPA [https://www.epa.gov/pesticide-science-and-assessing-pesticide-risks/strategic-vision-adopting-new-approach-2]
• 103. Roadmap Implementation: Skin Sensitization [https://ntp.niehs.nih.gov/whatwestudy/niceatm/natl-strategy/rdmp-imp/imp-sensit]
• 104. About Test Guidelines for Pesticides and Toxic Substances [https://www.epa.gov/test-guidelines-pesticides-and-toxic-substances/about-test-guidelines-pesticides-and-toxic]
• 105. Alternatives to Animals in Testing | NAVS [https://navs.org/insight/alternatives-to-animals-in-testing/]
• 106. About ICCVAM - National Toxicology Program [https://ntp.niehs.nih.gov/whatwestudy/niceatm/iccvam]
• 107. Contributions to OECD activities - National Toxicology Program [https://ntp.niehs.nih.gov/iccvamreport/2023/leadership/intl/01-oecd]
• 108. International Cooperation on Alternative Test Methods [https://ntp.niehs.nih.gov/whatwestudy/niceatm/iccvam/international-partnerships/icatm]
• 109. and Validity of measurement instruments used in research reliability [https://pubmed.ncbi.nlm.nih.gov/19020196/]
• 110. and Reliability of Measurement – Research Methods in... Validity [https://opentextbc.ca/researchmethods/chapter/reliability-and-validity-of-measurement/]
• 111. Dose-response assessment of neuroactive botanical ... [https://pubmed.ncbi.nlm.nih.gov/41204423/]
• 112. The environmental neuroactive chemicals list of prioritized ... [https://www.sciencedirect.com/science/article/pii/S0013935124024411]