Defined Approaches to Identify Potential Skin Sensitizers

Evaluation of Substances of Regulatory Interest Using Non-Animal Skin Sensitization Test Methods (May 2025)

DOI for Publication: https://doi.org/10.22427/NICEATM-05
DOI for CEBS Data Page: https://doi.org/10.22427/NICEATM-DATA-NICEATM-05

This NICEATM report evaluates the performance of individual in chemico, in vitro, and in silico methods for predicting skin sensitization hazard or potency. It compares predictions derived from testing with each individual method to predictions from the murine local lymph node assay (LLNA) and three defined approaches (DAs): the Key Event 3/1 Sequential Testing Strategy (Nukada et al. 2013), the 2 out of 3 Strategy (Bauch et al. 2012; Urbisch et al. 2015) and the Integrated Testing Strategy (Takenouchi et al. 2015). Predictions from the DAs were also evaluated for concordance with one another and with human and LLNA reference data.

Data from Appendix A: In Vitro Results, In Silico Data, Physicochemical Data, In Vivo Reference Data, and Defined Approach Results

Skin sensitizers are substances with the potential to cause allergic contact dermatitis. Allergic contact dermatitis is often initiated through covalent binding of substances to skin proteins.

Potential skin sensitizers can be identified without animal testing using DAs for testing and assessment. A DA consists of:

Input data generated from identified methods.
A data interpretation procedure that is used to translate data produced from a defined set of information sources into a prediction.

An adverse outcome pathway can provide a framework for the development of a DA. An adverse outcome pathway is a model that identifies the sequence of biological events required to produce a toxic effect when an organism is exposed to a substance. In vitro methods can be used to observe or measure the biological events and predict whether a substance will produce a toxic effect.

The events that result in allergic contact dermatitis have been described in an adverse outcome pathway for skin sensitization. NICEATM, working with the NIEHS Division of Translational Toxicology (DTT), ICCVAM, and industry scientists, created and evaluated DAs to identify potential skin sensitizers based on this adverse outcome pathway.

Regulatory Applications of Defined Approaches

International Regulatory Acceptance

In June 2021, OECD issued Guideline 497, Defined Approaches on Skin Sensitisation. Drafted and sponsored by NICEATM and international partners, Guideline 497 is the first internationally harmonized guideline to describe a non-animal approach that can be used to replace an animal test to identify skin sensitizers. Guideline 497 was last updated in 2025 to include new information sources for existing DAs and new DAs for quantitative risk assessment. NICEATM developed the DASS App, a user-friendly web app that uses the DAs from OECD Guideline 497 to predict skin sensitization hazard and potency (To et al. 2024).

The DAs described in Guideline 497 were evaluated using primarily chemicals relevant to the cosmetics industry. NICEATM evaluated performance of the Guideline 497 DAs using 185 substances nominated by DTT, EPA, and CPSC. Results of the study are provided in the May 2025 technical report, "Evaluation of Substances of Regulatory Interest Using Non-Animal Skin Sensitization Test Methods".

U.S. Regulatory Acceptance

FDA finalized guidance in June 2023 that states that the agency 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 skin sensitization with an accuracy similar to existing in vivo methods.

In April 2018, EPA released a draft science policy to reduce animal use by using DAs to identify potential skin sensitizers. The draft policy is the result of national and international collaboration among ICCVAM, NICEATM, Cosmetics Europe, the European Union Reference Laboratory for Alternatives to Animal Testing, and Health Canada's Pest Management Regulatory Agency.

Defined Approach for Non-animal Risk Assessment for Isothiazolinones

In May 2020, EPA requested comment on draft human health and ecological risk assessments for a group of antimicrobial chemicals known as isothiazolinones. Isothiazolinones are used in a variety of products including plastics, household cleaners, and laundry detergents and frequently cause skin sensitization.

EPA requested comment on the use of an in vitro and artificial neural network-based DA instead of laboratory animal data to evaluate skin sensitization risks for products containing isothiazolinones. This is the first use of this type of DA in regulatory risk assessment.

EPA Draft Document: Hazard Characterization of Isothiazolinones in Support of FIFRA Registration Review (April 2020)

The draft risk assessments rely heavily on work done by NTP, NICEATM, and ICCVAM. NTP conducted the in vitro testing. NICEATM analyzed the in vitro data and ran the artificial neural network-based DA for determining points of departure. The ICCVAM Skin Sensitization Expert Group reviewed the NTP testing report and the NICEATM analyses before data were provided to EPA for development of the risk assessments. Strickland et al. (2022) describes the application of DAs to potency classification of isothiozolinones.

NICEATM Report

Application of Non-animal Test Methods and Defined Approaches to Skin Sensitization Assessment of Isothiazolinone Compounds (June 2020)

Report sections:

Main body
Appendix A: Burleson Research Technologies, Inc. In Vitro Testing Results
Appendix B: Institute for In Vitro Sciences, Inc. h-CLAT Testing Results
Appendix C: LLNA Data
Appendix D: In Vivo, In Vitro, In Silico, and Defined Approach Results
Appendix E: Calculation of Weighted LLNA EC3 for CMIT/MIT

Defined Approaches Developed by NICEATM and Partners

DAs developed by NICEATM and ICCVAM use non-animal data to predict skin sensitization hazard and potency. These DAs combine data from the following sources:

Direct peptide reactivity assay (DPRA).
KeratinoSens assay.
Human cell line activation test (h-CLAT).
Read-across predictions generated by the QSAR Toolbox software package.
Physical properties (e.g., partition coefficients).

The DAs are described in papers published in 2016 and 2017:

Strickland et al. 2016 describes prediction of murine local lymph node assay outcomes.
Strickland et al. 2017 uses data from human exposures to predict human skin sensitization hazard.
Zang et al. 2017 describes the prediction of human or animal skin sensitization potency. These predictions enable the classification of skin sensitizers as "weak" or "strong" without animal tests.

To expand the potential applicability of these DAs to a broader range of chemical types, ICCVAM agencies nominated over 200 chemicals for additional testing in the DPRA, KeratinoSens, and h-CLAT tests. Chemicals tested included pesticide ingredients and formulations, industrial chemicals, and personal care product ingredients. The NIEHS Division of Translational Toxicology (DTT) conducted this testing, the results of which are described in a 2025 report. Strickland et al. (2022) describes the application of DAs to potency classification of isothiozolinones, which are antimicrobial compounds that are widely used in consumer products.

NICEATM and Cosmetics Europe collaborated to evaluate multiple DAs for skin sensitization safety assessment that had been submitted to OECD as case studies. The collaboration produced two publications:

Hoffman et al. 2018 describes a database including data from human, animal, and five non-animal tests for 128 substances. The substances in the database have a variety of chemical structures and use categories. The database is a proposed point of reference for the evaluation and development of new non-animal approaches to skin sensitization safety assessment.
Kleinstreuer et al. 2018 describes an analysis of multiple defined approaches for skin sensitization safety assessment of cosmetic ingredients using the database described above. Many of these approaches were found to perform as well or better than animal methods to predict human skin sensitization hazards.

The ICCVAM and Cosmetics Europe projects followed an earlier integrated testing strategy developed by NICEATM and collaborators (Pirone et al. 2014; Jaworska et al. 2015).