Per- and polyfluoro toxicity (LC(50) inhalation) study in rat and mouse using QSAR modeling

Area of Science:

• Environmental Chemistry
• Toxicology
• Computational Chemistry

Background:

• Per- and polyfluorinated chemicals (PFAS) are widespread environmental contaminants due to their use in various products.
• Concerns exist regarding the toxicity of both long-chain and short-chain PFAS, with limited experimental data available.
• Toxicity profiles vary significantly across different animal species.

Purpose of the Study:

• To develop quantitative structure-activity relationship (QSAR) models for predicting the inhalation toxicity (LC50) of short and long-chain PFAS.
• To model and predict PFAS toxicity in two rodent species: rat (Rattus) and mouse (Mus).
• To establish the structural applicability domain (AD) for the developed QSAR models.

Main Methods:

• Application of quantitative structure-activity relationship (QSAR) modeling for the first time to a combined set of short and long-chain PFAS.
• Utilized multiple linear regression (MLR) with ordinary-least-squares (OLS) and genetic algorithm (GA) for descriptor selection.
• Prepared distinct training and prediction sets to ensure model robustness and external predictivity.
• Verified the structural applicability domain (AD) using extensive PFAS data from databases and literature.

Main Results:

• Developed statistically robust and predictive QSAR models for PFAS toxicity in rats and mice.
• Identified key molecular descriptors influencing toxicity and discussed species-specific differences.
• Employed chemometric methods (PCA, MDS) to identify highly toxic compounds within the models' AD for experimental testing.

Conclusions:

• QSAR modeling provides a valuable tool for predicting PFAS toxicity, aiding in risk assessment.
• The study highlights species-specific variations in PFAS toxicity, necessitating tailored evaluations.
• Identified priority compounds for experimental validation under the EU CADASTER project.