Skin sensitization quantitative QSAR models based on mechanistic structural alerts. (28th February 2022)

• Record Type: Journal Article
• Authors: Chayawan, ; Selvestrel, Gianluca; Baderna, Diego; Toma, Cosimo; Caballero Alfonso, Ana Yisel; Gamba, Alessio; Benfenati, Emilio
• Is Part Of: Toxicology. Volume 468(2022)
• Publication Date: 2022-02-28
• Subjects: Skin sensitization -- Local lymph node assay -- Allergic contact dermatitis -- Reactivity domains -- QSAR models; Toxicology -- Periodicals; Chemicals -- Physiological effect -- Periodicals; 615.9005
• DOI: 10.1016/j.tox.2022.153111
• Languages: English
• ISSNs: 0300-483X

Abstract

Highlights: Quantitative skin sensitization LLNA EC3 data of organic molecules was collected. Two-Dimensional descriptors were computed to model the experimental LLNA EC3 data. Multilinear regression models were developed for reactivity domain based categories. Chemicals with unknown and multiple reactivity alerts were identified and modelled. A total of eight QSAR models were proposed, for a full dataset of 366 chemicals. Abstract: Allergic contact dermatitis is increasingly of interest for the hazard characterization of chemicals. in vivo animal testing is usually adopted but in silic o approaches are becoming the new frontier due to their swiftness and economic efficiency. Indeed, in silico models can rationalise the experimental outcomes besides having predictive ability. The aim of the present work was to explore the electrophilic chemical behaviour responsible for allergic contact dermatitis using quantitative QSAR regression models. Eight models were proposed, using an experimental LLNA dataset of 366 chemicals. Each model is unique to encode a type of electrophilic reactivity domain. The models were obtained using autocorrelation, electro-topological and atom centered fragment based on two-dimensional descriptors, which incorporated the electronic and stereochemical features of substances interacting with skin proteins to induce skin cell proliferation. Finally, simple steps were proposed to integrate the eight models for the application on the test chemicals.