Classification models for identifying skin sensitizers using in vitro alternatives to animal testing
Citation & Export
Hide
Simple citation
Lee, Serom.
Classification models for identifying skin sensitizers using in vitro alternatives to animal testing. Retrieved from
https://doi.org/doi:10.7282/T3WM1C26
Export
Description
TitleClassification models for identifying skin sensitizers using in vitro alternatives to animal testing
Date Created2014
Other Date2014-10 (degree)
Extent1 online resource (xiv, 109 p. : ill.)
DescriptionAllergic contact dermatitis (ACD) is an inflammatory disease that occurs when chemicals known as sensitizers come in contact with the skin. Recent European legislation prohibits animal based screens of cosmetic ingredients. Current alternatives to animal testing are limited by their poor ability to identify a subset of non-innate contact sensitizers known as pre-/pro-haptens which require transformation in the skin. Furthermore, these approaches only evaluate a single cell type with 1 or 2 biomarkers. To address this, we performed an initial study using RealSkin, a full thickness skin equivalent, in co-culture with MUTZ-3 derived Langerhan’s cells (MUTZ-LCs). This co-culture was treated with model pro-/pro-haptens from an irritant control and multiple cellular metrics were evaluated. A novel feature selection method was developed using a support vector machine (SVM) to rank the margin distances of each metric and identify biomarkers of sensitization. A panel (IL-12, IL-9, VEGF, IFN-γ) was identified by SVM and predicted sensitizers with over 90% accuracy. Although promising, this method is costly and resource intensive. Thus, we designed a more economic, high throughput screening approach to metabolize pro-hapten sensitizers. MUTZ-LCs were cultured alone and in parallel with a co-culture of HaCaT keratinocytes, dermal fibroblasts, and MUTZ-LCs. Both cultures were treated with a panel of pre- and pro-hapten sensitizers and non-sensitizers. The secretome of both cultures were evaluated for 27 cytokines, chemokines, and growth factors. Feature selection by SVM identified predictive signatures of sensitization for each culture type. These cellular metrics was used to develop a classification model of sensitization. The MUTZ-LCs classification model was 83.3% accurate at identifying pro-hapten sensitizers using MIP-1β, MIP-1α, RANTES, IL-8, and IL-9. The co-culture classification model was 89.6% accurate at identifying pro-hapten sensitizers using a panel of IL-8, GM-CSF, and RANTES. The presence of the keratinocytes and fibroblasts enhanced the identification of pre- and pro-haptens to sensitize the MUTZ-LCs. This approach also preserves the cross-talk signals between all three skin cell types. Thus, the co-culture of HaCaT keratinocytes, dermal fibroblasts, and MUTZ-LCs is an attractive, high throughput in vitro alternative to animal testing for the identification of pre- and pro-hapten skin sensitizers.
NotePh.D.
NoteIncludes bibliographical references
Noteby Serom Lee
Genretheses, ETD doctoral
Languageeng
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.