Raman Spectroscopic Biodosimetry Using Protein in Murine Hair.

Abstract

Hair is an attractive sample for determining exposure to ionizing radiation due to its non-invasive nature. A biological tissue comprised mainly of keratin protein, hair is susceptible to oxidative or reductive stress by direct or indirect damage mechanisms. In this report, changes observed in the Raman spectra associated with hair protein from ionizing radiation were assessed for biodosimetry. Raman spectra were obtained from the hairs of a mixed sex cohort of irradiated C57BL/6 mice (N = 32 total) with doses of gamma rays ranging from 0-4 Gy. Radiation-dependent changes in the Raman spectra of the hairs provided molecular-specific signals that can inform about the damage mechanism. Partial least-squares discriminant analysis (PLS-DA) models incorporating automated variable selection for each sex showed classification of controls or exposed at 80% accuracy based on cross-validation. Models show only slight differences in performance based on the mouse's sex from which the sample originated. This slight difference is consistent with PLS-DA models that show marginal cross-validation sensitivity (∼60%) in predicting the sex of the mouse from the Raman hair spectrum. Utilizing PLS regression, a dose-response model including both sexes showed root-mean-squared error (RMSE) ±1 Gy. The ability to determine dose or exposure from plucked hair with Raman spectroscopy would provide a needed tool for rapid medical triage after unexpected exposure.