Guinle_2026_Mar.Pollut.Bull_233_120117

Per- and polyfluoroalkyl substances (PFAS) are persistent contaminants of emerging concern known to induce various sublethal effects in aquatic organisms. This study explored the potential of near-infrared spectroscopy (NIRS) as a rapid, non-destructive method for assessing PFAS-induced biochemical alterations in blue mussels (Mytilus spp.). Mussels were exposed for 28 days to three experimental conditions: control, dietary PFAS exposure, and combined dietary and aqueous PFAS exposure. NIR spectra were acquired on 90 samples of gill and mantle tissues, and biological responses were assessed using conventional biochemical biomarkers, including protein and lipid hydroperoxide (LOOH) concentrations, and activities of superoxide dismutase (SOD), glutathione-S-transferase (GST), and acetylcholinesterase (AChE). Classification models combining principal component analysis and discriminant analysis achieved high accuracies (81-91%) in discriminating between exposed and non-exposed mussels. Partial least square regression models demonstrated that NIRS could accurately quantify protein contents (R(2) = 0.86, RPD = 2.45) and provide semi-quantitative predictions of SOD activity (R(2) = 0.73, RPD = 1.83). However, model's predictive performances were limited for LOOH concentrations (R(2) = 0.47, RPD = 1.24), GST activity (R(2) = 0.23, RPD = 1.09), and AChE activity (R(2) = 0.05, RPD = 1.04). Overall, these findings provide proof-of-concept that NIRS can capture contaminant-induced molecular changes in mussel tissues, offering a promising alternative method for diagnosing the general physiological status of individuals. Future work, including further experimental studies with various contamination scenarios and field-based validation studies, could enable the deployment of NIR-based tools for rapid and environmentally friendly assessment of contaminant impacts in bivalve biomonitoring programs.