Abstract: Particulate matter (PM) presents numerous chemical and physical properties, such as chemical composition, size, and surface area, that may induce several adverse effects on human health. For instance, fine particles (PM2.5) can oxidize target molecules and result in oxidative stress and (systemic) inflammation. The oxidative potential (OP) quantifies particle reactivity that is expected to correlate with biological responses induced by exposure to PM. The Metropolitan Area of São Paulo (MASP), Southeastern Brazil, is the fourth largest urban area in terms of population. Little is known about the toxicological properties of PM2.5 from MASP. Therefore, we evaluated the OP, chemical composition, and emission sources of PM2.5 collected during an extensive sampling campaign (from May 2021 to August 2022) at the Faculty of Medicine of the University of São Paulo (FMUSP), located in an area highly impacted by vehicular emissions. We sampled PM2.5 using a mobile laboratory at the FMUSP equipped with a Partisol 2025i Sequential Air Sampler. We analyzed the OP using electron spin resonance (ESR) and dithiothreitol (DTT) assays, while the chemical composition was measured via energy-dispersive X-ray fluorescence (EDXRF). The primary sources identified with Principal Component Analysis (PCA) were vehicular emissions, resuspension, and biomass burning (factor 1), engine oil burning (factor 2), sea salt (factor 3), vehicle exhaust and non-exhaust (brake and tire wear) emissions (factor 4). These factors explained 46, 16, 11, and 27% of the PCA variance. Our results show that PM collected in April and May of 2022 exhibited the highest monthly mean volume-based OP values for ESR (OP-ESRv: 35.9×103 and 39.3×103 a.u. m-3 , respectively) and DTT (OP-DTTv: 3.02 and 3.75 nmol min-1 m-3 ), despite the highest PM2.5 concentrations observed during the winter of 2021 (18.42 and 21.07 μg m-3 in July and August, respectively). Furthermore, there is no characteristic seasonal profile for OP-ESRv and OP-DTTv, as is observed for PM2.5, which presents higher values in winter due to reduced precipitation rates and stable atmospheric conditions. Thus, the composition and emission sources of PM2.5, rather than its concentration, significantly influence OP reactions. Regarding the days of the week, Sundays presented lower values for PM2.5 (13.25 μg m-3 ), OP-ESRv (13.7×103 a.u. m-3 ), and OP-DTTv (1.64 nmol min-1 m-3 ) compared to Monday to Saturday, attributed to a 70% reduction in vehicle traffic in the region. The lower emission levels observed on Sundays possibly influence PM2.5 atmospheric reactions and OP responses. PM2.5 at the FMUSP is mainly composed of OC > EC > S > Na > K > Fe > Si > Al > Zn. Applying a Generalized Linear Model (GLM) with gamma distribution showed that different PM2.5 species are statistically significant (p-value < 0.05) for OP-ESR and OP-DTT, such as OC, EC, Na, Mg, Cl, K, Fe, Cu, Br, Sr, Cd, and Sb. Therefore, OP is mainly influenced by PM2.5 composition and its emission sources. Moreover, the weak correlation between OP-ESR, OP-DTT, and PM2.5 mass indicates that the oxidative potential assays could be used as alternative metrics in observational studies.

Keywords: Oxidative stress, particulate matter pollution, health effects, chemical characterization, emission sources.