(P06-014-25) Unraveling Health Risks of Foodborne Titanium Dioxide and Silicon Dioxide Nanoparticles Across Various Populations by a Machine Learning-Based Approach
Professor University of Massachusetts Amherst Amherst, Massachusetts, United States
Objectives: Titanium dioxide (TiO2) and Silicon Dioxide (SiO2), commonly used as food additives (designated as E171 and E551, respectively), contain a significant proportion of nanoscale particles, raising concerns about their potential risks to human health. Accumulating evidence demonstrated that different populations have different susceptibilities to foodborne nanoparticles, such as obese vs. non-obese and males vs. females.
Methods: Herein, we determined the potential adverse effects of TiO2 and SiO2 NPs in non-obese and obese mice (both male and female mice). Mice (3-week-old) were fed either a low-fat diet or a high-fat diet containing TiO2 NPs (30 nm), SiO2 NPs (20 nm), E171, or E551 for 13 weeks.
Results: The study found that dietary TiO2 NPs, SiO2 NPs, E171, and E551 led to the dose-dependent accumulation of TiO2 and SiO2 in mouse organs, causing cellular damage and oxidative stress. Additionally, the NPs induced reproductive toxicity, reducing sperm count and oocyte number and altering gene expression linked to reproduction and oxidative stress. Obese mice exhibited higher levels of TiO2 and SiO2 accumulation and stronger adverse effects than non-obese mice, with female mice being more affected than males. The NPs exacerbated obesity-induced colon inflammation, reduced short-chain fatty acid levels in the gut, and caused gut microbiota dysbiosis, with more potent effects in obese mice. Innovatively, we constructed machine learning models to conduct an in-depth analysis of the relationships between gut microbiota profiles and NPs-induced outcomes, including tissue oxidative damage, chronic inflammatory responses, hormonal imbalances, and reproductive toxicity. The machine learning models confirmed that obese and female mice exhibit higher sensitivity to the effects of NPs. Additionally, Bifidobacterium, Lactobacillus, Ruminococcus, and Pseudomonas were identified as potential biomarkers for detecting nanoparticle toxicity in the gut. The NPs-induced alterations in gut microbiota were closely associated with sex hormones, potentially contributing to reproductive toxicity in mice.
Conclusions: The study presented the practical application of machine learning models in predicting nano-toxicity through gut microbiome profiles, underscoring the potential of computational methods in nano-safety evaluation.
