A Deep Learning Analysis Reveals Nitrogen-Doped Graphene Quantum Dots Damage Neurons of Nematode Caenorhabditis elegans

Abstract

Along with the rapidly increasing applications of nitrogen-doped graphene quantum dots (N-GQDs) in the field of biomedicine, the exposure of N-GQDs undoubtedly pose a risk to the health of human beings, especially in the nervous system. In view of the lack of data from in vivo studies, this study used the nematode Caenorhabditis elegans (C. elegans), which has become a valuable animal model in nanotoxicological studies due to its multiple advantages, to undertake a bio-safety assessment of N-GQDs in the nervous system with the assistance of a deep learning model. The findings suggested that accumulated N-GQDs in the nematodes’ bodies damaged their normal behavior in a dose- and time-dependent manner, and the impairments of the nervous system were obviously severe when the exposure dosages were above 100 μg/mL. When assessing the morphological changes of neurons caused by N-GQDs, a quantitative image-based analysis based on a deep neural network algorithm (YOLACT) was used because traditional image-based analysis is labor-intensive and limited to qualitative evaluation. The quantitative results indicated that N-GQDs damaged dopaminergic and glutamatergic neurons, which are involved in the neurotoxic effects of N-GQDs in the nematode C. elegans. This study not only suggests a fast and economic C. elegans model to undertake the risk assessment of nanomaterials in the nervous system, but also provides a valuable deep learning approach to quantitatively track subtle morphological changes of neurons at an unbiased level in a nanotoxicological study using C. elegans.