Abstract
Nanoparticles (NPs) capable of transporting and releasing therapeutic agents to target tissues constitute one of the most exciting areas in nanomedicine, especially magnetic mesoporous silica nanoparticles (M-MSN). M-MSNs may be addressed to tumors thanks to their magnetism and can act as drug carriers thanks to their high specific surface area. Nevertheless, the safety of these NPs with decorations, conferring them specific properties, must be assessed in order to avoid harmful effects on healthy tissues, in particular on the liver, the organ of xenobiotics metabolism.The goal of this thesis was therefore to evaluate the potential toxicity of M-MSN either pristine, or coated with polyethylene glycol (PEG), or surrounded by a lipid bilayer of 1,2-dimyristoyl-sn-glycero-3- Phosphocholine (DMPC). To this end, the human hepatic cell model HepaRG was chosen to realize in vitro toxicity testing and to elucidate the intracellular mode of action of these various NPs.The physico-chemical properties of pristine and covered M-MSNs were measured using different techniques such as dynamic light scattering (DLS), transmission electron microscopy (TEM) and atomic force microscopy (AFM). NPs toxicity was first evaluated by viability testing and real-time cell impedance analysis (xCELLigence).Gene expression profiles were then performed through very high density oligo microarrays (8x60k, Agilent) to evaluate, in a dose- and time-dependent manner, the toxicity of these NPs. In addition, the use of an original methodology for comparative analysis of large biological data allowed us to demonstrate the molecular mechanisms triggered by the NPs in the hepatocytes. We were able to determine the dose not triggering any toxicity as well as the dose inducing a slight transient toxicity after 24h. We thus defined this latter value as a threshold of biocompatibility with HepaRG cells. We also showed by TEM a slower uptake of PEGylated NPs by cells as well as their delayed effects on the transcriptome compared to the pristine and DMPC NPs. Nevertheless, a dose of 80 μg/cm² of pristine or covered M-MSNs triggers the chain of events of the hepatic cholestasis AOP (Adverse Outcome Pathway). This result demonstrates that this methodology is suitable for predictive toxicology by analysis of cellular biological responses after exposure to exogenous substances.Furthermore, NPs tend to be covered with proteins in the presence of serum (corona). Cell impedance analysis shows that M-MSNs surrounded by human or bovine serum proteins coronas do not trigger the same toxicity on human cells. This result raises the problem of a potential overestimation of NPs toxicity to human cells in in vitro testing by using fetal bovine serum in culture media.We undertook a dynamic analysis (between 30 s and 7 days) of the corona formation by tandem mass spectrometry has highlighted three groups of protein with distinct behaviors. The first cluster contains some abundant proteins that desorb over time, the second cluster comprises some protein families such as apolipoproteins, and the third cluster contains late enrichment proteins attracted by other proteins already present in the corona. A dynamic network of protein-protein interactions inside the corona, namely the interactome, was built from the data. This work opens the way to a possible control of the corona in order to provide the nanocarriers with stealth properties allowing them to reach target organs without being opsonized.These techniques used during this thesis and based on analyses of biological big data might be part of the future standards on nanosafety evaluation.