Deep insight into PEGylation of bioadhesive chitosan nanoparticles: sensitivity study for the key parameters through artificial neural network model

Abstract

lonically cross-linked chitosan nanoparticles have great potential in nanomedicine due to their tunable properties and cationic nature. However, low solubility of chitosan severely limits their potential clinical translation. PEGylation is a well-known method to increase solubility of chitosan and chitosan nanoparticles in neutral media; however, effect of PEG chain length and chitosan/PEG ratio on particle size and zeta potential of nanoparticles are not known. This study presents a systematic analysis of the effect of PEG chain length and chitosan/PEG ratio on size and zeta potential of nanoparticles. We prepared PEGylated chitosan chains prior to the nanoparticle synthesis with different PEG chain lengths and chitosan/PEG ratios. To precisely estimate the influence of critical parameters on size and zeta potential of nanoparticles, we both developed an artificial neural network (ANN) model and performed experimental characterization using the three independent input variables: (i) PEG chain length, (ii) chitosan/PEG ratio, and (iii) pH of solution. We studied the influence of PEG chain lengths of 2, 5, and 10 kDa and three different chitosan/PEG ratios (25 mg chitosan to 4, 12, and 20 mu moles of PEG) for the synthesis of chitosan nanoparticles within the pH range of 6.0-7.4. Artificial neural networks is a modeling tool used in nanomedicine to optimize and estimate inherent properties of the system. Inherent properties of a nanoparticle system such as size and zeta potential can be estimated based on previous experiment results, thus, nanoparticles with desired properties can be obtained using an ANN. With the ANN model, we were able to predict the size and zeta potential of nanoparticles under different experimental conditions and further confirmed the cell-nanoparticle adhesion behavior through experiments. Nanoparticle groups that had higher zeta potentials promoted adhesion of HEK293-T cells to nanoparticle-coated surfaces in cell culture medium, which was predicted through ANN model prior to experiments. Overall, this study comprehensively presents the PEGylation of chitosan, synthesis of PEGylated chitosan nanoparticles, utilizes ANN model as a tool to predict important properties such as size and zeta potential, and further captures the adhesion behavior of cells on surfaces prepared with these engineered nanoparticles.