Using artificial neural network for design and development of PVA/chitosan/starch/heparinized nZnO hydrogels for enhanced wound healing

Abstract

A study of many individual parameters is required to provide a robust investigation of a system in biomedical applications. A comprehensive understanding of these parameters is achievable by an appropriate experimental model as a valid description to predict the outputs (responses). A combination of response surface methodology and artificial neural network has been employed to design hydrogel dressings including polyvinyl alcohol, chitosan, and starch. The optimal ratio of components of hydrogels as input data based on the water vapor transmission rate, gel content, swelling ratio, and porosity properties as output parameters was determined using the quick propagation algorithm. Zinc-oxide nanoparticles were coupled with heparin and applied in the optimal formulation to investigate its effect on physical and mechanical properties, cytotoxicity, and antibacterial activities as well as in vivo wound healing. Mechanical strength improved in the presence of zinc-oxide nanoparticles. Heparin release reached with the saturation state in a longer period after conjugation onto zinc-oxide nanoparticles. Minimum inhibitory concentration decreased significantly by conjugation of heparin to the nanoparticles and current systems could protect wounds against infections. In vivo wound healing and immunohistochemistry assay indicated accelerated wound closure, re-epithelialization, and skin regeneration for hydrogel dressings containing heparin functionalized zinc-oxide nanoparticles.