Parameters Influencing Gene Delivery Efficiency of PEGylated Chitosan Nanoparticles: Experimental and Modeling Approach

• Published in: Advanced NanoBiomed Research (Online) 2022-01, Vol.2 (1), p.n/a
• Main Authors: Dogan, Nihal Olcay, Bozuyuk, Ugur, Erkoc, Pelin, Karacakol, Alp Can, Cingoz, Ahmet, Seker-Polat, Fidan, Nazeer, Muhammad Anwaar, Sitti, Metin, Bagci-Onder, Tugba, Kizilel, Seda
• Format: Article
• Language: English
• Subjects: Artificial neural networks; Biocompatibility; Chitosan; Data points; Datasets; Efficiency; Fluorescence; Gene therapy; Gene transfer; Genetic engineering; Green fluorescent protein; Independent variables; Mathematical models; Molecular weight; Nanoparticles; Nanotechnology; Neural networks; Parameters; PEGylated chitosan nanoparticles; plasmid DNA; Plasmids; Polyethylene glycol; Polymers; Scanning electron microscopy; Sodium triphosphate; Sodium tripolyphosphate; Toxicity; Transfection
• ISSN: 2699-9307; 2699-9307
• DOI: 10.1002/anbr.202100033

Experimentation of nanomedicine is labor‐intensive, time‐consuming, and requires costly laboratory consumables. Constructing a reliable mathematical model for such systems is also challenging due to the difficulties in gathering a sufficient number of data points. Artificial neural networks (ANNs) are indicated as an efficient approach in nanomedicine to investigate the cause‐effect relationships and predict output variables. Herein, an ANN is adapted into plasmid DNA (pDNA) encapsulated and PEGylated chitosan nanoparticles cross‐linked with sodium tripolyphosphate (TPP) to investigate the effects of critical parameters on the transfection efficiencies of nanoparticles. The ANN model is developed based on experimental results with three independent input variables: 1) polyethylene glycol (PEG) molecular weight, 2) PEG concentration, and 3) nanoparticle concentration, along with one output variable as a percentage of green fluorescent protein (GFP) expression, which refers to transfection efficiency. The constructed model is further validated with the leave‐p‐out cross‐validation method. The results indicate that the developed model has good prediction capability and is influential in capturing the transfection efficiencies of different nanoparticle groups. Overall, this study reveals that the ANN could be an efficient tool for nanoparticle‐mediated gene delivery systems to investigate the impacts of critical parameters in detail with reduced experimental effort and cost. The plasmid DNA (pDNA) encapsulated and PEGylated chitosan nanoparticle (CS–PEG–pDNA NP) system is adapted into an artificial neural network (ANN) model to indicate the impacts of critical parameters for such an NP‐mediated gene delivery system which could be accurately predicted with minimal experimental effort, resources, cost, and time.