AI‐Guided Design of pH‐ and Redox‐Responsive Lignin@GO@ZIF‐8 Nanocarriers for Targeted Co‐Delivery of 5‐Fluorouracil and Metformin in Prostate Cancer Therapy

Abstract

Prostate cancer therapy is limited by systemic toxicity and inefficient tumor-selective delivery. Here we report a multi-stimuli-responsive nanocomposite, Lignin@GO@ZIF-8, that co-delivers 5-Fluorouracil (5-FU) and metformin and couples pH- and redox-responsive release. We integrate machine learning (ML) to guide formulation: trained on 500 formulation property pairs with cross-validation and a held-out test set, XGBoost achieved the highest predictive performance (R² = 0.86-0.89) for drug loading, encapsulation efficiency, and release rate, with influential features including the ZIF-8 fraction, graphene oxide (GO) content, particle size, zeta potential, pH, and glutathione concentration. ML-optimized Lignin@GO@ZIF-8 exhibited improved loading and tunable release relative to pre-optimized controls. In vitro, the platform sustained drug release and produced potent anticancer activity, reducing LNCaP viability to 18 ± 3% at 72 h, while showing low cytotoxicity toward non-malignant cells. These results support a data-driven framework for rational design of multifunctional, stimuli-responsive nanomedicine platforms and may help inform future translational strategies for prostate cancer.