機械学習によるPANoptosis阻害カーボン化ポリマーの構築：化学療法誘発性腎毒性の軽減に向けた可能性

Chemotherapy-induced nephrotoxicity, particularly acute kidney injury (AKI), is a severe complication that is driven by multiple regulated cell death (RCD) pathways. However, current nephroprotective strategies predominantly focus on apoptosis inhibition by relieving oxidative stress. PANoptosis is a newly discovered form of RCD pathway incorporating key features of pyroptosis, apoptosis, and necroptosis, providing a potential target for synergistic multi-mechanistic nephroprotection. Herein, the enhanced sampling method is introduced for molecular dynamics based on trained machine learning force fields to guide the construction of bioactive carbonized polymer dots (Lu-CDs) with specific pharmacophoric moieties of the flavonoid compound. Unlike conventional carbonized polymer dots with solely antioxidative activity, Lu-CDs demonstrate both radical scavenging and PANoptosis-inhibiting activities. In chemodrug-induced AKI mice, a very low dose of Lu-CDs can elicit prolonged renal accumulation and superior nephroprotective efficacy compared to the small-molecule flavonoid compound and the antioxidative carbonized polymer dots. The findings of this study highlight a design strategy for bioactive nanodrugs based on the machine learning force fields for molecular dynamics and propose PANoptosis inhibition as a promising approach to mitigate chemodrug-induced nephrotoxicity.