Machine-Learning-Assisted Exploration of High Entropy-Atom Nanozyme for Anti-Tumor Immunotherapy by Enhancing Enzyme Activity and Disrupting Dual Energy Metabolism

Abstract

Despite its potential in cancer therapy, single-atom nanozyme (SAzyme) faces challenges like low atomic loading and rapid cancer metabolism. Here, a high-entropy atom nanozyme (HEAzyme), PtNiBiSnSb-anti-CD36, is designed for efficient anti-tumor immunotherapy. The PtNiBiSnSb HEAzyme, incorporating five SAzyme, demonstrates enhanced peroxidase (POD)-like activity due to its abundant active sites, showing a 7.2-fold increase in catalytic efficiency compared to the structurally analogous PtBi nanozyme. Further integrating of density functional theory calculations and machine learning analysis reveals that the main reason is due to the addition of Ni, Sn, and Sb, which reduced the reaction energy barrier and changed the adsorption energy of surface hydroxyl groups. Moreover, PtNiBiSnSb-anti-CD36 not only inhibits normal metabolic processes by depleting nicotinamide adenine dinucleotide (NADH) through its enhanced NADH-POD-like activity, but also suppresses lipid metabolism by blocking CD36-mediated uptake. The dual metabolic inhibition facilitates the mild photothermal therapy induced by PtNiBiSnSb. In this process, the anti-tumor immune response is amplified by inducing immunogenic cell death and inhibiting the growth of immunosuppressive cells through blocking fatty acid uptake. The proposal of HEAzyme for cancer therapy provides a novel approach to enhance the catalytic activity of SAzymes and deepens the theoretical understanding of HEAzyme in the POD-like reaction process.