Androgenetic alopecia (AGA) is a common form of hair loss, which is mainly caused by oxidative stress induced dysregulation of hair follicles (HF). Herein, a highly efficient manganese thiophosphite (MnPS3) based superoxide dismutase (SOD) mimic was discovered using machine learning (ML) tools. Remarkably, the IC50 of MnPS3 is 3.61 μg·mL-1, up to 12-fold lower than most reported SOD-like nanozymes. Moreover, a MnPS3 microneedle patch (MnMNP) was constructed to treat AGA that could diffuse into the deep skin where HFs exist and remove excess reactive oxygen species. Compared with the widely used minoxidil, MnMNP exhibits higher ability on hair regeneration, even at a reduced frequency of application. This study not only provides a general guideline for the accelerated discovery of SOD-like nanozymes by ML techniques, but also shows a great potential as a next generation approach for rational design of nanozymes.
Keywords: androgenetic alopecia; machine learning; nanozyme; superoxide dismutase; transition-metal thiophosphate.