An AI-assisted designed supramolecularly engineered nanoplatform reverses pigmentation by triggering an ineffective compensatory melanin production program

The clinical applications of natural compounds are limited by their inherent physicochemical properties. This study reports a hierarchical supramolecular engineering strategy for constructing a dual-assembly nanosystem for the treatment of skin hyperpigmentation. Using an AI-assisted computational screening model, tranexamic acid was identified as a suitable molecular partner of the hydrophobic and active, baicalin. Subsequent dual assembly processes yielded a stable hybrid nanoplatform (DHBTC) that enhanced the solubility and delivery efficiency of baicalin. Single-cell transcriptomics revealed an unexpected mechanism of "functional inhibition"; despite the depigmenting efficacy of DHBTC, the melanogenesis-related gene network in melanocytes was upregulated. This was identified as compensatory transcriptional feedback triggered by drug-induced autophagy. DHBTC functionally inhibits pigment accumulation by accelerating melanosome degradation, which elicits ineffective transcriptional activation as the cell attempts to restore homeostasis. Furthermore, the platform remodeled the cutaneous immune microenvironment toward an anti-inflammatory state. This study presents a strategy for designing drug delivery systems, from computational prediction to supramolecular assembly, and describes a therapeutic mechanism based on the modulation of post-translational and organellar homeostasis.