Integrated high-content bioactivity profiling for rapid annotation of neuroprotective constituents in Curcumae Radix

Curcumae Radix (CR), the roots derived from multiple Curcuma species, has long been used for treating neurological disorders. However, morphological similarity and chemical heterogeneity complicate efficacy evaluation, and the systematic elucidation of its neuroprotective constituents remains limited.

To establish an integrated strategy combining high-content bioactivity profiling with untargeted chemical characterization for the systematic identification of neuroprotective compounds from four CR species (C. wenyujin, CWR; C. longa, CLR; C. kwangsiensis, CKR; C. phaeocaulis, CPR).

A ponatinib-induced zebrafish ischemia model evaluated extract efficacy. Extracts were fractionated by preparative liquid chromatography to construct a component library and characterized using UHPLC-QTOF MS. Multiparametric phenotypes in OGD/R-injured PC12 cells were captured using high-content imaging (HCI). A Phenotypic Feature Mapping (PFM) approach integrated m/z features with subcellular phenotypes, generating activity-guided "m/z-phenotype" fingerprints. Activity Scores (AS) and Cluster Scores (CS) quantified feature-phenotype associations for annotating bioactive compounds and inferring mechanisms.

CR extracts reduced thrombus formation, behavioral deficits, and inflammation, with CLR showing the strongest efficacy. Based on the calculated (AS) and CS, 582, 884, 835, and 730 candidate active m/z features for CWR, CPR, CLR, and CKR were screened, respectively. Finally, a total of 69 compounds were identified. Sesquiterpenoids dominated CWR, CPR, and CKR, whereas curcuminoids were core constituents in CLR. Subcellular analysis revealed species-specific regulatory preferences: nuclear morphology (CWR), mitochondrial structure (CPR, CKR), and cytoskeletal integrity (CLR).

The untargeted metabolomics integrated high-content screening strategy enables the rapid identification of neuroprotective compounds from four chemically distinct CR. This approach provides a scalable framework for identifying bioactive compounds and functionally annotating them in complex herbal systems.