Yangkun Zhou, Meng Zhan, Ying Zhang, Tingting Li, Hongfeng Quan, Lin Dong
Stellariae Radix is a representative medicinal herb that is used clinically for treating yin deficiency fever, bone-steaming fever, and infantile malnutrition fever, but its quality control has not been fully investigated; however, this technique relies solely on single evaluation indicators of pharmacopeia and noncharacteristic ingredients. This study aimed to predict potential quality markers (Q-markers) of Stellariae Radix to establish a scientific and reasonable quality evaluation system. Our study established an integrated strategy that progressed from chemical characterization using ultra-performance liquid chromatography quadrupole time-of-flight mass spectrometry (UPLC-Q-TOF/MS) to bioactivity-guided screening via Pearson correlation analysis and culminated in the final selection of the most relevant quality markers by modeling their nonlinear relationships with efficacy using a back-propagation neural network (BP-NN) algorithm. First, the chemical constituents of Stellariae Radix were characterized via UPLC-Q-TOF-MS. By integrating a deficiency-heat syndrome rat model and serum pharmacochemistry-metabolomics, the absorbed constituents, metabolites and related biomarkers of Stellariae Radix were further elucidated. Subsequently, Pearson correlation analysis between the differentially expressed metabolites and blood-absorbed components revealed eight bioactive constituents linked to specific biomarkers. The five candidate Q-markers, namely, curcumenol, stellarine C, dichotomine B, dichotomine A, and wogonin, were selected via BP-NN analysis on the basis of their correlation with heat-clearing efficacy. Overall, the integrated approach combining UPLC-Q-TOF/MS-based metabolomics with Pearson correlation analysis and BP-NN is an effective tool for investigating the efficacy of Stellariae Radix, and these five components could be considered potential quality markers.