Identification of molecular targets of Hypericum perforatum in blood for major depressive disorder: a machine-learning pharmacological study

Abstract

Abstract Background Major depressive disorder (MDD) is one of the most common psychiatric disorders worldwide. Hypericum perforatum (HP) is a traditional herb that has been shown to have antidepressant effects, but its mechanism is unclear. This study aims to identify the molecular targets of HP for the treatment of MDD. Methods We performed differential analysis and weighted gene co-expression network analysis (WGCNA) with blood mRNA expression cohort of MDD and healthy control to identify DEGs and significant module genes (gene list 1). Three databases, CTD, DisGeNET, and GeneCards, were used to retrieve MDD-related gene intersections to obtain MDD-predicted targets (gene list 2). The validated targets were retrieved from the TCMSP database (gene list 3). Based on these three gene lists, 13 key pathways were identified. The PPI network was constructed by extracting the intersection of genes and HP-validated targets on all key pathways. Key therapeutic targets were obtained using MCODE and machine learning (LASSO, SVM-RFE). Clinical diagnostic assessments (Nomogram, Correlation, Intergroup expression), and gene set enrichment analysis (GSEA) were performed for the key targets. In addition, immune cell analysis was performed on the blood mRNA expression cohort of MDD to explore the association between the key targets and immune cells. Finally, molecular docking prediction was performed for the targets of HP active ingredients on MDD. Results Differential expression analysis and WGCNA module analysis yielded 933 potential targets for MDD. Three disease databases were intersected with 982 MDD-predicted targets. The TCMSP retrieved 275 valid targets for HP. Separate enrichment analysis intersected 13 key pathways. Five key targets (AKT1, MAPK1, MYC, EGF, HSP90AA1) were finally screened based on all enriched genes and HP valid targets. Combined with the signaling pathway and immune cell analysis suggested the effect of peripheral immunity on MDD and the important role of neutrophils in immune inflammation. Finally, the binding of HP active ingredients (quercetin, kaempferol, and luteolin) and all 5 key targets were predicted based on molecular docking. Conclusions The active constituents of Hypericum perforatum can act on MDD and key targets and pathways of this action were identified.

Keywords

• Hypericum perforatum
• Major depressive disorder
• Machine learning
• Pharmacological targets