Discovery of Herbal Remedies and Key Components for Major Depressive Disorder Through Biased Random Walk Analysis on a Multiscale Network

Abstract

Major depressive disorder (MDD) is a widespread psychiatric condition with substantial socioeconomic impacts, yet single-target pharmacotherapies often yield responses. To address its multifactorial nature, this study employed a multiscale network analysis of herbs, their active components, and MDD-associated protein targets. Using a biased random walk with restart, we calculated interactions between disease-related and herb-derived targets, identifying herbs highly correlated with MDD. Enrichment analysis further revealed key signaling pathways, including oxidative stress, neuroinflammation, and hormone metabolism, underlying these herbs' therapeutic effects. We identified Ephedrae herba, Glehniae radix, Euryales semen, and Campsitis flos as promising candidates, each containing multiple bioactive compounds (such as ephedrine, psoralen, xanthine, and ursolic acid) that modulate critical processes like oxidation-reduction, inflammatory cytokine regulation, and transcriptional control. Network visualization showed how these herbs collectively target both shared and distinct pathways, supporting a synergistic, multi-target therapeutic strategy. This approach underscores the significance of network-based methodologies in addressing complex disorders such as MDD, where focusing on a single target may overlook synergistic interactions. By integrating diverse molecular data, this study provides a systematic framework for identifying novel interventions. Future experimental validation will be crucial to confirm these predictions and facilitate the translation of findings into effective MDD therapies.

Keywords: active components; herbal medicine; major depressive disorder; multiscale network.

Figure 1

Schematics for identifying candidate herbs and active compounds for MDD. This schematic depicts the application of multiscale network analysis to identify herbs and their active compounds with potential efficacy against MDD. Herb–compound and compound–target interactions were mapped, followed by the identification of disease-associated proteins. Diffusion profiles for the herbs and disease-related proteins were computed and compared, leading to the prioritization of herbs based on their correlation scores. Enrichment analysis was performed to uncover pivotal biological pathways, while individual active components of the top-ranked herbs were analyzed to identify key protein targets. The lower panel illustrates the key mechanism of the selected herbs against MDD by visualizing edges between the herbs’ active components and the disease, with proteins and biological functions as intermediaries.

Figure 2

Herb–target interaction network of the top 10 candidate herbs with high correlation scores for MDD. Green hexagons represent herbs, and gray circles represent protein targets. Edges indicate interactions between herbs and targets, with the size of hexagons and circles reflecting interaction frequency.

Figure 3

Gene Ontology enrichment analysis for core protein targets. A Gene Ontology enrichment analysis was conducted on 55 key protein targets, categorized into three domains: biological processes (top), cellular components (middle), and molecular functions (bottom). The x-axis displays the adjusted p-value, representing the significance of the associations, while the bubble size indicates the odds ratio. The bubble color reveals the combined score, which indicates the statistical significance of each term. This visualization emphasizes the most significantly enriched terms within each category.

Figure 4

Active compounds of Ephedrae herba and their association with MDD. (A) The table lists five active compounds—ephedrine [17], pseudoephedrine, epicatechin [18], tetramethylpyrazine [19], and tyrosine—along with each compound’s correlation score and protein overlap p-value in relation to MDD. (B) The network diagram illustrates how these compounds (blue diamond nodes, left) connect to their target proteins and the relevant biological processes (gray and purple, center), ultimately linking to the MDD node (red hexagon, right). Edges represent interactions or functional relationships, highlighting potential mechanisms through which Ephedrae herba may modulate MDD pathophysiology.

Figure 5

Active compounds of Glehniae radix and their association with MDD. (A) The table lists four active compounds—psoralen [20], xanthotoxin [21], isoimperatorin [22], and falcarindiol—along with each compound’s correlation score and protein overlap p-value in relation to MDD. (B) The network diagram reveals how these compounds (blue diamond nodes, left) connect to their target proteins and the relevant biological processes (gray and purple, center), ultimately linking to the MDD node (red hexagon, right). Edges represent interactions or functional relationships, highlighting potential mechanisms through which Glehniae radix may modulate MDD pathophysiology.

Figure 6

Active compounds of Euryales semen and their association with MDD. (A) The table lists three active compounds—xanthine [23], adenine, and thymidine—along with each compound’s correlation score and protein overlap p-value in relation to MDD. (B) The network diagram reveals how these compounds (blue diamond nodes, left) connect to their target proteins and the relevant biological processes (gray and purple, center), ultimately linking to the MDD node (red hexagon, right). Edges represent interactions or functional relationships, highlighting potential mechanisms through which Euryales Semen may modulate MDD pathophysiology.

Figure 7

Active compounds of Campsitis flos and their association with MDD. (A) The table lists four active compounds—oleanolic acid [24], maslinic acid, corosolic acid, and ursolic acid [25]—along with each compound’s correlation score and protein overlap p-value in relation to MDD. (B) The network diagram reveals how these compounds (blue diamond nodes, left) connect to their target proteins and the relevant biological processes (gray and purple, center), ultimately linking to the MDD node (red hexagon, right). Edges represent interactions or functional relationships, highlighting potential mechanisms through which Campsitis flos may modulate MDD pathophysiology.

Figure 8

Transcriptomic profile and GO enrichment of oleanolic acid-related genes. (A) Volcano plot illustrating log2 (fold-change) in gene expression between upregulated (red) and downregulated (blue) transcripts following treatment with oleanolic acid. The y-axis represents −log10(adjusted p-value), highlighting the most statistically significant differentially expressed genes (DEGs). (B) Gene Ontology (GO) enrichment analysis of the DEGs, grouped by domain: biological processes (red), cellular components (green), and molecular functions (blue).