Network pharmacology and molecular docking analysis on the mechanism of Baihe Zhimu decoction in the treatment of postpartum depression

Baihe Zhimu decoction (BZD) has significant antidepressant properties and is widely used to treat mental diseases. However, the multitarget mechanism of BZD in postpartum depression (PPD) remains to be elucidated. Therefore, the aim of this study was to explore the molecular mechanisms of BDZ in treating PPD using network pharmacology and molecular docking. Active components and their target proteins were screened from the traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP). The PPD-related targets were obtained from the OMIM, CTD, and GeneCards databases. After overlap, the targets of BZD against PPD were collected. Protein–protein interaction (PPI) network and core target analyses were conducted using the STRING network platform and Cytoscape software. Moreover, molecular docking methods were used to confirm the high affinity between BZD and targets. Finally, the DAVID online tool was used to perform gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of overlapping targets. The TCMSP database showed that BZD contained 23 active ingredients in PPD. KEGG analysis showed that overlapping genes were mainly enriched in HIF-1, dopaminergic synapses, estrogen, and serotonergic synaptic signalling pathways. Combining the PPI network and KEGG enrichment analysis, we found that ESR1, MAOA, NR3C1, VEGFA, and mTOR were the key targets of PPD. In addition, molecular docking confirmed the high affinity between BZD and the PPD target. Verified by a network pharmacology approach based on data mining and molecular docking methods, the multi-target drug BZD may serve as a promising therapeutic candidate for PPD, but further in vivo/in vitro experiments are needed.


Introduction
Postpartum depression (PPD) can be defined as non-psychotic depression occurring within a year of childbirth, characterized by low mood, unusual thoughts, feelings of guilt, unexplained anxiety, worthlessness, and other depressive symptoms. [1] The rate of PPD among women may be as high as 15% and result in a high death rate from self-harm. The rising numbers could make matters worse. [2][3][4][5] Thus, PPD is a significant public health problem. Medicine Traditional Chinese medicine (TCM) is an oriental traditional medicine that is characterized by a holistic concept. It was the main medical method used in ancient China, with thousands of years of accumulated practical experience. [12] In ancient China, TCM was used to treat many mental illnesses, such as restlessness, lily disease, and depression. Reports show that many TCMs, such as Baihe Zhimu decoction (BZD), Chaihu decoction, and Zhizichi decoction Xiaoyao san, are used to treat PPD.
BZD is a classic traditional Chinese medicine prescription, which was first recorded in Treatise on Febrile and Miscellaneous Diseases (200-210 ad). Consisting of two herbs, Lilii Bulbus (Baihe BH) and Anemarrhenae Rhizoma (Zhimu ZM), BZD has become a classic Chinese medicine formulation for treating depression, insomnia, anxiety and other mental and neurological diseases. [13] Some studies have shown that BZD can increase 5-hydroxytryptamine, norepinephrine, and dopamine levels and improve the behavioral indicators of depression. [14][15][16][17] Clinical data indicate that Morita therapy combined with BZD can improve the symptoms, quality of life, and ability of daily living for first-episode depression, which may be related to the decrease in serum BDNF and an increase in DOPAC levels. [18] However, the potential mechanisms underlying the treatment of PPD are not fully understood.
Network pharmacology is an emerging method that analyses the components, targets, diseases, and related pathways of traditional Chinese medicine. It combines pharmacology, molecular biology, electronic technology, and bioinformatics and has great advantages for studying the complex components, targets, and pathways of traditional Chinese medicine prescriptions. [19][20][21] Therefore, this research aims to apply network pharmacology to identify the active ingredients and targets of BZD and to discover the common targets and possible signaling pathways of BZD in the treatment of PPD. The corresponding workflow is shown in Figure 1.

Chemical ingredient acquisition and processing
The components of the two herbs were found in the pharmacology of Traditional Chinese Medicine Systems Pharmacology (TCMSP, https://tcmspw.com/tcmsp.php) [22] and then screened by integrating oral bioavailability (OB ≥ 30%) and drug-likeness (DL ≥ 0.18). [23] Next, each target of the active ingredient was obtained from this website. The aggregated target was input into UniProt (https://www.uniprot.org/) to obtain information, such as gene symbols and gene IDs. [24,25]

Construction of drug-compound-target genes network
We input the information obtained above of BZD-related drugs, components, and targets into Cytoscape 3.7.2 software to construct a visualized network diagram of the drug-component-target. The graph depicts drugs, ingredients and targets as nodes and the lines connecting them as edges.

Venn diagram of targets between drugs and disease
The targets of BZD and PPD-related target gene lists were input into Venny 2.1.0 (http://bioinformatics.psb.ugent.be/webtools/ Venn/) to obtain their intersection with a Venn diagram.

Protein-protein interaction data
A protein-protein interaction (PPI) network for disease and drug mapping targets was constructed using the STRING database (https://string-db.org/, version 10.5). According to the corresponding calculation method, the species was set to "Homo sapiens," and the confidence score was set to >0.4. [30] 2.6. Gene ontology and pathway enrichment analysis The Database for Annotation, Visualization, and Integrated Discovery (DAVID; http://david.abcc.ncifcrf.gov/) [31,32] was used for gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses.

Molecular docking
Based on the results of the GO and KEGG pathway analyses, we selected the key protein receptor and ligand associated with the protein receptor. PubChem (https://pubchem.ncbi.nlm.gov/) [33] was used to obtain the 2D chemical structures of the small-molecule ligands. ChemOffice software was used to construct the 3D chemical structures of small molecular ligands, [34] and the 3D chemical structures of protein receptors were acquired from PDB (http://www.rcsb.org/). [35] After the molecular ligands and water molecules of the protein receptors were removed using PyMol 2.4.0 software (https://pymol.org.), [36] the format of the protein receptors and small-molecule ligands was transformed into pdbqt format. [37] AutoDock Vina was used for molecular docking. [38] Based on the binding energy value, the lowest binding energy value was selected as the docking affinity. Finally, PyMol software was used to visualize the 3D structures of the molecular ligand and protein receptor bonding.

Chemical components of BZD
The active compounds of BZD were retrieved from TCMSP. There were 165 related components in BZD; Baihe contained 84 (50.9%), and Zhimu had 81 (49.1%) components. The values of OB and DL (OB ≥ 30% and DL ≥ 0.18) were used to screen potential active compounds from Baihe and Zhimu, and 22 active compounds met the screening standards. The properties of the compounds are listed in Table 1. The 364 targets and the corresponding Uniprot IDs are listed in Table 2.

Screening targets of PPD
We searched for PPD targets in several publicly available databases: CTD, OMIM, and GeneCards. After deleting duplicate values, we finally obtained 663 disease targets.

Drug-compound-target network
The drugs, components, and targets of BZD were drawn into a drug-component-target network diagram using the Cytoscape software to show the relationship between them more intuitively (Fig. 2). In this figure, orange represents two drugs, green rectangle represents 22 active ingredients, and yellow circle represents 364 target genes. Among the three circles, the degree value of the inner circle is higher than that of the outer circle, which has a higher correlation. In the drug-compound-target network, according to the degree value, the top 10 targets were ESR1, AR, F2, PTGS2, NR3C1, DPP4, PGR, NOS2, PTGS1, and CDK2. Figure 3 shows the set relationship between the targets of BZD and PPD. We input the 364 targets of BZD and 663 targets of PPD into the Venny2.1 online software mapping tool platform to draw the Venn diagram. Among them, 41 target proteins were targets of BZD acting on PPD, and 25 were common targets of Baihe, Zhimu, and PPD. Figure 4 shows the PPI diagram of the 41 intersecting target proteins drawn using STRING. Red, evidence of fusion; green, evidence of proximity; yellow, evidence of text mining; Figure 1. Workflow of the study design. Medicine light blue, evidence of database; blue, evidence of coexistence; black, evidence of co-expression; purple, evidence of experiment.

GO and pathway enrichment analysis
The KEGG analysis and GO enrichment analysis of biological process (BP), molecular function (MF), and cell component (CC) obtained by DAVID 6.8 for 41 selected target genes are shown in Figure 5, which clarifies the multiple mechanisms of BZD in treating PPD. There is information on the number of genes, selections, and rich factors. P in the figure represents the significance of enrichment; the redder the color, the higher the significance. The five most affected BPs (P < .01) were response to drugs (GO:0042493), oxidation-reduction process (GO:0055114), response to lipopolysaccharide (GO:0032496), response to estradiol (GO:0032355), and response to hypoxia (GO:0001666) (  Table 3. The relevant reference values are shown in Figure 5D. According to Figure 6, HIF-1, neuroactive ligand-receptor interaction, dopaminergic synapse, estrogen, and serotonergic synapse signaling pathways, which were filtered out as prominent and conspicuous enriched pathways, contributed significantly to the PPD response.

Molecular docking analysis
According to the five related signaling pathways from the results of KEGG pathway enrichment analysis, we identified five key genes: ESR1, MAOA, NR3C1, VEGFA, and MTOR. These key genes were mainly bound to diosgenin, isopimaric acid, stigmasterol, and beta-sitosterol, which are the main components of BZDs (Table 4). The molecular docking method verified the binding sites of BZD target genes and their corresponding compounds, showing that the above five target genes have a high affinity for the main components of BZD (Fig. 7).

Discussion
PPD is a common and serious mental health problem that causes personal suffering and interferes with parenting. However, the effects of antidepressant medications remain controversial. In the meantime, numerous people worry about the potential adverse effects of antidepressant medications. TCM has been used to treat mental illness in China for thousands of years. Therefore, the application of Chinese medicine can help develop new ideas and methods for the treatment of PPD.
Owing to the multi-target treatment effects of TCM, it can serve as a significant repository to develop drugs for the treatment of PPD. This study utilized network pharmacology and molecular docking simulations to reveal the molecular mechanisms of BZD in PPD treatment. BZD plays a potential role in treating PPD by regulating multiple target genes and pathways.
For this study, we selected 22 main compounds of BZD, among which icariin, timosaponin B-III (TB-III), and others are shown to have antidepressant effects. The potential target genes are ESR1, MAOA, NR3C1, VEGFA, and mTOR. Moreover, GO annotation and KEGG pathway enrichment analyses confirmed that these target genes were associated with HIF-1, dopaminergic synapse, estrogen, and serotonergic synapse signaling pathways, which are closely involved in the treatment of PPD. Molecular docking showed that the five core targets had a certain affinity for the main compounds of BZD.
TB-III is a steroidal saponin isolated from the rhizome of Baihe, which exhibits antidepressant activity through the regulation of inflammatory cytokines, BDNF signaling, and synaptic plasticity. However, this study was only conducted in a mouse model of PPD, and there have been no human reports. [39] Icariin has potential preventive and therapeutic effects in various neurological diseases such as cerebral ischemia, depression, Parkinson's disease (PD), and multiple sclerosis (MS). The mechanism by which icariin improves depression may be related to the promotion of cell proliferation, peripheral nerve regeneration, improvement of Table 1 Compounds of Baihe Zhimu decoction.

Mol ID
Molecule name OB (%) DL Herbs  the function of damaged nerve regulation, decrease in glucocorticoid receptors (GRs) and 5-hydroxytryptamine 1A (5-HTR1A) receptors in the hippocampus and prefrontal cortex, regulation of the central neuroendocrine system, or restoration of the negative feedback regulation of the hypothalamic-pituitary-adrenal (HPA) axis. [40][41][42] Furthermore, icariin may ameliorate prenatal restraint stress-induced depression-like behavior. [43] The network pharmacology results confirmed that the potential target genes of PPD regulated by BZD were mainly ESR1, MAOA, NR3C1, VEGFA, and mTOR. Simultaneously, the PPI network results showed that these targets had close interactions. ESR1 plays an important role in mediating hormonal differences during pregnancy and postpartum. One clinical experiment suggested a role for ESR1 in the etiology of PPD, possibly through modulation of serotonin signaling. [44] Previous work has demonstrated that exposure to and withdrawal from normal levels of gonadal steroids results in depressive symptomatology in women previously diagnosed with PPD. [45,46] The MAOA gene, located on the short arm of the X-chromosome (Xp11.4-p11. 3), has been the focus of research in the field of mental disorders in recent years. Two studies showed that MAOA was positive at 6 weeks postpartum with PPD. [47,48] Women with PPD appear to have an abnormal HPA axis response to stress, which may involve genetic

No.
Target UniprotID No. Target  UniprotID  No.  Target  UniprotID  No.  Target  UniprotID   70  PRKACA  P17612  161  BCL2  P10415  252  ESR1  P03372  343  VEGFA  P15692  71  ADH1C  P00326  162  ALOX5  P09917  253  AR  P10275  344  AR  P10275  72  IGHG1  P01857  163  PTGS2  P35354  254  SCN5A  Q14524  345  PGR  P06401  73  CTRB1  P17538  164  AKR1C3  P42330  255  OPRD1  P41143  346  NR3C2  P08235  74  PRSS1  P07477  165  TNFSF15  O95150  256  ADRA1B  P35368  347  NR3C1  P04150  75  NCOA2  Q15596  166  ESR2  Q92731  257  OPRM1  P35372  348  PTGS1  P23219  76  NCOA1  Q15788  167  MMP1  P03956  258  GABRA1  P14867  349  F2  P00734  77  BCL2  P10415  168  JUN  P05412  259  DPP4  P27487  350  CHRM1  P11229  78  PON1  P27169  169  SELE  P16581  260  GSK3B  P49841  351  ESR1  P03372  79  JUN  P05412  170  CDK1  P06493  261  CDK2  P24941  352  ADRB1  P08588  80  MAP2  P11137  171  VCAM1  P19320  262  PIK3CG  P48736   variants, as reported previously. The NR3C1 gene, which is located on the long arm of chromosome 5 (5q31), encodes the GR. Epigenetic studies have shown that decreased NR3C1 gene expression activity caused by methylation can impair the negative feedback regulation of the HPA axis, alter an individual's response to SLE, and induce depression. [49] However, the mechanisms of interaction between these targets are not clear. Other targets have not been directly documented to improve PPD, but the literature supports an improvement in depression, which follows the same pathway as PPD. For example, VEGFA can affect the complex processes of learning and memory [50] and plays a role in regulating neurite growth and maturation during brain development. [51] The role of VEGFA in neurogenesis may be mediated by its interactions with downstream effector genes. [52] In the present study, our data showed that VEGF mRNA and protein expression in hippocampal tissues and serum were downregulated in depression model rats, suggesting that downregulation of VEGFA plays a key role in depression in rats. mTOR is a serine/threonine kinase that controls related signalling pathways to regulate many integrated physiological functions of the nervous system. Many studies have shown that it is tempting to hypothesize that the activation of mTOR function followed by enhanced mTOR-dependent protein synthesis may underlie the action of antidepressants, such as ketamine. [53][54][55][56] GO annotation and pathway enrichment analyses were conducted to identify the potential biological functions of PPD targets. GO enrichment analysis revealed that the major biological  processes included response to drugs, oxidation-reduction processes, response to lipopolysaccharide, response to estradiol, and response to hypoxia. The results of the KEGG enrichment analysis showed that HIF-1, dopaminergic synapse, estrogen, and serotonergic synapse signaling pathways were the leading signaling pathways for the treatment of PPD by BZD. These data suggest that the HIF-1 pathway might play an important role in antidepressant effects, and that altered mRNA expression of HIF-1 and its target genes in peripheral blood cells is associated mainly in a state-dependent manner with mood disorders, especially major depressive disorder (MDD). [57] The estrogen signaling pathway plays an important role in PPD. Studies have shown that changes in gonadal steroid levels, including estrogen levels, may contribute to the depressive symptoms of PPD. [58] Other data suggest that PPD symptoms are affected by estrogen levels. [59,60] Serotonin regulates several basic biological functions relevant to depression, including sleep and appetite. [61] Women with PPD have lower plasma serotonin levels than non-depressed controls, which are modulated by estrogen. [62] Thus, fluctuating estradiol levels during pregnancy and the postpartum period may cause depressive symptoms in vulnerable women by destabilizing the serotonin system. Dopamine is recognized and proven to be an important factor, and diminished dopaminergic function may also play a role in PPD. Studies have shown that dopamine activity and estradiol levels are positively correlated; therefore, they can synergistically affect PPD symptoms. [63]   Molecular docking verified that ESR1, MAOA, NR3C1, VEGFA, and mTOR have high affinities for the main active ingredients of BZD, diosgenin, isopimaric acid, stigmasterol, and beta-sitosterol, providing data support for BZD as a potential drug for the treatment of PD; however, further experimental studies are needed to confirm its effectiveness.

Conclusion
In this study, we used network pharmacology combined with molecular docking to elucidate the mechanism by which BZD regulates PPD through multiple targets and channels. The main target proteins of BZD in PPD treatment have been shown, constructing a target protein network. BZD mainly affected HIF-1, dopaminergic synapse, estrogen, and serotonergic synapse signaling pathways while regulating the key target proteins of ESR1, MAOA, NR3C1, VEGFA, and mTOR.
However, this article also has some limitations. The composition of traditional Chinese medicine is complex, the data of this article is obtained through some databases, and the information of the databases needs to be further improved. The validity of this article lacks experimental support, and its mechanism needs further experimental verification.

Author contributions
Qiong Zhao, Wengu Pan, and Guomin Si performed the main analyses and drafted the manuscript. Hongshuo Shi and Fanghua Qi designed the study. Yuan Liu helped with the introduction and discussion. Tiantian Yang and Hao Si assisted in the preparation of the manuscript. All authors wrote, read, and approved the manuscript.   Table 4 Results of the molecular docking of the five core genes with compounds of BZD.

Number
Core genes Compound Docking affinity