The Pharmacological Activity of the Wenjing Decoction in Recurrent Spontaneous Abortion

Background Recurrent spontaneous abortion (RSA) is intractable infertility and can be ameliorated with the use of traditional Chinese medicine preparation, the Wenjing decoction. This study aimed to identify the therapeutic mechanism of Wenjing decoction on specific target proteins involved in RSA. Methods Wenjing decoction contains Wuzhuyu, Danggui, Chuanxiong, Guizhi, Shengjiang, Banxia, Gancao, Ejiao, Mudanpi, Chishao, Dangshen, and Maidong. Using TCMSP and BATMAN databases, we queried for active ingredients and predicted their target proteins by BATMAN. Using the edgeR package, we analyzed the differentially expressed genes (DEGs) in the GSE121950 database between control samples and RSA (n = 3). The interaction between DEGs and the predicted target proteins was identified by the Venn diagram. Using the Cytoscape software and clusterProfiler package, enrichment analysis was conducted for the intersected target proteins. Additionally, the protein-protein interaction (PPI) network and pharmacological network were generated using the Cytoscape software. Results In total, 31, 2, 7, 7, 5, 13, 93, 11, 29, and 21 active ingredients were identified from Wuzhuyu, Danggui, Chuanxiong, Guizhi, Shengjiang, Banxia, Gancao, Mudanpi, Chishao, and Dangshen, respectively. Additionally, 100 intersected target proteins were revealed by the Venn diagram. Moreover, 98 functional terms and 24 pathways (including C-type lectin receptor signaling pathway, chemokine signaling pathway, leukocyte transendothelial migration, fluid shear stress, and atherosclerosis, and AGE-RAGE signaling pathway in diabetic complications) were enriched. In the PPI network, 10 proteins involved in these five pathways were identified, namely, TNF-α (tumor necrosis factor-α), IL-10 (interleukin-10), TLR4 (Toll-like receptor 4), JUN (Jun proto-oncogene), IL-1B (interleukin-1-beta), CYBB (cytochrome b558 heavy chain gene), PTGS2 (prostaglandin-endoperoxide synthase 2), APOE (apolipoprotein E), SPI1 (salmonella pathogenicity island 1), and MPO (myeloperoxidase) which showed higher degrees. Conclusion The abovementioned genes and pathways might be involved in the pharmacological activity of Wenjing decoction in RSA.


Introduction
Abortion involves pregnancy termination before 28 weeks, with fetal weight below 1000 g [1]. Abortion can be classified as spontaneous or induced abortion, and recurrent spontaneous abortion (RSA) involves two or more consecutive spontaneous abortions [2]. Abortion manifests as postmenopausal vaginal bleeding and abdominal pain, although some women do not exhibit any clinical symptoms [3].
Nearly 50% of women with RSA have identifiable etiologies, such as chromosomal abnormalities, cervical insufficiency, immune dysfunction, maternal endocrine abnormalities, maternal reproductive tract abnormalities, reproductive tract infection, and thrombotic susceptibility [4,5]. RSA is intractable infertility clinically and poses a global therapeutic challenge, accounting for 1-5% of total pregnancies [6]. erefore, it is vital to understand RSA pathogenesis to improve therapy. Some traditional Chinese medicines (TCMs) have shown promise in women with recurrent miscarriage by improving embryo development and live birth rate [7]. e TCM Wenjing decoction improves blood circulation and eliminates stasis and has been used for alleviating gynecological diseases, including amenorrhea, dysmenorrhea, and paramenia [8]. Wenjing decoction could significantly alleviate pain in women with primary dysmenorrhea and could be used for primary dysmenorrhea treatment [9]. e modified Wenjing decoction repairs abnormal ovarian function in gynecological cold coagulation blood stasis by reducing ovarian oxidative damage [10]. Wenjing decoction is involved in vascular remodeling, microvascular contraction, and microcirculation and helps restore the normal ovarian function in cold coagulation blood stasis syndrome [11]. erefore, Wenjing decoction alleviates RSA by warming the channel, dispelling cold, nourishing blood, and removing blood stasis. However, the mechanisms underlying Wenjing decoction activity in RSA have not been identified.
TCM typically has multiple components and multiple targets. e active ingredients and candidate targets can be explored by network pharmacology [12,13]. In this study, the Wenjing decoction was evaluated by active ingredient screening, target protein prediction, cross-validation, enrichment analysis, and protein-protein interaction (PPI) network and pharmacological network construction. Our findings might help understand the pharmacological mechanisms involved in the Wenjing decoction effect in RSA.

Screening of TCM Active Ingredients.
e active ingredients were identified using the (ADME) parameters [15] from the TCMSP database. ADME studies evaluate the absorption, distribution, metabolism, and elimination of exogenous chemicals by the body and included DL, HL, and OB. is analysis used conventional screening parameters for drug composition. e cutoff criteria were DL ≥ 0.18 and OB ≥ 30%.

Target Protein Prediction for the Active Ingredients.
Using the BATMAN (http://bionet.ncpsb.org/batman-tcm/) [16] and TCMSP [14] databases, the target protein prediction for the active ingredients was performed. e BATMAN database first selected the target proteins from the Drug-Bank, erapeutic Target Database (TTD), and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases and ranked the predicted targets in descending order. e target proteins were selected for each small molecule drug candidate using a threshold of a score of >20. e BATMAN database was used to predict the differentiating drug-target interactions using binary classification for target protein prediction of active ingredients. e classification feature eigenvalue of the predicted drug-protein interaction was considered as the maximum eigenvalue of the similarity scores between the predicted drug-protein interaction and all known interactions.

Cross-Validation of the Target Proteins of the Active Ingredients.
e raw data from the GSE121950 was downloaded from the Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) database. is dataset was deposited by Huang et al. [17] and included RNA-seq data from decidua tissues on three RSA samples and three control samples. e original reads were filtered as follows: the reads with sequencing adaptor were removed, and the reads with N-base number exceeding one were filtered out. Subsequently, the reads with low-quality (Q ≤ 20) bases exceeding 40% of the total bases were deleted, and the reads with length <35 were omitted. Using the Tophat Software (v2.1.0, http://ccb.jhu.edu/software/tophat, the default parameters) [18], the clean reads were mapped to the human reference genome (GRCh38). Based on featureCounts software (v1.6.0, http://subread.sourceforge.net) [19], read count information for the alignment of each gene was obtained following the human gene annotation information provided by GENCODE (Release 25, https://www. gencodegenes.org/) [20].
Using the edgeR package (version 3.4, http://www. bioconductor.org/packages/release/bioc/html/edgeR.html) [21], the raw counts were standardized and converted to log2 counts per million (logCPM) values. e differential expression analysis for RSA vs. control groups was conducted. All mRNAs/lncRNAs were analyzed to obtain corresponding p values and log fold change (FC) values. e mRNAs/lncRNAs with p values of <0.05 and |logFC| > 1 were considered as differentially expressed genes (DEGs).
e target protein intersection of the active ingredients and DEGs was analyzed subsequently.

Enrichment Analysis of the Intersected Target Proteins.
e Gene Ontology (GO)_biological process (BP) enrichment analysis of the intersected target proteins was conducted using the ClueGO + CluePedia plug-in Cytoscape software (http://www.cytoscape.org) [22]. e adjusted p value ≤0.01 was the threshold for GO_BP enrichment analysis. In addition, the R package cluster-Profiler [23] was used for the pathway enrichment analysis of the intersected target proteins, and a p value ≤0.01 was set as the threshold.

PPI Network for the Intersected Target Proteins and Pharmacological Network Construction.
e interactions among the intersected target proteins were predicted by the STRING database (http://string-db.org/) [24], and the PPI network was built using the Cytoscape software [22]. e required confidence (combined score) > 0.4 was the cutoff criterion for the PPI selection. In addition, the pharmacological network involving the TCM-active ingredient-target proteinpathway relationships was built by the Cytoscape software [22].

Target Protein Prediction and Cross-Validation.
e BATMAN online tool was used to predict the target proteins for the active ingredients in the TCMSP database, and the target records for 10 herbs, including Wuzhuyu, Danggui, Chuanxiong, Guizhi, Shengjiang, Banxia, Gancao, Mudanpi, Chishao, and Dangshen, were obtained. In total, we identified 933 target proteins of 49 active ingredients from the 10 TCMs after screening the targets with a score larger than 20 in BATMAN.
From the raw data of GSE121950, a total of 17163 mRNAs were annotated. Differential expression analysis showed that there were 1371 DEGs (including 453 upregulated and 918 downregulated genes) between RSA and the control samples (Figure 1). e Venn diagram showed the presence of 100 intersecting target proteins among the 933 target proteins of the active ingredients and the 1371 DEGs ( Figure 2 and Supplementary Table 1).

Discussion
In this study, we screened 31, 2, 7, 7, 5, 13, 93, 11, 29, and 21 distinct active ingredients from the TCM preparations of Wuzhuyu, Danggui, Chuanxiong, Guizhi, Shengjiang, Banxia, Gancao, Mudanpi, Chishao, and Dangshen. In total, 933 target proteins for the active ingredients in the TCMSP database were predicted and 1371 DEGs between RSA and control samples were identified from GSE121950. Venn diagram showed 100 intersected target proteins between the 933 target proteins of the active ingredients and 1371 DEGs. Among the 100 intersected target proteins, 98 GO_BP terms and 24 pathways were enriched significantly. e PPI network of the intersected target proteins, TNFα, IL-10, TLR4, JUN, IL-1B, CYBB, PTGS2, APOE, SPI1, and MPO, showed higher degrees. e etiologies of RSA have been extensively identified previously, including chromosomal abnormalities, cervical insufficiency, immune dysfunction, maternal endocrine abnormalities, maternal reproductive tract abnormalities, reproductive tract infection, and thrombotic susceptibility [4,5]. TNF-α overexpression in the decidual tissues and peripheral blood of RSA patients was detected, showing correlation with disease occurrence, and might serve as a candidate indicator for RSA [25,26]. e IL-10 levels were significantly lower in RSA than during normal pregnancy, and decreased IL-10 levels could be correlated to pathologic pregnancies, while increased IL-10 to healthy pregnancies [27]. e involvement of TLR4 and C-C chemokine ligand 2 (CCL2) in RSA development was shown Evidence-Based Complementary and Alternative Medicine  before, and the serum TLR4 and CCL2 levels in pregnant women could serve as RSA indicators [28]. Reduced TLR4 and increased TLR2 induced by ligand treatment of spermatozoa influenced unexplained RSA [29]. erefore, in RSA, TNF-α, IL-10, and TLR4 might be targets for Wenjing decoction.
e IL-1B (C-511T) polymorphism may promote the development of recurrent miscarriages, while the IL-6     (G-634C) polymorphism may reduce its risk [30]. Danggui-Honghua by modulating 42 targets, including PTGS2, might play a pharmacological role in treating blood stasis syndrome [31]. Apo E3/4, E4/4 genotypes are more frequent in women with recurrent pregnancy loss (RPL) compared to control women, and these genotypes might cause RPL by increasing the thrombophilic risk factors [32]. e levels of the oxidative stress biomarkers, including MPO, are higher in women with RPL exhibiting metabolic syndrome, and the oxidative stress, subclinical inflammatory state, and metabolic syndrome contribute to pregnancy outcome prediction in women with RPL [33], thus suggesting that in RSA, IL-1B, PTGS2, APOE, and MPO might also be targeted by Wenjing decoction. Enrichment analysis of the 100 intersected target proteins showed enrichment of "AGE-RAGE signaling pathway in diabetic complications," "TNF signaling pathway," "osteoclast differentiation," "C-type lectin receptor signaling pathway," "chemokine signaling pathway," "leukocyte transendothelial migration," and "MAPK signaling pathway." A previous study showed the contribution of the C-type lectin receptors in tailoring the adaptive responses and innate recognition, and several C-type lectin receptors function in protective antifungal immunity [34]. Systemic immune responses and abnormal maternal local have shown correlations with RSA; especially, antiphospholipid antibodies and maternal-fetal immune tolerance disturbance affect both autoimmune RSA and alloimmune RSA [35,36]. Chemokines are a family of chemoattractant proteins, which can trigger immune and inflammatory responses, cell polarization, and prevention of HIV-1 infection [37]. e migration of  Figure 6: e pharmacological network. Brown diamond, blue square, red and green circles, circles with a red border, and red hexagon represent TCMs, active ingredients, target proteins, pathway-associated target proteins, and pathways, respectively. leukocytes through activated venular walls is a basic immune response, which is necessary for the entry of effector cells into the inflamed tissues [38]. Investigations on the role of AGEs and the receptor of AGEs (RAGE) in vascular disease and blood pressure showed the importance of the AGE-RAGE axis in vascular disease, micro-and macrocirculation [39].
us, these pathways could be involved in the pharmacological activity of the Wenjing decoction against RSA. TNF-α, IL-10, TLR4, JUN, IL-1B, CYBB, PTGS2, APOE, SPI1, and MPO were involved in the pharmacological network pathways and might be responsible for the pharmacological activity of Wenjing decoction.
is study is a preliminary research to screen the primary targets of Wenjing decoction. RSA is a serious disease involving fetus/baby and mother. Misinformation about RSA may lead to more RSA. erefore, the results of this study should be interpreted cautiously. Further clinical studies as well as in vitro or in vivo studies are still warranted to investigate the effects of Wenjing decoction on RSA.

Conclusions
We selected 100 intersected target proteins of the active ingredients by cross-validation. During RSA, the TNF-α, IL-10, TLR4, JUN, IL-1B, CYBB, PTGS2, APOE, SPI1, and MPO pathways might be targeted by the Wenjing decoction. Furthermore, "C-type lectin receptor signaling pathway," "chemokine signaling pathway," "leukocyte transendothelial migration," "TNF signaling pathway," "MAPK signaling pathway," and "AGE-RAGE signaling pathway in diabetic complications" might correlate with the pharmacological activity of Wenjing decoction. However, the genes and pathways involved and their significance in women with RSA treated by Wenjing decoction should be explored further.

Additional Points
(1) In total, 100 intersecting target proteins were selected by the Venn diagram.