Multi‐omics analysis reveals the interaction between the complement system and the coagulation cascade in the development of endometriosis

Endometriosis (EMS) is a disease that shows immune dysfunction and chronic inflammation characteristics, suggesting a role of complement system in its pathophysiology. To find out the hub genes and pathways involved in the pathogenesis of EMs, three raw microarray datasets were recruited from the Gene Expression Omnibus database (GEO). Then, a series of bioinformatics technologies including gene ontology (GO), Hallmark pathway enrichment, protein–protein interaction (PPI) network and gene co-expression correlation analysis were performed to identify hub genes. The hub genes were further verified by the Real-time quantitative polymerase chain reaction (RT-PCR) and Western Blot (WB). We identified 129 differentially expressed genes (DEGs) in EMs, of which 78 were up-regulated and 51 were down-regulated. Through GO functional enrichment analysis, we found that the DEGs are mainly enriched in cell adhesion, extracellular matrix remodeling, chemokine regulation, angiogenesis regulation, epithelial cell proliferation, et al. In Hallmark pathway enrichment analysis, coagulation pathway showed great significance and the terms in which included the central complement factors. Moreover, the genes were dominating in PPI network. Combined co-expression analysis with experimental verification, we found that the up-regulated expression of complement (C1S, C1QA, C1R, and C3) was positively related to tissue factor (TF) in EMs. In this study, we discovered the over expression complement and the positive correlation between complement and TF in EMs, which suggested that interaction of complement and coagulation system may play a role within the pathophysiology of EMS.

diseases at the molecular level 12 . Some meaningful researches have been conducted to identify genes involved in the development of EMS 13,14 . In this study, we integrated data sets from multiple studies and performed deep bioinformatic analyses to identify differentially expressed genes (DEGs) and important pathways participating in the development of EMS. Through pathway enrichment analysis, the coagulation process containing the central complement factors was indicated as the significant component of EMS. Real-time quantitative polymerase chain reaction (RT-PCR) and Western Blot (WB) were adopted to validate these results in women with or without EMS. After evaluating and comparing the expression of hub genes in EMS and control women, we found that the overexpressed state of complement was positively related to tissue factor (TF) in EMS. These suggested the existence of crosstalk between complement and coagulation in EMS.

Method
Data acquistion and differentially expressed genes (DEGs) identification. Three datasets (GSE23339, GSE7305, and GSE25628) were selected for our analysis. In detail, the gene set GSE23339 was annotated by GPL6102 platform, which included 10 endometriosis tissue samples and 9 normal endometrial tissue samples (as of control samples); GSE7305 was annotated by GPL570 platform, the data in which consist of 10 endometriosis tissue samples and 10 normal endometrial tissue samples; GSE25628 was annotated by GPL571 platform, which included 16 endometriosis and 6 normal endometrial samples. The platform annotation and series matrix files were downloaded for follow-up analyses.
Next, the orignial file obtained from GEO databases were preprocessed in R software. Samples of each dataset were normalized by the Robust Multi-array Average (RMA) function in the Affy package 15 . And then we used SVA package to eliminate the inter-batch difference between groups 16 . Multiple probes relating to the same gene were summarized as the median value for further analysis 17 . In our analysis, DEGs were defined with a threshold of p value < 0.05 and |log (FC)|> 1. DEGs in each database were displayed with the volcano map. Overlapping DEGs within three databases were shown with Venn diagrams.
Protein-protein interaction (PPI) network analysis. The PPI of DEGs-encoded proteins was confirmed by STRING 18 (version 11.0), with search limitation to Homo sapiens and a score > 0.400 correspondingsto high confidence and statistical significance. After that, Cytoscape software (version 3.7.1) was used to construct the network 19 . Moreover, the cytohubba plug-in unit of Cytoscape software was used for the identification of the top 20 hub genes according to the count of interactions between genes.
Gene ontology and pathway enrichment analysis. Gene ontology (GO) Biological Processes of DEGs were analyzed using "clusterProfiler" package in R software 20 . P-value < 0.05 was defined as the cutoff value. GO considers annotates gene function from three aspects: molecular function(MF), cellular component(CC) and biological process(BP) 21 . Among these, Molecular functions define molecular processes, Cellular components define locations where molecular processes occur and Biological processes define biological programs comprised of regulated factors. To remove repetition and ensure the authenticity of the results, we merged the terms using simplify function in the GOplot package. The Hallmark pathway enrichment analysis was performed in Metascape16 (https:// metas cape. org/ gp/ index. html#/ main/ step3) (22). P-value < 0.05 was defined as the cutoff value.
Gene co-expression correlation analysis. To identify the correlation between hub genes, GSE141549 was downloaded as a validation set. GSE141549 was annotated by GPL10558 platform, the data in which consist of 198 endometriosis samples and 147 normal samples. We used R to build a co-expression network and visualize it.
Clinical sample collection. Eutopic endometrial tissues and control endometrium were obtained respectively from three patients with ovarian endometriosis (OE) and three women with benign ovarian teratoma, which were diagnosed by pathology after surgery. According to the American Fertility Society revised (AFS-r), these 3 EMs cases were all stage IV. All participants have written informed consent under the study protocol approved by the Ethics Committee of Changzhou Second People's Hospital (Changzhou, China). All of the endometrial tissues were collected at the proliferative phase during the menstrual cycle. None of the recruited women smoked or had taken any anticoagulant drugs or steroid hormones 3 months prior to the operation. There was no significant difference in age between the two groups (P > 0.05).
Quantitative-Polymerase chain reaction (qPCR). Firstly, we isolated total RNA by using Trizol (Invitrogen) according to the manufacturer's protocol. The PrimeScript RT Master Mix (Takara, JPN) was used for the synthesis of cDNA. Quantitative-Polymerase chain reaction was performed using the SYBR Green assay to evaluate the expression of hub genes on the basis of manufacturer's instructions (Applied Biosystems, USA). The typical qPCR cycle were as followed: Pre-denaturation at 95˚C for 30 s, followed by 40 cycles of denaturation at 95˚C for 5 s, annealing at 60˚C for 10 s and extension at 70-72˚C for 30 s 23 . We used GAPDH to normalize the expression of hub genes. And then, The relative expression level of the target amplicon was calculated between the EMs group and the control group. Hub genes and GAPDH primers were designed and synthesized by Sangon Biotech Co., Ltd. (Shanghai, China). The primers' sequences were showed in Table 1.

Identification of differentially expressed genes (DEGs) in EMS using integrated bioinformatics.
Samples in each dataset (GSE7305, GSE23339, and GSE25628) were normalized using Robust Multiarray Average (RMA) function in the Affy package ( Fig. 1). All datasets were normalized using SVA package to eliminate potential inter-batch difference (Fig. 2). Then we used the Limma package to perform differential expression analysis on the three data sets with a definition criterion based on p < 0.05 and |log (FC)|> 1 to select statically statistically DEGs. In GSE7305, 1262 DEGs were identified, of which 694DEGs were up-regulated and 568 down-regulated (  Table 2). Then, we converted the chart into the heat map to show the expression of DEGs intuitively (Fig. 4).
Gene Ontology (GO) functional enrichment in DEGs. Thereafter, we performed gene ontology (GO) enrichment analysis of DEGs in endometriosis by using clusterprofiler package. The results were divided into three parts: biological process (BP), cellular component (CC), and molecular functions (MF). Then, we use R to import the top ten results of every sector into the figure (Fig. 5C). In the BP, DEGs were mainly involved in cell adhesion, epithelial cell proliferation, granulocyte chemotaxis regulation, and extracellular matrix remodeling.
In the CC, cell connection, cell adhesion, and extracellular matrix collagen-containing were included. In the MF, terms were mainly involved in binding and extracellular matrix structural constituent conferring compression resistance.
Hallmark signaling pathway enrichment in DEGs. We use Metascape16 to perform signaling pathway enrichment of DEGs in endometriosis, and then we submit the most significantly enriched pathways to Hallmark genes hit analysis. Through Hallmark pathway enrichment analysis, epithelial-mesenchymal transition  TTC CCC AAG CCT TAC CCC AA  GCT GGA AGA CGA GCT TCA CC   C1S  ACT GTG CGT ATG ACT CAG TGC  GGG GAT TGT TAC TGC TCC TCT   C1QA  TCT GCA CTG TAC CCG GCT A  CCC TGG TAA ATG TGA CCC TTTT   C3  CGC AAC AAG TTC GTG ACC G  GAT GCC TTC CGG GTT CTC AAT   P2RY14  AAT CTA GCC GCA ACA TAT TCAGC GTC TGA CTC TTT GTG TAG GGGAT   SEPRING1  CTG GCT GGG GAT AGA GCC T  GAG ATA ACT GTT GTT GCG ACCT   TF  www.nature.com/scientificreports/ (EMT), coagulation, estrogen response late, KRAS signaling up, and apical junction were identified (Fig. 5D). Among them, EMT and coagulation pathways showed great significance according to the P Value (Table 3).
Bioinformatics analysis shows that SERPING1, C1S, C1QA, C1R, C3, and TF are up-regulated in EMs while P2RY14 is down-regulated in EMs group. Combined Top20 hub genes with pathway enrichment analysis, coagulation pathway showed great significance. Meanwhile, SERPING1, P2RY14, C1S, C1QA, C1R, and C3 showed better coagulation pathway correlation and EMs correlation. Among them, SERP-ING1, C1S, C1QA, C1R, and C3 were up-regulated in EMs while P2RY14 was down-regulated ( Figure S2a). To verify whether the expression of genes involved in the coagulation cascade was synchronized with complement factors, we selected tissue factor (TF) for verification, which was the important factor in the exogenous coagulation pathway and EMs 25,26 . After analyzing the data set, TF was found to be up-regulated in EMs (P = 0.044) ( Figure S2b).   www.nature.com/scientificreports/ C1R, and C3 were up-regulated in EMs group (Fig. 6A-E) while P2RY14 was down-regulated (Fig. 6G). This result was also validated by western blotting at the protein level (Fig. 7). The verification of the hub genes both at mRNA level and protein level suggested the abnormal status of complement in EMs. We also found that the expressions of TF were up-regulated both at the mRNA and protein levels of EMs group (Fig. 6F, 8).
Bioinformatics analysis shows that Central complement factors (C1S, C1QA, C1R, C3) are positively related to TF. Further, we analyzed the correlation between the expression of hub genes and TF in GSE141549 data set. In co-expression correlation analysis, the complement showed a positive correlation with TF while SERPING1 and P2RY14 showed a negative correlation with TF. Interestingly, SERPING1 was up-regulated in EMs synchronized with complement components (Fig. 8). Also, these genes were involved in the coagulation, proposing the existence of crosstalk between complement and coagulation cascades in EMs.

Discussion
Here, through integration and analysis of the data from 3 microarray datasets, we identified 129 DEGs, of which 78 up-regulated and 51 down-regulated in EMs. GO functional enrichment analysis showed that DEGs are mainly enriched in cell adhesion, extracellular matrix remodeling, chemokine regulation, angiogenesis regulation, epithelial cell proliferation, which have been proved the essential steps in the development of EMs [26][27][28][29] . In this study, Hallmark pathway enrichment analysis disclosed coagulation and epithelial-mesenchymal transition (EMT) were the most significant pathways in EMs [30][31][32] . To EMT, many studies have revealed its main role in the development of EMS and Yihua Wang et al. have also confirmed this by bioinformatic analysis 33 . So, in this study, we focused on the coagulation pathway of EMS and found that complement factors were enriched in this way according Hallmark pathway enrichment analysis. It is widely known that the relationship between the immune system and EMS is intimate 34,35 . Swati et al. revealed that chronic inflammation in endometriosis is dominated by complement 7 . Sikora et al. had proved that abnormal activation of the complement pathway occurred in EMS and resulted in the overexpression of C1Q, MBL, C1INH 36 . In our study, we found that the up-regulated expression of the central complement factor (C1S, C1QA, C1R, and C3) in EMS, indicating that the abnormal status of complement might be involved in the pathogenesis of EMS. Meanwhile, Vigano et al. confirmed the existence of hypercoagulable state in EMS patients 37 . Researches conducted by Guo et al. also revealed that stromal cells in EMS lesions could secrete thrombin and Thromboxane A2 (TXA2) to induce platelet activation and participate in coagulation 38,39 . And then, Thrombin in the endometrial matrix could be activated in response to repeated bleeding from highly vascularized lesions in EMs and combined with PAR-1 to mediate thrombin-induced angiogenesis 40,41 . By now, few researches focused on EMs indicated the existence of the crosstalk between complement and coagulation. www.nature.com/scientificreports/ Though, the crosstalk between the complement and coagulation cascades has been proved in septicemia. Rittirsch revealed that complement activation could enhance coagulation through up-regulating the expression of tissue factor 42 . Hence, we proposed that TF might be the interaction between complement and coagulation and the crosstalk between complement and coagulation might also play an essential role in the development of EMS. Previously, Ding et al. found that the concentration of TF in the peritoneal fluid of EMS patients is higher than that of normal women 43 . TF had also been proved to be overexpressed in ectopic lesions 44 . Lin et al. found that TF was overexpressed in epithelial cells and stromal cells of EMs 45 . Graciela et al. found that markedly increased expression of TF in glandular epithelial cells of ectopic and eutopic endometrium derived from patients with EMS 25 . It is generally known that TF could combine with circulating factor VII-a to participate in the exogenous coagulation pathway. Therefore, we considered that TF might be a bridge link between complement and coagulation in EMS. Two mechanisms might account for such an association. Firstly, when membrane attack complex (MAC) was anchored on the cell surface, MAC could stimulate oxidation of protein disulfide isomerase and induce exposure of functionally active procoagulant TF 46 . Secondly, MAC could enhance the platelet procoagulant reaction by generating procoagulant particles or inverting the membrane to expose phosphatidylserine 47 . Importantly, in addition to participating in blood clotting, activated platelets could be recruited to lesions and derive TGF-β to promote the development of diseases 32,48 .
Moreover, TF could induce the formation of blood vessels through AKT-ETS1 and promotes the stability of blood vessels through PAR2-SMAD3, which might be beneficial to the survival of lesions 49,50 . Nowadays, TF targeted treatments have shown a good therapeutic effect, which could significantly inhibit the EMS lesion 51-53 . In our experimental verification, TF was also found to be overexpressed in EMS. Combined with co-expression www.nature.com/scientificreports/ analysis, we demonstrated that the up-regulated expression of complement factors was positively related to TF. The result indicated the existence of crosstalk between the complement and coagulation cascade. So far, however, there is still little evidence focused on the interaction of complement and coagulation cascade in EMS. Subsequently, the encouraging result discovered by our study may provide a novel direction for future researches. Besides, SERPING1 and P2RY14 were also genes of great significance in our analysis. SERPING1 is the coding gene of C1NH, which is the inhibitor of the complement pathway 54,55 . And, SERPING1 was found to be overexpressed in EMS synchronized with C1S, C1QA, C1R, and C3, which showed a dual tendency to proinflammatory and anti-inflammatory. It is widely considered that EMS is a chronic inflammatory disease while Zhou et al. proposed that anti-inflammatory was equally important in EMS and that Anti-inflammatory factors involved in EMS could promote the growth, invasion, angiogenesis and immune escape of lesions 56 . This phenomenon has also been reported at the cell level. In general, the M2 macrophage phenotype accounts for anti-inflammation, whereas the M1 macrophage phenotype is proinflammatory. Interestingly, the ratio of M2 increased as EMS progressed 57 . Up-regulated expression of SERPING1 suggested that there may be protective mechanisms for ectopic lesions in vivo. P2RY14 is a membrane receptor for UDP-glucose, which is exclusively expressed in the epithelium. It could induce innate mucosal immunity in the female reproductive tract independent of the pathogen recognition pattern 58 . In our research, P2RY14 was down-regulated, which might be related to immune surveillance escape.
Through bioinformatics analysis, our research proposed that TF might be a significant bridge link between complement and coagulation in EMS. Meanwhile, we proved the abnormal state of complement in EMS again. To the best of our knowledge, this is an important study concentrating on the interaction of TF and the central complement factor in EMS. However, our study inevitably had certain limitations. Firstly, the amount of data published in GEO was limited, which might lead to potential errors or bias. In subsequent studies, the samples should be further expanded to reduce bias. Secondly, there were regional differences in data sources, so the research results might not apply to all patients. Hence, an in-depth research is needed to be verified.

Conclusion
Taken as a whole, through serial bioinformatics analysis and verification experiments, we confirmed the abnormal status of complement in EMs. And then, SERPING1, P2RY14, C1S, C1QA, C1R, and C3 might become potential immunotherapy targets for EMs. Besides, TF connected the central complement factors and coagulation cascades and was confirmed to be positively related to complement. Hence, the crosstalk between the complement and coagulation might play an essential role in the development of EMs.