Screening Of Key Immune Genes And Analysis Of Immune Infiltration In The Endometrium Of Patients With Recurrent Implantation Failure

Background: Recurrent implantation failure(RIF)is a complex clinical condition involving multiple etiologies and mechanisms,and the lack of specific markers for early diagnosis makes it a thorny problem in the field of assisted reproduction.Data show that RIF affects 10% of couples undergoing assisted reproductive technology(ART)to conceive.Several domestic and foreign studies have shown that immune factors play an important role in embryo implantation and maintenance of maternal-fetal immune tolerance.With the development of gene sequencing technology,numerous studies have identified altered gene expression profiles in the endometrium of RIF patients,but these studies usually have small sample sizes and lack a comprehensive analysis of endometrial immune cell and immune function infiltration.Therefore,this study analyzed multiple RIF related sequencing datasets in public databases,systematically evaluated the infiltration of endometrial immune cells and functions,and explored the different immune subtypes that RIF may exist,providing a theoretical basis for the early diagnosis and treatment of RIF.Objective: To screen endometrial immune key genes in RIF patients and analyze their immune infiltration,to search for different possible immune subtypes of RIF and to analyze the biological functions of their differential genes,in order to provide new ideas for the early diagnosis and treatment of RIF.Methods:1.Downloading the RIF endometrial tissue m RNA sequencing dataset from the Gene Expression Omnibus(GEO)database,using R software to screen differentially expressed genes(DEGs),construct weighted gene co-expression network analysis(WGCNA)to screen disease hub(hub)genes,immune genes were downloaded from Gene Cards and MSig DB databases,and RIF immune-differentially expressed genes(imm-DEGs)were obtained by intersecting DEGs,hub genes,and immune genes.2.Based on imm-DEGs,four RIF diagnosis machine learning models,extreme gradient boosting(XGB),random forest(RF),support vector machine(SVM)and generalized linear model(GLM)were constructed using the R software packages“xgboost” “random Forest” and “kernlab” and the glm function.Finally,the immune key genes were obtained from the optimal diagnostic model,and the nomogram model for predicting the risk of RIF occurrence was constructed based on the immune key genes.3.Consensus clustering analysis of RIF samples based on immune key genes was performed using the “Consensus Cluster Plus” R package to find different immune subtypes of RIF endometrium and to perform Gene Ontology(GO),Disease Ontology(DO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)analysis.Potential immune key genes for RIF were screened according to the pathway target gene network and ROC curves were plotted using the validation dataset to determine their diagnostic predictive value.4.Single-sample gene set enrichment analysis(ss GSEA)was performed using the “GSVA” R package to assess endometrial immune cell and immune function infiltration,and the rank sum test was performed to find immune cells and functions differentially expressed between RIF patients and controls and between different RIF immune subtypes.Spearman correlation analysis was used to evaluate the correlation of immune key genes with immune infiltrating cells and functions.5.Thirty endometrial tissues were collected from each of normal control and RIF patients in the secretory phase,and inflammation was assessed using hematoxylin-eosin(HE)staining,and the key immune gene PRF1 was detected by immunohistochemistry(IHC)and Western blot(WB)experiments protein levels.Results:1.A total of 54 samples were included after the dataset was merged,and 1143 DEGs were screened out by differential analysis;53 hub genes were obtained by WGCNA analysis;13519 immune genes were obtained from Gene Cards and MSig DB databases;a total of 43 imm-DEGs were obtained after the intersection of the above genes.2.Among the four machine learning models,the residual error of the XGB model is small and the diagnostic value is high,and 5 important explanatory variables(TNFAIP8L2,GPR4,FRYL,EHF,VCAM1)were obtained as the key genes of RIF immunity.3.Consistency clustering analysis yielded two different immune subtypes of RIF(cluster1 and cluster2),which were analyzed differently to obtain 522 DEGs.GO analysis showed that DEGs are mainly involved in embryonic organogenesis and development,anatomical structure and development of the reproductive system,placental development and other biological processes,and their cellular components are mainly located in the extracellular matrix containing collagen,vesicle cavity,secretory granules and other sites,and their molecular functions are mainly related to receptor ligand activity,endopeptidase regulatory activity,amino acid transmembrane transporter activity,etc.;DO analysis showed that DEGs were associated with polycystic ovary syndrome,endometriosis,coronary artery disease,etc.;KEGG pathway analysis showed that DEGs were mainly involved in JAK-STAT signaling pathway,NK cell cytotoxic activity,allograft rejection,complement and coagulation cascade signaling pathways.Five important target genes were screened from the signaling pathways,and three immune key genes(FASLG,HLA-DOB,PRF1)were obtained after validation of the dataset GSE92324.4.The evaluation of endometrial immune cells showed that MΦ,neutrophils,NK cells,Th1 cells,Th2 cells,TIL cells,and Treg cells were lowly expressed in RIF patients(P < 0.05),and MC and Th cells were highly expressed(P < 0.05);The analysis of immune cells in RIF subtypes showed that cluster1 subtype MΦ,neutrophils,NK cells,plasmacytoid dendritic cells,Th cells,Th1 cells,TIL cells,and Treg cells were highly expressed(P < 0.05).The evaluation of endometrial immune function showed that antigen presentation co-stimulation,low-grade inflammation,and T cell co-stimulation immune function decreased in RIF patients(P < 0.05);immune function evaluation in RIF different immune subtypes showed that cluster1 subtype antigen presentation co-inhibition,antigen presentation co-stimulation,CC chemokine receptors,checkpoints,cell activating factors,HLA,inflammation promotion,MHC Ⅰ molecules,low-grade inflammation,T cell co-suppression,IFN-I immune function enhancement(P < 0.05).Correlation analysis showed that the key immune gene PRF1 had a strong positive correlation with various immune cells and functions.5.HE staining found that the number of blood vessels in the endometrial tissue of RIF patients increased significantly,accompanied by immune cell infiltration,which initially suggested that the inflammatory response was enhanced.IHC and WB experiments found that the protein level of the immune key gene PRF1 was significantly increased in the endometrium of RIF patients(P < 0.05).Conclusion: TNFAIP8L2,GPR4,FRYL,EHF,VCAM1,FASLG,HLA-DOB and PRF1 can be used as key endometrial immune genes in the diagnosis of RIF,among which PRF1 is closely related to various immune cells and functions.There are two different immune subtypes(cluster1 and cluster2)in the endometrium of RIF patients,and their DEGs are related to biological functions such as substance metabolism,inflammatory response and immune response.There is abnormal immune infiltration in the endometrium of RIF patients.In the cluster1 subtype,various immune infiltrating cells increase and immune functions are enhanced;in the cluster2 subtype,various immune infiltrating cells decrease and immune functions decrease.Bioinformatics analysis shows that the key immune gene PRF1 is up-regulated in cluster1 subtype and down-regulated in cluster2 subtype.Experimental verification shows that PRF1 protein is up-regulated in RIF,which may be related to the number of samples and individual differences,and needs to be further expanded Study sample size.