Identification Of Prognostic Biomarkers And Screening Of Targeted Inhibitors For Triple-Negative Breast Cancer Based On Bioinformatics Analysis

Objective:To identify the prognosis biomarkers and therapeutic targets for TNBC,bioinformatics analysis was used based on TCGA and GEO public databases,and computational study of virtual screening was performed to screen natural compounds as the targeted inhibitors.Methods:The gene expression profile data of 122 TNBC samples and 13 paired adjacent normal tissue samples were obtained from TCGA database.Meanwhile,the gene expression profile data of 155 TNBC samples and 33 paired adjacent normal tissue samples were obtained from GSE76250 gene microarray of GEO database.Gene differential expression analysis was performed through “Limma” package in RStudio 4.2.2,and the result showed that there were2606 differentially expressed genes in TCGA database and 931 differentially expressed genes in GEO database.Then WGCNA analysis was performed to calculate the correlation coefficient between genes,and construct hierarchical clustering and degree-free network.Genes of TCGA and GEO were separately divided into different modules according to whether there was synergistic expression between them.The module with the highest correlation with TNBC tumorigenesis was screened out as the key module.In order to obtain hub genes with differential expression and high tumor correlation,EVenn online tool was used to intersect TCGA differential genes,GEO differential genes,TCGA key module genes and GEO key module genes.The DAVID database was adopted to perform GO functional annotation and KEGG pathway enrichment analysis of the hub genes,so as to preliminarily explore the key biological processes and pathways in the tumorigenesis and development of TNBC.STRING database and Cytoscape 3.9.1 software were used to construct a protein-protein interaction network(PPI network)between the protein products of hub genes,and the most important core genes in the network were screened according to PPI scores.Kaplan Meier plotter database was used to explore the prognostic value of these core genes.Four modules in Discovery Studio 4.5 software,including Lib Dock,ADME/TOPKAT,CDOCKER and MD Simulation,were used to screen potential drug candidates targeting core genes.The structural biological properties,pharmacological properties,binding affinity and complex stability of the compounds in the natural compound database were screened and analyzed.Results:(1)A total of 2606 differentially expressed genes were obtained from TCGA database,including 754 up-regulated genes and 1852 down-regulated genes.A total of 931 differentially expressed genes were screened from the GEO database,of which 454 were up-regulated and 477 were down-regulated.The results of WGCNA showed that the module with the highest correlation with TNBC tumorigenesis in the TCGA database was Turquoise module,including 4132 genes.The key module in the GEO database was the Turquoise module,which contains 622 genes.243 hub genes with differential expression and high correlation with tumor were obtained after intersecting these four parts.(2)The results of GO functional annotation of hub genes indicated that hub genes were mainly enriched in biological processes such as cell division,chromosome segregation,mitotic cell cycle,mitotic sister chromatid separation,mitotic spindle assembly checkpoint,et al.As for molecular functions,the hub genes were enriched in protein binding,microtubule junction,ATP binding,ATP-dependent microtubule driven function,microtubule driven function.KEGG pathway enrichment analysis results showed that hub genes were significantly enriched in biological pathways such as cell cycle,oocyte meiosis,cell senescence,progesterone-mediated oocyte maturation,and Fanconi anemia pathway.(3)The interactions between protein products encoded by hub genes were constructed into a visual PPI network,and the top 10 genes(CDK1,CCNA2,BUB1,CCNB1,TTK,TOP2 A,CCNB2,ASPM,CDC6 and CDCA8)were selected according to the PPI network score.The expression levels of these 10 genes in TNBC cells were higher than those in normal cells,and their high expression was related to the shorter OS and RFS of TNBC patients,which indicated that they were prognostic risk factors for TNBC patients.(4)CDK1 was selected as the key therapeutic target,and the structural biology of candidate drugs in ZINC15 database was analyzed by Libdock module in DS 4.5 software.A total of 7696 candidate drugs that could successfully dock with the crystal structure of CDK1 were screened out.ADME and TOPKAT tests were performed on the top 20 drug candidates with Libdock scores.Four compounds(ZINC000013451339,ZINC000049784088,ZINC000008689961 and ZINC000027646086)with good safety and pharmacological properties were screened out.Subsequently,CDOCKER module was used to dock the above four compounds with CDK1 with high accuracy.ZINC000013451339 and ZINC000049784088 were regarded as the key candidate drugs of this study because they could form stable complexes with CDK1.Molecular dynamics simulations of ZINC000013451339-CDK1 and ZINC000049784088-CDK1 complexes under normal physiological conditions demonstrated that the trajectories of the two complexes reached equilibrium at 46 ps,and the potential energy and RMSD of the complexes reached and maintained at a steady state over time.Taken together,ZINC000013451339 and ZINC000049784088 were screened as promising drug candidates for CDK1.Conclusion:243 hub genes with differential expression and high tumor correlation were identified by gene differential expression analysis and WGCNA analysis based on TCGA and GEO databases.Among them,CDK1 was screened out as the most important prognostic biomarker and therapeutic target of TNBC.Two novel natural inhibitors targeting CDK1,ZINC000013451339 and ZINC000049784088,were screened from the natural compound database by computational virtual screening technology.