| BackgroundOvarian Cancer(Ov Ca)ranked as one of the first lethal gynecologic malignancies.Ovarian cancer has a hidden onset and is difficult to diagnose in the early stage.Due to the lack of reliable biomarkers,the early diagnosis and evaluation of prognosis of ovarian cancer is a essential clinical problem which is need to be solved urgently.Although the targeted therapies for ovarian cancer is in continuous development,a significant proportion of patients still fail to benefit from the existing targeted therapies because of occurrence of low response rates and drug resistance.Thus,the research of biomarkers which can help to diagnose and predict the prognosis of ovarian cancer patients in the early stage become an urgent problem to be solved in the field of ovarian cancer diagnosis and treatment.In addition,the screened biomarker is expected to be a target for the treatment of ovarian cancer and applied in clinical work.Bioinformatics is a new interdisciplinary subject formed by the combination of life science and computer science.In recent years,high-throughput molecular technology has been widely applied in the field of tumor research.Therefore,a large amount of research data has been generated.With the establishment and continuous improvement of GEO,TCGA and other global tumor databases,the information extraction and in-depth analysis of tumor databases have become the focus of tumor research.Using bioinformatics methods to do data mining in tumor databases,is a efficient way to obtain effective biomarkers and filter the signal pathway,which will reveal the underlying mechanism of tumorigenesis and progression.Using this method can also greatly improve the screening efficiency and reliability of screening the biomarkers which have the potential to be diagnosis,prognosis molecules or therapeutic targets.Therefore,in this project,we selected ovarian cancer gene chip in the GEO database and screened the differentially expressed genes,in combination with a variety of bioinformatics methods screening hub genes from these differentially expressed genes.We proposed that the hub genes would help to evaluate the survival of patients and predict the prognosis.In addition,we hope to learn more about the role the screened gene play in the tumorgenesis and progression of ovarian cancer through the research of gene function and regulation network.Thus,our research will provide theoretical basis for the screened gene to be a diagnostic,prognostic factor or further as a therapeutic target which is applied to clinical work.Materials and methodsGEO ovarian cancer gene chip data mining,differentially expressed gene analysis and target gene screening1.Retrieval and download the data of four selected GEO ovarian cancer gene chip;2.R studio software was used to conduct quality analysis of gene chips from the multiple aspects as gray scale,weight,residuals,residuals signs,RLE,NUSE and RNA degradation in a R 3.5.1 environment;3.R studio software was used for data processing of GEO ovarian cancer gene chip.After removing unqualified samples,RMA algorithm was used for background correction,standardization and log2 processing;4.Differentially expressed gene(DEG)analysis of gene chip data was conducted using R software and limma package;5.The online functional enrichment tool DAVID was used to conduct GO and KEGG enrichment analysis on the screened DEGs;6.The proliferation-ralated DEGs were screened according to the results of GO enrichment analysis;7.Kaplan-meier Plotter public data platform was used for survival analysis of certain candidate gene and drawing survival curve;8.Oncomine public data platform was used for the verification the overexpression of the candidate gene in ovarian cancer tissue;9.GEPIA database was used to analyze the expression of candidate genes in ovarian cancer at different stages.Bioinformatic and experimental validation of KIF15 expression in ovarian cancer tissue and functional research of KIF15 both in vitro and in vivo1.The RNA-seq data from GTEx database and TCGA database was used to analyze the KIF15 expression level in female normal tissues and the differential expression between normal ovarian and ovarian cancer tissues;2.The WGCNA method and m RNAsi score were used to reveal the correlation between KIF15 expression and the proliferation of cancer stem cells;3.Immunohistochemistry(IHC)method was used to detect the expression of KIF15protein in ovarian cancer tissue chips and the immunohistochemical staining results was evaluated by score;4.The Cancer Cell Line Encyclopedia(CCLE)tumor Cell database was used to find the KIF15 m RNA expression data of a total of 55 ovarian Cancer Cell lines.Graphpad Prism6 software was used to construct the expression profile map of KIF15 m RNA in ovarian Cancer Cell lines;5.Real-time PCR was used to detect the expression of KIF15 m RNA in various tumor cell lines including Ovarian cancer cell line SKOV3,A2780,HO8910 and Ovcar-3,cervical carcer cell line Hela,Siha and C33a,lung adenocarcinoma cell line A549,pancreatic cancer cell line PANC-1 and neuroglioblastoma cell line U87.SKOV3 and HO8910 cell lines were used in subsequent experiment;6.Sh RNA lentivirus transfection was conducted on ovarian cancer cell lines,and lentivirus transfection efficiency was detected by real-time PCR and Western Blot;7.Celigo Cell count and Cell Counting kit-8(CCK8)method were used to detect Cell proliferation after KIF15 knock-down;8.FACS and caspase3-7 method was used to detect apoptosis after KIF15knock-down;9.Tumor formation experiments and animal imaging in nude mice were performed to observe the tumor formation.Anti-KIF15 immunohistochemical staining was performed on the tumor resected from the xenograft model.mRNA expression profiling and bioinformatic analysis of KIF15-KD ovarian cancer cells1.Lentiviral transfection on SKOV3 cells was performed and RT-PCR method was used to evaluate the efficiency of RNAi;2.The quality of total RNA was evaluated by A260/A280 ratio,RIN value and28s/18s ratio;3.The 3'IVT reaction method was used to conduct the gene chip detection;4.R studio software was used to analyze the quality of gene chip from the aspects of gray scale,weight,residual,residual signs,RLE and NUSE;5.Raw data processing and differentially expressioned gene analysis was performed using R studio software and limma package;6.Funrich,Clue GO and GSEA were used to perform functional and pathway enrichment analysis and screen apoptosis-related pathways;Phospho-antibody array detection and bioinformatic analysis of KIF15-KD ovarian cancer cells1.The phospho-antibody array was used to detect the phosphorylated protein in ovarian cancer cells after KIF15 knock-down;2.The proteins on the apoptosis-related pathways Apoptosis,TNF signaling pathway and NF kappa B signaling pathway were selected to construct a core protein network;3.The consistent genes or proteins on the three apoptosis-related pathways were selected to verify the existance of crosstalk between pathways.ResultsGEO ovarian cancer gene profiling data mining,differentially expressed gene analysis and candidate gene screening1.The four GEO gene chip dataset selected in this study is reliable in quality and can be used in the subsequent data analysis;2.The criterion which was used to screen differentially expressed genes was more than 1.5 times of Log Foldchange(Log FC).The data set of GSE40595 contained 2,544up-regulated genes and 52 down-regulated genes.The data set of GSE18520 contained 582up-regulated genes and 580 down-regulated genes.The data set of GSE38666 contained1,487 up-regulated genes and 205 down-regulated genes.The GSE36668 dataset contained2216 up-regulated genes and 108 down-regulated genes.Overlapping analysis of up-regulated differentially expressed genes and down-regulated differentially expressed genes(Log FC greater than 1.5 or less than-1.5)in 4 data sets revealed that 183 genes were up-regulated and 7 genes were down-regulated;3.By selecting in the ten GO categories with the smallest p value,we found five proliferation-related GO categories as“cell division”,“mitotic nuclear division”,“cell proliferation”,“mitotic sister chromatid segregation”and“chromosome segregation”.Among 42 genes within the 5 GO categories,a total of 17genes(SAC3D1,NUF2,FAM83D,TPX2,KIF11,ZWINT,CDCA3,NDC80,PTTG1,BUB1B,KI F15,KIF18B,SPAG5,CENPF,CDC20,CDK1 and KIF2C)was included in two or more proliferation-related GO categories simultaneously;4.Kaplan-meier Plotter was used to analyze the survival rate of the 17proliferation-related genes and draw survival curves.BUB1B,CDK1,CENPF,FAM83D,TPX2 and KIF15 were negatively correlated with Overall Survival(OS,p<0.01)and Progression free Survival(PFS,p<0.05)of ovarian cancer patients;5.The overexpression of the 6 candidate genes in ovarian cancer was verified in the Oncomine database.In the ovarian cancer datasets of the Oncomine database,the differential expression of the 6 candidate genes are all supported by different datasets.TCGA dataset was included in the supporting data sets of 5 candidate genes except FAM83D.As the differential expression of FAM83D gene in ovarian cancer was only supported by one dataset(Yoshihara),FAM83D gene was abandoned for further analysis,considering the authority of the TCGA dataset.Survival analysis was performed using Kaplan-meier Plotter.KIF15 m RNA expression was negatively correlated with with progression-free Survival(PFS)(HR=1.15,95%confidence interval 1.01-1.3,Log Rank P=0.037)and Overall Survival(OS)(HR=1.21,95%confidence interval 1.05-1.39,Log Rank p value=0.0087)in ovarian cancer patients with statistically significance(p<0.05);6.The m RNA expressions of BUB1B,CDK1,CENPF,TPX2,FAM83D and KIF15were significantly negatively correlated with the PFS and OS of early-stage ovarian cancer patients.These genes were overexpressed in stage I-II ovarian cancer patients with a higher HR value than ovarian cancer patients in general;7.In GEPIA database,we found that the expression of BUB1B,CENPF and KIF15 in the Stage II ovarian cancer patients are higher than that of Stage III-IV.KIF15 expression in ovarian cancer tissue and functional research in vitro and in vivo1.Except for bone marrow,KIF15 m RNA was expressed at a low level in other normal tissues;KIF15 was overexpressed in ovarian cancer tissue compared to normal ovarian tissues;2.KIF15 was positively associated with the proliferation of cancer stem cells;3.Anti-KIF15 immunohistochemical staining was performed on ovarian cancer tissue chips and the protein was only exsist in the cytoplasm.In a total of 90 tumor tissues of different pathological types,68 cases(75.6%)showed high KIF15 protein expression and22 cases(24.4%)showed low expression.In a total of 10 cases of adjacent tissues,only 2cases(20%)showed high expression and the rest of 8 cases(80%)all showed low expression.Compared to adjacent tissues,the KIF15 protein expression in ovarian cancer tissue are higher with a statistically significance(P<0.05);4.Among the selected 10 cancer cell lines,KIF15 m RNA expression of SKOV3 was the highest,and HO8910 expression was the lowest;5.The RNAi efficiency of Sh RNA lentivirus on SKOV3 was 83.1%and that of HO8910 was 61.0%,which could meet the needs of subsequent experiments;6.Celigo cell count results:The amount of SKOV3 cells at Day5 was compared with that of Day1.In Sh Control group,proliferation rate was 7.63±0.11 times,while in Sh KIF15group the proliferation rate was 1.24±0.03 times.Significant differences exsist between the groups(t test,P=6.82701×10-8).The proliferation rate of Sh Control of HO8910 cells was5.65±0.19 times,and that of Sh KIF15 group was 1.72±0.07 times.Significant differences also exsist between the groups(t test,P=7.10173×10-11);7.CCK8 test results:The OD value of SKOV3 cells at Day5 was compared with that of Day1.The fold change of OD value in SKOV3 Sh Control group was 4.095±0.0294 times,while that of Sh KIF15 group was 2.483±0.038 times.In the HO8910 Sh Control group,the fold change of OD value was 5.11±0.0291 times,while that of Sh KIF15 group was2.416±0.0268 times,showing significant differences(t test,SKOV3 P=1.11593×10-12,HO8910 P=3.85188×10-15);8.Apoptosis results detected by FASC method:Apoptosis of SKOV3 and HO8910cells was detected 72h after KIF15 depletion.The apoptosis proportion of SKOV3 cells in Sh KIF15 group was 6.4±0.255%,while that of Sh Control group was 2.98±0.192%.The apoptosis proportion of HO8910 Sh KIF15 group was 10.58±0.577%,while that of Sh Control group was 4.03±0.142%.Both the apoptosis proportion of two cell lines in the Sh KIF15 group was significantly higher than that in the sh Control group(t test,SKOV3 P=5×10-5,HO8910 P=4×10-5);9.Apoptosis results detected by caspase-3-7 method:The intensity of this luminous signal was proportional to the activity of caspase-3/7.In both SKOV3 and HO8910 cells,the luminous signal intensity(i.e.the indicator of Caspase activity)of sh KIF15 group was significantly higher than that of Sh Control group(t test,SKOV3 P=8.47283×10-4,HO8910 P=2.83606×10-4);10.Animals imaging:All the 10 nude mice xenograft model in the NC group injected with Sh Control SKOV3 showed fluorescence expression at different intensity,indicating subcutaneous tumor had formed.And a total of 10 nude mice in the KD group injected with Sh KIF15 SKOV3 showed fluorescence signal only on the No.13 nude mice xenograft model,while no obvious fluorescence signal was observed in the other nude mice;11.Subcutaneous tumor formation in nude mice:The tumor weight of the NC group was 1.482±0.273g.Only one subcutaneous tumor sample was obtained in the KD group and the tumor weight was 0.061g.Significant difference exsisted between the the NC group and the KD group(t test,P=8.442×10-11)in tumor weight;12.The anti-KIF15 immunohistochemical staining of tumor paraffin section of NC group showed tan while that of KD group showed light yellow,suggesting that the expression of KIF15 protein in tumor of KD group was significantly lower than that of NC group.Results of m RNA expression profiling and Bioinformatics analysis in KIF15knock-down ovarian cancer cells1.Total RNA quality analysis:the 6 samples in the project had a quite high RNA purity,good integrity and qualified quality;2.Quality analysis of gene chips:the gene chips and samples used in this study are of reliable quality,which is also the guarantee of reliability of data analysis in subsequent experiments;3.Analysis results of differentially expressed genes:After KIF15 knockdown treatment on SKOV3,three samples of KD group and three samples of NC group were compared with each other.The differential expression genes were analyzed and screened by the significant difference criterion of|FC|greater than 1.5 and p-value<0.05;134up-regulated genes and 309 down-regulated genes were screened;4.GO and KEGG enrichment analysis by Funrich:In upregulated genes,it was observed that the main biological processes of the differential expression genes are concentrated on nucleic acid metabolism,cell communication and signal transduction.Signal pathways are mainly concentrated on tumor related pathways and cell adhesion related pathways such as Er Bb,TRAIL and PI3K/Akt.Among the downregulated genes,the three biological processes enriched the highest proportion of genes were protein metabolism,anti-apoptosis and protein localization.Four of all the first 11 pathways were associated with apoptosis,as the apoptotic factor-mediated reponse,SMAC-mediated dissociation of IAP:caspase complex,SMAC binding to IAPs and SMAC-mediated apoptotic response;5.Functional and pathway enrichment analysis by Clue GO:Apoptosis,TNF signaling pathway and NF kappa B signaling pathway were the main and apoptosis-related pathways among the DEGs;6.Functional and pathway enrichment analysis by GSEA:Apoptosis and TNF?signaling via NFkb were the apoptosis-related hallmarks of the DEGs;Results of Phospho-antibody array detection and Bioinformatics analysis in KIF15 knock-down ovarian cancer cells1.At the protein level,KIF15 knock-down leaded to the activity of the apop-related pathway as Apoptosis,TNF signaling pathway and NF kappa B signaling pathway;2.Apoptosis,TNF signaling pathway and NF kappa B signaling pathway had a crosstalk with each other at both m RNA and protein level and constructed a regulatory network of apoptosis;3.MAP3K14 and TNF were the consistent gene of the three apoptosis-related pathway at m RNA level;4.NFk B-p105/p50(Phospho-Ser337),IKK-beta(Phospho-Tyr199),NFk B-p65(Phospho-Ser529)and IKK-gamma(Phospho-Ser31)were the consistent proteins of the three apoptosis-related pathway at protein level;These proteins were all significantly phosphorylated,indicating a functional activity;5.The apoptosis-related regulatory network contained both pro-apoptotic pathways and anti-apoptotic pathways.ConclusionWe have identified KIF15 as a proliferation-related biomarker with early prognostic and pathological diagnostic value in ovarian.Reduced KIF15 inhibited proliferation and promoted apoptosis of ovarian cancer cells.The apoptosis of OC cells was activated through the crosstalk among apoptosis pathway,TNF signaling pathway and the NFkb signaling pathways.Therefore,KIF15 may act as a potential therapeutic target of ovarian cancer. |
PDF Full Text Request