| BACKGROUND&OBJECTIVE:Bladder cancer is a common malignant tumor of the urinary system and can be divided into non-muscle invasive bladder cancer(NMIBC)and muscle invasive bladder cancer(MIBC);studies have shown that up to 45%of NMIBCs progress to MIBC,and MIBC patients require ureteral skin stoma after radical surgery,which brings inconvenience and financial burden to patients.Therefore,the mechanism of studying disease progression in patients with NMIBC is of clinical value.Studies have shown that the progress of NMIBC is related to tumor stem cells.Mature somatic cells can enhance their pluripotent differentiation and maintain self-renewal ability by inducing specific gene expression,and the degree of somatic cell dedifferentiation is positively correlated with bladder tumor tissue sternness.At present,researchers mostly study the role of single genes in cell dedifferentiation by stemness validation;however,this method is difficult to find new sternness genes and the genes that have been verified are limit This project will use machine learning algorithms to quantify the stemness genes in NMIBC tissues and further study the sternness-related regulatory mechanisms through tumor tissue gene expression,and to find score signaling pathways and transcription factors that regulate tissue sternness,that will lead to prevent NMIBC and provides potential biomarkers in this progression.MATERIALS AND METHODS:This object downloaded the NMIBC gene expression data encoded as "EGAS00001001236" from the EGA database official website,which contains 460 NMIBC samples and 16 MIBC samples;we also downloaded the corresponding clinical follow-up information online.Scored the NMIBC full-spectrum gene using a machine learning algorithm and the cox regression model was used to screen genes associated with disease progression.The stemness gene with high correlation coefficient was regarded as the core stemnes gene according to the Pearson correlation coefficient.Subsequently,the gene expression characteristics of key sternness genes expression between different samples were studied after group homogenization.The appropriate algorithm was used to divide NMIBC patients into different subtypes of sternness,and the relationship between disease progression among subtypes was studied.Exploring the correlation between different organizational sternness and age,gender and various molecular types;using the Reactome database to study core stemness gene functions and their enrichment pathways and to find key genes for regulation based on spatial location of gene distribution transcription factors calculating the sternness correlation coefficient of each transcription factor and the number of key sternness genes regulated by it,screening for multiple transcription factors with high correlation with tissue sternness;studying transcription factors and expression of genes involved in their regulation between co-expressing highly consistent transcription factors as a potential factor associated with tissue sternness.Results:This study calculated the sternness of 43204 genes detected and screened 187 sternness phases,the related 7 genes which have a correlation coefficient with sternness of less than 0 are MMP2,TFEB,APCDD1 PCDHGC,LTBP4,SYNPO and PDE4A;138 genes with positive correlation with sternness and P<0.001 are defined as score stemness genes.According to the Manhattan distance algorithm and the Ward.D2 clustering method,the expression levels of these key sternness genes in NMIBC can be divided into two types:sternness high group and sternness low group,the key sternness gene co-expression was great inside the group.The survival analysis with the follow-up time longer than 60 months showed that the disease progression in high stemness group was faster than that in low sternness group(Chi-square test,P<0.001).Based on clinical pathology and molecular typing information,the stemness of the tissue was found to be independent of the patient’s age(Chi-square test,P=0.103)and gender(Chi-square test,P=0.333);tumors less than 3 cm were mainly found in the low stemness group(Chi-square test,P<0.001),and the Ta stage tumor tissue was mainly gathered in low-stemness subtypes,T2-4 is mainly concentrated in sternness tissues(Chi-square test,P<0.001);low-grade tumors were more likely to appear in sternness lower groups,while high-grade tumors were just on the opposite(Chi-square test,P<0.001).The pattern growth of low sternness group is mostly papillary,however,the solid type is more in the high sternness group(Chi-square test,P<0.001).The BCG treatment has no difference in the treatment effect of the two groups with different sternness(Chi-square test,P=0.333).Samples with EORTC risk score of 1 were mainly concentrated in the high stemness group,while those with a score of 0 were mainly gathered in the low stemness group(Chi-square test,P<0.001).Those with unstable genes and sternness in the Lund classification were in molecular typing.There was a high degree of coincidence in the basal-like type A and unscreened type in the low sternness group(Chi-square test,P<0.001).Among the 12 gene features,the high-risk group mainly coincided with the high stemness group(76.51%),in contrast,the high risk group only accounted for 37.00%and the low risk group accounted for 63%in the low sternness group(Chi-square test,P<0.001).The CIS characteristics were mainly collected in the sternness high group,but not the CIS(Chi-square test,P<0.001).In the CLASS classification,the luminal in situ carcinomas were mainly distributed in the high stemness group,the luminal and early basal samples were mainly aggregated in the low stemness group(Chi-square test,P<0.001).Based on the Reactome database,key sternness genes were concentrated in the cell cycle pathway,followed by cell cycle and Mitotic pathway,Mitotic Prometaphase pathway and Resolution of Sister Chromatid Cohesion pathway.According to the key stemness genes,the regulatory factors were searched for in the range from-2500 bps to+1000 bps in the spatial position of the chromosome,and finally 205 transcription factors regulating key stemness genes were obtained.The transcription factors regulating more than 120 key stemness genes are CEBPG,HOXB2,FLI1,MEIS1,FOXJ3,FOXJ2,ETV2,MYBL1,FOXO1,ETV5,ERF,GCM1 and TEAD4.Among these transcription factors,genes with a correlation coefficient greater than 0.3 were CEBPG,HOXB2,FLI1 and MEIS1.The number of key stemness genes co-regulated by four transcription factors reached 111 with a ratio of 84.09%.Comparing these four regulatory factors with the expression of 111 genes regulated by them,HOXB2 has the highest correlation coefficient with the genes regulated by it.Conclusion:The machine learning method can be used to quantify the sternness of NMIBC tissues.It is found that patients with high sternness are more likely to progress and have shorter survival time.Clinical pathology shows that high stemness tissue has higher stage and higher grade.Most of the samples with high EORTC risk score were clustered in the high stemness group.The frequency of genomic instability was higher than that in the low sternness group.The high-risk group based on the 12-genome typing mainly coincided with the high sternness group.Those results revealed that the molecular typing results and organization sternness grouping are highly consistent,further,which suggested tissue sternness scoring may be a new marker of tissue dedifferentiation.The key sternness genes mainly collects in the cell cycle pathway and the genomic instability during cell division may be the cause of tissue dedifferentiation;84.09%of the key stem genes are regulated by HOXB2,so it is speculated that HOXB2 may be bladder cancer stemness maintenance and key regulators of tissue development.The results of the study found new potential sternness biomarkers for bladder cancer progression,but subsequent experiments are needed to validate the relevant conclusions.Figure 14 table 1 reference 81... |
PDF Full Text Request