| Being the second most common type of cancer in men,prostate cancer is gradually becoming a major clinical burden.It is estimated that over 350,000 people are died from prostate cancer each year,and the number of new diagnosed cases for prostate cancer is increasing worldwide.Due to the elevated incidence and mortality of prostate cancer,there is an urgent need to determine the key mechanisms of disease development,and then identification of potential biomarkers and therapeutic targets for disease progression to reduce mortality in patients.Prostate cancer is asymptomatic in the early stages of the disease,so it is difficult to be diagnosed,and advanced prostate cancers that progress to tumor metastases are often considered incurable or difficult to treat.Current diagnosis relies on prostate-specific antigen(PSA)testing,but this misses some aggressive tumours,and leads to the overtreatment of non-harmful disease.Hence,there is an urgent unmet clinical need to identify new diagnostic and prognostic biomarkers.As prostate cancer is a heterogeneous and multifocal disease,it is likely that multiple biomarkers will be needed to guide clinical decisions.In response to this clinical problem,we aimed to explored new biomarkers for prostate cancer in terms of long-noncoding RNA and m~6A modification,and hoped to build a comprehensive diagnostic scoring model based on these biomarkers.To achieve this,multi-omics data mining(at transcriptome and epigenome levels)was performed based on TCGA prostate cancer cohort.First,an lnc RNA-m RNA co-expression network was constructed by weight correlation network analysis.Two unknown lnc RNAs(LINC00683 and LINC00857)and a new m RNA(CCDC178)were discovered as potential prognostic factors for prostate cancer.Next,we revealed that lnc RNA LINC00683 was highly correlated with GNAO1 using Pearson’s correlation analysis,and the lnc RNA might be involved in the Wnt pathways and transcription process by influencing GNAO1.Then,in this study,we systematically analyzed the expression of 14 widely reported m~6A methylation regulators in 551 prostate cancer sample from m~6A modification perspective,as well as the association between clinicopathological characteristics.A diagnostic model of prostate cancer was constructed by a three-gene risk signature(YTHDF2,METTL14 and HNRNPA2B1).Multivariate Cox analysis and ROC analysis showed that it is not only an independent prognostic factor but can also predict the clinicopathological features of prostate cancer.In particular,two molecular subgroupss of prostate cancer were identified by unsupervised consensus clustering.And we proved that METTL14 has important prognostic value in the prostate cancer for the first time.Finally,we tried to construct a comprehensive diagnostic score model with better predictive performance by using the previously identified prognostic biomarker for prostate cancer(3 lnc RNA+2m RNA+3 m6A regulator).It was found that the LASSO-Cox regression model based on these six risk genes(LINC00683,LINC00857,FENDRR,CCDC178,SERPINA5,and HNRNPA2B1)had a better prediction effect in the clinical diagnosis of prostate cancer,which was reflected in the higher accuracy(AUC=0.827),the clinical diagnosis potential is greater.In addition,this study also attempted to combine m~6A modification with mechanical microenvironment for the first time,hoping to explore the bone metastasis mechanism of prostate cancer mediated by m~6A methylation regulators under fluid shear stress.In order to simulate the effect of real blood shear stress on prostate cancer metastasis,two fluid shear stress loading devices(Circulation flow chamber,Microfluidic chips)were designed and built for crawling and transmigration,which will lay a solid foundation for the follow-up exploration of m~6A mechanical experiments.In summary,the results of this study will help to develop new treatment methods for prostate cancer,and also provide new therapeutic targets for the development of related anti-cancer drugs. |
PDF Full Text Request