Integrated Multi-omics Data Study Reveals The Novel Molecular Subtypes Of Prostate Cancer And Explores Its Implications For The Clinical Guidance Of AR Signaling Pathway Inhibitors

Background and Objectives: Prostate cancer（PCa）is the most frequent malignant tumor in male,and also as the fifth leading matter of cancer-specific death places notable burdens on medical resources.Even if the risk of PCa can be stratified by combining prostate-specific antigen,clinical stage,and Gleason score,low-risk PCa still has the potential to progress to high-risk PCa,which limits the overall management of PCa patients.Recently,due to the general heterogeneity of genomes,transcriptomes and proteomes among individuals,the proposal of tumor molecular subtype has provided new ideas for precise treatment and prognostic evaluation of PCa.However,most of the established molecular subtype of PCa only focuses on one or few omics data,or ignores the heterogeneity among the genomes of PCa patients with different genetic backgrounds.Therefore,it is essential to precisely identify the specific molecular features and judge potential clinical outcomes from the details of multi-omics.Material and Method: In the current study,we first identified the prostate cancer multi-omics classification（PMOC）system derived from the data of m RNA,mi RNA,lnc RNA,DNA methylation,and gene mutation from the TCGA-PRAD cohort,using a total of ten leading edge clustering principles.And were further verified in cohorts of GSE116918,GSE70770,MSKCC,GSE54460,and Anhui medical university-patients cohort（AHMU-PC）.At the same time,the specific characteristics of each subtype were revealed by the method of single-sample gene set enrichment analysis to explore the application value of the "PMOC" system in the clinical treatment and prognosis evaluation of PCa.Results: Three PMOC subtypes associated with relapse-free survival were identified.Compared with PMOC 1 subtype and PMOC 3 subtype,PMOC 2 subtype had the highest proportion of patients with high Gleason score（≥ 9 points）（61.8% vs.23.0%vs.9.7%,P = 0.012）and with poor pathological T（≥ T3 stage）（86.8% vs.54.9% vs.52.6%,P < 0.001）.Patients in the PMOC 2 subtype had the shortest recurrence-free survival,and even after adjusting for confounding bias due to clinical factors,they showed the worst clinical outcomes（P < 0.05 in all cohorts）.Further analysis showed that in the PMOC 1 subtype,TNF-α,IL6/JAK/STAT3 and IL2/STAT5 signaling pathways were activated,and in the PMOC 2 subtype,G2 M checkpoint,E2 F target pathway,two MYC target pathways and DNA repair pathways were significantly activated,while in the PMOC 3 subtype,the activated pathways includes hypoxia,angiogenesis,epithelial mesenchymal transition,and PI3K/AKT as well as IL6/JAK/STAT3,IL2/STAT5,Notch and TNF-α signaling.Notably,diverse metabolic pathways and immune cell infiltration differences were also present in the tumor microenvironment of the three subtypes.In addition,the number of patients with gene mutations was the largest among the PMOC 2 subtypes,and the total tumor mutational burden was the highest（TP53: 23.6%,SPOP: 18.7%,APC: 5.7%,all higher than PMOC 1 subtype and PMOC 3 subtype,and P < 0.05）.Additionally,the luminal（Lum B）/PMOC 2 subgroup has higher infiltration of anti-inflammatory immunocytes,M2 macrophages,Tregs and follicular helper T cells,as well as the poor clinical outcome compared with Lum B/PMOC 1 + 3 subgroups.Further analysis reveals PMOC 3 subtype linked with the activation of the androgen response and the high response rate of AR signaling inhibitor treatment,which was initially validated in a real-world cohort.Conclusions: Taken together,we defined the PMOC system for PCa patients via multi-omics data and consensus results of ten algorithms.The PMOC 1 subtype,the tumor-inflammatory subtype,contains the highest expression levels of immune checkpoint proteins,moderate activated immune-associated pathways,and its patients are more likely to benefit from immunotherapy.The PMOC 2 subtype,tumor-activated subtype,characterized by the activated cell cycle pathway,activated DNA repair pathway,high gene mutation rate and copy number alteration of chromosome 8q24.21 region,and its patients have the worst outcome progression.The PMOC 3 subtype is likely to be a tumor-balance subtype,with the activated oncogenic signaling pathways as well as the activated proinflammatory pathways,its patients have the longest relapse-free survival and might benefit from ARSI.It has been verified that this multi-omics consensus PCa molecular classification is not limited by the racial heterogeneity of PCa,and it can not only identify the middle and high risk population of Lum B subtype,but also be expected to provide potential guidance for the clinical application of AR signaling inhibitors.