Proteomics And Phosphoproteomics-based Study On The Tumorigenesis Of Colorectal Cancer

Background:Colorectal cancer is an epithelially derived gastrointestinal malignancy,which is the third contributor of cancer incidence and the second contributor of cancer-related mortality in the world.The five-year survival rate of early-diagnosed colorectal cancer patients is more than 90%,while the survival rate of patients with advanced diagnosis is less than 10%.The current solution for colorectal cancer is surgery based comprehensive treatment,including radiotherapy,chemotherapy,and targeted immunotherapy.Due to its high heterogeneity,the therapeutic effect of these treatment is very different.Therefore,it is of great significance to elucidate the mechanism and develop new biomarkers for early diagnosis,treatment as well as precision medicine for colorectal cancer.Objective:We aim to identify the key proteins and phosphorylation sites in association with tumorigenesis and development of colorectal cancer through the combined proteomics and phosphoproteomics,in order to provide evidence for clinical treatment and prognosis.Materials and methods:8 colorectal cancer patients without neoadjuvant chemotherapy were enrolled from the First Affiliated Hospital of Zhengzhou University.The total proteins of colorectal adenocarcinoma and adjacent tissues were isolated.1)The protein expression profile was obtained by shotgun proteomics using high resolution mass spctrometry;2)the phosphorylated proteomics data was obtained after TiO2 enrichment;3)R software was used to normalize the data and paired t-test to identify the deferentially expressed proteins and phosphoproteins;4)Bioinformatics analysis methods including GO,KEGG,GSVA,GSEA,and STRING were used to analyze the interaction and network of proteo-genomics,and to identify the changes of disease-related biological pathways;5)motif analysis and protein interaction network were performed to explore the phosphorylation networks;for instance,Kinase-substrate enrichment analysis was performed to construct colorectal cancer related kinase-substrate interaction network;6)Consensus clustering analysis was used to analyze the network and biological pathways enriched in the modification data,and the results were further validated by experimental data from TCGA;7)the biological pathway,mutation status and prognosis in different sub-types were compared and the immune response in sub-types were analysed through gene set enrichment and ESTIMATE algorithm;8)the random forest model was used to explore the most important predictors,and the prognosis model was established to predict the survival of patients combined with clinical features.Result:1.A total of 11595 proteins were identified in 8 pairs of colorectal cancer and adjacent tissues.After filtration with |log2FC|>1 and p<0.05,222 differential proteins were identified,including 144 up-regulated proteins and 78 down-regulated proteins.It is found that S100P,CTSH,RPS28,CGREF1,and TIMM8A are the key functional proteins.However,the expression of DMD,MYL9,PRPH,UCHL1 and other proteins were decreased in cancer tissues.In addition,HNRNPC,TRMT10C,ALYREF and YBX1 were significantly up-regulated in tumors.Pathway enrichment showed that the progression of colorectal cancer was related to RNA processing,including RNA methylation and extracellular matrix pathway.2.By phosphoproteomics 13856 phosphorylation sites were identified on 4246 proteins.After filtration with |log2FC|>1 and p<0.05,431 differentially phosphorylated peptides were identified.Among them,273 were up-regulated and 158 down regulated.The results of kinase-substrate enrichment analysis showed that MAPK8/JNK1 and MAPK9/JNK2 were abnormally activated in tumor.Combined with proteomics,it was found that JNK-AP1-MMP7 axis played an important role in colorectal carcinogenesis,which related to extracellular matrix pathway.3.According to the consensus cluster analysis based on the RNA methylation modification regulators,patients with colon cancer were divided into two subgroups.The two groups had different prognosis,immune response and gene mutation characteristics.Functional analysis showed that subgroup 1 was more active in KRAS pathway,hypoxia pathway and epithelial mesenchymal transition;Subgroup 2 was mainly manifested in DNA replication,mismatch repair and the CRD mediated mRNA stability related pathways.4.By random forest algorithm,we obtained 2 key prognostic factors:NSUN6 and DNMT2.The prognosis model based on the above genes,age and tumor stage had better accuracy of 3-,5-,and 10-year overall survival rate than TNM stage model(AUC:3-year OS:0.764 vs 0.717;5-year OS:0.716 vs 0.644;10-year OS:0.689 vs 0.574).ConclusionIn this study,the shotgun proteomics and TiO2 phosphorylation enrichment were used and 222 differentially expressed proteins and 431 phosphorylation peptides between the colorectal cancer tissues and adjacent normal tissues were identified.The pathway analysis showed the extra-cellular matrix and RNA processing including RNA methylation play an important role in tumorigenesis of colorectal cancer.The JNK-AP1-MMP7 pathway may affect the development of colorectal cancer through extra-cellular matrix related pathways.We found the two sub-types by the consensus cluster of COAD based on RNA methylation regulators,which had different biological pathways and immune responses.Finally,we identified that DNMT2 and NSUN6 were the most important prognostic factors through random forest algorithm.The two genes combined with age and tumor stage could well predict the prognosis of patients(AUC:3-year OS:0.764;5-year OS:0.716;10-year OS:0.689).