Multi-omics profiling of metastatic colorectal cancer reveals the transcriptional network of focal adhesion and immune suppression and the role of p-RPS6.

BACKGROUND: Colorectal cancer (CRC) is one of the deadliest malignancies worldwide characterized by rapid progression, high metastasis propensity. Our study aimed to identify the driving biological factors of metastatic CRC. METHODS: We obtained frozen tumor tissues of 8 metastatic CRC (mCRC) patients and 10 non-metastatic CRC (nmCRC) patients from Ruijin Hospital for proteome analysis. FFPE tumor and adjacent normal tissues of another 8 metastatic CRC patients and 8 non-metastatic CRC patients were collected for transcriptome and whole exome sequencing. Mutational burden and signatures were revealed and differentially expressed genes and proteins were analyzed. Molecular Complex Detection was used to build the core network. KEGG and GO pathway enrichment analysis were performed. IHC staining against p-RPS6 and subsequent quantification were performed on human samples. RESULTS: We identified 53,917 SNPs by WES with a median of 1154 variants per sample and 23.08 mutations per megabase (Mb). We observed the mutation burdens were similar between mCRC tumor and nmCRC tumor tissues (p = 0.57), as well as the mutation frequencies of HRR and MMR related genes. All mCRC samples were affected by RTK-RAS, NOTCH and WNT pathway mutations. We constructed a 16-hub-gene network in mCRC which was characterized by dis-modulation of cell adhesion (SELE, SELL and SELP) and immune exhaustion (CXCR2, CCR7, CXCR1, CXCL13, CCL7, CCL19, CXCL11 and CD19) in mCRC tumor microenvironment. We detected 22 differentially expressed proteins, 54 phosphorylated proteins and 6 tyrosine-phosphorylated proteins between mCRC and nmCRC tumors. Phosphorylated RPS6 (p-RPS6) was the most differentially expressed protein between mCRC and nmCRC tumor tissues, which was found to be positively correlated with EMT proteins and poor prognosis in CRC. The IHC staining against p-RPS6 on human samples supported the strong expression in mCRC tumor samples. CONCLUSIONS: We identified the key transcriptome network in mCRC and confirmed the important role for RPS6 phosphorylation in mCRC. Our study suggested that the features of mCRC tumors were not driven by gene mutations. We revealed the EMT feature and immune exhaustion of the mCRC tumor microenvironment.