Tumour-promoting role of SOCS1 in colorectal cancer cells

The protein SOCS1 was first uncovered as a negative-feed back regulator of cytokine receptors that signal via the Janus family of tyrosine kinases (JAK) and the signal transducers and activators of transcription (STAT) proteins. Since then, SOCS1 is mainly recognized for its tumour-suppressing role1,2,3,4,5,6. Consistent with this idea, the gene encoding SOCS1 is frequently silenced in many cancers, such as by hypermethylation of its promoter in >50% hepatocellular carcinomas (HCC), pancreatic cancers, acute myeloid leukaemia and multiple myeloma2,7,8,9. In addition to blocking pro-inflammatory cytokine signalling, the tumour suppressor activity of SOCS1 has been linked to inhibition of hepatocyte growth factor (HGF)/Met receptor signalling and functions in hepatocytes5,10,11. Moreover, SOCS1 increases p53 phosphorylation (Ser15), DNA binding and transcriptional activity by forming a ternary complex with the DNA damage-regulated kinases ATM or ATR, which is critical for p53-dependent activation of genes involved in DNA repair, senescence and apoptosis12. These findings provide compelling evidence that SOCS1 works as a dominant tumour suppressor in HCC and underlie the molecular mechanisms involved. In sharp contrast, the role of SOCS1 in other types of cancer, including CRC, is understudied and conflicting. For example, increased SOCS1 expression have been reported in human melanoma, neuroendocrine, breast and epidermal cancers2,13,14. Inconsistent with its alleged dominant tumour suppressor role, SOCS1 expression has been linked with tumour invasion and disease stage in melanoma15,16. Likewise conflicting observations with respect to its expression and function have emerged amongst the limited experimental and clinical studies that have investigated SOCS1 relevance in CRC. For instance, SOCS1 gene methylation is rather uncommon in sporadic CRCs, ranging between 8–15% of the cases17,18. Nonetheless, methylation of the SOCS1 promoter gene, together with that of the CpG island loci of other tumour suppressor genes, is a marker of a subset of CRCs referred to as the CpG island methylator phenotype (CIMP)17,18,19. Notably, CIMP colorectal tumours are associated with specific genetic features and poor clinical outcomes20,21, but SOCS1 methylation in CIMP CRCs has been linked to better overall patient survival than those without18. Only two recently published studies have so far probed the abundance of SOCS1 mRNA or protein in relatively small cohorts of human CRC samples22,23. Their findings are somewhat contradictory. In the David et al. study, the highest SOCS1 mRNA and protein levels were seen in normal colon and early-stage adenomas, whereas the lowest levels were detected in advanced and poorly differentiated carcinomas22. Nonetheless, high SOCS1 protein level was still noted in 63% of advanced stage IV CRC tumours. Likewise, Ayyildiz et al. observed positive expression of SOCS1 in CRC tissues in nearly half of the cases by immunohistological analysis, but no association between SOCS1 protein level and clinicopathologic tumour characteristics23. Conflicting with a dominant tumour suppressor role for SOCS1 in CRC, elevated SOCS1 protein levels in CRC tumours did not predict better patient survival23. Functional relevance of SOCS1 in CRC cells remains unresolved. Mouse studies indicate that SOCS1 influences CRC progression in a cell lineage-dependent manner. While mice with Socs1 deletion in all tissues, except T and B cells, spontaneously developed colon inflammation and tumours24, its silencing in antigen-presenting macrophages and dendritic cells fostered anti-tumour immunity25,26. The role of SOCS1 in CRC cells has so far been investigated in a single published study by David et al.22. The authors concluded that SOCS1 works as a suppressor of metastasis, on the basis that its overexpression in human SW620 CRC cells reduced morphological transformation, invasion and metastasis without affecting proliferation, anchorage-independent growth or tumourigenesis22. However, an obvious increase in size of the colonies formed by SOCS1-expressing SW620 cells compared to control cells was not accounted for. Overall, these limited numbers of studies have not yet settled whether or not SOCS1 is working as a tumour suppressor or as an oncogene in CRC. In this study, we have queried the clinical relevance of SOCS1 in human CRC patients by analysing gene expression datasets from The Cancer Genome Atlas (TCGA)27. Furthermore, SOCS1-regulated functions were investigated by gain- and loss-of-function studies in murine and human colon carcinoma cell models. Our results show a predominant up-regulation of SOCS1 expression in human CRC tumours, but which did not correlate with better patient survival. Notably, we provide the first experimental evidence, both in vitro and in vivo, that SOCS1 fosters pro-oncogenic features in CRC cells.