CRC is a leading cause of cancer-related mortality globally. It ranks third in terms of incidence rate among all malignant tumors worldwide 1. The prognosis of advanced CRC is poor. it is crucial to explore and validate new molecular markers for CRC diagnosis. In this study, we integrated analyzed RNA-seq data and clinical information from GEO and TCGA databases. A total of 5408 DEGs consisting of 2779 upregulated DEGs and 2629 downregulated DEGs were identified between CRC tissues and normal tissues in the TCGA-CRC dataset. Based on WGCNA, we screened the modules related to CRC. The MEbrown module was the most associated with CRC and contained 1639 CRC-associated genes, of which 926 CRC-associated genes were differentially expressed. Subsequently, we conducted a great predictive capability of prognostic risk model. By LASSO and multifactorial Cox analysis, we obtained five characteristic genes of CRC, including TIMP1, PCOLCE2, MEIS2, HDC, and CXCL13. The Kaplan-Meier analysis showed that patients with high-risk scores had significantly lower OS than those with low-risk scores. The time-dependent ROC analysis for OS obtained an AUC of 0.7 which indicated relatively high specificity and sensitivity of the prognostic signature for CRC. This prognostic risk model still had strong predictive power in both the internal dataset and the GSE39582 external dataset. In addition, the DCA and PCA results demonstrated the strong predictive power of the prognostic risk model in the training set, internal dataset, and external validation set. These five characteristic genes were also differentially expressed in GSE32323 dataset, further demonstrating the superiority of the model.
The functional enrichment analysis demonstrated that the DEGs were involved in some biological processes, such as cell adhesion, cell communication. In accordance with previous research, these biological processes play crucial role in CRC tumorigenesis and development 21. KEGG analysis on DE-CRC-related genes revealed that they were also enriched in cGMP-PKG signaling pathway, circadian rhythm, cAMP signaling pathway, and PI3K-Akt signaling pathway related to CRC, the results were consistent with previous knowledge. According to Zhan Ma et al, PHLDA2 regulates epithelial-mesenchymal transition (EMT) and autophagy in CRC via the PI3K/AKT signaling pathway 3. Si-Yang Li et al has demonstrated that Diosgenin suppresses CRC cells through cAMP/PKA/CREB pathway 22. Our research revealed that these DEGs may be implicated in the tumorigenesis and development of CRC.
Univariate, multivariate Cox regression, and LASSO analyses all show that TIMP1, PCOLCE2, MEIS2, HDC, and CXCL13 are significantly related to CRC patient’s survival. According the previous studies, PCOLCE2 encodes a functional procollagen c-protease enhancer, and it can promote the enzymatic cleavage of type I procollagen to produce mature structured fibrils 23. Our findings are consistent with previous research, Chen et al. developed a prognostic gene signature made up by 9 genes, including PCOLCE2 and T1MP1, and they accurately predicted the overall survival in CRC patients 24. Although the specific mechanism of PCOLCE2 in CRC is less known, according to recent research, PCOLCE2 has been identified as the main gene driving the development of endometrial cancer, and our findings are supported by this research 25. MEIS2 is a member of the MEIS protein family that regulates neural crest and limb development 26, and it has been implicated in the development of human cancer 27. Recent research showed that in prostate cancer and ovarian cancer, the degree of MEIS2 protein expression was related to the development of clinically metastatic illness and the absence of biochemical recurrence 28,29. Ziang Wan et al. has firstly demonstrated that the MEIS2 promotes cell migration and invasion in CRC, and acts as a promoter of metastasis in CRC 30. Histamine dihydrochloride (HDC) is an inhibitor of NOX2-derived ROS 31, and exerts anti-cancer efficacy in experimental tumor models. Hanna et al. propose that anti-tumor properties of HDC may comprise the targeting of MDSCs 32. In addition, Chen et al. demonstrated that HDC + granulocytic myeloid cells influence CD8 + T cells both directly and indirectly through modulating Tregs, and which hence appear to play crucial roles in suppressing tumoricidal immunity in murine colon cancer 33. CXCL13, a homeostatic chemokine, is secreted by the stromal cells in the B-cell area of the secondary lymphoid tissues. CXCL13 plays a significant part in the growth of tumor34. The previous study demonstrates that CXCL13 can promote prostate cancer cell proliferation through JNK signalling and invasion through activation of ERK35. The same findings as our study, Qi XW et al, indicate that CXCR5 and CXCL13 appear to be independent predictors of survival for patients with CRC36. Senlin Zhao et al demonstrated that polarized M2 macrophages could induce premetastatic niche formation and promote CRLM by secreting CXCL13, which activated a CXCL13/CXCR5/NFκB/p65/miR-934 positive feedback loop in CRC cells37.
Among these genes, TIMP1 attract our attention. TIMP1 encodes a 931 base-pair mRNA and a 207 amino acid protein. This protein may inhibit the proteolytic action of matrix metalloproteinases (MMPs), which are thought to be crucial for the tumor invasion and development of metastatic disease 38,39. T1MP1 has been showed that its expression is upregulated in colon cancer 19,20, and it also plays an important role in the regulation of cell proliferation and anti-apoptotic function 40,41. A previous study indicated that TIMP1 is a key role in promoting progression and metastasis of human colon cancer, and function as a potential prognostic indicator for colon cancer 14. Through in vitro experiments, we verified the inhibitory effect of blocking TIMP1 expression on growth of colorectal cancer cells. Moreover, we investigated the apoptotic effect of TIMP1 on CRC cells, TIMP1 knocked down can promote apoptosis of CRC cells. In summary, we preliminarily investigated the biofunctions of TIMP1 in CRC, which shed lights on CRC treatment.
However, it is important to note that there are limitations to this study. Firstly, to confirm the biological functions of TIMP1 in CRC, in vivo experiments are required. Secondly, it is worth mentioning that the study did not investigate whether a TIMP1 inducer can reverse the stimulative effect of TIMP1 on the growth and apoptosis of colorectal cancer cells.