Collagen X a-1 (collagen X) is a member of the collagen family, in which the COL10A1 gene encodes collagen involved in the formation of the extracellular matrix. Notably, recent studies have found that overexpression of COL10A1 enhances the epithelial-mesenchymal transition of tumors and promotes tumor aggressiveness and disease progression9. There is growing evidence that COL10A1 is associated with the development of several human cancers, such as breast, colon, and stomach cancers. However, the pathogenicity of COL10A1 in pancreatic cancer is unknown9,10,12. We first analyzed the expression pattern and prognosis of COL10A1 in pancreatic cancer. Combined with validation from multiple databases, the expression of COL10A1 in pancreatic cancer tissues was significantly higher than that in normal tissues. Correlation with clinicopathological features indicated that COL10A1 expression was significantly higher in the high stage/grading subgroup. Subsequently, survival curves showed that high COL10A1 expression predicted shorter survival in PAAD patients. Univariate analysis showed that COL10A1, Histologic grade, and Pathologic stage were important predictors of survival in pancreatic cancer patients. In conclusion, COL10A1 promotes the malignant transformation process of PAAD and may be a biomarker of PAAD.
Previous studies have shown that miRNAs can cause gene silencing by binding to mRNAs, while ceRNAs can regulate tumorigenesis, invasion, metastasis, and drug resistance by competitively binding miRNAs35. We used multiple online prediction tools to predict the possible upstream miRNAs of COL10A1 and finally obtained 12 miRNAs. According to the endogenous competing RNA (ceRNA) hypothesis, there should be a negative correlation between microRNAs and COL10A1. Therefore, expression correlation analysis was performed for the predicted 12 miRNAs, and only miR-144-3p was negatively correlated with COL10A1 in PAAD. Next, expression analysis and prognostic analysis were performed for miR-144-3p in PAAD. hsa-miR-144-3p was significantly downregulated in PAAD, and higher levels of hsa-miR-144-3p were associated with better overall survival. This suggests that miR-144-3p is a protective factor in PAAD, and according to the endogenous competing RNA (ceRNA) hypothesis, miR-144-3p may be the most promising upstream miRNA for COL10A1. In addition, several previous studies demonstrated the protective effect of miR-144-3p against tumors, such as lung cancer36, liver cancer37, colorectal cancer38, renal carcinoma39, and breast cancer40. miR-144-3p expression is down-regulated in pancreatic cancer tissues and cell lines41,42. Transfection of pancreatic cancer cell lines (PANC-1) using miR-144-3p resulted in diminished colony formation of pancreatic cancer cells and a significant decrease in cell invasion and migration (P < 0.01)41. Li J et al found that miR-144-3p could arrest pancreatic cancer cells in the S phase of the cell cycle by activating the mitogen-activated protein kinase pathway, and its inhibitory effect on pancreatic cancer cell proliferation could be reversed using miR-144-3p inhibitors42.
The ceRNA hypothesis reveals a novel mechanism of lncRNA/miRNA/mRNA interactions43, This study predicts 96 possible lncRNAs for the 144-3p/COL10A1 axis. Based on the ceRNA hypothesis, Tug1 was identified as the most likely upstream lncRNA to be upregulated. In addition. Several studies have shown the relevance of TUG1 to pancreatic cancer. For example, Qin et al demonstrated that lncRNA TUG1 can promote the malignant progression of pancreatic cancer via the EMT pathway44. Yang et al TUG1 affect tumor invasion and gemcitabine resistance in pancreatic cancer 45. Liang et al demonstrated that TUG1 may be a potential target for the treatment of PAAD46. Taken together, TUG1/miR-144-3p/COL10A1axis were identified as potential regulatory pathways in PAAD.
Tumor immune cells are part of the tumor microenvironment and, in addition to their antitumor effects, also lead to immune escape from tumors, which in turn promotes tumor development15,16. Infiltration of tumor immune cells affects the efficacy and prognosis of chemotherapy, radiotherapy, or immunotherapy in tumor patients47–49. Correlation analysis showed that the expression level of COL10A1 was negatively correlated with the tumor purity of PAAD, indicating its relative enrichment in the tumor microenvironment. COL10A1 expression was closely associated with CD8 + T cells and their markers. CD8 + T cells are effector cells for immunotherapy50. Activated CD8 + T cells can kill tumor cells by perforating proteases of the Fas-Fas ligand pathway51. Further analysis showed that COL10A1 expression levels were associated with M1 macrophages, M2 macrophages, tumor-associated markers of macrophage (TAM), and T-regulatory (Treg) cells. In early cancer progression, TAM tends to favor the M1 phenotype and promotes antitumor activity. In contrast, as the tumor progresses, TAM has the opposite effect, TAM prefers the M2 phenotype and promotes tumor immunoregulation, thus promoting tumor invasion and malignant progression52. This may explain the higher expression of COL101 in intermediate, advanced pancreatic cancer. Treg cells have immunosuppressive effects and can reduce the activity of effector cells. In addition, we found that COL10A1 was significantly associated with CD11c + DC, which is a key target of Treg cells and is involved in immunosuppression in the tumor microenvironment.53. Taken together, tumor immune infiltration may partially explain the oncogenic effect of COL10A1-mediated pancreatic cancer.
Tumor immunotherapy, such as immune checkpoint inhibitors based on monoclonal antibodies, clears tumors by rebooting the body's normal anti-tumor immune response54. PD-1/PD-L1 and CTLA-4 are immune checkpoints for immunotherapy55. We investigated the correlation between COL10A1 and immune checkpoints. The results showed that COL10A1 was significantly correlated with PD-L1 and CTLA-4 in pancreatic cancer, suggesting that COL10A1 may be associated with immune escape. COL10A1 may serve as a new target for immunotherapy.
Our study also has shortcomings. The main limitation of this study is that the results were based on bioinformatics analysis and the data were mainly obtained from public databases, and there was the heterogeneity of the data and platform differences among the different databases. We finally validated the promoting role of COL10A1 in cancer development by validating it against several different databases. However, this result needs to be verified by further laboratory experiments