Prognostic Roles of ceRNA Network-Based Signatures in Gastrointestinal Cancers

Abstract

Gastrointestinal cancers (GICs) are high-incidence malignant tumors that seriously threaten human health around the world. Their complexity and heterogeneity make the classic staging system insufficient to guide patient management. Recently, competing endogenous RNA (ceRNA) interactions that closely link the function of protein-coding RNAs with that of non-coding RNAs, such as long non-coding RNA (lncRNA) and circular RNA (circRNA), has emerged as a novel molecular mechanism influencing miRNA-mediated gene regulation. Especially, ceRNA networks have proven to be powerful tools for deciphering cancer mechanisms and predicting therapeutic responses at the system level. Moreover, abnormal gene expression is one of the critical breaking events that disturb the stability of ceRNA network, highlighting the role of molecular biomarkers in optimizing cancer management and treatment. Therefore, developing prognostic signatures based on cancer-specific ceRNA network is of great significance for predicting clinical outcome or chemotherapy benefits of GIC patients. We herein introduce the current frontiers of ceRNA crosstalk in relation to their pathological implications and translational potentials in GICs, review the current researches on the prognostic signatures based on lncRNA or circRNA-mediated ceRNA networks in GICs, and highlight the translational implications of ceRNA signatures for GICs management. Furthermore, we summarize the computational approaches for establishing ceRNA network-based prognostic signatures, providing important clues for deciphering GIC biomarkers.

Keywords: ceRNA network; circRNA; gastrointestinal cancer; lncRNA; prognostic signature; translational implication.

Figure 1

Schematic diagram of representative ceRNA crosstalks function in GC (A), CRC (B) and EC (C). (A) ceRNA interaction regulates tumor cell proliferation, migration, invasion, apoptosis, or chemoresistance through PI3K/ATK, MAPK or Wnt/β-catenin signaling pathways, thereby exerting carcinogenic or tumor suppressor effects in GC. (B) ceRNA interaction contributes to CRC progression or chemoresistance by regulating autophagy process or pivotal pathways, such as Wnt/β-catenin, PI3K/ATK and JAK2/STAT3 signaling pathway. (C) ceRNA interaction promotes or inhibits EC progression by modulating cancer cell proliferation, migration, invasion, or apoptosis.

Figure 2

Computational strategy for construction, validation and functional annotation of ceRNA network-based prognostic signature in cancer. (A) Cancer-specific ceRNA network was constructed based on expression and interaction information. (B) Prognostic signature was developed by employing univariate Cox regression analysis and LASSO Cox regression analysis. (C) Prognostic performance of the signature should be evaluated and validated by Kaplan-Meier survival curve analysis, time-dependent ROC curve analysis and multivariate Cox regression analysis, and the expression pattern of genes that make up the model can be verified in other independent datasets. (D) Biological role of the prognostic signature could be investigated by functional enrichment analysis and immune infiltration analysis.