Abstract
BackgroundColorectal cancer (CRC) is one of the leading causes of cancer death worldwide. Emerging evidence has revealed that risk factors and metastatic patterns differ greatly between colon and rectal cancers. However, the molecular mechanism underlying their pathogenic differences remains unclear. Therefore, we here aimed to identify non-coding RNA biomarkers based on lncRNA-associated ceRNA network (LceNET) to elucidate the carcinogenic heterogeneity between colon and rectal cancers.MethodsA global LceNET in human was constructed by employing experimental evidence-based miRNA-mRNA and miRNA-lncRNA interactions. Then, four context-specific ceRNA networks related to cancer initiation and metastasis were extracted by mapping differentially expressed lncRNAs, miRNAs and mRNAs to the global LceNET. Notably, a novel network-based bioinformatics model was proposed and applied to identify lncRNA/miRNA biomarkers and critical ceRNA triplets for understanding the carcinogenic heterogeneity between colon and rectal cancers. Moreover, the identified biomarkers were further validated by their diagnostic/prognostic performance, expression pattern and correlation analysis.ResultsBased on network modeling, lncRNA KCNQ1OT1 (AUC>0.85) and SNHG1 (AUC>0.94) were unveiled as common diagnostic biomarkers for the initiation and metastasis of colon and rectal cancers. qRT-PCR analysis uncovered that these lncRNAs had significantly higher expression level in CRC cell lines with high metastatic potential. In particular, KCNQ1OT1 and SNHG1 function in colon and rectal cancers via different ceRNA mechanisms. For example, KCNQ1OT1/miR-484/ANKRD36 axis was involved in the initiation of colon cancer, while KCNQ1OT1/miR-181a-5p/PCGF2 axis was implicated in the metastasis of rectal cancer; the SNHG1/miR-484/ORC6 axis played a role in colon cancer, while SNHG1/miR-423-5p/EZH2 and SNHG1/let-7b-5p/ATP6V1F axes participated in the initiation and metastasis of rectal cancer, respectively. In these ceRNA triplets, miR-484, miR-181a-5p, miR-423-5p and let-7b-5p were identified as miRNA biomarkers with excellent distinguishing ability between normal and tumor tissues, and ANKRD36, PCGF2, EZH2 and ATP6V1F were closely related to the prognosis of corresponding cancer.ConclusionThe landscape of lncRNA-associated ceRNA network not only facilitates the exploration of non-coding RNA biomarkers, but also provides deep insights into the oncogenetic heterogeneity between colon and rectal cancers, thereby contributing to the optimization of diagnostic and therapeutic strategies of CRC.
Highlights
Colorectal cancer (CRC) is one of the most commonly diagnosed malignant tumors with high prevalence and mortality rates worldwide
Based on the identified ceRNA interaction triplets, the global lncRNA-associated ceRNA network (LceNET) composed by 3,102 long non-coding RNAs (lncRNAs), 1,085 miRNAs and 16,490 mRNAs was constructed in human
We explored the differential expression profiles of lncRNAs, miRNAs and mRNAs in the four comparable groups, including M0 vs. normal in colon cancer, M1 vs. normal in colon cancer, M0 vs. normal in rectal cancer and M1 vs. normal in rectal cancer
Summary
Colorectal cancer (CRC) is one of the most commonly diagnosed malignant tumors with high prevalence and mortality rates worldwide. MicroRNAs (miRNAs) are a class of well-characterized ncRNAs that post-transcriptionally regulate target gene expression through complementary pairing [6,7,8]; long non-coding RNAs (lncRNAs) are a class of recently discovered ncRNAs longer than 200 nucleotides with emerging roles in diverse cancer-related processes, e.g., proliferation, invasion, metastasis, and metabolism [9, 10]. Mounting evidence has demonstrated that lncRNAs could act as oncogenic or tumor-suppressive genes in the initiation and progression of CRC through a ceRNA mechanism. LncRNA SNHG7 was found to be significantly over-expressed in CRC It accelerates tumor proliferation and metastasis by serving as a miR-216b-mediated ceRNA of GALNT1 [13]. Emerging evidence has revealed that risk factors and metastatic patterns differ greatly between colon and rectal cancers. We here aimed to identify noncoding RNA biomarkers based on lncRNA-associated ceRNA network (LceNET) to elucidate the carcinogenic heterogeneity between colon and rectal cancers
Published Version (Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have