Long noncoding RNA associated-competing endogenous RNAs in gastric cancer

Abstract

Some long noncoding RNAs (lncRNAs) play important roles in the regulation of gene expression by acting as competing endogenous RNAs (ceRNAs). However, the roles of lncRNA associated-ceRNAs in oncogenesis are not fully understood. Here, based on lncRNA microarray data of gastric cancer, bioinformatic algorithm miRcode and microRNA (miRNA) targets database TarBase, we first constructed an lncRNA-miRNA-mRNA network. Then, we confirmed it by data of six types of other cancer including head and neck squamous cell carcinoma, prostate cancer, papillary thyroid carcinoma, pituitary gonadotrope tumors, ovarian cancer, and chronic lymphocytic leukemia. The results showed a clear cancer-associated ceRNA network. Eight lncRNAs (AC009499.1, GACAT1, GACAT3, H19, LINC00152, AP000288.2, FER1L4, and RP4-620F22.3) and nine miRNAs (miR-18a-5p, miR-18b-5p, miR-19a-3p, miR-20b-5p, miR-106a-5p, miR-106b-5p, miR-31-5p, miR-139-5p, and miR-195-5p) were involved. For instance, through its miRNA response elements (MREs) to compete for miR-106a-5p, lncRNA-FER1L4 regulates the expression of PTEN, RB1, RUNX1, VEGFA, CDKN1A, E2F1, HIPK3, IL-10, and PAK7. Furthermore, cellular experimental results indicated that FER1L4-small interfering RNA (siRNA) simultaneously suppressed FER1L4 and RB1 mRNA level. These results suggest that lncRNAs harbor MREs and play important roles in post-transcriptional regulation in cancer.

Figure 1. Competing endogenous RNAs (ceRNAs) function as microRNA (miRNA) sponges sequester miRNAs to regulate expression level of other transcripts sharing common miRNA response elements (MREs).

(A) Downregulation of long noncoding RNAs (lncRNA) leads more miRNA molecules free to bind to mRNA that contain the same MREs, thus its protein expression level decreases. (B) Overexpression of lncRNA leads fewer miRNA molecules to bind to mRNA, thus its protein expression level increases.

Figure 2. ceRNA network in gastric cancer.

Coloured nodes represent miRNAs; black nodes represent lncRNAs; gray nodes represent mRNAs; coloured edges indicate miRNA-target interactions. mRNAs' names are not shown.

Figure 3. Linear regression of ceRNAs' expression level.

Dashed lines represent 95% confidence interval. (A) FER1L4 vs RUNX1 (head and neck squamous cell carcinoma, n = 36). (B) FER1L4 vs RUNX1 (prostate cancer, n = 15). (C) LINC00152 vs THBS1 (papillary thyroid carcinoma, n = 18). (D) LINC00152 vs THBS1 (pituitary gonadotrope tumors, n = 23). (E) H19 vs MYCN (chronic lymphocytic leukemia, n = 52). (F) H19 vs MYCN (ovarian cancer, n = 15).

Figure 4. Relative Firefly/Renilla luminescence (mean ± SD, n = 3) mediated by luciferase plasmid harboring the wild-type or mutant FER1L4 sequence upon transfeciton with miR-106a-5p expression plasmid.

It indicates the direct interaction between FER1L4 and miR-106a-5p. ***P < 0.001.

Figure 5. qRT-PCR analysis of expression levels (mean ± SD, n = 3) of FER1L4 and RB1 in GES-1, AGS, MGC-803 and SGC-7901 treated with siRNA against FER1L4.

NC, negative control. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 6. A flowchart of ceRNA network construction.

(i) lncRNAs that are fold change ≥ 3.0 and P-value < 0.05 were retained; (ii) lncRNAs that have not been recorded in ENCODE were removed; (iii) miRNA-lncRNA interactions were predicted by miRcode; (iv) mRNAs that targeted by miRNAs were captured from TarBase; (v) ceRNA network construction.