Identification of a competing endogenous RNA axis "SVIL-AS1/miR-103a/ICE1" associated with chemoresistance in lung adenocarcinoma by comprehensive bioinformatics analysis

Abstract

Background: Chemotherapy is an important treatment for lung cancer. The molecular mechanism of lung adenocarcinoma (LUAD) chemoresistance is not completely understood.

Methods: Weighted gene co-expression network analysis (WGCNA) was applied to screen the modules related to chemosensitivity using the data of LUAD patients receiving chemotherapy in The Cancer Genome Atlas database. GDCRNATools package was used to establish competing endogenous RNA (ceRNA) network based on the key chemotherapy-related module. Kaplan-Meier and risk models were used to analyze the influence of genes in the ceRNA network on the prognosis of LUAD patients receiving chemotherapy. Cell counting kit-8, reverse transcription-quantitative PCR, and dual-luciferase reporter assay were used to detect the effects of abnormal expression of genes in the ceRNA network on the proliferation and IC50 of cisplatin (DDP)-resistant LUAD cells, and the targeting relationship of genes in the ceRNA network. The signaling pathways and functions of ICE1 in LUAD were analyzed by LinkOmics and CancerSEA databases, and validated by Western blot.

Results: Midnightblue module was the only WGCNA module positively correlated with chemosensitivity, in which the function of genes was related to cancer progression. SVIL-AS1/miR-103a/ICE1 was constructed based on midnightblue module. High expression of SVIl-AS1 and ICE1 corresponded to a favorable prognosis. High expression of miR-103a corresponded to a dismal prognosis. SVIl-AS1 was downregulated in DDP-resistant LUAD cells. SVIL-AS1 overexpression retarded the proliferation and DDP resistance of DDP-resistant LUAD cell. miR-103a was sponged by SVIL-AS1 and directly targeted ICE1. miR-103a overexpression and ICE1 knockdown overturned the suppressive effect of SVIL-AS1 overexpression on cell proliferation and DDP resistance. Further bioinformatics analysis and experimental verification showed that SVIL-AS1/miR-103a-3p/ICE1 axis can enhance DNA damage caused by chemotherapeutic agents.

Conclusions: SVIL-AS1 inhibited chemoresistance by acting as a sponge for miR-103a and upregulating ICE1 expression, which may be a potential therapeutic target for chemotherapy in LUAD.

Keywords: chemoresistance; competing endogenous RNA; lung adenocarcinoma; prognosis; weighted gene co-expression network analysis.

FIGURE 1

WGCNA of LUAD patients receiving chemotherapy in TCGA. (A,B) Scale‐free topology fit (A) and mean connectivity (B) for various soft threshold power. (C) Eigengene adjacency dendrogram of the merged module. (D) Heatmap of module–traits relationship. The correlation and p value (numbers in parentheses) of each module were shown in each table. (E) Correlation of eigengenes in midnightblue module with chemotherapy. LUAD, lung adenocarcinoma; WGCNA, weighted gene co‐expression network analysis

FIGURE 2

Functional analysis of genes in midnightblue module via Metascape. (A) The top 12 significantly representative terms. (B) The similarities of enriched terms for eigengenes in midnightblue module. Each term was represented by a circular node, the size of which was proportional to the number of genes in the term

FIGURE 3

ceRNA network based on the eigengenes of midnightblue module. (A) SVIL‐AS1/miR‐103a‐3p/ICE1 axis. (B) Correlation analysis of SVIL‐AS1, miR‐103a‐3p, and ICE1 expression. ceRNA, competing endogenous RNA

FIGURE 4

Overall survival analysis of SVIL‐AS1, miR‐103a‐3p, and ICE1 based on Kaplan–Meier plotter database (A–C) and TCGA database (D–F). TCGA, The Cancer Genome Atlas

FIGURE 5

Multivariate survival analysis of SVIL‐AS1/miR‐103a‐3p/ICE1 axis. (A) Risk score, survival time, and SVIL‐AS1/miR‐103a‐3p/ICE1 expression of LUAD patients. (B) Survival probability of patients in high‐risk and low‐risk groups. (C) Time‐dependent ROC curve. LUAD, lung adenocarcinoma; ROC, receiver operating characteristic

FIGURE 6

Effect of SVIL‐AS1 overexpression on DDP‐resistance and proliferation in A549/DDP and H1975/DDP cells. (A) IC50 value of DDP‐sensitive cells (A549 and H1975) and DDP‐resistant cells (A549/DDP and H1975/DDP). (B) Relative expression level of SVIL‐AS1 in A549, H1975, A549/DDP, and H1975/DDP cells. (C) Relative expression level of SVIL‐AS1 in A549/DDP and H1975/DDP cells transfected with pcDNA3.1 or pcDNA3.1‐SVIL‐AS1. (D) Effect of SVIL‐AS1 overexpression on IC50 of DDP in A549/DDP and H1975/DDP cells. (E,F) Effect of SVIL‐AS1 overexpression on cell proliferation in A549/DDP (E) and H1975/DDP (F) cells. **p < 0.01; ***p < 0.001. DDP, cisplatin

FIGURE 7

Targeting relationship of SVIL‐AS1/miR‐103a‐3p and miR‐103a‐3p/ICE1 pairs. (A) The predicted binding sites of miR‐103a at 3′‐UTR of SVIL‐AS1. (B) Luciferase activity of SVIL‐AS1‐WT or SVIL‐AS1‐MT in cells transfected with NC mimic or miR‐103a mimic. (C) Relative expression level of miR‐103a in A549/DDP and H1975/DDP cells transfected with pcDNA3.1 or pcDNA3.1‐SVIL‐AS1. (D) The predicted binding sites of miR‐103a at 3′‐UTR of ICE1. (E) Luciferase activity of ICE1‐WT or ICE1‐MT in cells transfected with NC mimic or miR‐103a mimic. (F) RT‐qPCR analysis of ICE1 expression in A549/DDP and H1975/DDP cells transfected with inhibitor control (in‐miR‐NC) or miR‐103a inhibitor (in‐miR‐103a). (G) RT‐qPCR analysis of ICE1 expression in A549/DDP and H1975/DDP cells transfected with pcDNA3.1 or pcDNA3.1‐SVIL‐AS1. **p < 0.01; ***p < 0.001. RT‐qPCR, reverse transcription‐quantitative PCR

FIGURE 8

Rescue experiments of SVIL‐AS1 overexpression in A549/DDP and H1975/DDP cells. (A) IC50 of DDP in A549/DDP and H1975/DDP cells transfected with pcDNA3.1, pcDNA3.1‐SVIL‐AS1, co‐transfected with pcDNA3.1‐SVIL‐AS1 and miR‐103a mimic, or co‐transfected with pcDNA3.1‐SVIL‐AS1 and si‐ICE1. (B,C) CCK‐8 assay of cell proliferation in A549/DDP (B) and H1975/DDP (C) cells transfected with pcDNA3.1, pcDNA3.1‐SVIL‐AS1, co‐transfected with pcDNA3.1‐SVIL‐AS1 and miR‐103a mimic, or co‐transfected with pcDNA3.1‐SVIL‐AS1 and si‐ICE1. **p < 0.01, SVIL‐AS1+si‐ICE1 versus SVIL‐AS1; ^## p < 0.01, SVIL‐AS1+miR‐103a mimic versus SVIL‐AS1. DDP, cisplatin; CCK‐8, cell counting kit‐8

FIGURE 9

Identification and functional analysis of ICE1 co‐expressed genes in LUAD. (A,B) Heatmap of the top 10 genes positively (A) and negatively (B) correlated with ICE1 expression. (C) GO analysis of ICE1 co‐expressed genes in LUAD. Blue, GO terms for genes positively correlated with ICE1 expression. Orange, GO terms for genes negatively correlated with ICE1 expression. (D) KEGG analysis of ICE1 co‐expressed genes in LUAD. Blue, GO terms for genes positively correlated with ICE1 expression. Orange, GO terms for genes negatively correlated with ICE1 expression. (E) Single‐cell sequencing analysis was performed using CancerSEA database. (F) Western blot was applied to detect the level of γ‐H2AX in A549/DDP and H1975/DDP cells transfected with si‐control, si‐ICE1, or pcDNA3.1‐SVIL‐AS1. **p < 0.01. GO, Gene Ontology; KYOTO, Kyoto Encyclopedia of Genes and Genomes; LUAD, lung adenocarcinoma