Silencing LINC01021 inhibits gastric cancer through upregulation of KISS1 expression by blocking CDK2-dependent phosphorylation of CDX2

Abstract

Gastric cancer remains one of the most dangerous cancers, bringing suffering and economic burden to people worldwide. Long noncoding RNAs (lncRNAs) exhibit great potentials for targeted therapy of various cancers. In this investigation, we tested mechanisms by which LINC01021 may regulate gastric cancer progression. We collected gastric cancer tissues and procured cell lines to explore the potential factors by which LINC01021 had effects on angiogenesis, invasion, and migration, by quantitative reverse-transcription polymerase chain reaction (qRT-PCR), Transwell assay, and western blot analysis. Relationships among LINC01021, Caudal-type homeobox 2 (CDX2), and KISS1 were validated by dual-luciferase gene reporter, RNA pull-down, and RNA immunoprecipitation assays. Additionally, a murine model was developed to further explore the impact of LINC01021 on tumors in vivo. LINC01021 was upregulated in gastric cancer tissues and cells. LINC01021 regulated KISS1 through CDK2, which promoted phosphorylation and nuclear export in CDX2. Inhibition of LINC01021 suppressed the tumorigenesis of gastric cancer. Further, silencing LINC01021 exerted an inhibitory effect on cancer cell migration, invasion, and angiogenesis by promoting the binding between CDX2 and KISS1, while inhibiting that between CDK2 and CDX2. Taken altogether, high LINC01021 expression in gastric cancer promotes malignant cell migration and angiogenesis by downregulation of KISS1 through CDK2-mediated CDX2 phosphorylation.

Keywords: CDK2; CDX2; KISS1; LINC01021; gastric cancer; phosphorylation.

Graphical abstract

Figure 1

Bioinformatics analysis predicted lncRNA with differential expression in gastric cancer (A) GEO: GSE13911 volcano plot for differential gene expression. The abscissa represents p value of log10, and the ordinate represents the log(fold change), with upregulated genes (red) and downregulated genes (green) in tumor samples. (B) Heatmap displaying expression profile of top 10 differentially expressed genes in microarray dataset GEO: GSE13911. (C) Expression of top 10 differentially expressed genes in microarray dataset GEO: GSE13911 determined by qRT-PCR. (D) Predicted sites of LINC01021 expression in the nucleus (below abscissa) and cytoplasm (above abscissa).

Figure 2

Highly expressed LINC01021 is found in gastric cancer tissues and cells (A) qRT-PCR analysis of LINC01021 in cancer tissues and adjacent normal tissues (n = 78). (B) Kaplan-Meier curve for LINC01021 and the overall survival time of patients with gastric cancer. (C) qRT-PCR analysis of LINC01021 expression in four gastric cancer cell lines and normal gastric mucosal cell lines. ∗p < 0.05 versus GES1 cells. Measurement data were expressed as mean ± standard deviation. Comparisons among multiple groups were conducted by one-way ANOVA with Tukey’s post hoc test. The data between two groups were analyzed by paired t test. Survival of patients was calculated using Kaplan-Meier method, and the survival difference was analyzed by log rank test.

Figure 3

LINC01021 is involved in cell migration, invasion, and angiogenesis in gastric cancer cells (A) qRT-PCR analysis of LINC01021 expression in BGC823 and SGC-7901 cells treated with si-NC and si-LINC01021. (B) Transwell assay of cell migration after treatment with si-NC and si-LINC01021. (C) Transwell assay of cell invasion after treatment with si-NC and si-LINC01021. (D) Western blot analysis of E-cadherin, N-cadherin, and vimentin in cells after treatment with si-NC and si-LINC01021. (E) Tube formation assay for number of tubes after treatment with si-NC and si-LINC01021. (F) Western blot analysis of VEGF and CD34 protein expression in cells after treatment with si-NC and si-LINC01021. ∗p < 0.05 versus si-NC treatment. Measurement data were expressed as mean ± standard deviation. Comparisons among multiple groups were conducted by one-way ANOVA. The data between two groups were analyzed by paired t test.

Figure 4

LINC01021 promotes tumor growth in gastric cancer in vivo (A) Average tumor volume of mice bearing treated cells. (B) Average tumor weight of mice bearing treated cells. (C) qRT-PCR of LINC01021 expression in tumors. (D) Western blot analysis of VEGF and CD34 protein expression. (E) IHC of positive expression of VEGF, Ki67, and CD34 in tumors. ∗p < 0.05 versus sh-NC treatment. Measurement data were expressed as mean ± standard deviation. Comparisons among multiple groups were conducted by one-way ANOVA. The data between two groups were analyzed by paired t test. n = 10.

Figure 5

LINC01021 inhibits KISS1 expression through CDX2 (A) LncMAP prediction of the target genes of CDX2 modulated by LINC01021. (B) RNA pull-down assay of binding between LINC01021 and CDX2 in BGC823 and SGC-7901 cells. (C) RIP assay of binding between LINC01021 and CDX2; ∗p < 0.05 versus IgG. (D) qRT-PCR analysis of KISS1 in gastric cancer tissues and adjacent normal tissues (n = 78). (E) qRT-PCR analysis of CDX2 and KISS1 expression in cells treated with oe-CDX2. ∗p < 0.05 versus oe-NC treatment. (F) Dual-luciferase reporter gene assay of binding between CDX2 and KISS1. ∗p < 0.05 versus oe-NC treatment. (G) ChIP assay of binding between CDX2 and KISS1. ∗p < 0.05 versus IgG. (H) ChIP assay of binding between CDX2 and KISS1 in BGC823 and SGC-7901 cells at the presence of si-LINC01021 or si-NC. ∗p < 0.05 versus IgG; ^#p < 0.05 versus si-NC treatment. (I) qRT-PCR of KISS1 expression in BGC823 and SGC-7901 cells after silencing LINC01021. ∗p < 0.05 versus si-NC treatment; ^#p < 0.05 versus si-LINC01021 + si-NC. Measurement data were expressed as mean ± standard deviation. Comparisons among multiple groups were conducted by one-way ANOVA. The data between two groups were analyzed by paired t test.

Figure 6

LINC01021 downregulates the expression of KISS1 through CDX2, thereby promoting cell migration, invasion, and angiogenesis in gastric cancer (A) qRT-PCR of LINC01021, CDX2, and KISS1 expression. (B) Western blot analysis of CDX2 and KISS1 expression. (C) Transwell assay to detect cell migration. (D) Transwell assay to detect cell invasion. (E) Tube formation assay to measure blood vessels. (F) Western blot analysis of VEGF and CD34 protein expression. ∗p < 0.05 versus si-NC + oe-NC treatment; ^#p < 0.05 versus si-LINC01021 + si-CDX2 treatment; ^&p < 0.05 versus si-LINC01021 + oe-KISS1 treatment. Measurement data were expressed as mean ± standard deviation. Comparisons among multiple groups were conducted by one-way ANOVA. The data between two groups were analyzed by paired t test.

Figure 7

Silencing LINC01021 promotes CDX2 expression and stability (A) qRT-PCR analysis of CDX2 expression in cancer tissues and adjacent normal tissues (n = 78); ∗p < 0.05. (B) IHC of CDX2 expression in cancer tissues and adjacent normal tissues. (C) Western blot analysis of CDX2 protein expression in BGC823 cell lines after CHX treatment. CDX2 expression was quantified relative to GAPDH expression at indicated times and normalized to the 0-hour time point (before CHX treatment); ∗p < 0.05, ∗∗p < 0.01 versus si-NC treatment. Measurement data were expressed as mean ± standard deviation. The data between two groups were analyzed by paired t test.

Figure 8

LINC01021 plays roles in CDX2 phosphorylation, nuclear export, and transcriptional activity of downstream target genes via CDK2 (A) The website (http://gps.biocuckoo.org/links.php#l1) predicted sites that CDK2 promoted CDX2 phosphorylation while highest score was observed in S100 and S60. (B) The website (http://www.phosphonet.ca/) predicted the phosphorylation enzymes at the S100 and S60 sites of CDX2, including CDK2. (C) coIP showed that CDK2 can bind to CDX2 in BGC823 cells. ∗p < 0.05 versus IgG. (D) Western blot analysis of CDX2 expression in BGC823 after IP with anti-phosphoserine. ∗p < 0.05 versus oe-NC treatment; ^#,∗p < 0.05 versus si-NC treatment. (E) Western blot of CDX2 expression in cytoplasm and nucleus after overexpressing CDK2 in cytoplasm and nucleus with GAPDH used as internal reference in neoplasm and Lamin A/C in nucleus. (F) qRT-PCR analysis of KISS1 expression and western blot analysis of CDX2 in BGC823 cells treated with oe-CDK2. ∗p < 0.05 versus oe-NC; ^#p < 0.05 versus oe-CDK2 treatment. (G) Western blot analysis of CDX2 expression in BGC823 cells treated with si-LINC01021 after IP with anti-phosphoserine. ∗p < 0.05 versus IgG. (H) CoIP of binding between CDK2 and CDX2 in BGC823 cells treated with si-LINC01021. ∗p < 0.05 versus IgG.

Figure 9

Molecular schematic diagram concerning LINC01021 in gastric cancer LINC01021 was upregulated in gastric cancer and could promote phosphorylation and nuclear export of CDX2 by CDK2. Silencing LINC01021 inhibited the binding of CDK2 and CDX2, as well as CDX2 phosphorylation. Additionally, silencing LINC01021 suppressed cell invasion, migration, and angiogenesis by upregulating the expression of KISS1.