Circular RNA circLMO7 acts as a microRNA-30a-3p sponge to promote gastric cancer progression via the WNT2/β-catenin pathway

Abstract

Background: Gastric cancer (GC) is one of the most common malignant tumors worldwide. Currently, the overall survival rate of GC is still unsatisfactory despite progress in diagnosis and treatment. Therefore, studying the molecular mechanisms involved in GC is vital for diagnosis and treatment. CircRNAs, a type of noncoding RNA, have been proven to act as miRNA sponges that can widely regulate various cancers. By this mechanism, circRNA can regulate tumors at the genetic level by releasing miRNA from inhibiting its target genes. The WNT2/β-Catenin regulatory pathway is one of the canonical signaling pathways in tumors. It can not only promote the development of tumors but also provide energy for tumor growth through cell metabolism (such as glutamine metabolism).

Methods: Through RNA sequencing, we found that hsa_circ_0008259 (circLMO7) was highly expressed in GC tissues. After verifying the circular characteristics of circLMO7, we determined the downstream miRNA (miR-30a-3p) of circLMO7 by RNA pull-down and luciferase reporter assays. We verified the effect of circLMO7 and miR-30a-3p on GC cells through a series of functional experiments, including colony formation, 5-ethynyl-2'-deoxyuridine and Transwell assays. Through Western blot and immunofluorescence analyses, we found that WNT2 was the downstream target gene of miR-30a-3p and further confirmed that the circLMO7-miR-30a-3p-WNT2 axis could promote the development of GC. In addition, measurement of related metabolites confirmed that this axis could also provide energy for the growth of GC cells through glutamine metabolism. We found that circLMO7 could promote the growth and metastasis of GC in vivo by the establishment of nude mouse models. Finally, we also demonstrated that HNRNPL could bind to the flanking introns of the circLMO7 exons to promote circLMO7 cyclization.

Results: CircLMO7 acted as a miR-30a-3p sponge affecting the WNT2/β-Catenin pathway to promote the proliferation, migration and invasion of GC cells. Moreover, animal results also showed that circLMO7 could promote GC growth and metastasis in vivo. CircLMO7 could also affect the glutamine metabolism of GC cells through the WNT2/β-Catenin pathway to promote its malignant biological function. In addition, we proved that HNRNPL could promote the self-cyclization of circLMO7.

Conclusions: CircLMO7 promotes the development of GC by releasing the inhibitory effect of miR-30a-3p on its target gene WNT2.

Keywords: Gastric cancer; Glutaminolysis; HNRNPL; WNT2; circRNA; miRNA.

Fig. 1

Differentially expressed circRNAs in GC were detected by NGS. a, b. Volcano map and cluster heat map showing the differences in the expression of circRNA in GC tissues and adjacent normal tissues. c. qRT-PCR showed that the expression of circLMO7 in GC tissues was significantly higher than that in adjacent normal tissues. d. Different expression levels of circLMO7 in cell lines. All data are presented as the mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

Fig. 2

CircLMO7 acts as a sponge of miR-30a-3p. a. Venn diagram showing the target miRNAs of circLMO7 predicted by TargetScan7.2 and RegRNA2.0. b. Target miRNAs overlapping between TargetScan7.2 and RegRNA2.0. c. Pull-down efficiency of the circLMO7 probe. d After pull-down experiments, we found that the binding efficiency between circLMO7 and miR-30a-3p was the highest. e. Sequence diagram of the binding sites between circLMO7 and miR-30a-3p. f. Map of the mutations of the binding sites between circLMO7 and miR-30a-3p. g. A luciferase reporter assay demonstrated that circLMO7 and miR-30a-3p bound to each other. h. RNA fluorescence in situ hybridization confirmed the binding relationship between circLMO7 and miR-30a-3p; scale bar = 10 μm. i. qRT-PCR showed that the expression of miR-30a-3p in GC tissues was significantly lower than that in adjacent normal tissues. j. qRT-PCR showed that circLMO7 and miR-30a-3p levels were negatively correlated. All data are presented as the mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

Fig. 3

Overexpression of circLMO7 or knockdown of miR-30a-3p promotes the proliferation, migration and invasion of GC cells. a, b. Colony formation and EdU assays showed that overexpression of circLMO7 or knockdown of miR-30a-3p promoted the proliferation of GC cells, while knockdown of circLMO7 or overexpression of miR-30a-3p inhibited this process; EdU scale bar = 25 μm. c. Transwell assays showed that overexpression of circLMO7 or knockdown of miR-30a-3p promoted the migration and invasion of GC cells, while knockdown of circLMO7 or overexpression of miR-30a-3p had the opposite effects; scale bar = 100 μm. d. Overexpression of circLMO7 or knockdown of miR-30a-3p promoted the growth of human GC organoids, while the growth of human GC organoids was inhibited after circLMO7 had been knocked down or miR-30a-3p had been overexpressed; scale bar = 20 μm. e. Overexpression of circLMO7 or knockdown of miR-30a-3p promoted the migration and invasion of GC cells through the EMT pathway, while knockdown of circLMO7 or overexpression of miR-30a-3p showed the opposite results. All data are presented as the mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

Fig. 4

CircLMO7 promotes the expression of WNT2. a. Pattern diagram showing the downstream target gene of miR-30a-3p predicted by TargetScan7.2 and miRWalk3. b. Western blotting showed that the expression levels of miR-30a-3p and WNT2 were negatively correlated. c. Sequence map of the binding site between miR-30a-3p and WNT2. d. A luciferase reporter assay demonstrated that miR-30a-3p and WNT2 bind to each other. e. The KmPlot database showed that patients with lower WNT2 had better overall survival rates and disease-free survival rates than patients with higher WNT2. f. Immunohistochemistry showed that WNT2 was highly expressed in GC tissues; scale bar = 100 μm. g. The TCGA database showed that the expression of WNT2 in GC tissues was significantly higher than that in normal tissues. h. qRT-PCR showed that WNT2 was highly expressed in GC tissues. i. qRT-PCR showed that the expression levels of circLMO7 and WNT2 were positively correlated. j. Overexpression of circLMO7 promoted the expression of WNT2, while knockdown of circLMO7 inhibited the expression of WNT2; scale bar = 50 μm. All data are presented as the mean ± SD * P < 0.05, ** P < 0.01, *** P < 0.001

Fig. 5

CircLMO7 can promote the growth and metastasis of GC in vivo. a. Overexpression of circLMO7 promoted the growth of xenograft tumors, while knockdown of circLMO7 inhibited tumor growth. b. Measurement of xenograft tumor volume and weight. c. Overexpression of circLMO7 promoted lung metastasis, while knockdown of circLMO7 inhibited lung metastasis. d. Hematoxylin-eosin staining showed the size of lung metastatic tissues; scale bar = 200 μm. All data are presented as the mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

Fig. 6

CircLMO7 promotes GC malignant biological functions through glutamine metabolism. a. TCGA database analysis showed that GLS was highly expressed in GC tissues. b. qRT-PCR verified that GLS was highly expressed in GC tissues. c, d. qRT-PCR and Western blot analysis showed that circLMO7 and GLS were positively related. e-h. Overexpressing circLMO7 increased the expression levels of GLN, GLS, GLU and α-KG. Knocking down circLMO7 blocked this phenomenon. i, j. Colony formation and EdU assays showed that the promoting effect of ov-circLMO7 on GC cell proliferation was rescued after cotransfection with si-GLS; EdU scale bar = 25 μm. k. Transwell assays showed that si-GLS rescued the promoting effect of ov-circLMO7 on GC cell migration and invasion; scale bar = 100 μm. l. The active oxygen assay showed that the inhibitory effect of ov-circLMO7 on the production of reactive oxygen species in GC cells was rescued after cotransfection with si-GLS; scale bar = 50 μm. All data are presented as the mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

Fig. 7

HNRNPL promotes the production of circLMO7 in GC. a. RBPmap showed the binding sites between the flanking introns of circLMO7 exons and HNRNPL. b, c. RNA pull-down and RNA immunoprecipitation assays showed that HNRNPL can bind to the flanking introns of circLMO7 exons. d, e. qRT-PCR showed that the wild-type plasmid could significantly increase the expression of circLMO7, while the mutant plasmids could not. f. qRT-PCR showed the expression levels of circLMO7 and pre-mLMO7 after knocking down HNRNPL. g. Immunohistochemistry showed that HNRNPL was highly expressed in GC tissues; scale bar = 100 μm. h. The TCGA database showed that the expression of HNRNPL in GC tissues was significantly higher than that in normal tissues. i, j. qRT-PCR showed that HNRNPL was highly expressed in GC tissues and was positively correlated with circLMO7. All data are presented as the mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001