Circ-ATIC regulates esophageal squamous cell carcinoma growth and metastasis through miR-1294/PBX3 pathway

Abstract

Esophageal squamous cell carcinoma (ESCC) is a digestive tract malignancy associated with poor clinical outcome. Growing evidence have elucidated that circular RNAs (circRNAs) play important roles in the pathological process of ESCC. However, the detailed mechanisms how circRNAs modulate the development of ESCC remain largely unknown. Our study aimed to decipher the role and mechanism of circ-ATIC (also termed as circRNA_0058063) in regulating the progression of ESCC. We found that circ-ATIC and its host gene ATIC were significantly increased in ESCC tissues and cells compared with the adjacent noncancerous tissues or normal esophagus epithelial cell. Circ-ATIC knockdown substantially reduced proliferation and the number of invaded ESCC cells and retarded EMT process, reflecting by the decreased N-cadherin and elevated E-cadherin. However, the level of host gene ATIC was not changed under circ-ATIC suppression. It was predicted that circ-ATIC could bind to miR-1294 and serve as a sponge RNA. The luciferase reporter assay and RNA immunoprecipitation (RIP) assay confirmed their relations. MiR-1294 was decreased in ESCC tissues and cells, which was reversely correlated with circ-ATIC level. Furthermore, PBX3 was predicted and proved to be a downstream direct target of miR-1294. PBX3 mRNA and protein were obviously upregulated in ESCC tumor tissues and cells. PBX3 overexpression could reverse the suppressive roles of miR-1294 mimics on ESCC proliferation and invasion. In an xenograft nude mice model, stable transfection of sh-circ-ATIC significantly retarded the growth of tumor and suppressed VEGF and Ki67. Collectively, circ-ATIC promoted ESCC proliferation and invasion by regulating miR-1294/PBX3 axis.

Keywords: Circ-ATIC; Esophageal squamous cell carcinoma; MiR-1294; PBX3.

Fig. 1

Expression of circ-ATIC in ESCC tissues and cells and RNA stability. (a) Schematic illustration showing the genomic region of circ-ATIC derived from exons of the ATIC host gene; (b–c) RT-PCR measurement of circ-ATIC and ATIC levels in ESCC tissues (tumor) and the matched adjacent noncancerous tissues (control) (n = 36), *P < 0.05 compared with control; (d) Level of circ-ATIC in normal esophagus epithelial cells (HET-1A) and ESCC cells (KYSE-510 and KYSE-520), *P < 0.05 compared with HET-1A; (e–f) ESCC cells were treated with actinomycin D and the expressions of circ-ATIC and ATIC mRNA were evaluated at indicated time points; (g–h) Total RNA was extracted from ESCC cells and RNase R or vehicle was added, then the remaining circ-ATIC and ATIC mRNA were detected using RT-PCR, *P < 0.05 compared with Mock.

Fig. 2

Circ-ATIC affects proliferation and invasion of ESCC cells. (a) Schematic illustration of the position and sequence of siRNA for circ-ATIC; (b) level of circ-ATIC in ESCC cells after transfection of siRNA or negative control siRNA (si-NC); (c–d) the effect of circ-ATIC interfering on mRNA and protein levels of ATIC; (e) ESCC cellular activity evaluated using CCK8; (f) ESCC cell proliferation detection using EdU staining; (g) Cell invasion after circ-ATIC inhibition was measured using transwell assay; (h) Western blot analysis of N-cadherin and E-cadherin expressions after circ-ATIC siRNA transfection.

Fig. 3

Circ-ATIC binds to miR-1294 to act as a ceRNA. (a) Eight predicted target miRNA of circ-ATIC were selected and their expressions were assessed in KYSE-510 cell transfected with si-ATIC or si-NC, *P < 0.05 compared to si-NC; (b) the predicted binding sites between circ-ATIC and miR-1294; (c–d) KYSE-510 and KYSE-520 cells were co-transfected with miR-1294 mimic or miR-NC and luciferase reporter plasmids carrying circ-ATIC-wt or circ-ATIC-mut. The relative luciferase activity was detected using dual luciferase reporter system; (e–f) Ago2 RNA immunoprecipitation and RT-PCR assays were performed to analyze the expression of circ-ATIC and miR-1294 in Ago2 immunoprecipitation and IgG immunoprecipitation; (g) miR-1294 level in ESCC tissues and the matched adjacent noncancerous tissues (n = 36), *P < 0.05 compared to normal tissues; (h) level of circ-ATIC in normal esophagus epithelial cells and ESCC cells, *P < 0.05 compared to HET-1A cells. (i) Scattering plot of circ-ATIC and miR-1294 expression in cancer tissues. (j) RT-PCR detection of miR-1294 in KYSE-510 and KYSE-520 cells transfected with si-circ-ATIC and anti-miR-1294, sole or in combination. *P < 0.05; (k) the effect of miR-1294 on circ-ATIC was evaluated in KYSE-510 cell line after overexpressing miR-1294.

Fig. 4

MiR-1294 directly targets to PBX3 mRNA in ESCC cells. (a) Seven predicted target genes of miR-1294 were selected and their expressions were assessed in KYSE-510 cell transfected with miR-1294 or miR-NC, *P < 0.05 compared to miR-NC; (b) RT-PCR measurement of PBX3 mRNA in ESCC tumor tissues and adjacent noncancerous tissues (n = 36); (c) three representative western blotting images and quantitative results of PBX3 in tumor tissues and adjacent noncancerous tissues of ESCC patients; (d–e) mRNA and protein levels of PBX3 in tumor cell lines and normal cells; (f) the diagram showed the binding sites between PBX3 mRNA 3′-UTR and miR-1294; (g–h) KYSE-510 and KYSE-520 cells were co-transfected with miR-1294 mimic or miR-NC and luciferase reporter plasmids carrying PBX3 3′-UTR-wt or PBX3 3′-UTR-mut. The relative luciferase activity was detected using dual luciferase reporter system; (i) scattering plot of miR-1294 and PBX3 mRNA expressions in cancer tissues; (j–k) RT-PCR and Western blot analysis of PBX3 expression in KYSE-510 and KYSE-520 cells transfected with miR-1294 mimics, sole or in combination with PBX3 overexpressing plasmid. *P < 0.05.

Fig. 5

MiR-1294/PBX3 regulates proliferative and invasive capabilities of ESCC cells. (a–b) MiR-1294 mimics were introduced to KYSE-510 and KYSE-520 cells alone or in combination with PBX3 overexpressing plasmid, cell viability was evaluated at indicated time using CCK8 assay. (c–d) Proliferation of treated ESCC cells was detected using EdU staining and the cells were counted; (e–f) cell invasion capability was measured by transwell assay and number of invaded cells were counted; (g representative Western blot images and quantitative analysis of N-cadherin and E-cadherin expressions after miR-1294 mimic and PBX3 plasmid transfections. *P < 0.05.

Fig. 6

Circ-ATIC-miR-1294 regulates expression of PBX3. (a–b) si-circ-ATIC was transfected into KYSE510 and KYSE520 cells alone or in combination with miR-1294 inhibitor (anti-miR-1294) for 48 h, the PBX mRNA and protein were detected using RT-PCR and Western blot.

Fig. 7

Circ-ATIC silencing retarded the growth of ESCC xenografts. (a–b) KYSE510 cells with stable transfection of sh-circ-ATIC or sh-NC lentivirus vectors were subcutaneously injected into nude mice (n = 5 per group), the tumor volumes were measured and calculated every four days from day 7 to day 28. After the last assessment, the xenografts were dissected and weighted; (c–d) RT-PCR analysis of circ-ATIC and miR-1294 in xenograft tissues of the two groups of mice; (e–f) RT-PCR and Western blot analysis of PBX3 expression in xenograft tissues; (g) IHC staining of cell proliferative markers VEGF and Ki67 in two groups of xenografts. *P < 0.05 compared with sh-NC.