microRNA-99a-5p induces cellular senescence in gemcitabine-resistant bladder cancer by targeting SMARCD1

Abstract

Patients with advanced bladder cancer are generally treated with a combination of chemotherapeutics, including gemcitabine, but the effect is limited due to acquisition of drug resistance. Thus, in this study, we investigated the mechanism of gemcitabine resistance. First, gemcitabine-resistant cells were established and resistance confirmed in vitro and in vivo. Small RNA sequencing analyses were performed to search for miRNAs involved in gemcitabine resistance. miR-99a-5p, selected as a candidate miRNA, was downregulated compared to its parental cells. In gain-of-function studies, miR-99a-5p inhibited cell viabilities and restored sensitivity to gemcitabine. RNA sequencing analysis was performed to find the target gene of miR-99a-5p. SMARCD1 was selected as a candidate gene. Dual-luciferase reporter assays showed that miR-99a-5p directly regulated SMARCD1. Loss-of-function studies conducted with si-RNAs revealed suppression of cell functions and restoration of gemcitabine sensitivity. miR-99a-5p overexpression and SMARCD1 knockdown also suppressed gemcitabine-resistant cells in vivo. Furthermore, β-galactosidase staining showed that miR-99a-5p induction and SMARCD1 suppression contributed to cellular senescence. In summary, tumor-suppressive miR-99a-5p induced cellular senescence in gemcitabine-resistant bladder cancer cells by targeting SMARCD1.

Keywords: SMARCD1; miR-99a-5p; bladder cancer; cellular senescence; gemcitabine resistance.

Fig. 1

Establishment of gemcitabine‐resistant BC cell lines. (A) BOY and (B) T24 were cultured with gemcitabine at concentrations ranging from 1 to 450 µg·mL⁻¹ for 12 months. Gemcitabine effects on parental and derived lines (C) BOY/GEM‐R‐BOY and (D) T24/GEM‐R‐T24. Comparison of tumor volumes after subcutaneous injection of 100 mg gemcitabine·kg⁻¹ per mouse per twice a week, n = 5 mice per group). *P < 0.05, Mann–Whitney U tests. The error bars indicate SEM. Photograph shows excised tissue in the Xenograft mouse model on day 36. Scale bar, 10 mm. Calculated IC50 for parental and derived lines (E) BOY/GEM‐R‐BOY and (F) T24/GEM‐R‐T24. n = 6. The error bars indicate SEM. (G) Cell proliferation measured by XTT assays of parental BC and derived GEM‐R BC strains. n = 6. *P < 0.05, Mann–Whitney U tests. The error bars indicate SEM. (H) Cell migration activities of parental BC and derived GEM‐R BC strains as measured by wound healing assay. n = 4. *P < 0.001, Mann–Whitney U tests. The error bars indicate SEM. (I) Cell invasion activities of parental BC and derived GEM‐R BC strains as measured by Matrigel invasion assay. Invasion cells were counted and compared. n = 6. *P < 0.001, Mann–Whitney U tests. The error bars indicate SEM. GEM, gemcitabine; GEM‐R BC, gemcitabine‐resistant bladder cancer.

Fig. 2

Expression level of miR‐99a‐5p and its restorative effect on GEM‐R BC cell lines and BC specimens. (A) A Venn diagram and in silico analysis of the miRNA sequences showed nine putative candidate miRNAs. (B) miRNA expression in the BLCA cohort of TCGA. miR‐99a‐5p expression in BLCA samples (n = 404) was compared with those in normal samples (n = 19). P < 0.0001, Mann–Whitney U tests. (C) The expression level of miR‐99a‐5p, as determined by qRT‐PCR, was significantly lower in the GEM‐R BC cell line than in the parental BC cell line. n = 3. *P < 0.001. The error bars indicate SEM. (D) Cell proliferation measured by XTT assay. n = 6. *P < 0.05; **P < 0.0001; ns, not significant. The error bars indicate SEM. (E) Cell migration activity measured by wound healing assay. n = 4. *P < 0.0001; ns, not significant. The error bars indicate SEM. (F) Cell invasion activity measured by Matrigel invasion assay. Invasion cells were counted and compared. n = 6. *P < 0.05; **P < 0.01; ns, not significant. The error bars indicate SEM. Transfection of miR‐99a‐5p increases the sensitivity of GEM‐R‐BC cell line to GEM. Each IC50 concentration of gemcitabine was given to parental cells and gemcitabine‐resistant cells. Transfection of 10 nm miR‐99a‐5p enhanced gemcitabine sensitivity of (G) GEM‐R‐BOY cells and (H) GEM‐R‐T24 cells by XTT assay. n = 6. *P < 0.0001. The error bars indicate SEM. Transfection of miR‐99a‐5p and administration of gemcitabine had a clear additive effect in (I) GEM‐R‐BOY and (J) GEM‐R‐T24 cells, significantly inhibiting cell proliferation. n = 6. *P < 0.01; **P < 0.0001. The error bars indicate SEM. The relationships between two groups were analyzed using Mann–Whitney U tests. The relationships between three or more groups were analyzed using the multiple comparison test with the Bonferroni/Dun method. These experiments were repeated at least three times. GEM, gemcitabine; GEM‐R BC, gemcitabine‐resistant bladder cancer; miRNA, microRNA; BLCA, bladder urothelial carcinoma.

Fig. 3

Identification of SMARCD1 mRNA as a target regulated by miR‐99a‐5p in the GEM‐R‐BC cell line. (A) Venn diagram and in silico analysis of mRNA sequences showed that 16 putative target candidate genes of miR‐99a‐5p are key molecules for gemcitabine‐resistant BC. (B) The mRNA expression level of SMARCD1 in the parental BC and GEM‐R BC strains was measured by qRT‐PCR and (C) the protein expression level was measured by western blot. The expression level of SMARCD1 was significantly higher in the GEM‐R BC strain. n = 3. *P < 0.0001, Mann–Whitney U tests. The error bars indicate SEM. imagej was used for protein levels. (D) mRNA expression of SMARCD1 in miR‐99a‐5p transfectants was measured by qRT‐PCR and (E) protein expression was measured by western blot. The expression of SMARCD1 was lower than that in mock or miRNA control transfectants. n = 3. *P < 0.0001, Bonferroni/Dun method. The error bars indicate SEM. imagej was used for protein levels. (F) Presumed miRNA target sites in WT or deleted regions. (G) Dual‐luciferase reporter assay using vectors encoding putative miRNA target sites in WT or deleted regions. The luminescence intensity was significantly reduced by cotransfection of miR‐99a‐5p and a vector with the 3′‐UTR of WT. n = 3. *P < 0.0001, Bonferroni/Dun method. The error bars indicate SEM. (H) SMARCD1 mRNA expression in the BLCA cohort of TCGA. SMARCD1 expression in BLCA samples (n = 404) was compared with that in normal samples (n = 19). P < 0.0001, Mann–Whitney U tests. These experiments were repeated at least three times. GEM‐R BC, gemcitabine‐resistant bladder cancer; BLCA, bladder urothelial carcinoma; miRNA, microRNA; WT, wild‐type.

Fig. 4

Effect of SMARCD1 knockdown in GEM‐R BC cells. The knockdown efficiency of si‐SMARCD1 was verified by evaluating (A) the expression level of SMARCD1 mRNA measured by RT‐qPCR and (B) the SMARCD1 protein level measured by western blot analysis. n = 3. *P < 0.0001. The error bars indicate SEM. imagej was used for protein levels. (C) Cell proliferation by XTT assay. n = 6. *P < 0.01; **P < 0.0001; ns, not significant. The error bars indicate SEM. (D) Cell migration activity measured by wound healing assay. n = 4. *P < 0.01; **P < 0.0001; ns, not significant. The error bars indicate SEM. (E) Cell invasion activity measured by Matrigel invasion assay. Invasion cells were counted and compared. n = 6. *P < 0.05; **P < 0.01; ***P < 0.0001; ns, not significant. The error bars indicate SEM. All experiments were performed in quadruplicate and si‐SMARCD1 transfectants were compared with mock or si‐control transfectants. Transfection of si‐SMARCD1 increased the sensitivity of the GEM‐R BC cell line to gemcitabine. IC50 concentrations of gemcitabine were given to BC cells. Ten nanomolar si‐SMARCD1 transfection increased gemcitabine sensitivity in (F) GEM‐R‐BOY cells and (G) GEM‐R‐T24 cells, as determined by XTT assay. n = 6. *P < 0.0001. The error bars indicate SEM. The combination of si‐SMARCD1 transfection and gemcitabine administration had a clear additive effect on (H) GEM‐R‐BOY cells and (I) GEM‐R‐T24 cells and significantly inhibited cell proliferation. n = 6. *P < 0.001; **P < 0.0001. The error bars indicate SEM. The relationships between two groups were analyzed using Mann–Whitney U tests. The relationships between three or more groups were analyzed using the multiple comparison test with the Bonferroni/Dun method. These experiments were repeated at least three times. GEM, gemcitabine; GEM‐R BC, gemcitabine‐resistant bladder cancer.

Fig. 5

Overexpression of miR‐99a‐5p and knockdown of SMARCD1 inhibit tumor growth in vivo. After transfection in GEM‐R BC cells with miR‐99a‐5p or si‐SMARCD1, the cells were injected into the flanks of nude mice (n = 5 mice per group). The animals with tumors were sacrificed in experimental observation for 19 days, and tumors were removed and weighed. Comparison of tumor volume trends and explant weights of (A) GEM‐R‐BOY cells and (B) GEM‐R‐T24 cells transfected with miR‐99a‐5p. *P < 0.01; **P < 0.0001, Mann–Whitney U tests. The error bars indicate SEM. Photograph shows excised tissue in the Xenograft mouse model on day 19. Scale bar, 10 mm. Ki67 expression in the xenograft tumors transfected with miR‐99a‐5p was detected by immunostaining. Micrographs of (C) GEM‐R‐BOY cells and (D) GEM‐R‐T24 cells (DAB, 100×). Positive cells were counted and compared. n = 6. *P < 0.05, Mann–Whitney U tests. The error bars indicate SEM. Scale bar, 250 μm. Comparison of tumor volume trends and explant weights of (E) GEM‐R‐BOY cells and (F) GEM‐R‐T24 cells transfected with si‐SMARCD1. *P < 0.01. **P < 0.0001, Mann–Whitney U tests. The error bars indicate SEM. Photograph shows excised tissue in the Xenograft mouse model on day 19. Scale bar, 10 mm. Ki67 expression in the xenograft tumors transfected with si‐SMARCD1 was detected by immunostaining. Micrographs of (G) GEM‐R‐BOY cells and (H) GEM‐R‐T24 cells (DAB, 100×). Positive cells were counted and compared. n = 6. *P < 0.05, Mann–Whitney U tests. The error bars indicate SEM. Scale bar, 250 μm. GEM‐R BC, gemcitabine‐resistant bladder cancer.

Fig. 6

Relationship between SMARCD1 and cellular senescence in the GEM‐R BC cell line. Senescence induction by si‐SMARCD1 transfection was confirmed by β‐galactosidase staining. Quantification of β‐galactosidase staining in (A) GEM‐R‐BOY cells and (B) GEM‐R‐T24 cells. n = 6. *P < 0.005, Bonferroni/Dun method. The error bars indicate SEM. (C) A typical micrograph of (A) and (B). Scale bar, 100 μm. (D) Western blot confirming the expression of p21^waf1/cip1 in GEM‐R BC cells transfected with si‐SMARCD1. Quantification of β‐galactosidase staining of (E) GEM‐R‐BOY cells and (F) GEM‐R‐T24 cells in miR‐99a‐5p transfection. n = 6. *P < 0.0001, Bonferroni/Dun method. The error bars indicate SEM. (G) A typical micrograph of (E) and (F). Scale bar, 100 μm. (H) Western blot confirming the expression of p21^waf1/cip1 in GEM‐R BC cells transfected with miR‐99a‐5p. imagej was used for protein levels and quantification of β‐galactosidase staining. These experiments were repeated at least three times. GEM‐R BC, gemcitabine‐resistant bladder cancer; SA‐β‐gal, senescence‐associated beta‐galactosidase.