TGFβ1 Promotes Gemcitabine Resistance through Regulating the LncRNA-LET/NF90/miR-145 Signaling Axis in Bladder Cancer

Abstract

High tumor recurrence is frequently observed in patients with urinary bladder cancers (UBCs), with the need for biomarkers of prognosis and drug response. Chemoresistance and subsequent recurrence of cancers are driven by a subpopulation of tumor initiating cells, namely cancer stem-like cells (CSCs). However, the underlying molecular mechanism in chemotherapy-induced CSCs enrichment remains largely unclear. In this study, we found that during gemcitabine treatment lncRNA-Low Expression in Tumor (lncRNA-LET) was downregulated in chemoresistant UBC, accompanied with the enrichment of CSC population. Knockdown of lncRNA-LET increased UBC cell stemness, whereas forced expression of lncRNA-LET delayed gemcitabine-induced tumor recurrence. Furthermore, lncRNA-LET was directly repressed by gemcitabine treatment-induced overactivation of TGFβ/SMAD signaling through SMAD binding element (SBE) in the lncRNA-LET promoter. Consequently, reduced lncRNA-LET increased the NF90 protein stability, which in turn repressed biogenesis of miR-145 and subsequently resulted in accumulation of CSCs evidenced by the elevated levels of stemness markers HMGA2 and KLF4. Treatment of gemcitabine resistant xenografts with LY2157299, a clinically relevant specific inhibitor of TGFβRI, sensitized them to gemcitabine and significantly reduced tumorigenecity in vivo. Notably, overexpression of TGFβ1, combined with decreased levels of lncRNA-LET and miR-145 predicted poor prognosis in UBC patients. Collectively, we proved that the dysregulated lncRNA-LET/NF90/miR-145 axis by gemcitabine-induced TGFβ1 promotes UBC chemoresistance through enhancing cancer cell stemness. The combined changes in TGFβ1/lncRNA-LET/miR-145 provide novel molecular prognostic markers in UBC outcome. Therefore, targeting this axis could be a promising therapeutic approach in treating UBC patients.

Keywords: bladder cancer.; cancer stem-like cells; lncRNA-LET; miRNA biogenesis; tumor recurrence.

Figure 1

Cancer stemness markers in cells treated with chemotherapeutic agents in vivo, and downregulation of lncRNA-LET enhances CSC-like properties. (A) In vivo GEM chemotherapy simulates clinical regimen with multiple GEM treatment cycles (dashed boxes) and gap periods. Tumor sizes of T24 xenografts were measured for GEM treatment and vehicle control group (n = 6 per group). (B) Sphere formation assay of primary cells derived from T24 xenografts of control group (Veh) and GEM (n=3 per group). (C) Representative H&E and IHC data showing the expression levels of CSC markers (CK5 and CK14) in T24 xenografts of control and GEM groups (n = 6 per group). (D) Western blotting of CSC markers (ALDH1A1, CD44, KLF4, and HMGA2) in T24 xenografts of control and GEM groups. (E) The levels of lncRNA-LET in T24 and 5637 cells treated with GEM or vehicle. (F) Knockdown efficiency of lncRNA-LET in T24 cells and forced expression of lncRNA-LET in J82 cells. (G-H) ALDH^high population (G) and CSC markers (H) were determined by flow cytometer and Western blotting in T24 with knockdown and J82 cells with overexpression of lncRNA-LET (n=3 per group). (I) Knockdown efficiency of lncRNA-LET in T24 cells infected with shRNA targeting lncRNA-LET virus (shLET), compared to control (shCTL). (J) Tumor volume of shCTL and shLET T24 cells with limiting dilution. Each dot represents an individual mouse (n=6 per group). Data are shown as mean ± SD and represent three independent experiments with similar results. * P < 0.05; ** P < 0.01, *** P < 0.001 (Student's unpaired two-tailed t-test). (K) CSC frequency was calculated using extreme limiting dilution analysis (ELDA). There was a significant difference in CSC frequency between two groups, with a P value of 0.025.

Figure 2

The role of lncRNA-LET in GEM-induced cancer stemness. (A) qRT-PCR showed stable overexpression of lncRNA-LET in T24 cells (pLET). (B) CSC markers were determined by Western blotting in T24 cells overexpressed lncRNA-LET and control vector. (C-E) Tumor size change (C), the levels of lncRNA-LET (D) and representative IHC (E) of xenografts from control (Vec) and overexpression (pLET) cells after GEM or vehicle treatment (Vec+Veh group n=5, other group n=6). Dashed boxes indicate the time frame of each GEM chemotherapy cycle. Data are shown as mean ± SD and represent at least two independent experiments with similar results. * P < 0.05, ** P < 0.01, *** P < 0.001 (One-way ANOVA followed by Newman-Keuls multiple comparison test).

Figure 3

TGFβ1 represses lncRNA-LET expression in UBC cells. (A) Western blotting showing the levels of p-SMAD2 and total SMAD2 in xenografts treated with or without GEM. (B-C) mRNA levels of lncRNA-LET were measured by qRT-PCR in T24 and 5637 cells treated with or without TGFβ1 or TGFβR1 inhibitor, SB-431542, as well as transfected with control (siNC) or 2 different RNAi of SMAD4 (siSMAD4#1, siSMAD4#2), followed by vehicle or TGFβ1 treatment. (D) Conserved SMAD4 binding element (SBE) in the promoter of human and Rhesus Monkey lncRNA-LET. (E) Relative luciferase activity in T24 cells transfected with plasmids of wild type (pGL3-LET-WT) and SBE deleted (pGL3-LET-ΔSBE) lncRNA-LET promoter, respectively, followed by treatment with vehicle or TGFβ1. (F-G) Sphere numbers (F) and expression of CSC markers (G) in T24 and 5637 cells treated with TGFβ1 alone or TGFβ1 and forced expression of lncRNA-LET. Data are shown as mean ± SD and represent at least two independent experiments with similar results. * P < 0.05, ** P < 0.01, *** P < 0.001 (Student's unpaired two-tailed t-test).

Figure 4

The stabilized NF90 by the reduced lncRNA-LET is required for cancer cell stemness. (A) Western blotting showing NF90 protein level in T24 with depletion of lncRNA-LET and J82 cells with overexpression of lncRNA-LET. (B) Protein level of NF90 in J82 cells over-expressed lncRNA-LET, followed by 50 μg/ml cycloheximide (CHX) treatment for the indicated time points. (C-D) ALDH^high population (C) and expression of CSC markers (D) were determined by flow cytometer and Western blotting in control (siNC), lncRNA-LET knockdown (siLET#2), and simultaneous knockdown lncRNA-LET and NF90 (siLET#2 + siNF90#1 or siLET#2 + siNF90 #2) of T24 cells (n=3 per group). (E) Protein level of NF90 was determined by Western blotting in untreated and GEM resistant T24 cells. (F) The ALDH^high population was determined by flow cytometer in control (siNC) and NF90 knockdown (siNF90 #1 and #2) T24 cells treated with or without GEM. (G) Western blotting showing the levels of CSC markers in control (siNC) and NF90 knockdown (siNF90 #1 and #2) T24 cells, followed by the treatment with or without GEM. Data are shown as mean ± SD and represent at least two independent experiments with similar results. * P < 0.05, ** P < 0.01, *** P < 0.001 (Student's unpaired two-tailed t-test).

Figure 5

NF90 regulates miR-145 biogenesis. (A) Cluster map of microarray showed the altered miRNAs ≥ 2 folds in NF90 knockdown cells, compared with control group (siNC). (B) mRNA level of miR-145 was measured by qRT-PCR in control (siNC) and NF90 knockdown (siNF90) T24 and 5637 cells. (C) qRT-PCR showing the level of miR-145 in control (siNC) and lncRNA-LET knockdown (siLET-#1 and siLET-#2) T24 cells. (D) The levels of miR-145 and miR-143 in control (shCTL) and lncRNA-LET stable knockdown (shLET) T24 cells. (E) RIP combined with qRT-PCR assays of NF90 binding to pri-miR-143 or pri-miR-145 in control (shCTL) or lncRNA-LET knockdown (shLET) T24 cells. Relative enrichment of pri-miRNAs in anti-NF90 group, compared with IgG group. The control GAPDH mRNA level was used for normalization. (F) Schematic illustration of wild-type (WT) and mutant (Mut) miR-145 with consensus binding motif of NF90. (G) qRT-PCR analysis of exogenous miR-145 levels in T24 cells, transfected with wild-type or mutant NF90 binding site pri-miR-145 vector. Expression of FLAG in Western blotting was used to represent the NF90 transfection efficiency. Data are shown as mean ± SD and represent at least two independent experiments with similar results. * P < 0.05, ** P < 0.01, *** P < 0.001 (Student's unpaired two-tailed t-test).

Figure 6

Inhibition of TGFβ1 signaling pathway enhances chemosensitivity to GEM treatment in T24 UBC xenografts. (A) Tumor size change of xenografts from chemoresistant T24 cells treated with vehicle, LY2157299, GEM or both LY2157299 and GEM (n=6 per group). (B) Tumor volumes calculated on day 30 (n=6 per group). (C) ALDEFLUOR assay. The primary cells derived from xenografts were harvested on day 30, and analyzed for ALDH activity. (D) Western blotting of pSMAD2 and total SMAD2 in xenografts of the four groups. Data are shown as mean ± SD. ** P < 0.01 (Student's unpaired two-tailed t-test).

Figure 7

Clinical relevance of lncRNA-LET/NF90/miR-145 axis in UBC specimens. (A). The levels of lncRNA-LET, TGFβ1, miR-145, and that of NF90 in human UBC (Tumor) and the adjacent normal tissues (Normal). Mann Whitney test, Unpaired t test or Wilcoxon signed rank test was used for analysis of statistical significance. (B) Western blotting showing NF90 expression in paired UBC (T) and adjacent normal tissues (N). (C) The correlations among lncRNA-LET, TGFβ1, miR-145 and NF90 were determined by Pearson correlation analysis. ΔCT values of qRT-PCR were used as the levels of lncRNA-LET, TGFβ1 and miR-145 transcripts. NF90 protein bands on Western blotting were semi-quantified by Image J. (D) Kaplan-Meier survival analysis of overall survival of 60 UBC patients with expression profile of TGFβ1^high vs TGFβ1^low, lncRNA-LET ^high vs lncRNA-LET^low, miR-145 ^high vs miR-145^low. Subgroup analysis of UBC patients according to the expression profile of TGFβ1^high/lncRNA-LET^low/miR-145^low signature versus the other combinations. Median value was chosen as cut-off point. The log-rank (Mantel-Cox) test was used to calculate P-values. P < 0.05 was considered statistically significant. (E) Schematic illustration of enhancement of cancer stem-like cell properties in GEM resistant UBC cells by the dysregulation of lncRNA-LET/NF90/miR-145 axis via TGFβ1.