An E2F1/DDX11/EZH2 Positive Feedback Loop Promotes Cell Proliferation in Hepatocellular Carcinoma

Abstract

Hepatocellular carcinoma (HCC) accounts for one of the leading causes of cancer-related death, and is attributed to the dysregulation of genes involved in genome stability. DDX11, a DNA helicase, has been implicated in rare genetic disease and human cancers. Yet, its clinical value, biological function, and the underlying mechanism in HCC progression are not fully understood. Here, we show that DDX11 is upregulated in HCC and exhibits oncogenic activity via EZH2/p21 signaling. High expression of DDX11 is significantly correlated with poor outcomes of HCC patients in two independent cohorts. DDX11 overexpression increases HCC cell viabilities and colony formation, whereas DDX11 knockdown arrests cells at G1 phase without alteration of p53 expression. Ectopic expression of DDX11 reduces, while depletion of DDX11 induces the expression of p21. Treatment of p21 siRNA markedly attenuates the cell growth suppression caused by DDX11 silence. Further studies reveal that DDX11 interacts with EZH2 in HCC cells to protect it from ubiquitination-mediated protein degradation, consequently resulting in the downregulation of p21. In addition, E2F1 is identified as one of the upstream regulators of DDX11, and forms a positive feedback loop with EZH2 to upregulate DDX11 and facilitate cell proliferation. Collectively, our data suggest DDX11 as a promising prognostic factor and an oncogene in HCC via a E2F1/DDX11/EZH2 positive feedback loop.

Keywords: DDX11; E2F1; EZH2; hepatocellular carcinoma; p21.

Figure 1

DDX11 expression is increased in HCC tissues. (A) The mRNA expression of DDX11 was determined in 24 pairs of fresh HCC and nontumorous (NT) samples, using qRT-PCR in SYSUCC cohort (Matched pair t test). (B) The increased expression of DDX11 mRNA was validated in 50 paired HCC tissues in TCGA cohort (Matched pair t test). (C) GEO studies indicate the level of DDX11 mRNA in HCC cases was higher than that in nontumorous (NT). (D) Proteins extracted in clinical samples were subjected to western blot to examine the DDX11 protein expression. β-actin was used as a loading control. T, tumor samples; N, nontumorous samples. (E) A large cohort containing 328 HCC cases was recruited. The protein level of DDX11 was measured by IHC in paraffin-embedded tissues. Representative images and relative IHC scores were presented in the left and right panels, respectively. (F) Another cohort consisting of 47 paired primary HCC and corresponding portal vein tumor thrombus tissues was collected to detect the expression of DDX11 by IHC.

Figure 2

High expression of DDX11 is correlated with poor prognosis in HCC. (A, B) Patients with HCC in TCGA cohort were divided into DDX11^high and DDX11^low, according to the median expression of DDX11 mRNA. Kaplan-Meier survival analyses (log-rank test) were conducted to indicate the clinical value of DDX11 mRNA in overall (A) and disease-free (B) survivals. (C, D) The implication of DDX11 protein expression was determined in patients with HCC in SYSUCC cohort. The median IHC score of DDX11 in HCC tissues was used as the cut-off value to separate cases into groups of DDX11^high and DDX11^low. (E) The impact of DDX11 mRNA on overall survival (OS) and disease-free survival (DFS) was tested in TCGA cancers (Mantel–Cox test). Significant p values were indicated by red and blue frames.

Figure 3

DDX11 promotes cell proliferation in HCC. (A) Cells were transfected with DDX11 shRNA and selected by G418 to construct stable cell lines with DDX11 knockdown. qRT-PCR and western blot were used to confirmed the downregulation of DDX11 mRNA and protein in HepG2 and PLC8024 cells. *P < 0.05. (B) The effect of DDX11 silence on HCC cell growth were determined by MTT assays. Stable cells were placed into 96-well plates to continue to grow for 4 days. Cell growth rates were calculated by the absorbance at OD490 nm. *P < 0.05, **P < 0.01. C. Colony formation was performed to validate the role of DDX11 on cell proliferation. Stable cells were cultured in 60 mm plates with G418-contained DMEM for 10 days. Colonies were pictured and counted under a microscope (left panel). The fold change (FC) of colony formation was indicated by histogram (right panel). *P < 0.05. (D) Transwell assays were performed to assess whether DDX11 affects cell migration. Cells transferred to the bottom side of the Transwell chamber were stained by 0.1% crystal violet and counted. The fold change (FC) of cell migration were shown. *P < 0.05. (E–H) Stable cells with DDX11 overexpression were constructed (E) and used in the determination of DDX11-mediated cell growth by MTT (F), colony formation (G) and Transwell (H) assays. *P < 0.05, **P < 0.01.

Figure 4

DDX11 depletion induces G1 phase arrest in HCC cells. (A) GSEA based on TCGA data indicated that in cases with high expression of DDX11, pathways involved in cell cycle regulation were activated or suppressed. (B) Cells with DDX11 knockdown by shRNA were stained with PI and subjected to cytometry analyses to indicate the alteration of cell cycle. The percentages of cells at each cell cycle or under apoptosis were shown. (C) The expressions of factors contributing to cell cycle regulation, such as p53, MDM2, p21, cyclin E, cyclin D1, and phosphorylated Rb were determined by western blot, in cells with DDX11 silence or overexpression. (D) Stable HepG2 and PLC8024 cells were transfected with p21 siRNA for 36 h. The expression of p21 and DDX11 was determined. (E, F) The effect of p21 on DDX11-mediated G1 phase arrest and cell growth suppression was examined by rescue experiments. Stable cells with DDX11 silence and/or p21 knockdown were subjected to cytometry analyses (E) and colony formation (F). *P < 0.05.

Figure 5

DDX11 suppresses p21 via enhance the protein stability of EZH2 in HCC cells. (A) GSEA indicated that EZH2 signaling was activated in patients with high expression of DDX11. (B) In TCGA cases, a positive correlation between DDX11 mRNA and EZH2 mRNA expression was found. (C) The protein expression of DDX11 was associated with EZH2 protein expression in 24 fresh HCC specimens in SYSUCC cohort. (D) HepG2 and PLC8024 cells were transfected with DDX11 shRNA or overexpression vectors. The mRNA expression of EZH2 was examined by qRT-PCR. (E) The expression of DDX11, EZH2 and p21 in stable cells with DDX11 knockdown or overexpression was examined by western blot. (F) DDX11-expressing cells were transfected with EZH2 siRNA for 36 h. The effect of EZH2 on DDX11-mediated p21 suppression was tested. (G) The protein binding of EZH2 and DDX11 was confirmed by co-IP experiments. (H) The EZH2 protein stability was measured by CHX treatment in cells with DDX11 knockdown. *P < 0.05, **P < 0.01. (I) Cells were cultured with 20 mg/L CHX for indicated time, with or without pretreatment of MG132 (20 μM). The protein degradation of EZH2 was determined by western blot. (J) HCC cells with or without DDX11 knockdown were transfected with Ub for 48 h. The ubiquitination of EZH2 protein was examined by co-IP mediated by Ub antibody. (K) The role of EZH2/p21 axis was assessed by rescue experiment, using colony formation. *P < 0.05.

Figure 6

E2F1/DDX11/EZH2 forms a positive feedback loop in HCC cells. (A) GSEA indicated that DDX11 may be a downstream target of E2F transcription factors. (B, C) HepG2 and PLC8024 cells were transfected with siRNAs for E2F family members, including E2F1, E2F2, and E2F3. The expression of E2Fs and DDX11 mRNA was determined by qRT-PCR (B) and western blot (C). (D) E2F1 was overexpressed in HCC cells. Expression of E2F1, DDX11, EZH2, and p21 was examined. (E) Dual luciferase reporter assays were performed in HepG2 cells with E2F1 overexpression or knockdown to indicate the effect of E2F1 on the activity of DDX11 promoter. **P < 0.01, ***P < 0.001. (F) ChIP assays were used to detect the enrichment of E2F1 on DDX11 promoter. *P < 0.05. (G) Correlation between DDX11 mRNA and E2F1 was determined in 24 HCC tissues (Pearson correlation analysis). (H) The positive correlation of E2F1 and DDX11 protein expression was confirmed in 303 paraffin-embedded HCC tissues. Patients with high expression of E2F1 were accompanied with more DDX11 expression. (I) Cells with E2F1 silence were transfected with DDX11 overexpression vector. Colony formation was performed to examine the role of DDX11 in shE1F1-mediated cell growth suppression. *P < 0.05. (J) Cells were transfected with E2F1 siRNA and DDX11 overexpression vector for 36 h. The mRNA expression of EZH2 was examined. ns, not significant. (K) According to the published data (GDS2445), DDX11 mRNA was downregulated in EZH2^-/- cells. (L) The association of EZH2 and DDX11 was determined in TCGA cases. (M) Cells were overexpressed with EZH2 and/or knockdown of E2F1. The mRNA expression of DDX11 was examined. *P < 0.05.