MYB Proto-oncogene-like 1-TWIST1 Axis Promotes Growth and Metastasis of Hepatocellular Carcinoma Cells.

MYB proto-oncogene-like 1 (MYBL1) has been reported to be a strong activator of transcription and plays an important role in the development of cancer. However, the precise biological function and molecular mechanism of MYBL1 in hepatocellular carcinoma (HCC) cells remain unclear. In the present study, we found that the expression of MYBL1 was markedly overexpressed in HCC cell lines and HCC samples, respectively. Moreover, MYBL1 expression positively correlated with tumor progression and inversely correlated with patient survival in 368 human HCC tissue samples. Overexpression of MYBL1 induced, whereas knockdown of MYBL1 reduced, HCC proliferation and metastasis both in vitro and in vivo. Furthermore, we demonstrated that HCC patients with high MYBL1 expression had significantly shorter overall and poorer disease-free survival than those with low MYBL1 expression. MYBL1 transcriptionally upregulated TWIST1 expression by directly targeting the TWIST1 promoter. More importantly, the in vitro analysis was consistent with the significant correlation between MYBL1 and TWIST1 expression observed in a large cohort of human HCC specimens. Taken together, our results demonstrate that MYBL1 plays an important role in HCC growth and metastasis and reveal a plausible mechanism for upregulation of TWIST1 in HCC.

Introduction

Hepatocellular carcinoma (HCC) is a frequently occurring malignancy and is a hazard to human health. Due to its continuously increasing incidence, especially in China, which account for 55% of all HCC cases worldwide, HCC ranks the third-leading cause of cancer-related death because of the poor prognosis and survival time., Several factors, such as hepatitis virus infection, metabolic abnormality, and liver fibrosis, can significantly increase the risk of developing HCC. Until now, radiotherapy, surgical hepatic resection, and liver transplantation might play a critical role in the treatment of HCC patients. Although after curative resection, the outcome for the treatment of HCC remains unsatisfactory, with survival rates of only 10%–20% at 5 years because of the late diagnosis and poor prognosis due to the high rate of recurrence, such as intrahepatic metastasis originating from the primary carcinoma.4, 5, 6 Therefore, further studies on the mechanisms underlying the HCC progression are urgently needed to develop effective diagnostic methods and therapeutic strategies to provide a potential target for the treatment of HCC.

TWIST1 is a transcription factor that belongs to the basic helix-loop-helix (bHLH) family, plays pivotal roles in multiple stages of embryonic development, and significantly contributes to tumor metastasis and primary tumor growth.7, 8, 9, 10 It is reported that overexpression of TWIST1 correlates with a worse prognosis of HCC patients and with a high expression of TWIST1 in HCC cell lines, shows a greater capacity for invasiveness/metastasis. Roberts et al. show that via upregulation of GAS6, L1CAM, and Akt signaling, TWIST1 confers cisplatin resistance and cell survival in HCC. Chen et al. also demonstrate that inhibition of Twist1 expression by A20 can suppresses hepatocellular carcinoma proliferation and metastasis. Therefore, TWIST1 may serve as a novel prognostic biomarker and potential therapeutic target for HCC patients, and the illumination of the molecular mechanism of TWIST1 in HCC is urged need.

MYB proto-oncogene-like 1 (MYBL1) is a novel tumor-related gene, which is located on chromosome 8q13.1 and codes for an open-reading frame of 733 amino acids. Previously, it has been found that the MYBL1 protein is a strong activator of transcription and that this activity depends on both the DNA-binding and acidic domains. Player et al. outline the discovery that six candidate genes, including MYBL1, as candidate genes associated with triple-negative breast cancer. By use of RNA sequencing (RNA-seq) analysis, Brayer and colleagues demonstrated that the recurrent fusions in MYB and MYBL1 define a common transcription factor-driven oncogenic pathway in salivary gland adenoid cystic carcinoma. Moreover, recurrent oncogenic truncating rearrangements in the transcription factor MYBL1 were found in the pediatric low-grade gliomas by genomic analysis. The above research indicates the association between MYBL1 and the development of cancer; however, the function and molecular mechanism of MYBL1 in cancer, particularly in HCC, were rarely reported.

Results

MYBL1 Overexpression Correlates with Progression and Poor Prognosis in Hepatocellular Carcinoma Patients

By analyzing the published mRNA expression profiles (The Cancer Genome Atlas [TCGA] datasets), we found that MYBL1 mRNA was significantly upregulated in hepatocellular carcinoma tissues compared with adjacent noncancer tissues (ANTs) (Figures 1A and 1B). Moreover, hepatocellular carcinoma patients with higher MYBL1 expression had a shorter survival time and demonstrated an earlier relapse disease-free survival time (p < 0.001; Figure 1C). Consistently, real-time PCR analyses revealed that MYBL1 was markedly overexpressed in all 9 hepatocellular carcinoma cell lines at both the protein and mRNA levels, compared with the immortalized, nonmalignant hepatocyte cell line (THLE3) (Figures 1D and 1F). Furthermore, comparative analyses showed that mRNA expression of MYBL1 was elevated in the eight hepatocellular carcinoma samples compared with nontumor liver specimens (Figure 1E), suggesting that MYBL1 is upregulated in hepatocellular carcinoma cells. To determine the clinical relevance of MYBL1 in hepatocellular carcinoma, MYBL1 expression was examined in 368 paraffin-embedded, archived hepatocellular carcinoma tissues by immunohistochemistry (IHC) assay. As shown in Figure 1H, the increased expression of MYBL1 was detected in the clinical hepatocellular carcinoma tissue samples but not detectable in adjacent normal liver tissue from the same HCC patient. Importantly, statistical analysis showed that hepatocellular carcinoma patients with high MYBL1 expression had significantly worse overall and disease-free survival than those with low MYBL1 expression (Figure 1I; Tables S2 and S3). These results suggest that MYBL1 has potential clinical value as a predictive biomarker for disease outcome in hepatocellular carcinoma.

Figure 1

Overexpression of MYBL1 Correlates with HCC Progression and Poor Prognosis

(A) Expression profiling of mRNAs showing that MYBL1 is upregulated in all 354 HCC tissues (T) compared to 50 adjacent non-cancer tissues (ANTs). (B) Expression profiling of mRNAs showing that MYBL1 is upregulated in 50 pairs HCC tissues (T) compared to the corresponding adjacent non-cancer tissues (ANT). (C) Kaplan-Meier analysis of overall (left) or progression-free (right) survival curves from TCGA dataset for HCC patients with low MYBL1 expression or high MYBL1 expression. *p < 0.05. (D) Real-time PCR analysis of MYBL1 expression in normal liver surface epithelial cell lines and in HCC cell lines. Transcript levels were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression. Each bar represents the mean ± SD of three independent experiments. (E) Real-time PCR analysis of MYBL1 expression in 8 pairs HCC tissues with match non-tumor tissues (N). Transcript levels were normalized to GAPDH expression. Each bar represents the mean ± SD of three independent experiments. (F) Western blotting analysis of MYBL1 expression in normal liver surface epithelial cell lines and cultured HCC cell lines. α-tubulin was used as a loading control. (G) Western blotting analysis of MYBL1 expression in 8 pairs in normal tissues and HCC samples. α-Tubulin was used as a loading control. (H) IHC staining indicating the MYBL1 protein expression in normal tissues and HCC. (I) The Kaplan-Meier survival curves compare HCC patients with low and high MYBL1 expression levels.

Upregulation of MYBL1 Promoted Proliferation and Metastasis on HCC In Vitro

Interestingly, gene set enrichment analysis (GSEA) revealed that MYBL1 overexpression strongly correlated with gene signatures associated with proliferation and metastasis in TCGA dataset of HCC, suggesting that MYBL1 overexpression might contribute to growth and metastasis in HCC (Figure 2A). To investigate the pro-growth and pro-metastasis role of MYBL1 in HCC progression, Huh7 and HepG2 cell lines, which stably express MYBL1, were established (Figure 2B). A 3-(4, 5-dimethyl-2-thiazolyl)-2, 5-diphenyl-2H-tetrazolium bromide (MTT) assay showed that ectopic expression of MYBL1 significantly increased the growth rate of Huh7 and HepG2 cells (Figure 2C). An anchorage-independent growth assay revealed that Huh7 and HepG2 cells stably expressing MYBL1 showed more and larger-sized colonies than control cells (Figure 2D). Furthermore, bromodeoxyuridine (BrdU) incorporation assay showed that the percentage of cells in S phase was dramatically increased in MYBL1-overexpressing HCC cells compared with the control cells (Figure 2E). Meanwhile, MYBL1-overexpressing cells also displayed a stronger metastasis and invasive capacity compared with the control cells (Figures 2F and 2G). These results indicated that MYBL1 overexpressing is involved in regulation of the pro-growth and pro-metastasis capability process of HCC cells in vitro.

Figure 2

Ectopic Expression of MYBL1 Promoted HCC Proliferation and Metastasis In Vitro

(A) GSEA plot, indicating a significant correlation between the mRNA levels of MYBL1 expression in HCC and the proliferation/metastasis gene signatures in published datasets. (B) Western blot analysis of MYBL1 in the indicated HCC cells. (C) MTT assay revealed that MYBL1 upregulation promoted HUH7 and HepG2 stable cell lines at indicated times after seeding. (D) Representative micrographs (left) and quantification of colonies >0.1 mm (right) were scored. Indicated cells (2 × 103) were suspended in soft agar and cultured for 10 days, and then colonies > 0.1 mm in diameter were counted. (E) Representative micrographs (left) and quantification of BrdU-positive signaling in the cells transfected with MYBL1 or Vector. (F and G) Quantification of indicated migrating (F) or invading cells (G) in ten random fields, analyzed by Matrigel-noncoated or -coated Transwell assays, respectively.

Inhibition of MYBL1 Repressed Proliferation and Metastasis on HCC In Vitro

We further examined the effect of MYBL1 inhibition on HCC cell proliferation and metastasis. Endogenous MYBL1 expression was silenced using a short hairpin RNA (shRNA; Figure 3A). Consistent with the overexpression results, MTT and anchorage-independent growth assays showed that MYBL1 suppression dramatically decreased the growth rate of HCC cells compared with that of control cells (Figures 3B and 3C). In addition, the BrdU incorporation assay revealed that MYBL1 suppression drastically decreased the percentage of HCC cells in the S peak compared with control cells (Figure 3D). Furthermore, MYBL1-inhibition cells displayed a lower metastasis and invasive capacity compared with the control cells (Figures 3E and 3F). Taken together, these results suggest that MYBL1 downregulation inhibited the proliferation and metastasis of HCC cells in vitro.

Figure 3

Inhibiting MYBL1 Expression Enhanced HCC Cell Proliferation

(A) Western blot analysis of MYBL1 in the indicated HCC cells. (B) MTT assay revealed that inhibition of MYBL1 repressed HUH7 and HepG2 stable cell lines at indicated times after seeding. (C) Representative micrographs (left) and quantification of colonies > 0.1 mm (right) were scored. Indicated cells (2 × 103) were suspended in soft agar and cultured for 10 days, and then colonies > 0.1 mm in diameter were counted. (D) Representative micrographs (left) and quantification of BrdU-positive signaling in the indicated cells. (E and F) Quantification of indicated migrating (E) or invading cells (F) in ten random fields analyzed by Matrigel-noncoated or -coated Transwell assays, respectively.

MYBL1 Upregulation Contributed to HCC Progression In Vivo

To investigate whether MYBL1 promoted HCC progression in vivo, we used a xenografted tumor model to examine the biological effect of MYBL1. As shown in Figures 4A–4C, the tumors formed by MYBL1-overexpression cells were larger in both size and weight than the tumors formed by control cells. Conversely, the tumors formed by MYBL1-inhibition-HCC cells were smaller and lighter than the control tumors. Furthermore, IHC analysis revealed that MYBL1-overexpressed tumors showed higher percentages of Ki-67-positive cells, whereas MYBL1-inhibition tumors displayed lower percentages of Ki-67-positive cells than the control tumors (Figure 4D). Collectively, these results demonstrate that MYBL1 functions as a tumor oncogenic gene in HCC in vivo.

Figure 4

MYBL1 Upregulation Contributed to HCC Progression In Vivo

(A) Representative images of tumor-bearing mice (left) and images of the tumors from all of the mice in each group (right). (B) Tumor volumes were measured on the indicated days. (C) Mean tumor weights. (D) IHC staining showing that MYBL1 upregulation increased the percentages of Ki-67-positive cells, whereas MYBL1 downregulation inhibited the percentages of Ki-67-positive cells. Each bar represents the mean ± SD of three independent experiments. ∗p < 0.05.

MYBL1 Directly Upregulates TWIST1

The molecular mechanism underlying the pro-growth and pro-metastasis capability process of MYBL1 was further explored in HCC cell lines. Interestingly, analysis of the TWIST1 promoter region predicted the presence of three MYBL1-specific nucleotide-binding sites (NBSs), suggesting that MYBL1 might transcriptionally regulate TWIST1 (Figure 5A). Indeed, it was confirmed by a chromatin immunoprecipitation (ChIP) assay in which MYBL1 was most significantly associated with the TWIST1 promoters, and overexpression of MYBL1 increased, but the silencing of MYBL1 decreased the luciferase activity of the TWIST1 promoter in HCC cells (Figures 5B and 5C). The expression of TWIST1, at both mRNA and protein levels, was dramatically increased in MYBL1-transduced cells but decreased in MYBL1-silenced cells (Figures 5D and 5E). Furthermore, we found that overexpressing MYBL1 increased, but silencing MYBL1 decreased, the luciferase activity driven by the second NBS of the TWIST1 promoter. However, neither overexpression nor knockdown of MYBL1 had any effect on the luciferase activities of the TWIST1 promoter that contained mutated second NBSs (Figure 5F). Moreover, the ChIP assay revealed that endogenous MYBL1 protein bound to the second NBS in the TWIST1 promoter (Figure 5G), demonstrating that MYBL1 regulates TWIST1 by directly targeting the TWIST1 promoter.

Figure 5

MYBL1 Directly Upregulates TWIST1

(A) Schematic of three MYBL1 binding sites in the TWIST1 promoter. (B) ChIP assay demonstrating that MYBL1 physically associated with the TWIST1 promoter. (C) Analysis of luciferase reporter activity in the indicated cells. (D) Real-time PCR analysis of TWIST1 mRNA expression in the indicated cells. Transcript levels were normalized to GAPDH. (E) Western blotting analysis of TWIST1 expression in the indicated cells. α-tubulin served as a loading control. (F) Transactivities of MYBL1 on serial TWIST1 promoter fragments as indicated in HCC cells. The luciferase activities of the promoter constructs were normalized to Renilla luciferase activity. (G) ChIP assay demonstrating that MYBL1 physically associated with the second NBS of the TWIST1 promoter. Each bar represents the mean ± SD of three independent experiments. ∗p < 0.05.

TWIST1 Is Required for MYBL1-Induced Pro-proliferation and Pro-metastasis in HCC

Next, we investigated whether MYBL1-induced pro-proliferation and pro-metastasis occurs through transcriptional activation of TWIST1 in HCC. As show in Figures 6A–6D, inhibition of TWIST1 significantly abrogates the pro-proliferation and pro-metastasis effect of MYBL1 on HCC in vitro. Importantly, inhibition of TWIST1 significantly repressed the effects of MYBL1 in vivo, as determined by quantification of the bioluminescence signal and proportion of Ki-67 +-cells, compared with that in the control group (Figures 6E and 6F). Taken together, these results indicate that TWIST1 is vital for MYBL1-induced pro-proliferation and pro-metastasis on HCC.

Figure 6

TWIST1 Is Required for MYBL1-Induced Pro-proliferation and Pro-metastasis in HCC

(A) Western blot analysis of TWIST1 in the indicated HCC cells. (B) Representative micrographs (left) and quantification of colonies > 0.1 mm (right) were scored. Indicated cells (2 × 103) were suspended in soft agar and cultured for 10 days, and then colonies > 0.1 mm in diameter were counted. (C) Representative micrographs (left) and quantification of BrdU-positive signaling in the indicated cells. (D) Quantification of indicated invading cells in ten random fields analyzed by Matrigel-coated Transwell assays. (E) The luminescence of the tumor xenografts from different treatment groups at the indicated weeks. (F) IHC staining demonstrated the expression of Ki-67-positive cells in the indicated tissues. ∗p < 0.05.

Clinical Relevance of MYBL1 and TWIST1 in HCC

Finally, we examined whether the MYBL1/TWIST1 axis identified in HCC cell models could be also verified in clinical HCC tumors. Analysis of 368 HCC tissue specimens using IHC analysis showed that MYBL1 expression was significantly correlated with the expression levels of TWIST1 (p < 0.001) (Figure 7A). Consistently, MYBL1 expression in 10 freshly collected clinical HCC samples also correlated with the protein levels of TWIST1 (r = 0.93, p < 0.01) (Figure 7B). These data further supported the notion that upregulation of MYBL1 in HCC induces TWIST1 expression and promotes proliferation and metastasis of HCC cell lines, ultimately leading to poor clinical outcomes for human HCC.

Figure 7

Clinical Relevance of MYBL1 and the Expression of TWIST1 in HCC

(A) MYBL1 levels were positively associated with TWIST1 expression in 368 primary human HCC specimens. Two representative cases are shown. (B) Expression analysis (left) and correlation (right) between MYBL1 expression and TWIST1 in 10 freshly collected human HCC samples. Each bar represents the mean ± SD of three independent experiments. ∗p < 0.05.

Discussion

Our study provides evidence that MYBL1 plays an important role in the malignant progression of HCC via transcriptional regulation of TWIST1 expression. IHC analysis revealed that MYBL1 was significantly upregulated in HCC and was associated with the clinical features and poor overall survival of HCC patients. Overexpression of MYBL1 induced, whereas knockdown of MYBL1 reduced, HCC proliferation and metastasis both in vitro and in vivo. Furthermore, we demonstrated that MYBL1 transcriptionally upregulated TWIST1 expression by directly targeting the TWIST1 promoter. Taken together, our results demonstrate that MYBL1 plays an important role in HCC growth and metastasis and reveal a plausible mechanism for upregulation of TWIST1 in HCC.

Due to the high morbidity of hepatitis B virus (HBV) and hepatitis C virus (HCV) infection, HCC leads to significant morbidity and mortality in Asian countries, especially in China, accounting for approximately 55% of all HCC cases worldwide. The lack of effective early diagnosis biomarkers impels HCC as the most common cause of cancer-related deaths with a very low overall survival rate, which is only 20%–65% for 1 year, 10%–30% for 3 years, and 10%–20% for 5 years. Additionally, the percentage of HCC patient survival, with a diagnosis of spinal metastasis, for 3 months, 6 months, 1 year, 2 years, and 5 years was 95.2%, 83.0%, 28.6%, 2.0%, and 1.4%, respectively. Therefore, identification and functional analysis of novel potential cancer-associated genes are of great importance for developing diagnostic, preventive, and therapeutic strategies for HCC treatment and management.

The transcription factors of the MYB family have been strongly implicated in the regulation of cell growth and differentiation in the development of nonhematopoietic tumors. For example, Brabender et al. suggest that upregulation of c-myb mRNA expression is an early event in the development of Barrett’s esophagus and Barrett’s-associated adenocarcinoma and that high c-myb mRNA expression levels may be a clinically useful biomarker for the detection of occult adenocarcinoma. Moreover, elevated MYB levels that result from the duplication clearly contribute to the differentiation block in these cases and together with NOTCH1 mutations, to proliferation and viability of the leukemic cells and showed MYB gene amplification and overexpression in 29% of tumors from women with BRCA1 mutations. More importantly, Guérin et al. reported nearly 2 decades ago that 64% of breast tumors expressed MYB and crucially, noted that MYB expression correlated strongly with the presence of estrogen receptor α (ERα) (∼70% of breast tumors are ERα+), which has been reproduced in several mRNA profiling studies., Herein, we found that MYBL1 was upregulated in HCC and that MYBL1 overexpression promoted HCC aggressiveness both in vitro and in vivo, which is in agreement with the oncogenic effect of MYB family members. However, the mechanism by which MYBL1 mediates transcriptional activation in HCC remains unknown. Interestingly, via performing the Coomassie brilliant blue assay and mass spectrum analysis, we found several transcriptional-activated-related factors, such as HAT1 and PRMT5 were the interaction proteins of MYBL1 (data not shown), suggesting that histone acetylation or methylation may involve regulation of MYBL1-mediated transcriptional activation in HCC. Thus, further studies of the mechanism by which MYBL1 mediates TWIST1 transcriptional activation in HCC will be of great interest.

TWIST, which was characterized by two α helices connected by a loop, is a bHLH transcription factor encoded by the Twist1 gene located on human chromosome 7p21. It is well known that Twist1 plays an essential role in malignant development of cancer through a variety of signal transduction pathways, including Akt, STAT3, mitogen-activated protein kinase (MAPK), Ras, and Wnt signaling, mainly via functioning as one of the major inducers of the Epithelial-Mesenchymal Transition (EMT) process,, For instance, Twist is phosphorylated at Ser42 in response to a serine/threonine-specific protein kinase Akt (PKB), subsequently inhibiting p53 activity in response to DNA damage, leading to aberrant cell-cycle regulation and inhibition of apoptosis through the phosphatidylinositol 3-kinase (PI3K)/Akt pathway during intrahepatic metastasis of HCC., Furthermore, several studies suggest that constitutively activated STAT3 was significant associated with Twist expression, which mediates HCC invasiveness and metastasis by regulating downstream genes (such as metastatic gene MMP-9 and the potent angiogenic gene VEGF) involved in antiapoptosis, proliferation, and angiogenesis suggesting that an abnormal p-STAT3/Twist signal axis may predict poor prognosis of HCC patients., The above-mentioned studies emphasized the important role of TWIST1 in the development HCC, and the regulation of Twist overexpression is also an important phenomenon in cancers. Twist protein stability is largely regulated by MAPK-mediated phosphorylation at the Ser68 position, and prevention of Ser 68 phosphorylation by an alanine (A) mutation (Ser 68A) dramatically accelerates Twist1 ubiquitination and degradation. There are also other proteins, such as HOXA5, claudins, and BMP7, involved in Twist regulation. Interestingly, in this study, we show that TWIST1 was transcriptionally regulated by MYBL1 and contributed to proliferation and metastasis of HCC cells, and histone acetylation or methylation may involve regulation of MYBL1-mediated TWIST1 transcriptional activation in HCC.

Our study shows that MYBL1 was markedly upregulated in HCC cells and clinical HCC samples, and a positive correlation was evident between MYBL1 expression and the recurrence-free survival of HCC patients. Overexpression of MYBL1 induced, whereas knockdown of MYBL1 reduced, HCC proliferation and metastasis both in vitro and in vivo. Furthermore, we demonstrated that MYBL1 transcriptionally upregulated TWIST1 expression by directly targeting the TWIST1 promoter. Taken together, our results demonstrate that MYBL1 plays an important role in HCC growth and metastasis and establish MYBL1 as a potential diagnosis biomarker and may serve as a putative target for HCC diagnosis and therapy.

Materials and Methods

Cell Culture

Normal liver epithelial cells, THLE3 and LO2, were from the American Type Culture Collection (ATCC; Manassas, VA, USA) and were cultured under the conditions stated by the manufacturer. The HCC cell lines were grown in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Carlsbad, CA, USA), supplemented with 10% fetal bovine serum (FBS; Invitrogen), at 37°C in a 5% CO2 atmosphere in a humidified incubator.

Tissue Specimens

Our study included 368 patients with HCC, diagnosed as HCC through pathological examinations, and treated from January 2005 to January 2010 in The First Affiliated Hospital of Sun Yat-sen University (Table S1). For the use of the clinical materials for research purposes, prior patient consent and approval were obtained from the Institutional Research Ethics Committee. The eight HCC tissues and the matched adjacent noncancerous tissues were frozen and stored in liquid nitrogen until further use.

Vectors, Retroviral Infection, and Transfection

pMSCV/MYBL1-overexpressing human MYBL1 was generated by subcloning the PCR-amplified human MYBL1 coding sequence into the pMSCV vector (Clontech, Mountain View, CA, USA). To silence endogenous MYBL1, two small interfering RNA (siRNA) oligonucleotides (detailed sequences showed in Table S4) were cloned to generate pSuper-retro-MYBL1-RNAi(s), respectively. The different regions of the human TWIST1 promoter were cloned into the NheI/BglII sites of the pGL3-basic luciferase reporter plasmid (Promega, Madison, WI, USA). Transfection of siRNA or plasmids was performed using the Lipofectamine 3000 reagent (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instruction. Stable cell lines expressing MYBL1 or MYBL1 RNAi were selected for 10 days with 0.5 μg/mL puromycin, 48 h after infection. Primers were listed in Table S4

Xenografted Tumor Model, IHC, and H&E Staining

BALB/c nude (BALB/c-nu) mice (4–5 weeks of age, 18–20 g) were purchased from the Center of Experimental Animal of Guangzhou University of Chinese Medicine. All experimental procedures were approved by the Institutional Animal Care and Use Committee of The First Affiliated Hospital of Sun Yat-sen University. The BALB/c nude mice were randomly divided into two groups (n = 5/group). One group of mice was inoculated subcutaneously with HepG2/vector cells (5 × 106) in the left dorsal flank and with HepG2/MYBL1 cells (5 × 106) in the right dorsal flank per mouse. Another group was inoculated subcutaneously with HepG2/RNAi-vector cells (5 × 106) in the left dorsal flank and with HepG2/MYBL1-RNAi cells (5 × 106) in the right dorsal flank. Tumors were examined twice weekly; length (L) and width (W) measurements were obtained with calipers, and tumor volumes were calculated using the equation (L × W2)/2. On day 40, tumors were detected by an IVIS imaging system; animals were euthanized; and tumors were excised, weighed, and paraffin embedded. Serial 6.0 μm sections were cut and subjected to IHC staining using anti-MYBL1 or Ki-67 antibodies. The images were captured using the AxioVision Rel.4.6 computerized image analysis system (Carl Zeiss, Germany).

IHC

IHC analysis was performed on the 368 paraffin-embedded HCC tissue sections, as previously described. The degree of immunostaining of formalin-fixed, paraffin-embedded sections was reviewed and scored separately by two independent pathologists. The scores were determined by combining the proportion of positively stained tumor cells and the intensity of staining. Tumor cell proportions were scored as follows: 0, no positive tumor cells; 1, <10% positive tumor cells; 2, 10%–35% positive tumor cells; 3, 35%–75% positive tumor cells; 4, >75% positive tumor cells. Staining intensity was graded according to the following standard: 1, no staining; 2, weak staining (light yellow); 3, moderate staining (yellow brown); 4, strong staining (brown). The staining index (SI) was calculated as the product of the staining intensity score and the proportion of positive tumor cells. With the use of this method of assessment, we evaluated protein expression in benign esophageal epithelia and malignant lesions by determining the SI, with possible scores of 0, 2, 3, 4, 6, 8, 9, 12, and 16. Samples with a SI ≥ 8 were determined as high expression, and samples with a SI <8 were determined as low expression. Cutoff values were determined on the basis of a measure of heterogeneity using the log-rank test with respect to overall survival.

ChIP

Cells (2 × 106) in a 100-mm culture dish were harvested and treated with 1% formaldehyde to cross link the proteins to DNA and quenched with 0.125 M glycine for 5 min. The cells were lysed in 1 mL lysis buffer 1 (1 M HEPES-KOH, 1 M NaCl, 0.5 M EDTA, 50% glycerol, 10% Nonidet P-40 [NP-40], 10% Triton X-100, pH 7.5) and rotated for 15 min at 4°C. Cell lysates were centrifuged at 2,300 × g for 5 min at 4°C to isolate the nuclei. Nuclei were suspended in 1 mL of lysis buffer 2 (1 M Tris-HCl, 1 M NaCl, 0.5 M EDTA, 0.5 M EGTA, pH 8.0). Cell lysates were centrifuged at 2,300 × g for 5 min at 4°C to isolate precipitate and suspended in 1 mL of lysis buffer 3 (1 M Tris-HCl, 1 M NaCl, 0.5 M EDTA, 0.5 M EGTA, 5% Na-deoxycholate, 5% N-lauroylsarcosine, pH 8.0). The cells were subjected to sonication to shear chromatin fragments to an average size between 400 bp and 700 bp on the Branson Sonifier 450 ultrasonic processor (40% amplitude, 0.5 s on 1 s off for 30 s). Fragmented chromatin was centrifuged at 12,000 × g for 10 min at 4°C. Supernatant was then incubated with 1 μg of anti-flag antibody (Sigma) or an anti-immunoglobulin G (IgG) antibody (Millipore, Billerica, MA, USA) overnight at 4°C with rotation. After overnight incubation, Dynabeads were washed twice with 1 mL of wash buffer (1 M HEPES-KOH, 5 M LiCl, 0.5 M EDTA, 5% Na-deoxycholate, 10% NP-40, pH 8.0) and twice with 1 mL of buffer (1 M Tris-HCl, 0.5 M EDTA, 1 M NaCl, pH 8.0). The chromatin was eluted in SDS elution buffer (10% SDS, 0.5 M EDTA, 1 M Tris-HCl, pH 8.0), followed by reverse crosslinking at 65°C overnight. The ChIP DNA was treated with RNase A (5 mg/mL) or protease K (0.2 mg/mL) at 37°C for 30 min and purified using QIAquick Spin Columns (QIAGEN). After reverse crosslink of protein/DNA complexes to free DNA, PCR was performed, and primers were listed in Table S4.

Western Blotting

Western blotting was performed, according to a previously reported method. The membranes were probed with polyclonal rabbit antibodies, anti-MYBL1 antibody (Abcam, Cambridge, MA, USA), and anti-TWIST1 (Cell Signaling Technology, Danvers, MA, USA). The membranes were then stripped and re-probed with an anti-α-tubulin mouse monoclonal antibody (Cell Signaling Technology) as a loading control.

Luciferase Assay

Cells were seeded in triplicate in 24-well plates and allowed to settle for 24 h. One hundred nanograms of pGL3-TWIST1-luciferase plasmid was transfected into HCC cells using the Lipofectamine 3000 reagent, according to the manufacturer’s instruction. Luciferase and control signals were measured at 48 h after transfection using the Dual Luciferase Reporter Assay Kit (Promega), according to a protocol provided by the manufacturer. Three independent experiments were performed, and the data were presented as the mean ± standard deviation (SD).

MTT Assay

800 cells were seeded on 96-well plates and stained at indicated time points with 100 μL sterile MTT dye (0.5 mg/mL; Invitrogen) for 4 h at 37°C, followed by removal of the culture medium and addition of 150 μL of dimethyl sulfoxide (DMSO; Sigma). The absorbance was measured at 570 nm, with 655 nm as the reference wavelength. All experiments were performed in three independent experiments.

Bromodeoxyuridine Labeling and Immunofluorescence

HCC cells grown on coverslips (Fisher, Pittsburgh, PA, USA) were incubated with BrdU for 1 h and stained with anti-BrdU antibody (Upstate, Temecula, CA, USA), according to the manufacturer’s instruction. Signal intensities from BrdU immunostaining and 4′,6-diamidino-2-phenylindole (DAPI) staining were determined at exposures in the linear range by a densitometry program (AxioVision Rel.4.6; Carl Zeiss, Germany) and analysis by Image-Pro Plus 6.0.

Anchorage-Independent Growth-Ability Assay

Cells (1 × 103) were trypsinized and suspended in 2 mL complete medium plus 0.33% agar (Invitrogen). The mixture was plated on top of a bottom layer comprising a 0.66% complete medium-agar mixture. After 10 days of incubation, colony sizes were measured with an ocular micrometer, and colonies greater than 0.1 mm in diameter were scored. All experiments were performed in three independent experiments.

Migration Assay

HCC cells were plated into the top side of polycarbonate Transwell filter of the BioCoat Invasion Chambers (BD, Bedford, MA, USA), and condition media collected form indicated cells were added in the lower compartment. HCC cells were incubated at 37°C for 22 h, followed by removal of cells inside the upper chamber with cotton swabs. Migratory cells on the lower membrane surface were fixed in 1% paraformaldehyde, stained with hematoxylin, and counted (ten random 100 × fields per well). Cell counts are expressed as the mean number of cells per field of view. Three independent experiments were performed, and the data are presented as mean ± SD.

Invasion Assay

Cells (2 × 104) were plated on the top side of a polycarbonate Transwell filter (precoated with Matrigel) in the upper chamber of a BioCoat Invasion Chamber (BD Biosciences) and incubated at 37°C for 22 h. The cells remaining on the upper surface were removed with cotton swabs. Cells that had migrated to the lower membrane surface were fixed in 1% paraformaldehyde, stained with hematoxylin, and counted under an optical microscope (×100 magnification). Cell counts are expressed as the mean number of cells from ten random fields per well.

Statistical Analysis

Student’s t test was used to evaluate the significant difference between two groups of data in all of the pertinent experiments. The chi-square test was used to analyze the relationship between MYBL1 expression and the clinical pathological characteristics. Statistical analyses were performed using the SPSS 21.0 statistical software package. Data were represented as the mean ± SEM. A p value <0.05 (using a two-tailed paired t test) was considered statistically significant.

Microarray Data Process and Visualization

Microarray data were downloaded from the GEO database (https://www.ebi.ac.uk/arrayexpress/) or TCGA dataset (http://software.broadinstitute.org/software/igv/tcga);

GSEA was performed using GSEA 2.0.9 (http://www.broadinstitute.org/gsea/).

Author Contributions

J.L., Q.D., and H.L. conceived and designed the experiments and conceived and supervised the project. B.X., Yao Liu, and Z.Z. performed the experiments. B.X., Yao Liu, Z.Z., Q.L., X.W., and Y.X. analyzed the data. Yanhong Liu, Yuwen Liu, and Y.Y. contributed reagents/materials. H.L. and J.L. contributed clinical specimens. Q.D. and H.L. wrote the paper.

Conflicts of Interest

The authors declare no competing interests.