Yes-associated protein (YAP) binds to HIF-1α and sustains HIF-1α protein stability to promote hepatocellular carcinoma cell glycolysis under hypoxic stress

Between July 2012 and November 2014, we obtained 54 tumor tissues and the paired adjacent noncancerous tissues from patients who had undergone surgical hepatectomy at the Liaoning Cancer Hospital and Institute (Shenyang, China) and who has been diagnosed with HCC by pathological examination. The patients had not undergone preoperative chemotherapy or radiotherapy. The China Medical University ethics committee approved this study (NO. 2014PS132K). All patients granted informed consent. The RNA sequencing (RNA-Seq) data from TCGA were obtained from Gene Expression Profiling Interactive Analysis (GEPIA) [23], which is based on the University of California Santa Cruz (UCSC) Xena project (http://xena.ucsc.edu).

The HepG2 and Huh7 human HCC cell lines were purchased from Shanghai Institute of Cell Bank (Shanghai, China). HepG2 cells were grown in Eagle’s Minimum Essential Medium and Huh7 cells were grown in Dulbecco’s Modified Eagle Medium. All medium containing 10% fetal bovine serum (FBS), penicillin (100 units/ml), streptomycin (100 μg/ml), 2 mM glutamine, and 10 mM HEPES buffer. Cells were cultured at 37 °C under 20% O₂ (normoxia) or 1% O₂ (hypoxia), balanced with N₂ in a 3-gas incubator.

We extracted RNA with TRIzol (Invitrogen, Carlsbad, CA, USA) and performed real-time PCR as has been described previously [24]. The primers used are listed in Additional file 1: Table S1.

All HCC cell transfections were performed in 6-well plates using Invitrogen Lipofectamine 3000 (Thermo Fisher Scientific, Shanghai, China) as per instructions from the manufacturer. We harvested the cells after 48 h to perform real-time PCR or western blotting. Expression plasmid for YAP5SA was obtained from Dr. Dawang Zhou and Dr. Lanfen Zhen [25]. The sense sequences of the YAP small interfering RNAs (siRNAs) (siYAPs, Sigma, Shanghai, China) were as follows:

We performed immunohistochemical staining and western blotting using methods that have been described previously [24]. Anti-YAP (1:100, immunohistochemistry), Anti–phosphorylated-YAP (p-YAP) (Ser127) (1:1000, western blot), and anti–p-LATS1 (Ser909) (1:1000, western blot) were from Cell Signaling Technology (Shanghai, China). Anti-YAP (1:1000, western blot), anti-LDHA (1:1000, western blot), anti-PKM2 (1:1000, western blot), anti-Glut1 (1:1000, western blotting), anti-PGK1 (1:1000, western blot), anti–large tumor suppressor kinase 1 (LATS1, 1:1000, western blotting), and anti–HIF-1α (1:1000, western blot) rabbit monoclonal antibodies were from Proteintech (Wuhan, China). Anti–matrix metalloproteinase 2 (MMP2, 1:1000, western blot), anti-MMP9 (1:1000, western blot), horseradish peroxidase–linked secondary antibody (1:5000, western blot), and anti–β-actin (1:3000, western blot) were from Abbkine (Wuhan, China). The protein bands were analyzed with Image J software (National Institutes of Health, Bethesda, MD, USA).

The HCC cells were cultured, plated in 6-well plates, washed with phosphate-buffered saline (PBS), and fixed in 4% polyformaldehyde. Primary antibody against YAP (1:100) or HIF-1α (1:100) was added on the plates of cells or frozen sections of HCC tissues, and the plates were incubated at 4 °C overnight. After washing, we added fluorescein isothiocyanate (FITC)-labeled (1:100) secondary antibody (Beyotime, Shanghai, China) and incubated the samples for 2 h. We counterstained the cells using diaminophenylindole (DAPI, Beyotime, Shanghai, China) and visualized them under a confocal microscope.

We assessed HCC cell migration and invasive capability using 24-well Transwell plates with or without Matrigel (BD), respectively. Then cells (1 × 10⁵) in serum-free culture medium were plated in the top chamber. High-glucose Dulbecco’s modified Eagle’s medium (DMEM) with 10% FBS was added to the bottom chamber. After 24-h incubation, the membrane was washed with PBS, fixed using 4% paraformaldehyde, and stained using 0.1% crystal violet solution. The cells were counted in five random fields under microscopy. We performed the experiment in triplicate and repeated it three times.

The Seahorse XFe 96 Extracellular Flux Analyzer (Seahorse Bioscience) was used to determine the extracellular acidification rate (ECAR). ECAR was examined with a Seahorse XF glycolysis stress test kit according to the manufacturer’s protocols. In brief, cells (1 × 10⁴ cells / well) were seeded into a Seahorse XF 96 cell culture plate. After baseline measurements, glucose, oligomycin, and 2-DG were sequentially injected into each well at the time points specified. ECAR data were assessed by Seahorse XF-96 Wave software and shown in mpH/ minute.

Lactate Assay Kit II (Sigma, Shanghai, China) and High Sensitivity Glucose Assay Kit (Sigma, Shanghai, China) were used according to the manufacturer’s instructions to detect HCC cell lactate and glucose levels, respectively.

After 24-h culture under hypoxia (1% O₂), the HCC cells were collected and incubated with 300 μl lysis buffer with protease inhibitors for 40 min on ice. Then, the supernatant was collected, and 2 μg HIF-1α, YAP, or immunoglobulin G (IgG) (Proteintech) antibody was added and incubated at 4 °C overnight. Next, 20 μl protein A/G-agarose beads (Santa Cruz Biotechnology, Shanghai, China) was added and rocked for 3 h at 4 °C. The pelleted cells were collected and washed three times with lysis buffer. Finally, the precipitate was boiled with 40 μl loading buffer for 5 min and analyzed by western blotting.

HCC tissues or HCC cells cytoplasmic and nuclear extracts were separated using a Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime, Shanghai, China) according to the manufacturer instructions. Then, YAP expression was analyzed by western blotting.

CHX chase assay was used to determine the half-life of HIF-1α. YAP knockdown or control HCC cells were individually seeded in 60-mm dishes for 24 h. Then, the cells were cultured under hypoxia (1% O₂) for 24 h and treated with CHX (100 μg/ml) for 0 h, 2 h, and 4 h. The cells were collected at the indicated time points, and HIF-1α expression was analyzed by western blotting.

Cells or cells transfected with YAP siRNA were cultured under hypoxic condition (1%O₂) for 24 h. Then ChIP assay was performed using the EZ-ChIP kit (17–371, Millipore, Billerica, MA, USA) according to the manufacturer’s instructions. The immunoprecipitated DNA was purified and analyzed by real-time PCR. Primer information used in ChIP assays was listed as follows:

The PKM2 promoter reporter plasmids were amplified from a human genomic DNA template and inserted into pGL3-Promoter Vector (Promega, Madison, WI). Cell were seeded in a 96-well plate and co-transfected with PKM2 promoter reporter plasmids (wt) or mutation of PKM2 promoter reporter plasmids and the 8xGTIIC-luciferase plasmid using Invitrogen Lipofectamine 3000 (Thermo Fisher Scientific, Shanghai, China) as per instructions from the manufacturer. Then cells were cultured under hypoxic condition (1%O₂) or nomoxic condition (20%O₂) for 24 h. Luciferase activities were analyzed using the Dual-luciferase reporter assay (Promega, Madison, WI, USA) according to the manufacturer’s instructions.

The HCC tissues data from TCGA databases were grouped into two groups based on the expression of HIF-1α: HIF-1α high expression and HIF-1α low expression. The gene expression values of the two groups samples were then put through GSEA v3.0 to analyze Hippo signaling genes signatures. Hippo signaling gene sets were obtained from the MSigDB database v6.1 [26].

We analyzed the data with SPSS 17.0 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 6 (GraphPad Software, San Diego, California, USA). The data are reported as the mean ± SD; Student’s t-test or analysis of variance was used for the statistical analyses. The correlation between YAP mRNA and HIF-1α mRNA was analyzed by Pearson’s correlation coefficient using GEPIA. p-values < 0.05 indicates statistical significance. We repeated the experiments three times at minimum.

We firstly investigated YAP expression in 54 cases of HCC tissues and their paired adjacent nontumor tissues by real-time PCR and immunohistochemistry. The results obtained from real-time PCR showed significantly increased YAP mRMA levels in the tumor tissues as compared with the adjacent noncancerous tissues (p = 0.006, Fig. 1a); Immunohistochemical staining revealed 53.7% (29/54) HCC tissues were positive for YAP expression, whereas only 18.5% (10/54) adjacent normal tissues were positive for YAP expression (Fig. 1b). Then immunofluorescence staining and cell fractionation assays were used to investigate YAP localization, among the 29 cases of HCC tissues overexpression YAP, 68.9% (20/29) HCC tissues showed stronger nuclear YAP staining as opposed to cytoplasmic staining (Fig. 1c and d). In addition, we also analyzed the mRNA levels of two canonical YAP transcription target genes (CTGF, CYR61) in the same HCCs displaying higher YAP mRNA levels. The results revealed that expression of CTGF mRNA and CYR61 mRNA were all increased in the HCC, along with the up-regulation of YAP mRNA (Additional file 2: Figure S1A). These results suggest that, in HCC tissues, YAP expression is high and that YAP is localized to the nucleus.

As a solid tumor, the abnormal new vasculature of the tumor and the increased consumption of oxygen in the cell proliferation in HCC are imbalanced, so the inner of HCC tissues is always hypoxic. Hypoxia promotes HCC invasion and migration, and hypoxia inducible factor-1α (HIF-1α) is also up-regulated in HCC. So in order to evaluate whether HIF-1α correlated with YAP, we firstly performed gene set enrichment analysis and found enriched expression of Hippo signaling genes in HCC tissues from The Cancer Genome Atlas (TCGA) database (Fig. 2a). Since YAP is a transcriptional activator in the Hippo signaling, and can also be inhibited by Hippo signaling [8], we then retrieved data on 369 HCC cases from The Cancer Genome Atlas (TCGA) to analyze the relationship between YAP mRNA and HIF-1α mRNA expression, and found a positive correlation between the two (R = 0.5, P = 1.7E-24) (Fig. 2b). Furthermore, we analyzed the relationship between the expression of YAP mRNA and HIF-1α mRNA in 54 cases of HCC tissues. It showed that YAP mRNA expression correlated positively with that of HIF-1α mRNA (R = 0.64, p < 0.01) (Fig. 2c). And western blotting yielded similar results both in the HCC tissues (Fig. 2d) and in the nuclear fraction of HCC (Additional file 2: Figure S1B). Our results suggest that YAP overexpression correlated strongly with HIF-1α in HCC tissues under hypoxia.

To determine whether hypoxia activates YAP, we incubated HCC cells under hypoxic and normoxic conditions for 24 h. The expression of p-YAP (Ser127) was significantly decreased under hypoxic conditions, while total YAP was not increased, and the protein levels of LATS1 and p-LATS1 (Ser909), the upstream regulators of YAP in the Hippo pathway, were decreased (Fig. 3a and b). It implied that Hippo signaling might be inhibited by hypoxia. As YAP nuclear localization is the key regulatory mechanism for activating YAP, we performed immunofluorescence staining to examine the cellular localization of YAP. Hypoxia triggered significant YAP nuclear translocation in the HCC cells (Fig. 3c and d). Consistent with that, cell fractionation showed more YAP protein accumulation in the nuclear fraction and less YAP protein in the cytoplasmic fraction of hypoxic cells (Fig. 3e and f). In addition, hypoxia also increased the mRNA expression of four canonical YAP transcription target genes (CTGF, CYR61, AREG and EDN1) (Fig. 3g and h). The results suggest that hypoxia activates YAP and induces YAP translocation to the nucleus by inhibiting the Hippo pathway in HCC cells.

Recent studies have reported that hypoxia promotes cancer cell glycolysis [8, 9]. To confirm this, the glucose and lactate assays were performed and the results showed that hypoxia significantly increased HCC cell glucose uptake and lactate production rates, respectively, compared with normoxia (Fig. 4a and b). Hypoxia also showed increased extracellular acidification rate (ECAR), which reflects overall glycolytic flux (Fig. 4c and d). Usually, hypoxia promotes cancer cell glycolysis by up-regulating the expression of key glycolysis enzymes [27, 28]. So we then performed real-time PCR to detect the expression of key glycolysis enzymes in HepG2 and Huh7 cells under 24 h normoxia or hypoxia. LDHA mRNA, GLUT1 mRNA, PGK1 mRNA, and PKM2 mRNA expression levels were up-regulated in the hypoxic cells than in the normoxic cells, and PKM2 mRNA had a significantly increased among the key glycolysis enzymes (Fig. 4e), western blotting yielded similar results (Fig. 4f). Besides that, as a potent micro-environmental factor, hypoxia also promotes tumor migration via up-regulation the expression of matrix metallo proteinases 2 and 9 (MMP2 and MMP9) [29‐32]. We then examined HCC cell invasion and migration under normoxia and hypoxia using Transwell assays. Hypoxia significantly promoted HCC cell migration and invasion when compared with normoxia (Fig. 4g and h). Moreover, the hypoxic cells had higher MMP2 and MMP9 expression than the normoxic cells (Fig. 4i and j). These results suggest that hypoxia contributes to HCC cell glycolysis, invasion, and migration.

To investigate the role of YAP in hypoxic HCC cells, we knocked down YAP expression in HCC cells using YAP siRNA. The knockdown efficiency was examined using real-time PCR and western blotting (Fig. 5a and b). Silencing YAP under hypoxia significantly decreased the HCC cell glucose uptake and lactate production rates compared with control cells (Fig. 5c and d). In addition, silencing YAP also decreased extracellular acidification rate (ECAR) (Fig. 5e and f) in HCC cells under hypoxic condition. These results indicated that YAP knockdown might inhibit HCC cell glycolysis under hypoxia. To further confirm the hypothesis, we examined the expression of PKM2 and activated YAP in YAP knockdown cells under hypoxia. It showed that YAP knockdown also decreased the expression of PKM2 and HIF-1α protein (Fig. 5g and h). However, it did not significant change the level of HIF-1α mRNA, although it decreased the PKM2 mRNA level (Additional file 3: Figure S2A and B). Moreover, activating YAP restored the expression of HIF-1α protein and PKM2 mRNA (Additional file 3: Figure S2C and E) and HCC cell glycolysis under hypoxia (Additional file 3: Figure S2F and G). Furthermore, silencing YAP significantly inhibited the invasion and migration of hypoxic HCC cells (Fig. 5i and j) and decreased MMP2 and MMP9 protein expression in hypoxic cells (Fig. 5k and l). Our findings indicate that YAP can mediate hypoxia-induced glycolysis in HCC cells.

To investigate the molecular mechanism of YAP in contributing to hypoxia-induced glycolysis in HCC cells, the cellular localization of YAP and HIF-1α was examined using immunofluorescence staining. Hypoxia greatly promoted YAP and HIF-1α accumulation in the nucleus (Fig. 6a and b). As hypoxia triggered both YAP and HIF-1α nuclear translocation, we believe that YAP may complex with HIF-1α in the nucleus under hypoxia. Co-immunoprecipitation confirmed that YAP could bind HIF-1α under hypoxic conditions (Fig. 6c). However, silencing YAP significantly decreased HIF-1α protein expression under hypoxia and did not change the level of HIF-1α mRNA, so we wondered whether YAP could sustain HIF-1α stabilization. We used CHX chase assays to explore the effect of YAP on HIF-1α stability, and found that YAP knockdown cells had significantly shortened HIF-1α half-life compared with the control cells (Fig. 6d). The results indicated that YAP bound HIF-1α to form a complex and sustained HIF-1α stability. As shown in Figs. 4 and 5, we found the expression of PKM2 was significant changed, so we further wondered whether YAP/HIF-1a complexes could bind target PKM2 gene to promote HCC cells glycolysis in hypoxia. Since YAP has no DNA-binding domains, and binds to its target genes promoters by interacting with DNA-binding transcription factors [33, 34], so we focused on HIF-1a and analyzed HIF-1a and PKM2 gene sequence. Finally, we found a candidate HRE (HIF-1a binding site 5’-ACGTG-3′) in PKM2 promoter. To determine whether YAP/HIF-1a complexes bind at this site, chromatin immunoprecipitation (ChIP) assay was performed in HCC cells under hypoxia. The results showed that ChIP with HIF-1a or YAP antibody enriched the putative HRE sequence > 5 fold compared to IP with IgG (Fig. 6e). In contrast, after silencing YAP, the enrichment of PKM2 with HIF-1a or YAP antibody was significantly decreased compared with IgG (Fig. 6f). These data indicated that YAP/HIF-1a complexes colud bind to the promoter of PKM2 gene. In addition, we also used luciferase reporter assay to test whether this putative HRE in the PKM2 gene is functional. As shown in Fig. 6g, this putative HRE in the PKM2 gene significantly increased luciferase activity in hypoxic HCC cells, however, mutation HRE significantly decreased hypoxia-induced luciferase activity. Taken together, these data indicated that YAP/HIF-1a complexes bound to PKM2 gene promoter and directly activated its transcription to accelerate glycolysis in hypoxia.