FLIP_L is critical for aerobic glycolysis in hepatocellular carcinoma

HepG2, MHCC97-H, Huh-7, SMMC-7721, BEL-7704 HCC, and HEK293 cell lines were obtained from the Department of Biochemistry and Molecular Biology (The Fourth Military Medical University) and cultured in a humidified incubator under 5 % CO₂ in Dulbecco’s modified Eagle’s medium (DMEM) (Life Technologies, Carlsbad, CA, USA) supplemented with 10 % fetal bovine serum (FBS), 2 mM L-glutamine, 100 U/ml penicillin, and 100 mg/ml streptomycin at 37 °C. The proteasome inhibitor, MG132, was from Sigma-Aldrich (Shanghai, China). Mouse anti-GLUT1, rabbit anti-SGLT1 and mouse anti-FLIP antibodies were obtained from Abcam (New Territories, HK). Rabbit anti-FLIP and mouse anti-HA-Tag antibodies were from Cell Signaling Technology (Danvers, MA). 2-NBDG (2-[N-(7-Nitrobenz-2-oxa-1,3-diazol-4-yl)amino]-2-deoxy -D-glucose) was purchased from Cayman Chemical Company (Michigan, US).

FLIP_L expression plasmid (Flag-FLIP_L), FLIP_S expression plasmid (Flag-FLIP_S), SGLT1 expression plasmid (Myc-SGLT1), HA-ubiquitin expression plasmid, shRNA targeting FLIP_L, and shRNA targeting SGLT1 were constructed by Jikai biotechnology (Shanghai, China). Cells (1 × 10⁶) were seeded into six-well plates and transfected using Lipofectamine 2000 (Invitrogen Life Technologies, Carlsbad, USA) according to the manufacturer’s instructions. The target sequences of shRNA of FLIP_L and SGLT1 are 5’ CAGAATAGACCTGAAGACA3’ and 5’ CTTCCGCATCCAGGT CAAT3’, respectively.

RNA was reverse transcribed with the PrimeScript™ 1st Strand cDNA Synthesis Kit (Takara, Dalian, China) according to the manufacturer's instructions. Real-time PCR was performed using SYBR Premix Ex Taq™ II (Takara, Dalian, China). The relative gene expression was calculated using the 2^−ΔΔCt method, in which Ct represented the threshold cycle. β-actin RNA was used as the reference. The primers were used as following. FLIP_L: Forward 5’GGCTCCCCCTGCATCAC3’, Reverse 5’ TTTGGCTT CCTGCTAGATAAGG3’; SGLT1: Forward 5’ATGGACAGTAGCACCTTGAGCC3’; Reverse 5’TAGCCCCAGAGAAGATGTCTGC3’; GLUT1: Forward 5’ AAGTCTCC TTTACCCACATCC 3’; Reverse 5’ GAGTGTCCGTGTCTTCTTGAGT 3’.

Cells were harvested and lysed with RIPA buffer (0.15 M NaCl, 1 % NP40, 0.01 M deoxycholate, 0.1 % SDS, 0.05 M Tris-HCl pH 8.0, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 10 mg/ml aprotinin, 10 mg/ml pepstatin and 10 mg/ml leupeptin). Protein concentrations were determined using the BCA Protein Assay kit (Beyotime, China). The total protein was electrophoresed on 8 or 12 % SDS-PAGE gels and transferred to polyvinylidene fluoride membranes. Membranes were blocked with 5 % nonfat dry milk or BSA in TBST buffer (20 mM Tris-HCl pH 7.4, 8 g/L NaCl and 0.1 % Tween 20) for 1 h at room temperature, and then incubated with the indicated primary antibody in TBST buffer containing 2 % BSA at 4 °C overnight. The membrane was then probed with the appropriate horseradish peroxidase-conjugated secondary antibody at room temperature for 1 h, and the proteins were detected with a chemiluminescence imaging analysis system.

Cells were plated at 1×10⁶/well in 6-well plates and cultured for 24 h. After the specified treatments, cells were refreshed with serum-starved (0.1 % FBS) and glucose-free DMEM. 16 h later, cells were grown in the presence of 50 mM 2-NBDG for 30 or 60 min, respectively, and glucose uptake was quantified using FACS analysis.

5×10⁶ of HepG2 cells were injected into the hind legs of athymic male nude mice (8 weeks, seven mice per group). The tumor size was measured weekly with a caliper, and the tumor volume was determined with the standard formula: 0.5×length×width². The tumor growth curve was then derived from these data. ¹⁸F-FDG was injected into the mice via a tail vein and PET/CT images were collected on a nanoScan PET/CT system (Mediso, Budapest, Hungary).

1×10⁵ cells were plated in 6-well plates and cultured overnight. Culture medium was removed from cells and the lactate concentration was determined using the Lactate Kit (Nangjing Jiancheng Bioengineering, Nanjing, China) according to manufacturer’s protocol. Cells were then harvested, stained with trypan blue, and the number of viable cells was counted using a hemocytometer. The rate of lactate production was determined (lactate production rate=lactate concentration/cells/time) and normalized to the rate detected in the control group.

Extracellular acidification rate (ECAR) was conducted using a Seahorse XF24 Analyzer. 20,000 cells were seeded into Seahorse 24-well microplates and cultured at 37 °C with 5 % CO₂ for 24 h. Culture medium was replaced with 700 μl of assay medium, composed of DMEM without FBS and sodium bicarbonate, and cells were then incubated at 37 °C without CO₂ for 1 h. To measure the glycolytic capacity of the cells, additions were injected to the wells at final concentrations of 11 mM glucose, 1 μM oligomycin, and 20 mM 2-DG.

Cells were fixed in a freshly prepared 4 % paraformaldehyde, rinsed with 0.1 % PBS and permeabilized with Triton X-100. Cells were then incubated with horse serum in PBS to block nonspecific binding. After being washed with PBS, cells were incubated overnight at 4 °C with mouse anti-FLIP antibody (1:200), rabbit anti-SGLT1 antibody (1:200). Subsequently, fluorescein isothiocyanate-conjugated anti-mouse antibody (diluted 1:400; Sigma-Aldrich) or cyanine 3-conjugated anti-rabbit antibody (diluted 1:400; Sigma-Aldrich) were incubated with the cells at room temperature for 2 h. The isotype mouse and rabbit IgG were used as negative controls. Fluorescence detection was performed using a laser confocal microscope (Nikon, US) after cell nucleus counterstaining with 4’, 6-diamidino-2-phenylindole.

Cells were incubated with 1 ml of lysis buffer containing (50 mM HEPES, pH 7.4; 150 mM NaCl; 0.1 % Triton X-100; 1.5 mM MgCl₂; 1 mM EDTA; and 1 mM phenylmethylsulfonyl fluoride), as well as protease inhibitors for 30 min at 4 °C. The insoluble fraction was eliminated through centrifugation at 12,000 rpm for 30 min at 4 °C. The lysates were incubated with the antibody after centrifugation and protein A/G-conjugated sepharose (Santa Cruz Bio Technology, USA) at 4 °C overnight. 20 μl of protein was saved to quantify the total amount of protein loaded. Beads were washed four times with lysis buffer. Proteins were eluted in SDS-PAGE sample buffer and separated by SDS-PAGE for immunoblot analysis.

Cells were seeded into 96-well cell culture plates. Each well contained 1×10³ cells in 100 μl culture media. After the specified treatments, the medium in each well was replaced with 100 μl fresh medium with 10 % CCK8 (Dojindo, Japan), and the cells incubated at 37 °C for additional 2 h. The absorbance was measured at 450 nm wavelength.

Cells were cultured in MEM containing low glucose (0.75 mM) or high glucose (25 mM) for 48 h after cell transfection. For Annexin staining, cells were trypsinized and resuspended in 1 ml Annexin-binding buffer (BD Biosciences, US) to which was added 5 μl Annexin V-FITC (BD Biosciences, US). After incubation in the dark at room temperature for 15 min, 50 μl PI (Sigma, US) was added to discriminate dead cells and the samples were analyzed on a FACS Caliber flow cytometer.

All patients from whom we obtained the 79 pairs of HCC and adjacent normal liver tissue specimens provided their full consent to participate in the study at the Xijing Hospital of the Fourth Military Medical University (Xi’an, China). IHC was performed using the avidin-biotin-peroxidase complex method. All sections were deparaffinized in xylenes and dehydrated through a gradient concentration of alcohol before endogenous peroxidase activity was blocked using 0.5 % H₂O₂ in methanol for 10 min. After nonspecific binding was blocked, the slides were incubated with FLIP_L antibody (1:200) or SGLT1 antibody (1:200) in phosphate-buffered saline (PBS) at 4 °C overnight in a humidified container. Biotinylated goat anti-rabbit immunoglobulin G (IgG) (1:400; Sigma-Aldrich) was incubated with the sections for 1 h at room temperature and detected using a streptavidin-peroxidase complex. The brown color indicative of peroxidase activity was developed by incubation with 0.1 % 3, 3-diaminobenzidine in PBS with 0.05 % H₂O₂ for 5 min at room temperature. An immunoreactivity score system was provided on the basis of the proportion and intensity of positively stained cancer cells. The extensional standard used was as follows: (1) the number of cells with positive staining (0–5 %: 0; 6–25 %: 1; 26–50 %: 2; 51–75 %: 3; and>75 %: 4) and (2) the staining intensity (colorless: 0; pallide-flavens: 1; yellow: 2; brown: 3). We multiplied the numbers scored (1) and (2) from the extensional standard, and the staining grade was stratified as absent (0 score), weak (1–4 score), moderate (5–8 score) or strong (9–12 score). Two pathologists blinded to the clinicopathologic information and outcome of the patients scored the multiple tissue arrays. Tumors with moderate or strong immunostaining intensity were classified as staining positive (+), whereas tumors with absent were classified as staining negative (−).

Data are reported as the mean ± S.D. or S.E.M. analyses were performed using GraphPad Prism software. The significance of differences was determined by one-way analysis of variance followed by Scheffe’s post hoc test, independent samples t-test or Student’s t-test. Associations between protein expression and categorical variables were analyzed using the χ² test or Fisher’s exact test as appropriate. P < 0.05 was considered statistically significant.

To investigate whether FLIP could affect aerobic glycolysis in HCC, FLIP_L and FLIP_S, the two main isoforms of FLIP, were overexpressed in HepG2 cells (Fig. 1a, b). Overexpressed FLIP_L significantly increased aerobic glycolytic capacity in response to glucose stimulation compared with the control group, but overexpressed FLIP_S had no effect on glycolysis (Fig. 1c, d). We also found that FLIP_L, but not FLIP_S, induced lactate production (Fig. 1e), which was consistent with the ECAR results. These findings suggest that FLIP_L is involved in aerobic glycolysis in HCC.

Glucose uptake is a key step in aerobic glycolysis, which is critical for cancer cell survival. Most cancer cells generate high levels of ATP through increasing the glucose uptake [23]. Therefore, we examined the effect of FLIP on glucose uptake. Ectopic expression of FLIP_L significantly induced glucose uptake in HepG2 cells, but glucose uptake was not altered in cells overexpressing FLIP_S (Fig. 1f). Moreover, to evaluate the influence of FLIP_L on glucose uptake and HCC progression in vivo, nude mice bearing FLIP_L down-regulated HepG2 cells (Fig. 1g, h) were imaged by ¹⁸F-FDG PET/CT scanning. The scan images demonstrated that ¹⁸F-FDG accumulation was notably decreased with the deletion of FLIP_L (Fig. 1i, j). We also analyzed the effect of FLIP_L on HCC development by continuously measuring the size of tumors in the xenograft mouse model. As shown in Fig. 1k, the growth of transplanted tumors with FLIP_L silenced was significantly slower, compared with the control group. These results indicated that FLIP_L promotes tumor cell survival and aerobic glycolysis by increasing glucose uptake both in vivo and in vitro.

Glucose is transported into cells by two families of transporters, the facilitative-type glucose transporters and the active-type glucose transporters [24]. GLUT1 and SGLT1, two well-studied glucose transporters in the respective families, are highly expressed in several cancers and participate in cellular glucose uptake. To determinate the mechanisms by which FLIP_L promotes the aerobic glycolysis and survival of HCC cells, we measured the GLUT1 and SGLT1 mRNA and protein levels after shRNA-mediated depletion of FLIP_L expression in HepG2 cells. Although the mRNA level of SGLT1 did not change, its protein expression was notably decreased following FLIP_L loss. However, neither mRNA nor protein levels of GLUT1 altered with the down-regulation of FLIP_L (Fig. 2a, b). These findings suggested that SGLT1, but not GLUT1, is involved in FLIP_L-mediated HCC glucose uptake.

Dysregulated expression of SGLT1 has been identified in various human cancers. However, the expression status of SGLT1 in HCC has not been investigated. To assess SGLT1 protein expression levels and its potential association with FLIP_L in HCC, we compared 79 pairs of HCC and adjacent normal tissue samples using immunohistochemistry (IHC). The results showed that SGLT1 and FLIP_L were highly expressed in liver cancer tissue compared with adjacent normal liver tissue (Fig. 2c). In normal versus liver cancer tissue, FLIP_L levels were 0.41 versus 2.97 (Fig. 2d), and SGLT1 were 0.28 versus 2.47 (Fig. 2e). Moreover, in these 79 pairs of HCC and adjacent normal tissue samples, SGLT1 and FLIP_L expression were correlated with tumor stage (Tables 1 and 2) in liver cancer tissues, but the correlation between the expression and sex or age was not statistically significant in both liver cancer tissues and their adjacent normal specimens (Tables 1 and 2, Additional file 1: Table S1 and Table S2).

Because knocking down FLIP_L expression led to the suppression of SGLT1 protein levels, we speculated whether there is a correlation between the protein levels of FLIP_L and SGLT1. We analyzed the correlation between FLIP_L and SGLT1 expression using spearman analysis. The expression of FLIP_L was demonstrated to positively correlate with SGLT1 expression (Fig. 2f), suggesting that FLIP_L is involved in the post-transcriptional regulation of SGLT1.

Given that FLIP_L could regulate the protein level of SGLT1, but not the SGLT1 transcript, we investigated whether the FLIP_L protein interacts with the SGLT1 protein. To address this question, we examined the subcellular distribution of FLIP_L and SGLT1 in HepG2 cells using immunofluorescence. Both FLIP_L and SGLT1 were predominantly localized in the cytoplasm and there was a strong co-localization of the two proteins (Fig. 3a). The co-localization of FLIP_L and SGLT1 prompted us to further analyze whether FLIP_L could bind with SGLT1. Co-immunoprecipitation was performed using lysates prepared from HepG2 cells, and showed that endogenous FLIP_L interacted with endogenous SGLT1, and vice versa (Fig. 3b).

FLIP_L has been shown to up-regulate the levels of some proteins by inhibiting the ubiquitin-proteasome degradation pathway [19]. We hypothesized that FLIP_L stabilizes SGLT1 protein by regulating its degradation. First we determined the protein degradation pathway of SGLT1. As shown in Fig. 3c, FLIP_L increased the SGLT1 protein level and MG132, a proteasome inhibitor, induced further accumulation of SGLT1 protein, indicating that the degradation of SGLT1 protein occurs via a proteasome associated pathway. In addition, we examined the effect of FLIP_L on SGLT1 ubiquitination in HEK293 cells and HepG2 cells. In HEK293 cells, exogenous FLIP_L inhibited the ubiquitination of exogenous SGLT1 and increased the total levels of SGLT1 protein (Fig. 3d). Endogenous SGLT1 ubiquitination was also examined in HepG2 cells and, consistent with the results with the exogenous proteins in HEK293 cells, we found that down-regulation of the expression of FLIP_L led to increased SGLT1 ubiquitination and accelerated SGLT1 degradation (Fig. 3e). Collectively, these results suggested that FLIP_L interacts with SGLT1 in the cytoplasm and increases the stability of SGLT1 protein by inhibiting its ubiquitination and proteasome-mediated degradation. This promotes the membrane translocation and function of SGLT1.

High level of glycolysis could not only rapidly provide ATP for cancer cell proliferation but also allow cancer cells to adapt low glucose condition [9]. We investigated whether and how FLIP_L regulates the tolerance for low glucose of HCC cells. We first assessed the ability of several HCC cell lines to tolerate low glucose. The apoptosis levels of HepG2, MHCC97-H, Huh-7, SMMC-7721 and BEL-7704 were detected under high glucose condition (25 mM) and low glucose condition (0.75 mM). Of these cell lines, HepG2, MHCC97-H, and Huh-7 cells exhibited higher tolerance of the low glucose condition (Fig. 4a). We next examined the FLIP_L and SGLT1 expressions in these cell lines. FLIP_L and SGLT1 protein levels were higher in two of the cell lines tolerant to low glucose, HepG2 and Huh-7 cells (Fig. 4b), so we used the HepG2 cells to perform the following experiments.

To determine whether FLIP_L and SGLT1 affect the susceptibility of HCC cells to low glucose conditions, FLIP_L and SGLT1 were separately down-regulated in HepG2 cells (Figs. 1g, h, and 4c, d). Examination of cell apoptosis, as shown in Fig. 4e, f, showed that apoptosis rates were significantly increased in FLIP_L-silenced or SGLT1-silenced HepG2 cells. In addition, we found that the loss of FLIP_L inhibited glucose uptake in HepG2 cells (Fig. 4g), which is consistent with the results of FLIP_L overexpression in Fig. 1f. We next rescued the SGLT1 expression in FLIP_L-deficient HepG2 cells (Fig. 4h, i), and the reduced glucose uptake was substantially recovered (Fig. 4g). The increased apoptosis rate was also attenuated with the restoration of SGLT1 expression in FLIP_L-deficient HepG2 cells (Fig. 4j). In addition, we examined the survival of HepG2 cells in a low glucose environment. FLIP_L-deficient HepG2 cells exhibited a decreased cell proliferation rate, and rescued SGLT1 expression significantly recovered cell survival under low glucose treatment (Fig. 4k). These results demonstrated that FLIP_L promotes glucose uptake and the survival of HCC cells under low glucose conditions through regulation of SGLT1.