The first two authors contributed equally to this work.
In vitro and in vivo study of phloretin-induced apoptosis in human liver cancer cells involving inhibition of type II glucose transporter
Article first published online: 26 NOV 2008
Copyright © 2009 Wiley-Liss, Inc.
International Journal of Cancer
Volume 124, Issue 9, pages 2210–2219, 1 May 2009
How to Cite
Wu, C.-H., Ho, Y.-S., Tsai, C.-Y., Wang, Y.-J., Tseng, H., Wei, P.-L., Lee, C.-H., Liu, R.-S. and Lin, S.-Y. (2009), In vitro and in vivo study of phloretin-induced apoptosis in human liver cancer cells involving inhibition of type II glucose transporter. Int. J. Cancer, 124: 2210–2219. doi: 10.1002/ijc.24189
- Issue published online: 24 FEB 2009
- Article first published online: 26 NOV 2008
- Manuscript Accepted: 19 NOV 2008
- Manuscript Received: 19 MAY 2008
- The National Science Council of Taiwan. Grant Numbers: NSC 95-2320-B-038-011, NSC 95-2314-B-038-006, NSC 96-2317-B-038-001
- glucose transporter;
- HepG2 cell;
- hepatocellular carcinoma
Phloretin (Ph), a natural product found in apples and pears with glucose transporter (GLUT) inhibitory activity, exerts antitumor effects. However, little is known about its effects on human liver cancer. The purpose of this study is to test the cytotoxic effects of Ph on HepG2 cells and to identify the underlying molecular pathways. Human hepatocellular carcinoma specimens and HepG2 show a high level of GLUT2 transporter activity in the cell membrane. Real-time PCR and MTT assays demonstrate that Ph-induced cytotoxicity correlates with the expression of GLUT2. Flow cytometry and DNA fragmentation studies show that 200 μM Ph induces apoptosis in HepG2, which was reversed by glucose pretreatment. GLUT2 siRNA knockdown induced HepG2 apoptosis, which was not reversed by glucose. Western blot analysis demonstrates that both intrinsic and extrinsic apoptotic pathways in addition to Akt and Bcl-2 family signaling pathways are involved in Ph-induced cell death in HepG2 cells. Furthermore, using flow cytometry analysis, a mitochondrial membrane potential assay and Western blot analysis, we show that cytochalasin B, a glucose transport inhibitor, enhances the Ph-induced apoptotic effect on HepG2 cells, which was reversed by pretreatment with glucose. Furthermore, we found significant antitumor effects in vivo by administering Ph at 10 mg/kg intraperitoneally to severe combined immune deficiency mice carrying a HepG2 xenograft. A microPET study in the HepG2 tumor-bearing mice showed a 10-fold decrease in 18F-FDG uptake in Ph-treated tumors compared to controls. Taken together, these results suggest that Ph-induced apoptosis in HepG2 cells involves inhibition of GLUT2 glucose transport mechanisms. © 2008 Wiley-Liss, Inc.
Consumption of apples has been associated with prevention of cancer and chronic diseases.1, 2 The phytochemicals found in apples have potent antioxidant activity and antiproliferative effects on cancer cells.3–5 Ph is a dihydrochalcone found in apples and other fruits,6 which is known to suppress both transmembrane glucose transport7 and inhibition of protein kinase C.8 Ph has been shown to inhibit human leukemia cell growth9 and to induce apoptosis of melanoma cells by suppressing transmembrane glucose transport.6 Ph also inhibits the growth of bladder cancer and rat mammary adenocarcinoma cells in vivo.10 These in vitro and in vivo findings suggest that Ph has antitumor activity. However, little is known about the effects of Ph on human hepatoma cells.
Facilitated glucose transporters (GLUTs) mediate the energy-independent transport of glucose across the plasma membrane.11 Thirteen GLUTs have been identified and classified into 3 groups according to their sequence similarities and characteristics.12 The GLUTs have tissue-specific patterns of expression reflecting the various glucose requirements of different tissues.13 GLUT2, for example, is distributed in liver cells, pancreatic islet β-cells, intestinal epithelial cells and kidney proximal renal tubule cells.14, 15 Malignant cells have been shown to express high levels of glucose transporters in connection with an enhanced rate of glycolysis.16 Increased glucose uptake in hepatocellular carcinoma (HCC) compared with adjacent nontumoral liver tissue has been documented,17–19 and previous reports demonstrated that GLUT2 mRNA expression was increased in human liver tumors compared with nontumorous liver tissue,20 while no expression of GLUT1 was detected in normal human liver tissue, adenoma, or HCC.17, 21 These findings suggest that GLUT2 expression is important in HCC tumorigenesis.
Recent reports suggest that the suppression of glucose transport is linked to the earliest steps of apoptosis.22 Glucose deprivation may induce 3 major pathways, including the ATP depletion-induced mitochondrial death pathway, the oxidative stress-related cell death pathway, and induction of hypoxia-inducible factor-1α with concomitant activation of the p53-induced cell death pathway.22 It has been suggested that Ph is a specific GLUT2 inhibitor.23, 24 In the present study, we demonstrate that human HCC and hepatoma cells express high levels of GLUT2. Treatment with Ph induces apoptosis in HepG2 cells which depends on GLUT2 expression and suppresses liver tumor growth in a xenograft mouse model. Both of these effects involve suppression of glucose transport.
Material and methods
Cell culture, chemicals and patient samples
The malignant human cell line HepG2 (HB-8065, American Type Culture Collection) was derived from a human hepatoma. The cells were grown in Eagle's minimal essential medium supplemented with 10% fetal calf serum (FCS), penicillin (100 U/ml), streptomycin (100 μM/ml) and 0.3 mg/ml glutamine in a humidified incubator (37°C, 5% CO2). Ph (Sigma Chemical, St. Louis, MO) dissolved in 0.05% dimethylsulfoxide (DMSO) was added at the doses indicated. For control specimens, the same volume of 0.05% DMSO without Ph was added.
The cells used for determining GLUT2 expression were as follows: CA9-22, oral cancer; FHC, normal colon cell; A431, epidermoid carcinoma; Hep3B, hepatocellular carcinoma; MCF-10A, mammary gland epithelial fibrocystic cell; LNCAP, prostate carcinoma; COLO 205, colorectal adenocarcinoma; AU565, breast adenocarcinoma; CL (Calu-1), lung epidermoid carcinoma; RAW, macrophage leukemia; HT 29, colorectal adenocarcinoma; NIH3T3, mouse embryo fibroblast; COLO-DLD, colorectal adenocarcinoma; MCF-7, breast adenocarcinoma; PC3, prostate adenocarcinoma; HBL 100, breast-derived nonmalignant cell; A549, lung carcinoma; KA, keratoacanthoma; MB231, breast adenocarcinoma; HL60, promyeloblastic leukemia; MCEC, murine cardiac endothelial cell; U373, human astrocytoma. The cells were grown and routinely maintained in the following media: minimal essential medium (for Hep3B, MCF-7, CL, KA), Dulbecco's modified Eagle's medium (DMEM) (for CA9-22, A431, RAW, NIH3T3, A549, MCEC, U373), DMEM-F12 (for FHC, MCF-10A, MB231), and RPMI-1640 (for LNCAP, COLO 205, HT 29, AU565, COLO-DLD, PC3, HBL 100, HL 60). In all cases, the medium was supplemented with 10% FCS, 2 mM L-glutamine, 100 units/ml penicillin and 100 mg/ml streptomycin. Cells wee incubated at 37°C in a 5% CO2 atmosphere incubator.
Human HCC samples (n = 14) were obtained as anonymous specimens from Taipei Medical University Hospital through the protocol approved by the Institutional Review Board (P950023) and were embedded in paraffin for immunohistochemistry. There are 12 male patients with mean age of 56.4 years (range, 39–77 years); 2 female patients with mean age of 57.5 years (53 and 62 years). Among the 14 patients, 9 of them are hepatitis B carrier; 2 of them are hepatitis C carrier; 1 patient has simultaneous hepatitis B and C infection; 2 of them are free of hepatitis B and C infection. Two of the patients also present history of heavy alcohol drinking, one is hepatitis B and the other is hepatitis C carrier.
Human HepG2 cells were immunostained for 2 hr with 1:100 dilutions of monoclonal antiGLUT2 or anticaveolin antibodies (Santa Cruz Biotechnology, CA). After washing with PBS/Tween 20 three times, the monoclonal antibody was visualized with fluorescent-conjugated secondary antibodies (FITC and rhodamine) diluted 1:50 in blocking buffer for 1 hr. Coverslips were washed, mounted, and viewed under a Leica TCS SP5 confocal laser-scanning microscope (Leica Microsystems, Heidelberg, Germany).
As described previously,25 paraffin-embedded blocks were sectioned at 5–7-μm thickness. After microwave pretreatment in citrate buffer (pH 6.0) for antigen retrieval, slides were immersed in 0.3% hydrogen peroxide for 20 min to block the endogenous peroxidase activity. After intensive washing with PBS, slides were incubated overnight at 4°C with the anti-GLUT2 antibody at a dilution of 1:100. After incubation with a biotinylated secondary antibody (1:200), slides were incubated with peroxidase-conjugated streptavidin (DAKO LSAB+kit; Dako, Carpinteria, CA). Reaction products were visualized by immersing slides in diaminobenzidine tetrachloride, and finally were counterstained with hematoxylin. As for the negative controls, sections were stained as above but with PBS substituted for the primary antibody.
mRNA detection using real-time PCR
Total RNA was extracted from cells using phenol chloroform extraction (TRIzol Reagent, Invitrogen, Carlsbad, CA). Primers for GLUT2 were the following: forward, 5′-AGTTAGATGAGGAAGTCAAAGCAA-3′; reverse, 5′-TAGGCTGTCGGTAGCTGG-3′. These primers yielded an amplicon of 164 bp. Primers for β-glucuronidase (GUS) (forward 5′-AGTGTTCCCTGCTAGAATAGATG-3′ and reverse 5′-AAACAGCCCGTTTACTTGAG-3′) were used to establish baseline mRNA content. A LightCycler thermocycler was used to conduct real-time PCR (Roche Molecular Biochemicals, Mannheim, Germany). Reaction reagents were purchased in a preformatted kit (LightCycler FastStart DNA Master SYBR Green I; Roche Diagnostics GmbH, Roche Molecular Biochemicals, Mannheim, Germany) containing 10× concentrations of Taq DNA polymerase, deoxynucleoside triphosphates, 10 mM MgCl2, SYBR Green I and reaction buffer. Cycling conditions were as follows: denaturation at 95°C for 10 min, followed by 40 cycles with denaturation at 95°C for 5 sec, primer annealing at 60°C for 5 sec, and a single fluorescence acquisition step during amplicon extension at 72°C for 10 sec. Data were analyzed by Roche LightCycler Software version 4. DNA amount was determined by the standard curve method with background subtraction. A cycle of melting curve analysis for each PCR product was performed to detect primer dimers. Results are expressed as copy number of GLUT2 per μg of mRNA loaded.
3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide assay
The 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay was performed as described previously.26 Cells were seeded in a 96-well plate at a density of 1 × 104 cells/well and allowed to adhere overnight. After removing the medium, 200 μl of fresh medium containing 10 mmol/l HEPES (pH 7.4) was added to each well. Then, 50 μl of MTT was added to the wells, and the plate was incubated for 2–4 hr at 37°C in the dark. The medium was removed, and 200 μl of DMSO and 25 μl of Sorensens's glycine buffer were added to the wells. Absorbance was measured using an ELISA plate reader at 570 nm with background subtraction at 650 nm.
Flow cytometry analysis
The cell cycle stages in the Ph- and DMSO-treated groups were measured by flow cytometry as previously described.26 Cells were harvested and stained with a reagent containing propidium iodide (50 μg/ml) and DNase-free RNase (2 U/ml). Measurement of nuclear DNA was performed using fluorescence-activated cell sorting (FACS). The population of nuclei in each phase of the cell cycle was determined using CellFIT DNA analysis software (Becton-Dickinson, San Jose, CA), and 15,000 events were analyzed for each sample.
Analysis of DNA fragmentation
Analysis of DNA fragmentation was performed as previously described.26 Briefly, Ph- and DMSO-treated cells were seeded in 100-mm dishes. The DNA was extracted twice with equal volumes of phenol and once with chloroform-isoamyl alcohol (24:1, v:v), precipitated with 0.1 volume of sodium acetate (pH 4.8) and 2.5 volumes of ethanol at −20°C overnight, and finally was centrifuged at 13,000g for 1 hr. Genomic DNA was quantified, and equal amounts of DNA sample per lane were electrophoresed in a 2% agarose gel. The DNA was visualized by ethidium bromide staining.
Plasmids and transfection assays
We ablated GLUT2 expression in liver cancer cells with at least 2 independent small interfering RNAs (siRNAs). The target sequences of GLUT2 siRNA (5′-ACCAATTCCAGCTACCGAC-3′, 5′-GTCGGTAGCTGGAATTGGT-3′) were selected to suppress GLUT2 gene expression and GLUT2 scramble siRNA (scRNA) (5′-CAAGCGCAACTACCTATCC-3′, 5′-GGATAGGTAGTTGCGCTTG-3′) were used as controls. After BLAST analysis to verify that there was no significant sequence homology with other human genes, the selected sequence was inserted into a BglII/HindIII-cut pSUPER vector to generate the pSUPER-glut2si and pSUPER-glut2sc vectors. All constructs were confirmed by DNA sequence analysis. The transfection protocol has been described previously.27 Briefly, 5 × 106 cells were washed twice with PBS and mixed with 12 μg plasmid. We applied 1 pulse of 30 msec under a fixed voltage of 1.1 kV for HepG2 and 1 pulse of 40 msec under a fixed voltage of 1 kV for Hep3B cells on a pipette-type microporator MP-100 (Digital Bio, Seoul, Korea).
Western blot analysis
Western blot analysis was performed as described previously.26, 28 Monoclonal or polyclonal antibodies were obtained from various sources as indicated: anti-p53, anti-caspase 8, anti-Bax, anti-Bcl-2, anti-cytochrome c, anti-Bid, anti-Bad, anti-PARP, anti-AIF, anti-AKT, anti-survivin and anti-XIAP antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA); anti-caspase 9 and anti-caspase 3 antibodies from Stressgen Biotechnologies (Victoria, British Columbia, Canada); and anti-p-AKT from Cell Signaling Technology (Beverly, MA,). Briefly, cell lysates were prepared, electrotransferred, immunoblotted with antibodies, developed using the SuperSignal chemiluminescence kit as indicated by the manufacturer (Pierce Biotechnology, Rockford, IL) and visualized by autoradiography. Expression of GAPDH (Abcam, Cambridge, UK) was used as a control to test for equal protein loading.
Mitochondrial transmembrane potential assay
Mitochondrial transmembrane potential (ΔΨm) was determined as described previously.28 Briefly, 5 × 105 cells were cultured in 10 cm dishes for 15 hr. After various treatments for 24 hr, the cells were harvested with 1× trypsin-EDTA and centrifuged in a 15 ml tube at 2,000 rpm for 5 min. After 2 washes with PBS, the cells were loaded with the cationic lipophilic fluorochrome, tetrachloro-tetraethylbenzimidazol carbocyanine iodide (JC-1, 5 μg/ml), for 20 min at room temperature. Cells were washed twice with PBS and submitted to flow cytometry analysis. Active mitochondria with high ΔΨm form JC-1 aggregates, which are red (FL2, 590 nm), whereas in mitochondria with low ΔΨm, JC-1 remains in a monomeric, green form (FL1, 527 nm). The ratio of red to green (FL2/FL1) reflects the change in ΔΨm.
Treatment of HepG2-derived xenografts in vivo
Male severe combined immune deficiency (SCID) mice (5–6-weeks old, purchased from National Science Council Animal Center, Taipei, Taiwan) were injected subcutaneously between the scapulae with 5 × 106 HepG2 cells in 0.2 ml PBS. After transplantation, tumor size was measured using calipers, and tumor volume was estimated according to the formula: tumor volume (mm3) = L × W2/2, where L is the length and W is the width.29 Once tumors reached a mean size of 200 mm3, animals received intraperitoneal injections of either DMSO (n = 5) or 10 mg/kg Ph (n = 5) three times per week for 6 weeks. All animal studies were performed according to the local guidelines for animal care and protection.
Following a previously described protocol,30 microPET images of the HepG2 tumor-bearing mice were obtained using the R4 system (Concorde Microsystems, Knoxville, TN). MicroPET scanning was performed at 6 weeks after tumor implantation. On the day of imaging, DMSO (n = 5) and Ph-treated (n = 5) tumor-bearing mice were anesthetized with isoflurane (Abbott Laboratories, Queensborough, Kent, England) using a vaporizer system (A.M. Bickford, Wales Center, NY) and intravenously injected with ∼14.8 MBq of 18F-fluorodeoxyglucose (18F-FDG) in 0.2 ml of saline through the lateral tail vein by bolus injection. Each mouse was imaged in the prone position. The mice were scanned for 10 min at each time point (30, 60, 90 and 120 min postinjection of the 18F-FDG), with the long axis of the mouse parallel to the long axis of the scanner. All images were reconstructed using the ordered-subsets expectation maximization reconstruction method, with a 128 × 128 pixel image matrix, 16 subsets, 4 iterations, and use of a Gaussian filter. Transmission scanning was performed at 130 min postinjection. Acquisition time was 600 sec, and images were reconstructed using filtered backprojection.
Results are expressed as mean ± S.E. for each study. Comparisons were subjected to one-way analysis of variance (ANOVA) followed by the Student's t-test. Statistical significance was accepted at p < 0.05.
Localization of GLUT2 transporter in HepG2 cells and human liver tissues and expression of GLUT2 mRNA in cell lines correlate with Ph cytotoxicity
We first examined the expression and localization of the GLUT2 transporter in HepG2 cells. Cells were immunostained with GLUT2 and caveolin-specific antibodies and analyzed by confocal microscopy. As shown in Figure1a, the yellowish dots in the cell membrane indicate the colocalization of GLUT2 and caveolin proteins, suggesting the presence of GLUT2 transporter in the HepG2 cell membrane. For human HCC and adjacent normal liver tissue, an immunohistochemistry study was performed. As shown in Figure 1b, the normal hepatocytes showed no apparent GLUT2 protein expression, whereas adjacent HCC tissue contained substantial GLUT2 protein expression in the cell membrane (arrows). We next analyzed GLUT2 mRNA expression in 23 primary cells or cell lines using real-time RT-PCR. As shown in Figure 1c, GLUT2 mRNA was highly expressed in HepG2 cells, followed by Hep3B cells. The remaining cells expressed a low level of GLUT2 mRNA. We then tested Ph, a GLUT2 inhibitor, for cytotoxic effects on cancerous and normal cells expressing either high or low levels of GLUT2. As shown in Figure 1d, treating HepG2 cells with 50–150 μM Ph for 48 hr in the MTT assay significantly suppressed cell viability in a dose-dependent manner. Treatment of Hep3B cells with 100–150 μM Ph for 48 hr showed significant suppression of cell viability in the MTT assay. However, the Ph-induced effect is less prominent for Hep3B as compared to HepG2 cells. In contrast, no significant effect was observed on normal breast epithelial cells (MCF 10A) or breast cancer cells (AU565), suggesting that the cytotoxicity of Ph correlates with expression of GLUT2.
GLUT2 was involved in the Ph-induced liver cancer cell death
To test the mechanisms of Ph-induced HepG2 cytotoxicity, flow cytometry analysis was performed as shown in Figure 2a. Treatment of HepG2 cells with 100 and 150 μM Ph for 24 hr slightly increased the populations in G0/G1 (64.99%) and G2/M (39.58%), respectively. In contrast, treatment of HepG2 cells with 200 μM Ph for 24 hr significantly induced HepG2 cell death, evidenced by an increase in the sub-G1 population to 21.55%. To further clarify the mechanism of HepG2 cell death induced by Ph, a DNA fragmentation assay was performed as shown in Figure 2b. Treatment of HepG2 cells with 200 μM Ph for 24 hr induced significant DNA ladder formation, suggesting apoptotic cell death.
Ph has been shown to competitively inhibit glucose transport into tumor cells. To test whether Ph-induced HepG2 cell death was associated with an interference with glucose transport, the flow cytometry study in Figure 2c was performed. Treatment of HepG2 cells with 200 μM Ph for 24 hr increased the sub-G1 population to 22.54%. However, as the glucose concentration increased from 20, to 50, then to 100 mM, the sub-G1 population decreased from 12.24%, to 7.56%, then to 5.69%, respectively, suggesting that Ph-induced HepG2 apoptosis involved suppression of glucose intake into the cells.
To confirm that GLUT2 is directly involved in the Ph-induced liver cancer cell death, siRNA knockdown of GLUT2 expression for HepG2 and Hep3B cells were used for further study (Fig. 2d). As shown in the left panel of flow cytometry, Figure 2e, similar sub-G1 population was induced for 200 μM Ph-treated HepG2 cells (the second bar) and GLUT2 siRNA knockdown HepG2 cells (the third bar). This effect cannot be further enhanced by treating GLUT2 siRNA knockdown HepG2 cells with Ph (the fourth bar). In contrast, scramble siRNA (the fifth bar) showed no significant effect to induce HepG2 apoptosis as compared to control. Interestingly, treatment of 100 mM glucose cannot reverse apoptotic effect in GLUT2 siRNA knockdown HepG2 cells (the sixth bar) suggesting a direct involvement of GLUT2 in the Ph-induced HepG2 cell death. Similar results were found for Hep3B as shown in the right panel of Figure 2e. But the effect of Ph-induced apoptosis in Hep3B cells was less prominent compared to HepG2 cells.
Both intrinsic and extrinsic apoptotic pathways are involved in Ph-induced apoptosis in HepG2 cells
Induction of apoptosis requires the activation of caspases.31 Therefore, Western blot analysis was performed to investigate the involvement of caspase activation in Ph-induced apoptosis in HepG2 cells. Treatment of HepG2 cells with 200 μM Ph activated caspase 3, as evidenced by a decrease in the level of procaspase 3 protein and the appearance of the cleaved form of caspase 3, and degradation of the caspase 3 substrate, poly-ADP-ribose polymerase (PARP). The addition of glucose at concentrations ranging from 20 to 100 mM reversed the activation of caspase 3 (Fig. 3a). To further elucidate the apoptotic pathways involved in the activation of caspase 3, we examined changes in the levels of caspases 8 and 9 proteins in the Ph-treated HepG2 cells. Treatment of HepG2 cells with 200 μM Ph activated caspases 8 and 9, as evidenced by the disappearance of procaspases 8 and 9, as well as the appearance of their cleaved forms. The addition of glucose at concentrations ranging from 20 to 100 mM reversed the activation of caspases 8 and 9 (Fig. 3a). These findings suggest that caspases 3, 8 and 9 are involved in Ph-induced apoptosis in HepG2 cells, and that the process depends on extracellular glucose concentration.
Activation of caspase 9 occurs during the release of cytochrome c from mitochondria into the cytosol. To test whether this release occurred in Ph-induced apoptosis in HepG2 cells, cytochrome c translocation was examined. As shown in Figure 3b, treatment of HepG2 cells with 200 μM Ph induced translocation of cytochrome c from the mitochondria into the cytosol. The addition of glucose at concentrations ranging from 20 to 100 mM reversed this translocation. Proteins of the Bcl-2 family are also involved in the control of apoptosis by regulating cytochrome c release. As shown in Figure 3b, treatment of HepG2 cells with 200 μM Ph induced translocation of Bax from the cytosol to mitochondria. It also increased expression of the Bax upstream regulator p53 protein and decreased expression of Bcl-2 protein. The addition of glucose at concentrations ranging from 20 to 100 mM reversed these effects. The proapoptotic protein Bid can be activated by caspase 8, leading to the formation of truncated Bid, which is subsequently translocated to the mitochondria during the activation of apoptosis. As shown in Figure 3b, treatment of HepG2 cells with 200 μM Ph induced Bid activation, leading to the formation of truncated Bid, which was reversed by the addition of glucose at concentrations ranging from 20 to 100 mM. These findings suggest that Bax, Bcl-2 and Bid proteins are all involved in Ph-induced apoptosis in HepG2 cells.
Akt activation can phosphorylate Bad, which leads to its inactivation through binding of a 14-3-3 protein. As shown in Figure 3c, treatment of HepG2 cells with 200 μM Ph inhibited Akt activation, evidenced by its decreased phosphorylation. It also activated Bad, as evidenced by the protein's translocation from the cytosol to the mitochondria. The addition of glucose at concentrations ranging from 20 to 100 mM reversed these effects. These findings suggest that Akt-Bad pathways are involved in Ph-induced apoptosis in HepG2 cells.
Cytochalasin B enhances Ph-induced apoptosis in HepG2 cells
Cytochalasin B is a glucose transporter inhibitor and has been shown to regulate a specific glucose-sensitive binding site in liver plasma membrane.32 We tested the cytotoxic effects of cytochalasin B alone or in combination with Ph on HepG2 cells using a flow cytometry assay. As shown in Figure 4a, treatment of HepG2 cells with 20 μM cytochalasin B induced G2/M cell cycle arrest (63.72%) without significant cell death. Treatment of HepG2 cells with 200 μM Ph increased the sub-G1 population to 22.54%, indicating significant cell death. Combined treatment of HepG2 cells with 200 μM Ph and 0.5–20 μM cytochalasin B increased the sub-G1 population from 30.76 to 70.37% in a dose-dependent fashion, indicating enhancement of cell death. To test whether cytochalasin B enhancement of Ph-induced HepG2 cell death was associated with glucose transport interference, a flow cytometry study (Fig. 4b) was performed. Combined treatment of HepG2 cells with 200 μM Ph and 20 μM cytochalasin B increased the sub-G1 population from 23.04 to 68.85%; this population was reduced to 7.53% by treatment with 100 mM glucose. These findings indicate that cytochalasin B enhancement of Ph-induced HepG2 apoptosis involves perturbation of glucose transport.
Decreased ΔΨm has been linked to apoptotic cell death. To determine the effects of Ph and cytochalasin B on the ΔΨm change in apoptotic HepG2 cells, flow cytometry analysis was performed. HepG2 cells were stained with a cationic lipophilic fluorescent dye, JC-1, which emits red or green light in response to a high or low ΔΨm, respectively. Therefore, the mitochondrial membrane potential can be measured by the red/green fluorescence intensity ratio. As shown in Figure 4c, treatment of HepG2 cells with 20 μM cytochalasin B for 24 hr showed no significant ΔΨm change compared with the control. Treatment of HepG2 cells with 200 μM Ph for 24 hr caused a significant decrease in ΔΨm, evidenced by a shift from red (FL2) to green (FL1) fluorescence. Combined treatment of HepG2 cells with 200 μM Ph and 20 μM cytochalasin B further decreased ΔΨm. Cotreatment with 100 mM glucose reversed the cytochalasin B enhancement of the Ph-induced decrease of ΔΨm in HepG2 cells. These findings indicate that cytochalasin B enhancement of the Ph-induced decrease in ΔΨm of HepG2 cells involves regulation of glucose transport.
To verify that the cytochalasin B enhancement of Ph-induced HepG2 cell death involved glucose transport, Western blot studies of apoptotic regulatory proteins were performed. As shown in Figure 4d, treatment of HepG2 cells with 20 μM cytochalasin B did not activate caspases or apoptotic regulatory proteins. However, treatment of HepG2 cells with 200 μM Ph activated caspases 3, 8 and 9, along with increased expression of p53 protein, translocation of Bax from the cytosol to the mitochondria, decreased expression of Bcl-2 protein, formation of truncated Bid and translocation of cytochrome c from the mitochondria to the cytosol. This treatment also led to inhibition of Akt phosphorylation, Bad translocation from the cytosol to mitochondria, suppression of XIAP and survivin expression, and AIF translocation from the mitochondria to the nucleus. Combined treatment of HepG2 cells with 200 μM Ph and 20 μM cytochalasin B further enhanced the activation of caspases 3, 8 and 9, and increased both the expression of p53 protein and the translocation of Bax from the cytosol to the mitochondria. This treatment also decreased expression of Bcl-2 protein, formation of truncated Bid and translocation of cytochrome c from the mitochondria to the cytosol. All of these effects were reversed by cotreatment with 100 mM glucose.
Ph inhibition of HepG2 xenograft growth in mice involves suppression of glucose uptake
We further examined the antitumor effect of Ph in vivo by treating SCID mice bearing HepG2 tumor xenografts. After growth of palpable tumors (mean tumor volume, 200 mm3), animals received Ph at a dose of 10 mg/kg or DMSO control 3 times per week. The tumor size of Ph-treated group began to decrease significantly compared with control group from the second to the sixth week (Fig. 5a), and the final tumor weight (Fig. 5b) was significantly smaller in Ph-treated group than in control group. In mice receiving these treatment regimens, no gross signs of toxicity were observed based on body weight, visible inspection of general appearance or microscopic examination of individual organs (data not shown). To determine whether Ph-induced suppression of liver tumor growth involved regulation of glucose transport, an 18F-FDG microPET study was performed. As shown in Figure 5c, 18F-FDG uptake in the Ph-treated HepG2 tumor-bearing mice was significantly suppressed relative to control mice. With microPET analysis in these mice, tumor volumes were significantly smaller in the Ph-treated groups compared to controls (Fig. 5d). After intravenous injection of the same dosage of 18F-FDG into Ph-treated and control mice (Fig. 5e), quantitative tumor uptake decreased ∼10-fold in the Ph-treated, tumor-bearing mice compared to the control mice (Fig. 5f). These findings suggest that suppression of glucose uptake into tumor tissues is involved in the Ph-induced inhibition of hepatoma xenograft growth in mice.
An increased rate of glucose uptake into the transformed cells is one of the alterations of cellular function that occurs when the sarcoma virus transforms normal cells into malignant cells.33 The elevated glucose transport into malignant cells is associated with increased expression of GLUTs.11 Positron emission tomography has been used to show that high-grade tumors take up more FDG than low-grade tumors, and this increased uptake is associated with more aggressive behavior and poor prognosis.34, 35 Suppressing glucose uptake by tumors, therefore, is a reasonable strategy to pursue in the search for improved cancer therapies. Our cellular and animal studies clearly show that Ph selectively suppresses hepatoma cell growth by activating apoptosis through inhibition of glucose transport mechanisms. To our knowledge, this is the first demonstration that Ph exerts an antitumor effect on hepatoma cells both in vitro and in vivo.
Ph has been shown to stimulate transcriptional activity of the human estrogen receptor and to compete for the binding of radiolabeled 17β-estradiol to this protein,36 suggesting that Ph can act as a nonsteroid estrogen. However, previous reports have demonstrated that treatment of HepG2 cells with 17β-estradiol induces cell proliferation,37 suggesting that the potential effect of Ph on the estrogen receptor is not involved in Ph-induced cell death in HepG2 cells. Ph has been shown to suppress the protein kinase C (PKC) pathway in melanoma cells. The effects on HepG2 cells of modulating PKC pathways are complex.38 Nevertheless, it is clear from the present work that pretreatment with glucose reverses the apoptotic effect of Ph on HepG2 cells and that siRNA knockdown of GLUT2 expression induced HepG2 cell apoptosis which cannot be reversed by glucose treatment suggesting a direct involvement of GLUT2 in the Ph-induced cytotoxicity in liver cancer cells.
In our study, 200 μM Ph was used to induce HepG2 cell apoptosis. Using primary rat hepatocytes, previous reports demonstrated that Ph at 0.4 to 200 μM significantly reduced tert-butyl hydroperoxide-induced cytotoxicity,39 and reported that the LD50 of Ph is 500 ± 51 μM.40 One report showed that Ph at 30 μM protected Aβ peptide-induced plasma membrane damage in PC12 cells, while 300 μM Ph became toxic to the cells.41 Ph at 50 to 500 μM blocked Helicobacter pylori cytotoxin VacA-mediated urea/ion transport and the toxin's major biological effects.42 Importantly, Ph of 10 mg/kg in our in vivo HepG2 xenograft mice model showed significant antitumor effects without apparent toxicity. This dosage of Ph has been reported to be nontoxic in animal study using female rats.43 These data suggest that the doses of Ph used in our in vitro and in vivo studies on hepatoma cells might be feasible approaches.
Malignant cells have an increased metabolic rate.11 Furthermore, the activity of lactate dehydrogenase44 and pentose phosphate pathway enzymes45, 46 is increased in malignant cells. These changes mean that a reduced amount of pyruvate is available to mitochondria, resulting in a low oxidation rate via the tricarboxylic acid cycle.45 The high energy requirements of tumors, together with their low energy production, leads to an increased rate of glycolysis and, therefore, glucose uptake. These observations are compatible with our findings that both human hepatoma cells and HCC tissue show increased expression of GLUT2 in the cell membrane.
ATP depletion is an important mechanism of apoptosis.22, 47 Mitochondrial membrane potential, the driving force of ATP production, is decreased during apoptosis.48 Partial ATP depletion induced the activation of the caspase 8-associated extrinsic apoptotic pathways in kidney cells.49 Our results show that Ph induces apoptosis in hepatoma cells by inhibiting glucose transport, and subsequently by the possible suppression of ATP production, resulting in mitochondrial membrane potential reduction (Fig. 6). Both intrinsic and extrinsic apoptotic pathways are implicated in this study, which may reflect the fact that 200 μM Ph causes more severe ATP depletion23 and, therefore, leads to the breakdown of mitochondria, activating downstream apoptotic pathways.
Elevation of facilitated glucose transporter mRNA in human HCC has been demonstrated using northern blot analysis.50 In a study of various human gastrointestinal cancers, GLUT2 mRNA was primarily restricted to the liver, with increased expression of GLUT2 transcripts in tumorous vs. normal tissues.20 In contrast, GLUT2 expression was reduced in hepatic preneoplastic or neoplastic lesions, with expression of GLUT1 in HCC.51 However, no expression of GLUT1 has been reported in human HCC or nontumorous tissues.21 Our results showed significantly increased expression of GLUT2 protein in human HCC and hepatoma cells in the cell membrane, suggesting translocation of the protein.24, 52, 53 The difference between our results and those of Grobholz et al.51 may be that they used an animal model of chemically induced hepatocarcinogenesis (N-nitrosomorpholine) rather than the human HCC tumor model used in other studies.
On the basis of our results, we propose a model for the molecular mechanism of Ph-induced apoptosis on HepG2 cells (Fig. 6). Our experimental findings demonstrate that Ph, a naturally occurring compound with glucose transporter inhibitory activity, exerts antitumor activity through the induction of apoptosis in hepatoma cells both in culture and in tumor-bearing SCID mice. The results of these in vitro and in vivo studies shed light on the molecular mechanisms of selective Ph-induced cytotoxicity in hepatoma cells. These findings may have potential applications in the treatment of liver cancer.
This work was supported by research grants from the National Science Council of Taiwan to Dr. Ho, Dr. Wu and Dr. Lin. Technical support was obtained from Molecular and Genetic Imaging Core/NRPGM.