|Home | About | Journals | Submit | Contact Us | Français|
Hepatitis C virus (HCV) infection induces a state of oxidative stress by affecting mitochondrial-respiratory-chain activity. By using cell lines inducibly expressing different HCV constructs, we showed previously that viral-protein expression leads to severe impairment of mitochondrial oxidative phosphorylation and to major reliance on nonoxidative glucose metabolism. However, the bioenergetic competence of the induced cells was not compromised, indicating an efficient prosurvival adaptive response. Here, we show that HCV protein expression activates hypoxia-inducible factor 1 (HIF-1) by normoxic stabilization of its α subunit. In consequence, expression of HIF-controlled genes, including those coding for glycolytic enzymes, was significantly upregulated. Similar expression of HIF-controlled genes was observed in cell lines inducibly expressing subgenomic HCV constructs encoding either structural or nonstructural viral proteins. Stabilization and transcriptional activation of HIF-1α was confirmed in Huh-7.5 cells harboring cell culture-derived infectious HCV and in liver biopsy specimens from patients with chronic hepatitis C. The HCV-related HIF-1α stabilization was insensitive to antioxidant treatment. Mimicking an impairment of mitochondrial oxidative phosphorylation by treatment of inducible cell lines with oligomycin resulted in stabilization of HIF-1α. Similar results were obtained by treatment with pyruvate, indicating that accumulation of intermediate metabolites is sufficient to stabilize HIF-1α. These observations provide new insights into the pathogenesis of chronic hepatitis C and, possibly, the HCV-related development of hepatocellular carcinoma.
Hepatitis C virus (HCV) infection is a major cause of chronic liver disease worldwide (43). Its progression to cirrhosis and hepatocellular carcinoma (HCC) may take decades, but the precise pathogenic mechanisms are unknown (20). HCV is a positive-strand RNA virus with a 9.6-kb genome encoding a large polyprotein (2). This is translated on the endoplasmic reticulum (ER) and processed by viral and host proteases into 10 individual membrane-associated proteins comprising structural (core, E1, and E2) and nonstructural (p7 and NS2 to NS5B) proteins (28). Evidence derived from experimental systems, as well as from liver biopsy specimens from patients with chronic hepatitis C (CHC), point to HCV-related dysfunction of mitochondria (32, 33, 47).
Warburg first proposed that the prime cause of cancer was impaired energy metabolism (50), which involved irreversible injury to cellular respiration followed by a gradual dependence on fermentation energy to compensate for the lost energy from mitochondrial respiration. Indeed, elevated glycolysis is the metabolic hallmark of nearly all tumors (23). Although Warburg's hypothesis has been overshadowed by support for the mutational theory of carcinogenesis, his original observations still have merit, even though the exact mechanisms have yet to be clarified.
In mammalian cells, an essential control element of the metabolic state is hypoxia-inducible factor 1 (HIF-1) (25). A well-defined oxygen-sensitive pathway regulates the activity of HIF-1 by posttranslational prolyl hydroxylation within the α subunit (38). Since the HIF prolyl hydroxylases (PHDs) have an absolute requirement for molecular oxygen, this process is suppressed in hypoxia, allowing HIF-1α to escape its degradation route (46) and to activate transcription. However, non-hypoxia-related factors have been shown to contribute to the activation of HIF-1, providing additional interfaces that may be important in regulating cellular stress adaptation (6). Almost all the enzymes of the glycolytic pathway are encoded by genes whose expression is under the control of HIF-1 (40). In cancer, the metabolic phenotype is activated by a variety of genetic and environmental mechanisms, most strikingly by stabilization of HIF-1 (34), which accounts for the classical tumor-associated properties of deregulated glycolysis and angiogenesis (34, 40).
In a recent report (29), it was shown that HCV infection elicits oxidant-mediated HIF-1α stabilization, leading to expression of vascular endothelial growth factor (VEGF). In the present study, we show that long-term HCV protein expression causes depression of mitochondrial oxidative phosphorylation (OXPHOS). Cell survival, however, is preserved by enhanced nonoxidative glucose utilization. Consistent with the results of Moradpour et al. (28), this adaptive response to HCV-induced mitochondrial injury proved to be mediated by stabilization of HIF-1α and, in consequence, upregulation of glycolytic enzymes. Comparable results were obtained in two different cell culture systems and, more importantly, in liver biopsy specimens from HCV-infected patients. Further, our study suggests the involvement of intermediate-metabolite-mediated inhibition of HIF-1 α-prolyl hydroxylation as a cause of the observed HCV-related alterations, and it provides novel insights into the pathogenesis of HCC.
UHCVcon-57.3 is a U-2 OS human osteosarcoma-derived tetracycline-regulated cell line inducibly expressing the HCV polyprotein (37). Cell lines UCp7con-9.10 and UNS3-5Bcon-27 inducibly express the HCV structural proteins and p7 or the nonstructural proteins 3 to 5B, respectively, derived from the HCV H consensus clone (15; D. Moradpour, unpublished data). As a control, the UGFP-9.22 cell line inducibly expressing the nonrelevant green fluorescent protein (GFP) was used (37). Inducible cell lines were cultured in complete Dulbecco's modified Eagle's medium (DMEM) containing 5.5 mM glucose or, when indicated, 10 mM galactose. Huh-7.5 human HCC cells (3) (kindly provided by Charles M. Rice, The Rockefeller University, New York, NY) were transfected by microelectroporation (Digital Bio Technology, Seoul, Korea) with a J6-JFH-1 (Jc1) chimeric full-length HCV genome harboring a GFP insertion in domain III of NS5A, as described previously (36) (kindly provided by R. Bartenschlager, University of Heidelberg, Heidelberg, Germany).
Huh-7.5 cells were electroporated with in vitro-transcribed Jc1 RNA. Supernatants were collected 48 h postelectroporation, and 50% tissue culture infectious doses were determined as described previously (19). The Huh-7.5 cells were infected with Jc1 virus at a multiplicity of infection of about 7.5. Four days postinfection, the cells were lysed in a buffer containing 150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 50 mM Tris-Cl, pH 8.0, and protease inhibitors.
For mitoplast (i.e., an isolated mitochondrion devoid of the outer mitochondrial membrane) preparation, UHCVcon-57.3 cells were harvested with 0.05% trypsin and 0.02% EDTA and washed in phosphate-buffered saline (PBS), pH 7.4, with 5% calf serum. The cells in PBS were exposed for 10 min on ice to 0.2 mg of digitonin/mg cellular protein. The mitoplasts were pelleted at 14,000 × g and resuspended in PBS. For the mitochondrion-enriched fraction, the harvested UHCVcon-57.3 cells were subjected to nitrogen cavitation in a cell disruption bomb (Parr Instrument Company, Moline, IL; model 4639). Briefly, 20 × 106 to 30 × 106 cells were suspended in 1 to 2 ml of 250 mM sucrose, 1 mM EGTA, 5 mM HEPES, 3 mM MgCl2, 40 μl of protease inhibitor cocktail (Roche), pH 7.4, and equilibrated for 20 min in the bomb at a nitrogen pressure of 700 lb/in2. Following rapid decompression, the ejected disrupted cells were centrifuged at 400 × g for 5 min. The pelleted undisrupted cells were resuspended and subjected to a second cycle of nitrogen compression-decompression. The pooled supernatants were centrifuged at 600 × g for 5 min to remove nuclei and cell debris and further centrifuged at 12,000 × g for 15 min. The pelleted mitochondrion-enriched fraction was resuspended in 50 to 100 μl of 250 mM sucrose, 0.2 mM EDTA, 10 mM Tris, pH 7.8. The protein contents of the cell suspension and subcellular fractions were assayed by the Bradford method using albumin as a standard.
From March 2007 to October 2007, patients with CHC referred to the outpatient liver clinic of the University Hospital Basel were asked for permission to use part of their diagnostic liver biopsy specimens for research purposes. Patients' biopsy specimens with no histological signs of cirrhosis (or other inflammatory alterations) were selected. All of the patients were Caucasians. The protocol was approved by the Ethics Committee of the University Hospital Basel. Written informed consent was obtained from all patients. As non-CHC controls, 14 patients who underwent ultrasound-guided liver biopsies of focal lesions gave informed consent for a biopsy from the normal liver tissue outside the focal lesion. These control patients had normal liver values, and the absence of liver pathology was histologically confirmed.
Cells were trypsinized, washed in PBS, and immediately assessed for O2 consumption by high-resolution respirometry (Oroboros Instruments) as described previously (33). The activities of complexes I, III, and IV were assayed as described previously (33). Complex V activity was measured spectrophotometrically by a coupled assay of the mitoplast fraction of ultrasound-treated cells in 50 mM Tris, 5 mg/ml serum albumin, 20 mM MgCl2, 50 mM KCl, pH 8.0, supplemented with 0.75 μM antimycin A, 1 μM carbonyl cyanide m-chlorophenylhydrazone, 1.5 mM phosphoenolpyruvate, 2.5 units/ml of lactate dehydrogenase (LDH), 1.12 units/ml of pyruvate kinase, 350 μM ATP. The reaction was started by adding 50 μM NADH, whose oxidation was recorded from the absorbance decrease at 340 nm.
Cellular ATP was extracted by a one-step procedure in boiling water and assayed by bioluminescence using a luciferin-luciferase reaction system as described previously (33). Lactate was assayed spectrophotometrically in culture medium and quantified by comparison with a calibration curve using standard titrated lactate according to the manufacturer's instructions (Vinci-Biochem). The values reported were normalized to the cell number.
Huh-7.5 cells were suspended at 0.25 × 106 cells/ml in DMEM plus 10% fetal bovine serum and placed in the spectrofluorometric cuvette with stirring. The instrument settings were as follows: λex = 485 nm and λem = 535 nm. To remove the green fluorescence background of the transfected Huh-7.5 cells harboring the GFP insert, the fluorescence was instrumentally zeroed, and then 10 μM 2′,7′- dichloro-fluorescein diacetate (DCF-DA) was added. DCF-DA is not fluorescent, but following its entry into the cell, it is hydrolyzed by intracellular esterase to generate DCF, which becomes fluorescent following reaction with hydrogen peroxides. The fluorescence changes were recorded for 30 min.
Cells were resuspended in lysis buffer (20 mM HEPES, pH 7.2, 150 mM NaCl, 1 mM EGTA, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl2, 2 mM sodium phosphate, and protease inhibitor cocktail). The lysates were run on a 10% SDS-polyacrylamide gel electrophoresis gel, and Western blotting was performed by standard transferring procedures. The polyvinylidene difluoride membranes were incubated with monoclonal antibody (MAb) 5B-3B1 against HCV NS5B (26), MAb C7-50 against HCV core (27), MAb 11H against HCV NS5A (4), MAb 1B6 against HCV NS3 (51), 1:500 rabbit anti-HIF-1α antibody (Santa Cruz Biotechnology), 1:10,000 mouse anti-β-actin (Sigma), and 1:300 rabbit anti-SERCA (Santa Cruz Biotechnology) overnight at 4°C and finally treated with 1:8,000 horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology) before analysis with a chemiluminescence kit (Pierce) using the VersaDoc Imaging System (Bio-Rad).
U-2 OS-derived inducible cell lines or transfected Huh-7.5 cells were seeded onto fibronectin-coated glass bottom dishes. Then, samples were fixed, permeabilized, blocked, and incubated with 1:100-diluted rabbit anti-human HIF-1α (Acris) overnight at 4°C. After being washed in PBS-bovine serum albumin, the samples were incubated with 8 μg/ml of fluorescein isothiocyanate-labeled goat anti-rabbit immunoglobulin G or rhodamine-labeled goat anti-mouse immunoglobulin G (Santa Cruz Biotechnology). ROS production was evaluated by treating the cells with 10 μM DCF-DA as described previously (33). After being washed in PBS, the cells were immediately analyzed. Imaging of labeled cells was performed with a Nikon TE 2000 microscope (the images were collected using a 60× objective, 1.4 numerical aperture) coupled to a Radiance 2100 laser scanning confocal microscopy (LSCM) system (Bio-Rad). Acquisition, storage, and analysis of data were done using LaserSharp and LaserPix software from Bio-Rad or ImageJ version 1.37.
Total cellular RNA was isolated from cell cultures with an Absolutely RNA miniprep kit (Stratagene) with an on-column DNase treatment. First-strand cDNA synthesis was carried out using 300 ng of random hexamer primers (Invitrogen) by Accuscipt High Fidelity Reverse Transcriptase (Stratagene) and Ribolock RNase Inhibitor (Fermentas), starting from 1 μg RNA. Reverse transcription (RT) quantification was performed with 1.5 μl cDNA using Brilliant SYBR green QPCR Master Mix (Stratagene) in a 25-μl reaction volume on Mx3000P (Stratagene) with 300 nM primers (see Table S1 in the supplemental material). The quantification of transcript abundance in derepressed cells was done using the ΔΔCT method, with the repressed cells as the reference sample (calibrator) and β-actin as the internal control. For liver biopsy specimens from CHC and control patients, samples were stored at −75°C after having been stabilized in RNAlater solution (Ambion/Applied Biosystems, Rotkreuz, Switzerland) at 4°C overnight. Total RNA was extracted from liver samples using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. RNA was aliquoted and stored at −75°C. The RNA was reverse transcribed with Moloney murine leukemia virus reverse transcriptase (Promega Biosciences, Inc., Wallisellen, Switzerland) in the presence of random hexamers (Promega) and deoxynucleotide triphosphates. The reaction mixtures were incubated for 5 min at 70°C and then for 1 h at 37°C and were stopped by heating them at 95°C for 5 min. The SYBR PCRs were performed using the SYBR green PCR master mix (Applied Biosystems) and primers (data not shown), The difference in the cycle threshold (ΔCT) value was derived by subtracting the CT value for GAPDH (glyceraldehyde 3-phosphate dehydrogenase), serving as an internal control, from the CT values for the transcripts of interest. All reactions were run in duplicate, using an Applied Biosystems Prism 7000 Sequence Detection System. mRNA expression levels were calculated relative to GAPDH from the ΔCT values, using the formula 2−ΔCT. The primer pairs used for quantitative RT (qRT)-PCR were for HIF-1α, LDH-A, hexokinase 1 (HK1), VEGF, VHL, and PHD2; β-actin or GAPDH amplicons were used as internal controls.
The two-tailed Student's t test was applied, with a P value of <0.05, to evaluate the statistical significance of differences measured throughout the data sets from cell culture experiments. For statistical analyses of the mRNA levels in human liver biopsy specimens, the two-tailed Mann-Whitney test was used.
Previously, we reported the impact of 2 days of induction of HCV protein expression on mitochondrial-respiratory-chain (RC) activity and OXPHOS efficiency (33). In this study, we extended these observations to evaluate longer-term effects. Figure Figure1A1A shows the results of a respirometric analysis carried out on stably transfected intact U-2 OS-derived cell lines. Induction of HCV proteins for 2 and 5 days resulted in the following similar outcomes: (i) 40% inhibition of the resting O2 uptake, (ii) a slight increase in respiratory activity in the presence of oligomycin, and (iii) a marked reduction of the respiratory control ratio (RCR) from 5 to 2.
The mitochondrial membrane potential (mtΔΨ) is known to control the rate of O2 consumption at the level of the RC. Thus, in the presence of oligomycin, which inhibits mtΔΨ-consuming H+-ATP synthase, a rise in the mtΔΨ and, consequently, a reduction of the respiratory rate are expected. The RCR (attained by dividing the resting rate of respiration by that in the presence of oligomycin) is an index of the OXPHOS efficiency. In HCV-induced U-2 OS-derived cells, the occurrence of a combined effect contributed by inhibition of the RC activity and by a dissipatory pathway for mtΔΨ was observed. The consequent depression of the RCR indicated a severe impairment of the OXPHOS yield.
Analysis by blue native/SDS bidimensional polyacrylamide gel electrophoresis of the protein profile of the mitoplasts showed no significant change either in the amount or in the composition of the OXPHOS complexes between noninduced and induced cells (data not shown). The measurement of the specific activities of the OXPHOS complexes revealed 40% and 50% inhibition of complex I and FoF1-ATPase, respectively, at both 2 and 5 days of HCV protein induction, whereas the activities of cytochrome c reductase and cytochrome c oxidase remained unchanged (Fig. (Fig.1B).1B). Immunoblots of the HCV core, NS3, NS5A, and NS5B proteins verified their expression following the 5-day induction protocol and displayed, in addition, a significant accumulation of HCV core and NS5A, and to a lesser extent of NS3 and NS5B, in mitochondria as assessed by comparing their relative amounts in cell lysates and organelle-enriched fractions (Fig. (Fig.1C).1C). The effect of HCV protein expression on cell growth was evaluated by cell density analyses (Fig. (Fig.1D),1D), demonstrating significant changes from the second day of induction compared to noninduced cells. However, cell growth was not as severely depressed as expected from the extent of the mitochondrial OXPHOS crisis.
The effect of HCV protein expression on cell bioenergetics was assessed by measuring the steady-state intracellular level of ATP. Figure Figure2A2A shows that 5 days of HCV protein induction resulted in an 80% increase in ATP relative to noninduced cells. This effect depended on the energy substrate present in the culture medium. Indeed, replacing glucose with galactose (a condition forcing cells to rely largely on mitochondrial OXPHOS ) resulted in a significant 30% decrease in the ATP content in HCV-induced cells. Consistently, the growth of HCV-induced cells in galactose-based medium was largely inhibited compared to that of cells grown in glucose (data not shown). Measurement of the amount of lactate released into a glucose-based medium following 5 days of HCV protein induction resulted in an almost threefold increase relative to noninduced cells (Fig. (Fig.2B).2B). This observation indicates an HCV protein-induced metabolic shift toward glycolysis.
To explain the glycolytic phenotype in the HCV protein-expressing cells, we postulated an upregulation of the glycolytic enzymes, whose expression is largely controlled by HIF-1. Therefore, we analyzed by LSCM the level and subcellular localization of the HIF-1α subunit. As shown in Fig. Fig.3A,3A, 5 days of HCV protein induction resulted in a 2.4-fold enhancement of the HIF-1α-related immunofluorescence, which largely localized to the nuclear compartment. Immunoblot analyses of cell lysates confirmed the HCV protein-dependent upregulation of HIF-1α. Treating noninduced cells with the hypoxia-mimetic desferroxiamine (or CoCl2 [not shown]) resulted in comparable outcomes. Of note, the HIF-1α stabilization in induced cells occurred under normal oxygen tension, thus suggesting a hypoxia-independent mechanism of HIF-1α stabilization. After 2 days of HCV protein induction, no significant increase in HIF-1α was observed. However, it became clearly detectable after 4 days of induction (data not shown). This indicated the need for a progressive setting of conditions for HIF-1α stabilization.
To assess whether the HCV protein-linked HIF-1α stabilization was functional, we measured the transcriptional level of known HIF-controlled genes by qRT-PCR. Figure Figure3B3B shows a significant 2- to 2.5-fold increase relative to noninduced cells of the glycolytic enzymes LDH-A and HK1, as well as of VEGF, encoded by a well-characterized HIF target gene (25). Of note, the transcript level of HIF-1α itself and that of the VHL protein, which controls HIF-1α stability (46), were unaffected by the viral-protein expression, suggesting HCV-linked regulation of HIF-1 at the posttranslational level.
To obtain insights into the possible involvement of a specific HCV protein in HIF activation, we examined U-2 OS-derived cell lines inducibly expressing either the structural or the nonstructural proteins. No significant difference in the induction of the HIF-controlled genes between the UHCVcon-57.3, UCp7con-9.10, and UNS3-5Bcon-27 cell lines was observed (Fig. (Fig.3B).3B). Hence, these results did not allow us to attribute the observed effect to either the structural or nonstructural region.
To confirm the specificity of our observations, we used a U-2 OS-derived cell line inducibly expressing the GFP as a nonrelevant control (Fig. (Fig.4A).4A). After 5 days of GFP induction, the levels of HIF-1α protein and of the related gene expression were comparable in cells cultured in the presence and absence of tetracycline, as shown in Fig. 4B and C. Moreover, the expression of the GFP did not cause any mitochondrial dysfunction, as assessed by probing the mtΔΨ (Fig. (Fig.4A4A).
To verify that the observed HCV protein-dependent activation of HIF-1 was not limited to a specific in vitro cell system, we extended the analyses to Huh-7.5 human HCC cells electroporated with a GFP-tagged HCV genome allowing the full replication cycle and production of infectious virus (37, 3). As shown in Fig. Fig.5A,5A, Huh-7.5 cells harboring HCV displayed both stabilization and nuclear localization of HIF-1α, as assessed by Western blotting and immunofluorescence microscopy, respectively. qRT-PCR analyses confirmed a significant upregulation of HK1 and VEGF, but not of LDH-A (Fig. (Fig.5B).5B). Of note, the transcript levels of HIF-1α, PHD2 (isoform 2 of PHD), and VHL were all significantly enhanced compared to nontransfected cells. Furthermore, normoxic stabilization of HIF-1α was also verified in Huh-7.5 cells infected with cell culture-derived HCV, resulting in a 4.5-fold increase in the protein level assayed 4 days after infection (Fig. (Fig.66).
Among the nonhypoxic factors known to stabilize HIF-1α are ROS (6). Since expression of HCV proteins was shown to cause enhanced mitochondrial ROS production, we tested their involvement in the upregulation of HIF-1α. The results shown in Fig. Fig.77 ruled out this possibility. Indeed, in spite of the clear antioxidant effects exerted by N-acetyl-cysteine (NAC) or by Tiron on the HCV-dependent redox alteration (Fig. (Fig.7A),7A), no significant change in the HIF-1a protein or in the expression of HK1, LDH-A, and VEGF was observed in induced cells treated with both ROS scavengers (Fig. (Fig.7B).7B). HCV protein expression is linked to deregulation of the calcium fluxes between ER and mitochondrial stores (21, 33). Increased mitochondrial calcium influx is known to activate a mitochondrial isoform of nitric oxide synthase and thereby to cause nitric oxide release (33), which has been reported to stabilize HIF-1 (5). Furthermore, activation of mitogen-activated protein kinase and phosphatidylinositol 3-kinase (PI3K) signaling pathways was shown to induce HIF-1 (9). However, treatment of UHCVcon-57.3 cells with inhibitors of the ER Ca2+ channel (dantrolene), mitochondrial Ca2+ uniporter (ruthenium red), nitric oxide synthase (N-methyl-l-arginine [l-NNMA]), and the PI3K/Akt pathway (wortmannin) were ineffective in modifying the upregulated expression profile of HIF-1 target genes (Fig. (Fig.7C).7C). Importantly, the inefficiency of ROS scavenging in preventing the HCV protein-related upregulation of HIF was also fully confirmed in Huh-7.5 cells transfected with the GFP-tagged genome, as shown in Fig. 8A to C. Under conditions that prevented ROS generation in transfected Huh-7.5 cells (measured by probe-assisted fluorimetry), both NAC and Tiron had no significant effect on the stabilization of HIF-1α protein and on the transcription of the HIF-controlled VEGF gene.
To clarify the cause-effect relationship between the observed normoxic stabilization of HIF-1α and the metabolic change caused by HCV protein expression, we mimicked an impairment of the mitochondrial OXPHOS system in noninduced UHCVcon-57.3 cells and assessed its impact on the HIF system. As shown in Fig. Fig.9,9, normoxic treatment of UHCVcon-57.3 cells with a subcytotoxic concentration of oligomycin, a specific inhibitor of the mitochondrial ATP synthase, resulted in a remarkable stabilization and nuclear localization of HIF-1α. As inhibition of the terminal catabolic flux is expected to cause accumulation of intermediate metabolites, we tested the effect of pyruvate on noninduced UHCVcon-57.3 cells. Figure Figure99 shows that incubation with a concentration of pyruvate, in a range compatible with its physiologic accumulation in proliferating glycolytic cells (49), caused stabilization of HIF-1α as observed in cells expressing HCV proteins. The effect of oligomycin or pyruvate treatment was also tested in Huh-7.5 cells, confirming their capability to induce stabilization of HIF-1α to levels comparable to that attained following HCV transfection (Fig. 10A and B).
To validate the HCV-linked upregulation of HIF-1 observed in in vitro cell culture systems, we extended our analyses to human liver biopsy specimens. A cohort of 19 patients affected by CHC was selected and compared with 14 control patients. Figure Figure1111 shows the comparative results of qRT-PCR analyses. It shows that the levels of HIF-1α, VHL, HK1, and VEGF transcripts were significantly higher in the liver biopsy specimens from HCV-infected patients than in the controls. Conversely, no significant difference was observed for the LDH transcripts. Interestingly, the differential expression profile observed in the HCV-infected patients closely resembled that described above in Huh-7.5 cells harboring cell culture-derived HCV (Fig. (Fig.5B).5B). Importantly, a statistical analysis comparing the HIF-1α expression levels and the inflammatory grade (using the Metavir grading/staging system) allowed us to exclude any correlation between them (data not shown).
HCV infection is associated with mitochondrial stress. In a previous study, we showed that the expression of the HCV proteins elicits a sequence of events primed by enhanced uptake of Ca2+ into mitochondria (33). The increased intramitochondrial concentration of Ca2+ was found to stimulate nitric oxide production and to inhibit RC complex I. This led to a progressive inhibition of the RC activity and of the mtΔΨ, accompanied by production of ROS. Interestingly, the ATP content of HCV protein-expressing cells was found to be increased despite these profound alterations.
To obtain insights into this cell adaptive response, we extended our studies to cells expressing HCV proteins for a longer time. The results obtained confirmed the remarkable depression of the OXPHOS system, further characterized by reduced oligomycin-sensitive FoF1 ATPase activity (Fig. 1A and B). No compensatory upregulation of the biogenesis of the OXPHOS complexes was observed. In this study, we confirmed previous reports showing partial localization of some HCV proteins on mitochondria (Fig. (Fig.1C)1C) (37, 39, 47), thus supporting the evidence that a direct physical interaction was causally related to the HCV-linked mitochondrial alterations. However, cell bioenergetics was not compromised even after longer-term induction of HCV protein expression because of an enhanced glycolytic catabolic flux, resulting in an increased level/production of intracellular ATP (Fig. (Fig.2).2). This paradoxical overproduction of ATP, hardly explainable by an allosteric regulation of glycolysis, was a clue arguing in favor of changes in the glycolytic-gene expression profile.
HIF-1 is the major transcription factor controlling the expression of practically all the enzymes of the glycolytic pathway, as well as of the inducible glucose transporters (25, 41). It is activated under conditions of low O2 tissue tension and provides a first adaptive metabolic response at the cellular level. If the hypoxic insult persists, HIF-1 upregulates the expression of factors involved in angiogenesis and erythropoiesis, thereby allowing an adaptive response at the systemic level (25). In addition to these early-recognized tasks, the involvement of HIF-1 in controlling a wide variety of cell functions has been unveiled over time. Importantly, conditions other than hypoxia proved to regulate the basal activity of HIF-1, which provides prosurvival adaptation to a variety of cells (6).
In this study, we showed that induction of HCV protein expression enhances the amount and the nuclear localization of the HIF-1α subunit and the transcriptional level of glycolytic enzymes, as well as of VEGF (Fig. (Fig.3).3). Comparable results, on a qualitative basis, were obtained in a completely different cell culture system fully competent in replicating the HCV genome and producing infectious virus (Fig. (Fig.55 and and6A)6A) and, more importantly, in liver biopsy specimens from patients affected by HCV infection (Fig. (Fig.11).11). This last observation, which to our knowledge has not previously been reported in HCV-infected individuals, might be linked to the results of a study showing hypoxia-independent overexpression of HIF-1α as an early change in a murine model of chemically induced hepatocarcinogenesis (45).
Some differences were observed among the HCV protein-expressing cell samples examined in this study. In particular, the transfected Huh-7.5 cells and liver biopsy specimens showed significant upregulation of HIF-1α and VHL at the transcriptional level, whereas the level of LDH-A did not change (Fig. (Fig.5B5B and and11).11). In contrast, the UHCVcon-57.3 cells expressing the entire HCV polyprotein displayed significant transcriptional upregulation of LDH-A, HK1, and VEGF, whereas HIF-1α and VHL remained unaltered (Fig. (Fig.3B).3B). A diverse phenotypic background may reasonably account for the observed differences. Concerning the difference in LDH-A transcript levels, the possibility that the control of its expression is not exclusively dependent on HIF-1 must be considered. Indeed, its interaction with a cyclic AMP response element on the LDH-A promoter is needed to achieve maximal trans-activation efficiency (6). Although HIF-1α was initially reported to be constitutively expressed, recent evidence indicates that its mRNA transcription is under the control of the PI3K/Akt pathway (9), as well as of NF-κB (48) and the chromatin-remodeling complex SWI/SNF (14). It is therefore plausible that in Huh-7.5 cells and liver, HCV infection may tissue specifically foster the PI3K/Akt and/or NF-κB signaling pathway, leading to upregulation of HIF-1α. The increased levels of the VHL and PHD2 mRNAs might be consistent with a described feedback mechanism whereby HIF-1α positively controls the expression of VHL and PHD2 by binding to their promoters (8, 13). This mechanism, which under normal conditions negatively self-controls the basal activation of HIF, failed in HCV protein-expressing cells. The last argument, together with the evidence provided here that activation of HIF was also observed in another cell type inducibly expressing HCV proteins in the absence of change in the HIF-1α transcript level, strongly argues in favor of HIF stabilization resulting from impaired posttranslationally linked degradation rather than from enhanced transcription.
The absence of differences in the regulation of the transcription of HIF-controlled genes between cells expressing either the entire HCV polyprotein or subgenomic constructs comprising either the structural or the nonstructural proteins excludes direct interaction of components of the HIF system with a single specific HCV protein (Fig. (Fig.3B).3B). On the other hand, overexpression of an unrelated HCV protein in the inducible cell system did not cause effects on either HIF or mitochondria (Fig. (Fig.4).4). All these observations argue that the HCV-mediated activation of HIF-1 via stabilization of its subunit HIF-1α occurs via different HCV proteins or an indirect mechanism.
In a recent report, it was shown that HCV-infected Huh-7 cells release angiogenic factors as a consequence of normoxic HIF-1α stabilization (29). It was further demonstrated that the stabilization of HIF-1α was mediated by oxidative stress induced by HCV gene expression. In the present study, however, the treatment of both HCV-inducible and -transfected cell lines with antioxidants proved to be inefficient in preventing the overexpression of HIF-controlled genes (Fig. (Fig.77 and and8).8). The reasons for this discrepancy are not clear but may rely on the different tumor-derived cell hosts used (U-2 OS and Huh-7.5 in this study versus Huh-7 ) and/or on the features of the in vitro HCV protein/RNA expression/replication (HCV-inducible and -transfected cells in this study versus HCV-infected cells ). Therefore, in our cell culture systems, factors other than ROS apparently cause stabilization of HIF-1α. We showed that pharmacological inhibition of the FoF1 ATP synthase induces HIF-1α stabilization and that treatment of both UHCVcon-57.3 cells and transfected Huh-7.5 cells with an excess of pyruvate, even in the absence of inhibitors of the OXPHOS, is in itself sufficient to stabilize HIF-1α (Fig. (Fig.99 and and10).10). The last evidence strongly suggests that the accumulation of metabolic intermediates, as a consequence of deregulation of the mitochondrial terminal metabolism, is likely to mediate activation of HIF-1.
PHDs play fundamental roles in the mechanism of HIF activation (37). They are nonheme Fe(II)- and 2-oxoglutarate (2-OG)-dependent dioxygenases. During catalysis, the splitting of molecular oxygen is coupled to hydroxylation of HIF-1α and oxidative decarboxylation of 2-OG to yield succinate and CO2 (10). It has been reported that carboxylic acids and 2-oxo acids, like succinate, oxalacetate, and pyruvate, are efficient competitors of 2-OG with Kis comparable to their physiological concentrations (15, 16, 24). This led us to propose for these Krebs cycle and glycolytic intermediates a role of modulators of PHD activity, thus establishing a link between mitochondrial dysfunction involving the OXPHOS system and activation of HIF-1. It is worth noting that in the case of the 2-oxo acids a mechanism of inhibition that is not simply competitive has been described, in which inactivation of the dioxygenases is accomplished by oxidation of the active metal center (15). Rereduction of the Fe(III) center (and consequent recovery of the PHD activity) specifically requires ascorbate (42). The maintenance of a reduced pool of ascorbate needs reduced glutathione (22), which is the main intracellular redox buffer. Depletion of glutathione to counteract a chronic oxidative insult, such as that induced by HCV protein expression (1, 33), is expected to affect the recovery of PHD activity.
Figure Figure1212 illustrates a plausible pathogenic mechanism integrating the results presented here with evidence reported in the literature. It shows that the HCV-related calcium-mediated depression of mitochondrial OXPHOS causes accumulation of Krebs cycle and glycolytic intermediate metabolites. These may inhibit PHDs, allowing HIF-1α to escape degradation. Inactivated/oxidized PHDs can recover full activity by ascorbate-dependent rereduction of their catalytic metal center. In a pro-oxidant condition, like that set by HCV infection, the availability of ascorbate decreases, and this progressively enhances the life span of the PHD inactivation. The accumulating HIF-1α translocates into the nucleus, where, upon binding to its cognate, HIF-1β, it forms the active heterodimer transcription factor. This positively controls the transcription of genes coding for the glycolytic enzymes, which promote aerobic glycolysis, thereby producing pyruvate. Accumulation of pyruvate might be also favored by a limited or unaffected overexpression of LDH. Thus, a positive feed-forward mechanism is, in turn, fueled. The model foresees that ROS may indirectly upregulate the expression of HIF-1α itself by activating a PI3K/Akt- and/or NF-κB-linked signaling pathway(s) (18, 30).
HIF controls the expression of a variety of antiapoptotic and cell growth factors, thereby establishing a prosurvival condition. This sequence of events may turn out to be favorable to HCV, which can proceed through its replicative and viral-particle assembly phases in a bioenergetically competent cell host. The importance of providing ATP buffering to sustain viral replication has recently been shown (11). In this study, interaction of creatine kinase B, a key ATP-generating enzyme that regulates ATP in subcellular compartments of nonmuscle cells, with NS3-4A was found to be important for efficient replication of the HCV genome and propagation of infectious virus. Noticeably, it was reported that the hepatitis B virus X protein enhances the transcriptional activity of HIF-1α through activation of the mitogen-activated protein kinase pathway (52). Thus, upregulation of HIF might be part of a more general viral strategy established during infection. On the other hand, the growth and proliferative advantage acquired by the infected cell may facilitate its clonogenic expansion, which, under the mutagenic pressure of ROS, may eventually lead to carcinogenic cell transformation (12, 31, 44).
Manipulation of HIF-1 activity by genetic or pharmacological means has been shown to markedly affect tumor growth (42). Accordingly, the results of the present study suggest that direct or indirect inhibitors of HIF-1 might represent promising components of novel combination therapies to prevent or treat HCC in patients with CHC.
We are grateful to Ralf Bartenschlager and Charles M. Rice for kindly providing the Jc1-GFP construct and Huh-7.5 cells, respectively.
The work was supported by the University of Foggia (Local Research Funds 2007-2008), Fondazione Banca del Monte Domenico Siniscalco Ceci-Foggia-Italy, the Swiss National Science Foundation (3100AO-122447), the Swiss Cancer League (OCS-01762-08-2005), and the Leenaards Foundation.
Published ahead of print on 21 October 2009.
†Supplemental material for this article may be found at http://jvi.asm.org/.