Curcumin inhibits hepatitis B virus via down-regulation of the metabolic coactivator PGC-1α

Abstract

Hepatitis B virus (HBV) infects the liver and uses its cell host for gene expression and propagation. Therefore, targeting host factors essential for HBV gene expression is a potential anti-viral strategy. Here we show that treating HBV expressing cells with the natural phenolic compound curcumin inhibits HBV gene expression and replication. This inhibition is mediated via down-regulation of PGC-1α, a starvation-induced protein that initiates the gluconeogenesis cascade and that has been shown to robustly coactivate HBV transcription. We suggest curcumin as a host targeted therapy for HBV infection that may complement current virus-specific therapies.

1 Introduction

Hepatitis B virus (HBV) is a small DNA virus infecting the liver almost exclusively [1]. Chronic infection with HBV might lead to severe liver pathologies including chronic hepatitis, cirrhosis and a fatal cancer designated hepatocellular carcinoma [2]. HBV-induced liver disease is directly correlated with the level of viral activity, making an efficient anti-viral therapy an extremely important component in the management and surveillance of chronically infected patients [3]. The current anti-HBV drugs of choice are members of the nucleotide/nucleoside analogues, which suppress HBV through inhibition of its polymerase/reverse-transcriptase activity. The virus specificity of these efficient agents is a double edge sword, since although almost devoted of unwanted adverse effects, the prolonged use of polymerase inhibitors carry a substantial risk for emerging of escape mutants and a potential flare-up of the disease. Indeed, the inability of the reverse-transcriptase inhibitors to totally eliminate the HBV cccDNA pool usually necessitates a prolonged or life-long therapy [4]. Therefore, an ideal therapy should be the combination of both virus-specific and host-directed therapies.

Recently, we have shown that the metabolic regulator PGC-1α, a coactivator of key gluconeogenesis genes [5, 6], robustly coactivates the transcription of HBV through the nuclear receptor HNF4α [7] and the forkhead transcription factor FOXO1 [8]. Physiologically, we have shown that under starvation when PGC-1α is induced HBV expression is dramatically enhanced, an effect that is largely reversible upon re-feeding. Notably, knocking-down PGC-1α under both fed and starvation states results in a significant down-regulation of HBV expression [7]. Therefore, as a “metabolovirus” that is largely dependent on the metabolic regulator PGC-1α for its gene expression [9], HBV is extremely susceptible to any manipulation in PGC-1α level. Thus, targeting PGC-1α is a potential anti-HBV strategy.

The natural phenolic compound curcumin has been shown to have anti-inflammatory, anti-oxidant and anti-proliferative effects on cells, thereby profoundly affecting their metabolism and proliferative potential [10, 11]. Furthermore, curcumin has been shown to have an inhibitory effect against a variety of organisms such as bacteria, fungi and viruses, including HBV, HCV and HIV [12-15]. The mechanism by which curcumin mediates its anti-viral activity differs from virus to virus and may involve a direct inhibition of the viral replication machinery, such as in the case of HIV [14], or inhibition of a cellular signaling pathway essential for viral replication, such as in the case of HCV [13]. Interestingly, in addition to its ability to inhibit cell signaling at multiple levels and to affect cellular enzymes, curcumin has been shown to specifically target cellular proteins and promote their degradation, as was recently shown with p53 [16]. This degradation is mediated through a unique “degradation by default” pathway common to short-lived regulatory proteins that are intrinsically disordered in their structure [17, 18]. As a short-lived regulatory protein [19] with a predicted disordered structure (unpublished data) we speculated that the HBV coactivator PGC-1α is susceptible to curcumin-induced protein degradation.

Therefore, we examined the potential inhibitory effect of curcumin on HBV expression and replication and further investigated whether curcumin-induced PGC-1α protein degradation is mechanistically involved in HBV inhibition.

2 Materials and methods

2.1 Cell culture and treatments

HepG2, HepG2215 and HEK293 cells were maintained in Dulbecco's modified Eagle's minimal essential medium as previously described [20]. Cells were seeded at about 60% confluence 18–24 h prior to transfection, which was carried out by the CaHPO₄ method as previously described [20]. For hormonal treatments cells were maintained in serum-free DMEM medium and subsequently were treated with 10–100 μM of forskolin (Sigma, catalogue number F6886) as previously described [7] with or without 50–150 μM of curcumin (dissolved in EtOH) for 4 h as previously described [16]. Cyclohexamide (60 μM, Sigma catalogue number 01810) was added to cells for up to 70 min for translation inhibition experiments. Lamivudine (1 mM, GlaxoSmithKline, Australia) was added to cells for 4 h for viral inhibition. Cell viability was assessed by trypan blue exclusion. Briefly, cells were counted in a hemocytometer following 1–2 min incubation with 1:1 0.4% trypan blue. Counted number of unstained cells represents the percentage of viable cells.

2.2 Plasmid constructs

For 1.3× HBV construction, an over-length HBV genome (adw strain) of 4195 bp was produced, harboring a 5′ terminus of the unique EcoRV site (nt 1043, considering EcoRI unique site in the original 3.2 kb HBV construct as nt number 1) and a 3′ terminus of the unique Taq1 site (nt 2017). This EcoRV-TaqI fragment was inserted between the SmaI-AccI unique sites of a pGEM-3Z plasmid, respectively. The 1.3× HBV-Luc plasmid has been previously described [7]. The pSG5-PGC-1α plasmid is a gift from B.M Spiegelman (Dana-Farber, Harvard).

2.3 Protein analysis

Proteins were extracted from cells by RIPA extraction according to published protocols, and subsequently fractioned on 7% (for pgc1) or 12.5% (for HBV Core) SDS–PAGE. For western blot analysis, gels were electro-blotted to a nitrocellulose membrane, which was later soaked for 1 h on a blocking solution [Phosphate buffer saline (PBS) containing 5% non-fat milk and 0.01%v/v tween-20 (Sigma)], and incubated for 1–2 h at RT in the presence of either one of the following antibodies: monoclonal mouse anti HBcAg (clone 22, diluted 1:5000), mouse anti pgc1 antibody (Calbiochem #ST1202, diluted 1:2000), mouse anti HA antibody (Covance #MMS-101R, diluted 1:2000), and mouse anti actin antibody (Santa Cruz sc47778, diluted 1:5000). After incubation, the membrane was washed three times, and Goat anti mouse conjugated with horse radish peroxidase (Jackson #115-035-062, diluted 1:5000) was added and incubation was allowed to proceed for an additional 1 h. Antibody-antigen complexes were visualized by ECL on an X-ray film.

2.4 ELISA analysis

Medium from HBV stably-transfected HepG2215 cells was collected and analyzed for HBsAg level using the ADVIA Centaur machine (Bayer HealthCare LLC, Tarrytown, NY) with HBsAg reagent kit (ref#03393362).

2.5 RNA analysis

RNA was extracted from cells using EZ-RNA isolation kit (Biological Industries, Israel, Beit Haemek) according to the manufacturer's protocol. After treatment with RNase-free DNaseI (Promega #M610A), RNA was subjected to quantitative RT-PCR analysis using the following primers:

HBV FW: TGTGGATTCGCACTCCTCCAGC;HBV Rev: TGCGAGGCGAGGGAGTTCTT;

HPRT1 (h/m) FW: TGACACTGGCAAAACAATGCA;HPRT1 (h/m) Rev: GGTCCTTTTCACCAGCAAGCT.

3 Results and discussion

3.1 Curcumin inhibits HBV expression and replication in stably-transfected hepatoma cells

We first investigated the inhibitory effect of curcumin on HBV in a stably-transfected HepG2215 cell line that continuously expresses HBV [21] and that more authentically simulates a HBV-infected liver. We quantified the secreted HBV surface antigen (HBsAg) level from the medium of HepG2215 cells as a marker for HBV replication. Cells were treated for 4 h each day for three consecutive days with various concentrations of curcumin. At the end of each daily treatment medium was collected, analyzed for HBsAg level and was subsequently replaced by fresh medium. Cell viability was assessed by the trypan blue exclusion method to rule-out a non-specific toxic effect. As shown in Fig. 1 A, whereas no change in the secreted HBsAg level was detected upon curcumin treatment at the end of the first day of treatment, a significant dose-dependent reduction of up to 65% from baseline in HBsAg levels was detected in curcumin-treated cells at the end of the second and the third days of treatment. Notably, one day after treatment cessation HBsAg levels in curcumin-pre-treated cells continued to decrease and reached their nadir level (up to 73% reduction in HBsAg level as compared to non-treated cells). However, on the second post-treatment day an increase in HBsAg levels of ∼10% was detected (Fig. 1A), suggesting that the anti-HBV effect in HepG2215 cells is curcumin-dependent.

Figure 1

Curcumin suppresses both HBV and PGC-1α proteins. (A) HepG2215 cells were either left untreated or alternatively treated with increasing amounts (100 and 150 μM) of curcumin for 4 h each day for three consecutive days after which all cells were left untreated for additional 2 days. Medium from each day was collected and analyzed for HBsAg level. HBsAg level in medium of untreated cells from each of the five experiment days was considered as 100% HBsAg level. The results are an average ± S.D of three independent experiments, each performed in triplicates. (B) HepG2215 cells were either left untreated or alternatively treated with increasing amounts (100 and 150 μM) of curcumin for 4 h, after which cells were harvested, protein extracted and analyzed by a Western-blot. The relative intensities (RI) of the bands after normalization to actin levels are shown below the corresponding panels.

Compatible with these results, protein analysis of curcumin-treated HepG2215 cells revealed a significant dose-dependent decrease of up to 45% from baseline in HBV Core protein level (Fig. 1B). Interestingly, analysis of PGC-1α protein in those cells revealed a substantial level of PGC-1α at baseline that steadily decreased with curcumin treatment paralleling the decrease in HBV Core protein level (Fig. 1B).

Overall, these results strongly suggest that curcumin inhibits HBV and that this inhibition correlates with a significant decrease in PGC-1α protein level.

3.2 Curcumin treatment results in suppression of HBV expression in a PGC-1α dependent manner

We have previously shown that PGC-1α strongly coactivates HBV transcription and that during physiological conditions in which hepatic PGC-1α is induced, HBV gene expression is dramatically enhanced [7]. Based on the results presented in Fig. 1B, showing a correlation between the reduction in HBV Core protein and PGC-1α level, we asked whether curcumin suppresses HBV expression through inhibition of PGC-1α. For this, we employed the HBV-Luc construct in which the luciferase ORF is cloned downstream to HBV enhancer II and to the core promoter (Fig. 2 A, upper panel) [7]. As expected, over-expression of PGC-1α in HBV-Luc transfected HepG2 cells resulted in a significant induction of HBV transcription as reflected by increased luciferase activity. Notably, whereas curcumin treatment resulted in a modest but still a significant decrease in HBV-Luc activity under basal conditions in which PGC-1α level is low, it largely abrogated HBV-Luc induction upon PGC-1α over-expression (Fig. 2A, lower panel). These results suggest that curcumin suppresses HBV through inhibition of PGC-1α.

Figure 2

Curcumin suppresses HBV expression through inhibition of PGC-1α. (A) (Upper panel) a schematic presentation of the HBV-Luc construct used in this experiment. (Lower panel) HepG2 cells were transfected with the indicated plasmids. Two days after transfection cells were treated with curcumin (50 μM for 4 h) after which cells were harvested and analyzed for luciferase activity. Results were normalized to β Gal activity. The results are an average of three independent experiments, each performed in triplicates (^∗ P = 0.03, calculated by Student t-test). (B) HepG2 cells were transfected with the indicated plasmids. Two days later treatment with curcumin (50 μM for 4 h) was performed as indicated, after which cells were harvested, RNA was extracted and subjected to semi-quantitative (upper panel) and quantitative (lower panel) RT-PCR analysis. (C) The same experimental procedure as in B was performed, this time cells were harvested for protein analysis as indicated. (D) HepG2 cells were transfected as indicated. Two days later cells were treated as indicated (Forsk 10 μM, Curc 50 μM for 4 h) after which cells were harvested, protein extracted and analyzed by a Western-blot.

Next, we analyzed both HBV mRNA and HBV Core protein levels under basal conditions and upon PGC-1α over-expression. As expected, and consistent with the luciferase experiment results, PGC-1α over-expression resulted in up-regulation of HBV expression at both the mRNA (Fig. 2B) and at the protein (Fig. 2C) levels. This induction was completely abrogated upon curcumin treatment and was associated with a sharp decrease in PGC-1α protein level (Fig. 2C).

Hepatic PGC-1α is robustly induced upon starvation [6]. To investigate the inhibitory effect of curcumin on the endogenous PGC-1α level and on HBV expression under conditions mimicking starvation, HBV-transfected HepG2 cells were pre-treated with forskolin and were subsequently treated with curcumin. As shown in Fig. 2D, both PGC-1α and HBV Core protein levels were induced upon forskolin treatment, an induction largely abrogated by curcumin treatment. Notably, the level of a control GFP protein remained unchanged, ruling out a non-specific degradation or translational inhibition effect. Noteworthy is the observation that even under basal conditions curcumin treatment suppressed HBV expression, although to a much lower extent (Fig. 2A–D). This modest effect correlates well with the low level of hepatic PGC-1α under basal conditions (Fig. 2C and D) in which the gluconeogenesis cascade is not activated.

3.2.1 Curcumin inhibition of PGC-1α is at the protein level

We next investigated the mechanism by which curcumin inhibits PGC-1α. HEK293 cells were transfected with a PGC-1α expression plasmid and were subsequently treated with increasing amounts of curcumin. A co-transfected GFP expression plasmid was used as a control. A Western-blot analysis revealed that whereas GFP protein level remained stable, curcumin treatment resulted in a significant reduction in PGC-1α protein level in a dose-dependent manner (Fig. 3 A left panel). Curcumin effect on PGC-1α is at the protein level, since no change in PGC-1α mRNA level was detected upon curcumin treatment (Fig. 3A right panel). Based on its predicted intrinsically disordered structure (unpublished data), we speculated that PGC-1α protein is susceptible to curcumin-induced protein degradation, as was recently shown with p53 [16]. Therefore, we further investigated the effect of curcumin on PGC-1α protein stability. For this, cells were transfected with a PGC-1α expression plasmid, and were subsequently either treated with curcumin or left untreated. As shown in Fig. 3B, following treatment with the translation inhibitor cyclohexamide PGC-1α protein level decreased much more rapidly in curcumin pre-treated cells as compared to non-treated cells, indicating that curcumin treatment results in a significant shortening of PGC-1α protein half-life. These results strongly suggest that curcumin enhances PGC-1α protein degradation.

Figure 3

PGC-1α protein is degraded by curcumin. (A) (Left panel) HEK293 cells were transfected with the indicated plasmids. Two days after transfection cells were either left untreated or treated with increasing amounts of curcumin (20 and 50 μM) for 4 h after which cells were harvested, protein extracted and analyzed by western blot. (Right panel) the same experiment protocol, this time RNA was extracted and analyzed by semi-quantitative RT-PCR. (B). HepG2 cells transfected with a PGC-1α expression plasmid were either pre-treated with curcumin (50 μM for 4 h) or left untreated, after which 60 μM of cyclohexamide was added for the indicated times. Cells were harvested, protein extracted and analyzed by a Western-blot for PGC-1α protein level. A representative blot is shown. (C). HepG2 cells were treated with forskolin (10 μM) and dexamethasone (100 nM). Four hours after treatment, curcumin (100 μM) was added to the medium for additional 4 h. Eight hours after treatment initiation all cells were harvested, RNA extracted and subjected to RT-PCR analysis. The resulting cDNA was further amplified using real-time PCR with the indicated primers. Results were normalized to actin mRNA. The results are the average of three independent experiments (* indicates P = 0.006, calculated by student's t-test).

Next, we asked whether curcumin-induced PGC-1α protein degradation also translates to a decrease in the expression of the “classical” PGC-1α target genes. For this, we pre-treated HepG2 cells with dexamethasone and forskolin, two known inducers of PGC-1α in the liver that simulate a starvation state [6, 7], and subsequently treated those cells with curcumin. We analyzed the mRNA level of the key gluconeogenesis gene G6Pase, a well-known target of PGC-1α coactivation [6, 22]. As expected, the mRNA levels of both PGC-1α and G6Pase were significantly induced by the combination of dexamethasone and forskolin treatment. However, whereas PGC-1α mRNA level remained unchanged upon curcumin treatment, the mRNA level of its target gene G6Pase was significantly reduced (Fig. 3C), suggesting that curcumin inhibits PGC-1α at the protein level and that this inhibition results in a reduced expression of its target genes.

3.3 Curcumin and the reverse-transcriptase inhibitor, Lamivudine, synergistically suppress HBV expression

So far our results indicate that curcumin inhibits HBV through promoting the degradation of its coactivator PGC-1α. We next asked whether curcumin treatment could complement the anti-viral activity of the nucleotide/nucleoside analogues, which are considered as the gold standard for anti-HBV therapy. For this, HBV expressing HepG2215 cells were treated with either Lamivudine alone, curcumin alone or with the combination of both. As shown in Fig. 4A , a 4 h treatment with either Lamivudine or curcumin resulted in a significant suppression of HBV transcription by ∼35% and ∼62%, respectively. However, the combination of both treatments resulted in an enhanced suppression of HBV expression by up to 75%, as compared to non-treated cells. These results suggest that curcumin may work synergistically with the current anti-HBV nucleotide/nucleoside analogous, and that this combination may result in a better suppression of HBV.

Figure 4

Curcumin as an additive host-directed anti-HBV therapy. (A) HepG2215 cells were left either untreated or alternatively treated for 4 h with either Lamivudine (1 mM), curcumin (150 μM) or the combination of both. Four hour after treatment initiation cells were harvested and analyzed for HBV mRNA level using quantitative RT-PCR. Results are presented after normalization to Hypoxanthine-guanine phosphoribosyltransferase 1 (HPRT1) level as a mean ± S.D. of three experiments, each performed in triplicates. The P value was calculated using a standard student t-test. (B) A schematic presentation of HBV life cycle and the targets for conventional anti-viral (i.e. nucleotide/nucleoside analogues) and the potential host-directed curcumin therapy.

Taken together, in this study we show that PGC-1α protein is readily degraded by the natural phenolic compound curcumin. Mechanistically, we show that curcumin promotes the degradation of PGC-1α protein and significantly shortens its half-life (Fig. 3). We speculate that as a short-lived regulatory protein with predicted intrinsically disordered structure (unpublished data), PGC-1α joins a growing group of proteins called IDPs (intrinsically disordered proteins) that are rendered to degradation upon curcumin treatment [16, 18, 23]. However, although our study clearly shows that curcumin promotes the degradation of PGC-1α, the exact molecular pathway of this degradation is a subject for further studies.

Taking advantage of the dependency of HBV on its coactivator PGC-1α [7], we show that inhibition of PGC-1α by curcumin treatment results in a significant suppression of HBV gene expression and replication markers. Indeed, the inhibitory effect of curcumin on HBV expression is seen mainly under conditions in which hepatic PGC-1α is over-expressed or induced, such as upon forskolin treatment mimicking starvation. A much more modest effect, although still substantial, is seen under basal conditions in which hepatic PGC-1α levels is relatively low.

Thus, this study joins previous works and provides additional evidence for the potential importance of nutrition among HBV-infected patients [7, 9]. In addition, as a “metabolovirus” that is profoundly affected by nutritional cues and by hepatic metabolic pathways [7, 9], HBV is potentially susceptible to manipulations of key molecular players in hepatic metabolic processes such as gluconeogenesis.

Therefore, we suggest curcumin as a potential host-directed therapy that may complement the current nucleotide/nucleoside analogues that directly suppress HBV replication in a virus-specific manner (Fig. 4B). As shown experimentally in this study, this combination therapy synergistically suppresses HBV (Fig. 4A). Furthermore, the addition of curcumin to the standard treatment of nucleotide/nucleoside analogues may minimize the risk of viral flare-ups under physiological conditions, such as during a short-term starvation, in which hepatic PGC-1α is induced to coactivate HBV transcription [7].

Obviously, some major obstacles have to be overcome prior to the clinical use of curcumin as an anti-HBV drug in humans. First, whereas host-directed anti-viral therapy has the advantage of avoiding viral-resistance, it carries a substantial risk for unwanted adverse effects. Accordingly, targeting a cellular coactivator involved in metabolic and energy-related processes such as PGC-1α [5, 6] might result in clinically significant adverse outcomes. Indeed, following curcumin treatment a modest but still significant down-regulation of the key gluconeogenic enzyme G6Pase was observed (Fig. 3C). Therefore, further studies in animals and humans should be carried out to adjust curcumin dosages that are minimally toxic on the one hand, but still result in a substantial impairment of HBV gene expression and replication on the other hand.

Second, the clinical use of curcumin is hampered by its low solubility in water, its short half-life and its low bioavailability following oral administration [24, 25]. Therefore, the reduction in HBV expression observed in our in vitro studies might be insufficient in vivo in terms of efficient viral suppression. This obstacle can be overcome by adopting one of the emerging techniques to increase curcumin bioavailability, such as its oral co-administration with the alkaloid piperine [26], its liposomal encapsulation for intravenous administration [27] or using curcumin nanoparticles that render curcumin completely dispersible in aqueous media [28]. However, although the feasibility of these techniques in the context of curcumin-induced HBV inhibition is a subject of further studies, it holds promise for future use of the “golden spice” curcumin as an efficient anti-HBV therapy that may replace, or at least complement the “conventional” anti-viral drugs.

Acknowledgment

This work was supported by the following Grants: and Grant No. 2007285 from the United States-Israel Binational Science Foundation (BSF).