Activation of ATP Citrate Lyase by mTOR Signal Induces Disturbed Lipid Metabolism in Hepatitis B Virus Pre-S2 Mutant Tumorigenesis

ABSTRACT

The development of hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC) has been found to be associated with disturbed lipid metabolism. To elucidate the role of lipid metabolism in HBV tumorigenesis, we investigated the dynamic pattern of lipid metabolism in HBV pre-S2 mutant-induced tumorigenesis. Lipid and gene expression profiles were analyzed in an in vitro culture system and in transgenic mouse livers harboring HBV pre-S2 mutant. The pre-S2 mutant transgenic livers showed a biphasic pattern of lipid accumulation, starting from mild fatty change in early (1 month) transgenic livers, which subsided and then, remarkably, increased in HCC tissues. This biphasic pattern was synchronized with ATP citrate lyase (ACLY) activation. Further analyses revealed that the pre-S2 mutant initiated an endoplasmic reticulum (ER) stress-dependent mammalian target of rapamycin (mTOR) signalling cascade. The pre-S2 mutant-induced mTOR signal activated the sterol regulatory element binding transcription factor 1 (SREBF1) to upregulate ACLY, which then activated the fatty acid desaturase 2 (FADS2), mediated through ACLY-dependent histone acetylation. Such an ER stress-dependent mTOR signal cascade also is important for the proliferation of hepatocytes in vitro and is further validated in HBV-related HCC tissues.

IMPORTANCE Aberrations of lipid metabolism frequently occur in chronic HBV infection. Our results provide a potential mechanism of disturbed lipid metabolism in HBV pre-S2 mutant-induced tumorigenesis, which should be valuable for the design of HCC chemoprevention in high-risk HBV carriers.

INTRODUCTION

Chronic hepatitis B virus (HBV) infection is a major risk factor for hepatocellular carcinoma (HCC) (1). Several theories have been proposed to explain the mechanism of HBV hepatocarcinogenesis, including immune-mediated tumorigenesis and viral protein-driven tumorigenesis (2, 3). However, the underlying mechanism of HBV-associated tumorigenesis remains to be elucidated. Previously, we demonstrated that type II ground glass hepatocytes (GGHs) contain exclusively HBV pre-S2 deletion mutant large surface antigen (pre-S2 mutant), which is an immune escape mutant (4, 5). The retention of pre-S2 mutant in the endoplasmic reticulum (ER) can induce ER stress and oxidative DNA damage, as well as exhibit transforming capabilities (5, 6). Transgenic mice harboring the pre-S2 mutant can induce nodular dysplasia and HCC (7). Hence, type II GGHs have been identified as a preneoplastic lesion of HBV-related HCC, and pre-S2 mutant has been recognized as a potential viral oncoprotein (8, 9).

Growing evidence indicates that cancer cells show specific alterations in lipid metabolisms that are important for cell growth and proliferation (10, 11). Recently, HCC has been linked to nonalcoholic fatty liver, obesity, and related metabolic disorders (12). Numerous reports have uncovered aberrant lipidomic profiles in human HCCs and mouse HCC models (13, 14). Furthermore, aberrations of lipid metabolism often are seen in chronic HBV infection (15). However, the contributing role of disturbed lipid biosynthesis in HBV tumorigenesis remains to be clarified.

The mammalian target of rapamycin (mTOR) is a serine/threonine kinase that has an important role in regulating cell growth (16). In addition to promoting protein synthesis, it is now becoming clear that mTOR also controls the synthesis of lipids required for proliferating cells to generate membranes (17). To a large extent, mTOR promotes de novo lipogenesis through inducing the cleavage of the sterol regulatory element binding transcription factor 1 (SREBF1), which then translocates to the nucleus and induces the expression of many lipogenic genes (18). Previously, we demonstrated that HBV pre-S2 mutant can activate mTOR through the induction of ER stress-dependent vascular endothelial growth factor A (VEGFA)/AKT signaling in GGHs to promote tumorigenesis (19, 20). Therefore, this study was designed to investigate whether the pre-S2 mutant could promote lipid biosynthesis through mTOR activation during the process of HBV-related tumorigenesis. We demonstrated that the pre-S2 mutant transgenic mice exhibited increased lipid accumulation in HCC tissues. The activation of mTOR by the pre-S2 mutant initiated the SREBF1/ATP citrate lyase (ACLY) signaling cascade, which led to increased de novo lipogenesis and proliferation abilities of HCC cells in vitro. Such an mTOR-dependent lipogenic pathway also was validated in human HBV-related HCC tissues.

MATERIALS AND METHODS

Transgenic mice.

The transgenic mice expressing the HBV pre-S2 mutant in the liver were established by Ting-Fen Tsai's laboratory as described previously, with minor modifications (21). Briefly, the pre-S2 mutant transgenic mice were generated in the C57BL/6 background. The pre-S2 mutant with in-frame deletions over the pre-S2 region (deletion of nucleotides 4 to 57; Δ2) of the large surface gene and with a point mutation at the start codon (ATG-ATA) of the middle S gene was driven by the liver-specific albumin promoter. All of the animal experiments were performed in male mice under the approval of the institutional animal care and use committees of National Cheng Kung University College of Medicine and the National Health Research Institutes.

Histopathology and IHC studies.

For histopathological examination, paraffin-embedded liver sections were stained with hematoxylin-eosin (HE). For immunohistochemistry (IHC) staining, the sections were incubated with anti-HBsAg (HBV surface antigen) (Dako, San Ramon, CA) and assayed as described previously (7). Lipid deposition in livers was visualized by oil red O staining of cryostat frozen sections as described previously (22).

cDNA microarray analysis.

The microarray experiment and data analysis were performed by Welgene Biotech (Taipei, Taiwan) using the Agilent mouse whole-genome oligonucleotide microarray chips (4×44K; Agilent Technologies, Santa Clara, CA).

Real-time PCR.

Total RNA was extracted and converted to cDNA as described previously (23). Real-time PCR was performed with the LightCycler system and the mouse Universal ProbeLibrary system (Roche Applied Science, Indianapolis, IN). The primers and Universal ProbeLibrary probes used are shown in Table S1 in the supplemental material.

Plasmid, cell lines, and transient transfection.

The plasmid pIRES-Δ2, expressing hemagglutinin (HA)-tagged pre-S2 mutant, was established as described previously (7). The FADS2 promoter reporter plasmid was constructed by inserting the promoter fragment into the pG5luc vector (Promega, Madison, WI), followed by inserting a Renilla luciferase expression cassette, which was generated from the pRL-TK vector (Promega). The FADS2 promoter region was amplified by PCR with primers shown in Table S2 in the supplemental material. Rapamycin was purchased from Calbiochem (San Diego, CA) and used at a final concentration of 100 nM. Trichostatin A was purchased from Sigma (St. Louis, MO) and used at a final concentration of 500 nM. All short interfering RNAs (siRNAs) were obtained as ON-TARGETplus SMARTpool reagents (a mixture of 4 individual siRNAs) (Dharmacon, Lafayette, CO) and used at a final concentration of 50 nM. All transfections were performed in human HuH-7 and HepG2 hepatoma cell lines with a MicroPorator (Invitrogen Life Technologies, Carlsbad, CA) by following the manufacturer's instructions. Cell proliferation was measured by MTT assay using a cell counting kit 8 (Sigma) by following the manufacturer's instructions.

Western blot analysis.

Western blot analysis was performed as previously described (23). The primary antibodies used in this study were anti-pmTOR (Ser2448), anti-SREBF1, anti-FADS2 (Abcam, Cambridge, United Kingdom), anti-ACLY, anti-pACLY (Ser455) (Cell Signaling Technology, Danvers, MA), anti-acetyl-histone H3 (Millipore, Bedford, MA), anti-HA tag (Zymed Laboratories, South San Francisco, CA), and anti-ACTB (actin, beta) (Chemicon, Temecula, CA). The pIRES-Δ2 plasmid expresses large and small surface proteins but not the middle surface proteins because of the mutated start codon at the middle S gene. In pre-S2 mutant-transfected cells, two distinct bands with larger molecular sizes could be identified by anti-HA tag antibody, corresponding to the unglycosylated 39-kDa and glycosylated 42-kDa pre-S2 mutant large surface proteins (p39 and gp42). The small surface proteins (p24 and gp27) also were observed but were trimmed for figure editing purposes.

Acetyl-CoA, triglyceride, and cholesterol measurements.

Triglyceride and cholesterol contents in livers and cells were determined using commercial kits from Fortress Diagnostics (London, United Kingdom) by following the manufacturer's instructions. Intracellular acetyl-coenzyme A (CoA) levels were measured using the PicoProbe acetyl-CoA assay kit (Abcam). Each experiment was repeated at least three times.

Luciferase reporter assay.

Luciferase-expressed cells were assayed by the Dual-Luciferase reporter assay system (Promega) according to the manufacturer's instructions. Renilla luciferase activity was measured for normalization.

Human HCC liver tissues.

Freshly frozen human HBV-related HCC tissues were obtained from the Department of Pathology, National Cheng Kung University Hospital, Tainan, Taiwan, from 2007 to 2011 under the approval of the hospital's Institutional Review Board.

Statistical analysis.

The significance of selected biomarkers in transgenic mice and human livers was determined by unpaired and paired t tests, respectively. The in vitro data were analyzed by one-way analysis of variance (ANOVA) with Bonferroni's multiple-comparison posttest (*, P < 0.05; **, P < 0.01; ***, P < 0.001). Each experiment was repeated independently at least three times, and data represent the means and standard deviations (SD).

RESULTS

Pre-S2 mutant transgenic mice exhibited increased lipid accumulation in HCC tissues.

In the pre-S2 mutant transgenic mouse model, male mice developed HCCs at the mean age of 24.5 months with a 12% occurrence rate and expressed HBsAg in the hepatocytes around the typical central vein area (Fig. 1A, panels a, b, and c). The transgenic livers were sampled for histopathologic studies at 1 and 3 months (early stage), 6 months (middle stage), 12 months (advanced stage), and at tumor formation. By oil red O staining, mild and diffused fatty change in hepatocytes was observed in 1-month-old transgenic mice compared to the age-matched nontransgenic mice (Fig. 1A, panels d and e). The staining intensity weakened as the disease progressed (Fig. 1A, panels f and g) and eventually disappeared in the later stages (6 and 12 months; data not shown). Interestingly, large amounts of enlarged lipid droplets were found accumulated in transgenic tumors rather than the surrounding nontumors (Fig. 1A, panels h and i). Furthermore, the levels of triglycerides and cholesterol in transgenic livers were measured and showed consistency with the accumulation pattern found in the oil red O staining (Fig. 1B).

FIG 1.

Pre-S2 mutant transgenic mice exhibited increased lipid accumulation in HCC tissues. (A) Gross view of representative HCC (a), histological HE staining of tumor (b), immunohistochemical detection of HBsAg (c), and oil red O-stained lipid deposits (d to i). Shown are representative results of each stage. Original magnification, ×20. Scale bar, 50 μm. (B) Triglyceride and cholesterol levels in pre-S2 mutant (Δ2) and nontransgenic (N) livers analyzed according to different months (M) and nontumor (NT) or tumor (T) status.

mTOR, SREBF1, and ACLY signals were chronologically activated in pre-S2 mutant transgenic livers and HCCs.

The cDNA microarray data of pre-S2 mutant transgenic livers was adopted to identify the candidates of SREBF1 target lipogenic genes (17). Real-time PCR was performed to confirm the selected genes' transcription levels. Acly was the only gene that showed significant upregulation (≥1.5-fold) in transgenic HCCs compared with the nontransgenic livers (see Fig. S1 in the supplemental material). This finding led us to hypothesize that the pre-S2 mutant regulates lipid metabolism through the activation of the mTOR/SREBF1/ACLY signaling cascade. To test this hypothesis, Western blot analysis was performed to examine the expression of the signaling molecules. As shown in Fig. 2, the expression of the phosphorylated active form of mTOR (pmTOR) was significantly elevated throughout the study period. The precursor form (p) of SREBF1 was upregulated in the middle to late stage, 6 and 12 months, while the cleaved nuclear form (n) was found increased only upon tumor formation. Both total and phosphorylated ACLY exhibited biphasic overexpression at early and tumor stages.

FIG 2.

mTOR, SREBF1, and ACLY signals were chronologically activated in pre-S2 mutant transgenic livers and HCCs. (A) Western blot analysis of the indicated biomarkers in pre-S2 mutant and nontransgenic livers of different ages, as well as paired nontumorous livers (NT) and tumors. Six livers were used in each group except for the tumor-stage group, due to small tumor size and low tumor formation rate. (B) Quantitative results were normalized by age-matched control livers.

Pre-S2 mutant activated ACLY through mTOR/SREBF1 signaling to promote de novo lipogenesis and cell proliferation in HCC cells.

To confirm that the mTOR/SREBF1/ACLY signaling cascade was indeed induced by the pre-S2 mutant, HuH-7 cells were transfected with pre-S2 mutant or control plasmid, and the expression profiles of signaling molecules were analyzed by Western blotting. As shown in Fig. 3A, the expression of all of the examined signaling molecules was increased at 24 and 48 h posttransfection in pre-S2 mutant-transfected cells compared to the control cells. The pre-S2 mutant-induced signaling upregulation was mediated by mTOR, as this effect could be abolished by the mTOR inhibitor rapamycin (Fig. 3B). siRNA-mediated knockdown of SREBF1 also could diminish the elevated signals, even under conditions of mTOR activation. Altogether, the data demonstrated that the pre-S2 mutant could upregulate ACLY through mTOR/SREBF1 signaling. We next evaluated the effect of the mTOR signaling cascade on de novo lipogenesis in HuH-7 cells. As shown in Fig. 3C to E, pre-S2 mutant transfections resulted in elevated levels of acetyl-CoA, triglycerides, and cholesterol compared with control cells, and this effect could be abrogated by the inhibition of mTOR, SREBF1, and ACLY signals. Furthermore, pre-S2 mutant-transfected cells exhibited increased proliferation compared to the control cells (Fig. 3F). Following treatment of the transfected cells with inhibitors or siRNAs, targeting to the signaling molecules significantly slowed cell proliferation. Similar results were obtained with HepG2 cells (see Fig. S2 in the supplemental material). Collectively, these results suggest that the pre-S2 mutant could promote de novo lipogenesis through mTOR-mediated activation of ACLY.

FIG 3.

Pre-S2 mutant activated ACLY through mTOR/SREBF1 signaling to promote de novo lipogenesis and cell proliferation in HuH-7 cells. (A and B) HuH-7 cells were transfected with pre-S2 mutant or control plasmid (Ctrl). After 24 h, cells were left untreated or were treated with rapamycin (Rapa) and SREBF1 siRNA (siSREBF1) for another 24 h and analyzed by Western blotting for the indicated biomarkers. For functional in vitro assays, HuH-7 cells transfected with pre-S2 mutant or control plasmid with or without further treatment were assayed for acetyl-CoA (C), triglycerides (D), cholesterol (E), and cell proliferation (F). Data in each experiment are presented as values relative to those of the untreated control cells.

Pre-S2 mutant upregulated lipogenic enzyme FADS2 through ACLY-dependent histone acetylation in HCC cells.

We further explored the possibility that pre-S2 mutant regulates other lipogenic gene expression through ACLY. As shown in Fig. 4A, the transcript level of several key enzymes involved in triglyceride and cholesterol biosynthesis pathways was assayed by real-time PCR in pre-S2 mutant transgenic HCCs compared with the paired nontumors and age-matched nontransgenic liver tissues. We observed that levels of mevalonate kinase (Mvk), hydroxysteroid (17-beta) dehydrogenase 7 (Hsd17b7), and fatty acid desaturase 2 (Fads2) transcripts were increased in transgenic HCCs. To confirm whether these genes were upregulated by the pre-S2 mutant, HuH-7 cells were transfected with pre-S2 mutant or control plasmid. As shown in Fig. 4B, pre-S2 mutant-transfected cells expressed higher levels of FADS2 transcripts than control cells at 48 h posttransfection. Furthermore, we found that the upregulation of FADS2 by the pre-S2 mutant could be reversed by siRNA-mediated ACLY inhibition (Fig. 4C). The level of histone acetylation was increased in pre-S2 mutant-expressed cells; this could be restored by ACLY siRNA. The effect of ACLY siRNA on FADS2 transcripts and histone acetylation levels could be rescued by the histone deacetylase inhibitor trichostatin A. To confirm that the ACLY-mediated histone acetylation under pre-S2 mutant expression is specific for the FADS2 promoter, we next generated the FADS2 promoter reporter plasmid and performed a luciferase reporter assay. As shown in Fig. 4D, FADS2 promoter-driven luciferase activities were increased in pre-S2 mutant-expressed HuH-7 cells; this upregulation could be abrogated by ACLY inhibition and restored by trichostatin A treatment. Consistent results were observed in HepG2 cells (see Fig. S3 in the supplemental material). Taken together, these data suggest that transcriptional upregulation of FADS2 through ACLY by the pre-S2 mutant is dependent on histone acetylation.

FIG 4.

Pre-S2 mutant upregulated lipogenic enzyme FADS2 through ACLY-dependent histone acetylation in HuH-7 cells. Real-time PCR was performed to analyze transcript levels of selected lipid genes in pre-S2 mutant transgenic livers (A) and pre-S2 mutant-expressed HuH-7 cells (B and C) with or without treatment of ACLY siRNA (siACLY) or tricostatin A (TSA). The level of histone acetylation was detected by Western blotting with anti-acetyl-histone H3 (Ac-H3) as shown in the lower panel. (D) Luciferase activities of FADS2 promoter reporter plasmid were expressed as values relative to that of the control reporter plasmid.

mTOR/SREBF1/ACLY/FADS2 signaling was activated in human HBV-related HCCs.

To ascertain whether mTOR/SREBF1/ACLY/FADS2 signaling is associated with human HBV-related hepatocarcinogenesis, Western blot analysis was performed on 30 pairs of HBV-related HCCs and adjacent nontumorous liver tissues for the expression of components of the signaling pathway. As shown in Fig. 5, we demonstrated that pmTOR, SREBF1 (p), SREBF1 (n), ACLY, pACLY, and FADS2 were consistently and significantly expressed at higher levels in HCCs than in the paired nontumorous livers in 20 of 30 tissue pairs (cases 11 to 30). The data support the essential role of the lipogenesis-associated mTOR signal cascade in HBV tumorigenesis.

FIG 5.

mTOR/SREBF1/ACLY/FADS2 signaling was activated in human HBV-related HCCs. (A) Western blot analysis revealed enhanced expression of the indicated biomarkers in HBV-related HCCs at a level comparable to or even higher than that in the paired nontumorous livers. (B) Data were quantified and statistically analyzed.

DISCUSSION

This study demonstrated, for the first time, the contributing role of the pre-S2 mutant-induced mTOR signalling cascade in lipid metabolic disturbances, the dynamic patterns of lipid accumulation, and the expression of lipogenic genes in HBV tumorigenesis. In this study, we showed that the pre-S2 mutant-induced mTOR activation signaling cascade could not only promote de novo lipogenesis through activation of SREBF1 to upregulate ACLY but also increase cell proliferation, both of which can lead to HBV tumorigenesis.

The mTOR signal was activated throughout the period of the study, while the precursor form of SREBF1 was found upregulated at the middle to advanced stage (6 and 12 months), and the cleaved (activated) nuclear form of SREBF1 was elevated only until HCC developed. Therefore, although our in vitro system demonstrated SREBF1 upregulation by mTOR, the exact temporal relationship between mTOR and the switch of SREBF1 activation in vivo remains to be clarified. One intriguing finding of this study is the biphasic overexpression of ACLY at the early and tumor stages, which is synchronized with the elevated levels of triglyceride and cholesterol as well as the patterns of lipid accumulation in transgenic livers. The perfect synchronization of ACLY, triglyceride, and cholesterol is logical, because ACLY is a cytosolic enzyme that converts mitochondrion-derived citrate into acetyl-CoA, which is a vital building block for the de novo biosynthesis of triglyceride and cholesterol (24). Moreover, the activation of SREBF1 at the tumor stage may contribute to the overexpression of ACLY, because our in vitro data provided evidence for ACLY upregulation by the pre-S2 mutant through mTOR/SREBF1 signaling. Furthermore, our microarray data showed that the elevated Acly mRNA was detected only at the HCC stage (see Fig. S1 in the supplemental material), indicating that the overexpression of ACLY at the tumor stage was regulated at the transcriptional level. Alternatively, it has been suggested that AKT signal can stimulate ACLY activity through the phosphorylation of ACLY, which contributes to its protein stabilization (25). In support of this assumption, our previous study showed that AKT signal activation was found at the early stage of tumorigenesis in pre-S2 mutant transgenic mice and may be a result of ACLY stabilization through AKT signaling, which is independent of SREBF1 activation (26).

In this study, we demonstrated that mTOR/SREBF1/ACLY signaling was activated in pre-S2 mutant transgenic HCCs and human HBV-related HCCs, and its inhibition suppressed lipid biosynthesis and the proliferation of the pre-S2 mutant-expressed HCC cells. The suppression of HCC cell proliferation may be a result of cosuppression of lipid biosynthesis and glucose metabolism. Our previous study has shown that pre-S2 mutant-expressed HCC cells exhibit elevated levels of aerobic glycolysis (unpublished data). It has been reported that the proliferation of cancer cells displaying a high rate of glucose metabolism is more severely affected by ACLY inhibition (27). In addition, ACLY has been shown to play crucial roles in regulating global chromatin architecture and gene transcription in multiple mammalian cells by determining the total amount of histone acetylation (28). ACLY-dependent histone acetylation is thought to contribute to the selective regulation of genes involved in glucose metabolism as well as to the conversion of glucose into nonessential amino acids and fatty acids (28). In this study, we further identified that the pre-S2 mutant could upregulate FADS2, the enzyme that plays an important role in the biosynthetic pathway of triglyceride via catalyzing synthesis of long-chain polyunsaturated fatty acids (29), through ACLY. Given its role as the key enzyme that links glucose metabolism and lipid biosynthesis, ACLY is an ideal candidate for tackling the pre-S2 mutant-induced metabolic disturbances.

HCC studies often are categorized into three major fields: viral hepatitis, alcoholic steatohepatitis, and nonalcoholic steatohepatitis (30). Each field has its own pathogenesis paradigms based on findings that focus on the specific topics. In this study, we enrolled 30 chronic HBV-infected HCC patients and observed the activations of lipogenic genes that were consistent with the in vitro results and the pre-S2 mutant transgenic mouse model. Moreover, aberrant lipid metabolism frequently occurs with chronic HBV infection (15). We crossed the barriers between the fields and connected virus-related hepatitis with metabolic syndromes. In HBV infection, the pre-S2 mutant can stimulate aerobic glycolysis through mTOR activation in an ER stress-dependent pathway (unpublished data). The aerobic glycolysis by-product, citrate, could be converted into acetyl-CoA, the raw material of triglycerides and cholesterol, by ACLY catalysis. Thus, ACLY played an important role in linking the glucose metabolism to the endogenous biosynthesis of triglycerides and cholesterol at the advanced stage of tumorigenesis. Together, our findings suggest that the pre-S2 mutant plays a role in HBV tumorigenesis by disturbing normal metabolism, including elevating glycolysis and promoting lipid biosynthesis through the mTOR signal cascade, as summarized in Fig. 6.

FIG 6.

Schematic model for the upregulation of de novo lipogenesis by pre-S2 mutant in HBV tumorigenesis. Through the induction of ER stress-dependent VEGFA/AKT signaling, pre-S2 mutant activates mTOR and stimulates aerobic glycolysis. In addition, the activated mTOR signal favors the cleavage of SREBF1 at the advanced stage of HCC tumorigenesis. The cleaved form of SREBF1 then translocates to the nucleus, where it induces the expression of ACLY. By converting the mitochondrion-derived citrate into acetyl-CoA, ACLY links glucose metabolism to the de novo biosynthesis of triglycerides and cholesterol and upregulates FADS2 via the increase of histone acetylation, all of which contribute to metabolic disturbances and growth advantages of hepatocytes and HCC development.

In conclusion, this study demonstrates that the activated mTOR signal cascade plays an important role in promoting de novo lipogenesis through the activation of SREBF1 and ACLY signals in HBV-associated tumorigenesis. Targeting the mTOR/SREBF1/ACLY signaling pathway may be a promising therapeutic strategy for HCC therapy.