Potential conflict of interest: Nothing to report.
Additional Supporting Information may be found in the online version of this article.
Phosphatase and tensin homolog (PTEN) is a regulator of phosphoinositide 3-kinase signaling and an important tumor suppressor mutated/deleted in human cancers. PTEN deletion in the liver leads to insulin resistance, steatosis, inflammation, and cancer. We recently demonstrated that unsaturated fatty acids trigger steatosis by down-regulating PTEN expression in hepatocytes via activation of a mammalian target of rapamycin (mTOR)/nuclear factor kappa B (NF-κB) complex, but the molecular mechanisms implicated in this process are still unknown. Here, we investigated potential genetic and epigenetic mechanisms activated by fatty acids leading to PTEN down-regulation. Our results indicate that unsaturated fatty acids down-regulate PTEN messenger RNA expression in hepatocytes through mechanisms unrelated to methylation of the PTEN promoter, histone deacetylase activities, or repression of the PTEN promoter activity. In contrast, unsaturated fatty acids up-regulate the expression of microRNA-21, which binds to PTEN messenger RNA 3′-untranslated region and induces its degradation. The promoter activity of microRNA-21 was increased by mTOR/NF-κB activation. Consistent with these data, microRNA-21 expression was increased in the livers of rats fed high-fat diets and in human liver biopsies of obese patients having diminished PTEN expression and steatosis. Conclusion: Unsaturated fatty acids inhibit PTEN expression in hepatocytes by up-regulating microRNA-21 synthesis via an mTOR/NF-κB–dependent mechanism. Aberrant up-regulation of microRNA-21 expression by excessive circulating levels of fatty acids exemplify a novel regulatory mechanism by which fatty acids affect PTEN expression and trigger liver disorders. (HEPATOLOGY 2009.)
Phosphatase and tensin homolog (PTEN) is a protein and phosphoinositide phosphatase originally identified as a tumor suppressor gene frequently mutated/deleted in human cancers.1 Numerous studies have also strongly supported its role as a regulator of insulin/insulin-like growth factor 1 signaling.2 In mice, liver-specific PTEN deletion leads to steatosis, insulin hypersensitivity early in life, and hepatomegaly and hepatocellular carcinoma (HCC) at later stages.3, 4 Consistent with these studies, PTEN is deleted, or weakly expressed, in HCC having a poor prognosis.5 Interestingly, accumulating evidence indicates that deregulated PTEN expression in hepatocytes, rather than PTEN mutations/deletions, represents a critical factor in the development of nonalcoholic fatty liver diseases (NAFLD) and HCC.6, 7 PTEN expression is regulated at the posttranscriptional level by various signaling cascades affecting its activity, stability, or intracellular localization.8 At the transcriptional level, genetic and epigenetic mechanisms controlling PTEN expression and implicated in hepatic carcinogenesis have been described. These include: (1) 5′ untranslated region (UTR)-dependent transcriptional regulation of PTEN mRNA synthesis, (2) methylation of PTEN promoter; (3) histone hyperacetylation, and (4) aberrant expression of microRNAs, which mediates translational repression or degradation of PTEN messenger RNA (mRNA).9, 10 We have recently shown that free fatty acids (FFAs) down-regulate PTEN expression in hepatocytes through a mammalian target of rapamycin (mTOR)/nuclear factor kappa B (NF-κB)-dependent mechanism, which then promotes liver steatosis.6 However, the molecular mechanisms by which FFAs inhibit PTEN expression are still unknown.
Obesity and elevated circulating FFAs are important etiological risk factors for the development of hepatic insulin resistance, NAFLD, and HCC.11 Given the role of PTEN in insulin signaling and as a tumor suppressor, understanding the mechanisms by which FFAs trigger PTEN down-regulation in hepatocytes is of crucial importance to understand the molecular basis of these diseases and to identify new therapeutic targets. In this study, we therefore investigated potential genetic and epigenetic mechanisms by which FFAs mediate PTEN down-regulation in hepatocytes.
Antibodies against PTEN and NF-κBp65 were from Santa Cruz Biotechnology (Santa Cruz, CA); anti-phospho-Akt(Ser473) and anti-Akt were from Cell Signaling Technology (Danvers, MA). Rapamycin was from Calbiochem (Bad Soden, Germany) and protein transduction domain (PTD)-p65-P1 was from Imgenex (San Diego, CA). Control and inhibitor anti–miR-21 microRNAs were from Ambion (Rotkreuz, Switzerland). The fatty acids 5-aza-2′-deoxycytidine (5-aza-dC) and trichostatin A (TSA) were from Sigma (St Louis, MO).
Cell Cultures, Human Primary Hepatocytes, Human Liver Biopsies, and Animals.
Cells were maintained at 37°C and 5% CO2. HepG2 and HuH-7 cells were cultured in Dulbecco's modified Eagle medium (DMEM)/10% fetal bovine serum and immortalized human hepatocytes (IHHs)12 in DMEM/F-12/10% fetal bovine serum containing dexamethasone/insulin.
Human primary hepatocytes (HPHs) were isolated from patients undergoing liver resections as previously described.13 Formalin-fixed human liver biopsies were obtained from obese patients with or without steatosis and going through bypass gastric surgery as described.6 Experiments were performed in accordance with ethical standards of the Helsinki Declaration of the World Medical Association and approved by the Geneva University Hospital ethical committee. Liver tissues from Wistar rats under standard or high-fat (Kliba-Nafag 2154, 19.5% fat including 34.6% saturated fatty acids, 41.9% unsaturated fatty acids [UFAs] and 22.9% polyunsaturated fatty acids; Kaiseraugst, Switzerland) diets were described previously.6 Experiments were approved by the Office Vétérinaire Fédéral et Cantonal of Geneva.
Plasmids and Luciferase Assays.
pcDNA plasmid encoding Flag–NF-κBp65 was provided by G. Haegeman (Gent University, Belgium),14 pGL3-PTEN-5′-UTR (5′-UTR-PTEN regulatory sequence cloned upstream of luciferase complementary DNA [cDNA]) was provided by T. Virolle (Nice University, France),15 pGL3-PTEN-3′-UTR (3′-UTR-PTEN cloned downstream of luciferase cDNA) was provided by T. Patel (Ohio State University, Columbus, OH),7 and pmiR-21s-luciferase (−959 to +49 hDNA fragment of the T1 pri-miR-21 transcription start cloned upstream of luciferase cDNA) was provided by B.R. Cullen (Duke University, Durham, NC).16
For luciferase assays, 2 × 106 cells/well were transfected with 1 μg of luciferase reporter constructs and 1 μg of pRL-TK (Renilla luciferase expression construct from Promega, Wallisellen, Switzerland) using FuGene6 reagent (Roche, Basel, Switzerland). Luciferase assays were performed 48 hours after transfection using a dual-luciferase reporter assay (Promega). Firefly luciferase activity was normalized to renilla luciferase expression for each sample.
Cells were lysed in ice-cold lysis buffer containing 50 mM Tris-HCl pH-7.5, 150 mM NaCl, 1% NP-40, 0.5% DOC, 0.1% SDS, 10 mM NaF, 5 mM VO4, and protease inhibitors (Roche). Equal amounts of proteins were resolved by 5%-20% sodium dodecyl sulfate–polyacrylamide gel electrophoresis, blotted to nitrocellulose membranes, and detected with enhanced chemiluminescence. Quantifications were performed using the ChemiDoc XRS from Bio-Rad Laboratories (Hercules, CA).
Caspase 3 enzymatic activity was measured with a fluorometric immunosorbent enzyme assay from Roche, according to manufacturer's instructions.
Real-Time Polymerase Chain Reaction, Methylation-Specific PCR, and MicroRNA measurements.
RNA was isolated from cells/tissues using Trizol Reagent and from paraffin-embedded biopsies using the Qiagen RNeasy/formalin-fixed, paraffin-embedded kit (Basel, Switzerland). First-strand cDNA synthesis was performed with 10 ng/L random hexamer primers and the SuperScript-II Reverse Transcriptase. Quantitative polymerase chain reaction (PCR) was performed using the Quantitect SYBR Green PCR kit in a Light-Cycler (Roche Diagnostics). Glyceraldehye 3-phosphate dehydrogenase (GAPDH)/Po/cyclophilin or EEF1A1/GusB and GAPDH/Po/β-actin transcripts were used as internal controls for humans and rats, respectively. Primer sequences were previously reported.6
The cDNA syntheses for miR-21, miR-let7g, and miR-let7i were performed using specific primers and MultiScribe Reverse Transcriptase (Applied Biosystems Inc., Rotkreuz, Switzerland). Quantitative PCR was performed with ad hoc (hsa-miR-21, hsa-miR-let7g, hsa-miR-let7i, has-miR-122) TaqMan MicroRNA assays (Applied Biosystems Inc.) in a Light-Cycler. The miR-let7g/ miR-let7i and miR-let7i/miR-122 levels were used as internal controls for human and rat samples, respectively.17
For methylation-specific PCR (MSP), DNAs were isolated using the Sigma GenElute DNA Miniprep kit. Unmethylated cytosines were converted to uracils using the Qiagen EpiTect Bisulfite kit. Unmodified/bisulfite-converted DNAs were then used as a template for quantitative PCR. MSP was performed using a specific set of primers for unmethylated/methylated PTEN promoter.18
Results are expressed as means ± standard error (SE). Comparisons were made using Student t tests. Differences were considered significant when *P < 0.05, **P < 0.01, or ***P < 0.001.
UFAs Down-Regulate PTEN mRNA Expression in Hepatocytes.
We previously reported that monounsaturated oleic acid (OA) down-regulates PTEN expression in HepG2 cells and that PTEN levels are decreased in human and rat livers with steatosis.6 Here, we evaluated whether PTEN mRNA expression is specifically down-regulated by UFAs and whether this mechanism is a general feature of hepatocytes.
HepG2 cells were incubated for 12-24 hours with various concentrations of OA or saturated palmitic acid (PA). After 12 hours incubation, we did not observe any effect of these FFAs on PTEN mRNA expression, but high concentrations of FFAs (≥250 μM) significantly increased apoptosis (caspase 3 activity; Supporting Fig. 1A,B). In contrast, 50 μM FFAs were well tolerated by HepG2 cells (up to 7 days12); at this concentration, UFAs (oleic, palmitoleic, and linoleic acids), but not saturated PA, induced a ≥ two-fold decrease in PTEN mRNA expression after 24 hours (Fig. 1A).
We then investigated whether OA also modulates PTEN mRNA expression in other hepatic cells including Huh7 hepatoma cells, IHHs, and HPHs. Consistent with previous studies,19 PTEN expression was lower in Huh7 and IHHs as compared to HPHs (Fig. 1B,C). However, in all cells, OA induced a ∼50% down-regulation of PTEN mRNA expression (Fig. 1B), which was further reflected by a ∼30%-40% decrease in PTEN protein levels (Fig. 1C). As a functional consequence, increased Akt basal phosphorylation was measured in cells with down-regulated PTEN (Fig. 1C).
OA-Mediated PTEN Down-Regulation Does Not Occur Through Epigenetic Mechanisms.
Gene methylation and histone hyperacetylation are epigenetic mechanisms regulating PTEN expression and involved in hepatic carcinogenesis.20 However, whether UFAs inhibit PTEN expression through such mechanisms in hepatocytes is unknown.
PTEN promoter methylation was examined in HepG2 cells and HPHs using MSP analysis and primer sequences targeting regions of the PTEN promoter distinct from those of the PTEN pseudogene.18 Consistent with PTEN expression levels in HepG2 cells and HPHs (Fig. 1), PTEN promoter in HepG2 cells was hypermethylated, whereas methylation was not detectable in HPHs (Fig. 2A). HepG2 cells preincubated with 5-aza-dC, a drug blocking DNA methylation,21 exhibited decreased PTEN promoter methylation and dose-dependently increased PTEN mRNA and protein expression (Fig. 2A,C). However, 5-aza-dC could not prevent PTEN down-regulation and Akt activation in cells exposed for 24 hours to 50 μM OA (Fig. 2B,C). Consistent with these data, OA does not affect PTEN promoter methylation in hepatocytes (Fig. 2A).
To evaluate whether OA could inhibit PTEN expression by inducing histone deacetylation, we incubated HepG2 cells with the histone deacetylases inhibitor TSA. As described in kidney cells,9 TSA up-regulated PTEN expression in hepatocytes in a dose-dependent manner (Fig. 2D). However, TSA could not prevent PTEN down-regulation and Akt activation induced by OA in HepG2 cells (Fig. 2D,E).
These data demonstrate that although promoter methylation and histone acetylation can affect PTEN expression in hepatocytes, these processes are not involved in UFA-mediated PTEN down-regulation.
OA Does Not Repress PTEN Promoter Activity.
We recently showed that PTEN down-regulation by OA was mediated by NF-κB.6, 22 To investigate whether UFAs affect PTEN expression by repressing the PTEN promoter activity, we used a luciferase reporter construct under the control of the PTEN promoter (pGL3-PTEN-5′-UTR) (Fig. 3A).15 As expected, NF-κBp65 overexpression, or tumor necrosis factor-α (TNFα) stimulation, significantly decreases the luciferase activity in HepG2 cells transiently transfected with pGL3-PTEN-5′-UTR (Fig. 3B).23, 24 However, exposure of HepG2 cells to 50 μM OA for 24 hours, or to higher doses of OA for 12 hours, failed to inhibit the luciferase activity of pGL3-PTEN-5′-UTR (Fig. 3B,C).
These findings indicate that although OA-mediated PTEN mRNA down-regulation is NF-κB–dependent, the pool of NF-κB activated by OA does not repress the PTEN promoter activity.
OA Induces mTOR/NF-κBp65–Dependent miR-21 Production Triggering PTEN mRNA Down-Regulation.
PTEN mRNA has been identified through bioinformatics and experimental approaches as a target for microRNA-21 (miR-21), whose “seed” sequence (base 2-9 of the 5′ end) shares >85% homology with PTEN mRNA (seven of eight bases; nucleotides 2672-2678)7, 25, 26 (Fig. 4A). We thus evaluated the hypothesis that UFAs up-regulate miR-21 expression in hepatocytes to decrease PTEN mRNA expression. As shown in Fig. 4B, 24 hours incubation with 50 μM OA, but not PA, significantly increased miR-21 expression in HepG2 cells and HPHs. Up-regulation of miR-21 necessitates at least 24 hours incubation with OA, because only 12 hours incubation with unsaturated/saturated FFAs did not affect miR-21 expression in HepG2 cells (Supporting Fig. 1C). Consistent with our previous studies showing that OA-induced PTEN down-regulation requires mTOR/NF-κB activation,6 miR-21 up-regulation was prevented by a NF-κBp65 inhibitor (PTD-p65-p127) and partially inhibited by an mTOR inhibitor (rapamycin) (Fig. 4B). The specificity of miR-21 in targeting PTEN mRNA was assessed by transfecting HepG2 cells and HPHs with a construct encoding the 3′-UTR region of the PTEN gene in a luciferase reporter gene (pGL3-PTEN-3′-UTR)7 (Fig. 3A). Incubation for 24 hours with 50 μM OA significantly decreased the luciferase activity in cells expressing pGL3-PTEN-3′-UTR indicating that the PTEN 3′-UTR is likely a target of miR-21 following induction of its synthesis by OA (Fig. 4C). The specificity of miR-21 binding to the PTEN 3′-UTR was further confirmed by cotransfecting HepG2 cells with pGL3-PTEN-3′-UTR plasmids in combination with either a control or a specific miR-21 inhibitor (anti–miR-21). As shown in Fig. 4D, anti–miR-21 increased basal luciferase activity in cells expressing pGL3-PTEN-3′-UTR and fully prevented the decrease in PTEN 3′-UTR–driven luciferase activity triggered by OA (Fig. 4D).
Finally, the pathophysiological relevance of PTEN down-regulation by UFA-mediated miR-21 up-regulation was evaluated in the liver of Wistar rats fed a high-fat diet and in human liver biopsies of patients with hepatic steatosis. We previously demonstrated in the same samples that PTEN mRNA and protein expression were down-regulated concomitantly with an increased phosphorylation of Akt and NF-κB.6 Consistent with these data, we found an increased miR-21 expression in these samples, which strongly correlated with down-regulation of PTEN expression (Pearson's correlation coefficient > 0.8) (Fig. 4E,F).
These results indicate that OA triggers PTEN mRNA degradation in hepatocytes by up-regulating miR-21, which specifically targets PTEN mRNA, through mTOR/NF-κB–dependent mechanisms.
miR-21 Promoter Is Activated by NF-κBp65 in Response to OA.
The human miR-21 transcriptional promoter has been recently described.16 By sequence analyses (AliBaba 2 software), we identified three putative NF-κB binding sites on the miR-21 promoter, suggesting that miR-21 expression might be under the direct control of NF-κB (Fig. 5A). We thus examined whether OA could up-regulate miR-21 expression by stimulating its transcription. HepG2 cells and HPHs were transfected with a luciferase reporter vector coupled to the miR-21 promoter elements (pmiR-21s-luc16) and then were treated for 24 hours with 50 μM OA prior to record the luciferase activity driven by the miR-21 promoter. OA enhanced the luciferase activity in both HepG2 cells and HPHs, whereas PA had no effect (Fig. 5B,C). Interestingly, preincubation of cells with mTOR and NF-κBp65 inhibitors prevented the OA-induced luciferase activity (Fig. 5C), recapitulating what we observed for miR-21 transcripts levels (Fig. 4A). Consistent with a direct effect of NF-κBp65 on the miR-21 promoter, NF-κBp65 overexpression in HepG2 cells significantly enhanced the miR-21 promoter-driven luciferase activity (Fig. 5C).
These data indicate that UFAs trigger miR-21 production in hepatocytes through mTOR/NF-κBp65–dependent activation of the miR-21 promoter.
Down-regulation of the tumor suppressor PTEN in hepatocytes is a critical event promoting hepatic steatosis, steatohepatitis, and HCC.3, 4, 6, 7, 28 We previously showed that UFAs decrease PTEN expression in hepatocytes through molecular mechanisms involving the activation of an mTOR/NF-κB complex.6 Here, we demonstrate that UFAs decrease PTEN mRNA expression by up-regulating miR-21, a microRNA specifically targeting PTEN mRNA for degradation.7 UFAs stimulated miR-21 promoter activity in an mTOR/NF-κB–dependent manner, thus unraveling a new regulatory mechanism for miR-21 expression in metabolic disorders.
Genetic and epigenetic mechanisms have been described that regulate PTEN expression in pathophysiological conditions. Most of these studies have focused on the flanking 5′-UTR of PTEN gene, which elicits strong promoter activity.29 Epigenetic mechanisms including DNA methylation and histone acetylation likely contribute to regulate PTEN expression in hepatocytes. Indeed, cytosine-guanine dinucleotide islands in PTEN 5′-UTR were reported to be hypermethylated in HCC displaying low PTEN expression.20 Consistent with these observations, HepG2 cells have a hypermethylated PTEN promoter and low PTEN expression as compared to HPHs, a feature reversed by DNA methyltransferase inhibitors. We and others also showed that histone deacetylase inhibitors stimulate PTEN mRNA expression in hepatocytes.9 Although these mechanisms are critical for the silencing of important genes, e.g., tumor suppressors, in cancer, they are not implicated in UFA-dependent PTEN down-regulation. Indeed, DNA methyltransferases and histone deacetylase inhibitors could not prevent UFA-mediated PTEN down-regulation in hepatocytes. Individual or combined epigenetic therapies with such general inhibitors are envisaged for HCC treatment.30 However, these inhibitors could not prevent PTEN down-regulation and may lack specificity, leading to undesirable effects, e.g., increased oncogene expression. Their use as therapeutic tools should thus be considered with caution.
PTEN expression is also modulated by transcription factors, including early growth response 1 (Egr-1)15 and NF-κBp65,23 which stimulate PTEN promoter activity. Egr-1 expression in hepatocytes is not affected by OA (data not shown). The role of NF-κB in controlling PTEN expression is, however, more complex. We showed previously that OA down-regulates PTEN expression through mTOR/NF-κB–dependent mechanisms,6 but this study indicates that this is not occurring through inhibition of the PTEN promoter activity (Fig. 3). Interestingly, TNFα-mediated NF-κBp65 activation, or NF-κBp65 overexpression in HepG2 cells, increases the luciferase activity of cells expressing pGL3-PTEN-5′-UTR, but independently of NF-κBp65 binding on the PTEN 5′-UTR as assessed by transfecting cells with a pGL3-PTEN-5′-UTRΔ600 mutant construct lacking NF-κBp65 binding sites23 (data not shown). This observation is consistent with the study of Rangnekar and collaborators,23 but in conflict with a report showing mitogen-activated protein kinase kinase 4/SEK1-mediated NF-κB binding on PTEN 5′-UTR.31 Together, these studies suggest complex NF-κB–dependent mechanisms regulating PTEN expression at multiple levels. A further degree of complexity is uncovered by our study, which shows that UFA-mediated mTOR/NF-κB activation up-regulates miR-21 expression, a microRNA targeting PTEN mRNA for degradation.7
MicroRNAs are small noncoding RNAs that play important roles in posttranscriptional gene regulation, as regulators of metabolic processes and in cancers including HCC.7, 32, 33 The ability of microRNAs to repress a target mRNA depends on the perfect microRNA seed sequences match to the target mRNA.25 However, recent studies demonstrated that microRNA:mRNA pairing may occur with mismatches (G:U wobble base pairing; Fig. 4A).34, 35 PTEN was identified as a target of miR-21 according to this criterion and further confirmed experimentally.25, 26 Our results now indicate that miR-21 up-regulation is not only involved in HCC,7 but also in NAFLD because PTEN down-regulation induces steatosis.6 The regulation of microRNA transcription is still poorly understood and subject to intense investigation.36 Interestingly, the structure of the miR-21 gene and its promoter have been recently studied in depth,16 and binding sites for transcription factors (e.g., signal transducer and activator of transcription 3 [Stat3]) have been identified.37 By sequence analysis, we have identified three putative NF-κBp65 binding sites in the human miR-21 promoter, which are functionally relevant because OA/TNFα-mediated NF-κBp65 activation, or NF-κBp65 overexpression, enhances miR-21 promoter activity and triggers miR-21 production in hepatocytes. It still remains unclear why UFA-mediated NF-κB activation decreases PTEN mRNA levels by up-regulating miR-21 expression and does not have any effect on PTEN promoter activity. We previously showed that in hepatocytes, a pool of NF-κBp65 is complexed to mTOR and that mTOR activation is required to down-regulate PTEN expression.6 We can thus hypothesize that the specific effect of UFA-mediated NF-κB activation on miR-21 expression relies on subtle, but critical, regulatory mechanisms such as NF-κB interactions with other signaling molecules (e.g., mTOR), the intensity/duration of NF-κBp65 activation, NF-κBp65 posttranscriptional modifications, or intracellular compartmentalization as suggested by previous studies.38 The role of miR-21 up-regulation in previously described NF-κB–dependent mechanisms regulating PTEN expression23, 31 should be evaluated to shed light on these issues.
In conclusion, our work identifies miR-21 as a crucial factor affecting PTEN expression in the liver with the metabolic syndrome. Indeed, dysregulated metabolic factors such as excessive circulating UFAs alter miR-21 expression resulting in PTEN down-regulation and the development of steatosis. Whether UFA-mediated/miR-21–mediated PTEN down-regulation also promotes the development of more severe liver disorders (e.g., steatohepatitis and HCC) is currently unknown. Also, whether miR-21 affects hepatocyte viability and insulin sensitivity via PTEN down-regulation or other mechanisms remains to be established. Interestingly, strong evidence indicates that partial loss of PTEN function is sufficient to promote tumor development.10 Further studies are ongoing to establish whether alterations of miR-21/PTEN expression in hepatocytes are reliable prognostic markers for NAFLD and HCC development. Understanding the molecular mechanisms affecting PTEN expression in hepatocytes is also essential to design alternative new therapeutic strategies to cure these diseases. In this regard, miR-21 and PTEN are potential new therapeutic targets for liver diseases associated with the metabolic syndrome.
We thank W. Reith, B. Gauthier, and S. Nef for critically reading the manuscript and A. Gateau, F. Meng, T. Patel, and D. O'Carroll for help with sequences analyses. We thank the Genomics Platform, Swiss National Center for Competence in Research “Frontiers in Genetics” at the University of Geneva for access to real-time PCR equipment.