Acute ethanol exposure inhibits insulin signaling in the liver

Authors

  • Jiman He,

    1. Liver Research Center, Department of Medicine and Pathology, Rhode Island Hospital and Warren Alpert Medical School of Brown University, Providence, RI 02903
    Search for more papers by this author
  • Suzanne de la Monte,

    1. Liver Research Center, Department of Medicine and Pathology, Rhode Island Hospital and Warren Alpert Medical School of Brown University, Providence, RI 02903
    Search for more papers by this author
  • Jack R. Wands

    Corresponding author
    1. Liver Research Center, Department of Medicine and Pathology, Rhode Island Hospital and Warren Alpert Medical School of Brown University, Providence, RI 02903
    • Liver Research Center, 55 Claverick Street, Providence, RI 02903
    Search for more papers by this author
    • fax: 401-444-2939


  • Potential conflict of interest: Nothing to report.

Abstract

Chronic ethanol consumption may produce hepatic injury and impair the ability of the liver to regenerate principally through its action on insulin signaling. These effects are mediated by insulin receptor substrate-1 (IRS-1) via the mitogen-activated protein kinase/extracellular signal regulated kinase (MAPK/Erk) pathway and by survival signals through phosphatidylinositol-3 kinase (PI3K) and protein kinase B (Akt). Because a protein phosphatase, phosphatase tensin homolog deleted on chromosome 10 (PTEN), has been reported to block insulin signaling through PI3K, we explored acute ethanol effects on signaling in the context of PTEN function. We measured upstream components of the insulin signal transduction pathway and Akt phosphorylation as an indicator of signaling through PI3K, including the generation of survival signals via glycogen synthase kinase 3β (GSK3β) and Bcl-2–associated death promoter (BAD). In addition, the physical association between PTEN and PI3K regulatory (p85α) and catalytic (p110α) subunits was evaluated both in vitro and in vivo. In Huh-7 cells, there was no effect of acute ethanol exposure on tyrosyl phosphorylation of the insulin receptor, IRS-1, and the association of IRS-1 with PI3K. However, Akt phosphorylation was impaired. The association of PTEN with the PI3K p85α subunit was substantially increased and led to the inhibition of downstream insulin-mediated survival signals through Akt, GSK3β, and BAD; the ethanol effect was reversed by PTEN knockdown with small interfering RNA. These results were confirmed in the liver. Conclusion: Short-term ethanol exposure rapidly attenuates insulin signaling. The major cellular mechanism involves the increased association of PTEN with the PI3K p85α subunit, which results in reduced phospho-Akt formation and impaired downstream survival signaling. These findings may have relevance to acute toxic effects of ethanol on the liver. (HEPATOLOGY 2007.)

Insulin produces its effect on hepatocyte growth and survival by the activation of a signal transduction cascade, which involves ligand interactions with cell surface receptors and the sequential activation of tyrosine kinases.1–3 Important components in this cascade include the insulin receptor (IR), insulin receptor substrate-1 (IRS-1), and insulin receptor substrate-2 (IRS-2). IRS-1 transmits the insulin signal following tyrosyl phosphorylation (PY) by the IR tyrosine kinase. Following this event, PY–IRS-1 interacts with growth factor receptor bound protein 2,4 synaptophysin protein tyrosine phosphatase,5 and the regulatory p85 subunit of phosphatidylinositol-3 kinase (PI3K).6, 7 Signaling through growth factor receptor bound protein 2 (Grb2) and son of sevenless (SOS) activates mitogen-activated protein kinase (MAPK) and Erk MAPK,8–10 leading to hepatocyte proliferation and gene expression. In addition, the binding of PY–IRS-1 to PI3K p85 promotes survival signals by signaling through Akt, glycogen synthase kinase 3β (GSK3β), and Bcl-2–associated death promoter (BAD).11, 12 Thus, the functional consequences of insulin signaling are increased hepatocyte proliferation, survival, and energy metabolism.

The protein phosphatase tensin homolog deleted on chromosome 10 (PTEN) plays a critical role in the regulation of insulin signaling. This molecule is a lipid phosphatase that dephosphorylates and reverses the activation of PI3K,4, 13–17 and it was initially discovered as a tumor suppressor gene because inactivation through mutational events leads to increased phosphorylation of Akt and constitutive activation of downstream signaling events that are antiapoptotic in function. In contrast, high levels of PTEN and increased binding to phosphatidylinositol-3,4,5-triphosphate (PIP3) are associated with reduced cellular levels of phosphorylated Akt (p-Akt) and increased GSK3β and BAD activity,18–21 resulting in proapoptotic signals to hepatocytes.

Chronic ethanol consumption in experimental animal models alters insulin signaling events and produces insulin resistance in the liver. These adverse effects of ethanol have been shown to be due to the inhibition of insulin/insulin-like growth factor 1–induced DNA synthesis22–25 and the activation of proapoptotic signals through PI3K and Akt and increased levels of PTEN.26 However, little is known regarding acute ethanol effects on these signaling cascades and the mechanism involved. Because acute ethanol exposure is toxic to the liver,27, 28 the present studies explored its action on the IRS-1/PI3K/Akt/GSK3β cascade both in vitro and in vivo. It has been found that the short-term exposure of cells to ethanol strikingly down-regulates insulin signaling through this pathway, and PTEN plays a significant regulatory role for ethanol action.

Abbreviations

BAD, Bcl-2–associated death promoter; DMEM, Dulbecco's modified Eagle's medium; GSK3β, glycogen synthase kinase 3β; IP, immunoprecipitation; IR, insulin receptor; IRβ, insulin receptor-β subunit; IRS-1, insulin receptor substrate-1; IRS-2, insulin receptor substrate-2; MAPK, mitogen-activated protein kinase; mRNA, messenger RNA; p-Akt, phosphorylated Akt; PI3K, phosphatidylinositol-3 kinase; PIP2, phosphatidylinositol-4,5-biphosphate; PIP3, phosphatidylinositol-3,4,5-triphosphate; PTEN, phosphatase tensin homolog deleted on chromosome 10; PY, tyrosyl phosphorylation; RIPA, radio immunoprecipitation assay; SDS-PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis; siRNA, small interfering RNA.

Materials and Methods

Acute Alcohol Exposure.

For in vitro studies, Huh-7 hepatocellular carcinoma cells were cultured with Dulbecco's modified Eagle's medium (DMEM) in 10% fetal serum with the addition of L-glutamine and nonessential amino acids. Cells were exposed to ethanol (100 mM) for 10-30 minutes after overnight serum starvation (9-10 hours) in 0.5% serum containing DMEM. Insulin stimulation (0.075 U/mL) was accomplished by direct addition to the medium followed by rapid mixing with a pipette and incubation with cells for two minutes; thereafter, the medium was removed and washed with 10 mL of a new (insulin-free) culture medium 2 times, and this was followed by a final change to insulin-free DMEM for the various studies as indicated. Thereafter, cells were harvested in a radio immunoprecipitation assay (RIPA) buffer plus a proteinase inhibitor cocktail, as described later, and separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), which was followed by a western blot analysis.

Acute ethanol effects were also studied in 5-month-old female FVB mice (n = 20). After an overnight fast, the animals were injected intraperitoneally with 0.5 mL of ethanol diluted in saline. The concentration of ethanol that was administered was 1.2 g/kg. This dose is similar to that used in human volunteer studies in which ethanol was administered intravenously to evaluate acute ethanol effects on insulin resistance.29 After the intraperitoneal injection, mice were euthanized with isoflurane, and 10 minutes after administration, the livers were removed and placed in liquid nitrogen. Tissue lysates were prepared for immunoprecipitation studies and western blot analysis to assess PTEN binding to the regulatory p85α subunit of PI3K and the phosphorylation of Akt GSK3β and BAD. The acute ethanol exposure protocol for the mice was approved by the Lifespan Animal Care Committee.

Protein Studies.

The protein expression was examined by a western blot analysis or immunoprecipitation followed by immunoblotting. For a direct western blot analysis, liver tissue was homogenized in 5 volumes of an immunoprecipitation assay buffer (50 mmol/L trishydroxymethylaminomethane-HCl, pH 7.5, 1% Nonidet P40, 0.25% sodium-deoxycholate, 150 mmol/L NaCl, 1 mmol/L ethylene diamine tetraacetic acid, and 2 mmol/L ethylene glycol tetraacetic acid) containing protease and phosphatase inhibitors (1 mmol/L phenylmethylsulfonyl fluoride, 0.1 mmol/L tosyl phenylalanyl chloromethyl ketone, 1 μg/mL aprotinin, 1 μg/mL pepstatin A, 0.5 μg/mL leupeptin, 1 mmol/L NaF, 1 mmol/L Na4P2O7, and 2 mmol/L Na3VO4). Protein concentrations were determined with a bischloracetate assay (Pierce, Rockford, IL). For immunoblotting, samples containing 100 μg of protein were fractionated by SDS-PAGE and transferred to poly(vinylidene difluoride) membranes. Nonspecific binding sites were blocked with SuperBlock-TBS (Pierce), and the membranes were incubated with a primary antibody (0.5-1 μg/mL) overnight at 4°C with gentle agitation. The immunoreactivity was detected with the horseradish peroxidase–conjugated secondary antibody (1:100 dilution) and enhanced chemiluminescence reagents and then quantified with a Kodak Digital Science imaging station (NEN Life Sciences, Boston, MA). The levels of immunoreactivity are reported in arbitrary densitometry units.

For immunoprecipitation studies, Huh-7 cell or liver tissue lysates were homogenized in a Triton lysis buffer (50 mmol/L trishydroxymethylaminomethane-HCl, pH 7.5, 10 mmol/L ethylene diamine tetraacetic acid, and 0.5% Triton X-100) containing protease and phosphatase inhibitors, as indicated previously. Aliquots of 500 μg of protein diluted to 1 mg/mL in a Triton lysis buffer were precleared with Protein A Sepharose (Amersham, Pharmacia) for 40 minutes and then incubated overnight at 4°C with a 1 μg/mL concentration of the primary antibody with a constant rotation of the samples. Protein A Sepharose was then added, and the samples were further incubated for 2 hours at 4°C with constant rotation. The immune complexes were washed 3 times in a Triton lysis buffer for 1 minute. After the third wash, the immunoprecipitants were resuspended in an SDS-PAGE sample buffer with a loading dye for a western blot analysis.26

Small Interfering RNA (siRNA) Transfection Experiments.

The siRNA directed against PTEN (#6250) was purchased from Gene Therapy Systems as described.30 The Huh-7 cells were passaged into 12-well plates in DMEM supplemented with 10% bovine serum. On the second day, when the cells were 50%-70% confluent, transfections were performed with siRNAs targeting PTEN or with scrambled control RNAs. On day 3, cells were placed in DMEM with the addition of 0.5% serum for 10 hours and then exposed to ethanol for 20 minutes with or without insulin additions for another 5 minutes. Thereafter, cells were harvested in an RIPA buffer and centrifuged at 10,000 g at 4°C, and the supernatants were separated on SDS-PAGE gels and tested for the abundance of PTEN and p-Akt.

Source of the Reagents.

Antibodies to Akt, GSK3β, BAD, PTEN, p-Akt, phospho-GSK3β (Ser9), and phospho-BAD (Ser112) were purchased from Cell Signaling Technology (Beverly, MA), Santa Cruz (Santa Cruz, CA), and BD (Franklin Lakes, NJ). Antibodies to IRS-1, the p85α subunit of PI3K, and phosphotyrosine (PY20) were purchased from Transduction Laboratories (Lexington, KY). Antibodies to the insulin receptor-β subunit (IRβ; SC711), p110α (SC-7174), bovine anti-rabbit immunoglobulin G horseradish peroxidase (SC-2374), and bovine anti-goat immunoglobulin G horseradish peroxidase (SC-2350) were obtained from Santa Cruz.

Statistical Analysis.

The data depicted in the graphs represent means ± the standard error of the mean. The results were analyzed with the SigmaStat statistics program (Jandel Scientific, San Rafael, CA). Data derived from 3 groups or more were compared by a 1-way analysis of variance followed by a Tukey-Dunn test to identify the groups that differed. Differences at P < 0.05 were considered significant.

Results

Acute Ethanol Exposure Does Not Alter the Upstream Components of the Insulin Signaling Cascade.

Insulin binds to IR and results in the activation of the kinase, which brings IRS-1 to the receptor followed by PY. PY–IRS-1 binds to the PI3K p85 subunit through four motifs located in the C-terminus of the molecule, resulting in PI3K activation. In these experiments, we examined PY of IR, the association of IRS-1 with IR, PY of IRS-1, and the association of IRS-1 with PI3K p85 following insulin stimulation in the presence or absence of a 20-minute exposure to 100 mM ethanol.

Using coimmunoprecipitation experiments, we found that there was no change in PY of IR or PY of IRS-1, and as expected, no difference was observed in the association of IRS-1 with IR or with the PI3K p85 subunit, as shown in Fig. 1. Furthermore, the cellular levels of IRβ, IRS-1, PI3K p85, and the PTEN phosphatase (Fig. 1A-C) remained constant; there was no change in the level of PI3K p85 as well (data not shown). These investigations suggest that acute ethanol exposure to Huh-7 cells has little, if any, effect on insulin-stimulated PY of the IRβ and IRS-1 molecules. Subsequently, the binding of IRβ to IRS-1 and the binding of IRS-1 to PI3K p85 as the main upstream events required to transmit the insulin signal were unchanged.

Figure 1.

Effect of insulin stimulation in the presence or absence of acute ethanol exposure on the upstream components of the insulin signaling pathway in Huh-7 cells. (A,B) There was no change in the protein levels of IRβ and IRS-1 in Huh-7 cells exposed to ethanol for 20 minutes. In addition, there was no difference in the association of IRβ with IRS-1 or IRS-1 with the PI3K p85α subunit because ethanol had no effect on PY of IRβ or IRS-1. (C) There was no change in the PTEN levels by a direct western blot in a comparison of cells not exposed to ethanol (0 hours) and those cells exposed for 0.5, 3, and 16 hours. These experiments were repeated three times with similar results. IP indicates immunoprecipitation.

Acute Ethanol Exposure Inhibits Akt Phosphorylation.

Given the aforementioned observation, it was surprising that insulin-stimulated phosphorylation of Akt was inhibited by acute ethanol exposure, as shown in Fig. 2, because there was little, if any, effect on the upstream components of this signaling cascade (Fig. 1). Indeed, there was rapid phosphorylation of Akt 5 minutes after insulin stimulation, with a gradual fall-off at 15 and 40 minutes. In contrast, the baseline p-Akt levels were low in ethanol-exposed Huh-7 cells, and there was a significant reduction in p-Akt formation following insulin stimulation, which was particularly prominent 5 and 15 minutes after exposure. These results suggest that the ethanol inhibition of Akt phosphorylation may not be acting through upstream components of this signaling cascade. It could involve a different mechanism, such as the activation of a phosphatase downstream of the IR and IRS-1 signaling proteins. To address this issue, we explored the effect of insulin on the binding of PTEN to PI3K.

Figure 2.

Short-term ethanol exposure has a significant impact on the phosphorylation of Akt. Insulin stimulation substantially increased the phosphorylation of Akt (p-Akt), with peak levels observed 5 minutes after the addition of insulin. There was a gradual decrease in p-Akt formation at 15 and 45 minutes. In contrast, ethanol exposure substantially reduced the baseline p-Akt levels and blunted the insulin-stimulated rise in p-Akt at 5 and 15 minutes. The results are the means of three independent experiments.

Acute Ethanol Exposure Produces a Rapid Association of PI3K p85α with PTEN.

In this experiment, Huh-7 cells were exposed to ethanol for 5, 10, and 30 minutes. At each time, cells were harvested, and lysates were prepared; this was followed by immunoprecipitation with antibodies directed against PTEN followed by immunoblotting with antibodies directed against the PI3K p85α regulatory subunit. As shown in Fig. 3A, after as little as 5 minutes of exposure to ethanol, there was a rapid association of PTEN with PI3K p85α, and by 30 minutes of exposure, the interaction was back to pre-exposure levels. We also performed the reciprocal experiment with immunoprecipitation for the p85 subunit of PI3K and with a blot with antibodies to PTEN to further confirm these results. Identical results were obtained (data not shown). In this context, insulin stimulation enhanced the binding of PTEN to PI3K p85 to the same level observed with ethanol exposure for 10 minutes; this phenomenon may represent a homeostatic mechanism to attenuate the insulin signal. These results suggest that ethanol alone (in the absence of insulin stimulation) has the capacity to enhance the association of PTEN with PI3K p85α, even though the cellular levels of PTEN do not change at least up to 16 hours following ethanol additions, as shown in Fig. 1.

Figure 3.

Acute ethanol exposure is associated with the rapid binding of PTEN to the PI3K p85α subunit. (A) A time course is presented, indicating that at 5 minutes, there is a striking association of PTEN with the PI3K p85α subunit, which peaks at 10 minutes and has diminished binding at 30 minutes after ethanol exposure. (B) Insulin stimulation for 10 minutes in the absence of ethanol causes a strong association of PTEN with the PI3K p85α subunit. At 10 minutes, the addition of ethanol produces essentially the same effect as insulin stimulation alone. The exposure of Huh-7 cells to ethanol cells for 10 minutes and to insulin for 10 minutes results in the diminished association of the PI3K p85α subunit and PTEN. (C) Ethanol promotes the association of the p85α regulatory subunit and not the PI3K P110α catalytic subunit with PTEN. In these experiments, cells were exposed to ethanol for 0, 5, 10, and 30 minutes. Lysates were prepared and immunoprecipitated with antibodies to PTEN and were blotted with antisera specific for the p85α and p110α subunits of PI3K. Similarly to the effects found in parts A and B, ethanol produced a maximum association at 10 minutes of ethanol exposure. These experiments demonstrate that ethanol exposure alone increases the binding of PI3K p85α to PTEN to the same degree as insulin stimulation. A western blot was performed on a parallel aliquot to ensure an equal loading. The experiments were repeated (A) 5, (B) 4, and (C) 3 times. IP indicates immunoprecipitation.

Acute Ethanol Exposure Enhances PI3K Association with PTEN via a Specific Interaction with the p85α Regulatory Subunit.

In order to further confirm the association of PI3K with PTEN in the context of ethanol exposure for 5, 10, and 30 minutes, we determined if the interaction was specific for the p85α regulatory subunit of P13K. In these experiments, immunoprecipitations were performed with lysates followed by immunoblotting with antisera specific for either the regulatory p85α or catalytic p110α subunits. As shown in Fig. 3C, acute ethanol exposure of Huh-7 cells for 10 minutes was associated with maximal binding of PTEN with the p85α and not the p110α subunit. This rapid association was similar to the findings exhibited in Fig. 3A,B.

Reduced Phosphorylation of Akt by Ethanol Is Translated to Downstream Effects on GSK3β and BAD.

Because GSK3β and BAD are downstream molecules of Akt in this signaling cascade, we evaluated the consequences of reduced p-Akt induced by ethanol on the phosphorylation of these molecules. Figure 4 displays the results. In these experiments, Huh-7 cells were exposed to ethanol for 20 minutes, and this was followed by 5 minutes of stimulation with insulin. Figure 4A reveals that the cellular levels of GSK3β were unchanged in the presence or absence of ethanol exposure. However, there was a significant reduction in phosphorylated glycogen synthase kinase 3β (p-GSK3β) formation in the presence of both insulin stimulation and ethanol exposure. Similarly, ethanol exposure alone was associated with reduced phosphorylated Bcl-2–associated death promoter (p-BAD) in comparison with control levels (without insulin or ethanol exposure). Acute ethanol exposure reduced p-BAD formation following insulin stimulation. Reduced p-GSK3β and p-BAD levels activate these downstream signaling molecules and promote proapoptotic signals, as shown later in Fig. 7.

Figure 4.

Acute ethanol exposure in Huh-7 cells affects the phosphorylation of downstream components of the PI3K/Akt signaling pathway. (A) Huh-7 cells were stimulated with insulin for 5 minutes in the presence or absence of 100 mM ethanol exposure for 20 minutes. There was a significant reduction in the formation of p-GSK3β. (B) The effect of acute ethanol exposure on the phosphorylation of BAD (p-BAD) is shown in the presence or absence of insulin stimulation after the overnight serum starvation (left panel, 0.0%; right panel, 0.5%) of cells. The presence of ethanol alone reduced p-BAD formation; this effect was also seen with insulin stimulation (the latter two columns). These experiments were repeated 3 times with similar results.

Figure 7.

Hypothetical model of acute ethanol effects on insulin signaling through PI3K/Akt and the role of PTEN. In the absence of ethanol exposure, as shown on the left, the insulin signal is transmitted through IRS-1, and this is followed by the binding of the p85α subunit of PI3K. This event activates the kinase and leads to the phosphorylation of Akt. The phosphorylation of Akt results in the phosphorylation of GSK3β and BAD, which renders them inactive. The functional effects are increased prosurvival signals. In the presence of acute ethanol exposure, there is preferential binding of the p85α subunit of PI3K to PTEN, which competes for its binding to IRS-1. The net effect is reduced phosphorylation of Akt, GSK3β, and BAD, which renders the last two signaling molecules active. Under these conditions, there is the generation of proapoptotic signals, and hepatocytes are more susceptible to injury because of the acute effects of ethanol.

Acute Ethanol Exposure Promotes PTEN Association with PI3K p85α and Inhibits p-Akt and p-GSK3β Formation in the Liver of Mice.

In order to further explore the effects of alcohol on signaling through the PI3K/Akt/GSK3β cascade, we performed experiments in vivo to validate the in vitro findings. In these experiments, mice were given an intraperitoneal injection of ethanol (1.2 g/kg), and 10 minutes later, the liver was harvested for the assessment of the binding of the PI3K p85α subunit to PTEN and for the measurement of the p-Akt and p-GSK3β levels. As shown in Fig. 5, there was rapid binding of the PI3K p85α subunit to PTEN 10 minutes after acute ethanol administration, and this was similar to what was found in vitro. This phenomenon was associated with reduced phosphorylation of Akt and, subsequently, its downstream effector GSK3β. Therefore, acute ethanol exposure to the liver in vivo inhibits this endogenous insulin-responsive signaling cascade.

Figure 5.

Acute ethanol exposure in vivo promotes the association of PTEN with the PI3K p85α subunit and results in the reduced phosphorylation of Akt and downstream signaling molecules such as GSK3β in the liver. (A) Immunoprecipitation with antibodies to PTEN followed by a western blot analysis with antibodies to the PI3K p85α catalytic subunit in cell lysates prepared from the liver tissue of mice who were acutely exposed to ethanol. These experiments were performed 10 minutes after an acute intraperitoneal administration of ethanol (1.2 g/kg). (B) A western blot analysis demonstrating the reduction of p-Akt in the presence of acute ethanol exposure. (C) Reduced p-GSK3β in mice receiving acute ethanol exposure. These studies confirm that acute ethanol effects seen in vitro are operative in vivo as well. The experiments were performed three times with similar results. IP indicates immunoprecipitation.

Inhibition of PTEN Expression Reverses Ethanol's Inhibitory Effect on p-Akt Formation.

In order to obtain more direct evidence of the role of PTEN in acute ethanol-mediated effects on the insulin signaling cascade through PI3K/Akt, we performed knockdown experiments with siRNA directed against PTEN in comparison with cells transfected with a control RNA. As shown in Fig. 6, ethanol had inhibitory effects on p-Akt formation at the baseline and under insulin-stimulated conditions. However, the reduction of PTEN expression was associated with a loss of this inhibitory effect exhibited by ethanol at the baseline and under insulin-stimulated conditions. Thus, in this acute model, the critical role of PTEN was further demonstrated because there was prevention of acute ethanol effects on survival signaling through p-Akt.

Figure 6.

Knockdown of PTEN expression with siRNA prevents acute ethanol effects on p-Akt formation. In these experiments, Huh-7 cells were transfected with control RNA and siRNA specific for PTEN.30 After serum starvation overnight, the cells were exposed to ethanol for 20 minutes with or without insulin addition(s) for another 5 minutes. Thereafter, the cells were harvested in an RIPA buffer, and this was followed by a western blot analysis for PTEN and p-Akt. The results are the means of three independent experiments. The inhibition of PTEN expression abolishes the ethanol inhibitory effect on p-Akt formation both with and without insulin stimulation.

Discussion

It is well established that chronic ethanol consumption can cause sustained hepatocellular injury and alters the subsequent hepatic regenerative response.22, 23 There is increasing evidence that the toxic effects of ethanol may be mediated by impaired hepatocyte survival mechanisms by activating programmed cell death cascades.26 Previous studies in our laboratory revealed that chronic ethanol feeding of rats was associated with reduced levels of PY–IRS-1, and this event inhibited downstream PI3K kinase, Akt kinase, and Erk MAPK activities involved in hepatocyte proliferation and survival. In addition, there were increased levels of PTEN protein and messenger RNA (mRNA) and enhanced phosphatase activity in liver tissue. These results were supported in vitro, in which long-term ethanol exposure increased PTEN expression and function in primary hepatocyte cultures.26 Therefore, chronic ethanol exposure impairs survival mechanisms in the liver because of the inhibition of insulin signaling through PI3K and Akt, which is related, in part, to increased PTEN phosphatase expression and activity and IRS-1–dependent signaling pathways.26

The lipid kinase, PI3K, regulates cellular processes, including survival, growth, and proliferation. In this regard, PI3K phosphorylates the 3′-hydroxyl group on the inositol ring of the downstream lipid component phosphatidylinositol-4,5-biphosphate (PIP2) to produce a second messenger, namely, PIP3. This molecule activates Akt, which is a serine threonine kinase, and Akt enhances cell survival and proliferation by inhibiting the activities of the forkhead family of transcription factors. Thus, insulin signaling through PI3K inhibits apoptosis by activating Akt/protein kinase B,11, 12, 31, 32 which phosphorylates GSK3β33–35 and BAD,36 rendering them inactive. Low levels of Akt kinase and high levels of activated GSK3β and/or BAD are associated with increased programmed cell death.

There are several PIP3 phosphatases that act to inhibit PI3K signaling, such as PTEN. This protein phosphatase takes PIP3 back to the 4,5-biphosphate (PIP2) form. Therefore, insulin signaling may be inhibited by ethanol through alterations in PTEN expression at the transcriptional and protein level(s) in the liver.14, 37 The inactivation of PTEN leads to membrane recruitment and increased phosphorylation of Akt with attendant activation of Akt kinase.13, 16 Therefore, PTEN is considered to be a tumor suppressor gene and has been found to be altered in a number of malignancies.38, 39 The inactivation of the PTEN gene through mutational events allows constitutive up-regulation of the PI3K/Akt/GSK3β signal transduction cascade, which promotes cell survival signals and contributes to the generation of the malignant phenotype. In contrast, high levels of biologically active PTEN are associated with reduced levels of p-Akt, which subsequently increases GSK3β and BAD activity, thereby inhibiting survival signals.

In the normal liver, chronic ethanol consumption results in significantly increased levels of PTEN, mRNA, and protein expression and enhanced phosphatase activity. The higher levels of PTEN gene expression and increased levels of PTEN phosphatase activity may play a major role in chronic ethanol effects on hepatocyte cell survival. Indeed, an analysis of downstream signaling molecules related to PTEN function revealed increased levels of nuclear p53 and Fas receptor mRNA and increased levels of forkhead transcription factors in ethanol-exposed livers, although the protein was primarily localized in the cytosolic fraction.26 When these observations are taken together, one of the major effects of chronic ethanol consumption on the liver in vivo appears to be transcriptional up-regulation of PTEN and enhanced phosphatase activity leading to diminished signaling through Akt and downstream effector molecules to enhance proapoptotic pathways.

In acute ethanol exposure, the findings are somewhat different than those seen with long-term ethanol consumption with respect to alterations in these signaling cascades and the mechanism of action. Acute ethanol exposure in vitro does not change the IRβ protein and PY levels of this receptor in response to insulin stimulation. Similarly, acute ethanol exposure has little, if any, effect on the levels of PY–IRS-1 and the subsequent interaction with the p85α subunit of PI3K, as shown in Fig. 1. However, there is substantial evidence that acute ethanol exposure inhibits signaling through PI3K because, as shown in Fig. 2, there is a time-dependent reduction of insulin-stimulated phosphorylation of Akt in the presence of ethanol. Indeed, there is a significant reduction in p-Akt formation 5 and 15 minutes after insulin stimulation in the presence of ethanol additions. As expected, there is reduced phosphorylation of downstream molecules such as GSK3β and BAD, as shown in Fig. 4, which renders these molecules less active. The net effect, therefore, is to reduce survival signals. These results were extended and confirmed by in vivo studies, as shown in Fig. 6. Acute ethanol exposure in the intact animal reduces endogenous insulin signaling through Akt and GSK3β levels in the liver because it promotes the association of PTEN with PI3K p85α.

What are the possible mechanisms of ethanol's adverse effects on survival signaling? In contrast to chronic ethanol exposure,26 there is no increase in PTEN protein levels in Huh-7 cells, as shown in Fig. 1. However, there is a shift leading to an increased interaction of PI3K p85α with PTEN, as shown in Fig. 3. Furthermore, acute ethanol exposure alone rapidly increases this association in as little as 5 minutes in the absence of insulin stimulation. In this context, ethanol substantially augments the interaction of PI3K p85α with PTEN under conditions of insulin stimulation. Furthermore, it became of interest to determine if the binding to PTEN by PI3K is through the regulatory p85α subunit, the catalytic p110α subunit, or both. As shown in Fig. 3C, the association is clearly through the p85α regulatory subunit and not p110α.

To provide more direct evidence that PTEN mediates an ethanol-induced reduction in survival signals to the liver, we performed PTEN knockdown experiments. The results shown in Fig. 6 suggest that the reduction of PTEN expression abolishes the ethanol effect and enhances survival signaling, as shown by an increase in p-Akt formation. This study further supports the critical role of PTEN in the acute ethanol-mediated process for insulin survival signaling. These results are similar to those of Shulga et al.,30 who observed that the suppression of elevated PTEN levels that developed during the prolonged exposure of hepatocellular carcinoma cells to ethanol, through the use of RNA interference, substantially reversed ethanol-induced alterations in tumor necrosis factor signaling, resulting in the preservation of mitochondrial function and cell viability.30

Insulin is one of the key hormones that regulate liver growth and survival. Insulin binds to its receptor and induces by autophosphorylation the activation of the kinase, which, in turn, tyrosyl-phosphorylates IRS-1 and IRS-2. Following PY of the C-terminus of IRS-1 and IRS-2, the recruitment and activation of PI3K to specific binding motifs occur. It is known that PTEN becomes activated after binding to lipid substrate PIP240 through a direct interaction. However, it has not been previously revealed that the PI3K p85α subunit may also directly interact with PTEN, as shown herein, and how ethanol enhances this interaction. As demonstrated by the scheme presented in Fig. 7, under normal insulin signaling conditions, IRS-1 binds to the P85α subunit of PI3K, which, in turn, phosphorylates Akt, increasing its activity. This leads to downstream phosphorylation of GSK3β and BAD to send prosurvival signals. In the model presented in Fig. 7, it is proposed that acute ethanol exposure rapidly promotes the association of PTEN with a p85α subunit of PI3K, thus shifting the equilibrium from the IRS-1/PI3K p85α complex toward the PTEN/PI3K p85α interaction. It is hypothesized that enhanced PTEN binding to the p85α regulatory subunit results in reduced Akt phosphorylation (as shown by Figs. 2 and 6). The reduced phosphorylation of Akt inhibits the phosphorylation of downstream molecules such as GSK3β and BAD, changing them into an active form. Under these conditions, there is a shift from prosurvival signals to proapoptotic signals, and this renders the liver susceptible to the other toxic effects of ethanol produced by increased oxidative stress and lipid peroxidation.28

In summary, acute ethanol exposure both in vitro and in vivo has surprising and substantial effects on insulin-mediated survival signaling in the liver. By yet unknown mechanisms, acute ethanol exposure rapidly increases the association of PTEN with the p85α regulatory subunit of PI3K, even though the cellular levels of PTEN remain unchanged. This shifts the balance from the IRS-1/PI3K p85α complex toward the PTEN/PI3K p85α complex through a direct interaction that subsequently leads or contributes to reduced phosphorylation of Akt, thus activating GSK3β and BAD to promote programmed cell death pathways. Further studies of such signaling pathways and the adverse effects of acute ethanol exposure may lead to a better understanding of the toxic effects of acute ethanol exposure on the liver.

Acknowledgements

The authors thank Donna Pratt for her editorial assistance in producing this publication.

Ancillary