Acetate, the key modulator of inflammatory responses in acute alcoholic hepatitis

Authors


  • Supported by the Medical Research Council (UK) Clinical Research Training Fellowship, G0500473 (to S.F.W.K.); European Research Advisory Board Project Grant EA 06 39. Funders had no role in study design or the collection, analysis, and interpretation of data.

  • Potential conflict of interest: Nothing to report.

Abstract

Acute alcoholic hepatitis is characterized by disproportionate macrophage inflammatory cytokine responses to bacterial lipopolysaccharide. Lack of knowledge of the underlying mechanism has limited progress toward effective therapy. We postulated a novel mechanism by which ethanol increases histone acetylation, increasing proinflammatory gene transcription and cytokine synthesis. Cytokine responses to lipopolysaccharide in a human macrophage cell line cultured in 86 mM ethanol, 1 mM acetate, and normal media were measured by multiplex immunoassay. Changes in histone acetylation were determined by immunofluorescence microscopy and chromatin immunoprecipitation on presentation. The effect of ethanol and acetate on acetyl-coenzyme A (acetyl-coA) synthetases, which convert acetate to acetyl-coA, the substrate for histone acetylation, was determined by quantitative reverse-transcription polymerase chain reaction and immunoblotting. Knockdown of acetyl-coA synthetases by short hairpin RNA (shRNA) was used to determine their role in ethanol's enhancement of the inflammatory cytokine response. Ethanol-exposed macrophages developed enhanced interleukin 6 (IL6), IL8, and tumor necrosis factor alpha responses to lipopolysaccharide with time-dependent increases in histone acetylation that could be prevented by inhibition of ethanol metabolism. Chromatin immunoprecipitation confirmed increased histone acetylation at promoter regions of specific cytokine genes. The effect of ethanol was reproduced by incubation with acetate, the principal hepatic metabolite of ethanol, and both ethanol and acetate reduced histone deacetylase activity and up-regulated acetyl-coA synthetases. Knockdown of the acetyl-coA synthetases abrogated the effect of ethanol on cytokine production. Conclusion: Synthesis of metabolically available acetyl-coA from acetate is critical to the increased acetylation of proinflammatory gene histones and consequent enhancement of the inflammatory response in ethanol-exposed macrophages. This mechanism is a potential therapeutic target in acute alcoholic hepatitis. (HEPATOLOGY 2010)

Alcoholic liver disease (ALD) is a significant and growing global health problem. Clinical liver failure in ALD can result from chronic hepatocyte injury producing cirrhosis or from rapid, acute hepatocellular dysfunction secondary to inflammation in acute alcoholic hepatitis.AAH This acute inflammatory form of ALD carries a mortality of up to 35% on first presentation, killing patients before they have the opportunity to reap the benefits of appropriate health education and subsequent abstinence from alcohol.1

Our current understanding of the pathogenesis of AAH attributes hepatocellular dysfunction to the action of supraphysiological concentrations of proinflammatory cytokines on hepatocytes that are already suffering oxidative and endoplasmic reticulum stress due to the reactive products of ethanol metabolism.2 The major source of cytokine release is thought to be hepatic macrophage or Kupffer cells responding, by way of Toll-like receptors (TLRs), to the increased concentration of bacterial endotoxin in portal blood that results from an ethanol-mediated increase in gut permeability.3

Evidence for the role of endotoxin, TLRs, and cytokines in this mechanism is well established.4 Increased gut permeability is a feature of ALD and plasma lipopolysaccharide (LPS) is elevated in all stages of ALD, levels correlating with clinical severity and outcome. The principal LPS receptor, TLR4, is up-regulated by chronic ethanol treatment in humans and both C3H/HeJ mice lacking TLR4 and animals deficient in the CD14 coreceptor show relative protection from ethanol-induced liver injury in comparison with wildtype animals.5, 6 AAH patients have significantly elevated serum cytokines, particularly interleukin 6 (IL-6), IL-8, and tumor necrosis factor alpha (TNF-α), with levels correlating with prognosis.7-9 Their ex vivo monocyte responses to LPS are significantly enhanced relative to controls and this LPS hyperresponsiveness can be reproduced in vitro by exposure of the human macrophage cell line MonoMac6 to ethanol for 6 days.10

The enhanced and sustained inflammatory response seen in AAH is, however, in complete contradistinction to the normal processing of portal endotoxin by the liver.11 The liver is normally subject to tonic endotoxin exposure by way of the portal vein and it is effective at clearing this endotoxin from the blood without an inflammatory response. The phenomenon of “endotoxin tolerance” thereby renders endotoxin-exposed Kupffer cells refractory to further LPS stimulation, maintaining an anti- rather than proinflammatory cytokine output.12 It is therefore somewhat unexpected that the proinflammatory response to endotoxin in AAH should be so disproportionately high, particularly considering that it is the Kupffer cells themselves that are key to maintaining hepatic endotoxin tolerance.13 It has become increasingly clear, therefore, that the enhancement of cytokine gene expression and perpetuation of the inflammatory response is the key event in the pathogenesis of AAH.14

Despite its clear importance for the pathogenesis of AAH, the mechanism for enhanced inflammatory cytokine release in this disease remains unclear. In this study we address the novel hypothesis that the enhanced inflammatory cytokine response results from the direct actions of ethanol itself on the final common pathway of cytokine gene transcriptional regulation by histone acetylation.

In its untranscribed state DNA is tightly coiled around histone protein octamers and the resulting chromatin is compacted into a closed tertiary structure from which the histone tails protrude, but in which the DNA is inaccessible to polymerases involved in gene transcription. Gene activation by transcription factors involves coactivator proteins with histone acetyl transferase (HAT) activity that acetylate key lysine residues in the histone tails. The negatively charged acetyl groups cause a conformational change in chromatin that allows RNA polymerases access to the DNA, facilitating gene transcription. Termination of transcription is mediated through histone deacetylases (HDAC), which release free acetate and allow the chromatin to resume its closed, untranscribed conformation.15 Various HDACs are able to modulate inflammatory gene transcription, including class I and II HDACs, which can be recruited by transcriptional repressors such as the activated glucocorticoid receptor and class III HDACs, known as sirtuins (SIRT), which are active in the presence of nicotinamide adenine dinucleotide (NAD+).16 Ethanol has been demonstrated to increase total histone acetylation in rat liver17 with increased HAT and reduced HDAC activity18 and separate investigations have established that both SIRT expression and activity can be inhibited by ethanol in the liver.19, 20

Hepatocyte ethanol metabolism produces free acetate as its endproduct which, largely in other tissues, can be incorporated into acetyl-coenzyme A (acetyl-coA) for use in Krebs cycle oxidation, fatty acid synthesis, or as a substrate for protein acetylation.21 This conversion is catalyzed by the acetyl-coA synthetases,22 recently redesignated acyl-coenzyme A synthetase short-chain family members 1 and 2 (ACSS1 [UniProt Q9NUB1] and ACSS2 [Q9NR19]).23

The role of acetyl-coA synthesis in control of inflammation has not previously been investigated and could open a novel field of study into the relationship between cellular energy supply and inflammatory disease. In this study we test the hypothesis that ethanol enhances macrophage cytokine production by uncoupling gene transcription from its normal regulatory mechanisms through increased histone acetylation, and that the conversion of the ethanol metabolite acetate to acetyl-coA is crucial to this process.

Abbreviations

AAH, acute alcoholic hepatitis; ACSS, acyl-coenzyme A synthetase short-chain; ALD, alcoholic liver disease; ER, endoplasmic reticulum; HAT, histone acetyltransferase; HDAC, histone deacetylase; IL, interleukin; LPS, lipopolysaccharide; NAD, nicotinamide adenine dinucleotide; ROS, reactive oxygen species; SIRT, sirtuin; SREBP, sterol response element binding protein; TLR, Toll-like receptor; TNF, tumor necrosis factor.

Materials and Methods

Cell Culture.

The human monoblastic cell line MonoMac6 (DSMZ, Braunschweig, Germany, ACC124), an established human cell line that displays features of mature macrophages and has been used to model Kupffer cell responses in ethanol,10 was grown in RPMI-1640 medium supplemented with 2 mM l-glutamine (Lonza, Basel, Switzerland), 1 mM sodium pyruvate, 1× nonessential amino acids (both Gibco, Paisley, UK), 10% fetal calf serum (Lonza), and 9 μg/mL human insulin (Sigma, St. Louis, MO). Ethanol exposure was achieved in fresh media with 86 mM (0.5%, 400 mg/dL) ethanol (VWR, Poole, UK). This is five times the legal blood alcohol limit for driving in the UK and equivalent to heavy drinking in humans.24 The ethanol concentration was maintained by using ethanol vapor in the incubator to prevent evaporation of ethanol from culture media10 and monitored by potassium dichromate reduction assay (BioAssay Systems, Hayward, CA). For acetate culture experiments, media were supplemented with 1 mM sodium acetate (Sigma) and replaced every 48 hours to minimize fluctuations in acetate concentration. One mM is an achievable acetate concentration in an individual metabolizing ethanol at the concentration used.25

LPS Stimulation and Cytokine Determination.

After 7 days incubation cells were resuspended in fresh medium at 2 × 106/mL and stimulated with E. coli O111:B4 LPS (InvivoGen) at a final concentration of 10 ng/mL. For messenger RNA (mRNA) determination, nucleic acids were isolated in binding columns (Stratagene, La Jolla, CA) and cytokine transcripts measured by quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR) using AffinityScript reverse transcriptase with random primers (Stratagene) and TaqMan polymerase, primers, and probes in a AB7000 cycler using 18s RNA as the endogenous control (all Applied Biosystems, Foster City, CA). Cytokine protein determination was achieved by multiplex electrochemoluminescent immunosorbent assay using the MesoScale Discovery System (MesoScale, Gaithersburg, MD).

Immunofluorescence Microscopy.

After 0-6 days in 86 mM ethanol, cells were adhered to slides by centrifugation at 1000 rpm for 3 minutes (Cytospin3, Shandon, Runcorn, UK). Adherent cells were fixed and permeabilized in ice-cold methanol and acetone, blocked in 3% bovine serum albumin (BSA), and incubated with primary antibodies to total acetyl lysine (Cell Signaling Technology, Danvers, MA; 1 in 200), acetyl-histone H3 (Upstate, Lake Placid, NY; 5 μg/mL), and acetyl-histone H4 (Upstate, 10 μg/mL) at 4°C for 18 hours. Slides were washed and incubated with FITC-conjugated secondary antibody (goat antirabbit IgG, Sigma; 1 in 200) and counterstained with DAPI (Vectashield Hardset, Vector Laboratories, Burlingame, CA). The effect of ethanol metabolism on the staining pattern was assessed by incubation in 86 mM ethanol supplemented with the alcohol dehydrogenase inhibitor 4-methylpyrazole 1 mM (Sigma).

Protein Kinase Inhibition.

Ethanol culture was repeated in the presence of inhibitors of the stress-activated protein kinases MEK (U0126 5 nM, Calbiochem, Darmstadt, Germany) and JNK (SP600125 25 nM, Calbiochem). Western blotting was performed with antibodies to acetyl-histone H3 (Upstate) to determine global histone H3 acetylation, phospho-MEK, and phospho-JNK (Cell Signaling Technology) to confirm successful inhibition, and β-actin (Sigma) to confirm equal loading.

Chromatin Immunoprecipitation.

Chromatin immunoprecipitation (ChIP) was used to detect ethanol-induced changes in histone acetylation at specific proinflammatory cytokine gene promoter regions. Cells were lysed in a Dounce homogenizer and intact nuclei isolated by sucrose density centrifugation. Chromatin was digested by micrococcal nuclease (Amersham, Little Chalfont, UK) to yield a mononucleosome suspension that was precleared with Zysorbin staphylococcal protein A membranes (Invitrogen) and aliquots of the resulting supernatant incubated with anti-acetyl-histone H3 and anti-acetyl-histone H4 antibodies (Upstate) overnight at 4°C. Antibody-bound mononucleosomes were precipitated out using Zysorbin. DNA was extracted from the precipitates and from the unprecipitated input fraction and the relative concentration of IL-6 and TNF-α promoter DNA in the extracts determined by SYBRGreen quantitative PCR (primers: IL6 forward GAGCAGTGGCTTCGTTTCAT, reverse TTGGGGAAAGTGAGGTCATC; TNF-α forward TGTCCAGGGCTATGGAAGTC, reverse TTTCATTCTGACCCGGAGAC).

HAT and HDAC Activity.

HAT activity was measured by an enzyme-linked immunosorbent assay (ELISA)-based method (Millipore, Temecula, CA) and HDAC activity was measured by color change on deacetylation of an acetylated substrate (Biomol, Plymouth Meeting, PA) according to the manufacturers' instructions.

ACSS1 and 2 Determinations.

ACSS1 and 2 transcripts were measured by qRT-PCR as above. Protein determination was by western blotting using anti-ACSS1 (Abnova, Taipei, Taiwan) and anti-ACSS2 (Atlas, Stockholm, Sweden) primary antibodies with anti-β-actin (Abcam, Cambridge, UK) used to confirm equal loading. Secondary antibodies were horseradish peroxidase (HRP)-conjugated goat antimouse IgG and goat antirabbit IgG (Sigma) and bands were identified using SuperSignal West Pico chemiluminescent substrate (Thermo Scientific, Rockford, IL) for ACSS2 and Immobilon Western chemiluminescent substrate (Millipore, Billerica, MA) for ACSS1. Band densitometry was performed using Scion Image (Frederick, MD).

ACSS1 and 2 Knockdown.

The contribution of ACSS1 and 2 to the effect of ethanol on cytokine output was determined by knockdown with short hairpin RNA (shRNA, Sigma). shRNA for ACSS1 or 2 was delivered by a lentiviral vector at a multiplicity of infectivity (MOI) of 5. MonoMac6 cells were treated with hexadimethrine bromide (Sigma) 8 μg/mL to increase transduction efficiency and then incubated with the vector for 18 hours. Stably transduced lines were selected by cotransduced resistance to puromycin 5 μg/mL. After five passages in selection medium, knockdown was assayed by qRT-PCR and western blotting. Stable knockdowns of ACSS1, ACSS2, and a double-knockdown were prepared and subjected to ethanol incubation, LPS stimulation, and multiplex cytokine immunoassay as above. Transduced cells were compared to untransduced and effects of shRNA on cellular function were controlled for by comparison with lines transduced with irrelevant shRNA constructs at the appropriate MOI in accordance with the principles laid down by the 2003 Horizon symposium.26 The effect of ACSS1 and 2 knockdown on global histone H3 and H4 acetylation after 7 days culture in 86 mM ethanol or 1 mM acetate was determined by western blotting with antibodies to acetyl-histone H3 and H4 (both Upstate) and total histone H3 and H4 (both Abcam).

Statistical Analysis.

Numerical results were expressed as means of at least three samples and statistical significance assessed by the Mann-Whitney U test.

Results

Enhancement of Inflammatory Cytokine Responses by Ethanol.

Monomac6, an established human macrophage cell line modeling Kupffer cell responses in ethanol,10 was maintained in a validated constant-exposure ethanol culture system at an ethanol concentration of 86 mM, equivalent to human blood concentrations after heavy drinking. Using this system we demonstrated enhancement of the cytokine response to E. coli LPS 10 ng/mL compared to cells grown in normal medium. This was not seen with acute ethanol exposure, but after 7 days culture in ethanol we observed significant augmentation of IL6, IL8, and TNF-α release following LPS exposure (Fig. 1A). Cytokine mRNA expression was also increased (Fig. 1B). The effect of ethanol on cytokine output was reversible with transfer of hyperresponsive cells from ethanol to normal medium causing the cytokine response to LPS to normalize within 4 days (data not shown). The culture system adopted therefore replicates previous reports of the augmentation of inflammatory cytokine release in the context of chronic ethanol exposure in animal models and human patients with AAH, and demonstrates that this results, at least in part, from an increase in the level of inflammatory cytokine gene transcription.

Figure 1.

Enhancement of cytokine responses to LPS in ethanol and acetate. Monomac6 cells were cultured in normal media or in the presence of 86 mM ethanol (A,B) or 1 mM acetate (C) for 7 days and then stimulated with E. coli LPS 10 ng/mL. Cells were harvested for cytokine mRNA determination by qRT-PCR at 60 minutes (B) and supernatants for cytokine protein determination by multiplex immunoassay at 48 hours (A,C). Values shown are mean + SEM for three independent determinations.

Global Acetylation Increases in Ethanol.

Next we examined whether the enhanced inflammatory responses associated with chronic ethanol exposure were associated with increased histone acetylation. Immunofluorescence microscopy for total acetylated lysine residues, acetyl-histone H3, and acetyl-histone H4 revealed a time-dependent increase in acetylation over 6 days culture in 86 mM ethanol (Fig. 2A). Coculture with the inhibitor of ethanol metabolism 4-methylpyrazole in the ethanol-containing medium reduced the acetylation staining to baseline, suggesting that ethanol metabolism rather than simply ethanol exposure was responsible for the acetylation changes. Global increases in acetyl-histone H3 and H4 after 7 days ethanol culture were also demonstrated by western blotting. This effect was not abrogated by inhibition of the MEK and JNK stress-activated protein kinases previously demonstrated to increase histone H3 acetylation in the presence of ethanol,27 suggesting that a separate mechanism is responsible for the increased acetylation in this setting (Supporting online Figs. 1, 2). These observations demonstrate that ethanol metabolism by mononuclear cells is associated with increased histone acetylation, with a time course similar to the cytokine enhancement, and which is dependent on the metabolism of ethanol but not on MEK and JNK kinase signaling.

Figure 2.

Increased histone acetylation in ethanol. (A) Monomac6 cells were cultured in 86 mM ethanol and examined for acetylation modifications (green) after 0-6 days by immunofluorescence microscopy. Modifications stained were total acetylated lysine residues (top row), acetylated histone H3 (middle row), and acetylated histone H4 (bottom row). Nuclei were stained blue with DAPI. The effect of ethanol metabolism on the modifications was demonstrated by coculture with the alcohol dehydrogenase inhibitor 4-methylpyrazole (4-MP) for 6 days (right column). (B) Histone H3 and H4 acetylation at IL6 and TNF-α promoters after 6 days in 86 mM ethanol was studied by chromatin immunoprecipitation. Immunoprecipitates generated using anti-acetyl-histone H3 and H4 antibodies were assayed by qPCR for enrichment of promoter regions of IL6 and TNF-α genes relative to precipitates from cells cultured in normal media. Enrichment of the precipitate from a given sample with DNA from the IL6 and TNF-α promoters reflects the degree to which these were associated with histones bearing the modification of interest.

Increased Histone Acetylation at Specific Cytokine Gene Promoters.

The immunofluorescence microscopy revealed global increases in histone acetylation. To determine whether this specifically included increased acetylation of the crucial promoter regions of proinflammatory cytokine genes we performed chromatin immunoprecipitation on cells cultured in ethanol and control cells cultured in normal medium. The immunoprecipitates produced by anti-acetyl-histone H3 and anti-acetyl-histone H4 antibodies from the monococcal nuclease-digested chromatin of ethanol-exposed cells were enriched for DNA from the promoter regions of the IL6 and TNF-α genes relative to immunoprecipitates from unexposed cells (Fig. 2B). This confirmed that increased histone H3 and H4 acetylation was present at these proinflammatory cytokine gene promoters after 7 days culture in 86 mM ethanol, providing a mechanism for increased cytokine transcription in response to LPS stimulation.

Reproduction of the Ethanol Effect by Acetate.

A potential mechanism for the effect of ethanol exposure on histone acetylation status would be through increased exposure to acetate (the principal hepatic metabolite of ethanol). In order to address this mechanism we explored the extent to which coculture with acetate could replicate the ethanol effect on histone acetylation. When cells were cultured in 1 mM acetate for 7 days and then stimulated with LPS we observed an augmentation of cytokine release similar in magnitude to that seen in the ethanol-exposed cells for IL6, IL8, and TNF-α (Fig. 1C). Western blotting demonstrated that 7-day culture in 1 mM acetate or 86 mM ethanol produced similar global increases in acetylated histones H3 and H4 (Fig. 6, left-hand panel). That exposure to acetate can replicate both the enhanced cytokine responses and the increased histone acetylation seen following prolonged ethanol metabolism suggests that exposure to acetate (or one of its metabolites) is likely to be critical for increased histone acetylation in the context of ethanol exposure/AAH.

HAT and HDAC Activity.

We next tested whether ethanol or acetate were acting by influencing the balance of HAT and HDAC activity in the cells. Addition of 86 mM ethanol or 1 mM acetate to fresh lysate of MonoMac6 cells significantly reduced HDAC activity within 30 minutes and produced a nonsignificant increase in HAT activity, a situation favoring net increase in histone acetylation (Fig. 3). In lysates from cells exposed to 86 mM ethanol or 1 mM acetate for 7 days, assays revealed a nonsignificant trend toward reduced HDAC activity and increased HAT activity (Supporting online Fig. 3).

Figure 3.

HDAC and HAT activity in ethanol and acetate. Monomac6 cell lysates were incubated with control media, 86 mM ethanol, or 1 mM acetate for 30 minutes at 37°C in the presence of an acetylated substrate which absorbed light at 405 nm upon deacetylation (A) or a measured quantity of unacetylated histone H3, an excess of acetyl-coA and trichostatin A to inhibit competing HDAC activity, with acetylated histone H3 being assayed after incubation by ELISA with absorbance at 450 nm (B). Substrate deacetylation or acetylation was quantified by use of a standard curve in each case. Values shown are mean + SEM for three independent determinations.

Induction of ACSS1 and 2 by Ethanol and Acetate.

Free acetate has little metabolic activity and is more likely to influence cellular responses as the metabolically active acetyl-coA, synthesized from acetate by ACSS1 and 2. ACSS1 and 2 transcripts were significantly more abundant in cells incubated in 86 mM ethanol for 7 days than in control cells (Fig. 4A). At the protein level, western immunoblotting identified induction of ACSS1 and 2 from 6 days culture in ethanol. A similar induction was observed in 1 mM acetate but was apparent at 24 hours (Fig. 4B). This demonstrates, for the first time, that macrophages have the potential to increase synthesis of metabolically active acetyl-coA during ethanol exposure, making additional acetyl-coA available for use by HAT enzymes and the Krebs cycle.

Figure 4.

Induction of acetyl-coA synthetases (ACSS1 and 2) by ethanol and acetate. (A) Monomac6 cells were cultured in normal media or media with 86 mM ethanol for 7 days and assayed for ACSS1 and 2 mRNA by qRT-PCR. Values shown are mean + SEM for three independent determinations. (B) Monomac6 cells were cultured in 86 mM ethanol or 1 mM acetate for 0-7 days before lysis and proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and identified by immunoblotting for ACSS1 and 2 with β-actin to control for differences in loading. Increases in band density over untreated cells were quantified relative to β-actin by band densitometry. The increase in expression after 7 days in ethanol mirrors the enhancement of cytokine responses observed in ethanol.

ACSS1 and 2 Knockdown Abrogates the Effect of Ethanol.

To confirm that conversion of acetate to acetyl-CoA is crucial to the acetylation-mediated potentiation of inflammatory responses in ethanol we performed shRNA knockdown of ACSS1 and 2. Western immunoblotting confirmed stable knockdown of ACSS1, ACSS2, and the double ACSS1+2 knockdown at the protein level (Fig. 5A). The enhancement of cytokine output after incubation in 86 mM ethanol was markedly diminished by ACSS knockdown, most significantly in the double ACSS1+2 knockdown cells. Cytokine output from the double knockdown cells was significantly lower than from the cells transduced with irrelevant transcript shRNA constructs at an equal multiplicity of infectivity (Fig. 5B). Western blotting demonstrated that the double ACSS1+2 knockdown abrogated the increase in acetylated histone H3 and H4 induced by either ethanol or acetate (Fig. 6). These findings demonstrate that the augmented inflammatory cytokine response seen in alcoholic hepatitis occurs as a direct consequence of the metabolic effects of ethanol exposure modifying the acetylation-dependent transcriptional regulation of these cytokines by way of acetate and its conversion to acetyl-coA.

Figure 5.

Effect of ACSS1 and 2 knockdowns on inflammatory cytokine responses in ethanol. shRNA delivered by lentiviral vector was used to achieve knockdown of ACSS1, ACSS2, and a double ACSS1+2 knockdown in Monomac6 cells. Controls were transfected with irrelevant shRNA at an MOI of 5 for the single knockdowns and 10 for the double knockdown. Stably transfected lines were selected by cotransfected puromycin resistance. Effective knockdown of ACSS1 and 2 protein expression was confirmed by SDS-PAGE and immunoblotting with β-actin as a loading control (A). Stable knockdowns, controls, and untransfected cells were cultured in 86 mM ethanol for 7 days and then stimulated with E. coli LPS 10 ng/mL. Supernatants were harvested for cytokine protein determination by multiplex immunoassay at 48 hours. Cytokine output was compared between knockdowns and their relevant controls. Values shown are mean + SEM for three independent determinations.

Figure 6.

Effect of ACSS1 and 2 knockdown on histone acetylation changes in ethanol and acetate. Untransfected Monomac6 cells and stable knockdowns for ACSS1+2 were cultured in normal media, 86 mM ethanol, or 1 mM acetate for 7 days and then assayed for total and acetylated histone H3 and H4 by western blotting. The left-hand panel demonstrates the increase in global H3 and H4 acetylation in ethanol and acetate and the right-hand panel demonstrates the loss of this effect in the ACSS1+2 knockdown cells.

Discussion

Our observation of increased histone acetylation in a macrophage cell line after ethanol treatment is consistent with the findings of other groups that have recently demonstrated that ethanol increases histone acetylation in hepatocytes,28 hepatic stellate cells,29 and whole rat tissues.17 However, the present study is the first demonstration of ethanol modulation of gene expression in inflammatory cells by a mechanism dependent on histone acetylation. This increased acetylation could, in principle, arise through a number of routes. Ethanol metabolism, particularly at higher concentrations, produces a significant burden of reactive oxygen species (ROS)30 and endoplasmic reticulum (ER) stress.31 ROS can directly activate transcription factors such as nuclear factor kappaB (NF-κB)p6532 and oxidative and ER stress can favor a proinflammatory transcription factor milieu.33 NF-κBp65 will recruit HAT coactivators to proinflammatory gene promoters and increase histone acetylation. Additionally, oxidative stress is known to inhibit HDAC recruitment to actively transcribed chromatin.34 However, critically, we have shown that exposure to acetate, the principal hepatic endproduct of ethanol metabolism, can fully mimic the effects on cytokine production seen with ethanol. As cytokine potentiation can occur without the ROS-generating metabolism of ethanol to acetate, then oxidative stress cannot be solely responsible for the enhanced inflammatory response to ethanol.

We observed that both ethanol and its metabolite acetate could reduce HDAC activity in a cell-free system, with a similar pattern observed in cell lysates after 7 days culture. Free acetate is the endproduct of histone deacetylation, so acetate may increase histone acetylation through endproduct inhibition of HDACs. Acetate may also increase histone acetylation through increased HAT activity. This could be through increased substrate supply (although for this acetate must be in the form of acetyl-coA) or indirectly through reduced HDAC activity. HDACs have a role in deacetylation of NF-κBp65 leading to a reduction in its ability to recruit HAT coactivators, so reduced HDAC activity can lead to increased HAT recruitment.35

Ethanol and acetate might also influence total HDAC activity by modulating the activity of SIRTs. These are class III HDACs whose activity is dependent on the presence of NAD+ and which are increasingly recognized as a vital link between energy supply, gene expression, cellular activity, and cellular aging.16 There is recent evidence that ethanol can reduce total HDAC activity.18 Metabolism of ethanol to acetate results in NAD+ depletion which will reduce SIRT and hence total HDAC activity. Free acetate will not affect NAD+, but once converted to acetyl-coA it can enter the Krebs cycle and convert NAD+ to NADH in the same way as if it had come from glycolysis or fatty acid oxidation. Acetate is converted to acetyl-coA through the action of the ACSS1 and 2 enzymes whose activity is also dependent on SIRT activity.36 Acetyl-coA synthetases can be induced by acetate in prokaryotes,37 although free acetate can also down-regulate ACSS1 and 2 through reduced SREBP transcription.38 We have demonstrated up-regulation of ACSS1 and 2 in this human cell line by acetate and, at a slower rate, by ethanol. Knockdown of these enzymes by shRNA significantly diminished ethanol's enhancement of cytokine responses to LPS and we infer that the supply of acetyl-coA from free acetate by ACSS1 and 2 makes a significant contribution to the increased inflammatory cytokine responses seen after chronic ethanol exposure. ACSS1 and 2 play no role in acetyl-coA synthesis from glycolysis and, in the presence of ample glucose and insulin, their knockdown should not diminish the normal metabolic flux of glycolytic acetyl-coA but will reduce the additional acetyl-coA load derived from ethanol metabolism. They therefore represent a potential therapeutic target through which the additional acetyl-coA load (and hence the inflammatory response) could be reduced without interference with cellular respiration.

We have demonstrated that increased histone acetylation in ethanol is a nonspecific, genomewide phenomenon. The ChIP assays confirm that increased acetylation is present at inflammatory cytokine promoters, but in no way suggest that the process is specific for these regions. It seems likely that increased acetylation enhances the expression of those genes whose transcription is activated by the well-established LPS/TLR/oxidative stress pathways. Which genes are activated is determined by receptor-transcription factor signaling but it is the massive enhancement of the response to that signaling that we suggest to be acetylation-dependent.

This study focused on three cytokine genes previously demonstrated to play a significant pathogenic and prognostic role in AAH. However, the nonspecific nature of the increase in histone acetylation means that the expression of any number of other genes may be potentiated in the same way by chronic ethanol exposure. This is likely to include mediators with a more antiinflammatory action and/or profibrotic effect, which will influence the degree of inflammation and eventual organ dysfunction. The net effect of the ethanol-dysregulated cytokine milieu and its modulation by compounds affecting HDAC, SIRT, or ACSS activity will need to be studied at the whole-organism level.

In summary, these findings are evidence for a mechanism of enhanced inflammation in acute alcoholic hepatitis in which acetyl-coA synthetases are up-regulated and convert the ethanol metabolite acetate to an excess of acetyl-coA which increases proinflammatory cytokine gene histone acetylation by increased substrate concentration and HDAC inhibition, leading to enhanced gene expression and perpetuation of the inflammatory response. The clinical implication of these findings is that modulation of HDAC or ACSS activity might affect the clinical course of alcoholic liver injury in humans. We have already demonstrated that the drug theophylline, which can improve HDAC recruitment to actively transcribed chromatin, has the potential to improve ex vivo lymphocyte steroid sensitivity in acute alcoholic hepatitis.39 If HDAC activators or inhibitors of ACSS1 and 2 can modulate ethanol-associated histone changes without affecting the flow of acetyl-coA through the normal metabolic pathways, then they have the potential to become much-needed effective therapeutic options in acute alcoholic hepatitis.

Acknowledgements

S.F.W.K. conceived and designed the study, carried out experimental work, analyzed the data, and wrote the report. G.O'B. contributed to the experimental work, data analysis, and final report. J.M. developed the ChIP assays, contributed to data analysis, and final report. M.Z. performed histone extraction and immunoblots and contributed to data analysis. J.P. contributed to the design of cytokine assays and to the report. D.E.J.J. and C.P.D. contributed to study design, data analysis, and the report. All authors approved the final report for submission.

Ancillary