S-adenosylhomocysteine sensitizes to TNF-α hepatotoxicity in mice and liver cells: A possible etiological factor in alcoholic liver disease

Authors

  • Zhenyuan Song,

    1. Department of Medicine, University of Louisville College of Medicine, Louisville, KY
    Search for more papers by this author
  • Zhanxiang Zhou,

    1. Department of Medicine, University of Louisville College of Medicine, Louisville, KY
    Search for more papers by this author
  • Silvia Uriarte,

    1. Department of Medicine, University of Louisville College of Medicine, Louisville, KY
    Search for more papers by this author
  • Lipeng Wang,

    1. Department of Pharmacology and Toxicology, University of Louisville College of Medicine, Louisville, KY
    Search for more papers by this author
  • Y. James Kang,

    1. Department of Medicine, University of Louisville College of Medicine, Louisville, KY
    2. Department of Pharmacology and Toxicology, University of Louisville College of Medicine, Louisville, KY
    Search for more papers by this author
  • Theresa Chen,

    1. Department of Pharmacology and Toxicology, University of Louisville College of Medicine, Louisville, KY
    Search for more papers by this author
  • Shirish Barve,

    1. Department of Medicine, University of Louisville College of Medicine, Louisville, KY
    2. Department of Pharmacology and Toxicology, University of Louisville College of Medicine, Louisville, KY
    Search for more papers by this author
  • Craig J. McClain

    Corresponding author
    1. Department of Medicine, University of Louisville College of Medicine, Louisville, KY
    2. Department of Pharmacology and Toxicology, University of Louisville College of Medicine, Louisville, KY
    3. Department of Veterans Affairs Medical Center, Louisville, KY
    • Department of Medicine, University of Louisville Medical Center, Louisville, KY 40292
    Search for more papers by this author
    • fax: 502-852-0846


Abstract

In alcoholic liver disease, tumor necrosis factor-α (TNFα) is a critical effector molecule, and abnormal methionine metabolism is a fundamental acquired metabolic abnormality. Although hepatocytes are resistant to TNFα-induced killing under normal circumstances, previous studies have shown that primary hepatocytes from rats chronically fed alcohol have increased TNFα cytotoxicity. Therefore, there must be mechanisms by which chronic alcohol exposure “sensitizes” to TNFα hepatotoxicity. S-adenosylhomocysteine (SAH) is product of methionine in transsulfuration pathway and a potent competitive inhibitor of most methyltransferases. In this study, we investigated the effects of increased SAH levels on TNFα hepatotoxicity. Our results demonstrated that chronic alcohol consumption in mice not only decreased hepatic S-adenosylmethionine levels but also increased hepatic SAH levels, which resulted in a significantly decreased S-adenosylmethionine-to-SAH ratio. This was associated with significant increases in hepatic TNFα levels, caspase-8 activity, and cell death. In vitro studies demonstrated that SAH-enhancing agents sensitized hepatocytes to TNFα killing, and the death was associated with increased caspase-8 activity, which was blocked by a caspase-8 inhibitor. In addition, increased intracellular SAH levels had no effect on nuclear factor κB activity induced by TNFα. In conclusion, these results provide a new link between abnormal methionine metabolism and abnormal TNFα metabolism in alcoholic liver disease. Increased SAH is a potent and clinically relevant sensitizer to TNFα hepatotoxicity. These data further support improving the S-adenosylmethionine-to-SAH ratio and removal of intracellular SAH as potential therapeutic options in alcoholic liver disease. Supplementary material for this article can be found on the HEPATOLOGYwebsite (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). (HEPATOLOGY 2004;40:989–997.)

Alcoholic liver disease (ALD) continues to be an important health problem in the United States. Although much progress has been made over the past decade, we still do not have a complete understanding of its pathogenesis, and this impedes devising specific therapies. Compelling evidence generated over the past decade demonstrates that tumor necrosis factor-α (TNFα) is a critical effector molecule in ALD. In 1989, we first reported dysregulated TNFα metabolism in alcoholic hepatitis patients with the observation that cultured monocytes (which produce the overwhelming majority of systemic circulating TNFα and are a surrogate marker for Kupffer cells) from alcoholic hepatitis patients spontaneously produced TNFα and produced significantly more TNFα in response to a lipopolysaccharide stimulus than did control monocytes.1 Increased serum TNFα concentrations in alcoholic hepatitis patients were next reported by several groups, and the values correlated with disease severity and mortality.2–4 Concomitant with these human studies, complementary studies in rats, mice, and tissue culture evaluated the role of TNFα in experimental models of liver disease.5, 6 Initially, it was shown that rats chronically fed alcohol were more sensitive to hepatotoxic effects of injected lipopolysaccharide.7 Subsequently, investigations showed that rats chronically fed alcohol had much higher lipopolysaccharide-stimulated plasma levels of TNFα than control rats and that liver injury could be attenuated by agents that inhibited TNFα production.8 Perhaps the most convincing data relating TNFα to alcohol induced liver injury are the observation that anti-TNFα antibody prevented liver injury in alcohol fed rats and that in mice lacking the TNFα type I receptor, alcohol induced liver injury did not develop.6, 9

Although the fact that TNFα plays an etiological role in ALD is now widely accepted, the exact molecular mechanism(s) involved in its hepatotoxicity are not well defined. Normally, TNFα can induce both proliferative and cytotoxic responses in hepatocytes.10 Under normal conditions, hepatocytes are resistant to TNFα-induced killing. However, primary hepatocytes from rats chronically fed alcohol have increased TNFα cytotoxicity11; thus, there must be a process by which hepatocytes can be “sensitized” to TNFα hepatotoxicity.

Abnormal methionine metabolism is one of the major characteristics of ALD. Although previous studies have reported that chronic alcohol consumption caused hepatic S-adenosylmethionine (SAMe) deficiency and elevated homocysteine levels,12, 13 its effects on S-adenosylhomocysteine (SAH) are less well defined. SAH is a product of SAMe in the hepatic transsulfuration pathway in which methyl groups from SAMe are transferred to a vast number of molecules, including DNA, RNA, biogenic amines, phospholipids, histones, and other proteins, via specific methyltransferases. SAH is a potent competitive inhibitor of most methyltransferases. Either a decrease in SAMe levels or an increase in SAH levels, or both, results in a decrease in the SAMe-to-SAH ratio, which can inhibit transmethylation reactions.14, 15

In this study, we examined the in vivo effects of chronic alcohol consumption on hepatic SAH levels and liver injury. Based on our in vivo study, we investigated the effects of increased SAH levels on TNFα-induced hepatotoxicity and the molecular mechanism(s) involved.

Abbreviations

ALD, alcoholic liver disease; TNFα, tumor necrosis factor-α; SAMe: S-adenosylmethionine; SAH, S-adenosylhomocysteine; HPLC, high-performance liquid chromatography; GSH, glutathione; HA, homocysteine plus adenosine; DZA, 3-deaza-adenosine; NF-κB, nuclear factor κB.

Materials and Methods

Cells and Culture Conditions.

HepG2 cells, a human hepatoma cell line, and WRL68 cells, nontransformed human fetal hepatocytes, were obtained from the American Type Culture Collection (Manassas, VA) and were cultured in Dulbecco modified eagle medium containing 10% (vol/vol) fetal bovine serum, 2 mmol/L glutamine, 5 U/mL penicillin, and 50 μg/mL streptomycin at 37°C in a humidified O2/CO2 (19:1) atmosphere.

Animal Model and Experimental Protocol.

Male C57BL/6 mice weighing 20 ± 0.5 g (mean ± SEM) were obtained from the Jackson Laboratory (Bar Harbor, ME). The mice were housed in the animal quarters at the University of Louisville Research Resources Center, and the studies were approved by the Institutional Animal Care and Use Committee, which is certified by the American Association of Accreditation of Laboratory Animal Care. In the first week, 16 mice were pair-fed liquid diets containing 18% of energy as protein, 35% as fat, 17% as carbohydrate, and 30% as either ethanol (ethanol diet) or as an isocaloric maltose–dextrin mixture (control diet), according to Lieber et al.16 The energy from ethanol was increased by 2% in each subsequent week. Mice were maintained on the treatments for 4 weeks before being killed.

Histological Analysis and In Situ Apoptosis Detection.

At the time of killing, small pieces of liver tissue were harvested and fixed immediately in 10% buffered formalin. After paraffin embedding, 5-μm sections were deparaffinized in xylene and were rehydrated through a series of decreasing concentrations of ethanol. Sections were stained with hematoxylin and eosin. Apoptotic hepatocytes were detected by the terminal transferase dUTP end-labeling assay with a terminal transferase dUTP end-labeling assay kit (Intergen Company, Purchase, NY) as described previously, and numbers were counted in five high-power fields. Images of liver sections were captured with a Nikon Eclipse E600 microscope (Fryer Company, Inc., Cincinnati, OH) equipped with a digital camera (SPOT; Diagnostic Instruments, Sterling Height, MI).

SAMe and SAH Assays by High-Performance Liquid Chromatography.

Deproteinized cellular extracts or tissue homogenates (4% metaphosphoric acid) were prepared and SAMe and SAH were determined by a high-performance liquid chromatography (HPLC) method, using a 5-μm Hypersil C-18 column (250 × 4.6 mm). The mobile phase consisted of 40 mmol/L ammonium phosphate, 8 mmol/L heptane sulfonic acid (ion-pairing reagent, pH 5.0), and 6% acetonitrile and was delivered at a flow rate of 1.0 mL/min. SAMe and SAH were detected using a Waters 740 UV detector (Milford, MA) at 254 nm. An internal standard, S-adenosylethionine, was added to all samples and standard solutions to a concentration of 100 nmol/mL. Protein concentrations were measured by a protein assay kit from Bio-Rad Laboratories (Hercules, CA) in accordance with the manufacturer's instructions.

Glutathione Assay by HPLC.

Both reduced glutathione (GSH) and oxidized glutathione (GSSG) in cell extracts and tissue homogenates were quantified by HPLC with electrochemical detection according to the method of Richie and Lang,17 with slight modifications. In brief, 20-μL samples were injected onto a reversed-phase C18 column (Val-U-Pak HP, fully endcapped octadecyl silane, 5 μm, 250 × 4.6 mm; ChromTech Inc., Apple Valley, IN). The mobile phase, which consisted of a solution of 0.1 M monochloroacetic acid and 2 mmol/L heptane sulfonic acid at pH 2.8 (98%) and acetonitrile (2%), was delivered at a flow rate of 1 mL/min. The compounds were detected in the eluant with a Bioanalytical Systems (W. Lafayette, IN) dual LC4B amperometric detector using two Au/Hg electrodes in series with potentials of −1.2 V and +0.15 V for the upstream and downstream electrodes, respectively. Standard curves for the analytes were plotted as peak area versus concentration of the analyte.

TNFα Enzyme-Linked Immunosorbent Assay Analysis.

TNFα in homogenized tissues or conditioned medium was quantified using enzyme-linked immunosorbent assay kits in accordance with the manufacturer's instructions. The detection limit for TNFα is 13.5 pg/mL.

Lactate Dehydrogenase Assay.

Cell death was measured after 24 hours of cell treatment by the release of lactate dehydrogenase into the culture medium. Lactate dehydrogenase activity was determined spectrophotometrically at 340 nm by following the rate of conversion of oxidized nicotinamide adenine dinucleotide to the reduced form.

MTT Assay.

Cells were seeded in 96-well culture plates at a cell density of 20,000 cells/well, cultured overnight, and treated as required. After treatment, 25 μL of fresh MTT (5 mg/mL) was added. The cells were incubated at 37°C for 2 hours to allow incorporation and conversion of MTT to the formazan derivative. The formazan derivative was solubilized by the addition of 100 μL of lysis buffer (20% SDS in 50% dimethylformamide, pH 4.7). After incubation for 24 hours at 37°C, the absorbance values were measured at 570 nm using a Dynatech MR700 plate reader (Dynatech Laboratories, Inc., Chantilly, VA).

Enzyme-Linked Immunosorbent Assay for DNA Fragmentation.

For quantification of DNA fragmentation, we used a cellular DNA fragmentation enzyme-linked immunosorbent assay kit manufactured by Roche (Indianapolis, IN) in accordance with the manufacturer's instructions. Briefly, HepG2 cells were plated in 6-well plates and were pretreated with 1 mmol/L homocysteine plus adenosine (HA), 1 mmol/L SAH, or 100 μmol/L 3-deaza-adenosine (DZA) for 2 hours before TNFα (500 U/mL) was added to the media. At 6 hours after TNFα addition, cell membrane was lysed for 30 minutes at 20°C in a solution containing 10 mmol/L ethylenediaminetetraacetic acid and 0.1% Tween-20. After centrifugation for 10 minutes at 200g to remove cellular debris and nuclear components, 20 μL cytoplasmic lysate together with 80 μL immunoreagent containing 5% anti–histone-biotin and 5% anti-DNA-peroxidase were added to the 96-well plate precoated with streptavidin. After a 2-hour incubation, wells were rinsed and 100 μL of substrate solutions were added to each well. DNA fragmentation was measured at 405 nm against substrate solution as a blank.

Caspase Activities Assay.

Caspase-8 enzyme activities were measured using the assay kit from Calbiochem (San Diego, CA) according to manufacturer's instruction. In brief, HepG2 cells were lysed by adding 50 μL of chilled lysis buffer and incubated on ice for 10 minutes. The cell lysates were centrifuged at 15,000g for 10 minutes, and the supernatants were collected. Assay mixtures were prepared by mixing 40 μL supernatants with 20 μL 5X assay buffer and 10 μL caspase-8 substrate (IETD-pNA: Ile-Glu-Thr-Asp-para-nitroaniline). After incubation at 37°C for 2 hours in the dark, the absorbance of each sample was measured at 405 nm with a microtiter plate reader.

Cell Transfection and κB-Luciferase Activity Assay.

The transactivating activity of nuclear factor NF-κB was assayed by transfecting HepG2 cells with a luciferase reporter construct controlled by κB binding sites using a FuGENE reagent. After 48 hours, the medium was replaced and transfectants were further incubated with appropriate treatments. For cell lysis and detection of luciferase activity, a commercial luciferase assay kit (Promega, Madison, WI) was used, and luciferase activity was quantified in a Berthold luminometer.

Statistical Analysis.

All data were expressed as mean ± SD. Statistical analysis was performed using a one-way ANOVA and was analyzed further by Newman-Keul's test for statistical difference. Differences between treatments were considered to be statistically significant at P < 0.05.

Results

In Vivo Study.

Male C57BL/6 mice fed the Lieber-DeCarli liquid diet containing 4% alcohol for 4 weeks were used to document liver injury, hepatic TNFα production, products of hepatic methionine metabolism, and caspase-8 production and activity. Our results demonstrated that mice chronically fed alcohol showed modestly elevated serum alanine aminotransferase activity in comparison with pair-fed animals (Fig. 1A) and that alcohol feeding induced microvesicular steatosis and necrosis by hematoxylin and eosin staining (Fig. 1B). A terminal transferase dUTP end-labeling assay showed obvious hepatic apoptosis in mice chronically fed alcohol (Fig. 1C). To determine the effect of chronic alcohol consumption on the hepatic transsulfuration pathway, hepatic GSH, GSSG, SAMe, and SAH levels were measured by our well-established HPLC system. Alcohol feeding for 4 weeks did not significantly affect hepatic GSH and oxidized GSH levels (Fig. 2A), whereas SAMe levels were decreased significantly (Fig. 2B). More importantly, our results showed that chronic alcohol consumption significantly increased hepatic SAH levels (Fig. 2B), and thereby markedly decreased the ratio between SAMe and SAH (Fig. 2C), indicating disruption of hepatic transmethylation pathway in this animal model of ALD. Hepatic TNFα levels were increased significantly after ethanol feeding (see Supplemental Figure 1). Caspase-8 is one of the initiator caspases in apoptosis signaling induced by many receptor-mediated death signals, including TNFα. To determine the effects of alcohol consumption on hepatic caspase-8 activity and protein production, both a caspase-8 activity assay and a Western blot for its protein production were performed. Our results showed that alcohol consumption increased caspase-8 activity in liver (Fig. 3A). Moreover, a Western blot assay demonstrated that alcohol consumption increased protein level of the cleavage form of caspase-8 (Fig. 3B).

Figure 1.

Liver injury after chronic alcohol consumption. Male C57BL/6 mice were pair-fed (PF) liquid diets with or without ethanol for 4 weeks. (A) Plasma alanine aminotransferase (ALT) activity changes. Data represent mean ± SD (n = 6). *P < 0.01 compared with PF animals. (B) Formalin-fixed and paraffin-embedded liver sections were stained with hematoxylin and eosin. Livers from PF animals had normal architecture. In contrast, livers from alcohol-fed (AF) animals had obvious accumulation of microvesicular steatosis (arrows) and necrotic hepatocytes (arrowheads; original magnification, ×130). (C) Formalin-fixed and paraffin-embedded liver sections were used in terminal transferase dUTP end-labeling (TUNEL) assays to detect fragmented nuclear DNA. Liver from PF animals had very few detectable TUNEL positive hepatocytes. In contrast, livers from AF animals had more scattered hepatocytes with TUNEL-positive labeling of nuclei (arrowheads; original magnification, ×130). CV, central vein.

Figure 2.

Effect of chronic alcohol consumption on hepatic methionine metabolism. Male C57BL/6 mice were pair-fed liquid diets with or without ethanol for 4 weeks. Liver samples were homogenized with 4% metaphosphoric acid and hepatic glutathione, S-adenosylmethionine (SAMe), and S-adenosylhomocysteine (SAH) contents were assayed by high-performance liquid chromatography. (A) Hepatic glutathione (GSH) and oxidized GSH (GSSG) contents were not affected by chronic alcohol consumption. (B) Chronic alcohol consumption resulted in decreased hepatic SAMe levels and increased SAH levels. (C) Chronic alcohol intake significantly decreased hepatic SAMe-to-SAH ratio. Data represent mean ± SD (n = 6). *P < 0.01 compared with pair-fed animals. AF, alcohol-fed animals.

Figure 3.

Chronic alcohol consumption increased caspase-8 activity. (A) Hepatic caspase-8 activity was assayed by commercially available enzyme-linked immunosorbent assay kit. Chronic alcohol feeding for 4 weeks elevated hepatic caspase-8 activity. *P < 0.05 vs. pair-fed (PF) animals. The data are expressed as mean ± SD, n = 6. (B) Western blot assay for active form of caspase-8 protein production. Chronic alcohol feeding for 4 weeks enhanced protein production of active form of caspase-8. OD, optical density; AF, alcohol-fed animals.

Disruption of the Transmethylation Pathway Sensitizes Hepatocytes to TNFα-Induced Cell Death.

Under normal conditions, most of the SAMe generated is used in transmethylation reactions in which methyl groups are added to other compounds. SAMe is converted to SAH, which is a potent competitive inhibitor of most methyltransferases studied. Both an increase in the SAH level as well as a decrease in the SAMe-to-SAH ratio are known to inhibit transmethylation reactions. In our in vitro experiments, a well-defined cell culture system was used to evaluate the effects of disruption of hepatic transmethylation pathway on “sensitization” of hepatocytes to TNFα killing. The basic methodology used to disrupt hepatic transmethylation reactions was to increase intracellular SAH levels, thereby inhibiting intracellular methyltransferases. Three complementary treatments were used in the hepatocyte cell viability studies: (1) HA at 1 mmol/L; (2) SAH at 1 mmol/L; and (3) DZA at 80 αmol/L. Homocysteine and adenosine are the substrates for the hepatic synthesis of SAH catalyzed by SAH hydrolase; this treatment results in increased intracellular SAH synthesis. DZA is a widely used specific inhibitor of SAH hydrolase, and this treatment causes the accumulation of intracellular SAH. To confirm that the effects we observed were not limited to transformed cells (HepG2 cells), WRL68 cells and nontransformed human fetal hepatocytes also were used to validate our cell viability study. Cells were pretreated with these agents for 2 hours, then TNFα (500 U/mL) was added to the medium and cell viability was measured after 24 hours by both lactate dehydrogenase and MTT assay. Results are shown in Fig. 4. Without TNFα exposure, HA, SAH, and DZA did not cause significant cell death by themselves. Moreover, homocysteine or adenosine alone did not sensitize cells to TNFα killing. However, pretreatments with HA, SAH, or DZA significantly sensitized both HepG2 cells (Fig. 4A) and WRL68 cells (Fig. 4B) to TNFα-induced cell death. Also, pretreatment of HepG2 cells with HA caused a dose-dependent increase in TNFα-induced cell death (Fig. 4C). HepG2 cells treated with TNFα for 6 hours after 2-hour pretreatment with either HA, SAH, or DZA demonstrated significantly increased DNA fragmentation (see Supplemental Figure 2). Pretreatments with these three agents significantly increased TNFα-induced DNA fragmentation, which was verified by electrophoresis assay of DNA ladder (data not shown).

Figure 4.

Intracellular S-adenosylhomocysteine (SAH)-enhancing agents sensitized hepatocytes to tumor necrosis factor-α (TNFα)-induced cytotoxicity. (A) HepG2 cells and (B) WRL68 cells were cultured in media alone (UT, untreated), TNFα alone, SAH-enhancing agents alone (ADO, 1 mmol/L adenosine; HOMO, 1 mmol/L homocysteine), or the combination of TNFα and one of agents (HA, 1 mmol/L homocysteine plus adenosine; DZA, 80 μM 3-deaza-adenosine; SAH, 1 mmol/L S-adenosylhomocysteine). All SAH-enhancing agents were used for 2 hours before TNFα addition at a dose of 500 U/mL. Cell death was measured by lactate dehydrogenase release and cell viabilities were expressed as percentage of UT (mean ± SD; n = 3). These modest amount of TNFα by itself caused no hepatocyte injury. Moreover, the SAH-enhancing agents alone caused minimal toxicity. However, in combination with TNFα, they caused severe cell death. (C) Dose response of cell viability to TNFα-induced cytotoxicity in HepG2 cells pretreated with HA. HepG2 cells were pretreated with different doses of HA for 2 hours before stimulation with TNFα at 500 U/mL. Cell death was measured by lactate dehydrogenase release and cell viabilities were expressed as percentage of UT and data represent a mean of at least three experiments ± SD.

Effects of HA Treatment on Productivity of the Transsulfuration Pathway.

Time course effects of HA on intracellular GSH, SAMe, and SAH were investigated using HepG2 cells. HepG2 cells at a density of 0.5 × 106 were treated with 1 mmol/L HA, and cell pellets were collected at different time points for intracellular GSH, SAMe, and SAH analysis by HPLC. HA treatment increased intracellular GSH levels over time (see Supplemental Figure 3). Although HA treatment also elevated intracellular SAMe levels that peaked at 8 hours after HA treatment and returned to normal level at 24 hours (Fig. 5A), intracellular SAH levels were more significantly elevated by HA treatment (Fig. 5B), which resulted in an extremely low SAMe-to-SAH ratio (Fig. 5C).

Figure 5.

Time course changes of intracellular (A) S-adenosylmethionine (SAMe); (B) S-adenosylhomocysteine (SAH); (C) SAMe-to-SAH ratio after homocysteine plus adenosine (HA) treatment. HepG2 cells were treated with 1 mmol/L HA and samples were collected at different time points. Intracellular SAMe and SAH levels were assayed by a high-performance liquid chromatography system. Although HA treatment increased both intracellular SAMe and SAH levels, it resulted in a significantly lowered SAM-to-SAH ratio. The data represent a mean of at least three experiments ± SD. *P < 0.05 vs. corresponding untreated value.

Treatment of HepG2 Cells with HA Had No Effect on NF-κB Activity.

NF-κB activation induces the transcription of antiapoptotic proteins that renders liver cells resistant to TNFα-induced apoptosis. To assess NF-κB transcriptional activity, transient transfection assays were performed. HepG2 cells were cotransfected with an NF-κB-luciferase reporter, and cells were pretreated with HA at 1 mmol/L for 2 hours followed by treatment with TNFα (500 U/mL) for 24 hours. In comparison with untreated HepG2 cells, TNFα treatment significantly increased NF-κB activity. HA pretreatment also significantly enhanced NF-κB activity in response to TNFα. There was no difference between cells treated with TNFα either with or without HA pretreatment (Fig. 6).

Figure 6.

Pretreatment with homocysteine plus adenosine (HA) had no effect on tumor necrosis factor-α (TNFα)-induced nuclear factor κB (NF-κB) activation. The transactivating activity of NF-κB was assayed by transfecting HepG2 cells with a luciferase reporter construct controlled by κB binding sites using a FuGENE reagent for 48 hours, and then cells were pretreated with 1 mmol/L HA for 2 hours before TNFα addition at a dose of 500 U/mL. Luciferase activity was quantified 24 hours later. The data are expressed as mean of at least three experiments ± SD. Bars with different letter differ.

Involvement of Caspase-8 in TNFα Induced Apoptosis in HepG2 Cells Sensitized by HA.

To determine whether caspase-8 is involved in this TNFα induced apoptosis, HepG2 cells were pretreated with 1 mmol/L HA for 90 minutes, and then z-IETD-FMK, a specific caspase-8 inhibitor, was added to the medium at a dose of 100 μM, and TNFα at 500 U/mL was added 30 minutes later. The caspase-8 inhibitor almost completely inhibited cell death (Fig. 7). Based on this finding, we investigated the effects of HA on caspase-8 activity and protein production during TNFα exposure by both Western blot and enzymatic activity assay. Pretreatment of HepG2 cells with 1 mmol/L HA for 2 hours increased caspase-8 activities after TNFα exposure as assessed by both activity assay and Western blot for the active form of caspase-8 (see Supplemental Figure 4).

Figure 7.

An inhibitor of caspase-8 attenuated tumor necrosis factor-α (TNFα)-induced apoptosis in HepG2 cells sensitized by homocysteine plus adenosine (HA). HepG2 cells were pretreated with HA for 2 hours, then z-IETD-FMK, a specific caspase-8 inhibitor, was added to the medium at a dose of 100 μM. TNFα at 500 U/mL was added 30 minutes later. The caspase-8 inhibitor significantly attenuated cell death. The data are expressed as mean of at least three experiments ± SD. UT, untreated; HA, homo+ ado; T, tumor necrosis factor.

Discussion

In ALD, there is both increased TNFα production by monocytes and Kupffer cells and “sensitization” to TNFα-mediated hepatocyte cytotoxicity. The objective of this research was to determine whether disruption of the hepatic transmethylation pathway represented one clinically relevant mechanism by which hepatocytes are sensitized to TNFα-induced hepatotoxicity. Some factors that are known to sensitize cells to TNFα cytotoxicity include agents that inhibit protein synthesis or induce transcriptional arrest (e.g., cycloheximide, galactosamine, actinomycin D),18, 19 agents that cause total GSH depletion or selective mitochondrial GSH depletion,20, 21 and proteosome inhibition.22–24 Using the Lieber-DeCarli alcohol feeding model, we demonstrated that chronic alcohol consumption not only caused mild liver injury with increased hepatic TNFα production and apoptosis, but also resulted in disruption of the hepatic transmethylation pathway. Four weeks of alcohol feeding resulted not only in hepatic SAMe deficiency, but also in SAH elevation, and thereby caused a significantly decreased hepatic SAMe-to-SAH ratio. We postulate that the observed increased SAH levels and the decreased SAMe-to-SAH ratio sensitizes to TNFα hepatotoxicity.

Abnormal hepatic methionine metabolism is an acquired metabolic abnormality in ALD, and the effects of chronic alcohol intake on hepatic methionine metabolism are initially seemingly paradoxical (see Supplemental Figure 5). Whereas alcohol consumption causes hepatic SAMe deficiency,12 it elevates hepatic homocysteine levels, a product of SAMe metabolism.13 Accumulated evidence from multiple studies demonstrates that both decreased SAMe levels and elevated homocysteine levels may contribute to alcohol-induced liver injury. A pivotal clinical trial from Mato et al.25 reported that SAMe supplementation improved mortality in ALD. Moreover, a recent study from Ji and Kaplowitz26 reported that hyperhomocysteinemia induced endoplasmic reticulum stress with liver injury in alcohol-fed mice, and removal of homocysteine by betaine supplementation alleviated liver injury. Although the contribution of SAMe deficiency and hyperhomocysteinemia to alcohol-induced liver injury has been reported widely, the effect of alcohol on SAH levels, another metabolite in the methionine metabolism pathway and a potent inhibitor of methyltransferases, has received limited investigative attention. Our current study clearly demonstrates that alcohol consumption not only significantly decreases SAMe levels, which is consistent with previous studies in both animal models and humans,27 but more importantly, chronic alcohol consumption significantly elevates hepatic SAH levels, suggesting that alcohol consumption causes disruption of hepatic transmethylation reactions. This result is consistent with the previous study by Lieber et al.28 that showed that chronic alcohol consumption diminished hepatic phosphatidylethanolamine methyltransferase activity in a baboon model of alcohol-induced fibrosis. Recent research by Halsted et al.29 using the micropig ALD model showed that alcohol feeding increased hepatic SAH levels, which could be explained by the combination of increased glycine N-methyl transferase levels and reduced activity of SAH hydrolase, the enzyme catalyzing SAH metabolism to homocysteine and adenosine. Moreover, a study by Barak et al.30 demonstrated recently that primary hepatocytes isolated from rats chronically fed alcohol had elevated SAH levels that could be corrected by treating these cells with betaine. Thus, emerging evidence documents elevated SAH levels in animal models of ALD, and in vitro studies in L929 cells, suggest that elevated SAH enhances TNFα-induced cytolysis.31

TNFα is a pleiotropic cytokine that can trigger distinct signal pathways in liver cells through TNFα receptor 1 via adapter molecules, including the intracellular cascades leading to apoptosis, as well as NF-κB and Jun kinase activation. The initial step in TNFα signaling involves the binding of the TNFα trimer to the extracellular domain of TNFα receptor 1, and the resulting aggregated TNFα receptor 1 is recognized by the adaptor protein TNFα receptor-associated death domain, which recruits additional adaptor proteins such as receptor-interacting protein, TNFα-receptor-associated factor 2, and Fas-associated death domain. These latter proteins recruit key enzymes to TNFα receptor 1 that are responsible for initiating signaling events. Fas-associated death domain contains a death effector domain, which then recruits pro–caspase-8. Activation of pro–caspase-8 through self-cleavage leads to a series of downstream events, including the caspase cascade resulting in activation of caspase-3, or the mitochondrial pathway, leading to apoptosis. In addition to activating the death pathway, TNFα also activates NF-κB and c-Jun N-terminal kinase. Although the role of c-Jun N-terminal kinase activation during TNFα-induced apoptosis is less well defined, NF-κB activation induces the transcription of antiapoptotic proteins, which renders liver cells resistant to TNFα-induced cell death.32 Thus, an interesting feature of the TNFα signaling network is the existence of extensive cross talk and balance between death and survival signaling pathways that emanate from TNFα receptor 1. Any process that can interfere with this balance will affect sensitivity of cells to TNFα-regulated cytotoxicity. Enhanced apoptosis signals will result in cellular susceptibility to TNFα-induced apoptosis, whereas activation of NF-κB protects against apoptosis.

Our results showed that caspase-8 inhibition significantly attenuated TNFα-induced cell death. Both intracellular methylation status and oxidant stress can modify gene expression and production of critical proteins. The fact that both GSH and oxidized GSH were not changed by alcohol feeding for 4 weeks in our study suggests that oxidant stress was not the major factor involved in this hepatotoxicity. Moreover, antioxidants did not block SAH-enhanced TNFα hepatotoxicity in vitro.31 Methylation involves conversion of cytosine to 5-methylcytosine in reactions mediated by specific DNA methyltransferases. The extent of methylation of cytosine to 5-methylcytosine often is correlated with the gene activity. Hypermethylation is associated with nonexpression of the gene, whereas its hypomethylation is a necessary, but not sufficient, condition for gene expression.33 Considering the important position of caspase-8 in the TNFα signaling pathway and previous evidence that gene expression of caspase-8 is sensitive to intracellular methylation status,34, 35 we investigated the effects of disrupted transmethylation reactions resulting from increased intracellular SAH levels on caspase-8 production and its activity when exposed to TNFα. Our results suggest that increased intracellular SAH levels resulted in increased caspase-8 production and activity after hepatocyte stimulation by TNFα. To relate our in vitro data with the in vivo situation, we also measured caspase-8 production and activity in mice fed alcohol for 4 weeks and observed similar results. Although we do not have direct evidence showing that increased SAH levels play a causative role in TNFα-induced hepatotoxicity in vivo, our combined in vitro and animal data suggest that disrupted hepatic transmethylation reactions resulting from increased intracellular SAH from chronic alcohol consumption together with increased hepatic TNFα levels could be one of the mechanisms contributing to ALD. Consistent with this hypothesis, in mice in which the adenosine kinase gene has been “knocked out,” fatal neonatal hepatic steatosis develops, with elevated hepatic SAH levels and disrupted transmethylation reactions postulated to play an etiological role in the development of the postnatally lethal fatty liver.36 Similarly, carbon tetrachloride-induced hepatic injury in rats is associated with decreased SAMe and increased SAH levels, as well as global DNA hypomethylation.37 SAMe supplementation improved the altered SAMe-to-SAH ratio and attenuated liver injury.

In conclusion, our research clearly demonstrates that chronic alcohol consumption not only decreased hepatic SAMe levels, but also increased hepatic SAH levels and significantly decreased the SAMe-to-SAH ratio in association with increased hepatic TNFα and caspase-8 activities and hepatocyte apoptosis. In vitro studies demonstrated that multiple techniques that increased hepatic SAH sensitized to TNFα killing and that the hepatocytes death was associated with increased caspase-8 activity and was blocked by a caspase-8 inhibitor. These data provide important new information concerning potential interactions between abnormal methionine metabolism and TNFα induced hepatotoxicity. Our data suggest that removal of intracellular SAH by either increasing SAH hydrolase activity or through other biochemical reactions deserves further investigation as a potential therapy for ALD.

Acknowledgements

The authors thank Marion McClain, MS, and Steve Mahanes for assistance in the preparation of this manuscript and Marcia Liu for assistance in the HPLC assay.

Ancillary