Roles for C16-ceramide and Sphingosine 1-Phosphate in Regulating Hepatocyte Apoptosis in Response to Tumor Necrosis Factor- 

Roles for C16-ceramide and Sphingosine 1-Phosphate in RegulatingHepatocyte Apoptosis in Response to Tumor Necrosis Factor-�*□S

Received for publication, March 18, 2005, and in revised form, May 23, 2005Published, JBC Papers in Press, June 9, 2005, DOI 10.1074/jbc.M503002200

Yosuke Osawa‡, Hiroshi Uchinami‡, Jacek Bielawski§, Robert F. Schwabe‡, Yusuf A. Hannun§,and David A. Brenner‡¶

From the ‡Department of Medicine, Columbia University, College of Physicians and Surgeons, New York, New York 10032and the §Department of Biochemistry and Molecular Biology, Medical University of South Carolina,Charleston, South Carolina 29425

Tumor necrosis factor (TNF)-� signals cell death andsimultaneously induces the generation of ceramide,which is metabolized to sphingosine and sphingosine1-phosphate (S1P) by ceramidase (CDase) and sphingo-sine kinase. Because the dynamic balance between theintracellular levels of ceramide and S1P (the “ceramide/S1P rheostat”) may determine cell survival, we investi-gated these sphingolipid signaling pathways in TNF-�-induced apoptosis of primary hepatocytes. EndogenousC16-ceramide was elevated during TNF-�-induced apo-ptosis in both rat and mouse primary hepatocytes. Theputative acid sphingomyelinase (ASMase) inhibitor imi-pramine inhibited TNF-�-induced apoptosis and C16-ce-ramide increase as did the knock out of ASMase. Over-expression of neutral CDase (NCDase) inhibited theTNF-�-induced increase of C16-ceramide and apoptosisin rat primary hepatocytes. Moreover, NCDase inhibitedliver injury and hepatocyte apoptosis in mice treatedwith D-galactosamine plus TNF-�. This protective effectwas abrogated by the sphingosine kinase inhibitor N,N-demethylsphingosine, suggesting that the survival ef-fect of NCDase is due to not only C16-ceramide reductionbut also S1P formation. Administration of S1P or over-expression of NCDase activated the pro-survival kinaseAKT, and overexpression of dominant negative AKTblocked the survival effect of NCDase. In conclusion,activation of ASMase and generation of C16-ceramidecontributed to TNF-�-induced hepatocyte apoptosis.NCDase prevented apoptosis both by reducing C16-cer-amide and by activation of AKT through S1P formation.Therefore, the cross-talk between sphingolipids andAKT pathway may determine hepatocyte apoptosisby TNF-�.

Tumor necrosis factor (TNF)1-� is a multifunctional cytokinethat plays a role in inflammation, immunity, antiviral re-

sponses, and a variety of diseases. TNF-� is particularly im-portant in the pathophysiology of hepatocytes, mediating viralhepatitis, alcoholic liver disease, and fulminant hepatitis (1).TNF-� activates a variety of cellular signal transduction path-ways, some of which result in apoptosis (2, 3). Ceramides aresphingolipid-derived signaling and regulatory intermediatesthat play a role in the stress response and cell death (4). Stressstimuli such as TNF-�, Fas ligand, oxidative stress, growthfactor withdrawal, anticancer drugs, ionizing radiation, heatshock, or ultraviolet light induce an elevation in the endoge-nous cellular levels of ceramides (4, 5). Ceramides are gener-ated from the major membrane sphingolipid sphingomyelin byacid or neutral sphingomyelinases (ASMase), enzymes that areactivated in response to TNF-� and other cytokines (6). AS-Mase knock-out mice are resistant against TNF-�-induced liverinjury (7). Thus, ceramides generated by ASMase have animportant role in TNF-�-induced hepatocyte cell death.

Besides apoptotic signals, TNF-� also activates moleculesthat protect cells from apoptosis. For example, TNF-� trans-mits anti-apoptotic signals via nuclear factor (NF)-�B andphosphatidylinositol 3-kinase (PI3K)/AKT. Blocking these sig-naling pathways results in sensitization of hepatocytes to apo-ptosis induced by TNF-� (8, 9). In addition TNF-� activatessphingosine kinase (SphK), which converts sphingosine tosphingosine 1-phosphate (S1P). This lipid-derived mediatorprevents the cytotoxic action of TNF-� (8, 10). Moreover, amodel has been proposed in which the dynamic balance be-tween the intracellular levels of ceramide and S1P (the “cer-amide/S1P rheostat”) is an important factor that determineswhether a cell survives or dies (11).

Ceramide is deacylated by either acid or neutral ceramidase(ACDase or NCDase) and converted to sphingosine. Overex-pression of ACDase, which is localized in the lysosomes (12),protects from TNF-�-induced apoptosis in murine fibrosarcomaL929 cells by reduction of intracellular ceramide and thereforein a shift of the ceramide/S1P rheostat in the direction of cellsurvival (13). Besides an acidic form, a neutral ceramidase hasbeen identified and cloned (14–16). NCDase is a membrane-bound enzyme that catalyzes the hydrolysis of N-acyl linkagein ceramide (10). In rat hepatocytes, NCDase is localized in lateendosomes/lysosomes because of the presence of a functionaldi-Leu motif in the putative amino acid sequences in NCDases,which is a sorting signal for vesicular transport from plasma

* This work was supported by National Institutes of Health GrantPPG DK59340 and by the Mochida Memorial Foundation for MedicalPharmaceutical Research. The costs of publication of this article weredefrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

□S The on-line version of this article (available at http://www.jbc.org)contains Figs. S1 and S2.

¶ To whom correspondence should be addressed: Dept. of Medicine,Columbia University, College of Physicians and Surgeons, 622 West168th St., PH 8E-105J, New York, NY 10032. Tel.: 212-305-5838; Fax:212-305-9822; E-mail: [email protected].

1 The abbreviations used are: TNF, tumor necrosis factor; ASMase,acid sphingomyelinase; NF-�B, nuclear factor-�B; PI3K, phosphatidyl-inositol 3-kinase; SphK, sphingosine kinase; DMS, N,N,-dimethyl-sphingosine; S1P, sphingosine 1-phosphate; ACDase, acid ceramidase;

NCDase, neutral ceramidase; PTX, pertussis toxin; GFP, green fluores-cent protein, GalN, galactosamine; CA-AKT, constitutively activateform of AKT; DN-AKT, dominant negative form of AKT; TUNEL, ter-minal deoxynucleotidyltransferase nick-end labeling; ALT, alanineaminotransferase; MS, mass spectrometry; HA, hemagglutinin; m.o.i.,multiplicity of infection.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 30, Issue of July 29, pp. 27879–27887, 2005© 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org 27879

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

membrane to endosomes/lysosomes, (10). More importantly,the roles of NCDases in TNF-�-induced hepatocyte apoptosishave not been addressed.

In the present study, we evaluated the role of endogenousceramide in apoptosis of rat primary hepatocytes induced byTNF-�. Preventing formation of endogenous ceramide or accel-erating its clearance by overexpression of NCDase protected fromTNF-� induced hepatocyte apoptosis. We provide evidence for animportant role of AKT activation in this protective response.

EXPERIMENTAL PROCEDURES

Materials—Cell culture media for rat and mouse primary hepato-cytes, Waymouth’s MB 752/1 and RPMI 1640, and recombinant humaninsulin were purchased from Invitrogen. Recombinant mouse TNF-�was obtained from R & D Systems (Minneapolis MN). N, N,-Dimethyl-sphingosine (DMS), S1P, pertussis toxin (PTX), imipramine hydrochlo-ride, LY 294002, ASMase from human placenta, Hoechst 33258 (bis-benzimide), propidium iodide, and anti-�-actin (clone AC-15) antibodywere obtained from Sigma. Antibodies against cleaved caspase-3 (9661)and phosphorylated AKT (Ser473) (9271) were obtained from Cell Sig-naling Technology (Beverly, MA). Anti-mouse and -rabbit IgG horse-radish peroxidase-coupled secondary antibodies were obtained fromSanta Cruz Biotechnology (Santa Cruz, CA). L-[3-3H]Serine was fromAmersham Biosciences. D-Galactosamine hydrochloride (GalN) waspurchased from Nacalai Tesque (Tokyo, Japan). All other reagents used

were of the highest analytical grade available.Primary Hepatocyte Cultures and Treatments—Sprague-Dawley

male rats (200–250 g) or ASMase knock-out mice (ASMase�/�), main-tained in C57BL/6 background, were anesthetized with ketamine andxylazine administered by intraperitoneal injection. Hepatocytes werethen isolated by a nonrecirculating in situ collagenase perfusion oflivers cannulating through the portal vein by a procedure modified fromMoldeus et al. (17). Livers were first perfused in situ with 0.5 mM EGTAcontaining calcium-free salt solution, followed by perfusion with solu-tion containing 0.02% collagenase D (Roche Applied Science). The liverwas then gently minced on a Petri dish and filtered with polyamidemesh (3-60/42, Sefar America, Kansas City, MO). Hepatocytes werewashed three times and centrifuged at 50 � g for 1 min. Cell viabilitywas consistently 85% as determined by trypan blue exclusion. Cells (1 �105 and 1 � 106) were plated on 12-well and 60-mm dishes coated withrat collagen type I in Waymouth’s medium containing 10% fetal bovineserum containing antibiotics (plating medium). After 4 h, the culturewas washed with PBS and changed to serum-free RPMI 1640 contain-ing antibiotics in the presence or absence of recombinant adenoviruses(Ad5GFP, Ad5I�B, Ad5NCDase, Ad5CA-AKT, Ad5DN-AKT, andAd5NF�B-Luc). After 2 h of incubation, the culture media were changedto serum-free RPMI 1640 containing antibiotics, and the cells wereincubated for another 16 h. Before stimulation with 30 ng/ml mouseTNF-� or 1 �M S1P, the cells were washed with PBS and, if necessary,incubated for 1 or 2 h in serum-free RPMI 1640 containing the indicatedagent(s): 50 �M imipramine in H2O, 10 �M DMS in Me2SO, ASMase (1

FIG. 1. The I�B super repressor sensitizes rat hepatocytes to TNF-�-mediated cell death. A, NF-�B activation of primary rathepatocytes infected with or without Ad5I�B by TNF-� (30 ng/ml, for 8 h) was assessed by reporter gene assays using 3� �B luciferase [(�B)3-Luc]expressing adenovirus (10 m.o.i.). B, primary rat hepatocytes infected with Ad5GFP or Ad5I�B (10 m.o.i.) were treated with TNF-� for variousperiods of time. Cell death was determined by double staining with Hoechst 33258 (left panel) and propidium iodide (right panel) to detect apoptoticand necrotic cells, respectively. At least 500 cells were counted, and cell death was expressed as a percentage of total cells. Results are expressedas means � S.D. from at least three independent experiments. *, p � 0.01 using Student’s t test. C, primary rat hepatocytes infected with Ad5GFPor Ad5I�B were treated with TNF-� (30 ng/ml) for various periods of time. Extracted proteins were subjected to SDS-PAGE, and immunoblottingwas performed with anti-cleaved caspase (casp)-3 and �-actin antibodies. The results shown are representative of at least two independentexperiments. D, primary rat hepatocytes infected with Ad5GFP or Ad5I�B were incubated with TNF-� for the indicated times. The levels ofindividual ceramide molecular species were measured by using MS analysis. The results are expressed as picmoles of lipid/�g of Pi, and theyrepresent the mean from two independent experiments.

NCDase Protects Hepatocyte Apoptosis Dependence on AKT27880

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

IU/ml). In some experiments, hepatocytes were pretreated with PTX for12 h before adenovirus infection or S1P treatment.

Adenoviruses—The adenovirus five variants Ad5I�B, Ad5GFP,Ad5CA-AKT, Ad5DN-AKT, and Ad5NF-�B-Luc expressing hemagglu-tinin (HA)-I�B� (S32A, S36A), green fluorescent protein, HA constitu-tively active (CA)-AKT encoding an amino-terminal myristoylation sig-nal, HA-dominant negative (DN)-AKT, and an NF-�B driven luciferasereporter, respectively, have been described previously (9, 18–20). Therecombinant replication-deficient adenovirus Ad5NCDase was con-structed by AdEasyTM adenoviral vector system (Stratagene, La Jolla,CA). Briefly, the full-length of human NCDase DNA (GenBankTM ac-cession number AF449759) was subcloned into pTrack adenoviral vec-tor. The plasmid DNA was prepared by the alkaline lysis method andtransfected into BJ5183-AD-1 electroporation-competent cells. The vi-rus was grown in 293 cells and purified by banding twice on CsClgradients and then stored in 10% (v/v) glycerol at �20 °C.

Assessment of Cell Death—To assess cell death, cell cultures weredouble-stained with propidium iodide and Hoechst 33258. Propidiumiodide is a vital nucleic acid-staining dye that penetrates cells with com-promised plasma membrane (necrotic cells). Morphological changes in thenuclei of cells undergoing apoptotic cell death were determined by stain-ing with the DNA-binding fluorochrome Hoechst 33258. Apoptotic nuclearchanges include condensation, margination, and segmentation of the nu-clei into several fragments. Briefly, cells were stained with both dyes (100�M) for 20 min and examined under a fluorescent microscope (OlympusIX72, Tokyo, Japan). Quantitation of apoptotic and necrotic cells wasperformed by counting at least 500 cells and was expressed as a percent-age of total cells counted.

Animal Treatment—The experiments were conducted in accordancewith the institutional guidelines by Columbia University. Eight-week-old male BALB/c mice were infected with Ad5GFP or Ad5NCDase (1 �109 plaque-forming units) by intravenously injecting the tail vein. GalN

was dissolved in sterile, nonpyrogenic saline solution and was admin-istered to the mice intraperitoneally (20 mg/mouse), 2 days after ade-novirus administration. Recombinant mouse TNF-� was diluted withpyrogen-free saline and injected to mice intravenously (0.5 �g/mouse)30 min after GalN treatment. Treated animals were anesthetized andkilled by withdrawal of blood from the inferior vena cava. Hepatocellu-lar injury was monitored biochemically by measuring serum alanineaminotransferase (ALT) activity. The liver was excised, fixed with 10%buffered formalin, sectioned at a thickness of 5 �m, and stained withhematoxylin and eosin for light microscopic examination. Apoptoticcells were estimated by the terminal deoxynucleotidyltransferase nick-end labeling (TUNEL) assay, which relies on incorporation of labeleddUTP at sites of DNA breaks. For the TUNEL procedure, all reagents,including buffers, were parts of a kit (ApopTag, Chemicon, Temecula,CA). Procedures were carried out according to the manufacturer’sinstructions.

Western Blot Analysis—For the preparation of total cell proteins,cells or frozen liver were sonicated in radioimmunoprecipitation assaybuffer (50 mM Tris-HCl, pH 8.8, 150 mM NaCl, 10 mM EGTA, 1% TritonX-100, 0.1% SDS, 1% deoxycholic acid, 0.3 mM phenylmethylsulfonylfluoride, 1 mM sodium orthovanadate, 10 mM sodium fluoride, 0.1 mM

sodium molybdate, 0.5 mM 4-deoxypyridoxine). The proteins were sep-arated by SDS-PAGE and were electrophoretically transferred ontonitrocellulose membrane. The membranes were first incubated with theprimary antibodies and then incubated with the anti-mouse or -rabbithorseradish peroxidase-coupled secondary antibodies. Detection wasperformed with an ECL system.

Reverse Transcription-PCR Analysis—Total RNAs isolated fromhepatocytes infected with Ad5NCDase or Ad5GFP were reverse-tran-scribed using random hexamer mixed primers. cDNAs were amplifiedwith the following primer sets: human NCDase sense, 5�-AGAACCT-GATGGGTCCAATCG-3�, and human NCDase antisense, 5�-GGGA-

FIG. 2. ASMase inhibitor treatment, imipramine, or ASMase(�/�) hepatocytes are resistant to cell death induced by TNF-�. A, ratprimary hepatocytes infected with Ad5I�B were pretreated with or without imipramine (50 �M) for 1 h and then treated with TNF-� (30 ng/ml)for the indicated times (left panel). Mouse primary hepatocytes from wild type or ASMase(�/�) mice were infected with Ad5I�B and then exposedto TNF-� for the indicated times (right panel). Apoptotic nuclei cells stained with Hoechst 33258 (upper panel) and necrotic cells with propidiumiodide (lower panel) were counted, and the percentage of cell apoptosis and necrosis out of 500 cells was determined. Data are means � S.D. fromat least three independent experiments. *, p � 0.01 using Student’s t test. B, the C16- and total ceramide (Cer) contents of rat primary hepatocytes(left panel) or mouse primary hepatocytes (right panel) at 12 h after TNF-� treatment were measured by using MS analysis. Data are means fromtwo independent experiments. C, mouse primary hepatocytes from wild type or ASMase(�/�) mice were infected with Ad5I�B or Ad5GFP and thenexposed to TNF-� with or without ASMase (1 IU/ml) for 8 h. Apoptotic nuclei cells stained with Hoechst 33258. N. S., not significant.

NCDase Protects Hepatocyte Apoptosis Dependence on AKT 27881

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

CAAGTGCTATTGGCGTTA-3�; rat NCDase sense, 5�-TGAAGACGAC-GTGTAAAGCCGC-3�, and rat NCDase antisense, 5�-TGCGATAACGA-CAGTCATATCC, at an annealing temperature of 58 °C. After ampli-fication, PCR products were resolved on a 1.5% agarose gel andvisualized by ethidium bromide staining.

NF-�B Reporter Gene Assay—Hepatocytes were infected withAd5NF-�Bluc (10 m.o.i.) and Ad5I�B or Ad5GFP for 18 h. NF-�B-de-pendent gene transcription was analyzed 8 h later on a luminometer(FLUOstar OPTIMA, BMG LABTECH, Inc., Durham, NC) using anenhanced luciferase assay kit (Pharmingen) and adjusted forprotein content.

Mass Spectrometric Analysis of Lipids—These were performed using

electrospray ionization MS/MS analysis on a Thermo Finnigan TSQ7000 triple quadruple mass spectrometer, operating in multiple reac-tion monitoring positive ionization mode as described previously (21).

Measurement of S1P Formation—S1P formation was measured asdescribed previously (8) with slight modifications. For the radiolabelingof sphingolipids, the primary rat hepatocytes were incubated with theplating medium containing [3H]serine (2 �Ci/ml) for 12 h. The mediumwas then changed to serum-free RPMI 1640 medium containing[3H]serine, and the cells were incubated for another 18 h with orwithout adenovirus infection. The radiolabeled cells were stimulatedwith or without S1P, and the cellular lipids were extracted by themethod of Bligh and Dyer and separated on TLC plates in the solvent

FIG. 3. Overexpression of NCDase reduces hepatocyte apoptosis induced by TNF-�. A, cDNAs derived from Ad5GFP- or Ad5NCDase-infected rat primary hepatocytes RNA were analyzed by reverse transcription-PCR using specific oligonucleotide primers of human NCDase, ratNCDase, and �-actin. PCR with RNA was performed as negative control. B, primary rat hepatocytes infected with Ad5NCDase plus Ad5GFP orAd5I�B was treated with TNF-� (30 ng/ml) for 8 h. Apoptotic nuclei cells stained with Hoechst 33258 were counted among 500 cells, and thepercentage of cell apoptosis was determined (upper panel). Data are means � S.D. from at least three independent experiments. *, p � 0.01 usingStudent’s t test. Cellular protein was extracted 12 h after TNF-� administration and subjected to SDS-PAGE. Immunoblotting was performed withanti-cleaved caspase-3 and �-actin antibodies. A densitometric scan of the Western blot analysis was performed (lower panel). Cleaved caspase-3/�-actin ratio is indicated. Data are the means from two independent experiments. C, hepatocytes infected with Ad5I�B and, Ad5NCDase, orAd5GFP were treated with TNF-� for 12 h. The C16- and total ceramide contents were measured using MS analysis. Data are means from twoindependent experiments.

NCDase Protects Hepatocyte Apoptosis Dependence on AKT27882

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

system of 1-butanol/acetic acid/water (60:20:20, v/v). The radiolabeledS1P spot, identified by comigration with an authentic standard, wasscraped off from the plate, and the radioactivity was measured in aliquid scintillation counter.

RESULTS

Increase of C16-ceramide Levels during TNF-�-inducedHepatocyte Apoptosis—TNF-� induced NF-�B activation inrat primary hepatocytes as determined by luciferase assay.The activation was almost abolished when hepatocytes wereinfected with Ad5I�B but not with the control adenovirusAd5GFP (Fig. 1A). Normal hepatocytes are usually resistantto the cytotoxicity of TNF-�. However, infection with Ad5I�Bsensitizes hepatocytes to TNF-�-mediated apoptosis (8, 19,22, 23). To determine the extent of cell death induced byTNF-� under these conditions, hepatocytes were labeled withpropidium iodide, a vital stain, and the permeable DNA-binding fluorochrome Hoechst 33258. Ad5I�B infection sen-sitized hepatocytes to both apoptosis and necrosis induced by

TNF-� (Fig. 1B), as reported previously. Previous studieshave demonstrated that TNF-� causes activation of caspasesin hepatocytes infected with Ad5I�B (8, 22). Cleavage ofcaspase-3 occurred within 12 h after TNF-� treatment (Fig.1C). Moreover, Ad5CrmA infection, which expresses crmA, aserpin inhibitor of a subset of caspases including caspases-1and -8 (24, 25), completely prevented cell death (data notshown).

Previous studies demonstrated that TNF-� induces ceramideformation in hepatocytes by activation of ASMase (7, 26), andceramide is thought to be a lipid mediator involved in theapoptotic process (2, 27). Although total ceramide levels inresponse to TNF-� did not show remarkable change in bothAd5GFP- and Ad5I�B-infected hepatocytes (Fig. 1D), C16-cer-amide increased 1.7-fold in Ad5I�B-infected hepatocytes afterTNF-� treatment (Fig. 1D). Thus, TNF-�-induced cell deathwas accompanied with selective increase in the levels of onlyone ceramide species.

FIG. 4. Overexpression of NCDasereduces liver damage induced byTNF-�. BALB/c mice were infected withAd5GFP or Ad5NCDase (1 � 109 plaque-forming units/mouse) by intravenous ad-ministration. 48 h later, the mice weretreated with (per mouse) 20 mg of GalNintraperitoneally 30 min before TNF-�treatment. Recombinant mouse TNF-�was injected intravenously (0.5 �g/mouse). A, serum ALT levels were deter-mined 6 h after TNF-� injection. Data aremeans � S.D. from at least three inde-pendent experiments. *, p � 0.05 usingStudent’s t test. B, liver sections werestained with hematoxylin and eosin forlight microscopic examination (upperpanel). Apoptotic nuclei in mouse liverwere identified using TUNEL (lowerpanel). C, protein extracts from liver tis-sue were subjected to SDS-PAGE andimmunoblotting was performed withanti-cleaved caspase-3 (casp-3) and�-actin antibodies. The results shown arerepresentative of three independentexperiments.

TABLE IChanges in the sphingolipid profile in Ad5NCDase-infected hepatocytes

Rat primary hepatocytes were infected with Ad5GFP or Ad5NCDase. Sphingosine and the different ceramide species were examined by MSanalysis. The results are expressed as picomoles of lipid/�g of Pi, and they represent the mean from two independent experiments. Theabbreviations used are as follows: Cer, ceramide; dh-Cer, dihydroceramide.

Sph C14-Cer dhC16-Cer C16-Cer C18:1-Cer C18-Cer C20-Cer C24:1-Cer C24-Cer

Ad5GFP 55.88 15.12 5.34 251.75 57.23 168.87 14.88 763.55 1085.15Ad5NCD 28.81 10.99 3.09 151.66 45.14 134.74 13.21 754.86 1048.04

NCDase Protects Hepatocyte Apoptosis Dependence on AKT 27883

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

Protective Effect of ASMase Inhibition against HepatocyteApoptosis by TNF-�—We next aimed to determine the role ofC16-ceramide in cell death. We initially attempted to examine theeffects of exogenous C16-ceramide. However, rat primary hepato-cytes did not take up C16-ceramide even when dissolved in eth-anol/dodecane (98:2, v/v) (data not shown). Therefore, we re-sorted to a combination of pharmacologic, molecular, and geneticapproaches to investigate the role of endogenous ceramide.

It was reported previously (7) that ASMase knock-out miceare resistant to TNF-�-induced liver injury. Imipramine, atricyclic antidepressant, induces the proteolysis of the active72-kDa ASMase form and hence inhibits ASMase activity (28).Pretreatment with imipramine protected rat primary hepato-cytes from apoptosis and necrosis induced by Ad5I�B plusTNF-� (Fig. 2A, left panel). However, because imipramine mayhave pleiotropic effects, we also used ASMase knock-out mice.Indeed, primary hepatocytes from ASMase knock-out miceshowed less apoptosis and necrosis than wild type hepatocytes(Fig. 2A, right panel). Furthermore, imipramine treatment orASMase knock-out cells prevented the increase of C16-ceramdeby Ad5I�B plus TNF-� (Fig. 2B), and exogenous ASMase ad-ministration reversed susceptibility of apoptosis in ASMaseknock-out hepatocytes (Fig. 2C). These results suggest thatASMase activation and C16-ceramde accumulation contributeto hepatocyte apoptosis.

Protective Effects of NCDase Overexpression in Hepatocytes—Ceramide levels are regulated not only through formation (e.g.

by SMase) but also by degradation through the action ofCDases. Besides an acidic form, which is localized in the lyso-somes (12), a neutral ceramidase has been identified andcloned (14, 16). We hypothesized that this neutral ceramidasemay play a role in counterbalancing ceramide generation bythe sphingomyelinases. Thus, we constructed a human NC-Dase expressing adenovirus, and we investigated its effects onTNF-�-induced apoptosis. PCR analysis revealed that humanNCDase was expressed in Ad5NCDase-infected rat primaryhepatocytes and that endogenous rat NCDase levels were notaffected by Ad5NCDase infection (Fig. 3A). Ad5NCDase infec-tion increased the survival of rat primary hepatocyte apoptosisafter TNF-� and reduced caspase-3 cleavage (Fig. 3B), butAd5NCDase did not effect the expression of I�B mutant formby Ad5I�B (supplemental Fig. 1). Ad5NCDase induced thelargest decrease in C16-ceramide (40% reduction) (Table I) fol-lowed by C18:1- and C18-ceramide (20% reduction). On the otherhand, C24- and C24:1-ceramide levels were not significantlyaffected. The levels of C14-, dihydro-C16-, and C20-ceramidewere extremely low in hepatocytes. Ad5NCDase infection in-hibited the increase of C16-ceramide by Ad5I�B plus TNF-�(Fig. 3C). To assess further the protective role of NCDaseduring hepatocyte apoptosis, its effect on TNF-�-mediated liverinjury was examined in vivo. Administration of Ad5NCDasevia tail vein, which expresses both NCDase and GFP, resultedin GFP expression and NCDase mRNA expression in liver ofBALB/c mice (supplemental Fig. 2). In BALB/c mice infected

FIG. 5. Sphingosine kinase inhibitor, DMS, cancels the anti-apoptotic effect of NCDase overexpression. A, rat primary hepatocytesinfected with Ad5GFP, Ad5I�B, and/or Ad5NCDase were pretreated with or without 10 �M DMS for 2 h before 30 ng/ml TNF-� treatment and werefurther incubated for 8 h. Apoptotic nuclei cells stained with Hoechst 33258 were counted among 500 cells, and the percentage of cell apoptosis wasdetermined. Data are means � S.D. from at least three independent experiments. *, p � 0.01 using Student’s t test (upper panel). Cellular proteinextracted 12 h after TNF-� administration was subjected to SDS-PAGE, and immunoblotting was performed with anti-cleaved caspase-3 and�-actin antibodies (lower panel). A densitometric scan of the Western blot analysis was performed. Cleaved caspase-3/�-actin ratio is indicated.Data are the means from two independent experiments. B–D, rat hepatocytes infected with Ad5GFP or Ad5NCDase were pretreated with orwithout the indicated concentration of DMS for 2 h (B and C) or were treated with 1 �M S1P for the indicated times (D). E, rat hepatocytes werepretreated with or without the indicated concentration of PTX for 12 h and were then infected with Ad5NCDase for 18 h or treated with S1P for1 h. S1P levels were measured in [3H]serine-labeled hepatocytes (B). Cellular protein extracts were subjected to SDS-PAGE, and immunoblottingwas performed with anti-phosphorylated AKT and �-actin antibodies (C–E). The results shown are representative of at least two independentexperiments.

NCDase Protects Hepatocyte Apoptosis Dependence on AKT27884

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

with Ad5GFP, GalN plus TNF-� elicited liver damage, and theserum ALT level increased to about 7000 IU/liter at 6 h (Fig.4A). Histological analysis revealed that GalN plus TNF-� re-sulted in pronounced hepatocyte destruction in large areas ofthe liver (Fig. 4B, upper panel), with many TUNEL-positivehepatocytes (Fig. 4B, lower panel), and an increase in caspase-3cleavage (Fig. 4C). Overexpression of NCDase reduced serumALT levels by �70% (Fig. 4A). Moreover, overexpression ofNCDase reversed morphological changes induced by GalN plusTNF-� and strongly reduced the number of TUNEL-positivehepatocytes (Fig. 4B) and caspase-3 cleavage (Fig. 4C). Thus,overexpression of neutral ceramidase protects from TNF-�-induced hepatotoxicity.

Effects of SphK Inhibitor, DMS, on Hepatocyte Apoptosis—Ceramide is hydrolyzed to sphingosine, which is subsequentlyconverted to S1P by SphK. Previous studies demonstrated thatSphK activation can mediate anti-apoptotic actions via S1Pformation (8, 29). NCDase reduced sphingosine in rat hepato-cytes (Table I), suggesting that sphingosine was probably me-tabolized to another sphingolipid like S1P. Indeed, pretreat-ment with the SphK inhibitor DMS blocked the anti-apoptoticeffect of NCDase overexpression (Fig. 5A, upper panel), and itprevented the cleavage of caspase-3 (Fig. 5A, lower panel).Moreover, NCDase overexpression increased S1P level, and

DMS reversed this effect (Fig. 5B). Thus, the anti-apoptoticeffect of NCDase overexpression was not due to ceramide deg-radation alone, but appears to be contributed to by the forma-tion and phosphorylation of sphingosine and a change in thebalance of C16-ceramide and S1P.

To begin to elucidate the cytoprotective mechanisms ofNCDase, we characterized the effects of NCDase on TNF-�-induced cytoprotective pathways. Because NF-�B activationwas already completely blocked by Ad5I�B and not furthermodified by NCDase under the conditions of our study, wefocused on potential effects of NCDase on the anti-apoptoticmediator AKT (8, 23, 30, 31). Infection of Ad5NCDase inducedphosphorylation of AKT, and DMS inhibited this effect in aconcentration-dependent manner (Fig. 5C). DMS could not in-hibit AKT phosphorylation by insulin (data not shown), sug-gesting that this effect of DMS was specifically due to SphKinhibition. Moreover, exogenous S1P administration inducedphosphorylation of AKT (Fig. 5D). Thus, NCDase overexpres-sion resulted in AKT phosphorylation through S1P formation.

Accumulating evidence indicates that S1P formed by SphKacts not only as an autocrine and/or paracrine ligand via theEdg receptor but also as an intracellular second messenger(32). To examine the involvement of Edg receptor(s) in theactivation of AKT, hepatocytes were pretreated with PTX for

FIG. 6. Constitutively active AKT protects hepatocytes from TNF-�-induced apoptosis, and dominant negative AKT eliminates theanti-apoptotic effect of NCDase overexpression. A and B, rat hepatocytes infected with Ad5GFP, Ad5NCDase, Ad5CA-AKT, and/orAd5DN-AKT were treated with 30 ng/ml TNF-�. Eight hours after TNF-� administration, typical apoptotic cells stained with Hoechst 33258 werecounted among 500 cells, and the percentage of apoptotic cells was determined (upper panel). Data are means � S.D. from at least threeindependent experiments. *, p � 0.01 using Student’s t test. Cellular protein extracted 12 h after TNF-� administration was subjected toSDS-PAGE, and immunoblotting was performed with anti-cleaved caspase-3 (casp-3) and �-actin antibodies. A densitometric scan of the Westernblot analysis was performed. Cleaved caspase-3/�-actin ratio is indicated. Data are the means from two independent experiments (lower panel).

NCDase Protects Hepatocyte Apoptosis Dependence on AKT 27885

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

12 h, because Edg receptors are sensitive to PTX. AKT activa-tion by exogenous S1P was completely blocked by PTX pre-treatment (Fig. 5E, upper panel), indicating that extracellularS1P acted through the Edg receptor(s) coupled with Gi/o. How-ever the activation of AKT by NCDase overexpression was onlypartially inhibited (Fig. 5E, lower panel). Thus, the effect ofS1P on AKT phosphorylation is probably by both an intracel-lular effect and via Edg receptors, although the participation ofPTX-insensitive S1P receptor signaling cannot be ruled out.

Protective Effect of AKT Activation on TNF-�-induced Apo-ptosis—To confirm that AKT activation does rescue hepato-cytes from TNF-�-induced apoptosis, we infected hepatocyteswith Ad5CA-AKT. Specific AKT activation by CA-AKT overex-pression drastically reduced Ad5I�B plus TNF-� induced apo-ptosis and caspase-3 cleavage (Fig. 6A). More importantly,infection of dominant negative AKT, Ad5DN-AKT, eliminatedthe anti-apoptotic effect of NCDase overexpression (Fig. 6B,upper panel) and prevented the cleavage of caspase-3 (Fig. 6B,lower panel), demonstrating a causative link between NCDase-induced AKT activation and its anti-apoptotic effects. Thesefindings suggest that the mechanism by which NCDase over-expression protects hepatocytes includes S1P formation andAKT activation.

DISCUSSION

TNF-� is a potent mediator of hepatotoxicity in vivo and incultured hepatocytes. However, TNF-� simultaneously acti-vates cytoprotective and cytotoxic signaling pathways, andTNF-� alone does not induce cell death in normal hepatocytes(22, 23, 26). Although several TNF-�-induced pro- and anti-apoptotic pathway have been characterized in hepatocytes, in-cluding NF-�B, c-Jun NH2-terminal kinase/AP-1, and AKT, therole of sphingolipids in TNF-�-induced cell death is not com-pletely understood (8, 22, 23, 33, 34). The present study wasundertaken to address specifically the role of sphingolipids inthe apoptotic signaling of TNF-� in hepatocytes using Ad5I�B-infected rat and mouse primary hepatocytes. Our results dem-onstrate that overexpression of NCDase inhibits TNF-�-in-duced hepatocyte apoptosis via AKT activation. The resultsraise novel therapeutic possibilities for ameliorating acuteliver injury.

TNF-� has been reported to induce ceramide formation, atleast in part, by activation of ASMase (26). Ceramide genera-tion by ASMase is involved in TNF receptor or Fas-inducedsignals. Deoxycholic acid activates the c-Jun NH2-terminal ki-nase pathway via Fas translocation to the plasma membrane,which is induced by ceramide generation by ASMase activationin primary hepatocytes (35). ASMase knock-out mice are re-sistant to TNF-�-induced liver injury (7). Indeed, administra-tion of TNF-� increased C16-ceramide content in both rat andmouse primary hepatocytes, and the ASMase inhibitor imipra-mine or ASMase knock out blocked increase of C16-ceramideand apoptosis induced by TNF-�. Moreover, the de novo ceram-ide synthase inhibitor fumonisin B1 or the neutral sphingomy-elinase inhibitor N-acetylcysteine (36) did not inhibit this apo-ptosis (data not shown). These results suggest that activationof ASMase contributes to TNF-�-induced hepatocyte apoptosis.These results and previous reports lead us to speculate thatC16-ceramide formation is specifically required for apoptosisinduced by TNF-�.

TNF receptor I and Fas are associated with lipid rafts, and itis believed that the generation of ceramide by ASMase withinthese cholesterol- and sphingolipid-rich domains is required forTNF receptor- and Fas-mediated apoptosis (37–40). However,pretreatment of methyl-�-cyclodextrin, a cholesterol-depletingagent that is commonly used to disrupt lipid rafts, did notprevent hepatocyte apoptosis in our model (data not shown).

Another pathway by which ceramides may contribute to apo-ptosis is the generation of glucosylceramide and glycosphingo-lipids (such as GD3). It has been reported that GD3 has apo-ptotic effects by affecting mitochondria (41, 42). In hepatocytes,GD3 is targeted to mitochondria after TNF-� treatment, andadministration of exogenous GD3 induces apoptosis in hepato-cytes sensitized by depletion of mitochondrial glutathione (7).However, exogenous GD3 (50 �M) did not induce cell death inAd5I�B-infected hepatocytes (data not shown). Thus, both GD3formation and ceramide generation in lipid rafts do not appearto exert major effects on hepatocyte apoptosis in this model.

Ceramides are also converted into sphingosine by CDase,and sphingosine is phosphorylated by SphK to S1P. Althoughceramide has been regarded as a proapoptotic factor, S1P hasbeen implicated as a survival factor (43). Ceramide and itsmetabolite S1P often exert opposing effects, thus the balancebetween these two sphingolipids may be important for cell fate(11, 43, 44). CDase is classified into the following three groupsbased on optimum catalytic pH: ACDase, alkaline, and NC-Dase. We now report that overexpression of NCDase reducedhepatocyte apoptosis by TNF-� both in tissue culture as well asin vivo. Most interestingly, in NCDase overexpressing hepato-cytes, total ceramide levels were not modulated/changed be-cause C24-ceramide and C24:1-ceramide, the major contents ofceramides in hepatocytes, were not affected; however, C16-ceramide was significantly reduced, again implicating this spe-cific ceramide in mediating apoptosis. In addition to changes inceramide, overexpression of NCDase also decreased sphingo-sine and increased S1P levels, suggesting that the metabolitewas further converted to S1P. The SphK inhibitor DMS elimi-nated the anti-apoptotic effect of NCDase. Thus, the anti-apoptotic effect of NCDase is due to both C16-ceramide reduc-tion and S1P formation. S1P stimulates many signalingpathways, such as cAMP-dependent kinase, focal adhesionkinase, extracellular signal-regulated kinase, AP-1, and NF-�B(10, 32). S1P also activates the PI3K/AKT pathway in endothelialdifferentiation gene (EDG)3 receptor overexpressing Chinesehamster ovary cells (45) and human hepatocytes (8). The PI3Kproduct, phosphatidylinositol 3,4,5-triphosphate activatesphosphatidylinositol-dependent kinase, which activates AKT.Activation of AKT protects cells from apoptosis induced by TNF-�(8, 23, 30, 31) by phosphorylating MDM2, AFX, FKHR, mTOR,and BAD (46, 47). Reciprocally, ceramide has been shown to

FIG. 7. Model for sphingolipid signaling pathway by TNF-� inhepatocytes.

NCDase Protects Hepatocyte Apoptosis Dependence on AKT27886

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

induce dephosphorylation of AKT through activation of PP2Aserine/threonine protein phosphatase; thus, S1P and ceramideexert opposing effects on AKT, and the action of ceramidase isexpected to abrogate the effects of ceramide and enhance theeffects of S1P. Indeed, in rat primary hepatocytes, overexpressionof NCDase activated AKT, and this activation was blocked byDMS treatment. Administration of exogenous S1P also activatedAKT. Moreover, DN-AKT prevented the anti-apoptotic effect ofNCDase, indicating that the anti-apoptotic effect of NCDase de-pends on AKT activation by S1P. S1P is reported to act as aligand for the Edg receptor(s) and also as an intracellular secondmessenger (32, 43). In our system, PTX pretreatment inhibitedexogenous S1P-induced AKT phosphorylation. In contrast, PTXcaused partial inhibition of the AKT activation by NCDase over-expression. Thus, S1P produced by NCDase activates AKT intra-cellularly and also functions as an extracellular ligand for Edgreceptor(s) in primary hepatocytes. Although it has been shownthat ACDase also protects from TNF-�-induced apoptosis (13), itis not clear whether this protective effect is mediated throughAKT activation. Further studies need to examine the effects ofseveral kinds of CDases on apoptosis signals.

In summary, we have shown that overexpression of NCDaseprotects from TNF-�-induced hepatocyte apoptosis via AKTactivation by S1P (Fig. 7). Our results suggest that the regu-lation of NCDase may present a new therapeutic approach toblock apoptosis in liver disease.

Acknowledgments—We thank Cecilia Devlin and Yoshiko Banno forassistance with S1P measurements.

REFERENCES

1. Bradham, C. A., Plumpe, J., Manns, M. P., Brenner, D. A., and Trautwein, C.(1998) Am. J. Physiol. 275, G387–G392

2. Osawa, Y., Banno, Y., Nagaki, M., Nozawa, Y., Moriwaki, H., and Nakashima,S. (2001) Liver 21, 309–319

3. Nagata, S. (1997) Cell 88, 355–3654. Hannun, Y. A. (1996) Science 274, 1855–18595. Spiegel, S., Foster, D., and Kolesnick, R. (1996) Curr. Opin. Cell Biol. 8,

159–1676. Kolesnick, R. N., and Kronke, M. (1998) Annu. Rev. Physiol. 60, 643–6657. Garcia-Ruiz, C., Colell, A., Mari, M., Morales, A., Calvo, M., Enrich, C., and

Fernandez-Checa, J. C. (2003) J. Clin. Investig. 111, 197–2088. Osawa, Y., Banno, Y., Nagaki, M., Brenner, D. A., Naiki, T., Nozawa, Y.,

Nakashima, S., and Moriwaki, H. (2001) J. Immunol. 167, 173–1809. Hatano, E., and Brenner, D. A. (2001) Am. J. Physiol. 281, G1357–G1368

10. Maceyka, M., Payne, S. G., Milstien, S., and Spiegel, S. (2002) Biochim.Biophys. Acta 1585, 193–201

11. Cuvillier, O., Pirianov, G., Kleuser, B., Vanek, P. G., Coso, O. A., Gutkind, S.,and Spiegel, S. (1996) Nature 381, 800–803

12. Li, C. M., Hong, S. B., Kopal, G., He, X., Linke, T., Hou, W. S., Koch, J., Gatt,S., Sandhoff, K., and Schuchman, E. H. (1998) Genomics 50, 267–274

13. Strelow, A., Bernardo, K., Adam-Klages, S., Linke, T., Sandhoff, K., Kronke,

M., and Adam, D. (2000) J. Exp. Med. 192, 601–61214. Tani, M., Okino, N., Mori, K., Tanigawa, T., Izu, H., and Ito, M. (2000) J. Biol.

Chem. 275, 11229–1123415. Mitsutake, S., Tani, M., Okino, N., Mori, K., Ichinose, S., Omori, A., Iida, H.,

Nakamura, T., and Ito, M. (2001) J. Biol. Chem. 276, 26249–2625916. El Bawab, S., Roddy, P., Qian, T., Bielawska, A., Lemasters, J. J., and Han-

nun, Y. A. (2000) J. Biol. Chem. 275, 21508–2151317. Moldeus, P., Hogberg, J., and Orrenius, S. (1978) Methods Enzymol. 52, 60–7118. Bradham, C. A., Qian, T., Streetz, K., Trautwein, C., Brenner, D. A., and

Lemasters, J. J. (1998) Mol. Cell. Biol. 18, 6353–636419. Iimuro, Y., Nishiura, T., Hellerbrand, C., Behrns, K. E., Schoonhoven, R.,

Grisham, J. W., and Brenner, D. A. (1998) J. Clin. Investig. 101, 802–81120. Schwabe, R. F., Bataller, R., and Brenner, D. A. (2003) Am. J. Physiol. 285,

G949–G95821. Pettus, B. J., Bielawski, J., Porcelli, A. M., Reames, D. L., Johnson, K. R.,

Morrow, J., Chalfant, C. E., Obeid, L. M., and Hannun, Y. A. (2003) FASEBJ. 17, 1411–1421

22. Xu, Y., Bialik, S., Jones, B. E., Iimuro, Y., Kitsis, R. N., Srinivasan, A.,Brenner, D. A., and Czaja, M. J. (1998) Am. J. Physiol. 275, C1058–C1066

23. Osawa, Y., Nagaki, M., Banno, Y., Brenner, D. A., Nozawa, Y., Moriwaki, H.,and Nakashima, S. (2003) J. Immunol. 170, 4053–4060

24. Ray, C. A., Black, R. A., Kronheim, S. R., Greenstreet, T. A., Sleath, P. R.,Salvesen, G. S., and Pickup, D. J. (1992) Cell 69, 597–604

25. Zhou, Q., Snipas, S., Orth, K., Muzio, M., Dixit, V. M., and Salvesen, G. S.(1997) J. Biol. Chem. 272, 7797–7800

26. Schwandner, R., Wiegmann, K., Bernardo, K., Kreder, D., and Kronke, M.(1998) J. Biol. Chem. 273, 5916–5922

27. Obeid, L. M., Linardic, C. M., Karolak, L. A., and Hannun, Y. A. (1993) Science259, 1769–1771

28. Grassme, H., Gulbins, E., Brenner, B., Ferlinz, K., Sandhoff, K., Harzer, K.,Lang, F., and Meyer, T. F. (1997) Cell 91, 605–615

29. Xia, P., Wang, L., Gamble, J. R., and Vadas, M. A. (1999) J. Biol. Chem. 274,34499–34505

30. Madge, L. A., and Pober, J. S. (2000) J. Biol. Chem. 275, 15458–1546531. Takano, R., Hisahara, S., Namikawa, K., Kiyama, H., Okano, H., and Miura,

M. (2000) J. Biol. Chem. 275, 16360–1636532. Pyne, S., and Pyne, N. J. (2002) Biochim. Biophys. Acta 1582, 121–13133. Hatano, E., Bradham, C. A., Stark, A., Iimuro, Y., Lemasters, J. J., and

Brenner, D. A. (2000) J. Biol. Chem. 275, 11814–1182334. Ding, W. X., and Yin, X. M. (2004) J. Cell Mol. Med. 8, 445–45435. Gupta, S., Natarajan, R., Payne, S. G., Studer, E. J., Spiegel, S., Dent, P., and

Hylemon, P. B. (2004) J. Biol. Chem. 279, 5821–582836. Yoshimura, S., Banno, Y., Nakashima, S., Hayashi, K., Yamakawa, H.,

Sawada, M., Sakai, N., and Nozawa, Y. (1999) J. Neurochem. 73, 675–68337. Simons, K., and Ikonen, E. (1997) Nature 387, 569–57238. Grassme, H., Cremesti, A., Kolesnick, R., and Gulbins, E. (2003) Oncogene 22,

5457–547039. Ko, Y. G., Lee, J. S., Kang, Y. S., Ahn, J. H., and Seo, J. S. (1999) J. Immunol.

162, 7217–722340. Cottin, V., Doan, J. E., and Riches, D. W. (2002) J. Immunol. 168, 4095–410241. Malisan, F., and Testi, R. (2002) Biochim. Biophys. Acta 1585, 179–18742. De Maria, R., Lenti, L., Malisan, F., d’Agostino, F., Tomassini, B., Zeuner, A.,

Rippo, M. R., and Testi, R. (1997) Science 277, 1652–165543. Olivera, A., and Spiegel, S. (1993) Nature 365, 557–56044. Kolesnick, R. (2002) J. Clin. Investig. 110, 3–845. Banno, Y., Takuwa, Y., Akao, Y., Okamoto, H., Osawa, Y., Naganawa, T.,

Nakashima, S., Suh, P. G., and Nozawa, Y. (2001) J. Biol. Chem. 276,35622–35628

46. Mitsiades, C. S., Mitsiades, N., and Koutsilieris, M. (2004) Curr. Cancer DrugTargets 4, 235–256

47. Sen, P., Mukherjee, S., Ray, D., and Raha, S. (2003) Mol. Cell. Biochem. 253,241–246

NCDase Protects Hepatocyte Apoptosis Dependence on AKT 27887

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from

Supplemental figure1

β-actin

HA

GFPGFP

GFPIκB

NCDaseIκB

Adenovirus

Supplemental figure 1 Rat primary hepatocytes were infected with Ad5GFP, Ad5IκB, and/or Ad5NCDase. Cellular protein extracts were subjected to SDS-PAGE and immunoblotting was performed with anti-HA and β-actin antibodies. The results shown are representative of at least two independent experiments.

0

50

100

150

200

250

No virus Ad5GFP Ad5NCDase

GF

P e

xpre

ssio

n(f

old-

indu

ctio

n fro

m n

o vi

rus)

No virus Ad5GFP Ad5NCDase

Human NCDase

β-actin

Supplemental figure 2

Supplemental figure 2 (A) BALB/c mice were infected with Ad5GFP or Ad5NCDase (1 X 109 pfu / mouse) by intravenous administration. 48 h later, the mice were killed and liver sections were homogenized in Cell lysis Buffer (BD Biosciences Pharmingen, San Diego, CA). After centrifuge, GFP of supernatant was measured by fluorometer (BMG Labtechnology FLUOstar OPTIMA, Burham, NC) and adjusted for protein content. (B) Total RNAs isolated from liver of mouse infected with or without Ad5NCDase or Ad5GFP were reverse transcribed. RT-PCR was performed using specific oligonucleotide primers of Human NCDase and β-actin.

A

B

Hannun and David A. BrennerYosuke Osawa, Hiroshi Uchinami, Jacek Bielawski, Robert F. Schwabe, Yusuf A.

αApoptosis in Response to Tumor Necrosis Factor--ceramide and Sphingosine 1-Phosphate in Regulating Hepatocyte16Roles for C

doi: 10.1074/jbc.M503002200 originally published online June 9, 20052005, 280:27879-27887.J. Biol. Chem. 

  10.1074/jbc.M503002200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

Supplemental material:

  http://www.jbc.org/content/suppl/2005/06/15/M503002200.DC1.html

  http://www.jbc.org/content/280/30/27879.full.html#ref-list-1

This article cites 47 references, 21 of which can be accessed free at

by guest on March 27, 2016

http://ww

w.jbc.org/

Dow

nloaded from