were lasted k>r overnight and APAP (600 mg/kg BW) dissolved in saline was injected intraperitoneally (ip), In the SAMe treatment group, SAMe (lg/kg BW) dissolved in saline was injected i,p, either 4 hours before or 1 hour after APAP admmistration. Blood samples and hver were collected 6, 12 and 24 hours alter APAP treatment, APAP administration induced the typical changes of contluent centrilobular necrosis with microvesicular steatosis by histological examination and an elevation of serum hver enzyme (ALT) activity. APAP administration induced marked decreases ni both hepatic and whole blood SAMe levels. Moreover, APAP decreased intracellular (both cytosol and mitochondria) GSH levels and caused mitochondrial dysfmtction that was verified by a rapid occurrence of permeability transition in response to the presence of 300 ~,M calcium in the incubation buffer. A marked increase in lipid peruxidation was demonstrated by a 6-fold increased in hepatic TBARs, SAMe treatment (both fxef0re and after APAP) significantly' attenuated the liver injury verified by histological examination Treaunent with SAMe prevemed a decrease in SAMe levels in both liver and whole blood. Moreover SAMe treatment attenuated both cytosolic and mitochondrial GSH depletion as well as mitochondrial dysfunction and lipid peroxidation tbltowing APAP administration We conclude that SAMe prutects the liver from APAP- induced iqury by, preventing intracellular GSH depletmn and mitochondrial dysfunction, Thus, SAMe maybe effective therapy tbr APAP overdoses. SAMe is well tolerated when administrated orally which suggests the potential for SAMe as a treatment option for prevent- ing acetamniophen poisoning by incorporation of SAMe with acetaminophen, (Supported by VA NIH RO1 ~e~01762, NIH R01 AA i0496), $904 Hepatocyte Keratin Filament Disruption Up-Regulates Oxidative Stress-Related Genes Qm Zhou, M, Bishr Omary Background: Keratins (>20 proteins.), actins and tubulins make up the 3 major cytoskeletal proteins of epithelial cells, Hepatocytes express keratin 8 (l(8) and 18 (K18) and other epithelial cells express their own nmque kemtins. K8/K18 mutations pose a risk for develop- mere of cirrhosis. Mice that over-express an Arg89-to-Cys (R89C) K18 mutation have a marked predisposition to drug-induced liver injury, in association with fiepatoeyte keratin filament disruption, Aim: Examine mouse liver genetic changes in response to K18 R89C- induced keratin hlament disruption. Methods: We used mouse mtcroarrays (19137 genesd slide). Tntal liver RNA was isolated from previously described: 2 mouse transgenic lines that overexpress R89C K18 (F22/F50 lines) and a line that overexpresses comparable levels of wdd-type R18 (TG2 line). All mice were male (8-10 wk old), Ten experiments were done, 5 comparing TG2 with F22 and 5 companng TG2 with FS0. RNA was reverse transcribed and labeled with Cy3 (F22/FSO) and Cy5 (TG2), Slides *,,'ereincubated overnight (65oC), washed, seanned, then analyzed using cluster/treeview programs, Results: We choose a >2-told change cutoff in RNA to select genes tbr turther assessment in terms of up/down regrtlation Using such criteria, only 25 genes had altered expression (I0 down- and 15 up- regulated). The most prominent genes that were significantly up-regulated upon keratin filament disruption are those involved in oxidative metabolism and stress including cyto- chrome-P450, epoxide hydrdase, acybCoA oxidase, glntathione transterase, and cysteine sulfinic acid decarboxylase. Genes that were down-regulated included fatty, acid binding protein, tubulin-beta-5, and some signaling molecules, Some of the observed genetic profile cKanges were further confirrned by immunoblotting and quantitative PCR. Conclusion: Hepatocyte keratin filament disruption, as a result of a single point mutation, causes dramatic modulation of stress-related genes. This findmg supports the emerging role of keratin mutations m patients with liver disease and provides a tunctional role for keratins in protecting from oxidative injury $905 Endothelin-I Suppresses Plasma Membrane Ca + *-Atpase, Concomitant with Contraction of Hepatic Sinusoidal Endothelial Fenestrae Hiroaki Yokomori, Masaya Oda, Hiromasa ishii Background and Aims: lntmcytoplasmic free cakmm ions ([Ca + +/i) are maintained at a very low concentration in mammalian tissue by extruding Ca + + from the cytoplasm against a steep extracellular Ca + gradient, nrainly through the activity of plasma membrane Ca ++ pump ATPase The present study aimed to elucidate how endothelin-1 (ET-I) affects the morphology ot sinusoidaI endothelial tenestrae and uhrastructural distribution of plasma nmmbrane Ca + +-ATPases and [Ca+ +]i in isolated mt hepatic sinusoidal endothelial cell (SEC)s. Materials & Methods: SECs were isolated from rat livers by collagenase iniKsion method, These cultured endotbehal cells were dB,'ided into control, ET-1 (1 nM)-treated, and ET-1 + Bosetuan (10 ~tM)-treated groups. Tire uhrastructuml localization of Ca+ +-Mg + *-AIPase activity, was examined by the electron cytochemical method of Ando, and that of ~* + pumDATPase by" the electron immunogold post-embedding method. The protein level of Ca ++ pump-ATPase was identified by Western blot. The uhrastructud localization of [Ca + +]i was examined by potassium antimonate method. Results: Addition of ET-I to SEC significantly decreased Ca +*-Mg ~ + -ATPase activity and Ca ++ pump-ATPase expression and increased [Ca ++li concentration, concomitant with a decrease m diameter of smusoidal endothelial tenestrae. Ca ++ purnp-ATPase protein expression was high in control and ET- 1 + Bosentamtreated ceils, and low in ET-l-treated cells. Pre-treatment with Bosentan abol- ished the actions ot ET-1. Conclusions: These results suggest that ET-1 suppresses Ca ++- Mg~+-ATPase activity and Ca + + pump-ATPase expression on the plasma membrane of sinusoidal endothelial tenestrae, thereby attenuating the extrusion of intracellular Ca ++ into the extracellular space, leading to an increased intracytoplasmic free Ca ++ concentration and contractio,, of sinnsoidal endothehal fenestrae, $906 Liver Reproduction by Tissue Engineering : Control of Hepatocyte-Hepatocyte and I-lepatocyte-Scaffold Adhesion by Shear Stress and Extracellular Matrix Mitsuo Miyazawa, Takahiro Torii, lsamu Koyama Aims: For liver reproduction by tissue engineering, the environmental control that 3-dimen- sional scaffold and liver cells begin to make is necessary. Now to some extent a stud3," of a cell and a study of scaffold progress, but a study" to control the environment that cell and scaffold make actively is considerably late. In the present study, we investigated whether control of morpholugical behavior of liver cells was possible by shear stress (SS) and extracellular mamx (ECM) Materials and Methods: Using a rotating radial flow bioreactor (RRFB) (Able, Tokyo), SS was applied in stages to co<uhures of hepatic parenchymal cells (PC) and nonparenchymal cells (NPC) adhering to a scaffold to above a stirrer, by rotating the sea/fold to create centntugal force. PC and NPC were extracted from Fischer rats and seeded omo 5-mm thick pulp at a concentratkm ot 4X106/cm3 and 4X 106/cm3, respectively. 1)Morphological behavior of co-culture with SS or without SS : Thick pulp having collagen coat was tised as a scaffold. In SS cultures, the RRFB scaffold was rotated at 30 rpm to create SS of 5-20 dyne/cm2, in undisturbed cuhums, the cells were co cultured in 120 ml of W-E medium w~thout SS. The two types of culture were then compared by electron microscopy. 2) Morphological behavior of co-culture with ECM or without BCM : The cells were cultured with SS loading on pulp with collagen or without collagen, The two types of culture were then conrpared by ekctmn microscopy. Resuhs: In SS loaded culture on pulp without ECM, the cells were aggregated like spheroid. In SS loaded culture with ECM, the cells attached to pulp like a monolayer. The cells attached like a monolayler to pulp with collagen (ECM) either with SS loading or without SS loadmg. Only in SS loaded culture with ECM, hepatocytes maintained polarity and structure was kept well and microvilli, lipid droplets, bile canalicula and desmosome were recognized. In a culture without SS or without ECM, the hollow formation of hepatocytes stood out and structure destruction was recog- nized. Discussion and Conclusmns: Hepatoc3~'te cannot maintain a kmction if it does not maintain polarity. These results suggested that liver reproduction was possible by changing SS loading and ECM moderately, that could attach hepatocytes to both ECM and other hepatocytes and that could maintain polarity of hepatocytes and make built bile duct struc- ture. $907 Mitogen-Activated Protein Kinases (MAPKs) in Mouse lntrahepatic Immortalized Biliary Epithelial Cells (IBECs): Constitutive Expression and Activation by Cytokines Expressed in Inflamed Livers of Mice with Chronic Graft-vs-Host Disease (CGVHD) Haimei Wang, Marius Braun, John M. Vierling Background: Non-suppurative destructive cholangitis (NSDC) in the B10.D2 into BALB/c model of CGVHD is marked by T cell inflammation of small caliber bile ducts, apoptosis of BECs and ductopema. T cell mfihration of ducts is preceded by secretion of lFN% TNFu and IL-11~by portal inflammatory, cells. These cytokines induce BALB/c target strani IBECs to express chemokines and their receptors. The combination of IFNy + TNF~ also induces 1BEC apoptusis. Although IL-6 promotes profit~:ration of IBECs by activation of the I~s this pathway in BECs exposed to IFN% TNFe~ and IL-I~ has not been studied. Aim: To quantify the kinetics of p38, ERK1/2 and JNK activation in BALB/c IBECs exposed to IFN% TNFc~ and IL-l[3 Methods: IFN"/, TNFa, IL-I~ and IFN~/+ TNFc~ were added to cultured BALB/c intrahepatlc 1BECs. Control and cytokine-exposed IBECs were harvested at eight time points. Activation of ERK1/2, p38 and JNK in IBEC lysates was assessed by Western blotting. Results: Control IBBCs in basal culture constitutively, expressed activated and inactive forms of ERK1/2, p38 and JNK We found that rlFNy reduced activated ERK1/2 in all treatments, increased activated p38 11-fold by 60 min, and reduced activated JNK by 75% by 30 min and 63% thereafter. In contrast, rTNFc~ induced a bimodal increase of activated ERK1/2 (5.8-fold peak at 15 mm and 2 2&~id peak at 360 mio) and a 2.7-fold increase of activated p38 at 30 ram, md a 10-fold increase in actrvated JNK by 15 nrin. rlFN'y+rTNFa induced a bimodal increase of activated ERK1/2 (2.3-fold peak at 15 mm and 3.2-Ibld peak at 360 mitt), a 7 5-krld increase in activated p38 at 60 ram, and a 14- tbld increase in activated JNK at 15 min rig 113reduced activated ERK1/2 at all time points, increased activated p38 2-fold at 30-60 rain and activated JNK by up to 5.8-fold at 7.7-15 min. Conclusions: IFN% TNFa, IFN~,+ TNFc~ and 1L-I~ induced disparate changes in activated ERK1/2, p38, and jNK. Both IFNy and IL-I~ inhibited ERK1/2, while 1FN"/also inhibited JNK. in contrast, TNFc~ activated ERK1/2, p38 and jNK. Similarly, rlFN7 + rTNFc~ abundant in inflamed portal tracts during NSDC, activated ERK1/2, p3B and JNK. MAPK signaling in BECs by cytokines secreted by portal inflammatory cells likely alters the BECs targets of NSDC by modifying their gene expression, susceptibility to apoptosis and capacity" to repopulate inflamed bile ducts. Supported by NIH RO1DK53945~ $9o8 Co-induction of HO1 and ATP-dependent iron transport by stressful stimuli Yingda Wang, Susan Williams, Christopher Ferris Heine Oxygenase 1 (HOD is a heat-shock protein that is induced by stressful stimuli (Mames, M D, 1997, Annu, Rev, PMrmacoL Toxicol. 37, 517). Increasing evidence has accumulated in vitro and in vivo that HOI participates in cellular defense mechanisms against oxidative stress, Induced HO1 degrades haem, releasing substantial quantities of fi*ee iron, which would be expected to be toxic though Fenton Chemistry. We discovered that HO1 activity is linked to the extrusion of iron from cells, which explains this apparent paladox and provides a mechanism for cellular protection following HO1 induction, Inhibition of iron release by genetic deletion of HO1 and by the HO inhibitor SnPPIX indicates that haem itself may be the source iron of iron ielease or mobilization from cells (Ferris, CDet al Nature Cell Biol. 1999, 1, 152). We recently, identified an iron-transporting ATPase that is induced by iron, co-localizes with HO1 and may act in association with HO1 to mediate the release/mobilization of cellular iron (Baranano, D.E. et al. J. Biol. Chem. 2000, 275, AASLD Abstracts A-724