Inhibition of cytokine-induced nitric oxide synthase expression by gene transfer of adenoviral IκBα Bradley S. Taylor, MD, Lifang Shao, BS, Andrea Gambotto, PhD, Raymond W. Ganster, PhD, and David A. Geller, MD, Pittsburgh, Pa
Background. Nitric oxide is overexpressed in nearly every organ during sepsis and it has profound biologic effects. Previously, we showed that maximal inducible nitric oxide synthase (iNOS) expression is up-regulated by a combination of cytokines and that this effect is mediated by the transcription factor NFκB. Therefore the purpose of this study was to establish whether gene transfer of the inhibitory molecule IκB would result in the abrogation of cytokine-induced iNOS expression. Methods. Cultured hepatocytes were infected with an adenoviral vector containing the IκBα gene (Ad5IκB) and after an 18-hour recovery period were stimulated with the cytokine mixture of tumor necrosis factor-α (500 U/mL) plus interleukin 1β (200 U/mL) plus interferon gamma (100 U/mL). Results. As expected, cytokine mixture induced significant hepatocyte nitrite (NO2–) and iNOS messenger RNA production. Cells infected with the IκBα gene showed a dose-dependent decrease in NO2– and iNOS messenger RNA levels. Western blot analysis showed a marked decrease in iNOS protein levels in the presence of Ad5IκBα. Gel shift assays of nuclear extracts demonstrated that Ad5IκBα decreased the cytokine-induced DNA binding activity for NFκB. Conclusions. NFκB is an important regulator of cytokine-induced NO expression. These results identify a novel therapeutic approach where gene transfer of the inhibitory molecule IκBα can be used to downregulate cytokine-induced iNOS expression as well as other NFκB-dependent genes that are up-regulated during the inflammatory response. (Surgery 1999;126:142-7.) From the Department of Surgery, University of Pittsburgh, Pittsburgh, Pa
INDUCIBLE NITRIC OXIDE SYNTHASE (iNOS) is overexpressed in a number of pathologic conditions encountered by surgical patients and it is associated with the adverse consequences seen in hemorrhagic and septic shock. The regulatory mechanisms that govern iNOS expression and nitric oxide (NO) production are beginning to be understood and may serve as important therapeutic targets to modulate the induced NO synthesis pathway. Previously, we have shown that cytokines (tumor necrosis factor-α [TNF-α], interleukin 1β [IL-1β], and interferon gamma [IFN-γ]) have a synergistic effect on iNOS expression1 and that up-regulation of iNOS
Supported by National Institutes of Health grant Nos. GM52021 (D. A. G.), GM-37753 (R. L. S.), and GM-44100 (T. R. B.). B. S. T. is the recipient of the Association of Academic Surgery/Davis & Geck Surgical Research Award 1996-1998. Presented at the 60th Annual Meeting of the Society of University Surgeons, New Orleans, La, Feb 11-13, 1999. Reprint requests: Bradley S. Taylor, MD, 677 Scaife Hall, Department of Surgery, University of Pittsburgh, Pittsburgh, PA 15261. Copyright © 1999 by Mosby, Inc. 0039-6060/99/$8.00 + 0 11/6/98740
expression occurs through cytokine activation of the transcription factor NFκB.2-5 This is interesting from a clinical viewpoint because both NO production6 and NFκB DNA-binding activity are markedly elevated in sepsis, which has been shown to be a predictor of mortality.7 By elucidating the mechanisms of signal transduction involved in iNOS expression, specific therapeutic targets may be identified to modulate iNOS expression and subsequent NO production. NFκB is normally localized to the cell cytoplasm and its activation is inhibited by the regulatory molecule IκBα. In the presence of certain cytokines IκBα is phosphorylated at positions Ser 32 and Ser 36, which signals for the dissociation of IκBα from NFκB and the eventual degradation of the IκBα protein.8 The unbound NFκB heterodimer is then free to translocate into the nucleus of the cell. Once in the nucleus, NFκB binds to specific promoter elements within the 5´-flanking region of a number of inflammatory genes, including both the human and rodent iNOS promoter.9,10 This process results in the initiation of transcription and subsequent gene expression of NFκB-dependent genes.
Surgery Volume 126, Number 2
IκBα has been shown to be indirectly linked to NFκB activation in the regulation of iNOS gene expression in a number of studies. Pharmacologic agents such as dexamethasone, dithiocarbamates (pyrrolidine dithiocarbamate [PDTC] and diethyldithiocarbamate [DDTC]), protein-tyrosine phosphatase inhibitors (phenylaresine oxide, pervanadate), and the end product NO itself all decrease iNOS expression by blocking NFκB activation.2-5 Recently Jobin et al11 constructed and reported on an adenoviral vector composed of a mutant form of IκBα where the serine 32 and serine 36 amino acids are replaced by alanine residues (S32A/S36A). The amino acid changes result in a mutant IκB protein that is resistant to phosphorylation and degradation. Their work demonstrates that this vector is capable of suppressing the expression of a number of NFκBdependent genes in intestinal epithelial cells. Because of the availability of the adenoviral IκBα viral vector (Ad5IκB), we sought to develop a gene transfer strategy to directly modulate the iNOS pathway by blocking NFκB-mediated transcriptional activation. Here we report that the overexpression of the inhibitory molecule IκBα decreases iNOS gene expression and subsequent NO production by decreasing cytokineinduced NFκB DNA binding activity in primary rat hepatocytes. MATERIAL AND METHODS Reagents. Murine recombinant TNF-α was purchased from Genzyme (Cambridge, Mass). Human recombinant IL-1β was generously provided by Craig Reynolds, PhD (National Cancer Institute), and murine recombinant IFN-γ was obtained from Gibco BRL (Grand Island, NY). The Ad5IκBα viral vector was kindly provided by David A. Brenner, MD (University of North Carolina). Thyroxine (T4) polynucleotide kinase was purchased from United States Biochemical (Cleveland, Ohio). Gel shift antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, Calif). Cell culture. As previously described, rat hepatocytes were isolated from male 200- to 250-g SpragueDawley rats (Harlan Sprague-Dawley, Madison, Wis) with use of a modified in situ collagenase (type IV, Sigma, St Louis, Mo) perfusion technique.2 After an 18-hour recovery period, cells were infected with various titers of the Ad5IκB vector. After a 1-hour infection period, the media were changed and the cells were allowed to recover for 18 hours. After the recovery period, hepatocytes were treated with a cytokine mixture (CM) of TNF-α (500 U/mL), IL-1β (200 U/mL), and IFN-γ (100 U/mL) in the presence or absence of the Ad5IκBα. Cells were then incubated for the indicated time period before collection.
Taylor et al 143
Fig 1. Schematic of study design. Cultured rat hepatocytes were treated with cytokines in presence or absence of adenoviral IκB vector (Ad5IκB). Western blot analysis was performed to determine IκBα and iNOS protein production. Northern blot analysis and NO release were measured to determine iNOS expression. Electromobility assays were performed to determine NFκB DNA-binding activity.
Nitrite assay. As we have described previously, culture supernatants were collected 24 hours after cytokine treatment and assayed for nitrite (NO2–), the stable end products of NO oxidation, with an automated procedure based on the Greiss reaction.5 Northern blot analysis. Total cellular RNA was collected as previously described.12 Briefly, 20 µg of RNA was electrophoresed on 1% agarose gels. Blots were then transferred to Gene-Screen membranes and were ultraviolet auto-cross-linked. After overnight hybridization, the membranes were probed with a 2.3-kb fragment of human iNOS complementary DNA. After probing for iNOS, membranes were stripped with boiling 5 mmol/L EDTA (ethylenediaminetetra-acetic acid) and 0.1% sodium dodecyl sulfate (SDS) and rehybridized with a probe for 18S ribosomal RNA. Relative messenger RNA (mRNA) levels were quantitated by PhosphorImager scanning with ImageQuant software (Molecular Dynamics, Sunnyvale, Calif). SDS–polyacrylamide gel electrophoresis and Western blotting. Cytosolic proteins were prepared as described13 and quantitated with bicinchonic acid protein assay reagent (Pierce Chemical, Rockford, Ill). Western blot analysis was carried out by separating cytosolic proteins (75 µg) with 8% SDS–polyacrylamide gel electrophoresis (PAGE) and electrophoretically transferring the products to nitrocellulose membranes (Scleicher & Schuell, Keene, NH). Nonspecific binding to the membrane was blocked by 5% nonfat dry milk in phosphatebuffered saline solution (PBS)–Tween overnight at 4°C. Blots were washed in PBS-Tween and then incubated for 1 hour with the polyclonal mouse
144 Taylor et al
Fig 2. Ad5IκB infection results in increased expression of IκBα. Cultured rat hepatocytes were infected with Ad5IκB (MOI 0, 0.1, 1.0, and 10). Western blot analysis for IκB was performed 20 and 34 hours after infection. Ad5IκB gene contains additional DNA segment that codes for hemagglutinin gene to discriminate between endogenous (37 kd) and exogenous IκB (43 kd). CTL, Control. n = 3.
anti-iNOS antibody (1:5000 dilution, Transduction Laboratories, Lexington, Ky) or IκBα antibody (1:1000 dilution, Santa Cruz Biotechnology). The secondary antibody used was a peroxidase-conjugated goat antimouse immunoglobulin G in a 1:1000 dilution (Schleicher & Schuell). Membranes were developed with the ECL detection system (Dupont–New England Nuclear, Boston, Mass) and exposed to film. Electromobility shift assay. Hepatocytes were treated with cytokines for the indicated time period. Preparation of nuclear extracts and electromobility shift assay (EMSA) were carried out as described.14 The NFκB oligonucleotide used was derived from the murine iNOS promoter (positions –90 to –71)15 and contained a functional NFκB element (underlined): CAACTGGGGACTCTCCCTTTG. A mutant NFκB oligonucleotide was used as a mutant competitor: CAACTGGGTCCTCTCCCTTTG. Five micrograms of nuclear extracts were incubated with 0.5 (40,000 counts/min) of phosphorus 32–end-labeled oligionucleotide (T4 polynucleotide kinase) for 30 minutes at room temperature. DNA-protein complexes were electrophoretically resolved on a 5% nondenatured polyacrylamide gel, which was then dried and subjected to autoradiography. Statistical analysis. The significance of differences was determined by analysis of variance (ANOVA) with use of the Statview statistics program (Abacus Concepts, Berkeley, Calif). Statistical significance was established at a P value <.05. RESULTS In this study we used an adenoviral vector that expresses a phosphorylation and degradation resistant form of IκBα to examine the role of NFκB in CM-induced iNOS expression. Fig 1 is a schematic outlining the design of our study. High levels of IκBα are expressed with infection
Surgery August 1999
Fig 3. Ad5IκB infection blocks NFκB DNA-binding activity. Cultured rat hepatocytes were stimulated with combination of cytokines (TNF-α [500 U/mL] plus IFN-γ [100 U/mL] and IL-1β [200 U/mL] plus IFN-γ [100 U/mL]) or cytokine mix (TNF-α [500 U/mL] plus IL-1β [200 U/mL] plus IFN-γ [100 U/mL]), and effect of Ad5IκB (MOI 10) and AdLacz (MOI 10) on NFκB binding was analyzed. Nuclear extracts were collected 30 minutes after stimulation and incubated with 32P-end-labeled oligonucleotide probes containing binding site for NFκB. Resulting complexes were resolved by EMSA. CTL, Control. n = 3.
of Ad5IκB viral vector in rat hepatocytes. To document that infection with the Ad5IκB vector resulted in increased expression of the mutant-form of IκBα, Western blot analysis was performed on cytoplasmic extracts from infected rat hepatocytes 20 or 34 hours after infection. Untreated cells demonstrated basal levels of wild-type IκBα (Fig 2). After Ad5IκB infection with increasing titers (multiplicity of infection [MOI] ranging from 0.1 to 10), there was a marked dose-dependent increase in steady-state Ad5IκB levels. The Ad5IκB gene contains extra DNA nucleotides coding for the hemagglutinin gene (YPYDVPDYA) to discriminate between endogenous and exogenous IκB.11 This accounts for the fact that the mutant adenoviral IκB protein is 43 kd and can be readily discriminated from the 37-kd endogenous IκBα protein. Interestingly, the endogenous IκBα gene is NFκB dependent, and as the amount of Ad5IκB protein levels increased we observed a significant dosedependent decrease in the expression of wild-type IκBα. Infection with an MOI of 10 resulted in complete suppression of wild-type IκBα. Therefore rat hepatocytes are readily infected in vitro with the Ad5IκB vector, and IκBα production increases in a dose- and time-dependent manner. Inhibition of cytokine-induced NFκB DNA-binding activity by Ad5IκB infection. Because the expression of iNOS by cytokines requires the activation of the transcription factor NFκB,4,10 EMSAs
Surgery Volume 126, Number 2
Fig 4. Overexpression of Ad5IκBa results in decreased iNOS mRNA production. Cultured rat hepatocytes were infected with Ad5IκB (MOI 10) or AdLacz in the presence or absence of CM consisting of TNF-α (500 U/mL) plus IL-1β (200 U/mL) plus IFN-γ (100 U/mL) or IL-1β plus IFN-γ (100 U/mL). Northern blot analysis for iNOS mRNA is shown 6 hours after CM treatment. Equal loading of RNA in each lane was confirmed by 18S ribosomal RNA probe (not shown). Ctl, Control. n = 4 per group.
were performed to determine whether the Ad5IκB protein was functional and would block cytokineinduced NFκB activation. In all experiments nuclear extracts from resting hepatocytes showed a basal NFκB complex, and cytokine stimulation (CM: TNF-α plus IL-1β plus IFN-γ or TNF-α plus IFN-γ or IL-1β plus IFN-γ) resulted in a marked increase in NFκB DNA-binding activity (Fig 3). In all groups the addition of Ad5IκB blocked the formation of the NFκB complex in a concentrationdependent manner regardless of whether the stimulus was a double- or triple-cytokine combination. The addition of the negative control AdLacZ had no effect on NFκB DNA binding activity. Thus the mutant inhibitor IκBα blocked the cytokineinduced NFκB DNA-binding activity in rat hepatocytes, demonstrating that in vitro infection with the viral vector expressing IκBα results in a loss of NFκB DNA activation and subsequent nuclear translocation. iNOS mRNA levels are decreased by IκBα gene transfer. To determine whether overexpression of IκBα had an effect on cytokine-induced iNOS mRNA production, rat hepatocytes were infected with the Ad5IκB vector and Northern blot analysis was performed for iNOS mRNA 6 hours after cytokine treatment. CM treatment resulted in a significant increase in iNOS mRNA levels compared with control cells, whereas the addition of Ad5IκB resulted in a marked decrease in iNOS mRNA expression (Fig 4). Infection with Ad5LacZ had no effect either alone or on cytokine-induced iNOS mRNA expression. These data reveal that iNOS mRNA expression is depen-
Taylor et al 145
Fig 5. Suppression of iNOS protein levels by in vitro AdIkB expression. Cultured rat hepatocytes were exposed to Ad5IkB or AdLacZ with CM of TNF-α (500 U/mL) plus IL-1β (200 U/mL) plus IFN-γ (100 U/mL). Western blot analysis was performed for iNOS 8 hours after cytokine stimulation. Lipopolysaccharide-stimulated RAW 264.7 cells served as positive control. CTL, Control. n = 3.
dent on NFκB activation and that IκBα gene transfer is an effective method to down-regulate iNOS expression. Infection with IκBα decreases cytokine-induced iNOS protein levels. To investigate the effects of Ad5IκB infection on CM-induced iNOS protein production, we performed Western blot analysis for iNOS protein levels. These studies demonstrated that infection with Ad5IκBα resulted in a decrease in iNOS protein levels (Fig 5). At 8 hours after treatment, Western blots showed that CMstimulated iNOS protein expression was markedly increased. In addition, double-cytokine combinations of IL-1β plus IFN-γ and TNF-α plus IFN-γ resulted in an increase in steady-state iNOS protein production. Ad5IκB (MOI 10) infection in the presence of cytokines resulted in complete abrogation of iNOS protein accumulation. These data indicate that cytokine-induced iNOS protein expression in hepatocytes is inhibited by Ad5IκB and suggests that NFκB is essential for maximal iNOS protein expression. Infection with IκBα decreases cytokineinduced NO release. To determine whether Ad5IκB had an effect on cytokine-induced NO release, rat hepatocytes were infected with the Ad5IκB vector. After infection and an 8-hour recovery period, hepatocytes were exposed to cytokines. CM stimulation alone resulted in a marked increase in hepatocellular nitrite (NO2–) production, whereas NO2– levels were decreased in a concentration-dependent fashion in those groups infected with the IκBα gene (Fig 6). Infection with AdLacZ did not result in any decrease in CM-induced NO production, demonstrating that the adenoviral vector itself was not responsible for the decrease in NO production.
146 Taylor et al
Fig 6. Overexpression of IκBα inhibits NO release. Cultured rat hepatocytes were infected with AdIκBα (MOI 1 and 10) or AdLacZ with CM of TNF-α (500 U/mL) plus IL-1β (200 U/mL) plus IFN-γ (100 U/mL). Graph, Nitrite (NO2–) production. Culture supernatants were collected 24 hours after treatment and NO2– release was quantitated. n = 6 per group. Asterisk, P < .01 versus control (CTL); two asterisks, P < .01 versus CM (by ANOVA).
DISCUSSION Because the expression of iNOS has such profound effects in sepsis, hemorrhagic shock, and transplant biology, we believe it is important to define the signal transduction mechanisms required for its expression. We and others have previously shown that the transcription factor NFκB is required for iNOS gene expression. The purpose of this study was to determine the utility of adenoviral gene transfer of the IκBα gene. In this study we demonstrated that cultured hepatocytes are readily infected with the Ad5IκB gene and that the IκB protein is produced and is functionally capable of inhibiting NFκB DNA binding. In addition, we demonstrated that a gene transfer strategy can be used to inhibit iNOS mRNA, protein, and NO production. The data in this study support the hypothesis that NFκB is required for iNOS gene transcription and that adenoviral gene transfer is a potential tool to block iNOS gene expression. NFκB appears to be a key regulatory transcription factor and represents a potential master switch that governs gene expression in the inflammatory response. NFκB nuclear translocation and DNA binding to the 5´-flanking promoter region is required for the gene expression of cell adhesion molecules (endothelial cell adhesion molecule-1, vascular cell adhesion molecule-1, intercellular adhesion molecule-1) and cytokines (interleukin 1, interleukin 6, interleukin 8) as well as IκBα and iNOS.4 Previously, our group has shown that iNOS expression can be suppressed pharmacologically by blocking NFκB activation. Dexamethasone, PDTC, protein-tyrosine phosphatase inhibitors, and NO
Surgery August 1999
preserve IκB in the cytoplasm, thereby blocking NFκB activation.2-5 The results in this study illustrate the role of IκB in regulating iNOS expression and represent direct evidence that NF-κB is an essential transcription factor that is required for iNOS expression. The signal transduction pathway leading to NFκB transcriptional activation has become a target for a number of therapeutic interventions in an effort to down-regulate the NFκB-responsive genes. Because pharmacologic approaches that target mechanisms that block IκB phosphorylation and degradation have significant side-effect profiles and may be cytotoxic in high concentrations, a number of gene transfer strategies that use antisense oligonucleotides16 or gene transfer of wild-type IκB have been used.17,18 However, overexpression of wild-type IκB result in only a partial block of NFκB responsive genes because the wild type is not resistant to degradation. The Ad5IκB vector used in this study has been shown to produce an inhibitory protein, which is not degraded and therefore has greater utility in trials of gene therapy. The appropriate use of this vector in different clinical scenarios remains to be determined and in vivo experiments need to be performed. Similar to NO, NFκB appears to have both beneficial and detrimental effects in the liver.19-25 A number of studies have demonstrated the importance of hepatic NFκB in models of liver injury. NFκB has been shown to be antiapoptotic19-21 and important in liver regeneration,22 whereas others have demonstrated that it may be involved in the fibrotic response seen in cirrhosis.23 The findings in these studies suggest that a product of an NFκB dependent gene must be important for successful tissue regeneration and to prevent apoptosis. Recent evidence suggests that the gene product iNOS and its end product NO may be the factor responsible for this antiapoptotic effect, seen in the aforementioned studies, that block the NFκB pathway.24,25 Data from Saavedra et al24 demonstrated that NO-stimulated cyclic guanosine monophosphate production inhibits TNF-α–induced apoptosis in hepatocytes. Additional in vivo evidence to support the antiapoptotic hepatoprotective role of NO uses the iNOS-null mice. In this study the hepatocyte proliferative response to partial hepatectomy was severely inhibited in the transgenic mice lacking the iNOS gene. Partial liver resection led to increased caspase 3 activity, hepatocellular death, and liver failure despite NFκB, interleukin 6, and TNFα induction, suggesting that NO protects hepatocytes from cytokine-mediated death.25 In conclusion, this study demonstrates that the
Surgery Volume 126, Number 2
transcription factor NFκB is critical for cytokineinduced iNOS gene expression and that its activation is dependent on the phosphorylation and degradation of IκBα. These results suggest that adenoviral gene delivery of the IκBα gene is an attractive therapeutic strategy to down-regulate cytokine-induced iNOS expression and also is likely to be applicable to other NFκB-dependent genes, which are up-regulated during inflammation. We thank Melina Kibbe and Debra Williams for excellent technical assistance. We also thank David A. Brenner (UNC) for providing the Ad5IκB vector. REFERENCES 1. Geller DA, Nussler AK, Di Silvio M, Lowenstein CL, Shapiro RA, Wang SC, et al. A central role for interleukin-1b in the in vitro and in vivo regulation of hepatic inducible nitric oxide synthase: Il-1b induces hepatic nitric oxide synthesis. J Immunol 1995;155:4890-8. 2. Taylor BS, Kim YM, Wang Q, Shapiro RA, Billiar TR, Geller DA. Nitric oxide down-regulates hepatocyte-inducible nitric oxide synthase gene expression. Arch Surg 1997;132:1177-83. 3. de Vera ME, Taylor BS, Wang Q, Shapiro RA, Billiar TR, Geller DA. Dexamethasone suppresses inducible nitric oxide synthase gene expression by upregulating I-κBα and inhibiting NF-κB. Am J Physiol 1997;273:G1290-6. 4. Taylor BS, de Vera ME, Ganster RW, Wang Q, Shapiro RA, Morris SM, et al. Multiple NFκB enhancer elements regulate cytokine induction of the human inducible nitric oxide synthase gene. J Biol Chem 1998;273:15148-56. 5. Taylor BS, Liu S, Villavicencio RT, Ganster RW, Geller DA. The role protein of phosphatases in the expression of inducible nitric oxide synthase in the rat hepatocyte. Hepatology 1999;29:1199-1207. 6. Ochoa JB, Udekwu AO, Billiar TR, Curran RD, Cerra FB, Simmons RL, et al. Nitrogen oxide levels in patients after trauma and during sepsis. Ann Surg 1991;214:621-6. 7. Bohrer H, Qiu F, Zimmerman T. Role of NF-κB in the mortality of sepsis. J Clin Invest 1997;5:972-85. 8. Brown K, Gerstberger S, Carlson L, Granzoso G, Sienblest U. Control of IkappaB-alpha proteolysis by site-specific, signal-induced phosphorylation. Science 1995;267:1485-8. 9. de Vera ME, Shapiro RA, Nussler AK, Mudgett JS, Simmons RL, Morris SM, et al. Transcriptional regulation of human inducible nitric oxide synthase (NOS2) gene by cytokines: initial analysis of the human NOS2 promoter. Proc Natl Acad Sci U S A 196;93:1054-9. 10. Xie QW, Kashiwabara Y, Nathan C. Role of transcription factor NF-κB/Rel in induction of nitric oxide synthase. J Biol Chem 1994;269:4705-8. 11. Jobin C, Panja A, Hellerbrand C, Limuro Y, Didonato J, Brenner DA, et al. Inhibition of proinflammatory molecule production by adenovirus-mediated expression of a nuclear factor κB super-repressor in human intestinal epithelial cells. J Immunol 1998;160:410-8.
Taylor et al 147
12. Geller DA, Nussler AK, Di Silvio M, Lowenstein CJ, Shapiro RA, Wang SC, et al. Cytokines, endotoxin, and glucocorticoids regulate the expression of inducible nitric oxide synthase in hepatocytes. Proc Natl Acad Sci U S A 1993;90:522-6. 13. Geller DA, Lowenstein CJ, Shapiro RA, Nussler AK, Di Silvio M, Wang SC, et al. Molecular cloning and expression of inducible nitric oxide synthase from human hepatocytes. Proc Natl Acad Sci U S A 1993;90:3491-5. 14. de Vera ME, Kim YM, Wong HR, Billiar TR, Geller DA. Heat shock response inhibits cytokine-inducible nitric oxide synthase expression in rat hepatocytes. Hepatology 1996;24: 1238-45. 15. Xie Q, Whisnan R, Nathan C. Promoter of the mouse gene encoding calcium-independent nitric oxide synthase confers inducibility by interferonγ and bacterial lipopolysaccharide. J Exp Med 1993;177:1779-84. 16. Goodman DJ, von Albertini MA, McShea A, Wrighton CJ, Bach FH. Adenoviral-mediated overexpression of I kappa B alpha in endothelial cells inhibits natural killer cell-mediated endothelial cell activation. Transplantation 1996;62:967-72. 17. Wrighton CJ, Hofer-Warbinek R, Moll T, Eytner R, Bach FH, de Martin R. Inhibition of endothelial cell activation by adenovirus-mediated expression of I kappa B alpha, an inhibitor of the transcription factor NF kappa B. J Exp Med 1996;183:1013-22. 18. Neurath MF, Petterson S, Meyer zum Bushenfelde KH, Strober W. Local administration of antisense phosphorothioate oligonucleotides to the p65 subunit of NF kappa B abrogates established experimental colitis in mice. Nat Med 1996;2:998-1004. 19. Beg AA, Sha WC, Bronson RT, Ghosh S, Baltimore D. Embryonic lethality and liver degeneration in mice lacking the RelA component of NF kappa B. Nature 1995;375:167-70. 20. Van Antwerp DJ, Mayo MW, Baldwin AS Jr. TNF- and cancer therapy-induced apoptosis: potentiation by inhibition of NF-kappa B. Science 1996;274:784-7. 21. Arsuro M, Fitzgerald MJ, Fausto N, Sonenshein GE. Nuclear factor-kappaB/Rel blocks transforming growth factor beta 1-induced apoptosis of murine hepatocyte cell lines. Cell Growth Differ 1997;8:1049-59. 22. Cressman DE, Greenbaum LE, Haber BA, Taub R. Rapid activation of post-hepatectomy factor/nuclear factor kappa B in hepatocytes, a primary response to regenerating liver. J Biol Chem 1994;269:30429-35. 23. Hellerbrand C, Jobin C, Iimuro Y, Licato L, Sartor RB, Brenner DA. Inhibition of NF-κB in activated rat hepatic stellate cells by proteosome inhibitors and a IκB superrepressor. Hepatology 1998;27:1285-95. 24. Saavedra JE, Billiar TR, Williams DL, Kim YL, Watkins SC, Keefer LK. Targeting nitric oxide (NO) delivery in vivo: design of a liver-selective NO donor prodrug that blocks tumor necrosis factor-α–induced apoptosis and toxicity in the liver. J Med Chem 1997;40:1947-54. 25. Rai RM, Lee FYJ, Rosen A, Yang SQ, Lin HZ, Koteish A, et al. Impaired liver regeneration in inducible nitric oxide synthase-deficient mice. Proc Natl Acad Sci U S A 1998;95: 13829-34.