Inhibition of NFκB enhances the susceptibility of cancer to TRAIL-mediated apoptosis and is suggested as a strategy for cancer therapy. Because the role of NFκB in TRAIL-mediated apoptosis of hepatocytes is unknown, we investigated the influence of NFκB-inhibition in death ligand-mediated apoptosis in hepatitis. Adenoviral hepatitis resulted in upregulation of NFκB-activity, which could be inhibited by expression of IκBα-superrepressor. We treated mice after the onset of adenoviral hepatitis with adenoviruses expressing FasL (AdFasL), TRAIL (AdTRAIL), or GFP (AdGFP). In contrast to apoptosis induced by AdFasL, NFκB inhibition strongly enhanced AdTRAIL-mediated apoptosis of hepatocytes. Expression of IκBα inhibits adenoviral infection-mediated overexpression of bcl-xl, providing a molecular mechanism for TRAIL sensitization. In agreement with this hypothesis, downregulation of bcl-xl by siRNA enhanced susceptibility of hepatocytes to TRAIL, but not to FasL-mediated apoptosis, resulting in TRAIL-mediated severe liver damage after AdTRAIL application. Our data demonstrate that inhibition of NFκB in adenoviral hepatitis strongly sensitizes hepatocytes to TRAIL-mediated apoptosis. Bcl-xl, in contrast to bcl-2 and c-FLIP, is strongly upregulated after viral infection and represents an essential NFκB-dependent survival factor against TRAIL-mediated apoptosis. In conclusion, inhibition of NFκB or bcl-xl during TRAIL therapy may harbor a risk of liver damage in patients with viral hepatitis. (HEPATOLOGY 2005;41:280–288.)
Hepatotropic viruses induce liver injury by apoptotic cell death. The molecular machinery of apoptosis is triggered by signals released from the cytoplasm or from the cell membrane, leading to the activation of caspase cascades, which execute apoptotic cell death. In viral hepatitis, apoptosis of hepatocytes is mediated by engagement of death receptors belonging to the tumor necrosis factor (TNF) receptor gene superfamily. The mechanisms of TNFα- and FasL-mediated apoptosis in liver damage have been studied in humans and in several animal models.1–6
Recently, involvement of TRAIL-mediated apoptosis has been shown in hepatitis B and C7, 8 and in cholestatic liver disease.9, 10 The critical role of TRAIL in liver diseases has been confirmed in mouse models of adenoviral hepatitis8 and in Con-A-mediated liver injury.11 Whereas FasL provides a strong constitutional death signal in hepatocytes, TRAIL is unable to induce apoptosis in healthy hepatocytes in vivo. TRAIL needs triggering through viral infection or bile acids to activate the caspase cascade.
The molecular mechanisms of apoptosis by FasL and TRAIL show similar patterns. Fas and TRAIL-receptors 1 (DR4) and 2 (DR5) recruit the key adapter protein FADD to the cell membrane, thereby inducing the caspase cascade. In contrast to Fas, however, both TRAIL receptors also bind the adapter molecule TRADD,12, 13 which explains the more potent activation of NFκB by TRAIL compared with FasL.14
NFκB is a transcription factor that is activated by a large variety of bacteria and viruses and controls expression of many proteins involved in inflammation. In addition, NFκB is upregulated in other stress conditions, such as treatment with chemotherapy or tissue ischemia, indicating a more general participation of NFκB in stress responses and apoptosis independent of the immune system. However, the role of NFκB in apoptosis of hepatocytes appears to be ambivalent. NFκB has been described to counteract apoptosis and to be an essential strong survival factor in embryonal liver development and during TNF-mediated liver regeneration.15–17 Conversely, evidence exists that NFκB activity boosts apoptosis of hepatocytes in the early time course of acute viral hepatitis by transcriptional activation of the Fas receptor.18
Our study investigated the role of NFκB in FasL- and TRAIL-mediated apoptosis of hepatocytes in vivo to elucidate different roles of Fas and TRAIL in viral hepatitis. We showed that inhibition of NFκB activity strongly enhances the susceptibility of hepatocytes to TRAIL- but not to FasL-mediated apoptosis in hepatitis. We observed a strong upregulation of bcl-xL expression in the liver after adenoviral infection, which was strongly inhibited by IκBα. We showed that siRNA-mediated inhibition of bcl-xl expression strongly sensitizes the mouse liver to TRAIL-mediated apoptosis in viral hepatitis. These data suggest that NFκB-dependent bcl-xl upregulation confers resistance to TRAIL-mediated apoptosis in hepatocytes in viral hepatitis.
Cell Lines, Plasmids, and Detection of Apoptosis in Cell Culture.
The human hepatoma cell lines HepG2, Huh7, and the human embryonal kidney cell line 293 were obtained from the American Type Culture Collection. The cells were maintained in growth medium (Dulbecco modified Eagle medium/Glutamax, Gibco BRL, Gaithersburg, MD), supplemented with 10% heat-inactivated fetal bovine serum (Gibco BRL), 100 units/mL penicillin, and 100 μg streptomycin at 37°C in 5%. CO2. Adenoviral vectors AdGFP, AdTRAIL, and AdFasL were generated as described previously.8 The adenoviral vectors AdLacZ and AdIκBα were provided by Dr. D. Brenner, Chapel Hill, NC.
For detection of apoptosis in hepatoma cells after adenoviral application, cells were harvested 12 hours after infection for photometric histone enzyme-linked immunoassay (ELISA). Histone-ELISA was performed with the Cell Death Detection ELISA Plus® kit (Roche, Basel, Switzerland), according to the manufacturer's instructions.
Adenovirus Preparation and In Vivo Infection Experiments.
To generate high titer viral stocks, 2 × 108 293 packaging cells at 90% confluence were infected at a multiplicity of infection of 5 to 10 plaque-forming units (pfu) per cell. Adenovirus preparation and viral titering were performed as described previously.8, 18 Before infection, the virus was dialyzed twice against a solution containing 10 mmol/L Tris-HCl, pH 8,0, 1 mmol/L MgCl2, 140 mmol/L NaCl at 4°C. Infection of the mice was carried out by administration of 0.25 mL virus solution into the tail vein at total virus loads as indicated in the figures. Virus preparations were stored at −20°C in 25% glycerol, 10 m mmol/L Tris/HCl, pH 7,4, and 1 mmol/L MgCl2.
Animal Experiments and Preparation of Nuclear and Whole Cell Liver Extracts.
Pathogen-free balb\c mice (aged 4 to 8 weeks) or NMRI-nu/nu mice were obtained from the Animal Research Institute of the Medizinische Hochschule Hannover. All experiments were performed in agreement with the German legal requirements.
At the times indicated in the figure legends, mice were killed, and the livers were harvested for the preparation of nuclear extracts, whole cell extracts, total RNA, and cryosections, respectively. Whole cell liver extracts were prepared by homogenizing freshly harvested liver tissue in RIPA buffer (1% Nonidet P-40, 0.5% sodiumdeoxycholate, 0.1% SDS in phosphate-buffered saline [PBS], before use, 5 μL protease inhibitor cocktail (Sigma Chemicals, St. Louis, MO) was added per milliliter). The tissue was homogenized on ice with a few strokes in a Elvehjem-homogenizer and subsequently centrifuged at 12,000g for 10 minutes (4°C). The supernatant was stored at −80°C. Nuclear extracts were isolated from mouse liver as described by Lichtsteiner et al.19
Isolation of Total RNA and Semiquantitative Reverse Transcription Polymerase Chain Reaction Analysis.
Total RNA was isolated from liver tissue using the peqGOLD RNAPure™ kit (PeqLab, Erlangen, Germany) according to the manufacturer's instructions. First, strand synthesis for reverse transcription polymerase chain reaction (RT-PCR) was performed byusing Oligo(dT)15 primer (Promega, Madison, WI) and the Omniscript RT kit (Qiagen) in combination with 1 μg total RNA per reaction. Amplification of bcl-xl and GAPDH cDNA were carried out in one reaction using the specific primers (bcl-xl: sense:5′-cgacccagccaccacctcctc-3′ and anti-sense: 5′-tggggcctcagtcctattctc-3′, cDNA length 316bp; GAPDH: sense: 5′-tgatgacatcaagaaggtggtgaa g–3′ and anti-sense: 5′-tccttggaggccatgtaggccat-3′, cDNA length 249bp) and 1 μL of the RT reaction. Briefly, PCR was performed for 10 cycles with first strand-DNA and bcl-xl primers at a Tanneal of 58°C. Then the reaction was interrupted, primers for GAPDH were added, and the reaction was continued for another 25 cycles. PCR products were documented and quantitated by using a Gel Doc 1000 apparatus (Bio-Rad, Hercules, CA) and Molecular Analyst software (Bio-Rad).
Assessment of Adenoviral Liver Infection by Semiquantitative Duplex-PCR of Ad5 Fiber and Caspase 8 Gene Locus.
Livers from infected mice were harvested 6 hours after infection. Fifty-milligram portions of liver tissue were digested in a mixture of 750 μL solution A (50 mmol/L Tris, pH 8.0, 100 mmol/L ethylenediaminetetraacetic acid, 100 mmol/L NaCl, 1% SDS), and 50 μL Proteinase K (22 mg/mL, Roche Diagnostics) by overnight shaking at 56°C. Cell debris was pelleted in a tabletop centrifuge, and DNA in the supernatant was isolated by conventional isopropanol precipitation. The DNA was resolved in TE 8.0/RNAse by shaking for 2 hours at 37°C, and all DNA preparations from one liver lobe were pooled. Adenoviral infection was determined by using the primer pair fiber-fw (5′-CCCAAAATGTAACCACTGTGAGC-3′) and fiber-rev (5′-GTGTTTAGGTCGTCTGTTACATGC-3′) directed against the coding sequence of the fiber protein of adenovirus serotype 5, resulting in a 417-bp fragment. Mouse genome templates as control were determined by using the primers Intr2-s (5′-CCAGATGTATCTCTGTGCGTTTGC-3′) and Exon3-as (5′-GTTATTTTCTGCCAGCATGGTCC-3′), directed against the murine locus of Caspase-8 (intron 2/exon 3), resulting in a 520-bp fragment. PCR was performed by using the HotStarTaq Master Mix (Qiagen), 50 ng DNA, and murine genomic primers at a concentration of 400 nmol/L each. The PCR reaction was started as follows: 15′—95°C; [30”—94°C; 30”—57°C; 1'—72°C] × 5. The reaction was interrupted, fiber primers were added (same concentration), and a further 27 cycles were performed. PCR products were resolved on a 1.5% agarose gel, and mouse/adenovirus fragment ratio was visually evaluated.
Assessment of Liver Injury by TUNEL Assay, Haemalaun/Eosin Staining, Histone-ELISA, and Measurement of Serum Alanine Aminotransferase Activity.
Liver tissue of infected mice was embedded in OCT compound (Sakura, Netherlands) and shock frozen in liquid nitrogen. The samples were stored at −80°C. Seven-micron sections were prepared and fixed in PBS-buffered paraformaldehyde solution (4%) for 30 minutes at room temperature. After washing with PBS, endogenous peroxidase activity was blocked by covering the sections with 0.3% H2O2 in methanol for 30 minutes at room temperature. The sections were rinsed with PBS and incubated with 0.1 % Triton X-100 in 0.1% Na citrate for 30 minutes at room temperature. The sections were again washed with PBS and TUNEL-stained with the In Situ Cell Death Detection Kit (Roche Diagnostics), following the instructions of the manufacturer's protocol. The number of TUNEL-positive cells was assessed for each liver by counting TUNEL-positive cells in three sections. For each section, three fields (magnification ×200) were examined.
For hematoxylin-eosin (HE) staining, sections were incubated 10 seconds in Haemalaun solution (Sigma Chemicals), washed in warm water, and then stained 2 minutes in Eosin solution (Sigma Chemicals). After washing, the sections were incubated subsequently in 70%, 96%, and 100% ethanol.
A cell death detection kit (Boehringer Mannheim, Mannheim, Germany) was used according to the manufacturer's protocol for the qualitative and quantitative determination of cytoplasmic histone-associated DNA fragments in liver after adenoviral application by measuring of fluorescence. Liver tissue (50 mg) and 450 μL TBE buffer were homogenized and then centrifuged. The supernatants were diluted in incubation buffer (1:10). Ten microliters of this solution and 90 &μL Immunomix (supplied with the kit) were transferred into streptavidine-coated microtiter plates. After incubation for 2 hours at room temperature, we started evaluation according to the manufacturer's instructions.
Activity of liver-specific alanine aminotransferase was determined by an automated enzyme assay.
Electrophoretic Mobility Shift Assays.
For electrophoretic mobility shift assay experiments, nuclear extracts from liver tissue were used. Electrophoretic mobility shift assay experiments were performed by using nuclear extracts as indicated and 1 ng end-labeled DNA. The NFκB site ′5-TAG-TTG-AGG-GGA-CTT-TCC-CAG-GCA-3′ was used as the cognate DNA binding sequence for NFκB in electrophoretic mobility shift assays. Supershift experiments were performed with antibodies directed against p50 and p65 (NFκB p65 (A) sc-109 and NFκB p50 (NLS) sc-114x, Santa Cruz Biotechnology, Santa Cruz, CA).
SDS-Polyacrylamide-Gel Electrophoresis and Western Blot Analysis.
Protein concentrations of nuclear extracts were measured by Bio-Rad Microassay (Bio-Rad, Munich, Germany). Five to twenty micrograms whole cell extract were separated on a 10% SDS-polyacrylamide gel and blotted onto Hybond N membrane (Millipore, Frankfurt, Germany). As primary antibodies we used anti-Bcl-xs/l (sc634, Santa Cruz Biotechnology), anti-Bcl-2 (sc7382, Santa Cruz Biotechnology), and anti-FLIP (AF821, R&D Systems, Minneapolis, MN). The antigen–antibody complexes were visualized by using the ECL detection system as recommended by the manufacturer (Amersham, Buckinghamshire, United Kingdom).
SiRNA Duplexes, Transfection of siRNA, and Hydrodynamic Tail Vein Delivery of siRNA.
21-Nucleotide RNA with 3′-dTdT overhangs was synthesized by Dharmacon Research Inc. (Lafayette, CO) in the “ready to use” option C. AA-N19 mRNA Targets 5′-3′: Bcl-xl target sequence 1: AAA GGA UAC AGC UGG AGU CAG (human and mouse bcl-xl mRNA); Bcl-xl target sequence 2: AAC CGG GAG CUG GUG GUC GAC (mouse mRNA, one mismatch to human bcl-xl mRNA); Bcl-xl target sequence 3: AAG GAU ACA GCU GGA GUC AGU (human and mouse bcl-xl mRNA); Bcl-xl target sequence 4: AAC UGG GGU CGC AUC GUG GCC (mouse mRNA, one mismatch to human bcl-xl mRNA); As scrambled sequences we used AA-N19 mRNA Targets 5′-3′: AAU UUA ACC GCC AGU CAG GCU or 5′-3′: AAG CAA AAC ACC AGC AGC AGU. Transfection of siRNA duplexes was performed using Oligofectamine (Invitrogen, Carlsbad, CA) and Opti-MEM medium (Invitrogen) according to the manufacturer's recommendations. Huh7 cells grown to a confluence of 40% to 50% in 24-well plates were transfected with 60 pmol siRNA duplex per well.
In vivo delivery of siRNA duplexes was performed via the tail vein by high-volume hydrodynamic injection, using 1 nmol siRNA duplexes per gram body weight. For tail vein injection, siRNA was applied in a total volume of 2.5 mL (0.5 mL dH20 and 2.0 mL NaCl 0.9%). Injection time was 4 to 10 seconds.
Each experiment was repeated 3 times with three animals in each group. The Student t test was used to compare differences between groups. A difference with a P value equal to or less than .05 was considered significant.
Inhibition of NFκB Activity in Viral Hepatitis Strongly Enhances Sensitivity of Hepatocytes to TRAIL-, But Not to FasL-Mediated, Apoptosis.
The transcription factor NFκB is activated in many cell types through viral infection and thus may modulate FasL- or TRAIL-mediated apoptosis of hepatocytes in hepatitis. To explore the role of NFκB activity in virally infected livers in vivo, we used a mouse model of adenoviral hepatitis. Infection of the liver with adenovirus (AdLacZ or AdiκBα) was performed by administration of different amounts of infectious viral particles (pfu/g) into the tail vein of Balb/c mice. Activation of NFκB DNA binding in nuclear extracts of hepatocytes was evident at a broad range of viral titers (Fig. 1 A) and occurred rapidly after adenoviral transduction (Fig. 1B). Using an adenoviral vector expressing the IκBα-superrepressor activation of NFκB after viral infection of the liver could be inhibited significantly (Fig. 1C).
Comparable infection of mice with different adenoviral vectors was achieved by administration of the same particle load regarding the infectious viral titer. Equal infection of the livers by AdLacz and AdIκBα was confirmed by a semiquantitative duplex-PCR, comparing the ratio of viral genome (Ad5 fiber) versus mouse genome as internal control, using total DNA from infected mice as a template (Fig. 1D).
To investigate the consequences of NFκB inhibition on FasL- and TRAIL-mediated apoptosis, balb/c mice were pretreated with indicated doses of AdIκBα or AdLacZ to induce viral hepatitis. Subsequently, mice were infected with adenoviral vectors expressing the death ligands TRAIL or FasL. Inhibition of NFκB by IκBα strongly sensitizes hepatocytes to TRAIL-mediated apoptosis, as shown in Fig. 2. In contrast, inhibition of NFκB has no significant impact on the sensitivity of hepatocytes toward FasL. FasL-mediated apoptosis in mice pretreated with AdIκBα appears to be similar compared with that in mice treated with the control vector AdLacZ (Fig. 2). The same results were observed in nude mice, which indicates that adenoviral expression of the death ligands rather than potential death-ligands expressed by immune cells are responsible for the results (data not shown).
Bcl-xl Expression in Hepatocytes After Viral Infection Is Dependent on NFκB Activity and Mediates Resistance to TRAIL-Mediated Apoptosis.
NFκB has been implicated in the transcriptional control of several negative regulators of apoptosis such as cFLIP, bcl-2, and bcl-xl. To evaluate the molecular mechanisms of cellular resistance to TRAIL-mediated apoptosis in viral hepatitis, livers infected with AdLacZ or AdIκBα were analyzed for expression of cFLIP, bcl-2, and bcl-xl. No significant upregulation of cFLIP protein expression was observed in viral hepatitis. However, NFκB-inhibition resulted in slightly lower expression of c-FLIP 4 hours after adenoviral infection, but the level of c-FLIP expression remained constant during the further time course of hepatitis (Fig. 3 A). In contrast to c-FLIP, bcl-2 expression was upregulated after adenoviral infection. However, inhibition of NFκB resulted only in moderately delayed expression pattern of bcl-2 during the early time course of adenoviral hepatitis without affecting the maximal expression level (see Fig. 3A).
Compared with cFLIP and bcl-2, expression of bcl-xl was significantly upregulated more strongly in viral hepatitis (see Fig. 3B), suggesting a peculiar role of NFκB-dependent regulation of bcl-xl in viral hepatitis. Inhibition of NFκB effectively inhibited bcl-xl overexpression in viral hepatitis, demonstrating that regulation of bcl-xl in viral infection is highly dependent on NFκB activity (see Fig. 3B).
To investigate the role of bcl-xl in TRAIL-mediated apoptosis, we inhibited gene expression by RNA interference. TRAIL- and FasL-resistant Huh7 hepatoma cells were transfected with siRNA duplexes against bcl-xl or with siRNA-scrambled. Interestingly, treatment with siRNA-bcl-xl–sensitized Huh7 hepatoma cells to TRAIL- but not to FasL-mediated apoptosis (Fig. 4).
Because siRNA is effectively delivered into approximately 70% to 80 % of hepatocytes by hydrodynamic tail vein injection, RNA interference can also be used for silencing of liver gene expression in vivo.20–23 To investigate the contribution of bcl-xl on TRAIL-apoptosis in vivo, we inhibited bcl-xl expression in the mouse liver by hydrodynamic injection of siRNA-bcl-xl, as shown in Fig. 5 A. After silencing of bcl-xl by RNAi, viral hepatitis was triggered by application of AdlacZ. SiRNA-bcl-xl significantly inhibited bcl-xl expression in viral hepatitis and strongly sensitized the hepatocytes to TRAIL-, but not to FasL-mediated apoptosis (see Fig. 5). Application of AdTRAIL into siRNA-bcl-xl pretreated mice suffering from viral hepatitis led to severe liver damage as shown by TUNEL assay (Fig. 5C), HE-stained liver sections (Fig. 5D), histone-ELISA (Fig. 5E), and the strong release of liver transaminases into the serum (Fig. 5F). Compared with the siRNA-scrambled control, hydrodynamic injection of siRNA-bcl-xl also slightly enhanced liver cell apoptosis in experiments with high-dose (1 × 109 pfu/g) AdlacZ/AdGFP, which may be explained by the fact that high-dose adenoviral infection can trigger TRAIL expression in the liver.8
Our results demonstrate that downregulation of bcl-xl in viral hepatitis alone strongly sensitizes hepatocytes to TRAIL-mediated apoptosis, suggesting that bcl-xl is an important NFκB-dependent protection factor in TRAIL-mediated apoptosis in the liver.
NFκB is activated by viral infections as a part of the cellular defense mechanism and is important in modulating cell survival and cell death after cellular damage. NFκB protects cells from apoptosis by transcriptional activation of survival factors, such as c-FLIP, XIAP, c-IAP1, c-IAP2, Bfl-1/A1, Bcl-2, and Bcl-xL.24–30 NFκB is also capable of transactivating death receptors such as Fas or TRAIL-DR5,18, 31, 32 providing a molecular explanation for its proapoptotic functions. The ability to induce both—death receptors or antiapoptotic mediators—explains the dual role of NFκB as a mediator or inhibitor of cell death, whereby the modulation of apoptosis by NFκB appears to be largely determined by the nature of the death stimulus.
We investigated the role of NFκB in FasL- and TRAIL-mediated apoptosis of hepatocytes in viral hepatitis. Whereas inhibition of NFκB activity in viral hepatitis resulted in strongly enhanced sensitivity of hepatocytes to TRAIL-mediated apoptosis, no sensitization to FasL-mediated apoptosis was observed.
In some cell lines, TRAIL-mediated apoptosis is inhibited by overexpression of bcl-2 and bcl-xl33, 34; recently, Higuchi et al.10 showed involvement of c-FLIP in bile acid–induced TRAIL-mediated apoptosis of hepatoma cells.
In adenoviral hepatitis, expression of c-FLIP remained unchanged, whereas expression of bcl-2 was moderately affected and expression of bcl-xl was strongly upregulated by viral infection. However, inhibition of NFκB activity by IκBα led only to little delayed upregulation of bcl-2, but largely inhibited upregulation of bcl-xl in viral infection.
To investigate the hypothesis that bcl-xl is a crucial NFκB-dependent survival factor in TRAIL-mediated apoptosis in the liver, we used RNA interference to silence bcl-xl gene expression in vitro and in vivo. Inhibition of bcl-xl gene expression resulted in enhanced susceptibility of FasL- and TRAIL-resistant hepatoma cells to TRAIL-, but not to FasL-mediated apoptosis, emphasizing the predominant role of bcl-xl in TRAIL-mediated apoptosis. In agreement with the results obtained in vitro, in vivo down-regulation of bcl-xl sensitizes hepatocytes to TRAIL-, but not to FasL-mediated apoptosis.
Recently several studies suggested that inhibition of NFκB activity and bcl-xl expression may be used as an attractive therapeutic strategy to sensitize resistant tumor cells to TRAIL-induced apoptosis.32, 35, 36 However, our results demonstrate a potential limitation of this strategy, because inhibition of NFκB activity as well as inhibition of bcl-xl gene expression sensitize the liver to TRAIL-mediated apoptosis in adenoviral hepatitis and may therefore harbor a risk for liver damage during TRAIL therapy in patients with viral hepatitis.
In summary, viral infection triggers sensitivity to TRAIL-mediated apoptosis as a component of an innate defense mechanism. Overexpression of TRAIL in viral diseases appears to be a defense mechanism of the organism to eliminate infected cells and limit viral replication.8, 37 The molecular mechanisms involved in sensitizing virally infected cells to TRAIL-mediated apoptosis remain unclear, but protection against apoptosis during the early course of viral infection may be necessary for fine tuning the apoptotic machinery to specifically select damaged cells.