TAT-apoptosis repressor with caspase recruitment domain protein transduction rescues mice from fulminant liver failure


  • Potential conflict of interest: Nothing to report.


Acute liver failure (ALF) is associated with massive hepatocyte cell death and high mortality rates. Therapeutic approaches targeting hepatocyte injury in ALF are hampered by the activation of distinct stimulus-dependent pathways, mechanism of cell death, and a limited therapeutic window. The apoptosis repressor with caspase recruitment domain (ARC) is a recently discovered death repressor that inhibits both death receptor and mitochondrial apoptotic signaling. Here, we investigated the in vivo effects of ARC fused with the transduction domain of human immunodeficiency virus 1 (HIV-1) (TAT-ARC) on Fas- and tumor necrosis factor (TNF)-mediated murine models of fulminant liver failure. Treatment with TAT-ARC protein completely abrogated otherwise lethal liver failure induced by Fas-agonistic antibody (Jo2), concanavalin A (ConA), or D-galactosamine/lipopolysaccharide (GalN/LPS) administration. Importantly, survival of mice was even preserved when TAT-ARC therapy was initiated in a delayed manner after stimulation with Jo2, ConA, or GalN/LPS. ARC blocked hepatocyte apoptosis by directly interacting with members of the death-inducing signaling complex. TNF-mediated liver damage was inhibited by two independent mechanisms: inhibition of jun kinase (JNK)-mediated TNF-α expression and prevention of hepatocyte apoptosis by inhibition of both death receptor and mitochondrial death signaling. We identified JNK as a novel target of ARC. ARC's caspase recruitment domain (CARD) directly interacts with JNK1 and JNK2, which correlates with decreased JNK activation and JNK-dependent TNF-α production. Conclusion: This work suggests that ARC confers hepatoprotection upstream and at the hepatocyte level. The efficacy of TAT-ARC protein transduction in multiple murine models of ALF demonstrates its therapeutic potential for reversing liver failure. (HEPATOLOGY 2012)

Death of hepatocytes and other hepatic cell types is typically found in liver diseases such as cholestasis, viral hepatitis, ischemia/reperfusion, liver preservation for transplantation, and drug/toxicant-induced injury. Acute liver failure (ALF) is characterized by stimulus-dependent activation of tumor necrosis factor (TNF)-receptor family members and/or mitochondrial death signaling pathways triggering massive apoptotic and/or necrotic cell death.1, 2 A common event leading to both apoptosis and necrosis is mitochondrial permeabilization and dysfunction, although the mechanistic basis of mitochondrial injury may vary in different settings. A better understanding of the cascades leading to liver cell death will be important to develop effective interventions to prevent or treat ALF.

TNF-α, Fas ligand (FasL), and related members of the TNF cytokine family are implicated in hepatocyte killing but the signaling pathways contributing to initiation and progression of ALF are presently unclear.3 Fas-induced apoptosis is implicated in patients with fulminant hepatic failure.1, 2 The Fas receptor contains a domain called “death domain” which is essential for death-inducing signaling complex (DISC) formation.4 This multiprotein complex is required for binding and activation of procaspase-8 and necrotic RIP kinase-mediated signaling. Activated caspase-8 can cleave multiple intracellular substrates, such as downstream effector caspases-3 and -7 and Bid, thus engaging the mitochondrial death pathway.4

TNF-α is a proinflammatory cytokine that acts through two distinct transmembrane receptors: TNF receptor 1 (TNFR1) and TNF receptor 2 (TNFR2) mediating either cell survival, death, or proliferation.5, 6

Most apoptosis inhibitors antagonize only a single central death pathway. An exception is apoptosis repressor with caspase recruitment domain (ARC), which is predominantly expressed in long-living tissues such as heart, brain, and skeletal muscle.7 ARC was originally described as an inhibitor of the death receptor pathway because it blocks apoptosis induced by a variety of death receptors (CD95/Fas, TNFR1, TRAMP/DR3) and their adaptors (Fas-associated protein with death domain [FADD], tumor necrosis factor receptor type 1-associated death domain [TRADD]).7 We showed ARC's ability to block apoptosis induced by activators of the mitochondrial death pathway such as ischemia/reperfusion injury in the heart and doxorubicin-induced cardiotoxicity.8, 9 A recent investigation demonstrates that endogenous ARC inhibits both death receptor and mitochondrial apoptotic death pathways through nonhomotypic death-fold interactions. The death receptor pathway is disrupted by interactions between ARC and Fas, FADD, and procaspase-8.10 The mitochondrial death pathway is inhibited by ARC binding Bax.10 This suggests that ARC could be a treatment option for ALF. We thus tested its therapeutic potential in clinically relevant models of both Fas- and TNF-mediated ALF.


AcD, actinomycin D; ALF, acute liver failure; ARC, apoptosis repressor with caspase recruitment domain; βgal, β-galactosidase; CARD, caspase recruitment domain; Ced-3, C. elegans cell death protein 3; ConA, concanavalin A; DISC, death-inducing signaling complex; FADD, Fas-associated protein with death domain; FasL, Fas ligand; GalN, D-galactosamine; HIV-1, human immunodeficiency virus 1; HM, mitochondrial heavy membrane; IFN-γ, interferon-gamma; IL-4, interleukin 4; JNK, c-Jun N-terminal kinase; Jo2, Fas-agonistic antibody; LPS, lipopolysaccharide; NKT, natural killer cells; siRNA, small interfering RNA; TNF, tumor necrosis factor; TNF-α, tumor necrosis factor alpha; TNFR1, tumor necrosis factor receptor 1; TNFR2, tumor necrosis factor receptor 2; TRADD, tumor necrosis factor receptor type 1-associated death domain; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling.

Materials and Methods

TAT Fusion Proteins.

For the generation of recombinant proteins, pTAT-HA and pTAT-βgal (beta-galactosidase) vectors were obtained from S. Dowdy (Howard Hughes Medical Institute, La Jolla, CA). The pTAT-HA vector was used for cloning of TAT-ARC constructs. We produced TAT recombinant proteins as published.11 We lysed Escherichia coli BL21 or BL21(DE3)pLysS cells (Promega) transformed with pTAT-ARC, pTAT-ARC mutant (L31F; G69R), or pTAT-βgal (for His6-tagged proteins) encoding wildtype (WT) ARC, mutant ARC, and WT βgal, respectively, in 8 mol/L urea buffer, 1.0 mmol/L dithiothreitol (DTT), 10 mmol/L phenylmethylsulfonyl fluoride (PMSF), 15 mmol/L imidazole (Sigma), 100 mmol/L NaCl, 20 mmol/L Hepes, pH 8.0 (Calbiochem), and sonified six times for 30 seconds on ice. Cleared supernatant was subjected to Ni-NTA column (12 mL, GE Healthcare) connected to a fast protein liquid chromatography (FPLC; ÄKTA, GE Healthcare). TAT fusion protein was eluted in Z-buffer containing 500 mM imidiazole and subjected to ionic exchanger chromatography (Mono Q 5/10 column, GE Healthcare). TAT proteins were eluted with a single 2 mol/L NaCl step and desalted in phosphate-buffered saline (PBS; G-25 column, GE Healthcare). We measured the protein concentration by Bradford assay. Purified TAT proteins were adjusted to 10% (v/v) glycerol, aliquoted, and stored at −80°C.

Animals and Acute Liver Injury Models.

Animal experiments were conducted following standards and procedures approved by the local Animal Care and Use Committee. For the animal models of ALF we used age-matched both male and female Balb/c mice for Fas-agonistic antibody (Jo2) and concanavalin A (ConA) models and female Balb/c mice for D-galactosamine/lipopolysaccharide (GaIN/LPS) experiments. Adult 8-week-old Balb/c mice were injected intravenously with 0.25 μg/g of Jo2 diluted in pyrogen-free PBS; 25 mg/kg ConA (Sigma) was injected intravenously diluted in PBS. For GaIN/LPS experiments mice were injected intraperitoneally with 700 mg/kg GaIN (Sigma) plus 35 μg/kg LPS from E. coli 055:B5 (Sigma) diluted in pyrogen-free PBS. TAT proteins were injected intraperitoneally (10 mg/kg) or for rescue experiments intravenously (20 mg/kg). JNK inhibitor SP600125 (50 mg/kg) (Calbiochem) was administered by intraperitoneal injection 1 hour prior to ConA or 1 hour prior and 2 hours after GaIN/LPS injection.

Histologic Analysis.

We perfused animals with ice-cold PBS and then with 4% buffered paraformaldehyde. Tissues were further fixed in 4% buffered paraformaldehyde for 2 days, embedded in paraffin, and processed for sectioning. For histological staining, we stained paraffin-embedded sections of liver tissue with hematoxylin and eosin (H&E).

Subcellular Fractionation, Immunoprecipitation, and Immunoblotting.

Subcellular fractionation was performed as described.12 Immunoprecipitations of the CD95 DISC was done as described.13 We made protein extracts and performed immunoprecipitations as published.11 Protein extracts were mixed with antibodies (1-5 μg/mL) for 2 hours on a rotating wheel, followed by addition of 50 μL of proteinA or G Plus-Sepharose beads (Roche) or 30 μL of agarose conjugated JNK1 (sc-1648 AC) / JNK2 (sc-827 AC) for an additional hour at 4°C. Immunoprecipitates were washed four times with RIPA buffer (for activated Bax, we used CHAPS buffer as described12) and boiled in 50 μL sodium dodecyl sulfate (SDS) sample buffer. Samples were resolved over 12% or 15% SDS-polyacrylamide gels (PAGE) and transferred onto nitrocellulose membranes. We incubated blots with primary antibodies (0.5-5 μg/mL), followed by horseradish peroxidase (HRP)-conjugated secondary antibodies (diluted 1:2,500). Immunoreactive bands were visualized by incubation with LumiGLO (Cell Signaling) and exposed to light-sensitive film.

Detection of Caspase-3, -8, -9 Activities.

Caspase activities were detected using commercial assay kits (ClonTech) according to the kit instructions.

In Vitro Binding and Coimmunoprecipitation.

Purified recombinant full-length human His-JNK1 (2 μg) (Millipore) or GST-JNK2 (2 μg) (Santa Cruz) protein was incubated at 4°C for 5 hours to overnight, with each 5 μg TAT fusion protein (TAT-ARC or TAT-βgal) or the same volume PBS immobilized on agarose conjugated JNK1 (sc-1648 AC)/JNK2 (sc-827 AC) beads in 0.5 mL of buffer, containing 50 mM NaCl, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM PMSF, 2 μg/mL leupeptin, 2 μg/mL aprotinin, 25 mM glycerophosphate, 0.1 mM sodium orthovanadate, 1 mM sodium fluoride, 1% NP-40, and 10% glycerol. After the beads had been washed four times with 500 μL of the same buffer, the bound proteins were eluted from the beads and visualized by SDS-PAGE and immunoblotting.

TNF-α Enzyme-Linked Immunosorbent Assay (ELISA).

Cytokine levels were analyzed in serum samples. ELISA for TNF-α was performed according to manufacturer's recommendations (Duoset, R&D Systems).

Hepatocyte Culture.

Hepatocytes were isolated from mouse liver as described14; 2.5 × 105 hepatocytes were seeded into collagen-coated 6-well plates without or with coverslips for cell counting and fluorescence microscopy, respectively, and cultured for 4 hours. After medium change, cells were incubated with the Pan-JNK inhibitor SP600125 (20 μM), TAT-βgal (10 μg/mL), TAT-ARC (10 μg/mL), or PBS as control. Sixty minutes later cells were treated with TNF-α (30 ng/mL) and AcD (0.4 μg/mL) or GalN (700μg/mL) to induce apoptosis. After 15 hours cells were harvested and living cells were quantified by Trypan blue staining.

Statistical Analysis.

Results are presented as means ± standard error of the mean (SEM). Statistical comparisons were performed using one-way analysis of variance (ANOVA), or ANOVA on ranks with Tukey's or Dunn's posttest. P < 0.05 was considered significant.


Endogenous and Ectopic ARC Expression After Protein Transduction of the Liver.

First we confirmed that ARC protein is expressed endogenously in heart, but not liver-derived tissue e.g., murine and human liver by immunoblot7 (Fig. 1A). The therapeutic time window during lethal liver failure is limited; hence, we aimed to apply a protein-based therapy approach using the transduction domain of HIV-1 TAT.15 Earlier results demonstrated that intraperitoneally injection of the 120-kDa βgal protein, fused to the protein transduction domain derived from the HIV TAT domain, results in rapid delivery of biologically active fusion protein into mouse organs including the liver.16 Strong and rapid expression of TAT-ARC and TAT-βgal fusion proteins were detected in mouse liver lysates for 24 hours after single intraperitoneal injection (Fig. 1B) and subcellular fractions of cytoplasm and mitochondrial heavy membrane (data not shown). Of note, no adverse or toxic effects related to TAT fusion protein transduction were evident as indicated by normal serum transaminase levels following TAT protein transduction (Fig. 2B).

Figure 1.

TAT-ARC protein transduction to the liver. (A) Immunoblotting of endogenous and ectopic ARC expression in murine and human liver samples. ARC is neither expressed in hepatocytes nor liver at the endogenous level. Protein extracts from both murine heart and TAT-ARC transduced liver served as positive controls. (B) Western blot analysis showing TAT-βgal and TAT-ARC expression following transduction of each 10 mg/kg TAT-fusion protein in the liver by way of intraperitoneal injection. Both TAT-βgal and TAT-ARC fusion proteins contain tags against His and HA.

Figure 2.

TAT-ARC protects from Fas-induced ALF. (A) The survival curves of the different groups (n = 7/group) are shown following Jo2 injection. (B) Serum transaminase levels (alanine aminotransferase [ALT] and aspartate aminotransferase [AST]) are shown 6 hours after treatment with Jo2 or the vehicle control (n = 4-7/group, data are mean ± SEM, *P < 0.001 versus TAT-βgal or PBS stimulated with Jo2). The AST levels measured were beyond the maximal detection level (cutoff 8,000 U/I). (C) Macroscopic appearance of representative liver samples, H&E staining, and TUNEL assay 6 hours after Jo2 treatment of the different groups as indicated. In the TUNEL assay bright green cells indicate apoptotic cells.

TAT-ARC Protein Transduction Rescues From Fas-Mediated Fulminant Liver Failure.

Hepatocytes are highly susceptible to Fas-induced apoptosis.2 A prominent role of the Fas-FasL system has been reported in hepatic injury from diverse insults, including viruses, autoimmunity, and transplant rejection.17, 18 To determine whether ARC protects from Fas-mediated ALF in vivo, mice were injected intravenously with Jo2 2 hours after pretreatment with TAT-ARC, TAT-βgal, or PBS intraperitoneally, respectively. Jo2 stimulation resulted in death of TAT-βgal or PBS-pretreated mice within 12 hours (Fig. 2A). This was associated with extensive hepatocellular damage, as indicated by a massive increase of serum transaminases (Fig. 2B). In contrast to the PBS or TAT-βgal cotreated group, all TAT-ARC-pretreated mice survived a lethal dose of Jo2 challenge without signs of liver injury showing normal serum transaminase levels (Fig. 2A-C). Notably, all mice survived Fas-mediated ALF even when TAT-ARC fusion protein was given 1 hour after Jo2 stimulation (Fig. 2A). The protective effect of ARC was already detectable macroscopically on liver appearance, with strong hemorrhagic changes in livers derived from Jo2−, and Jo2+ TAT βgal-treated mice, but normal liver structure in Jo2+ TAT-ARC treated and untreated control mice (Fig. 2C). Staining of liver sections by H&E and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay confirmed extensive hepatocyte apoptosis in mice treated with Jo2 and TAT-βgal or PBS, whereas TAT-ARC pretreated mice appeared unaffected (Fig. 2C). Strong elevations of caspase-8, -9, and -3 levels after Jo2 injection demonstrated activation of both Fas-death receptor and mitochondrial apoptotic signaling pathways (Fig. 3A,B). Accordingly, mitochondrial cytochrome c release was detected in response to Jo2 stimulation (Fig. 3C). In contrast, TAT-ARC-treated mice challenged with Jo2 showed unaffected caspase-8 and -9 activities, with only mild elevation in caspase-3 activity in the proteolytic assay but neither caspase-3 cleavage nor mitochondrial cytochrome c release in the immunoblot (Fig. 3A-C).

Figure 3.

TAT-ARC abrogates Fas-induced ALF by inhibiting DISC assembly and activation of the mitochondrial death pathway through protein-protein interactions. (A) Caspase-8, -9, and -3 activities in liver homogenates of TAT-ARC, TAT-βgal, or PBS-pretreated mice 6 hours after stimulation with Jo2 and the respective controls as indicated (n = 4/group, data are mean ± SEM, *P < 0.05 versus TAT-βgal or PBS stimulated with Jo2; #P < 0.001 versus TAT-βgal or PBS stimulated with Jo2). IETD.fmk, LEHD.cho, and DEVD.fmk indicate caspase inhibitors. (B) Representative western blot showing caspase-8 and -3 activation in liver homogenates of TAT-ARC, TAT-βgal, or PBS-pretreated mice 2 hours after stimulation with Jo2. (C) Mitochondrial cytochrome c release in subcellular fractions of cytosol and mitochondrial heavy membrane (HM) after Jo2 treatment. (D) Interaction of TAT-ARC with endogenous members of the DISC complex and BH3-only proteins analyzed by immunoprecipitation-immunoblots of liver lysates. (E) Immunoprecipitation-immunoblot experiments showing inhibition of DISC assembly by TAT-ARC.

Activation of caspase-8 is essential for triggering Fas-mediated ALF and endogenous ARC was previously shown to interfere with assembly of the DISC.10 Immunoprecipitation experiments were performed to investigate the interaction of TAT-ARC with members of the DISC complex such as Fas, FADD, and procaspase-8. In contrast to PBS or TAT-βgal-treated controls, immunoprecipitations of ectopic ARC 1 hour after TAT-ARC administration demonstrated binding of ARC to Fas, FADD, and procaspase-8 in liver lysates, respectively (Fig. 3D). In addition, interactions of TAT-ARC could be detected with the proapoptotic BH3-only Bcl-2 family members Bax and Bad that are critical mediators of the intrinsic death pathway (Fig. 3D).

To prove the functional relevance of these observations we tested its effect on DISC formation. Although stimulation of PBS or TAT-βgal-treated mice with Jo2 resulted in rapid DISC assembly, TAT-ARC completely blocked Jo2-induced DISC formation as shown by immunoprecipitates of TAT-ARC-transduced livers containing ARC, but no Fas or FADD (Fig. 3E). These experiments demonstrate that TAT-ARC blocks Fas-mediated ALF by inhibiting DISC formation.

TAT-ARC Protein Transduction Rescues From TNF-Dependent Fulminant Hepatitis.

Besides Fas, other members of the TNF cytokine family have been implicated in hepatocyte killing in humans.1 TNF-dependent fulminant hepatic failure in mice can be induced after LPS application with the liver-specific transcription inhibitor, GalN, or treatment with the T-cell mitogen, ConA.19, 20 In both models TNF-α is essential for hepatocyte killing and death of the animals. Secreted TNF-α is critical in GalN/LPS-challenged mice, whereas both secreted and membrane-bound TNF-α contribute to hepatocyte destruction after ConA stimulation. To evaluate whether TAT-ARC protects from TNF-mediated ALF, mice were pretreated with TAT-ARC, TAT-βgal, or PBS and challenged 2 hours later by ConA intravenously or application of GalN/LPS intraperitoneally. In both models, TAT-βgal or PBS-treated mice died within 24 hours from ALF, as indicated by markedly elevated serum transaminases (Fig. 4A,B). In contrast, TAT-ARC-treated mice showed strong resistance to lethal doses of ConA and GalN/LPS, respectively (Fig. 4A,B). Notably, delayed TAT-ARC administration 2 hours following ConA and 15 minutes after GalN/LPS was able to rescue ConA- and GalN/LPS-challenged mice (Fig. 4A). In contrast to TAT-βgal or PBS-treated mice that showed activation of caspases-8 and -3 after ConA and GalN/LPS, respectively, no caspase activation was seen in TAT-ARC-pretreated mice (Fig. 4C).

Figure 4.

TAT-ARC protects from TNF-mediated ALF in the ConA and GalN/LPS models. (A) Mice were challenged with either ConA intravenously (left panel, n = 5-8/group) or GalN/LPS intraperitoneally (right panel, n = 5/group) and the survival curves for the different conditions are depicted. (B) Transaminase levels (ALT and AST) are shown and were measured 8 hours after ConA and 6 hours after GalN/LPS treatment (n = 5-8/group, data are mean ± SEM, *P < 0.05 versus TAT-βgal or PBS stimulated with ConA or GalN/LPS). (C) Western blots depicting caspase-8 and -3 activation in livers lysates of ConA and GalN/LPS treated mice as shown.

TAT-ARC Confers Protection in Hepatocytes and Nonparenchymal Liver Cells.

To investigate whether TAT-ARC protein transduction confers direct cell protective effects in hepatocytes, primary hepatocytes were pretreated with TAT-ARC and stimulated with TNF/actinomycin D (AcD) or TNF/D-galactosamine. In response to TAT-ARC pretreatment, survival was significantly improved in both models compared to PBS or TAT-βgal-pretreated hepatocytes. TAT-ARC-mediated protection of hepatocytes was in both models comparable to that of JNK-inhibitor SP600125 pretreated hepatocytes (Fig. 5A). TNF/AcD or TNF/GalN stimulation of hepatocytes resulted in JNK activation that was inhibited by TAT-ARC or JNK-inhibitor pretreatment (Fig. 5B). In both models, ConA- and GalN/LPS-induced hepatitis, TNF-α levels have been shown to be critical for hepatocyte killing and high mortality of the animals. Thus, we examined the effect of TAT-ARC on serum TNF-α levels in murine models of hepatitis caused by ConA and GalN/LPS. Importantly, in both models TAT-ARC significantly reduced serum TNF-α levels (Fig. 5C). Together, these data suggest that TAT-ARC prevents TNF-mediated hepatitis by inhibiting TNF-α expression, e.g., in nonparenchymal cells, but also directly protects hepatocytes from apoptosis.

Figure 5.

TAT-ARC protein transduction shows direct cell protective effects in hepatocytes and inhibits TNF-α production. (A) Hepatocyte survival determined by Trypan blue exclusion. Isolated hepatocytes were pretreated with PBS, TAT-βgal (10 μg/mL), TAT-ARC (10 μg/mL), or JNK inhibitor (SP600125 20 μM) and then stimulated for 15 hours with TNF-α (30 ng/mL) and AcD (0.4 μg/mL) or GalN (700μg/mL) (representative of three independent experiments; data are mean ± SEM, *P < 0.05 versus TAT-βgal or PBS stimulated with TNF-α/AcD or TNF-α/GalN). (B) Immunoblot showing JNK activation 1 hour after stimulation with TNF-α and AcD or GalN. Conditions as above. (C) TNF-α concentration in the serum of mice 60 minutes after challenge with ConA or GalN/LPS and pretreated with PBS, TAT-βgal, or TAT-ARC (n = 5-6/group, data are mean ± SEM, *P < 0.05 versus TAT-βgal or PBS stimulated with ConA or GalN/LPS).

ARC-Mediated Hepatoprotection in TNF-Mediated Hepatitis Is JNK-Dependent.

The crucial role of JNK signaling in TNF-dependent ALF was demonstrated by Maeda et al.,21 who showed that mice lacking either JNK1 or JNK2 are highly resistant to ConA-induced ALF. Thus, we investigated JNK activation by immunoblot analysis using liver lysates from both ConA- and GalN/LPS-treated mice. As shown in Fig. 6A, both p46- and p54-JNK phosphorylation, which are essential steps for JNK activation, were significantly induced after ConA or GalN/LPS stimulation. Interestingly, both p46- and p54-JNK phosphorylation were strongly reduced in TAT-ARC-treated mice following ConA stimulation and were completely abrogated after GalN/LPS administration (Fig. 6A). Because activated JNK translocates from cytosol to mitochondria to trigger cell death JNK translocation was assessed by subcellular fractionation and immunoblotting of liver lysates.22 TAT-ARC application completely blocked JNK activation and subsequent mitochondrial translocation following ConA or GalN/LPS, respectively (Fig. 6B). Because the death-promoting function of JNK-signaling in the liver is antagonized by p38 signaling, p38α phosphorylation was analyzed23 (Fig. 6A). Although no relevant activation of p38-signaling was detected 4 hours after GalN/LPS stimulation when JNK was already activated, TAT-ARC-mediated hepatoprotection following ConA stimulation was associated with a substantial concomitant activation of p38α signaling. It remains unclear whether p38α activation following ConA and TAT-ARC treatment plays a causal role for the observed protective effect seen or is rather a secondary phenomenon.

Figure 6.

ARC inhibits JNK activation and translocation in TNF-α-mediated ALF. (A) JNK and p38α activity is visualized by western blot analysis using lysates from mouse livers stimulated with ConA or GalN/LPS at the time points and conditions indicated. (B) TAT-ARC inhibits JNK activation and mitochondrial translocation. Western blot analysis from cytosolic subcellular fractions and mitochondrial HM were performed before and after ConA or GalN/LPS treatment. (C) ARC prevents activation of proapoptotic BH3-only proteins Bax and Bim following ConA or GalN/LPS stimulation in western blots of liver lysates. (D) The JNK inhibitor SP600125 protects from ConA- and GalN/LPS-induced ALF. The survival curves of the different conditions are shown (n = 5/group).

JNK specifically regulates the proapoptotic activity of BH3-only proteins Bax and Bim, which cooperate with Bid in hepatocyte killing.24 To assess the functional significance of ARC-mediated JNK inhibition we investigated the activity of BH3-only proteins Bax and Bim. Both Bax and Bim activation resulted in mitochondrial translocation triggering the intrinsic death pathway. ConA or GalN/LPS stimulation resulted in activation of Bax and Bim that was inhibited by TAT-ARC pretreatment (Fig. 6C). TAT-ARC application abrogated Bim mitochondrial translocation following ConA or GalN/LPS stimulation (data not shown) but no interaction of ARC and Bim was detected (data not shown). However, due to the direct ARC-Bax interaction it remains unclear whether abrogated Bax activation results from ARC's inhibition of Bax or JNK only or a combination of both. Thus, our results suggest that abrogated Bax activation might result from direct inhibition by ARC or, alternatively, from ARC-mediated JNK inhibition, whereas impaired Bim activation is most likely an indirect effect of ARC, probably mediated through JNK inhibition.

The pathophysiological relevance of JNK signaling in TNF-mediated models of ALF was demonstrated in mice treated with the small molecule JNK inhibitor, SP600125, showing JNK-dependent survival (Fig. 6D). These observations clearly show that JNK signaling is critically involved in mediating hepatotoxicity in both models.

Interaction Between ARC's CARD and JNK Inhibits JNK Activation and Translocation.

Our results demonstrated that in both models of TNF-dependent liver injury ARC-dependent protection is associated with JNK inhibition. Hence, we sought whether ARC/JNK interaction might be involved in mediating protection, and thus performed immunoprecipitation experiments to test this hypothesis. Immunoprecipitation of lysates from TAT-ARC-transduced livers demonstrated binding of TAT-ARC to endogenous JNK1 and JNK2, respectively (Fig. 7A). The interactions of ectopic ARC with both JNK1 and JNK2 were further confirmed using JNK1 and JNK2-specific antibodies (Fig. 7B).

Figure 7.

ARC inhibits JNK activation and translocation by binding JNK1/2 by way of the CARD domain in TNF-α-mediated ALF. (A,B) TAT-ARC interacts with endogenous JNK1/2 in vivo. Immunoprecipitation experiments of liver lysates were performed using the conditions as indicated. (C) TAT-ARC directly binds JNK1/2 in vitro. A cell-free system was used to prove ARC's specific and direct interaction with JNK1/2 by using both recombinant JNK1/2 proteins. (D) Endogenous ARC interacts with endogenous JNK1/2 in skeletal but not heart muscle. Immunoprecipitation experiments were performed with murine heart or skeletal muscle tissue lysates as indicated. (E) ARC interacts by way of its CARD domain with JNK1/2. Immunoprecipitation experiments were performed as indicated. (F) Both ConA- and GalN/LPS-induced fulminant liver failure is JNK-dependent. The TAT-ARC CARD mutated protein is not protective for ConA or GalN/LPS-induced ALF. The survival curves of mice pretreated with TAT-ARC mutant (n = 3/group) and challenged with ConA or GalN/LPS are shown.

To exclude unspecific antibody binding, because eight JNK isoforms exist at the messenger RNA level, and to investigate whether interactions between ARC and JNK are direct or indirect, a cell-free system was used (Fig. 7C). Applying a cell-free system with both recombinant JNK1 and JNK2 protein proved the specificity of ARC JNK1 and JNK2 interactions. Furthermore, our results demonstrated that ARC interacts directly with JNK1 and JNK2 (Fig. 7C). Although TAT-ARC interacted with JNK1 and JNK2, it did not bind other relevant mediators of TNF signaling such as Flip, RIP, TRADD, or TRAF2 (data not shown). These results suggest that ectopic ARC protein inhibits JNK activation and translocation in vivo by binding to endogenous JNK1 and JNK2 in the liver.

To elucidate the physiological occurrence of the ARC-JNK interaction, immunoprecipitations were performed using murine heart and skeletal muscle lysates that express ARC, JNK1, and JNK2 endogenously.7 Immunoprecipitation experiments confirmed interactions of endogenous ARC with endogenous JNK1 and JNK2 in skeletal muscle (Fig. 7D). However, no interactions could be detected in heart muscle despite endogenous ARC, JNK1, and JNK2 expression. The reason for the tissue-specific ARC-JNK interaction remains unclear. Because CARDs mediate protein-protein interactions and ARC's CARD was shown to interact with Fas, FADD, procaspase-8, and Bax, its functional importance in binding JNK was assessed.10 We disrupted ARC's CARD by mutating two residues (L31F; G69R) that are conserved in death-fold proteins back to Ced-3.25 Mutant TAT-ARC abrogated the interaction of ectopic TAT-ARC with JNK1 and JNK2 and showed no protection against TNF-α-mediated liver failure (Fig. 7E,F). Thus, the CARD of ARC mediates its interaction with JNK1 and JNK2. Thus, our results suggest that ARC inhibits JNK activation and translocation by a direct interaction between ARC's CARD and JNK1 and JNK2.


ARC is exceptional in its ability to antagonize both the extrinsic (death receptor) and the intrinsic (mitochondria / endoplasmic reticulum [ER]) death pathways.8-10 Here, we demonstrate highly efficient therapeutic in vivo delivery of ARC to the adult murine liver using the TAT protein transduction technique. Ectopic ARC delivery completely blocks Fas- and TNF-mediated hepatocyte apoptosis in vitro and in three different in vivo models of ALF protecting mice from death in preventive and therapeutic settings. Fas-induced apoptosis is triggered by way of Fas receptor-mediated DISC assembly.4 TAT-ARC blocks caspase-8-dependent cell killing by binding to members of the DISC, namely Fas, FADD, and procaspase-8. Additionally, it inhibits Fas-mediated Bax conformational activation and subsequent mitochondria-dependent death signaling. Hepatocytes are highly sensitive to Fas-induced apoptosis compared with other tissues and organs and absence of endogenous ARC might contribute to this observation.2 Previous in vivo studies demonstrated successful hepatic delivery of small interfering RNA (siRNA) targeting Fas or caspase-8 of mice with Fas-mediated hepatitis.25, 26 However, the relevance of those therapeutic approaches targeting hepatocyte injury in ALF is limited due to its delayed mode of action and the low delivery efficiency of siRNA into hepatocytes.26, 27 TAT-ARC does not have these limitations and therefore might be a more valuable candidate for treatment of ALF in humans.

Several studies have convincingly demonstrated a critical role of JNK during ConA or GalN/LPS-induced hepatocyte apoptosis.21, 28-30 These findings suggested JNK as a major therapeutic target and JNK-specific drugs are currently in clinical development. We demonstrate that administration of TAT-ARC prevents JNK activation in the liver upon ConA or GalN/LPS-induced hepatitis. In vitro experiments with recombinant JNK1 and JNK2 show binding with the CARD domain of ARC, indicating that ARC directly suppresses JNK activity, which has not been reported before. Traditionally, death-fold motifs use homotypic protein-protein interactions. The CARD of ARC engages in homotypic death-fold interactions as shown by ARC homodimerization.10 Additionally, ARC's CARD was shown to be involved in heterotypic interactions with the death domains of Fas and FADD and the C terminus of Bax.10 The absence of CARD and a death domain in JNK1 and JNK2 suggests other nonhomotypic death-fold interactions between JNKs and ARC. Further studies will aim to map the ARC binding domain in JNK1 and JNK2. In accordance with other studies, the application of JNK inhibitor abrogated ConA- and GalN/LPS-induced ALF, indicating a crucial role of JNK-dependent signaling in these models.31 Therefore, our results indicate that the interaction between JNK1 and JNK2 with ARC is involved in protecting mice from TNF-mediated ALF. However, at present the ultimate role of JNK1 and JNK2 in regulating cell death is still not completely defined.

ConA and GalN/LPS-induced hepatitis have been reported to be TNF-dependent,29 which is, e.g., evidenced by the fact that liver cell death in the ConA model is greatly reduced in Tnfr1−/−, Tnfr2−/−, and TNF−/− mice.21, 32 In the present study, ARC delivery strongly suppressed TNF-α serum levels in both ConA and GalN/LPS-induced hepatitis. Our observation that ARC prevents JNK activation suggests that suppression of TNF-α serum levels by ectopic ARC results from its JNK inhibitory function. Indeed, JNK plays an important role in TNF-α gene transcription.33 JNK1- and JNK2-deficient fibroblasts exhibit a severe defect in TNF-α mRNA expression and JNK1- and JNK2-deficient macrophages and T cells express profoundly reduced amounts of TNF-α in the culture medium.33 Previous reports also show that JNK in hematopoietic cells is critically required for TNF-α expression and that JNK is not required for TNF-stimulated cell death during development of hepatitis.29 Because membrane-bound TNF-α, but not soluble TNF-α, is required for ConA-induced hepatitis, it is likely that the hematopoietic cells that are the source of JNK-dependent TNF-α expression include resident inflammatory liver cells like Kupffer and natural killer (NKT) cells.34 Indeed, NKT cell-mediated expression of TNF-α, interferon-gamma (IFN-γ), and interleukin (IL)-4 have been implicated in ConA-induced hepatitis.35

Most antiapoptotic approaches are limited by interfering selectively only with either death receptor or mitochondrial apoptotic signaling, or operating at the postmitochondrial level like caspase-inhibitors. Multiple interactions of ARC with critical mediators of cell death receptor and mitochondrial death signaling at the premitochondrial stage result in a strong inhibition of apoptotic cell death. “Redundant” death repressing interactions of ectopic ARC protein with critical mediators of both death pathways guarantee interference with death signaling at different stages.10 Furthermore, TAT-ARC supposedly not only interferes with death signaling at the hepatocyte level but also upstream within the compartment of resident hepatic inflammatory cells. These multiple actions of ARC suggest that TAT-ARC might be a more powerful tool for the treatment of ALF than caspase- or small molecule JNK-inhibitors targeting only one pathway. In contrast to TNF-α-mediated hepatitis, apoptosis caused by Fas-ligand (FasL) is JNK-independent.36 Based on the crucial role of Fas- and TNF-mediated cell injury in a broad spectrum of immune-related liver diseases, TAT-ARC protein transduction or ARC-related small molecules could be of therapeutic benefit for preventing and treating acute and chronic liver injuries.17, 18 A short-term application of an antiapoptotic approach would also limit the risk of developing cancer. Although we applied TAT fusion protein intraperitoneally or intravenously, regional delivery of high TAT-ARC concentrations or small molecules by way of the hepatic artery or portal vein might be attractive in the therapeutic setting.


We thank Katarzyna Pogodzinski, Marlies Grieben, and Nadine Weser for excellent technical assistance. pTAT-HA and pTAT-βgal vectors were kindly provided by S. Dowdy (Howard Hughes Medical Institute, La Jolla, CA). This article is dedicated to Prof. Dr. Michael P. Manns on the occasion of his 60th birthday.