Potential conflict of interest: Nothing to report.
Induction of hepatic stellate cell (HSC) apoptosis attenuates hepatic fibrosis, and, therefore, mechanisms to induce HSC cell death are of therapeutic interest. Proteasome inhibitors induce apoptosis in transformed cells, especially those cells dependent upon nuclear factor kappa B (NF-κB) activation. Because stimulated HSCs also trigger NF-κB activation, the aim of this study was to determine if proteasome inhibitors induce HSC apoptosis. The immortalized human HSC line, LX-2, and primary rat HSCs were treated with the proteasome inhibitors bortezomib and MG132. Both proteasome inhibitors induced HSC apoptosis. Proteasome inhibition blocked NF-κB activation and, more importantly, NF-κB inhibition by Bay11-7082–triggered HSC apoptosis. Activated HSC survival is dependent upon the NF-κB target gene A1, an anti-apoptotic Bcl-2 family member, as siRNA targeted knockdown of A1-induced HSC apoptosis. In contrast, proteasome inhibition–induced alterations in TRAIL, death receptor 5, and Bim could not be implicated in the apoptotic response. The relevance of these findings was confirmed in the bile-duct–ligated mouse where bortezomib reduced hepatic markers of stellate cell activation and fibrosis. In conclusion, proteasome inhibition is a potential therapeutic strategy for inducing HSC apoptosis and inhibiting liver fibrogenesis. (HEPATOLOGY 2006;43:335–344.)
Liver fibrosis is a common response to chronic liver injury due to viral, metabolic, genetic, and cholestatic liver disease. Hepatic stellate cells (HSCs) are recognized as the primary cellular source of matrix components in chronic liver disease, and therefore play a critical role in the development and maintenance of liver fibrosis.1, 2 In normal liver, HSCs are in a quiescent state within the space of Disse, where they primarily store retinoids. However, in response to injury these cells transform into proliferative, fibrogenic, proinflammatory, and contractile myofibroblasts which express α-smooth muscle actin (α-SMA).3 These activated stellate cells secrete collagen type I and III, which are the collagens responsible for liver fibrosis. Increasing evidence suggests that hepatic fibrosis is reversible.4–6 Resolution of hepatic fibrosis requires cessation of activated HSC function. However, HSCs appear to undergo commitment steps to become activated; which prevent reversion to the quiescent phenotype. The resolution phase of fibrosis, therefore, requires loss of activated HSC via apoptosis.5, 6 Indeed, induction of HSC apoptosis by gliotoxin or sulfasalazine reduces hepatic fibrosis providing proof of principle that HSC apoptosis is a therapeutic antifibrogenic strategy.7, 8
HSC activation is associated with activation of nuclear factor kappa B (NF-κB), a potent prosurvival transcription factor.9 Inhibition of NF-κB activation has been associated with apoptosis-induction in many cancer cells.10 Although it has been reported that NF-κB inhibition does not result in HSC apoptosis,7 several studies have indicated that NF-κB inhibition is a potent mechanism to induce HSC apoptosis.8, 11 Thus, mechanisms to inhibit NF-κB in HSC have potential therapeutic relevance.
NF-κB is normally complexed to an endogenous inhibitor protein I-κB in the cytosol.12 NF-κB–activating stimuli result in phosphorylation of I-κB, leading to its dissociation from NF-κB and degradation by the proteasome.12 Freed from I-κB inhibition, NF-κB translocates to the nucleus and functions as a transcription factor. NF-κB induces expression of survival genes, including the anti-apoptotic Bcl-2 family proteins Bcl-xL and A112; loss of these proteins can result in cell death.13 Proteasome inhibition prevents I-κB degradation, blocking NF-κB activation, an effect which can culminate in loss of survival proteins and cell death.14 In addition to blocking NF-κB activation, proteasome inhibitors may also upregulate death receptor 5 (DR5) and the proapoptotic BH3-only protein, Bim.15, 16 Thus, proteasome inhibitors, which already are clinically available for the treatment of cancers, are a particularly attractive strategy for induction of HSC apoptosis.
The overall objective of this study was to examine the effects of proteasome inhibitors on HSC apoptosis. To address this objective, we formulated the following questions: (1) Do proteasome inhibitors induce apoptosis in activated HSCs? and if so, (2) what is the cellular mechanism? and (3) Does proteasome inhibition attenuate hepatic fibrogenesis in the bile duct ligated (BDL) mouse? The results demonstrate that proteasome inhibitors induce HSC apoptosis by inhibiting NF-κB activity, and thereby reduce hepatic fibrosis. Our data suggest that proteasome inhibitors are potential candidates of “chemotherapy” for liver fibrosis.
HSC, hepatic stellate cell; NF-κB, nuclear factor kappa B; DR5, death receptor 5; BDL, bile duct ligated; Bcl-2, B cell leukaemia-2; DR4, death receptor 4; α-SMA, α-smooth muscle actin.
Materials and Methods
Cell Lines and Culture.
The human stellate cell line, LX-2 cells, obtained from Dr. Scott Friedman, Mt. Sinai, NY, are spontaneously immortalized cells,17 which exhibit the typical features of activated stellate cells: expressing desmin and glial acidic fibrillary protein, and responsiveness to platelet-derived growth factor BB and transforming growth factor β1 (TGFβ1). The cells express α-SMA under all culture conditions and must be regarded as at least partially activated, even after immediate replating. LX-2 cells were cultured in Dulbecco's minimum essential medium (DMEM; Bio-Whittaker, Walkersville, MD) containing 10% fetal bovine serum, 100 U/mL penicillin, 100 μg/mL streptomycin, 50 μg/mL gentamicin, and 100 nmol/L insulin. Primary HSCs were isolated and cultured from normal Fisher rats, as previously described in detail.18 HSC cell purity was confirmed by: (1) immunocytochemical expression of α-SMA after several days in culture, (2) lack of expression of eNOS by Western blot analysis of HSC lysates, and (3) characteristic phase contrast morphology. Isolated primary HSCs were resuspended in culture media, plated on plasticware, and activated.19 These cells were used within the first two passages.
Apoptosis in cell cultures was quantified by assessing the characteristic nuclear changes of apoptosis using the nuclear binding dye DAPI (Molecular Probes Inc., Eugene, OR) or Annexin V FITC labeling (Calbiochem, San Diego, CA) with fluorescent microscopy.20
Cells were directly lysed for 30 minutes on ice with lysis buffer containing 50 mmol/L Tris-HCl (pH 7.4), 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mmol/L NaCl, 1 mM EDTA, 1 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L Na3VO4, 1 mmol/L NaF, and protease inhibitors (Complete Protease Inhibitor Cocktail; Boehringer Mannheim Biochemica, Mannheim, Germany), and centrifuged at 14,000 g for 15 minutes at 4°C. Protein content was determined by the Bradford assay (Bio-Rad, Hercules, CA). Samples were resolved by 12% or 15% SDS-PAGE, transferred to nitrocellulose membrane, and blotted with primary antibodies at a dilution of 1:1000. Peroxidase-conjugated secondary antibodies (Biosource International, Camarillo, CA) were incubated at a dilution of 1:2000. Bound antibodies were visualized using enhanced chemiluminescent substrate (ECL; Amersham, Arlington Heights, IL) and exposed to Kodak X-OMAT film. The following primary antibodies were used: rabbit anti-TRAIL-R1/death receptor 4 (DR4), rabbit anti-TRAIL, goat anti-I kappa B alpha (I-κBα), rabbit anti-Bcl-xL, goat anti-actin (Santa Cruz Biotechnology Inc., Santa Cruz, CA), goat anti-TRAIL-R2/DR5 (Alexis Corp., San Diego, CA), mouse anti-α-SMA (Sigma, St. Louis, MO), rabbit anti-caspase-3, and rabbit anti-Bim (BD Pharmingen, San Diego, CA).
Immunocytochemistry for Cellular NF-κB Localization.
Cells, cultured on collagen-coated glass coverslips, were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton-X in phosphate-buffered saline (PBS), and blocked with PBS containing 1% bovine serum albumin. To ascertain the cellular localization, cytoplasmic versus nuclear, for the p65 subunit of NF-κB, cells were next incubated with a polyclonal rabbit anti-p65 immunoglobulin G (sc-372, Santa Cruz Biotechnology) in blocking buffer (1:1500) at 4°C overnight. After washing, Cy3-conjugated goat anti-rabbit Ig (Jackson ImmunoResearch Labs, West Grove, PA) in blocking buffer (1:1000) was added for 1 hour at room temperature. Cells were imaged by confocal microscopy (Zeiss LSM 510, Carl Zeiss, Inc., Thornwood, NJ) as previously described in detail.21
NF-κB DNA Binding Activity.
The DNA binding activity of NF-κB in LX-2 cells was quantified by an ELISA assay using the trans-AM NF-κB p65 Transcription Factor Assay kit (Active Motif North America, Carlsbad, CA), according to the instructions of the manufacturer. Briefly, nuclear extracts were prepared by using the NE-PER nuclear and cytoplasmic extraction reagents (Pierce Biotechnology Rockford, IL). The protein content in the two fractions was quantitated by the Bradford assay. Nuclear extracts were incubated in 96-well plates coated with immobilized oligonucleotide (5′-AGTTGAGGGGACTTTCCCAGGC-3′) containing a consensus (5′-GGGACTTTCC-3′) binding site for the p65 subunit of NF-κB. NF-κB binding to the target oligonucleotide was detected by incubation with primary antisera specific for the activated form of p65. The ELISA assay was developed by employing an anti-IgG horseradish peroxidase conjugate and a developing solution provided by the kit. Optical density (OD) was determined at 450 nm with a reference wavelength of 655 nm. Background binding was subtracted from the value obtained for binding to the consensus DNA sequence.
Real-Time Polymerase Chain Reaction (PCR).
Total RNA was obtained from LX-2 cells and hepatic tissue using the TRIzol Reagent (Invitrogen, Carlsbad, CA). For each RNA sample, a 10-μg aliquot was reverse transcribed into cDNA using oligo-dT random primers and Moloney Murine Leukemia virus reverse transcriptase. Real-time PCR was performed using Taq polymerase (Invitrogen) as previously described.22 Primers used were as follows: A1 forward, 5′-CGAAAGTGACTAGGGAGGAAGG-3′, reverse, 5′-ACTCTGGACGTTTTGCTTGG-3′ (yielding a 280-bp product); COL1A1 forward, 5′-GAAACCCGAGGTATGCTTGA-3′, reverse, 5′-GACCAGGAGGACCAGGAAGT-3′ (yielding a 276-bp product); α-SMA, forward, 5′-ACTACTGCCGAGCGTGAGAT-3′, reverse, 5′-AAGGTAGACAGCGAAGCCAA-3′, (yielding a 452-bp product). 18S primers (Ambion, Austin, TX) were used as a control for RNA isolation and integrity. All PCR products were confirmed by gel electrophoresis. Real-time PCR was performed using the LightCycler (Roche Diagnostics, Mannheim, Germany) and SYBR green as the fluorophore (Molecular Probes). The results were expressed as a ratio of product copies/mL to copies/mL of housekeeping gene 18S from the same RNA (respective cDNA) sample and PCR run.
Transient Transfection of LX-2 Cells.
LX-2 cells were transiently transfected using Lipofectamine (Invitrogen); with plasmids encoding for dominant-negative (DN)-FADD green fluorescent protein (pcDNA3-GFP-ΔFADD) as previously described in detail.23 LX-2 cells were analyzed 24 hours after transfection. The transfection efficiency was approximately 40% to 60% as estimated by the percentage of cells expressing GFP as quantitated by fluorescence microscopy.
A specific double-stranded 21-nucleotide RNA sequence homologous to the target message was used to silence Bim and A1 expression. siRNA was designed and synthesized using the software (www.ambion.com) and Silencer siRNA Construction kit (Ambion, Austin, TX) according to the manufacturer's instructions. The sequences of the double-stranded RNA used to block expression in the current experiments are 5′-AAA TGA TTA CCG CGA GGC TGA CCT GTC TC-3′ (T7-promoter in bold) for Bim and 5′-AAG CTT CAG GAG AGA TAG CAT CCT GTC TC-3′ (T7-promoter in bold) for A1; the A1 sequence was targeted for both human and mouse. Inhibition of mRNA or protein expression and proteasome inhibitor–induced apoptosis were assessed after transient transfection of stellate cells with siRNA. Briefly, cells grown in 6- or 12-well dishes were transiently transfected with Bim [20 nmol/L] or A1 [50 nmol/L] siRNA using 6 μL/mL siPORT Lipid (Ambion) in a total transfection volume of 1.0 mL or 0.5 mL of DMEM containing 10% fetal bovine serum. After 5 hours of incubation at 37°C with 5% CO2, 1.0 mL or 1.5 mL of normal growth medium was added. After 48 hours, cells already transfected with A1 siRNA were re-transfected by the same procedure described in previous paragraph. Samples were then prepared and analyzed for apoptosis, immunoblot analysis or real-time PCR as described in this section.
Cellular TRAIL Secretion.
Cells were incubated for up to 48 hours with or without proteasome inhibitor. Supernatants were collected with a pipette and centrifuged to remove particulate debris. Concentrations of TRAIL in the supernatant were determined by an ELISA assay using a commercially available kit (R&D Systems, Minneapolis, MN). Cell lysates were prepared from the cells as described, and their protein content was quantitated by the Bradford assay.
Bile-Duct–Ligated Mouse Model of Liver Fibrosis.
The use and care of the animals were reviewed and approved by the Institutional Animal Care and Use Committee at the Mayo Clinic. C57/BL6 mice, 6 to 8 weeks of age, were used for these studies. Common bile duct ligation was performed as previously described in detail.24 Bortezomib (Mayo Clinic Formulary, Rochester, MN) was administered at 0.1 mg/kg body weight by intraperitoneal injection on day 4 after BDL. Bortezomib was dissolved in saline as a vehicle, and control animals received saline alone. On day 7 after bile duct ligation, animals were sacrificed and serum and liver samples were removed for analysis.
Immunolabeling was performed using formalin-fixed paraffin-embedded blocks (5-μm thick). Freshly cut sections were deparaffinized in xylene and rehydrated through sequential graded ethanol steps. Antigen retrieval was performed by incubating the slides in 1 mmol/L EDTA (pH 8.0) using a household vegetable steamer for 60 minutes. For detection of α-SMA immunoreactivity, slides were incubated with proteinase K (20 μg/mL) for 10 minutes. Slides were next incubated in 3% hydrogen peroxide for 10 minutes at room temperature and then incubated in blocking buffer (5% normal goat serum in PBS −0.05% Tween) for 30 minutes, also at room temperature. The samples were then placed in antisera blocking buffer overnight at 4°C. The slides were then incubated with α-SMA primary antibody (dilution 1:400; NeoMarkers, Fremont, CA). Nonimmune immunoglobulins were used as a negative control. After washing 3 times with PBS −0.05% Tween, the slides were incubated with the peroxidase-conjugated secondary antibody (DAKO Envision system, HRP [DAB] kit; DAKO Corporation, Carpinteria, CA) for 30 minutes at room temperature. Peroxidase activity was detected by incubating the samples with 3,3 DAB as the chromogen. Slides were subsequently counterstained with hematoxylin. To accurately quantitate immunostained positive cells, slides were viewed by digital microscopy (Axioplan 2; Carl Zeiss, Inc). Digital pictures were captured through a video archival system using a digital TV camera system (Axiocam High Resolution color, Carl Zeiss, Inc.). With an automated software analysis program (KS400, Carl Zeiss, Inc.), the percentage of immunostained/field areas of digital photomicrographs were quantified.
TUNEL and α-SMA Dual Staining.
The TUNEL assay was performed using a commercially available kit, following the manufacturer's instructions (In Situ Cell Death Detection Kit; Roche Diagnostics, Indianapolis, IN).25 After TUNEL staining, slides were incubated in blocking buffer (5% normal goat serum in PBS −0.05% Tween) for 30 minutes. The slides were next incubated with α-SMA primary antibody (dilution 1:400) as described in the previous pagragraph. After washing 3 times with PBS −0.05% Tween, the slides were incubated with the alkaline phosphatase-conjugated secondary antibody (Santa Cruz Biotechnology Inc.) for 30 minutes at room temperature. After washing 3 times, α-SMA positive cells were detected by fast blue (Sigma). TUNEL and α-SMA dual-positive cells were quantitated by counting the number of cells in 30 random microscopic high-power fields (40×).
MG132 and the I-κBα phosphorylation inhibitor Bay11-7082 were purchased from Calbiochem. Recombinant human TRAIL-R2: Fc was purchased from Alexis Corp.
All data are expressed as the mean ± S.E. unless otherwise indicated. Differences between groups were compared using two-tailed Student's t tests or chi-square test.
Do Proteasome Inhibitors Induce Apoptosis in HSCs?
We first determined if the proteasome inhibitors, bortezomib and MG132, can induce apoptosis in the human HSC line (LX-2) and primary rat HSCs. Both proteasome inhibitors induced apoptosis in a concentration-dependent manner in LX-2 cells (Fig. 1A). After 24 hours of exposure, the concentration-response curve was linear between 10 μmol/L and 50 μmol/L of inhibitor. The curvilinear response had an inflection point at 10 μmol/L for both inhibitors where significant apoptosis began to occur. Maximal apoptosis of 50% to 60% was observed at a concentration of 10 μmol/L. Cytotoxicity by both proteasome inhibitors was also time dependent at 10 μmol/L (Fig. 1B). Apoptosis increased over time and was >50% after 48 hours of exposure to the proteasome inhibitors. The occurrence of apoptosis was confirmed biochemically by demonstrating the disappearance of pro-caspase 3, consistent with its proteolytic activation (Fig. 1C), a biochemical hallmark of apoptosis. As a complementary assay of apoptosis, Annexin V-FITC labeling of externalized phosphatidylserine was also performed and demonstrated a similar rate of apoptosis by both proteasome inhibitors (Fig. 1D). Finally, both proteasome inhibitors also induced comparable levels of apoptosis in primary rat HSC (Fig. 1E). Thus, proteasome inhibitors induce apoptosis of activated mammalian HSCs.
Does Bortezomib Induce Apoptosis by Blocking NF-κB Activation?
To examine the NF-κB inhibitory effect of bortezomib in HSCs, the subcellular localization of the NF-κB RelA/p65 protein was assessed by immunocytochemistry and confocal microscopy (Fig. 2A). Nuclear fluorescence characteristics of translocated RelA/p65 protein were observed under basal condition consistent with the constitutive activated state of these cells.9 In contrast, RelA/p65 immunofluorescence was cytoplasmic and excluded from the nucleus in bortezomib-treated cells (Fig. 2A). Identical results were observed with the NF-κB inhibitor Bay11-7082 (data not shown). Next we performed immunoblot analysis for a NF-κB inhibitory protein, which complexes with NF-κB, preventing its nuclear translocation. Degradation of I-κBα is associated with NF-κB activation and conversely increases in this protein correlate with NF-κB inhibition.26 I-κBα protein levels increased in LX-2 cells treated with bortezomib consistent with NF-κB inhibition (Fig. 2B). To assess whether, indeed, NF-κB DNA binding activity was blocked by bortezomib in LX-2 cells, nuclear extracts were prepared from bortezomib-treated LX-2 cells and assayed for NF-κB DNA-binding activity (Fig. 2C). Nuclear extracts from control LX-2 cells demonstrated constitutive NF-κB DNA-binding activity, which was reduced by 75% after bortezomib treatment (Fig. 2C). Finally, we performed studies to ascertain whether specific inhibition of NF-κB is sufficient to induce apoptosis in LX-2 cells and primary rat HSC. Bay11-7082, an inhibitor of I-κBα phosphorylation,27 inhibited NF-κB DNA binding activity in LX-2 cells (Fig. 3A) and induced apoptosis in LX-2 cells (Fig. 3B) and primary rat HSC (Fig. 3C). Taken together, these data suggest that inhibition of NF-κB activity can induce HSC apoptosis.
Does Loss of Bcl-2 Proteins Mediate Proteasome Inhibitor-Induced HSC Apoptosis?
Apoptosis is regulated by the Bcl-2 family of proteins.28 The anti-apoptotic proteins, Bcl-xL and A1, are target genes for NF-κB.29 Moreover, NF-κB-dependent expression of A1 plays a key role in macrophage survival.30 Therefore, to determine if proteasome inhibitors regulate A1 expression in HSCs, A1 mRNA expression was examined by real-time PCR. Indeed, A1 mRNA expression was reduced by 50% not only by the proteasome inhibitors bortezomib and MG132, but also by selective NF-κB inhibitor Bay11-7082 (Fig. 4A). Unfortunately, because commercial antisera demonstrated nonspecific bands, the A1 knockdown could not be confirmed at the protein level. Transfection with A1 siRNA induced apoptosis in LX-2 cells (Fig. 4B) and primary rat HSC (Fig. 4C). In contrast to A1, BcL-xL protein expression was not reduced by bortezomib treatment (Fig. 4D). Bim, a pro-apoptotic Bcl-2 family member protein, can also be increased by proteasome inhibition.15, 16, 31 Therefore, we also examined Bim protein expression by immunoblot analysis following treatment of LX-2 cells with proteasome inhibitors (Fig. 5A-B). Both inhibitors increased cellular protein levels of these Bim isoforms. However, transfection with Bim siRNA, despite reducing cellular Bim expression, did not reduce bortezomib-induced apoptosis (Fig. 5C). Taken together, proteasome inhibitors by blocking NF-κB activation, reduce A1 cellular levels thereby causing apoptosis.
Does Bortezomib, Enhance TRAIL-R1/DR4, TRAIL-R2/DR5 and TRAIL/Apo2L Protein Expression?
We have recently demonstrated that activated HSC upregulate expression of the tumor necrosis factor related apoptosis inducing ligand (TRAIL) receptor-2 (TRAIL-R2).32 Furthermore, the activated HSC undergo enhanced TRAIL-mediated apoptosis. To examine if the TRAIL signaling pathway was involved in proteasome inhibitor-induced apoptosis of LX-2 cells, protein expression of TRAIL-R/DR4 and 5 and TRAIL/Apo2 ligand protein were examined by immunoblot analysis. Indeed, TRAIL-R2/DR5, but not TRAIL-R1/DR4, protein expression was increased by bortezomib treatment of LX-2 cells (Fig. 6A). Interestingly, we observed that TRAIL/Apo2 ligand protein was also increased in LX-2 cells following exposure to bortezomib (Fig. 6B). Cells treated with bortezomib or MG132 for 48 hours also released more TRAIL ligand into the culture media demonstrating that not only was the TRAIL protein increased, but that its secretion was also enhanced (Fig. 6C). However, when LX-2 cells were preincubated with recombinant human TRAIL-R2-Fc, which binds and neutralizes TRAIL, and then incubated with bortezomib, apoptosis was not reduced (Fig. 6D). To confirm this result, LX-2 cells were incubated with bortezomib and then transfected with DN-FADD, which inhibits the recruitment of FADD to the death receptor complex and blocks the TRAIL receptor signaling pathway. DN-FADD was also ineffective in reducing apoptosis during treatment of the cells with bortezomib (Fig. 6E). These data suggest that although bortezomib enhances TRAIL-R2/DR5 and TRAIL/Apo2L protein expression, its apoptosis induction is not TRAIL-dependent.
Does Bortezomib Reduce Activated HSCs and Hepatic Fibrosis in BDL Mice?
To investigate the effect of bortezomib on activated HSCs in vivo, liver specimens and total hepatic RNA were obtained from the sham operated and 7-day BDL mice treated with or without bortezomib. Bortezomib was administered on day 4 following the surgical procedure to permit activation of HSC prior to drug administration. The number of activated HSC present in the liver was assessed by immunohistochemistry for α-SMA (Fig. 7A) and quantitated using digital image analysis technology (Fig. 7B). α-SMA mRNA was quantitated using real-time PCR technology (Fig. 7C). α-SMA immunoreactivity was increased 9-fold in and its mRNA 10-fold in BDL mice as compared with sham-operated mice indicating a significant number of activated HSC. Bortezomib treatment on day 4 following BDL ligation reduced α-SMA immunoreactivity and mRNA expression by >80% and 70%, respectively (P < .05 and P < .01, respectively). Consistent with the reduction in activated HSC, collagen 1α (I) mRNA, a marker for stellate cell fibrogenic activity, was also reduced in bortezomib treated BDL versus untreated animals (Fig. 7D). To confirm that the reduction in α-SMA positive cells was owing to apoptosis, TUNEL and α-SMA dual staining assay was performed to document HSC apoptosis. Consistent with the in vitro observations, bortezomib treatment was associated with a significant increase in HSC apoptosis in the BDL mouse (Fig. 7E). Taken together, these data suggest that bortezomib reduces the number of activated HSC and liver fibrosis in vivo.
The principal findings of this study relate to mechanisms of apoptosis induction in HSCs, the principal, if not the only, source of collagen I in the liver.33 The results demonstrate that: (1) proteasome inhibitors induce apoptosis in activated primary rodent HSCs and LX-2 cells (human HSC cell line); (2) the proteasome inhibitor, bortezomib, inhibits NF-κB activation in LX-2 cells, causing a reduction in cellular levels of A1; (3) reduction of A1 induces LX-2 cell apoptosis; and (4) bortezomib reduces hepatic fibrosis in the BDL mouse. The implication of these observations is will be further discussed.
Apoptosis was induced in LX-2 cells and primary rat HSCs, not only by proteasome inhibitors but also by selective inhibition of NF-κB activation. These observations suggest that inhibition of NF-κB activation is sufficient to induce apoptosis in activated stellate cells. This concept is consistent with the prosurvival, anti-apoptotic effects of activated NF-κB.10 Indeed, numerous data now suggest that NF-κB inhibition is sufficient to eliminate HSC by apoptosis.34 Our current data extend these observations by demonstrating that blocking NF-κB activation by proteasome inhibition is an additional mechanism to induce HSC apoptosis in vitro and in vivo. Finally, the current observations demonstrating apoptosis in human, murine, and rat stellate cells by proteasome inhibitors further strengthen our findings by demonstrating that the results are species-independent.
Previous reports suggest that NF-κB activation induces anti-apoptotic genes such as Bcl-xL and A1.29, 35, 36 In this study, inhibition of NF-κB activation with either bortezomib or Bay11-7082, inhibited expression of A1, but not Bcl-xL, in LX-2 cells. Furthermore, targeted siRNA inhibition of A1 also induced LX-2 cell apoptosis. Taken together, these data suggest that activated HSC may be dependent upon A1 expression for survival. In this regard, activated HSC are similar to activated macrophages, which also require A1 expression for survival.30 Recent data suggest that, like macrophages, HSCs are also derived from hematopoietic stem cells.37–39 These seemingly divergent cellular phenotypes may, therefore, share a common origin, explaining their similar dependence on the A1 survival gene in the activated state.
BH3-only proteins such as Bim, Bad, Bid, Bik, PUMA, Hrk, Bmf, and Noxa have a critical role in the induction of cell apoptosis.40 Recent reports have demonstrated that bortezomib contributes to upregulation of Bim expression via inhibition of its proteasomal degradation and sensitizes tumor cells to apoptosis.15, 41 In this study, however, reduction of Bim expression by transfection of Bim siRNA did not decrease bortezomib-induced apoptosis in LX-2 cells. Recently, Kirschnek et al. demonstrated that upregulation of Bim expression by toll-like receptor-stimulation in macrophages is also not sufficient to induce apoptosis until a second stimulus is provided.42 Consistent with these data, the mere induction of Bim expression by proteasome inhibition may not be sufficient to induce LX-2 cell apoptosis.
We have previously reported that TRAIL R2/DR5 mRNA and protein expression increases during activation of LX-2 cells, thereby sensitizing the cells to TRAIL-mediated apoptosis.32 Consistent with its effects in other cell types,43–45 bortezomib enhanced TRAIL R2/DR5 expression in LX-2 cells. However, incubation with a TRAIL-R2-Fc, which binds and neutralizes TRAIL, did not attenuate apoptosis. DN-FADD transfection, which blocks death receptor–mediated apoptosis, also failed to reduce bortezomib-induced apoptosis. Therefore, neither a TRAIL autocrine nor paracrine process could be implicated in bortezomib-induced apoptosis of LX-2 cells.
Our studies in the BDL mouse demonstrate an in vivo correlate to our in vitro observations. Indeed, quantitative morphometry for α-SMA immunoreactivity (a marker for the number of activated HSCs) was reduced by bortezomib administration. Because bortezomib was administered 4 days after BDL to permit HSC activation, the loss of α-SMA immunoreactivity is most consistent with loss of HSCs, rather than inhibition of their activation. Indeed, HSC apoptosis, as assessed by the TUNEL assay, was increased in bortezomib-treated BDL mice, confirming this interpretation of the data. This reduction in activated HSCs was also accompanied by a decrease in a marker of hepatic fibrogenesis, collagen 1α(I) mRNA. In this regard, proteasome inhibition holds the promise of an antifibrogenic strategy. However, before such clinical studies are undertaken, the effect of proteasome inhibition on other cell types in the liver must be investigated. Potential inhibition of NF-κB activation in Kupffer cells would be anti-inflammatory and beneficial, whereas NF-κB inhibition in hepatocytes may inhibit survival pathways and promote liver injury. The net result of all these potential outcomes and the dominant effect of hepatic NF-κB inhibition will require further studies.
In conclusion, these studies have several major findings. The proteasome inhibitor bortezomib prevents HSC activation and reduces hepatic fibrogenesis in BDL mice. The well-documented stimulation of NF-κB, which occurs in activated HSCs, appears to play a critical role in their survival. Proteasome inhibition, by blocking NF-κB activation, induces activated HSC apoptosis. NF-κB–mediated survival is, in part, mediated by enhanced expression of A1. Inhibition of A1 by small molecules, as has been described for Bcl-2,46 is therefore a potential antifibrotic strategy. Also, given the clinical availability of bortezomib, it is a potential HSC targeted chemotherapeutic agent for the treatment of hepatic fibrosis.
The authors thank Erin Bungum for superb secretarial assistance.