Nonalcoholic fatty liver disease (NAFLD) is a serious health problem. Although NAFLD represents a form of lipotoxicity, its pathogenesis remains poorly understood. The aim of this study was to examine the cellular mechanisms involved in free fatty acid (FFA)-mediated hepatic lipotoxicity. FFA treatment of liver cells resulted in Bax translocation to lysosomes and lysosomal destabilization with release of cathepsin B (ctsb), a lysosomal cysteine protease, into the cytosol. This process was also partially dependent on ctsb. Lysosomal destabilization resulted in nuclear factor κB–dependent tumor necrosis factor α expression. Release of ctsb into the cytoplasm was also observed in humans with NAFLD and correlated with disease severity. In a dietary murine model of NAFLD, either genetic or pharmacological inactivation of ctsb protected against development of hepatic steatosis, liver injury, and insulin resistance with its associated “dysmetabolic syndrome.” In conclusion, these data support a lipotoxic model of FFA-mediated lysosomal destabilization. Supplemental material for this article can be found on the HEPATOLOGY website (http://www.interscience.wiley.com/jpages/0270-9139/suppmat/index.html). (HEPATOLOGY 2004;40:185–194.)
Obesity is associated with the so-called “dysmetabolic syndrome” characterized by insulin resistance, dyslipidemia, hypertension, and adverse health outcomes.1 This syndrome is also the main risk factor for development of nonalcoholic fatty liver disease (NAFLD), a common liver disease.2 NAFLD is a clinico-pathological syndrome with a wide spectrum of liver damage ranging from simple steatosis to steatohepatitis to advanced fibrosis and cirrhosis.3 The cellular mechanisms culminating in this syndrome remain unclear but are of great interest, because current therapies are limited and future therapies will be predicated by an understanding of its pathogenesis.
Tumor necrosis factor α (TNF-α) appears to play a central role in the development of hepatic steatosis.4 TNF-α, by mechanisms not completely defined, is overexpressed in the liver of obese mice and is an important mediator of insulin resistance in both diet-induced and ob/ob models of obesity.5–7 It is also required for alcohol to produce fatty liver in mice.7 The importance of TNF-α in NAFLD has been highlighted by the observations that anti–TNF-α therapy attenuates NAFLD in ob/ob mice,8 although the results in ob/ob mice are controversial.9 The mechanisms responsible for intrahepatic TNF-α generation in this syndrome remain unclear.
Free fatty acids (FFAs) appear to be important mediators of lipotoxicity, both as potential cellular toxins and by leading to lipid overaccumulation. When lipids overaccumulate in nonadipose tissue, they may enter nonoxidative deleterious pathways leading to cell injury and death.10–12 In humans with NAFLD, not only are circulating FFAs elevated, but their levels also correlate with disease severity.13 A single in vitro study in endothelial cells and cardiac myocytes has suggested that accumulation of FFAs may result in lysosomal permeabilization.14 Furthermore, lysosomal breakdown with cathepsin B (ctsb) release into the cytosol is a feature of TNF-α signaling cascades.15 The role of this potential lysosomal permeabilization in lipotoxicity, proinflammatory gene expression, and NAFLD pathogenesis has not been explored. Thus, the overall objective of this study was to determine the cellular mechanisms by which FFAs contribute to the pathogenesis of NAFLD.
NAFLD, nonalcoholic fatty liver disease; FFA, free fatty acid; ctsb, cathepsin B; TNF-α, tumor necrosis factor α; ctsb−/−, cathepsin B knockout mouse; ctsb-GFP, cathepsin B green fluorescent protein; ctsb-RFP, cathepsin B red fluorescent protein; Bax-GFP, Bax green fluorescent protein; I-κB, inhibitor κB; PCR, polymerase chain reaction;TNFR1−/−, tumor necrosis factor receptor 1 knockout mouse; ctsb−/−, cathepsin B wild type mouse; TNFR1+/+, tumor necrosis factor receptor 1 wild type mouse; ALT, alanine aminotransferase; mRNA, messenger RNA; NF-κB, nuclear factor κB; IKK-β, I-κB kinase β.
Materials and Methods
Cell Isolation and Culture.
Mouse hepatocytes from cathepsin B knockout mice (ctsb−/−) and littermate wild type mice were isolated and cultured as described previously.15 HepG2 cells were cultured as described previously.16 Mouse hepatocytes and HepG2 cells were incubated with or without 0.5-mM or 1-mM long-chain FFAs (2:1 oleate/palmitate, Sigma, St. Louis, MO) in media containing 1% bovine serum albumin. FFA levels in plasma are usually 0.2–2 mM.17, 18
Transient Transfection and Confocal Microscopy.
The ctsb green fluorescent protein (ctsb-GFP) expression vector, ctsb red fluorescent protein (ctsb-RFP) expression vector, and Bax green fluorescent protein (Bax-GFP) expression vector were transfected into HepG2 cells using lipofectamine.15 Confocal microscopy was performed with an inverted Zeiss Laser Scanning Confocal Microscope (Zeiss LSM 510, Carl Zeiss, Inc., Thornwood, NJ). Cells were counted in 30 random microscopic fields (magnification ×400) as described previously by us in detail.16
Apoptosis in cell culture was quantified by assessing the characteristic nuclear changes of apoptosis using the nuclear binding dye DAPI (Molecular Probes Inc., Eugene, OR) and fluorescent microspcopy.16
Measurement of Intracellular ctsb Activity.
Intracellular ctsb activity was assayed using the Magic Red detection kit (Immunochemistry Technologies, LLC, Bloomington, MN). According to the manufacturer's instructions, a 1× solution was prepared in Dulbecco's Modified Eagle Medium and 500 μL was added to each 35-mm well. Following a 30- to 60-minute incubation at 37°C, cells were rinsed twice with phosphate-buffered saline. Fluorescence was monitored at the single cell level using an inverted fluorescent microscope, and images were collected using a cooled, charge-coupled device camera (Photometerics, Tucson, AZ) and digitized by Metafluor software (Universal Imaging, Westchester, PA). Activity of ctsb was expressed as an increase in total cellular fluorescence (pixel number per average fluorescence intensity) per minute per cell.
Nile Red Staining.
Cells were incubated with or without 1 mM FFA for 24 hours, fixed with 3.7% paraformaldehyde, incubated for 5 minutes with 0.1 μm/mL Nile red (a hydrophobic dye that accumulates in lipid droplets) in phosphate-buffered saline, and examined by digitized fluorescent microscopy. Nile Red staining was expressed as an increase in total cellular fluorescence (pixel number per average fluorescence intensity) per cell.
Adenoviral Transduction With the Inhibitor κB (I-κB) Superrepressor.
The recombinant replication-deficient adenovirus Ad5I-κB superrepressor (a generous gift from Dr. David Brenner, Columbia University, New York, NY) and Ad5ΔE1 were grown and purified by banding twice in CsCl gradients as described previously.15, 19 Six hours after hepatocyte isolation, the medium was changed to Dulbecco's Modified Eagle Medium containing 2% fetal bovine serum, and Ad5I-κB or Ad5ΔE1 viral stock solutions were added to the medium at a multiplicity of infection of 100. After 12 hours, cells were incubated in the presence or absence of FFAs or TNF-α.
Cells grown on glass coverslips were fixed with 4% paraformaldehyde. Cells were permeabilized with 0.5% Triton-X and block with phosphate-buffered saline containing 1% bovine serum albumin. The cells were next incubated with a polyclonal rabbit anti–65 immunoglobulin G (sc-372, Santa Cruz Biotechnology, Santa Cruz, CA) in blocking buffer at 4°C overnight. After washing, Cy3-conjugated goat anti–rabbit Ig (Jackson ImmunoResearch Labs, West Grove, PA) in blocking buffer was added. Cells were imaged by confocal microscopy and counted in 30 random microscopic fields (magnification ×400) as described previously by us in detail.16
Immunoblot analysis was performed using whole cell lysates as described previously.20 Samples were resolved by 12% SDS–polyacrylamide gel electrophoresis, transferred to nitrocellulose membrane, and blotted with appropriate primary antibodies. The membrane was incubated with peroxidase-conjugated secondary antibodies (1:10,000 dilution, Biosource International, Camarillo, CA) and the bound antibody was visualized using a chemiluminescent substrate (ECL, Amersham, IL) and Kodak X-OMAT film (Eastman Kodak, Rochester, NY). Primary antibodies were goat anti–I-κB (1:500 dilution, Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-Bax (1:1000 dilution, Santa Cruz Biotechnology), rabbit anti-Bak (1:1000 dilution, Santa Cruz Biotechnology), goat anti–lysosomal-associated membrane protein 1 (1:500 dilution, Santa Cruz Biotechnology), mouse anti–cytochrome c oxidase (1:1000 dilution, Molecular Probes, Inc.) and goat anti-actin (1:2000 dilution, Santa Cruz Biotechnology).
Quantitative Real-Time Polymerase Chain Reaction.
Total RNA was isolated from liver tissue and HepG2 cells using the Trizol reagent (Invitrogen, Carlsbad, CA), and complementary DNA was prepared using an oligo-dT primer and Moloney murine leukemia virus reverse transcriptase.21 The complementary DNA product was amplified using polymerase chain reaction (PCR) with Thermus aquaticus DNA polymerase using standard protocols.21 PCR primers were as follows: human TNF-α, 5′-AGCCCATGTTGTAGCAAACC-3′ and 5′-GGA AGACCCCTCCCAGATAG-3′; mouse TNF-α, 5′ACGTGGAACTGGCAGAAGAG-3′ and 5′-CTCCTCCACTTGGTGGTTTG-3′. Primers for 18S ribosomal RNA (Ambion Inc, Austin, TX) were used as a control for RNA integrity. After electrophoresis in 1% low–melting temperature agarose gel, the expected base pair PCR products were identified and the bands were cut from the gel. Next, the PCR products were eluted into Tris-HCl using a DNA elution kit (Qiagen, Valencia, CA), and serial dilutions performed for a standard curve. Real-time PCR was performed with the Roche Light Cycler using SYBR green as the fluorophore (Molecular Probes, Inc.).21 All PCR conditions and primers were optimized to produce a single product of the correct base pair size.
These experimental protocols were approved by the Institutional Animal Care and Use Committee at the Mayo Clinic. C57BL/6 ctsb knockout (ctsb−/−) mice and TNF receptor 1 knockout (TNFR1−/−) and wild type (ctsb+/+ and TNFR1+/+) mice (20–25 g) were fed a high-carbohydrate diet ad libidum (consisting of 65% sucrose, 20% casein, 5% corn oil, 4% mineral mixture, 1% vitamin mixture, and 1% orotic acid, Teklad Mills, Madison, WI) for 16 weeks starting at 6 to 8 weeks of age (n = 15 in each group). Identical groups of animals (n = 15 in each group) received standard rodent chow to act as controls.
Pharmacological ctsb Inhibition.
In selected experiments, after 15 weeks on the diet, mice were treated with the ctsb inhibitor R-3032 (4.9 mg/mL in 30/70 polyethyleneglycol 400/H2O at pH 7.5) intraperitoneally twice a day for 1 week. This agent, a selective reversible ctsb inhibitor, was obtained from the Celera Cooperation (San Francisco, CA).
Under deep anesthesia, blood samples and liver tissue were collected after a 7-hour fast as described previously by us in detail.22 The liver tissue was fixed in 4% paraformaldehyde and embedded in Tissue Path (Fisher Scientific, Pittsburgh, PA). Hematoxylin-eosin–stained liver specimens were evaluated using light microscopy. Histology specimens were reviewed in a blinded manner and graded according to the degree of steatosis using a standard 3-point scale (<33%, 1; >33% and <66%, 2; >66%, 3) as we described previously.23
Human Studies and ctsb Immunofluorescence.
The study was approved by the Mayo Institutional Review Board. Liver samples from 35 untreated consecutive patients with nonalcoholic steatohepatitis and normal livers from 15 patients undergoing hepatic resection for colorectal metastasis were obtained. The sections were prepared by standard procedure and were blocked for 60 minutes in 37°C in blocking buffer (5% goat serum, 5% glycerol, 0.004% sodium azide). The specimens were next incubated for 2 hours with a goat polyclonal anti–ctsb antibody (1:100) (C-19, Santa Cruz Biotechnology), washed in phosphate-buffered saline three times for 10 minutes, and incubated with the secondary antibody (1:1000) fluorescein-isothiocyanate–conjugated swine anti–goat antibody immunoglobulins (Santa Cruz Biotechnology) at 37°C for 45 minutes. The slides were mounted with a ProLong antifade kit (Molecular Probes, Inc.). Livers were imaged using confocal microscopy and cells were counted according to their diffuse or punctuate appareance in 30 random microscopic fields (magnification ×400) as described previously by us in detail.16
Under deep anesthesia, blood samples were collected after a 7-hour fast.16 Blood glucose concentrations, serum alanine aminotransferase (ALT), serum insulin, leptin, and FFAs were measured as described previously.16 TNF-α levels were measured using an enzyme-linked immunosorbent assay kit (R&D Systems, Inc., Minneapolis, MN).
All data are expressed as the mean ± SD unless otherwise indicated. Differences between groups were compared using ANOVA followed by a post hoc Bonferroni test to correct for multiple comparisons.
HepG2 cells incubated with a mixture of FFA (2:1 oleate/palmitate) developed cellular “steatosis” as we described previously.16 To assess if FFA-treated cells manifest lysosomal permeabilization with ctsb translocation into the cytosol, the cellular compartmentation of ctsb-GFP was examined using confocal microscopy. In untreated cells transfected with ctsb-GFP, fluorescence was punctuated consistent with lysosomal localization (Fig. 1A). In contrast, in the “steatotic” FFA-treated cells, ctsb-GFP was diffusely distributed throughout the cytoplasm in more than 80% of cells (Fig. 1B). Modest rates of apoptosis were observed in the diffusely fluorescent cells, as assessed by DAPI staining and fluorescent microscopy; apoptosis rates were 12 ± 0.5% versus 0.9 ± 0.4% in controls (P < .01). The cytosolic ctsb was catalytically active, because ctsb activity was 76% greater in FFA-treated cells than in controls (Fig. 1C). These observations are consistent with lysosomal permeabilization with ctsb translocation into the cytoplasm during FFA exposure.
Next, to ascertain if ctsb is mechanistically linked to the FFA-induced lysosomal permeablization, ctsb-GFP transfected cells were incubated with FFAs alone or in combination with the selective ctsb pharmacological inhibitor R-3032 and scored as either demonstrating punctuate (lysosomal) or diffuse (cytosolic) fluorescence. Incubation with R-3032 significantly decreased the number of FFA-treated cells with a diffuse appearance (Fig. 1D), indicating that FFA-induced lysosomal permeabilization in HepG2 cells is partially dependent on ctsb. However, these observations are tempered by the fact that the inhibitor also reduced neutral lipid accumulation in the cells (Fig. 1E). To further examine the mechanisms of ctsb lysosomal release, we assessed whether or not there were common pathways between mitochondrial and lysosomal dysfunction. Permeabilization of the outer mitochondrial membrane is often associated with Bax (a Bcl-2 protein that induces channels in membranes) translocation from cytosol to mitochondria.24 The cellular compartmentation of Bax was examined over time by imaging Bax-GFP in HepG2 cells. Under basal conditions, Bax-GFP was heterogenously distributed within the cell, displaying a mixed cytosolic and organelle localization; however, following exposure to FFAs, Bax-GFP became more punctate and partially colocalized with ctsb-RFP (Fig. 1F). The colocalization preceded ctsb-RFP release into the cytosol (Fig. 1F and 1G), consistent with a role for Bax in this permeabilization process. To confirm these findings in vivo, we performed immunoblots for Bax and Bak in lysosomes isolated from mice fed a standard rodent diet (lean mice) or a high-carbohydrate diet associated with elevated serum FFAs and hepatic steatosis.16 Interestingly, while Bak was present in lysosomes from both groups of mice, Bax was only found in lysosomes from obese animals (Fig. 1H). Collectively, these data suggest that the FFA-mediated lysosomal permeabilization process involves Bax translocation to lysosomes and ctsb.
FFAs Induce TNF-α Expression in a ctsb-Dependent Manner.
Next, we tested the hypothesis that FFAs induce TNF-α expression. TNF-α messenger RNA (mRNA) increased more than 100-fold in FFA-treated HepG2 cells versus untreated cells (Fig. 2A). Because FFAs induce ctsb-dependent lysosomal destabilization, the relationship between lysosomal permeabilization and TNF-α expression was assessed. Primary hepatocytes from both wild type and ctsb−/− mice were incubated in the presence or absence of FFAs and TNF-α mRNA transcripts measured by real-time PCR. As with HepG2 cells, following incubation with FFAs, TNF-α mRNA increased more than 60-fold in wild type hepatocytes. In contrast, this increase in TNF-α expression did not occur in ctsb−/− hepatocytes and was abrogated in wild type hepatocytes treated with R-3032 (Fig. 2B). These data strongly implicate ctsb-mediated lysosomal permeabilization in FFA-associated TNF-α expression.
To ascertain if nuclear factor κB (NF-κB) contributes to FFA-associated TNF-α expression, NF-κB activation was inhibited with the superrepressor of I-κB. HepG2 cells were infected using the adenovirus vector Ad5I-κB superrepressor or the empty virus Ad5ΔE1 with or without FFAs. FFA-associated TNF-α expression was blunted in cells treated with the I-κB superrepressor but not in those infected with the empty vector (Fig. 2C). To assess if indeed NF-κB activation occurred in FFA-treated cells, the subcellular localization of the NF-κB RelA/p65 protein25 was assessed in HepG2 cells. Nuclear fluorescence characteristics of activated and translocated RelA/p65 protein were observed in more than 80% of cells treated with FFAs but were not present in either control cells or cells treated with FFAs plus R-3032 (Fig. 2D).
Next, we performed immunoblot analysis for I-κBα, a NF-κB–inhibitory protein.19 HepG2 cells were incubated in the absence (negative control) or presence of TNF-α (positive control), FFAs, or FFAs plus R-3032. As in cells exposed to TNF-α, FFA-treated cells manifested a significant decrease in cellular I-κBα protein levels (Fig. 2E). In contrast, this was not observed in cells treated with FFAs plus R-3032. Taken together, these data strongly indicates that ctsb-dependent lysosomal permeabilization may induce TNF-α expression through activation of NF-κB.
Redistribution of ctsb Into Cytosol Is Present in Human NAFLD and Correlates With Disease Severity.
To ascertain if human NAFLD is also associated with lysosomal permeabilization, immunofluorescence for ctsb was performed in livers of patients with NAFLD and in normal livers. In liver specimens from control patients, ctsb immunofluorescence was predominantly punctate, consistent with lysosomal localization; in contrast, the majority of hepatocytes in liver tissue from patients with NAFLD displayed diffuse immunofluorescence consistent with a redistribution of this protease from lysosomes to the cytosol (Fig. 3A and 3B). Furthermore, cytosolic localization of ctsb was significantly increased in patients with moderate to severe steatohepatitis compared with those with only mild inflammatory activity (Fig. 3C). Similar results were obtained in a murine model of fatty liver disease, as described below (Fig. 3D). These data confirm in human NAFLD and an in vivo murine model of fatty liver disease that hepatic steatosis is associated with lysosomal permeabilization and release of ctsb into the cytoplasm.
Inactivation of ctsb Protects Against Diet-Induced Fatty Liver Disease.
To investigate the role of ctsb in diet-associated fatty liver disease, ctsb−/− mice and their wild type controls were placed on a high-carbohydrate diet for 4 months. On this diet, both ctsb−/− and wild type mice developed marked obesity compared with control mice fed a standard rodent diet (Supplemental Fig. 1A). However, gross examination of the livers from obese wild type mice showed marked hepatomegaly relative to those of obese ctsb−/− and lean animals; in contrast, livers from obese ctsb−/− mice appeared grossly normal and similar to lean wild type and ctsb−/− animals (Supplemental Fig. 1B). The average liver weight and liver/body weight ratios from obese wild type mice were significantly greater compared with that of obese ctsb−/− mice, which was similar to the lean wildtype and ctsb−/− animals (Table 1). Histological examination of the liver tissue showed marked diffuse macrovesicular steatosis in obese wild type mice (Supplemental Fig. 1C and 1D); in contrast, no steatosis was present in lean wild type and ctsb−/− mice, and minimal steatosis was observed in obese ctsb−/− animals (see Supplemental Fig. 1D). Treatment of wild type mice on the high-carbohydrate diet for 16 weeks with addition of R-3032 during the last 7 days also virtually eliminated steatosis from the livers, while the body weight remained similar to that of nontreated obese mice (see Supplemental Fig. 1C and 1D). These R-3032–related improvements in hepatic histology were also associated with decreased serum ALT values, a marker of liver injury (Supplemental Fig. 1E).
Table 1. Tissue Weight in ctsb−/−, TNFR1−/−, and Wild-Type Littermates After 16 Weeks on Either High-Carbohydrate or Control Diet
Body Weight (g)
Liver Weight (g)
Liver/Body Weight (%)
NOTE. Values are mean ± SD (n = 5–7/group).
Abbreviation: NS, not significant.
27.9 ± 1.2
1.3 ± 0.2
4.7 ± 0.1
36 ± 1.5*
2.7 ± 0.1*
7.5 ± 0.3*
*P value vs. lean
36.6 ± 1.6†
1.5 ± 0.1†
4.1 ± 0.3†
†P value vs. lean
25.1 ± 1.1
1.2 ± 0.1
4.8 ± 0.1
35.6 ± 1.5‡
1.6 ± 0.1‡
4.6 ± 0.2‡
‡P value vs. lean
26.3 ± 1.3
1.2 ± 0.2
4.6 ± 0.1
35.9 ± 0.9§
1.4 ± 0.1§
3.9 ± 0.3§
§P value vs. lean
The wild type animals also developed the dysmetabolic syndrome observed in humans with NAFLD. Specifically, they manifested increased serum FFA, insulin, leptin, and TNF-α values (Table 2). These alterations were not observed in ctsb−/− mice, nor in wild type mice treated with R-3032. Thus, either pharmacological or genetic inactivation of ctsb protects from fatty liver disease and the obesity-associated dysmetabolic syndrome.
Table 2. Fasting Blood Measurements in ctsb−/−, TNFR1−/−, and Wild-Type Littermates After 16 Weeks on Either High-Carbohydrate or Control Diet
NOTE. Values are mean ± SD (n = 5–7/group).
Abbreviation: NS, not significant.
2.2 ± 0.2
1 ± 0.4
160 ± 6
509 ± 10
54.7 ± 9
23.4 ± 1
38.2 ± 5*
3.7 ± 0.3*
223 ± 12*
681 ± 4.2*
49 ± 8*
57.4 ± 2*
*P value vs. lean
18.9 ± 4†
2.3 ± 0.1†
178 ± 5†
542 ± 10†
65 ± 10†
28.5 ± 3†
†P value vs. obese
3.2 ± 0.2
0.8 ± 0.3
147 ± 3
515 ± 3.2
58 ± 2.9
27 ± 5
6.6 ± 1‡
2.1 ± 0.4‡
150 ± 18‡
591 ± 19‡
71 ± 25‡
35 ± 2‡
‡P value vs. obese wild type
2.9 ± 0.2
0.7 ± 0.3
155 ± 5
535 ± 7
59 ± 6
6.9 ± 0.7§
1.7 ± 0.4§
145 ± 7§
511 ± 4.5§
75 ± 3.7§
§P value vs. obese wild type
TNFR1 Is Essential for Hepatic Steatosis.
In obese wild type mice, TNF-α mRNA expression was eightfold greater than values from obese ctsb−/− and lean animals; moreover, treatment of wild type mice on the sucrose diet for 16 weeks with R-3032 during the last 7 days completely suppressed the increase in TNF-α expression (Supplemental Fig. 1F). These findings, coupled with the elevated serum TNF-α levels observed in wild type diet-fed animals (see Table 2), support a role for TNF-α in dietary hepatic steatosis. To more directly test this concept, TNFR1−/− mice were fed the high-carbohydrate diet. Although these mice became obese (see Table 1), they neither manisfested hepatic steatosis nor the associated metabolic syndrome (Fig. 4–4C; see Table 2). These data are compatible with the salutary benefits of anti–TNF-α therapy in genetic models of hepatic steatosis8 and strongly support a critical role for ctsb-induced TNF-α expression in hepatic steatosis syndromes.
Following treatment of cells with FFAs in vitro, lysosomal permeabilization with redistribution of ctsb from lysosomes to the cytosol was observed. The lysosomal permeabilization was associated with Bax translocation to lysosomes. Bax is a predominantly cytosolic protein that, upon certain stimuli, translocates to membranes, inducing channel formation. Bax effects on mitochondria in this regard have been extensively studied.24 In the current study, our data suggest Bax can also translocate to lysosomes and induce membrane permeabilization in this organelle. Thus, the cell appears to have consensus mechanisms for causing organelle permeabilization. Consistent with other models of lysosomal permeabilization, the lysosomal cysteine protease also appears to contribute to loss of lysosomal integrity. The potential interaction between Bax and ctsb in this process will require further investigation. The in vitro observations can be extrapolated to in vivo hepatic steatosis, because a redistribution of ctsb from lysosomes to the cytosol was observed in human liver biopsy specimens from patients with NAFLD. In the diet-associated fatty liver disease model, ctsb release into the cytosol and liver injury with elevated serum ALT values were also documented. Thus, the in vitro observations appear to be germane to obesity-related liver injury.
By a ctsb-dependent process, FFAs markedly stimulated TNF-α expression. TNF-α expression was NF-κB–dependent, because a superrepressor of I-κB blocked TNF-α up-regulation. Two distinct pathways of NF-κB activation have been convincingly elucidated. The so-called “canonical” processing pathway uses I-κB kinase β (IKK-β) to phosphorylate I-κB triggering its ubiquination-dependent degradation, resulting in activation of RelA:p50.19 In contrast, the “noncanonical” pathway involves IKK-α, which phosphorylates p100, resulting in its processing and generation of a RelB:p52-active complex.19 Our finding that IKK-β was reduced—and therefore likely degraded—following treatment of the cells with FFAs is consistent with activation of the canonical pathway. This activation was ctsb-dependent, because it was blocked by a pharmacological ctsb inhibitor. The precise mechanisms by which ctsb-dependent lysosomal permeabilization results in activation of this pathway will require further study.
TNF-α is an important mediator of insulin resistance in both humans and rodents.5, 26–28 Our current studies provide insight into how FFAs induce TNF-α generation. These results implicate lysosomal permeabilization and release of a lysosomal protease, namely ctsb into the cytosol in FFA-induced TNF-α expression. The model emanating from these studies (Fig. 4D) is one in which FFAs cause Bax translocation to lysosomes and lysosomal destabilization with ctsb redistribution from these vesicles into the cytosol. Once in the cytosol, ctsb triggers a cascade culminating in NF-κB activation with TNF-α generation. TNF-α promotes triglyceride accumulation and hepatic steatosis, perhaps in a paracrine or autocrine manner. TNF-α can promote insulin resistance by signaling IKK-β activation and c-jun-N-terminal kinase activation.29, 30 TNF-α may also further promote lysosomal destabilization, resulting in a feed-forward, self-perpetuating pathway, further accentuating liver injury. Supporting this model are the in vivo studies demonstrating that in a dietary murine model of NAFLD, which closely resembles NAFLD in humans, the development of fatty liver disease is dependent on ctsb and TNFR1, identifying a new potential target for treatment of NAFLD—namely ctsb.
The excellent secretarial help of Erin Bungum is gratefully acknowledged. The Bax-GFP plasmid was obtained from Anna-Liisa Nieminen, Case Western Reserve University, Cleveland, OH.