Cathepsin B deficiency ameliorates liver lipid deposition, inflammatory cell infiltration, and fibrosis after diet-induced nonalcoholic steatohepatitis

Abstract

Non-alcoholic steatohepatitis (NASH) is a severe form of non-alcoholic fatty liver disease characterized by fat accumulation and inflammation in liver. Yet, the mechanistic insight and diagnostic and therapeutic options of NASH remain incompletely understood. This study tested the roles of cysteine protease cathepsin B (CatB) in mouse NASH development. Immunoblot revealed increased liver CatB expression in NASH mice. Fructose-palmitate-cholesterol diet increased body weight gain, liver to body weight ratio, blood fasting glucose, plasma total cholesterol and alanine transaminase levels, and liver triglyceride, but decreased plasma high-density lipoprotein in wild-type mice. All these changes were blunted in CatB-deficient (Ctsb^−/−) mice. In parallel to reduced expression of genes involved in liver lipid transport and lipogenesis, liver CD36, FABP4, and PPARγ protein levels were also significantly decreased in Ctsb^−/− mice, although CatB deficiency did not affect liver gluconeogenesis and fatty acid beta-oxidation-associated gene expression. Mechanistic studies showed that CatB deficiency decreased liver expression of adhesion molecules, inflammatory cytokine, and chemokine, along with reduced liver inflammatory cell infiltration. CatB deficiency also promoted M2 macrophage polarization and reduced liver TGF-β1 signaling and fibrosis. Together, CatB deficiency improves liver function in NASH mice by suppressing de novo lipogenesis and liver inflammation and fibrosis.

INTRODUCTION

Increased caloric intake and decrease in physical activity lead to global obesity. Obese patients develop metabolic syndrome that is a significant risk factor of many chronic diseases including type 2 diabetes, lipid disorders, cardiovascular diseases, and non-alcoholic fatty liver disease (NAFLD)¹. NAFLD involves several liver disease stages, ranging from lipid accumulation (fatty liver), inflammatory cell infiltration and liver fibrosis referred as non-alcoholic steatohepatitis (NASH), to more advanced liver diseases such as NAFLD-associated cirrhosis and hepatocellular carcinoma in the absence of alcohol intake^2–9. The pathogenesis of NAFLD and its progression are multifactorial and influenced by environmental and genetic factors^1,7,10. Current understanding regarding the mechanisms leading to liver inflammation and fibrosis is limited. Prior studies support a role of lipid disorders that comes from insulin resistance and associated cytokine profile alterations¹¹. Recent evidences suggest that the oxidative stress and associated mitochondrial dysfunction, endoplasmic reticulum (ER) stress, lysosomal cathepsin release, and cell apoptosis together exacerbate inflammation and collagen deposition within the hepatic sinusoids¹².

Cathepsin B (CatB), a ubiquitously expressed cysteinyl cathepsin is a lysosomal protease that has implicated in multiple chronic human diseases, including NAFLD, liver fibrosis, atherosclerosis, Alzheimer’s disease, and cancer^5,9,11,13. Previous studies reported high serum levels of CatB and cathepsin L (CatL) in patients with cirrhosis and hepatocellular carcinoma^13–15. Accumulative evidences suggest a role of CatB in TGF-β1-dependent myofibrogenesis and subsequent regulation of extracellular matrix collagen and fibronectin deposition in murine models of lung or liver fibrosis^16–18. In carbon tetrachloride CCl₄-induced liver inflammation, CatB expression is increased in hepatic stellate cells but not in hepatocytes, and its inactivation mitigated liver inflammation, stellate cells activation, and collagen deposition¹⁹. CatB deficiency down-regulates the expression of tumor necrosis factor-α (TNF-α) and transcription inhibitor actinomycin D-induced mitochondrial release of cytochrome c and caspase activation that contribute to hepatocytes apoptosis²⁰. Yet, most of these observations were from in vitro cultured hepatocytes, hepatic stellate cells, cell lines, or from CCl₄-induced mice model of liver fibrosis¹⁹. Several reports have revealed considerable disadvantages from this chemical-induced liver fibrosis, regardless the different routes of CCl₄ administration, intraperitoneal or subcutaneous injection, inhalation, or gavage. Side effects include chronic peritonitis, respiratory arrest, necrosis at site of injection, and high mortality with inconsistent liver fibrosis²¹. In contrast, fructose-palmitate-cholesterol (FPC) diet-induced weight-gain model of NASH recapitulates many features of human metabolic syndrome and NASH with progressive fibrosis²². FPC diet is rich in fructose, palmitate, cholesterol, and trans-fat, together with other components as listed in Supplemental Table 1 and mimics the metabolic profile of NAFLD found in humans, including overweight and insulin resistance along with increased liver inflammation, fibrosis, and endoplasmic reticulum (ER) stress. Therefore, this diet-induced NASH model shares significant physiological characters to those of human NASH, although it takes about 16 weeks to develop all these histological changes²².

To study the role of CatB in hepatic lipid metabolism, inflammation, and fibrosis initiation, we produced NASH by challenging the CatB-deficient Ctsb^−/− mice and their littermate controls (Ctsb^+/+) with a FPC diet in this study. Mitigated NASH in Ctsb^−/− mice supports a possible therapeutic potential by targeting human CatB.

MATERIALS AND METHODS

Mice and NASH model

Male Ctsb^−/− mice and Ctsb^+/+ littermates (C57BL/6J) mice²³ were housed in 12-hour light/dark cycle and allowed to free food and water. At the age of 9 weeks, mice were split in two groups of each and fed with a low fat diet (LFD) or a FPC diet (components listed in Supplemental Table 1, Envigo. Madison, WI) for 6 and 18 weeks to produce NASH. Body weight and food intake were measured weekly. Each group used 10~15 mice. All mouse procedures conformed to the Guideline for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health and were approved by the Brigham and Women’s Hospital Standing Committee on Animals (protocol #2016N000442). At harvest, mice were euthanized and their liver, epididymal adipose tissue (EAT), subcutaneous adipose tissue (SAT), and brown adipose tissue (BAT) were excised. Fat and body weights were recorded immediately after isolation. Livers were embedded in OCT for 6 μm frozen section preparation or frozen in −80 °C refrigerator for mRNA determination and immunoblot analysis.

Real-time polymerase chain reaction (RT-PCR)

RT-PCR was performed to determine the mRNA levels of Ppar-α (peroxisome proliferator activated receptor-α), Cpt1α (carnitine palmitoyltransferase-1α), Acc1 (acetyl-CoA carboxylase), Acc2, Glut2 (glucose transporter-2), Pepck (phosphoenolpyruvate carboxykinase), CD36, CETP (cholesteryl esters and triglycerides transporter), Ppar-γ (peroxisome proliferator activated receptor-γ), Fabp4 (fatty acid binding protein 4), Scad1 (short-chain acyl dehydrogenase-1), Col1α1, Col1α2, Col3α1, Col4α1, αSMA (α-smooth muscle actin), Tgfβ1 (transforming growth factor-β1), Ctsb (CatB), Ctsk (CatK), Ctsl (CatL), Ctss (CatS), Mmp2, Mmp8, Mmp9, Mmp13, Tnfα, Mcp-1 (monocyte chemoattractant protein-1), Icam-1 (intercellular adhesion molecule-1), E-selectin, Arg1 (arginase-1), iNOS (inducible nitric oxide synthase) and F4/80 in liver tissues. Total RNA was extracted from liver tissues using the TRIzol reagent (Invitrogen, Carlsbad, CA) and treated with RNase-free DNase (Ambion, Austin, TX) to remove genomic DNA contaminants. Equal amounts of RNA were reverse-transcribed, and quantitative PCR was assessed in SYBR green RT-RCR detection system (Stratagene, La Jolla, CA). The mRNA level of above genes were normalized to β-actin. All primer sequences were listed in Supplemental Table 2.

Immunoblot

To detect CatB, CatK, CatL, CatS, CD36, PPAR-γ, FABP4, phosphorylated-AMPKα (p-AMPKα), AMPKα, phosphorylated-hormone-sensitive lipase (p-HSL), HSL, collagen-III, TGF-β1, lysyl oxidase, phosphorylated-Smad-2/3 (p-Smad-2/3), and Smad-2/3 expression in mouse liver tissue extract, equal amounts of protein from each mouse were separated on a 10% or 12% SDS-PAGE and transferred onto a PVDF membrane (Millipore, Billerica, MA), blotted, and detected with antibodies against CatB (1:1000, PC41, Sigma-Aldrich, St. Louis, MO), CatK (1:1000, PB9856, Bosterbio, Pleasanton, CA), CatL (1:1000, P07154, RayBiotech Inc, Peachtree Corners, GA), CatS (1:200)²⁴, importin-β (1:1000, NB120–2811, Novus Biologicals, Centennial, CO), Ranbp3 (1:1000, NB100–74647, Novus Biologicals), CD36 (1:250, ab133625, Abcam, Cambridge, MA), PPAR-γ (1:200, sc7196-A3113, Santa Cruz Biotechnology, Dallas, Texas), FABP4 (1:500, MAB3150, R&D systems, Minneapolis, MN), p-AMPKα (1:1000, 4185S, Cell Signaling Technology, Danvers, MA), AMPKα (1:1000, 2532, Cell Signaling Technology), p-HSL (1:1000, Cat# 4139S, Cell Signaling Technology), HSL (1:1000, Cat# 4107S, Cell Signaling Technology), collagen-III (1:1000, PA5–27828, Invitrogen), TGF-β1 (1: 1000, ab64715, Abcam), lysyl oxidase (1:1000, ABT112, Sigma-Aldrich), p-Smad-3 (1: 1000, ab52903, Abcam), p-Smad-2 (1:1000, 3108S) and Smad-2 (1:1000, 5339s), Smad-3 (1:1000, 9523s), β-Actin (1:3000, 4970S), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH, 1:3000, 2118s), all from Cell Signaling Technology or otherwise indicated.

Oil red O staining and Masson’s trichrome staining

OCT embedded liver tissues were cut into 6 μm frozen sections and stained with an oil red O (ORO) solution (26079–15, Electron Microscopy Sciences, Hatfield, PA). Masson’s trichrome staining kit was used to detect liver collagen deposition according to the manufacturer’s instruction (87019, Fisher Scientific, Hampton, NH). Quantitative data of ORO-positive areas taken in a 200x magnification were analyzed via color measurement using the software Image-Pro Plus 6.0.

Immunostaining

Liver frozen sections (6 μm) were prepared and immunostained for α-smooth muscle actin levels (α-SMA, fibroblast, 1:750, F3777, Sigma-Aldrich), CD4-positive T cells (CD4, 1:90, 553043), CD8-positive T cells (CD8, 1:100, 01041D), CD68-positive macrophages (CD68, 1:100, 137002, BioLegend, San Diego, CA), and MHC class-II-positive areas (MHC II, 1:350, 556999) (all from eBioscience, San Diego, CA). Quantitative data of CD4⁺ T-cell numbers, CD8⁺ T-cell numbers, MHC class-II- and CD68-positive areas were taken and analyzed via color measurement using the software Image-Pro Plus 6.0.

Plasma analysis

We collected whole blood samples and centrifuged them at 5000 rpm for 15 min at 4 °C. Plasma levels of triglycerides (T7532, Pointe scientific Inc, Canton, MI), high density lipoprotein (HDL, H7511, Pointe scientific Inc), total cholesterol (C7510, Pointe scientific Inc), aspartate transaminase (AST, EASTR-100, BioAssay Systems, Hayward, CA) and alanine transaminase (ALT, EALT-100, BioAssay Systems) were measured with a colorimetric method as followed in the manufacturer’s instructions. Hemolyzed samples were eliminated for AST or ALT detection. ELISA determined plasma IL6 (88–7064, Invitrogen) and MCP-1 (88–7391, Invitrogen) levels.

Liver triglyceride measurement

For liver triglyceride quantification, liver tissue was homogenized in 500 μl of phosphate-buffered saline (PBS). Color reagent solution from the Pointe Scientific Inc total triglyceride assay kit was added at a 2:200 ratio (v/v) to the liver lysates, swirled gently to mix, and incubate for 5 min, followed by reading the plate.

Statistical analysis

All data were presented as mean±SEM. To all data, those with normal distribution and homogeneity of variance, independent t-test, and one-way analysis of variance (ANOVA) LSD test were used for the comparison between two groups and among multiple groups, respectively. SPSS16 version was used for statistical analysis and P<0.05 was considered significant.

RESULTS

Increased expression of CatB in livers FPC diet-fed mice.

Earlier studies showed high levels of CatB in serum from patients with cirrhosis and hepatocellular carcinoma^13–15. Here we compared the expression of few common cysteinyl cathepsins in livers from normal mice and FPC diet-induced NASH mice. Wild-type (Ctsb^+/+) mice consumed a normal laboratory diet LFD or a FPC diet for 6 and 18 weeks to develop NASH^1,22. Immunoblot analyses showed that livers from mice consumed a LFD contained low levels of CatB and CatK. Yet, in livers from mice that consumed a FPC diet for 18 weeks, both CatB and CatK were significantly increased. In contrast, immunoblot analyses of CatL and CatS expression yielded opposite conclusions. Both CatL and CatS were highly expressed in livers from LFD-fed mice, but the expression of both cathepsins was significantly blunted after mice consumed a FPC diet for 6 or 18 weeks (Fig 1). These observations suggest differential roles of these cysteinyl cathepsins in FPC diet-induced NASH.

Fig 1.

Cathepsin expression in mouse liver on different diets. Immunoblot analyses of CatB, CatK, CatL, and CatS in liver extracts from male Ctsb^+/+ mice on a LFD or a FPC diet for 6 or 18 weeks. Gel density quantifications are shown to the right. n=4 per group. **P<0.01, *** P<0.001.

CatB deficiency increases body weight gain but improves liver function in FPC diet-induced NASH mice.

High liver CatB protein levels in NASH mice suggest the involvement of this protease in NASH development. To test this hypothesis, we produced FPC diet-induced experimental NASH in both Ctsb^−/− mice and their Ctsb^+/+ littermates. After 18 weeks of a FPC diet, Ctsb^−/− mice gained significantly more body weights than those of their Ctsb^+/+ littermates (Fig 2A). Such differences in body weight gain may not be due to their food intake. Ctsb^+/+ and Ctsb^−/− mice consumed comparable amounts of FPC diet throughout the course (Fig 2B). FPC diet consumption for 18 weeks significantly increased fasting blood glucose levels in Ctsb^+/+ mice, but such increase in Ctsb^−/− mice did not reach the significance (Fig 2C).

Fig 2.

CatB deficiency increased obesity but improved liver function in mice on a FPC diet. A/B. Body weight gain and food intake of male Ctsb^+/+ and Ctsb^−/− mice on a LFD or a FPC diet for 18 weeks. C. Fasting blood glucose levels in male Ctsb^+/+ and Ctsb^−/− mice on a LFD or a FPC diet for 18 weeks. D. Liver, EAT, SAT, and BAT weight to body weight ratios from above mice. E. Plasma total cholesterol (TC), high-density lipoprotein (HDL) and total triglyceride (TG) of different groups mice as indicated. F. Plasma ALT and AST levels of different groups mice as indicated. G. Triglyceride content of livers from above mice. *P<0.05, **P<0.01, ***P<0.001; mean±SEM; n=10–15 per group.

Decreased fasting blood glucose and increased body weight gain in Ctsb^−/− mouse suggest changes in adipose tissue depot and liver functions from these mice. After 18 weeks feeding of a FPC diet, both Ctsb^+/+ and Ctsb^−/− mice gained more liver weight, but such gain of liver weight was significantly reduced in Ctsb^−/− mice (Fig 2D). As expected, the EAT to body weight ratios in FPC diet-fed Ctsb^−/− mice were significantly higher than those from Ctsb^+/+ mice, although the SAT and BAT to body weight ratios did not differ between the groups (Fig 2D).

ELISA determined plasma lipid profiles of these mice. Although Ctsb^−/− mice gained significantly more body weight than the Ctsb^+/+ mice after 18 weeks of FPC, Ctsb^−/− mice had lower TC and higher HDL than the Ctsb^+/+ mice. Plasma TG did not differ between the groups (Fig 2E). To assess whether CatB deficiency affected liver functions, we measured plasma liver enzymes ALT and AST as indicators of liver damage. As expected, FPC diet feeding for 18 weeks significantly increased plasma ALT in Ctsb^+/+ mice, but such ALT increase was significantly blunted in Ctsb^−/− mice, although the AST changes were not affected by this diet or CatB expression (Fig 2F). Consistent with these observations, FPC diet feeding increased liver TG levels in Ctsb^+/+ mice, and such increase was also muted in livers from Ctsb^−/− mice (Fig 2G). Together, these observations from plasma biochemical analyses, adipose tissue depots, and liver functional analyses suggest an essential role of CatB in murine NASH development.

CatB deficiency decreases liver lipid accumulation in FPC diet-fed mice.

Liver triglyceride has been associated with nonalcoholic fatty liver disease among markers of hyperlipidemia²⁵. Lower hepatic triglyceride levels in FPC diet-fed Ctsb^−/− mice than the Ctsb^+/+ control mice (Fig 2G) support a role of CatB in FPC diet-induced NASH. Consistent with this conclusion, liver section ORO staining showed that the FPC diet-induced lipid accumulation in livers from Ctsb^+/+ mice was significantly blocked in livers from Ctsb^−/− mice (Fig 3A).

Fig 3.

CatB deficiency reduced liver lipid accumulation in FPC-fed mice. The following parameters were measured in male Ctsb^+/+ and Ctsb^−/− mice after 18 weeks on a LFD or a FPC diet (*P<0.05, **P<0.01, ***P<0.001; mean±SEM): A. Oil red O staining of liver sections. Representative images are shown to the left. Scale:100 μm. n=10–15 per group. B. RT-PCR analyses of liver lipolysis related genes Ppar-α, Cpt1α, Acc1 and Acc2. n=6 per group. C. RT-PCR analyses of liver gluconeogenesis related genes Glut2 and Pepck. n=6 per group. D. RT-PCR analysis of liver lipid transportation and lipogenesis related genes CD36, CETP, Ppar-γ, Fabp4 and Scad1. n=6 per group. E. Immunoblot detection of liver CD36, PPAR-γ, FABP4, p-AMPKα, AMPKα, p-HSL and HSL expression from different groups of male mice as indicated. Gel density quantifications of each blot were normalized to the levels of GAPDH, AMPKα, or HSL expression and shown to the right. n=3 per group.

To test the corresponding molecular mechanisms that accounted for the reduced liver TG content and lipid accumulation in FPC diet-fed Ctsb^−/− mice, we performed RT-PCR and analyzed the expression of liver lipid metabolism associated genes, including nuclear receptor protein Ppar-α, fatty acid oxidation enzyme Cpt1α, and fatty acid biosynthesis enzymes Acc1 and Acc2 from FPC diet-fed mice. As expected, FPC diet feeding reduced the expression of liver Ppar-α, Cpt1α, and Acc2, although FPC diet did not affect liver Acc1 expression in Ctsb^+/+ mice. CatB deficiency did not affect the expression of these genes (Fig 3B). RT-PCR also revealed no effect of CatB expression on liver gluconeogenesis genes, including Glut2 and Pepck from the gluconeogenesis metabolic pathway (Fig 3C).

Triglyceride is derived from the esterification of glycerol and free fatty acids (FFA) from the plasma, de novo synthesis, and dietary fat delivered by chylomicron remnants²⁶. FFA translocation into the hepatocyte is mediated by CD36, also known as free fatty acid translocase²⁷. PPARγ activation stimulates CD36 expression²⁸. Reduced liver triglyceride level from Ctsb^−/− mice suggests a role of CatB in liver lipid transport and lipogenesis. RT-PCR revealed that the expression of liver extracellular and mitochondrial lipid carrier CD36, lipogenic transcription factor Ppar-γ, FFA transport marker Fabp4, and lipogenesis marker Scad1 was all increased in Ctsb^+/+ mice after FPC diet consumption, but such increases were significantly blunted in Ctsb^−/− mice (Fig 3D). In contrast, the transcripts of liver intracellular cholesteryl esters and triglycerides transporter (CETP) were markedly decreased from Ctsb^+/+ mice after FPC diet consumption. CatB deficiency did not affect this reduction (Fig 3D). In parallel to reduced expression of genes involved in liver lipid transport and lipogenesis, immunoblot analysis showed that the protein levels of liver CD36, FABP4 and PPARγ were also significantly decreased in Ctsb^−/− mice (Fig 3E). In contrast, the activation (phosphorylation) of mitochondrial energy sensor protein AMPK and HSL increased in livers from both Ctsb^+/+ and Ctsb^−/− mice after FPC diet consumption and such increase did not differ between the genotypes (Fig 3E). Therefore, CatB played an essential role in liver lipogenesis and lipid metabolism during NASH development.

CatB promotes liver adhesion molecule expression and inflammatory cell accumulation in FPC diet-induced NASH.

Liver inflammation is one of the most important characteristics of NASH^1,22. The expression of liver adhesion molecules plays a pivotal role in inflammatory cell accumulation. To test a direct role of CatB in inflammatory response in NASH development, we compared the expression of adhesion molecules in livers from mice fed with a LFD or a FPC diet. In livers from FPC diet-fed Ctsb^+/+ mice, we detected significantly increased expression of adhesion molecules ICAM-1 and E-selectin, compared with those fed a LFD. Such increase was blunted in livers from Ctsb^−/− mice (Fig 4A). Different levels of adhesion molecules in livers from Ctsb^+/+ and Ctsb^−/− mice may affect inflammatory cell infiltration. Immunohistological analyses revealed significantly increased accumulation of CD4⁺ T cells (Fig 4B), CD8⁺ T cells (Fig 4C), CD68⁺ macrophages (Fig 4D), and MHC II-positive cells as indication of tissue inflammation (Fig 4E) in livers from Ctsb^+/+ mice that consumed a FPC diet for 18 weeks. FPC diet consumption also increased CD4⁺ T-cell, CD8⁺ T-cell, CD68⁺ macrophage, and MHC II-positive cell accumulation in livers from Ctsb^−/− mice, although the increase of CD4⁺ T cells and CD68⁺ macrophages did not reach statistical significance. Overall, the increases of liver inflammatory cells from Ctsb^−/− mice under the same dietary conditions was greatly suppressed. Therefore, CatB deficiency reduced liver inflammation in FPC-fed NASH mice.

Fig 4.

CatB activities on liver adhesion molecule expression and inflammatory cell accumulation in FPC-induced NASH. A. RT-PCR analyses of liver adhesion molecule ICAM-1 and E-selectin expression from male Ctsb^+/+ and Ctsb^−/− mice on a LFD or a FPC diet for 18 weeks. n=6 per group. B-E. Immunostaining detected liver CD4⁺ T cells (B), CD8⁺ T cells (C), CD68⁺ macrophages (D), and MHC class-II-positive areas (E) in mice from A. Representative images are shown to the left. Scale: 100 um. n=10–15 per group. *P<0.05, **P<0.01, *** P<0.001.

CatB activity controls systemic and hepatic inflammation and liver macrophage polarization.

Reduced accumulation of inflammatory cells (CD4⁺ and CD8⁺ T cells, and CD68⁺ macrophages) (Fig 4B–4D) in livers from Ctsb^−/− mice may indicate changes in local and systemic inflammatory cytokine and chemokine expression. RT-PCR analyses of liver tissues revealed FPC diet-induced expression of Tnfα and Mcp1 genes in Ctsb^+/+ mice and such inductions were significantly reduced in livers from Ctsb^−/− mice (Fig 5A). ELISA of plasma IL6 and MCP-1 showed same changes. FPC diet induced plasma IL6 and MCP-1 levels in Ctsb^+/+ mice, but CatB deficiency blocked these increases (Fig 5B). CatB activity in regulating the release of inflammatory cytokines and chemokines may modulate liver inflammatory cell accumulation, lipid metabolism, and fibrosis.

Fig 5.

CatB deficiency decreased FPC diet-induced systemic and hepatic inflammation in mice. A. RT-PCR analyses of liver inflammatory cytokine and chemokine TNF-α and MCP-1 expression from male Ctsb^+/+ and Ctsb^−/− mice on a LFD or a FPC diet for 18 weeks. n=6 per group. B. ELISA quantification of plasma IL6 and MCP-1 levels from different groups of mice as indicated. n=10–15. C. RT-PCR analysis of liver M2 macrophage marker (Arg1) and M1 macrophage markers (iNOS and F4/80) from the same mice as in A. n=6 per group.

Studies reported an association between macrophage polarization and NAFLD/NASH development^29,30. Increased expression of PPAR-γ shifts lipid-induced macrophage and Kupffer cell polarization towards a M2 phenotype. PPAR-γ agonist rosiglitazone restores diet-induced imbalance of Kupffer cells polarization, thereby alleviating hepatic steatosis and local inflammatory responses²⁹. Results from Fig 3D and 3E showed that FPC diet increased PPAR-γ mRNA and protein levels in livers from Ctsb^+/+ mice. Consistent with these results, FPC diet consumption reduced liver expression of M2 marker Arg1 and increased liver expression of M1 markers iNOS and F4/80 in Ctsb^+/+ mice. All these changes were significantly improved in livers from FPC diet-fed Ctsb^−/− mice (Fig 5C), supporting a role of CatB in NASH mouse liver M1 macrophage polarization.

CatB deficiency reduces liver fibrosis in FPC diet-fed mice.

Imbalances in collagen synthesis and degradation contribute to cirrhosis and liver failure³¹. Improved liver function in FPC diet-fed Ctsb^−/− mice relative to that in Ctsb^+/+ mice suggests a role of CatB in liver fibrosis. As expected, 18 weeks of FPC diet feeding increased the mRNA levels of all tested liver fibrosis genes, including collagen subtypes Col1α1 (8-fold), Col1α2 (6-fold), Col3α1 (4-fold), Col4α1 (1.5-fold), fibroblast activation marker αSMA (up to 1.5-fold) and Tgfβ1 (1.5-fold) in livers from Ctsb^+/+ mice. In contrast, all these changes except Col1α2 were significantly reduced or even fully blunted in livers from the same diet-treated Ctsb^−/− mice (Fig 6A). Differentiation and accumulation of myofibroblasts contribute to liver fibrosis^31,32. Myofibroblast differentiation during liver fibrosis leads to increased α-SMA expression and generation and secretion of extracellular matrix proteins such as collagen and fibronectin³¹. Decreased mRNA levels of type I, type-III, and type-IV collagens in Ctsb^−/− mouse livers also suggest that these livers contain fewer fibroblasts and are less prone to develop fibrosis than those in Ctsb^+/+ mice. Immunostaining of α-SMA and Masson’s trichrome staining supported this hypothesis. Livers from Ctsb^−/− mice contained significantly fewer α-SMA-positive myofibroblasts and less Masson’s trichrome-positive collagen deposition than those from Ctsb^+/+ mice after 18 weeks of FPC diet consumption (Fig 6B–6C). In line with the decreased collagen expression and deposition in livers from FPC diet-fed Ctsb^−/− mice, immunoblot analysis displayed reduced expression of collagen-III and collagen crosslinking enzyme lysyl oxidase in livers from Ctsb^−/− mice relative to those from Ctsb^+/+ mice after FPC diet treatment (Fig 6D). Immunoblot analysis of the same sets of livers reveled markedly lower protein levels of TGF-β1 and associated downstream signaling molecules p-Smad-2 and p-Smad-3 in Ctsb^−/− mice, compared with those in Ctsb^+/+ mice (Fig 6D).

Fig 6.

CatB deficiency reduced liver fibrosis in mice on a FPC diet. The following parameters were measured in male Ctsb^+/+ and Ctsb^−/− mice after 18 weeks on a LFD or a FPC diet: A. RT-PCR analyses of liver collagen synthesis-related genes Col1α1, Col1α2, Col3α1, Col4α1, αSMA and Tgfβ1. n=6 per group. B. Immunostaining detection of liver α-SMA expression. Representative images are shown to the left. Scale: 100 μm. n=10–15 per group. C. Masson’s trichrome staining of liver sections. Scale: 100 μm. D. Immunoblot detection of liver collagen-III, TGF-β1, lysyl oxidase, p-Smad-2/3, and Smad-2/3 expression. Quantification of the blots normalized to the level of β-actin or GAPDH expression. n=3 per group. Data are mean±SEM.

CatB deficiency blocks TGF-β signaling

Reduced collagen deposition in livers from Ctsb^−/− mice suggest a role of CatB in regulating collagen degradation or biosynthesis. Although CatB showed moderate collagenase activity,³³ CatB deficiency may affect the expression of other proteases that indirectly contributed to liver collagen proteolysis. To test this hypothesis, we performed RT-PCR analysis in livers from Ctsb^−/− and Ctsb^+/+ mice fed with a LFD or a FPC diet and assessed mRNA levels of several common cysteinyl cathepsins (CatB, CatK, CatL, CatS) and matrix metalloproteinases MMPs (MMP-2, MMP-8, MMP-9, and MMP-13). Although FPC feeding increased the expression of some of these proteases, CatB deficiency did not affect the expression of none of these tested proteases (Fig 7A). We also performed immunoblots to test the expression of several collagenolytic cysteinyl cathepsins, including CatK, CatL, and CatS.^33–36 FPC diet consumption reduced the expression of both the 25~28-kDa and 21-kDa active forms of CatL and CatS in livers from both Ctsb^+/+ and Ctsb^−/− mice (Fig 7B). In contrast, livers from Ctsb^+/+ mice showed increased expression of the 28 kDa active form of CatK after mice consumed a FPC diet. Such increase did not occur in livers from Ctsb^−/− mice (Fig. 7B). Therefore, possible changes in collagenolytic activity because of CatB and CatK deficiencies in livers from FPC-fed Ctsb^−/− mice did not explain reduced liver collagen production in these mice.

Fig 7.

CatB deficiency affects liver expression of cathepsins and p-Smad-2/3 nuclear membrane transporters. A. RT-PCR analyses of liver collagen degradation-related cathepsins (CatB, CatK, CatL, CatS) and MMPs (MMP-2, 8, 9, 13). n=6 per group. B. Immunoblot detection of liver CatK, CatL, and CatS. n=3 per group. C. Immunoblot detection of liver RanBP3 and importin-β. Quantification of the blots normalized to the level of GAPDH expression. n=3 per group. Data are mean±SEM.

Immunoblot analyses in Fig 6 showed that CatB deficiency reduced liver expression of TGF-β1 and downstream p-Smad-2 and p-Smad-3, supporting a role of CatB in TGF-β signaling. Our recent study showed that cyeteinyl cathepsins regulate TGF-β signaling by controlling the expression and activity of plasma membrane TGF-β receptors TGFBR1 and TGFBR2 and nuclear membrane importin-β and RanBP3 that mediate the translocation and exportation of Smad complex between the cytoplasma and nucleus.²³ Immunoblot analyses showed that FPC diet consumption did not affect the expression of Smad complex nucleus importing molecule importin-β in Ctsb^+/+ or Ctsb^−/− mice. In contrast, FPC diet significantly reduced the expression of Smad complex nucleus exporting molecule RanBP3 in Ctsb^+/+ mice. Such reduction did not occur in livers from Ctsb^−/− mice (Fig 7C). These observations suggest that reduced TGF-β signaling and collagen biosynthesis contributed to reduced liver collagen deposition in Ctsb^−/− mice.

DISCUSSION

Studies demonstrated that diets rich in fructose lead to oxidative stress, expression of proinflammatory cytokines and chemokines, triglyceride production, and elevated lipogenesis in liver as well as increased adipokine expression in visceral adipose tissue in dose- and time-dependent manners, which recapitulate the human NASH pathologies^37–40. This study used the FPC diet to induce mouse NASH with enhanced hepatic steatohepatitis and associated overweight, increased liver to body weight ratio, elevated EAT and SAT deposition, and reduced BAT amount. Using CatB-deficient Ctsb^−/− mice, here we established a central role for CatB in hepatic lipid accumulation, inflammatory cell accumulation, M1 macrophage/Kupffer cell polarization, TGF-β signaling, and fibrosis, although CatB may play many other untested roles to promote the NASH pathogenesis.

The expression of CD36 is increased in obese individuals and correlates with hepatic steatosis and NAFLD. Increased CD36 expression promotes lipid metabolism by enhancing fatty acid uptake^41,42. Further, the activation of PPARγ also stimulates CD36 expression and activities²⁸. In mice fed with a high fat diet, deletion of CD36 from either hepatocytes or Kupffer cells attenuates steatosis^43,44. Here we showed a direct role of CatB in regulating the mRNA and protein levels of CD36 and PPARγ. CatB deficiency blocked the expression of these lipid metabolism and hepatic steatosis essential molecules. Abolished liver FABP4 expression, as a common marker of adipocyte, in FPC diet-fed Ctsb^−/− mice supported a direct role of CatB in liver lipid accumulation. Although not tested, this activity of CatB in livers may also be true in high fat diet-induced obese mice.

As we recently summarized⁴⁵, CatB activates caspase-3, caspase-9, and BH3-interacting domain death agonist (BID), as a mechanism to promote cell apoptosis. CatB also degrades one of the NACHT, LRR, and PYD domains- containing protein (NLRP) molecules NLRP10 that is a negative regulator of NLRP3. NLRP3, pro-caspase-3, and pro-casapse-9 form the inflammasome complex that mediates the activation of caspase-1, pro-IL-1β, and pro-IL-18^46,47. In this study, we did not test whether these activities of CatB were involved in the NASH mouse liver pathology, but we found that CatB deficiency reduced liver chemokine expression and inflammatory cell accumulation. If inflammatory cell accumulation is pathogenic to NASH livers^1,22, CatB activities in regulating liver chemokine expression and inflammatory cell accumulation can be essential to the NASH liver pathology.

As a protease, expression of CatB may affect matrix protein degradation directly or indirectly. Although CatB has moderate collagenase activity, other cathepsins and MMPs have been shown essential in tissue fibrosis^23,33,48. Yet, we did not find an involvement of CatB proteolytic function in FPC diet-induced liver fibrosis. CatB deficiency or FPC diet consumption did not affect liver MMP expression or CatS and CatL levels. CatB deficiency blocked the FPC diet-induced expression of liver collagenolytic CatK³⁴, but also reduced liver collagen production. Therefore, CatB may contribute to liver collagen deposition independent to its protease activity. We recently reported a similar but opposite observation in kidney epithelial cells²³. Different from the livers in this study, in unilateral ureteral obstruction-induced renal injury model, CatB deficiency exacerbated post-injury kidney fibrosis. Mechanistic studies using the mouse kidney epithelial cells, we revealed a role of CatB in binding to TGFBR1 on the plasma membrane, thereby blocking the TGF-β signaling from the TGFBR2. In addition, CatB also binds to the nuclear membrane exporter protein RanBP3, but not nuclear membrane importer protein importin-β, thereby blocking the Smad complex nuclear translocation. Here we showed that FPC diet reduced RanBP3 expression in livers from WT mice, thereby reducing Smad complex release from the nucleus and enhancing fibrotic protein synthesis. In contrast, FPC diet did not affect RanBP3 expression in livers from Ctsb^−/− mice. Therefore, this study provided a direct evidence of CatB activity in increasing TGF-β1 and reducing RanBP3 expression, leading to enhanced canonical Smad signaling and fibrogenesis in NASH mouse livers. CatB deficiency markedly reduced liver fibrogenesis by reducing collagen synthesis. This is our second evidence that cysteinyl cathepsins regulate tissue fibrosis without using their activities in matrix protein degradation, although CatB displayed different activities in livers from those in the kidneys²³. Although not tested, other cathepsins such as CatL, CatS, and CatK may also act differently in livers from NASH mice from those in the injured kidneys.

In conclusion, our study established a pathogenic role of CatB in diet-induced NASH in mice. Targeting CatB is a novel concept of controlling diet-induced NASH in humans, although potential side effect in the kidneys or possibly in other organs will have to put into consideration when we design the therapeutic approaches to target this cathepsin.

Glance Commentary.

Background

Prior preclinical studies suggested essential roles of cysteinyl cathepsins in diet-induced obesity and diabetes. Here we report a pathogenic role of CatB in diet-induced non-alcoholic steatohepatitis (NASH) characterized by increased liver CatB expression, fat accumulation, and inflammation. CatB deficiency reduced body weight gain, liver-body weight ratio, blood glucose, total cholesterol, and alanine transaminase levels, and liver triglyceride, along with reduced liver lipid transport and lipogenesis gene expression. Mechanistic study supports a role for CatB in liver adhesion molecule and chemokine expression and inflammatory cell infiltration, blocking M2 macrophage polarization, and reducing liver TGF-β1 signaling and fibrosis.

Translational Significance

Targeting CatB with its selective inhibitor is a novel concept of controlling diet-induced NASH in humans.

NONSTANDARD ABBREVIATIONS

NASH

non-alcoholic steatohepatitis

NAFLD

non-alcoholic fatty liver disease

CatB

cathepsin B

CatL

cathepsin L

CatS

cathepsin S

CatK

cathepsin K

FPC

fructose-palmitate-cholesterol

RT-PCR

real-time polymerase chain reaction

TC

total cholesterol

ALT

alanine transaminase

AST

aspartate transaminase

TG

triglyceride

HDL

high-density lipoprotein

ER

endoplasmic reticulum

LFD

low fat diet

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

Ppar-α

peroxisome proliferator activated receptor-α

Cpt1α

carnitine palmitoyltransferase-1α

ACC

acetyl-CoA carboxylase

Glut2

glucose transporter-2

Pepck

phosphoenolpyruvate carboxykinase

CETP

cholesteryl esters and triglycerides transporter

Ppar-γ

peroxisome proliferator activated receptor-γ

Fabp4

fatty acid binding protein 4

Scad1

short-chain acyl dehydrogenase-1

α-SMA

α-smooth muscle actin

Tgfβ

transforming growth factor-β

Tnfα

tumor necrosis factor-α

Mcp-1

monocyte chemoattractant protein-1

Icam-1

intercellular adhesion molecule-1

Arg1

arginase-1

iNOS

inducible nitric oxide synthase

HSL

hormone-sensitive lipase

ORO

oil red O

PBS

phosphate-buffered saline

ANOVA

analysis of variance

EAT

epididymal adipose tissue

SAT

subcutaneous adipose tissue

BAT

brown adipose tissue

FFA

free fatty acids

MMPs

matrix metalloproteinases

TGFBR1

TGF-β receptor-1

TGFBR2

TGF-β receptor-2