Transforming growth factor beta signaling in hepatocytes participates in steatohepatitis through regulation of cell death and lipid metabolism in mice

Authors

  • Ling Yang,

    1. Division of Gastroenterology, Department of Medicine, School of Medicine, University of California San Diego, La Jolla, CA
    2. Division of Gastroenterology, Department of Internal Medicine, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
    Search for more papers by this author
  • Yoon Seok Roh,

    1. Division of Gastroenterology, Department of Medicine, School of Medicine, University of California San Diego, La Jolla, CA
    Search for more papers by this author
  • Jingyi Song,

    1. Division of Gastroenterology, Department of Medicine, School of Medicine, University of California San Diego, La Jolla, CA
    Search for more papers by this author
  • Bi Zhang,

    1. Division of Gastroenterology, Department of Medicine, School of Medicine, University of California San Diego, La Jolla, CA
    Search for more papers by this author
  • Cheng Liu,

    1. Division of Gastroenterology, Department of Medicine, School of Medicine, University of California San Diego, La Jolla, CA
    Search for more papers by this author
  • Rohit Loomba,

    1. Division of Gastroenterology, Department of Medicine, School of Medicine, University of California San Diego, La Jolla, CA
    Search for more papers by this author
  • Ekihiro Seki

    Corresponding author
    1. Division of Gastroenterology, Department of Medicine, School of Medicine, University of California San Diego, La Jolla, CA
    • Address reprint requests to: Ekihiro Seki, M.D., Ph.D., Division of Gastroenterology, Department of Medicine, School of Medicine, University of California San Diego, School of Medicine, 9500 Gilman Drive, MC 0702, Leichtag Biomedical Research Building, Room 118B, La Jolla, CA 92093-0702. E-mail: ekseki@ucsd.edu; fax: 858-822-5370.

    Search for more papers by this author

  • This study is supported by a National Institutes of Health grant (R01AA02172 [to E.S.] and R01DK085252 [to E.S.]). R.L. is supported, in part, by the American Gastroenterological Association (AGA) Foundation-Sucampo-ASP Designated Research Award in Geriatric Gastroenterology and by a T. Franklin Williams Scholarship Award, with funding provided by: Atlantic Philanthropies, Inc., the John A. Hartford Foundation, the Association of Specialty Professors, and the AGA and grants K23-DK090303-2 and P30CA23100-27. L.Y. was supported, in part, by the National Natural Science Foundation of China (no. 30500658 and 81370550).

  • Potential conflict of interest: Nothing to report.

Abstract

Transforming growth factor beta (TGF-β) signaling activates Smad- and TGF-β-activated kinase 1 (TAK1)-dependent signaling to regulate cell survival, proliferation, fibrosis, and tumorigenesis. The effects of TGF-β signaling on metabolic syndrome, including nonalcoholic fatty liver disease, remain elusive. Wild-type (WT) and hepatocyte-specific TGF-β receptor type II-deficient (Tgfbr2ΔHEP) mice were fed a choline-deficient amino acid (CDAA)-defined diet for 22 weeks to induce NASH. WT mice fed a CDAA diet displayed increased activation of Smad2/3 and had marked lipid accumulation, inflammatory cell infiltration, hepatocyte death, and fibrosis; in comparison, Tgfbr2ΔHEP mice fed a CDAA diet had suppressed liver steatosis, inflammation, and fibrosis. Both palmitate-induced steatotic hepatocytes and hepatocytes isolated from WT mice fed a CDAA diet had increased susceptibility to TGF-β-mediated death. TGF-β-mediated death in steatotic hepatocytes was inhibited by silencing Smad2 or blocking reactive oxygen species (ROS) production and was enhanced by inhibiting TAK1 or nuclear factor kappa B. Increased hepatic steatosis in WT mice fed a CDAA diet was associated with the increased expression of lipogenesis genes (Dgat1 and Srebp1c), whereas the decreased steatosis in Tgfbr2ΔHEP mice was accompanied by the increased expression of genes involved in β-oxidation (Cpt1 and Acox1). In combination with palmitate treatment, TGF-β signaling promoted lipid accumulation with induction of lipogenesis-related genes and suppression of β-oxidation-related genes in hepatocytes. Silencing Smad2 decreased TGF-β-mediated lipid accumulation and corrected altered gene expression related to lipid metabolism in hepatocytes. Finally, we confirmed that livers from patients with nonalcoholic steatohepatitis (NASH) displayed phosphorylation and nuclear translocation of Smad2/3. Conclusions: TGF-β signaling in hepatocytes contributes to hepatocyte death and lipid accumulation through Smad signaling and ROS production that promote the development of NASH. (Hepatology 2014;59:483–495)

Abbreviations
Ab

antibody

α-SMA

alpha-smooth muscle actin

ALT

alanine aminotransferase

BHA

butylated hydroxyanisole

BSA

bovine serum albumin

CDAA

choline-deficient amino acid defined

C/EBPβ

CCAAT/enhancer-binding protein beta

CSAA

choline-supplemented amino acid defined

CTGF

connective tissue growth factor

FFA

free fatty acid

GTT

glucose tolerance test

H&E

hematoxylin and eosin

HFD

high-fat diet

HNE

hydroxynonenal

HSC

hepatic stellate cell

IHC

immunohistochemistry

IL

interleukin

IP

intraperitoneal

IR

insulin resistance

ITT

insulin tolerance test

JNK

c-Jun N-terminal kinase

LDH

lactate dehydrogenase

mRNA

messenger RNA

NAFL

nonalcoholic fatty liver

NAFLD

nonalcoholic fatty liver disease

NAS

NAFLD activity score

NASH

nonalcoholic steatohepatitis

NF-κB

nuclear factor kappa B

NIH

National Institutes of Health

PAI

plasminogen activator inhibitor 1

PGC-1α

PPAR-γ coactivator 1 alpha

PPAR-γ

peroxisome proliferator-activated receptor gamma

qPCR

quantitative real-time polymerase chain reaction

ROS

reactive oxygen species

siRNA

small interfering RNA

TAK1

TGF-β-activated kinase 1

Tg

transgenic

TG

triglyceride

TGF-β

transforming growth factor-β

Tgfbr2

TGF-β receptor type II

Tgfbr2ΔHep mice

hepatocyte-specific Tgfbr2 deleted mice

TIMP-1

tissue inhibitor of metalloproteinase 1

TNF-α

tumor necrosis factor alpha

TUNEL

terminal deoxynucleotidyl transferase dUTP nick end labeling

UCSD

University of California San Diego

WT

wild type

Nonalcoholic fatty liver disease (NAFLD), the hepatic manifestation of metabolic syndrome, is currently a significant health concern in the United States in both adults and children.[1-5] Obesity and type II diabetes are strongly linked to the progression of NAFLD.[6] A disease spectrum of NAFLD ranges from simple steatosis to steatosis with liver inflammation and fibrosis, referred to as nonalcoholic steatohepatitis (NASH).[5] Ten to twenty percent of NAFLD patients have NASH, and 10%-15% of NASH patients eventually progresses into liver cirrhosis that significantly increases the risk for the development of hepatocellular carcinoma.[1-4] However, it is still unclear what triggers the progression of NAFLD to NASH accompanied with inflammation and fibrosis. Though simple hepatic steatosis is considered benign, NAFLD with the presence of histologic changes consistent with NASH significantly shortens human life expectancy. NAFLD is projected to replace hepatitis C as the leading cause for liver transplantation by 2020.[2] A deeper understanding of the molecular mechanisms underlying the development of NASH is necessary to develop effective alternative therapies for NASH that eliminate reliance on the insufficient number of donor livers.

Transforming growth factor beta (TGF-β) is a pleiotropic cytokine involved in cell survival, proliferation, differentiation, angiogenesis, and wound-healing response.[7-9] TGF-β binding to type II TGF-β receptor causes type II receptor to recruit and phosphorylate type I TGF-β receptor. Type I TGF-β receptor kinase then activates a canonical Smad-dependent pathway and a noncanonical Smad-independent TGF-β-activated kinase 1 (TAK1)- or phosphoinositide 3 kinase-dependent pathway. In Smad-dependent signaling, phosphorylated Smad2 and Smad3 associate with Smad4 to translocate into the nucleus to control transcriptional gene expression. In the liver, TGF-β signaling participates in fibrogenic response through hepatic stellate cell (HSC) activation.[9, 10] Thus, it is well established that TGF-β signaling in HSCs plays a role in progression of fibrosis in advanced NAFLD. Although previous studies demonstrated increases in hepatic TGF-β expression in NASH patients with fibrosis,[11, 12] the role of TGF-β signaling in hepatocytes in the development of NASH is not well understood. In addition, the study of TGF-β signaling in metabolic disease is still limited. Previous reports demonstrated that TGF-β/Smad3 signaling is involved in insulin gene expression in pancreatic β cells.[13] TGF-β/Smad3 signaling is associated with systemic insulin resistance (IR), obesity, and hepatic steatosis through regulation of expression of PPAR-γ coactivator 1 alpha (PGC-1α) and peroxisome proliferator-activated receptor gamma (PPAR-γ).[14, 15] Indeed, Smad3−/− mice were protected from IR, obesity, and hepatic steatosis induced by high-fat diet (HFD) feeding.[14, 15] Because the previous study used whole-body Smad3−/− mice, the specific role of TGF-β signaling in hepatocytes to the development of diet-induced NASH needs to be studied. In the present study, we used choline-deficient amino acid (CDAA)-defined diet that induces hepatic steatosis, inflammation, and fibrosis, all of which are comparable pathologies to human NASH. We tested functions of hepatocyte TGF-β signaling in the progression of CDAA diet-induced NASH in mice.

Materials and Methods

Mouse Colonies

Albumin-Cre recombinase transgenic (Tg) mice (Albumin-Cre Tg mice) and Tgfbr2flox/flox mice were purchased from The Jackson Laboratories (Bar Harbor, MA). The two lines were intercrossed to generate albumin-Cre/+ Tgfbr2flox/flox mice (Tgfbr2ΔHep) on C57BL6 background. Cre-negative animals were used as wild-type (WT) controls. Eight-week-old male mice were divided into three groups: standard chow group (catalog no. 5053; PicoLab Rodent Diet 20 series; LabDiet, Framingham, MA); choline-supplemented L-amino-acid-defined diet group (CSAA; catalog no. 518754; Dyets Inc., Bethlehem, PA); and CDAA-defined diet group (catalog no. 518753; Dyets Inc.). Both CSAA and CDAA diets contain higher calories than the standard chow (Supporting Table 1).[16-18] Mice were on these diets for 22 weeks, and CDAA diet intake was monitored for the first month. All mice received humane care according to National Institutes of Health (NIH; Bethesda, MD) recommendations outlined in the Guide for the Care and Use of Laboratory Animals. All animal experiments were approved by the University of California San Diego (UCSD) Institutional Animal Care and Use Committee (La Jolla, CA).

Mouse Tissue Processing

Mouse tissues were either snap-frozen in liquid nitrogen for RNA and protein preparations or fixed in 10% or 4% neutral-buffered formalin phosphate (Fisher Scientific, Pittsburgh, PA) to be embedded in paraffin or optimal cutting temperature compound and cut into 5-μm sections for hematoxylin and eosin (H&E), Oil Red O, Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL), and Sirius Red staining.[16-18] TUNEL-positive cells were counted on 10 high power (200×) fields/slide. The Sirius Red-positive area was measured on 10 low power (40×) fields/slide and quantified with the use of NIH imaging software. NAFLD activity score (NAS) was determined as described. Immunohistochemistry (IHC) for F4/80, alpha-smooth muscle actin (α-SMA), phosphorated Smad2/3, and 4-hydroxynonenal (HNE) were performed.[16-18]

Quantitative Real-Time Polymerase Chain Reaction

RNA extracted from livers and cells were subjected to reverse transcription and subsequently underwent quantitative real-time polymerase chain reaction (qPCR) with the use of the CFX96 real-time PCR system (Bio-Rad, Hercules, CA). PCR primer sequences are listed in Supporting Table 2. Genes were normalized to 18S RNA as an internal control.

Cell Isolation and Treatment

Hepatocytes were isolated, and cells with viability above 90% were used for the experiments.[16, 17] After hepatocytes attached (2 hours after plating), medium was changed to serum-free media (day 0). Then, in some experiments, cells were incubated with scramble or Smad2 small interfering RNA (siRNA; sc-38375; Santa Cruz Biotechnology, Santa Cruz, CA) or adenovirus encoding control-Lac-z or IκB super-repressor for an additional 16 hours (overnight). On the following day, palmitate or oleate (200 μM) was added for 6 hours (day 1). Subsequently, cells were treated with 10 ng/mL of murine TGF-β1 (R&D Systems, Minneapolis, MN) for 24 hours (day 2) and analyzed for cell death or lipid accumulation (total, 48 hours after plating). In some experiments, 250 nM of 5z-7-oxozeaenol (TAK1 inhibitor), 45 μM of fasudil (rock inhibitor), 10 μM of SB203580 (p38 inhibitor), 20 μM of SP600125 (c-Jun N-terminal kinase [JNK] inhibitor), or 100 μM of butylated hydroxyanisole (BHA; Sigma-Aldrich, St Louis, MO) was added 30 minutes before TGF-β1 treatment. Albumin-conjugated palmitate used in this study was prepared by dissolving palmitate in ethanol at 50oC and then conjugating with fatty-acid-free bovine serum albumin (BSA). To measure reactive oxygen species (ROS), cells were incubated with 10 μM of CM-H2DCFDA (Invitrogen, Grand Island, NY) for 30 minutes at 37°C and analyzed by a fluorescence microplate reader.[19]

Lipid Isolation and Measurement

Liver extracts were prepared by homogenization in 0.25% sucrose with 1 mmol/L of ethylenediaminetetraacetic acid. Lipids were extracted using chloroform/methanol (2:1, v/v) and suspended with 5% fatty-acid-free BSA. Triglyceride (TG), total cholesterol, and free fatty acid (FFA) contents were measured with the use of the Triglyceride Reagent Set (Pointe Scientific, Canton, MI), Cholesterol E (Wako, Richmond, VA), FFA, and the half micro test (Roche, Mannheim, Germany).[16] Hepatocyte TG accumulation was quantified by extraction of hepatocyte lipids from cell homogenates using chloroform/methanol (2:1), and TG was measured.

Immunofluorescence

Hepatocytes were labeled by Mito Tracker Red CMXRos (100 nM; Molecular Probes, Eugene, OR) and fixed with 4% paraformaldehyde in phosphate-buffered saline, followed by permeabilization and incubation with anti-Bax antibody (Ab; Cell Signaling Technology, Danvers, MA).[20]

Glucose and Insulin Tolerance Tests

For the glucose tolerance test (GTT), baseline glucose levels were measured from mice fasted for 16 hours (0 min). Then, 2 g/kg of glucose were administered by intraperitoneal (IP) injection to mice, and glucose levels were measured at 15-minute intervals over a span of 2 hours after the glucose load. For the insulin tolerance test (ITT), baseline glucose levels were measured after 4 hours of fasting. Blood-glucose concentration was monitored every 15-30 minutes for 90 minutes after administration of insulin at 0.5 U/kg by IP injection.

Human Liver Samples

Paraffin-embedded human liver tissues were acquired from liver biopsy samples of patients with NAFLD. All liver biopsies were read by a single hepatopathologist who was blinded to clinical data. Liver biopsies were scored using the NASH Clinical Research Network Histologic Scoring System.[21] NAS was documented that ranges from 0 to 8. NAS is a sum of three histologic scores, including steatosis (0-3), lobular inflammation (0-3), and ballooning degeneration (0-2). Patients were classified into two groups: nonalcoholic fatty liver (NAFL; NAS score between 0 and 3) without fibrosis (n = 10) versus those with NASH with fibrosis and hepatocyte ballooning (NAS score of 5 or higher; n = 10). Please visit the UCSD NAFLD registry for details on inclusion and exclusion criteria (fattyliver.ucsd.edu) for diagnosis of NAFLD.[22] The diagnosis of NASH was defined as previously reported by the San Diego Integrated NAFLD Consortium.[23] The study was approved by the UCSD Institutional Review Board.

Statistical Analysis

Differences between two groups were compared using Mann-Whitney's U test or two-tailed unpaired Student t test. Differences between multiple groups were compared using one-way analysis of variance using SPSS software (SPSS, Inc., Chicago, IL). P values <0.05 were considered significant.

Results

Deletion of Tgfbr2 in Hepatocytes Alleviates Steatohepatitis in Mice

Initially, we examined the activation of TGF-β signaling in mouse NASH livers induced by 22 weeks of CDAA diet feeding. NASH livers exhibited increased phosphorylation of Smad2/3 and its nuclear translocation in hepatocytes and liver nonparenchymal cells as well as increased expression of TGF-β (Fig. 1A). These results suggest an association between TGF-β signaling and NASH formation. Though TGF-β is well known to stimulate HSCs to induce fibrosis,[10, 24] the present study used WT and Tgfbr2ΔHEP mice to investigate the specific role of TGF-β signaling in hepatocytes in the development of NASH. WT and Tgfbr2ΔHEP mice fed on standard chow had similar metabolic phenotypes (Supporting Table 3). Twenty-two weeks of CDAA diet feeding induced significant hepatic steatosis, hepatocyte damage, and inflammatory cell infiltration in WT livers (Fig. 1B-D).[16-18] In contrast, CDAA diet Tgfbr2ΔHEP mice had a significant reduction in serum alanine aminotransferase (ALT) and TG levels and NAS, as determined by degree of hepatic steatosis, inflammation, and ballooning (Fig. 1B-D). Increased infiltration of hepatic macrophages and neutrophils in WT mice was reduced in Tgfbr2ΔHEP mice (Fig. 1B,E and Supporting Fig. 1). Consistent with histology results, hepatic expression of tumor necrosis factor alpha (TNF-α), interleukin (IL)-6, chemokine (C-C motif) ligand 2, and IL-1β was increased in WT mice fed on 22 weeks of CDAA diet, whereas these expression were decreased in Tgfbr2ΔHEP mice fed on the CDAA diet (Fig. 1F). These results indicate that TGF-β signaling in hepatocytes contributes to development of hepatic injury, steatosis, and inflammation in mice fed with a CDAA diet.

Figure 1.

TGF-β signaling in hepatocytes promotes NASH. (A-F) WT and Tgfbr2ΔHEP mice were fed a CSAA (CS; n = 10) or CDAA (CD; n = 8) diet for 22 weeks. (A, left) IHC staining of phosphorylated Smad2/3 in liver tissues from WT and Tgfbr2ΔHEP mice fed with 22 weeks of a CDAA diet. Original magnification: 200×. (A, middle) Immunoblots for phosphorylated Smad2 and Smad3, total Smad2, and Smad3 in liver tissues from WT and Tgfbr2ΔHEP mice are shown. (A, right) Hepatic Tgfb1 expression in WT mice fed CSAA and CDAA diets was measured by qPCR. (B) Liver sections were stained with H&E, Oil Red O, and IHC staining for F4/80. Original magnification: 200× for H&E and F4/80 staining and 100× for Oil Red O. (C, left) ALT level. (C, right) NAS. (D) Hepatic TG levels were measured. (E) F4/80-positive area was calculated. (F) Hepatic mRNA levels of inflammatory genes (Il1b, Il6, Tnf, Ccl2) were determined by qPCR. Data represent mean ± standard error of the mean. Open bar: WT mice; closed bar: Tgfbr2ΔHEP mice. *P < 0.05; **P < 0.01.

TGF-β Signaling in Hepatocytes Promotes Liver Fibrosis in NASH

Subsequently, we analyzed liver fibrosis in WT and Tgfbr2ΔHEP mice fed a CDAA diet for 22 weeks. Deposition of fibrillar collagen and HSC activation were significantly increased in WT mice, but were reduced in Tgfbr2ΔHEP mice, as assessed by Sirius Red staining and IHC for α-SMA (Fig. 2A-C). WT livers had increased messenger RNA (mRNA) expression of collagenα1(I), α-SMA, tissue inhibitor of metalloproteinase 1 (TIMP-1), TGF-β1, connective tissue growth factor (CTGF), and plasminogen activator inhibitor 1 (PAI-1), whereas these gene expressions were suppressed in Tgfbr2ΔHEP mice fed with the CDAA diet (Fig. 2D). These findings demonstrate that in addition to its known profibrogenic role in HSCs, TGF-β signaling in hepatocytes also possesses profibrogenic functions, inducing liver fibrosis in NASH.

Figure 2.

Tgfbr2 deficiency in hepatocytes attenuates liver fibrosis. (A-D) WT and Tgfbr2ΔHEP mice were fed a CSAA (CS; n = 10) or CDAA (CD; n = 8) diet for 22 weeks. (A) Liver sections were stained with Sirius Red and IHC staining for α-SMA. Original magnification: 200×. (B and C) Quantifications of Sirius Red staining (B) and staining for α-SMA (C). (D) Hepatic mRNA levels of fibrogenic markers (Col1a1, Acta2, Timp1, Tgfb1, Ctgf, and Pai1) were determined by qPCR. Data represent mean ± standard error of the mean. Open bar: WT mice; closed bar: Tgfbr2ΔHEP mice. *P < 0.05; **P < 0.01.

TGF-β Signaling Contributes to Hepatocyte Death in NASH

Because TGF-β signaling is associated with induction of cell death,[25] we investigated whether TGF-β signaling contributes to hepatocyte death in steatotic livers. CDAA diet feeding significantly increased the number of TUNEL-positive hepatocytes and expression of cleaved caspase-3 in WT mice fed a CDAA diet; however, these increases were not observed in Tgfbr2ΔHEP mice (Fig. 3A,B). Proapoptotic Bax gene was increased in livers of WT mice, but not in livers of Tgfbr2ΔHEP mice fed a CDAA diet (Fig. 3C). To examine the direct effect of TGF-β signaling in hepatocyte death, we purified primary hepatocytes from WT mice fed either a control CSAA diet or CDAA diet and challenged them with recombinant TGF-β. Similar viabilities of these hepatocytes were confirmed before TGF-β treatment. Expression of TGF-β receptor type I, but not type II, was increased in hepatocytes of CDAA-fed mice (Supporting Fig. 2). TGF-β treatment increased cell death only in lipid-accumulated hepatocytes purified from mice fed a CDAA diet, but not in hepatocytes with less lipid accumulation, purified from mice fed a CSAA diet (Fig. 3D). In response to TGF-β, lactate dehydrogenase (LDH) concentrations in the supernatant of hepatocytes from CDAA-diet-fed mice were elevated, when compared to supernatant from hepatocytes from CSAA-diet-fed mice (Fig. 3E). TGF-β treatment induced up-regulation and translocation of Bax into mitochondria in steatotic hepatocytes (Fig. 3F). These results indicate that fat accumulation increases susceptibility of hepatocytes to TGF-β-mediated cell death.

Figure 3.

TGF-β signaling promotes hepatocyte death in NAFLD. (A and B) WT and Tgfbr2ΔHEP mice were fed a CSAA (CS; n = 10) or CDAA (CD; n = 8) for 22 weeks. (A) TUNEL staining of liver tissues from WT and Tgfbr2ΔHEP mice fed with a CDAA diet. Original magnification: 200×. Representative pictures (left panel) and their quantification (right panel). (B) Immunoblots for cleaved and full-length caspase-3 of liver tissues of WT and Tgfbr2ΔHEP mice. (C) Hepatic expression of Bax mRNA was determined by qPCR. Open bar: WT mice; closed bar: Tgfbr2ΔHEP mice. (D-F) Primary hepatocytes were isolated from WT mice fed with a CDAA or CSAA diet and then treated with 10 ng/mL of TGF-β for 24 hours. (D) TUNEL staining of hepatocytes (left) and its quantification are shown (right). Pre, cell viability before TGF-β stimulation. (E) Concentrations of LDH released to supernatants were measured. (F, left) mRNA expression of Bax mRNA was determined by qPCR. (F, right) Hepatocytes were stained with Mito Tracker and anti-Bax Ab followed by assessment with confocal microscopy (original magnification: 600×). Data represent mean ± standard error of the mean; n = 3 for each group. *P < 0.05; **P < 0.01.

TGF-β Induces Cell Death in Lipid-Accumulated Hepatocytes Through Smad Signaling, but Independent of TAK1/Nuclear Factor Kappa B Signaling

Then, we tested whether TGF-β signaling induces cell death in steatotic hepatocytes developed by in vitro treatment with palmitate. As expected, TGF-β stimulation dramatically increased hepatocyte death with LDH released in palmitate-induced steatotic hepatocytes, but not in normal hepatocytes and oleate-treated hepatocytes (Fig. 4A and Supporting Fig. 3A,B). We found that TGF-β treatment induced Smad2 phosphorylation in normal hepatocytes, which was further potentiated in palmitate-treated steatotic hepatocytes (Fig. 4B). This led us to examine whether Smad activation is involved in TGF-β-induced cell death in fat-rich hepatocytes. We knocked down Smad2 by incubation of hepatocytes with siRNA-Smad2. When Smad2 was inactivated, increased TGF-β-induced cell death in lipid-accumulated hepatocytes was significantly suppressed with the decrease of LDH levels in culture supernatant (Fig. 4A). Thus, Smad signaling is an important pathway to contribute to cell death in lipid-accumulated hepatocytes in response to TGF-β. Then, we tested whether a Smad-independent TAK1-dependent pathway in TGF-β-mediated hepatocyte death is important. TAK1 activity was blocked in hepatocytes by treatment with 5z-7-oxozeaenol, a specific TAK1 inhibitor.[26] TAK1 inhibition increased cell death in TGF-β-treated hepatocytes and further increased cell death in hepatocytes treated with both palmitate and TGF-β (Fig. 4C). This result suggests that the TAK1-dependent pathway acts as a negative regulator for TGF-β-induced cell death in steatotic hepatocytes. Because TAK1 regulates activation of nuclear factor kappa B (NF-κB), JNK, and p38 in TGF-β signaling,[27, 28] we examined which downstream pathways contribute to the enhancement of TGF-β-mediated hepatocyte death. We blocked activation of NF-κB, JNK, and p38 by treatment with adenoviral IκB super-repressor, SP600125, and SB203580, respectively. Whereas inhibition of NF-κB increased TGF-β-induced death in steatotic hepatocytes, inhibition of JNK or p38 did not affect hepatocyte death (Fig. 4D and Supporting Fig. 4). Because TGF-β also activates the Rho/ROCK pathway, we examined its importance in TGF-β-mediated hepatocyte death.[29] Fasudil, a ROCK inhibitor, suppressed TGF-β-mediated death in steatotic hepatocytes (Fig. 4E). These results suggest that TGF-β-induced cell death in steatotic hepatocytes is dependent on Smad and Rho/ROCK signaling, but is negatively regulated by TAK1/NF-κB pathway.

Figure 4.

TGF-β-Smad signaling induces cell death in lipid-accumulated hepatocytes. (A-E) WT primary hepatocytes were treated with 10 ng/mL of TGF-β for 24 hours after pretretment with palmitate. (A) siRNA Smad2 was used for silencing Smad2 expression in hepatocytes. Quantification of TUNEL staining (left). Concentrations of LDH released to supernatants were measured (right). (B) Immunoblots for phosphorylated and total Smad2 and β-actin are shown. (C) TAK1 activity was inhibited by 250 nM of 5z-7-oxozeaenol. Quantifications of TUNEL staining (left) and concentrations of LDH in supernatants are shown (right). (D) Adenovirus IκB superrepressor (IκB-sr) was used for blocking NF-κB activation in hepatocytes. Quantifications of TUNEL staining (left) and concentrations of LDH in supernatants are shown (right). (E) Rho/ROCK activity was inhibited by 45 μM of fasudil. Quantifications of TUNEL staining (left) and concentrations of LDH in supernatants are shown (right). Data represent mean ± standard error of the mean; n = 3 for each group. *P < 0.05; **P < 0.01.

TGF-β and Palmitate Synergistically Produce ROS That Contributes to Hepatocyte Death

To investigate whether TGF-β-mediated ROS production in hepatocytes is involved in development of NASH, we examined lipid peroxidation in WT and Tgfbr2ΔHEP mice fed a CDAA diet by IHC for 4-HNE. Hepatic expression of 4-HNE increased in WT mice, but not in Tgfbr2ΔHEP mice fed a CDAA diet (Fig. 5A,B). Then, we investigated the effects of TGF-β-induced ROS in hepatocytes. Whereas TGF-β or palmitate alone induced ROS production, cotreatment with TGF-β and palmitate synergistically increased ROS production in hepatocytes (Fig. 5C). Antioxidant BHA inhibited ROS production and cell death induced by cotreatment with TGF-β and palmitate (Fig. 5D,E), suggesting that ROS production induced by TGF-β and palmitate cotreatment contributes to hepatocyte death.

Figure 5.

TGF-β-induced ROS production mediates hepatocyte death in NASH. (A and B) WT and Tgfbr2ΔHEP mice were fed a CSAA (CS; n = 10) or CDAA (CD; n = 8) for 22 weeks. (A) IHC for 4-HNE of liver tissues from WT and Tgfbr2ΔHEP mice fed with a CDAA diet. Original magnification: 320×. (B) Quantification of 4-HNE staining. (C) Primary WT hepatocytes were incubated with DCFDA (10 μM) and then stimulated with 10 ng/mL of TGF-β. Hepatocytes were preincubated with or without antioxidant BHA (100 μM) for 30 minutes before TGF-β stimulation. Fluorescent signals were quantified for 30 minutes using a fluorometer at excitation and emission wavelengths of 488 and 520 nm, respectively. (D and E) WT primary hepatocytes were treated with 10 ng/mL of TGF-β for 24 hours after pretretment with palmitate. (D) ROS production was inhibited by BHA (100 μM). Quantifications of TUNEL staining and (E) concentrations of LDH in supernatants are shown. DCFDA, dichlorofluorescin diacetate.

TGF-β Signaling Is Involved in Lipid Metabolism in Hepatocytes

Because hepatic lipid accumulation was suppressed in Tgfbr2ΔHEP mice, compared to WT mice, we analyzed the gene expression involved in lipid metabolism in the liver. Upon CDAA diet feeding, WT livers increased the gene expression associated with lipogenesis, including Dgat1, Pparg, and Srebp1c, whereas livers of Tgfbr2ΔHEP did not (Fig. 6A). In contrast, gene expression related to β-oxidation, such as Acox1, Cpt1, and Hmgcs2, increased in Tgfbr2ΔHEP mice, but not in WT mice (Fig. 6B). These findings prompted us to examine whether TGF-β signaling affects lipid accumulation in primary cultured hepatocytes. TGF-β stimulation alone had little effect on lipid accumulation. Interestingly, TGF-β stimulation potentiated hepatocyte lipid accumulation induced by palmitate (Fig. 6C). Consistent with lipid accumulation, hepatocyte TG levels were also potentiated by cotreatment with TGF-β and palmitate, but not oleate (Fig. 6D and Supporting Fig. 3C). These results indicate that TGF-β signaling has an effect on modulating lipid metabolism in hepatocytes.

Figure 6.

TGF-β signaling is associated with gene expression involved in lipogenesis and β-oxidation in fatty liver disease. (A-D) WT and Tgfbr2ΔHEP mice were fed a CSAA (CS; n = 10) or CDAA (CD; n = 8) for 22 weeks. (A and B) Hepatic mRNA expression of Dgat1, Pparg, Srebp1c (A), Acox1, Cpt1, and Hmgcs2 (B) were determined by qPCR. Open bar: WT mice; closed bar: Tgfbr2ΔHEP mice. (C and D) WT primary hepatocytes were treated with 10 ng/mL of TGF-β for 24 hours after pretretment with palmitate. (C) Representative pictures of Oil Red O staining are shown. Original magnifications: 400×. (D) TG concentrations in hepatocytes were measured. Data represent mean ± standard error of the mean. *P < 0.05; **P < 0.01.

Smad Signaling Is Crucial for TGF-β-Mediated Modulation of Lipid Metabolism in Hepatocytes

Next, we determined whether the TGF-β/Smad pathway is required for regulation of lipid metabolism in hepatocytes. Increased lipid accumulation and TG levels in hepatocytes treated with TGF-β and palmitate were decreased when Smad2 was knocked down by siRNA (Fig. 7A,B). In the context of decreased lipid accumulation and TG levels in Smad2-silenced cells, gene expression involved in lipogenesis, including Dgat1, Fas, and Srebp1c, were suppressed, whereas gene expression related to β-oxidation (Ppara, Acox1, and Cpt1) were increased in hepatocytes treated with TGF-β and palmitate when Smad2 was inactivated (Fig. 7C,D). These results suggest that TGF-β/Smad signaling regulates lipogenesis and β-oxidation-related gene expression that affects palmitate-induced lipid accumulation in hepatocytes.

Figure 7.

TGF-β-Smad signaling regulates lipid metabolism in lipid-accumulated hepatocytes. (A-D) WT primary hepatocytes were treated with 10 ng/mL of TGF-β for 24 hours after pretretment with palmitate. siRNA Smad2 was used for silencing Smad2 expression in hepatocytes. (A) Representative pictures of Oil Red O staining. Original magnifications: 400×. (B) TG concentrations in hepatocytes. (C and D) mRNA expression of Dgat1, Fas, Srebp1c (C), Ppara, Acox1, and Cpt1 (D) were measured by qPCR. n = 3 for each group. *P < 0.05; **P < 0.01.

Hepatic-Specific Deletion of Tgfbr2 Influences Systemic Insulin Sensitivity and Metabolic Phenotypes in CDAA-Fed Mice

To test whether hepatocyte TGF-β signaling affects systemic IR in mice fed on 22 weeks of a CDAA diet, GTTs and ITTs were examined. GTT demonstrated that increased blood-sugar levels of CDAA-diet fed WT mice started to decrease after 60 minutes of glucose administration and did not decrease to basal levels within the 120-minute time point. In contrast, Tgfbr2ΔHEP mice displayed a rapid reduction in blood-sugar levels, compared to WT mice, after glucose challenge (Supporting Fig. 5A). Moreover, WT mice fed a CDAA diet exhibited IR, but Tgfbr2ΔHEP mice fed a CDAA diet were more sensitive to insulin than WT mice (Supporting Fig. 5B). Whereas Tgfbr2ΔHEP mice exhibited lower body weight and liver/body-weight ratio than WT mice fed a CDAA diet (Supporting Table 3), food intake was similar between WT and Tgfbr2ΔHEP mice (Supporting Fig. 6). These results suggest that TGF-β signaling in hepatocytes may contribute to systemic IR and increased body weight in mice fed a CDAA diet during development of NASH.

TGF-β Signaling in Hepatocytes Is Activated in Human NASH Livers

Last, we examined activation of TGF-β signaling in liver biopsy specimens of patients with NAFL (NAS score between 0 and 3) without fibrosis and patients with NASH (NAS score of 5 or higher) with severe fibrosis and hepatocyte ballooning by IHC for phospho-Smad2/3. In liver tissue from mild NAFL patients, we observed weak and no phosphorylation of Smad2/3 in hepatocytes (Fig. 8). In contrast, we observed increased phosphorylation and nuclear translocation of Smad2/3 in liver parenchymal cells in liver tissue of NASH patients (Fig. 8), confirming the clinical relevance of activation of TGF-β signaling in hepatocytes in human NASH.

Figure 8.

Human NASH livers displayed activation of TGF-β signaling. IHC for phospho-Smad2/3 in liver biopsy samples from patients with mild NAFLD without fibrosis (n = 10) and samples from patients with NASH with severe fibrosis (n = 10) was performed. (A) Representative pictures are shown. Original magnification: 400×. (B) Quantification of phospho-Smad2/3-positive hepatocytes.

Discussion

Activation of TGF-β signaling is associated with formation of fibrotic scar tissue in the liver.[30] Activated HSCs are key players in responding to TGF-β, producing scar tissues consisting of collagen fiber. Alternatively, phosphorylation and nuclear translocation of Smad2/3 was increased in hepatocytes treated with TGF-β as well as in mouse and human NASH livers (Figs. 1, 4 and 8).[25] However, it was unclear that whether TGF-β signaling specifically in hepatocytes contributes to progression of NASH and whether TGF-β-mediated hepatic phenotypes affect systemic metabolic manifestations. The present study demonstrated that TGF-β signaling in hepatocytes promoted hepatic steatosis, hepatocyte damage, inflammatory cell infiltration, inflammatory cytokine production, HSC activation, and fibrosis. TGF-β signaling participated in ROS production and hepatocyte death in lipid-accumulated hepatocytes (Figs. 4 and 5). TGF-β signaling was also involved in lipid metabolism by regulation of gene expression involved in lipogenesis and fatty acid β-oxidation. Intriguingly, hepatocyte's TGF-β signaling influenced systemic IR and body-weight gain.

Previous studies demonstrated that Smad3 deficiency results in decreased obesity and systemic IR on an HFD.[14, 15] Systemic Smad3 deficiency does not modulate the gene expression associated with β-oxidation and lipolysis in skeletal muscles and the liver. The responsible site of Smad3 signaling for systemic metabolic manifestations is peripheral adipose tissue in HFD-fed mice.[15] In white adipose tissues of Smad3-deficient mice, expression of lipogenic PPAR-γ, sterol regulatory element-binding protein 1c, and fatty acid synthase was suppressed, whereas expression of molecules related to β-oxidation and lipolysis, such as Acyl-coenzyme A oxidase 1, carnitine palmitoyl transferase 1, and PPAR-δ, was increased. In WT adipocytes, CCAAT/enhancer-binding protein beta (C/EBPβ) interacts with histone deacetylase 1, a transcriptional repressor, which binds to PPAR-δ promoter to inhibit expression of β-oxidation-related genes.[15] C/EBPβ also binds to PPAR-γ to increase lipogenesis gene expression that results in adipogenesis in WT adipocytes.[15] In contrast, Smad3 deficiency increases C/EBP homologous protein-10 expression, which interferes with interaction of C/EBPβ to the promoter of PPAR-δ or PPAR-γ, leading to inhibition of adipogenesis.[15] Moreover, Smad3 can directly bind to the PGC-1α promoter to repress its promoter activity, contributing to mitochondrial biogenesis, including β-oxidation.[14] Alternatively, Smad3 deficiency leads to changes in phenotype of macrophages in adipose tissues from M1 to M2, suggesting that Smad3 signaling in macrophages is also important for adipogenesis.[14]

In addition to previous studies, our results clearly showed that TGF-β signaling regulates lipid metabolism through induction of lipogenesis genes and suppression of β-oxidation-related genes. TGF-β signaling itself did not possess a significant effect on hepatocyte lipid metabolism. In combination with FFA challenges, TGF-β signaling modulated the gene expression involved in lipogenesis and fatty acid β-oxidation, which resulted in increased TG synthesis and lipid accumulation in hepatocytes. TGF-β signaling activates the Smad2/3/4-signaling pathway, and inhibition of these pathways abolished changes in gene expression associated with lipid metabolism. In alcoholic liver disease, TGF-β signaling increases hepatic lipid accumulation through down-regulation of alcohol dehydrogenase 1.[25] In conclusion, these evidences indicate that hepatocytic TGF-β signaling enhances fat deposition in hepatocytes induced by excessive fat loading.

TGF-β signaling is also involved in hepatocyte death. Treatment with TGF-β alone does not cause hepatocyte death, but induces significant death in fat-accumulated hepatocytes. Inhibition of Smad2 inhibited TGF-β-induced death in fat-accumulated hepatocytes. As previously reported, inhibition of NF-κB increases susceptibility of hepatocytes to death induced by TNF-α or IL-1β.[31] Similarly, TGF-β is able to induce hepatocyte death by inhibition of NF-κB or TAK1, which is further enhanced in lipid-accumulated hepatocytes.[32] Importantly, expression of TGF-β receptor type I was increased in lipid-accumulated hepatocytes (Supporting Fig. 2). Moreover, TGF-β-induced ROS production was dramatically increased in palmitate-treated hepatocytes, and inhibition of ROS suppressed TGF-β-mediated hepatocyte death (Fig. 5). These findings suggest that lipid-accumulated hepatocytes enhance TGF-β-mediated ROS production, resulting in the enhancement of Smad activation and hepatocyte death.[33] In addition to the conventional intracellular signaling pathways activated by TGF-β, such as Smad- and TAK1-dependent pathways, a previous study demonstrated that p53-mediated signaling is also required for TGF-β-mediated hepatocyte death.[34] We additionally determined that Rho/ROCK signaling is an important component for TGF-β-mediated hepatocyte death in fatty liver cells.

The present study suggests that TGF-β-mediated liver pathology could influence body-weight gain and systemic IR, as analyzed by GTTs and ITTs. It has been shown that systemic inflammation influences metabolic manifestations, including weight gain and IR.[35, 36] Thus, liver inflammation caused by hepatocyte TGF-β-Smad signaling may influence systemic or adipose tissue inflammation that affect body weight and systemic IR. This may further contribute to increased liver steatosis and inflammation. Interestingly, whereas Tgfbr2ΔHEP mice are resistance to GTTs and ITTs, there are no significant differences in fasting blood-sugar levels between WT and Tgfbr2ΔHEP mice. This suggests that TGF-β signaling in adipose tissues is essential to regulate fasting blood-sugar levels, but hepatic IR determined by hepatic TGF-β signaling may contribute to systemic IR.

In summary, the results from our study revealed that TGF-β signaling in hepatocytes contributes to lipid accumulation by deregulation of lipid metabolism and enhancement of cell death in lipid-accumulated hepatocytes, resulting in the development of hepatic steatosis, hepatocyte death, inflammation, and fibrosis. Currently, fresolimumab, a human monoclonal Ab against TGF-β, is being tested for clinical trials for idiopathic pulmonary fibrosis, renal disease, and cancer.[37, 38] Our study provides evidence that targeting TGF-β signaling could be an effective therapy of NASH.

Ancillary