Role of vascular endothelial growth factor in the pathophysiology of nonalcoholic steatohepatitis in two rodent models

Authors

  • Stephanie Coulon,

    Corresponding author
    1. Department of Gastroenterology and Hepatology, Ghent University Hospital, Ghent, Belgium
    • Department of Hepatology and Gastroenterology, Ghent University Hospital, 1K12 IE, De Pintelaan 185, 9000 Ghent, Belgium
    Search for more papers by this author
    • fax: +32-9-3324984

  • Vanessa Legry,

    1. Laboratory of Hepato-Gastroenterology, Institut de Recherche Expérimentale et Clinique,Université Catholique Louvain, Brussels, Belgium
    Search for more papers by this author
    • S. Coulon received a scholarship from the Ghent University Research Fund (BOF 09/24J/012). Heindryckx F. received a scholarship (FWO09/ASP/161) from the Fund for Scientific Research (FWO-Flanders). I. Colle and H. Van Vlierberghe received a fundamentally clinical mandate of FWO Flanders. The Department Gastroenterology and Hepatology of the Ghent University Hospital received unrestricted funding from Bayer, Roche, Astellas, Ferring, and MSD. The work of P. Carmeliet is supported by long-term structural funding: Methusalem funding by the Flemish Government. The work of V. Legry is supported by the Fund for Scientific Research (FNRS/FRSM 3.4520.10). I. Leclercq is an FNRS research associate.

  • Femke Heindryckx,

    1. Department of Gastroenterology and Hepatology, Ghent University Hospital, Ghent, Belgium
    Search for more papers by this author
  • Christophe Van Steenkiste,

    1. Department of Gastroenterology and Hepatology, Ghent University Hospital, Ghent, Belgium
    Search for more papers by this author
  • Christophe Casteleyn,

    1. Department of Veterinary Sciences, University of Antwerp, Wilrijk, Belgium
    Search for more papers by this author
  • Kim Olievier,

    1. Department of Gastroenterology and Hepatology, Ghent University Hospital, Ghent, Belgium
    Search for more papers by this author
  • Louis Libbrecht,

    1. Department of Morphology, Ghent University Hospital, Ghent, Belgium
    Search for more papers by this author
  • Peter Carmeliet,

    1. Laboratory of Angiogenesis and Neurovascular Link, Vesalius Research Center, VIB, KU Leuven, Leuven, Belgium
    Search for more papers by this author
    • S. Coulon received a scholarship from the Ghent University Research Fund (BOF 09/24J/012). Heindryckx F. received a scholarship (FWO09/ASP/161) from the Fund for Scientific Research (FWO-Flanders). I. Colle and H. Van Vlierberghe received a fundamentally clinical mandate of FWO Flanders. The Department Gastroenterology and Hepatology of the Ghent University Hospital received unrestricted funding from Bayer, Roche, Astellas, Ferring, and MSD. The work of P. Carmeliet is supported by long-term structural funding: Methusalem funding by the Flemish Government. The work of V. Legry is supported by the Fund for Scientific Research (FNRS/FRSM 3.4520.10). I. Leclercq is an FNRS research associate.

  • Bart Jonckx,

    1. ThromboGenics NV, Heverlee, Belgium
    Search for more papers by this author
  • Jean-Marie Stassen,

    1. ThromboGenics NV, Heverlee, Belgium
    Search for more papers by this author
  • Hans Van Vlierberghe,

    1. Department of Gastroenterology and Hepatology, Ghent University Hospital, Ghent, Belgium
    Search for more papers by this author
    • S. Coulon received a scholarship from the Ghent University Research Fund (BOF 09/24J/012). Heindryckx F. received a scholarship (FWO09/ASP/161) from the Fund for Scientific Research (FWO-Flanders). I. Colle and H. Van Vlierberghe received a fundamentally clinical mandate of FWO Flanders. The Department Gastroenterology and Hepatology of the Ghent University Hospital received unrestricted funding from Bayer, Roche, Astellas, Ferring, and MSD. The work of P. Carmeliet is supported by long-term structural funding: Methusalem funding by the Flemish Government. The work of V. Legry is supported by the Fund for Scientific Research (FNRS/FRSM 3.4520.10). I. Leclercq is an FNRS research associate.

  • Isabelle Leclercq,

    1. Laboratory of Hepato-Gastroenterology, Institut de Recherche Expérimentale et Clinique,Université Catholique Louvain, Brussels, Belgium
    Search for more papers by this author
    • S. Coulon received a scholarship from the Ghent University Research Fund (BOF 09/24J/012). Heindryckx F. received a scholarship (FWO09/ASP/161) from the Fund for Scientific Research (FWO-Flanders). I. Colle and H. Van Vlierberghe received a fundamentally clinical mandate of FWO Flanders. The Department Gastroenterology and Hepatology of the Ghent University Hospital received unrestricted funding from Bayer, Roche, Astellas, Ferring, and MSD. The work of P. Carmeliet is supported by long-term structural funding: Methusalem funding by the Flemish Government. The work of V. Legry is supported by the Fund for Scientific Research (FNRS/FRSM 3.4520.10). I. Leclercq is an FNRS research associate.

  • Isabelle Colle,

    1. Department of Gastroenterology and Hepatology, Ghent University Hospital, Ghent, Belgium
    Search for more papers by this author
    • S. Coulon received a scholarship from the Ghent University Research Fund (BOF 09/24J/012). Heindryckx F. received a scholarship (FWO09/ASP/161) from the Fund for Scientific Research (FWO-Flanders). I. Colle and H. Van Vlierberghe received a fundamentally clinical mandate of FWO Flanders. The Department Gastroenterology and Hepatology of the Ghent University Hospital received unrestricted funding from Bayer, Roche, Astellas, Ferring, and MSD. The work of P. Carmeliet is supported by long-term structural funding: Methusalem funding by the Flemish Government. The work of V. Legry is supported by the Fund for Scientific Research (FNRS/FRSM 3.4520.10). I. Leclercq is an FNRS research associate.

  • Anja Geerts

    1. Department of Gastroenterology and Hepatology, Ghent University Hospital, Ghent, Belgium
    Search for more papers by this author

  • Potential conflict of interest: Dr. Stassen own stock and consults for ThromboGenics.

Abstract

The pathophysiology of nonalcoholic steatohepatitis (NASH) should be approached as a multifactorial process. In several stages of NASH, a link between disease progression and hepatic microvasculature changes can be made. In this study we investigated the role of angiogenesis in two mouse models for NASH, and the effect of a preventive and therapeutic antiangiogenic treatment in a diet-induced mouse model for NASH. Protein and RNA levels of angiogenic and inflammatory factors were significantly up-regulated in the liver of C56BL/6 and db/db mice with NASH at different timepoints. To examine the effect of angiogenic factors on the disease progression of NASH, a prevention and treatment study was set up, blocking the placental growth factor (PlGF) or vascular endothelial growth factor receptor 2 (VEGFR2). Our study showed that treatment prevents the progression of NASH by attenuating steatosis and inflammation, both in a preventive and therapeutic setting, thereby confirming the hypothesis that angiogenic factors play an early role in the disease progression from steatosis to NASH. Anti-PlGF (αPlGF) did not significantly improve liver histology. Vascular corrosion casting showed a more disrupted liver vasculature in mice with NASH compared to controls. Treatment with αVEGFR2 showed an improvement of the liver vasculature. Moreover, fat-laden primary hepatocytes treated with αVEGFR2 stored significantly less lipids. Conclusion: Our results demonstrate that there is an increased expression of angiogenic factors in the liver in different mouse models for NASH. We found that VEGFR2 blockage attenuates steatosis and inflammation in a diet-induced mouse model for NASH in a preventive and therapeutic setting. Our findings warrant further investigation of the role of angiogenesis in the pathophysiology in NASH. (HEPATOLOGY 2013)

Nonalcoholic steatohepatitis (NASH) is the most severe form of nonalcoholic fatty liver disease (NAFLD) and a serious consequence of the current obesity epidemic.1 NASH is present in more than one-third of the NAFLD cases and is recognized as a potentially progressive disease that may cause fibrosis, cirrhosis, and hepatocellular carcinoma (HCC).2 At present, a multimodal treatment plan that targets obesity, insulin resistance, hyperlipidemia, and hypertension appears to be the only effective means of improving NASH.3 The two-hit theory, proposed in 1998 by James and Day,4 is the first theory that gave a plausible explanation for the pathogenesis of NASH. This hypothesis suggests that the first hit is caused by steatosis and the second hit is a synergy of oxidative stress and inflammation. Recently, Tilg and Moschen5 described the inflammatory process as a multiple parallel theory. However, the pathogenesis of NASH is still not fully understood. The recognized mechanisms as stated above do not fully explain the range of symptoms and physiological processes found in the disease progression. Nonetheless, the pathophysiology of NASH should be approached as a multifactorial process. In several stages of NASH, a link might be made between disease progression and hepatic microvasculature changes such as angiogenesis.

Disruption of the liver vascular architecture, in particular angiogenesis, has been linked to progression of fibrosis, cirrhosis, and HCC in chronic liver diseases.6, 7 In previous reports, our group was able to demonstrate a role for vascular endothelial growth factor (VEGF) and placental growth factor (PlGF) in the pathogenesis of portal hypertension, fibrosis, and HCC.8-10 Some studies suggest that angiogenesis plays a role in the progression of NASH. This was first brought to light in a study by Kitade et al.,11 which suggested that leptin-mediated neovascularization, coordinated by VEGF, is important in the development of liver fibrosis and HCC in a rat model for NASH. A macro-array gene expression analysis on the liver of obese patients with severe NASH showed that VEGF, transforming growth factor beta, connective tissue growth factor, and fibroblast growth factor were overexpressed compared to control patients.12 Kitade et al.13 described a significant up-regulation of CD34 expression, which is widely used as a marker of neovascularization, in liver biopsies of patients with NASH. Recently, it was found that VEGF is up-regulated in the serum of biopsy-proven steatosis and NASH patients compared to healthy controls.14, 15 These new findings could give a new perspective to investigate the pathophysiology of NASH.

In this study we determined the role of angiogenesis at several timepoints in the pathophysiology of NASH in different mouse models. Moreover, we assessed whether inhibition of angiogenic factors could serve as a potential treatment of NASH. Therefore, we looked at the effect of anti-PlGF (αPlGF) and anti-VEGFR2 (αVEGFR2) in vivo, using a prevention and treatment study in a mouse model for NASH, and in vitro, using fat-laden primary hepatocytes.

Materials and Methods

Animal Experiments.

Ten-week-old C57BL/6 and homozygous db/db female mice (Charles River Laboratories, Brussels, Belgium) were kept under constant temperature and humidity on a 12-hour controlled dark/light cycle. Mice had ad libitum access to food and water. C57BL6/J and db/db mice were fed a methionine and choline-deficient (MCD) diet (MP Biomedicals, Brussels, Belgium) for 3 days, 1 week, 2 weeks, 4 weeks, or 8 weeks (n = 8/group) in order to develop NASH.16 Control groups received an identical diet to which choline bitartrate (2 g/kg) and DL-methionine (3 g/kg) was added (MP Biomedicals). Daily food intake was measured during the experiment. The Ethical Committee of Experimental Animals at the Faculty of Medicine and Health Sciences, Ghent University, approved the protocols.

Prevention and Treatment Study.

Anti-VEGFR2 (αVEGFR2; DC101) (ThromboGenics, Leuven, Belgium) and anti-PlGF (αPlGF; 5D11D4) (ThromboGenics) are known angiogenic inhibitors.17, 18 Anti-VEGFR2 (40 mg/kg body weight) and αPlGF (25 mg/kg body weight) were administered intraperitoneally to age- and weight-matched C57BL/6 mice for 8 weeks, two times a week (n = 10/group). Control mice were injected with phosphate-buffered saline (PBS) following the same dose and time schedule (n = 10/group). In the prevention study, mice were fed an MCD or control diet and were sacrificed after 8 weeks. For the treatment study, αVEGFR2 treatment was given 2 weeks after starting the MCD diet, when mice already developed steatosis, inflammation, and ballooning.

Tissue Sampling and Histology.

Mice were sacrificed under isoflurane anesthesia (Forene, Hoofddorp, The Netherlands) while blood was obtained from the carotid artery. The liver and spleen were rapidly excised and weighed. The left liver lobe was fixed in 4% phosphate buffered formaldehyde (Klinipath, Olen, Belgium). The right liver lobe was collected in RNAlater (Qiagen, Venlo, The Netherlands) and snap-frozen in liquid nitrogen. The left liver lobe was embedded in paraffin, histologically processed, and sections were cut and stained with hematoxylin and eosin staining (H&E) and Sirius Red. All stainings were performed using standard histology protocols and evaluated by an experienced pathologist. The degree of steatosis, lobular inflammation, and ballooning were defined as stated previously.19 Degree of steatosis, defined as the percentage of hepatocytes containing fat droplets, was scored using the following scale: 0 (<5%), 1 (5%-33%), 2 (>33%-66%), 3 (>66%). Foci of lobular inflammation were defined as two or more inflammatory foci (averaged from 3-4 200× fields) and scored as: 0 (no foci), 1 (<2 foci), 2 (2-4 foci), 3 (>4 foci). Ballooning was scored according to number of ballooned hepatocytes: 0 (none), 1 (few), 2 (many). The degree of fibrosis was evaluated separately and scored as: 0 (none), 1 (zone 3 perisinusoidal or portal fibrosis), 2 (zone 3 perisinusoidal and periportal fibrosis without bridging), 3 (bridging fibrosis), 4 (cirrhosis).

Cell Isolation, Culture, and Treatments.

Primary hepatocytes were prepared by in situ perfusion and collagenase digestion (Liberase Blendzymes, Roche) of livers of adult female C57BL/6 mice as described.20 Cells were plated at a density of 2.5 × 104 cells per well on collagen I-coated 96-well plates (Greiner Bio-One, Frickenhausen, Germany) and cultured in Dulbecco's modified Eagle's medium (DMEM)/F12 supplemented with 10% FBS, ITS (insulin 5 μg/mL, transferrin 5 μg/mL, selenium 5 ng/mL) and 1% streptomycin and penicillin in 5% CO2 at 37°C. Two hours after plating, medium was replaced by the same culture medium but without ITS. The cells were used for experiments after an overnight incubation.21 Oleic and palmitic acids stock solutions (100 mM) were prepared in 0.1 M NaOH at 70°C as previously described.20 A 5 mM free fatty acid (FFA) / 5% bovine serum albumin (BSA) working solution was prepared by complexing an appropriate volume of stock solution to 5% BSA (FFA-free low endotoxin; Sigma-Aldrich, Bornem, Belgium) in a 60°C water bath. After filtration and cooling, a mixture of oleic and palmitic acids was prepared at a molar ratio of 2:1. After a 3-hour serum deprivation, hepatocytes were treated 24 hours with 100 μg/mL IgG, 100 μg/mL αVEGFR2, 150 ng/mL VEGF, or 100 μg/mL αVEGFR2 and 150 ng/mL VEGF diluted in DMEM/F12. Thereafter, 1 mM of the oleic:palmitic acid mixture was added to cultures for a 6-hour treatment. To detect intracellular lipid droplets accumulation, the cells were brought to room temperature and the medium was replaced with 200 μL PBS. Five μL of AdipoRed reagent (Lonza, Walkersville, MD) were added in each well and the plates were incubated at room temperature for 10 minutes. The relative fluorescence was measured (λ excitation at 485 nm, λ emission at 572 nm) using a fluorescence spectrometer (HTS-7000 Plus-plaque-lecteur, Perkin Elmer). Each fluorescence value was normalized to DNA content. The analyses were performed on three independent cell isolates in sextuplicate.

Additional Methods.

Methods describing biochemical serum analysis, immunohistochemistry, protein expression, vascular corrosion casting, quantitative real-time polymerase chain reaction (PCR) and statistics are provided in the Supporting Materials and Methods.

Abbreviations

αPlGF, antiplacental growth factor; αSMA, alpha-smooth muscle actin; αVEGFR2, antivascular endothelial growth factor receptor 2; ALT, alanine aminotransferase; AST, aspartate aminotransferase; HCC, hepatocellular carcinoma; HSC, hepatic stellate cells; Il1b, interleukin 1b; L-fabp1, liver fatty acid binding protein 1; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; MCD, methionine choline-deficient; PlGF, placental growth factor; Scd1, stearoyl-CoA desaturase 1; TG, triglyceride; Tnf, tumor necrosis factor alpha; VEGF, vascular endothelial growth factor; Vwf, Von Willebrand factor

Results

General Characteristics of MCD Diet Fed C57BL/6 and db/db Mice at Different Timepoints.

Both C56BL6/J and db/db mice on an MCD diet displayed significant weight loss after 3 days of MCD diet (P < 0.01) (Table 1). After 8 weeks of the MCD diet both mouse models lost 40% of their initial body weight. Liver/body weight ratio significantly augmented after 3 days of MCD diet in C57/BL6 and after 2 weeks of MCD diet db/db mice (P < 0.05) (Table 1). These alterations can be taken into account for the onset of steatosis. Db/db mice had a significantly higher food consumption compared to C57BL6/J mice. Nonetheless, there was no significant difference in food consumption between mice fed an MCD or a control diet (Table 1). Biochemical analysis of serum of C57BL6/J and db/db mice showed significantly increased alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels when fed a MCD diet (Table 1).

Table 1. Clinical and Biochemical Characteristics of C57BL6/J and db/db Mice Given MCD or Control Diet for 3 Days, 1 Week, 2 Weeks, 4 Weeks or 8 Weeks. (n = 8 Mice/Group)
 Control3 Days MCD Diet1 Week MCD Diet2 Weeks MCD Diet4 Weeks MCD Diet8 Weeks MCD Diet
C57BL/6db/dbC57BL/6db/dbC57BL/6db/dbC57BL/6db/dbC57BL/6db/dbC57BL/6db/db
  • Data are presented as mean ± SEM. Asterisks represent P-values:

  • *

    P < 0.05;

  • **

    P < 0.01;

  • ***

    P < 0.001 versus control diet. Alanine aminotransferase (ALT), aspartate aminotransferase (AST), body weight (BW), methionine choline-deficient (MCD), not available (na).

Food/week (g)8.9 ± 1.2313.11 ± 1.48nana11.5 ± 2.112.8 ± 0.58.6 ± 0.512.1 ± 1.38.3 ± 1.912.3 ± 0.97.8 ± 1.711.89 ± 1.9
BW (g)20.7 ± 0.345.1 ± 0.917.6 ± 0.3*** 38.5 ± 1.8** 16.2 ± 0.3*** 41.2 ± 1.1* 14.9 ± 0.2*** 40.0 ± 1.5* 13.3 ± 0.3*** 38.8 ± 1.7* 12.1 ± 0.2*** 26.8 ± 1.0***
Liver/BW (%)4.6 ± 0.14.2 ± 0.24.9 ± 0.1* 5.9 ± 0.35.3 ± 0.2** 6.0 ± 0.15.0 ± 0.26.2 ± 0.2* 5.2 ± 0.2** 6.4 ± 0.1** 5.4 ± 0.2** 7.8 ± 0.1***
Spleen/BW (%)0.35 ± 0.012.16 ± 0.210.31 ± 0.01* 2.11 ± 0.270.32 ± 0.012.17 ± 0.150.27 ± 0.02** 1.88 ± 0.250.28 ± 0.03* 1.52 ± 0.13* 0.24 ± 0.01*** 1.58 ± 0.08**
ALT (U/I)24.7 ± 0.8112.8 ± 11.135.8 ± 3.1*** 85.3 ± 10.9133.4 ± 14.6*** 514.9 ± 102.7*** 173.9 ± 30.4*** 640.5 ± 47.5*** 235.6 ± 79.6*** 347.7 ± 27.6*** 204.7 ± 43.4*** 488.1 ± 41.4***
AST (U/I)138.6 ± 5.9167.7 ± 13.0154.7 ± 5.9152.6 ± 10.4254.8 ± 17.6*** 331.7 ± 52.43* 293.8 ± 29.6*** 526.3 ± 39.8*** 347.4 ± 54.4*** 378.7 ± 33.7*** 337.3 ± 53.9*** 417.4 ± 38.5***

Steatosis, inflammation, ballooning, and fibrosis were assessed histologically using H&E and Sirius Red staining. From 4 weeks onwards, liver sections of the C57BL6/J mice fed an MCD diet were scored as NASH (Fig. 1A). The liver of C57BL6/J mice fed the control diet were normal, whereas liver sections of mice on 8 weeks of the MCD diet showed large fat droplets, clusters of inflammatory cells, and ballooned hepatocytes (Fig. 1B,C). Liver histology of db/db mice fed a control diet showed steatosis and ballooning. Db/db mice on MCD diet developed NASH after 1 week of the MCD diet, which worsened visibly after 8 weeks (Fig. 1D-F). C57BL6/J and db/db mice developed mild portal to pericellular fibrosis after 8 weeks of the MCD diet (Supporting Fig. 1A-D).

Figure 1.

Liver histology of C57BL6/J and db/db mice fed an MCD diet or control diet. (A) Overview of the histological scores at different timepoints of C57BL6/J mice fed an MCD or control diet. Mice develop NASH after 4 weeks of MCD diet. (B,C) H&E staining of liver of mice C57BL6/J fed a control or MCD diet for 8 weeks (200×). (D) Overview of the histological scores of db/db mice fed an MCD or control diet at different timepoints. Mice develop NASH after 1 week of MCD diet. (E,F). H&E staining of liver of db/db mice fed a control or MCD diet for 8 weeks (200×). Scale bars = 100 μm.

Effect of MCD Diet on Inflammation and Lipid Metabolism in C57BL/6 and db/db Mice at Different Timepoints.

Inflammatory foci increased in the liver of C57BL5/J and db/db mice from 1 week of an MCD diet onwards compared to control mice. Furthermore, F4/80 staining significantly increased in C57BL6/J and db/db mice fed the MCD diet (P < 0.05) (Fig. 2A,B). Kupffer cells in animals fed a normal diet remained isolated, while the MCD diet caused recruitment of Kupffer cells into clusters (Supporting Fig. 2A,D). Increased tumor necrosis factor alpha (Tnf) and interleukin 1b (Il1b) gene expression confirmed inflammation. We found a significant up-regulation of these inflammatory genes after 1 week in C57BL6/J and db/db mice till 8 weeks of the MCD diet compared to controls (P < 0.05) (Fig. 2C,D).

Figure 2.

Inflammation and lipid metabolism during the disease progression of NASH in C57BL6/J and db/db mice fed an MCD or control diet. (A) F4/80 staining showed increased macrophage recruitment after 1 week in the liver of C57BL6/J mice fed an MCD diet compared to control mice. (B) F4/80 staining showed increased macrophage recruitment after 3 days in the liver of db/db mice fed an MCD diet compared to control mice. (C) Gene expression of inflammatory genes (Tnfα and Il1b) in C57BL6/J mice fed MCD or control diet at different timepoints (n = 8/group). (D) Gene expression of inflammatory genes (Tnfα and Il1b) in db/db mice fed MCD or control diet at different timepoints (n = 8/group). (E) Gene expression of lipogenic genes (Scd1 and L-fabp1) in C57BL6/J mice fed MCD or control diet at different timepoints (n = 8/group). (F) Gene expression of lipogenic genes (Scd1 and L-fabp1) in db/db mice fed MCD or control diet at different timepoints (n = 8/group). P-values are presented as: (a) P < 0.05; (b) P < 0.01; (c) P < 0.001 versus control diet.

Moreover, liver fatty acid binding protein 1 (L-fabp1), a gene involved in lipid transport, was significantly down-regulated in both mouse models after 4 weeks of MCD diet compared to controls (P < 0.05) (Fig. 2E,F). Stearoyl-CoA desaturase 1 (Scd1), a gene involved in lipogenesis, was 500-fold less expressed in C57BL6/J mice and decreased 20-fold in db/db mice after 8 weeks of the MCD diet compared to controls (P < 0.001) (Fig. 2E,F).

Increased Angiogenic Factors in MCD Diet Fed C57BL/6 and db/db Mice.

VEGF and PlGF are important angiogenic factors. VEGF expression was increased both in C57BL/6 and db/db fed an MCD diet. VEGF levels augmented after 3 days of MCD diet in db/db mice and after 1 week in C57BL6/J mice. In both mouse models the VEGF concentration peaked after 4 weeks of the MCD diet (P < 0.001) (Fig. 3A,B).

Figure 3.

Angiogenic factors during the disease progression of NASH in C57BL6/J and db/db mice fed an MCD or control diet. (A) Quantification by enzyme-linked immunosorbent assay (ELISA) of VEGF expression in the liver of C57BL6/J mice fed an MCD or control diet at different timepoints. (B) Quantification by ELISA of VEGF expression in the liver of db/db mice fed an MCD or control diet at different timepoints. (C) Quantification of CD105 staining of the liver of C57BL6/J mice fed an MCD diet or a control diet. (D) Quantification of CD105 staining of the liver of db/db mice fed an MCD diet or a control diet. (E) Gene expression of Vwf at different timepoints of C57BL6/J mice fed MCD or control diet (n = 8/group). (F) Gene expression of Vwf at different timepoints of db/db mice fed MCD or control diet (n = 8/group). P-values are presented as: (a) P < 0.05; (b) P < 0.01; (c) P < 0.001 versus control diet.

Expression of CD105, an endothelial cell marker that is up-regulated by angiogenic factors such as VEGF, was significantly increased after 1 week of the MCD diet in C57BL6/J mice and after 3 days in db/db mice (P < 0.05) (Fig. 3C,D). In control mice, CD105 showed a typical sinusoidal expression pattern. After 8 weeks of MCD diet, CD105 is more expressed in the liver tissue (Supporting Fig. 3A-D). Furthermore, gene expression of the Von Willebrand factor (Vwf) gene, another endothelial cell marker, increased steadily after 1 week in C57BL/6 mice on an MCD diet (Fig. 3E). Vwf gene expression in db/db mice peaks after 2 weeks of MCD diet (P < 0.001) (Fig. 3F).

The vascular structure of C57BL6/J mice was determined by scanning electron microscopy images of vascular corrosion casts of the liver. These casts showed that mice fed an MCD diet have a more disrupted liver vasculature compared to controls (Fig. 6C,F).

General Characteristics of Mice Treated With Antiangiogenic Therapy in a Prevention and Treatment Study.

To address the role of angiogenic factors in the pathophysiology of NASH, a prevention (8 weeks αVEGFR2) and a treatment study (6 weeks αVEGFR2) were set up. In the prevention study, C57BL6/J mice were fed an MCD or control diet and were treated at the same time with VEGFR2 or PlGF antibodies for 8 weeks. In the treatment study, C57BL6/J mice started αVEGFR2 treatment after 2 weeks of MCD diet, when they had already developed steatosis, inflammation, and ballooning.

Mice treated during 8 weeks with αVEGFR2 antibody had significantly higher bodyweights and spleen to body weight ratio compared to untreated MCD fed mice (P < 0.05). Serum levels of ALT, AST, triglyceride (TG), and cholesterol did not differ between untreated, 6 weeks αVEGFR2, 8 weeks αVEGFR2, and αPlGF treated groups (Table 2).

Table 2. Clinical and Biochemical Characteristics of Mice Given MCD or Control Diet for 8 Weeks in a Treatment and Prevention Set-up for αVEGFR2, αPlGF or PBS (n = 10 Mice/Group)
  Treatment StudyPrevention Study
 Control diet6w αVEGFR2 + 8w MCD diet8w αVEGFR2 + 8w MCD diet8w αPlGF + 8w MCD diet8w PBS + 8w MCD diet
  • Data are presented as mean ± SEM. P-values are presented as:

  • *

    P < 0.05 versus control diet;

  • **

    P < 0.05 versus PBS + MCD;

  • ***

    P < 0.01 versus 8w PBS + MCD.

  • d

    Antivascular endothelial growth factor (αVEGFR2), antiplacental growth factor (αPlGF), alanine aminotransferase (ALT), aspartate aminotransferase (AST), body weight (BW), methionine choline-deficient (MCD), phosphate-buffered saline (PBS), triglyceride (TG), total cholesterol (TC).

Food/week (g)7.63 ± 1.238.97 ± 1.54*9.58 ± 2.069.19 ± 1.978.83 ± 1.98
BW (g)24.56 ± 0.4411.16 ± 0.15*11.42 ± 0.19*,**11.20 ± 0.23*10.73 ± 0.18*
Liver/BW (%)4.59 ± 0.075.10 ± 0.19*4.89 ± 0.175.15 ± 0.15*5.28 ± 0.22*
Spleen/BW (%)0.38 ± 0.020.28 ± 0.02*0.28 ± 0.02*,***0.27 ± 0.01*0.21 ± 0.01*
ALT (U/I)28.31 ± 2.21247.26 ± 35.68*204.68 ± 43.38*224.01 ± 42.71*206.53 ± 26.09*
AST (U/I)126.86 ± 5.92361.12 ± 73.61*337.29 ± 53.94*311.38 ± 28.11*292.63 ± 34.66*
TG (mg/dl)74.20 ± 4.5854.50 ± 5.0264.54 ± 3.3654.89 ± 4.4156.27 ± 4.80
TC (mg/dl)73.67 ± 3.1745.37 ± 2.71*56.34 ± 2.29*47.48 ± 3.98*40.68 ± 4.02*

Steatosis, inflammation, ballooning, and fibrosis were assessed histologically using H&E and Sirius Red staining (Fig. 4A). As illustrated in representative sections, the liver of mice fed an MCD diet and treated with αVEGFR2 during 8 weeks had significantly lower grades of steatosis and inflammation compared to the PBS-treated group. Mice treated for 6 weeks with αVEGFR2, in a preventive setting, also showed significantly less steatosis and inflammation compared to untreated mice. This clearly shows that αVEGFR2 prevents the progression to NASH both in a preventive and a therapeutic setting. The liver of mice treated with αPlGF showed no significant changes in liver histology compared to the PBS-treated group (Fig. 4B-F).

Figure 4.

Liver histology of mice treated with antiangiogenic therapy in a prevention and treatment study. (A) Overview of the histological score of mice fed control or MCD diet for 8 weeks and treated 6 weeks with αVEGFR2, 8 weeks with αVEGFR2, 8 weeks with αPlGF, or 8 weeks with PBS. Anti-VEGFR2 treatment prevents the progression of NASH by attenuating steatosis and inflammation, both in a preventive and therapeutic setting. (B-F) H&E-stained liver slides of mice fed a control diet or MCD diet for 8 weeks and treated 6 weeks with αVEGFR2, 8 weeks with αVEGFR2, 8 weeks with αPlGF, or 8 weeks with PBS (200×). Scale bars = 100 μm.

Effect of Antiangiogenic Treatment on Inflammation and Lipid Metabolism.

The presence of inflammatory infiltrates in the liver was examined with F4/80 staining. The staining showed that Kupffer cells were more isolated and formed fewer clusters in mice treated with 6 or 8 weeks of αVEGFR2 compared to MCD-fed mice treated with PBS (Supporting Fig. 2A,B,E-G). Mice treated with αVEGFR2 for 8 weeks had significantly less F4/80 staining compared to untreated mice (Fig. 5A). Gene expression in the liver of Tnf and Il1b gene confirmed that αVEGR2 treatment reduced inflammation in the liver of MCD-fed mice for 8 weeks compared to mice treated with PBS (Fig. 5B).

Figure 5.

Effect of antiangiogenic treatment on inflammation and lipid metabolism in mice fed MCD or control diet. (A) F4/80 staining in the liver of mice fed control or MCD diet for 8 weeks and treated 6 weeks with αVEGFR2, 8 weeks with αVEGFR2, 8 weeks with αPlGF, or 8 weeks with PBS. (B) Gene expression of inflammatory genes (Tnfα and Il1b) in mice fed control or MCD diet for 8 weeks and treated 6 weeks with αVEGFR2, 8 weeks with αVEGFR2, 8 weeks with αPlGF, or 8 weeks with PBS (n = 10/group). These data confirm the data seen at the protein level. (C) Gene expression of scd1 and L-fabp1 of mice fed control or MCD diet for 8 weeks and treated 6 weeks with αVEGFR2, 8 weeks with αVEGFR2, 8 weeks with αPlGF, or 8 weeks with PBS (n = 10/group). P-values are presented as: (a) P < 0.05 versus control diet; (b) P < 0.05 versus 8-week PBS + MCD; (c) P < 0.01 versus 8-week PBS + MCD; (d) P < 0.001 versus 8-week PBS + MCD. (D) In vitro measurement of lipid accumulation with AdipoRed assay. Fat-laden primary hepatocytes treated with 100 μg/mL IgG, αVEGFR2, VEGF, or VEGF+αVEGFR2. (E) Dose-response curve for the effect of lipid accumulation in primary hepatocytes treated with different concentrations of αVEGFR2. P-values are presented as: (*) P < 0.05; (**) P < 0.01; (***) P < 0.001 versus IgG-treated cells.

Scd1 gene expression was significantly increased in mice treated with αVEGFR2 for 6 and 8 weeks compared to mice treated with PBS (P < 0.001) (Fig. 5C). Expression of L-fabp1, a gene involved in lipid transport, was not affected by any treatment (Fig. 5C). Lipid regulation in vitro was assessed by AdipoRed assay. Dose-response curves showed that a concentration of 100 μg αVEGFR2/mL was optimal and showed that αVEGFR2 therapy significantly decreased lipid accumulation in fat-laden primary hepatocytes (Fig. 5D,E).

Effect of Antiangiogenic Treatment on Vasculature.

Expression of CD105 was significantly decreased in C57BL/6 mice on 8 weeks of an MCD diet and treated with 6 or 8 weeks of αVEGFR2 compared to MCD-fed mice treated with PBS (Fig. 6A) (Supporting Fig. 3A,B,E-G). Only the group treated for 8 weeks with αVEGFR2 showed a reduced expression of the Vwf gene compared to untreated MCD fed mice (P < 0.001) (Fig. 6B). CD105 and Vwf gene expression of mice on 8 weeks of an MCD diet compared to mice on a control diet were increased in the αPlGF and PBS-treated group, confirming our previous data (Figs. 3C,E; 6A,B).

Figure 6.

Effect of antiangiogenic treatment on angiogenic factors in mice on MCD or control diet. (A) Immunohistochemical analysis of CD105 expression in the liver of mice fed a control or MCD diet for 8 weeks and treated 6 weeks with αVEGFR2, 8 weeks with αVEGFR2, 8 weeks with αPlGF, or 8 weeks with PBS. Inhibition of VEGFR2 for 8 weeks significantly normalized CD105 levels to the same level as in control mice. (B) Gene expression of Vwf and αSma of mice fed control or MCD diet for 8 weeks and treated 6 weeks with αVEGFR2, 8 weeks with αVEGFR2, 8 weeks with αPlGF, or 8 weeks with PBS (n = 10/group). (C-F) Scanning electron microscopy images of vascular corrosion casts of mice fed 8 weeks a control diet or MCD diet and treated 8 weeks with αVEGFR2, αPlGF or PBS. Typically, NASH is characterized by disorganized vessels, which are tortuous and show no hierarchical organization (F). Inhibition of VEGFR2 caused a partial normalization of the vasculature (D), αPlGF treatment did not induce this effect (E). P-values are presented as: (a) P < 0.05 versus control diet; (b) P < 0.05 versus 8-week PBS + MCD; (c) P < 0.01 versus 8-week PBS + MCD; (d) P < 0.001 versus 8-week PBS + MCD.

Hepatic stellate cell (HSC) activation was evaluated with alpha-smooth muscle actin (αSma) gene expression in the liver and visualized with an αSMA staining. Gene expression of αSMA was significantly up-regulated in mice fed the MCD diet. Anti-VEGFR2 treatment for 6 or 8 weeks significantly reduced αSMA expression (Fig. 6B). Immunohistochemical staining showed only very mild activation of HSC in mice fed an MCD diet for 8 weeks. Nonetheless, mice treated with antiangiogenic antibodies showed less HSC activation compared to untreated mice (Supporting Fig. 4A-E).

Images of vascular corrosion casts showed that mice treated with αVEGFR2 had a less disrupted vascular architecture than untreated or αPlGF-treated mice. Although the image still shows some disrupted vessels, the vascular structure is more conserved compared to untreated mice (Fig. 6C-F).

Discussion

The present study highlights the importance of the angiogenic factor, VEGF, in the pathophysiology of NASH. In current literature, the role of angiogenesis in the disease progression of NASH in both human and experimental settings is gaining more and more attention. VEGF and PlGF are reported to be one of the main factors involved in normal and pathologic angiogenesis in several chronic liver diseases.7 VEGF is a potent angiogenic growth factor that stimulates endothelial cell proliferation and induces microvessel permeability.24 However, the underlying mechanisms that initiate angiogenesis in NASH remain unclear. A number of provoking stimuli with potential relevance in NASH, including inflammation and hypoxia, have been proposed in other pathologic circumstances as initiators of angiogenesis.25 Nonetheless, detailed studies specifically addressing these molecular signals in NASH are not available. The aim of our study was to increase insight on the role of angiogenic factors in the progression from steatosis to NASH. Our study showed that VEGF increased during the transition from steatosis to NASH, peaking after 4 weeks of an MCD diet in two mouse models for NASH. Moreover, αVEGFR2 treatment prevents the progression of NASH, by attenuating steatosis and inflammation, both in a preventive and therapeutic setting.

In the first part of our study, we determined the role of angiogenic factors at different timepoints in the disease progression of NASH. The experiments were conducted in two frequently used mouse models for NASH. First, C57BL6/J mice were given an MCD diet to induce NASH. The main advantage of the MCD diet is that histological changes occur rapidly and are morphologically similar to those observed in human NASH.16 The MCD diet induced an increase in ALT and AST serum levels, prominent steatosis and ballooning of the hepatocytes, and infiltration of inflammatory foci in the liver. However, MCD-fed mice, contrary to NASH patients, lose a significant amount of body weight during the experiment and do not develop insulin resistance. To attenuate the inconsistencies between the MCD model and human NASH, a genetic db/db mouse model was used. These mice preserve peripheral insulin resistance and hepatic injury is accentuated compared to C57BL6/J fed an MCD diet. Despite the fact that the MCD dietary model has known disparities with human NASH, it is useful in exploring mechanisms of injury in NASH.16

Angiogenesis is initiated in the pathophysiology of NASH due to physiological and mechanical triggers. First, we will discuss the physiological part in the development of NASH. Lipotoxicity activates cytokines, which will subsequently induce recruitment of inflammatory cells and platelets. Inflammation triggers vascular permeability by recruiting monocytes, macrophages, platelets, mast cells, and other leukocytes. These cells can initiate angiogenesis through different pathways. In this manner, inflammation can contribute to the formation of new vascular structures in the liver.26 Second, angiogenesis could also be triggered mechanically as fat accumulation damages the hepatocytes leading to deregulation of the microvascular blood flow. Reduction in sinusoidal perfusion initially arises from the effects of hepatocytes loaded with accumulated lipids. This results in reduction of the intrasinusoidal volume, as well as altering the sinusoidal architecture.27 Vascular corrosion casting has recently been revived and has proven to be an excellent tool for detailed 3D morphological examination of normal and pathological microcirculation.28 Another factor that should be taken into account is the activation of HSC, as it has been associated with fibrosis and angiogenesis.29 It has been shown that angiogenic factors are up-regulated in various chronic liver diseases with endstage liver fibrosis.10 However, in our study HSC activation is probably not the main trigger for angiogenesis as immunohistochemical staining for aSMA only showed mild activation of HSC. This can be explained due to the fact that our mice only developed a stage 1 fibrosis after 8 weeks of the MCD diet. Furthermore, we found that levels of VEGF and CD105 are already significantly increased after 3 days and 1 week of MCD diet in db/db and C57BL6/J mice, respectively.24 At that time liver histology showed increased steatosis, ballooning, and inflammation. However, mice could not yet be classified as NASH. These results suggest that the molecular events associated with an up-regulation of angiogenic factors start very early in the pathophysiology of NASH.

Our study confirmed that during the pathophysiology of NASH there is both a physiological and a mechanical trigger that induces angiogenesis. As such, we found a significant increase of inflammatory and angiogenic factors in two mouse models early in the development of NASH. Electron microscopic images of the vascular corrosion casts of the liver of mice with NASH clearly show that the morphology of the vasculature is disrupted compared to controls. However, it is not clear what the primary trigger for angiogenesis in the pathophysiology of NASH is. Probably this phenomenon should be addressed as a multifactorial process in which the combination of forces affecting the mechanotransduction in the endothelium and physiological changes in the lipogenic and inflammatory metabolism causes the initiation of angiogenesis. Angiogenesis might improve oxygen and nutrient delivery to the hepatocytes or improve the removal of substances processed by hepatocytes, leading to the formation of new functional sinusoids.30

In the second part of our study we showed that αVEGFR2 treatment prevents the progression of NASH by attenuating steatosis and inflammation, both in a preventive and therapeutic setting, thereby confirming our hypothesis that angiogenic factors play an early role in the disease progression from steatosis to NASH. The grade of steatosis and inflammation was significantly less in the liver of αVEGFR2-treated mice in a preventive setting. Moreover, αVEGFR2 treatment was able to block the progression to NASH in mice with steatosis and inflammation. This could indicate that αVEGFR2 treatment could temper the effect of angiogenic stimuli. Moreover, in vitro we found that αVEGFR2 therapy significantly decreased lipid accumulation in fat-laden primary hepatocytes. This is in line with previous studies that the VEGF/VEGFR2 pathway is critical for both angiogenesis and adipogenesis during de novo adipose tissue formation from preadipocytes.31 Moreover, previous studies showed that angiogenic, inflammatory, and lipogenic processes are tightly crosslinked and are capable of sustaining each other.32 The reason why αPlGF treatment did not have a significant effect on the liver histology of mice with NASH compared to untreated mice with NASH could probably lie in the fact that PlGF signals only through the VEGFR1 pathway and not by way of VEGFR2.

To obtain a clear insight into the molecular events, we examined the transcript levels of lipogenic genes (Scd1 and L-fabp1). These experiments showed that αVEGFR2 treatment increased Scd1 gene expression in both a preventive and therapeutic setting. Scd1 plays a key role in the prevention of steatohepatitis by partitioning excess lipid into monounsaturated fatty acids that can be safely stored.33 The literature confirms our results that MCD feeding causes a significant decrease in hepatic Scd1 gene expression as well as the down-regulation of L-fabp1 gene expression compared to mice fed a control diet.34, 35 Our study showed that αVEGFR2 increases Scd1 gene expression to a more normal Scd1 expression in the liver of mice treated in a therapeutic and preventive setting compared to untreated mice fed an MCD diet. This could suggest that αVEGFR2 therapy improves lipid metabolism in the liver. However, the exact molecular mechanism responsible for MCD diet-induced down-regulation of Scd1 as well as changes in Scd1 expression in human NAFLD, and its relation to liver damage and disease progression, remain unknown and will require further investigation.

In summary, we demonstrated the role for VEGF in the pathophysiology in two mouse models for NASH. This study shows for the first time that αVEGFR2 treatment attenuates steatosis and inflammation in the liver of mice with NASH both in a preventive and therapeutic setting. This implies that angiogenic factors should be considered an important factor in the pathophysiology of NASH. NASH is a multifactorial process in which a number of diverse parallel processes might contribute to the development of liver inflammation and angiogenesis. In several stages of NASH a link between disease progression, inflammation, and hepatic microvascular changes can be made. The close relationship between angiogenesis and the progression of NASH could offer multiple clinical applications. Antiangiogenic therapies might be used to manage disease progression in NASH.

Acknowledgements

The authors thank the Ghent University Hospital, Department of Gastroenterology and Hepatology.

Ancillary