Main

Non-alcoholic fatty liver disease (NAFLD), which is associated with obesity, type 2 diabetes and dyslipidemia, has become a worldwide health problem during the last decades.1, 2, 3, 4 NAFLD encompasses a broad spectrum of pathological conditions ranging from simple steatosis to steatohepatitis, fibrosis and finally to cirrhosis.5, 6, 7, 8, 9 Molecular mechanisms involved in the pathogenesis of NAFLD still remain unclear and therapies are limited. A better understanding of the molecular mechanisms contributing to the development of the early stages of NAFLD is desirable to develop prevention and also new therapeutic strategies.

Throughout the last three decades, dietary fructose intake has markedly increased in the United States and in Europe, paralleling the prevalence of obesity and type 2 diabetes.10, 11, 12 Results of recent studies further suggest that a diet rich in fructose may also be a risk factor for the development of NAFLD in humans.13, 14, 15, 16 The hypothesis that fructose intake might be causally involved on the onset and progression of NAFLD is also bolstered by the results of several animal studies. For instance, it has been shown that in rodents a diet rich in fructose can cause steatosis and oxidative stress in the liver, as well as insulin resistance.17, 18, 19, 20 Results of our own group have suggested that the onset of hepatic steatosis caused by chronic moderate fructose intake (eg, 30% fructose solution as only fluid source) is associated with an increased translocation of endotoxin from the intestine into the portal vein, formation of reactive oxygen species (ROS) in the liver and an induction of hepatic tumor necrosis factor-α (TNF-α) expression;19, 21 however, despite the fact that markers of insulin resistance such as plasma RBP-4 levels were found to be elevated in this model,21 blood glucose levels did not differ between fructose-fed mice and controls.19 An increased expression of TNF-α has repeatedly been reported to be associated with the development of NAFLD in different animal models.22, 23, 24 Indeed, in line with the findings of other groups, we recently showed that TNF-α receptor 1−/− (TNFR1−/−) mice were protected from the onset of fructose-induced steatosis and also insulin resistance.25 However, mechanisms involved in the deleterious effects of TNF-α on the liver have not yet been fully understood.

Plasminogen activator inhibitor-1 (PAI-1) is an acute-phase protein, which lately has also been linked to the onset and to the progression of liver diseases of various etiologies (eg, alcohol and endotoxin; for overview, see ref. 26). For instance, we and others reported that patients with NAFLD have elevated hepatic and plasma levels of PAI-1 in comparison with controls.13, 27 Furthermore, in animal studies, it was shown that PAI-1 expression in the liver and plasma is increased in genetically obese mice.28 Ma et al29 showed that PAI-1−/− mice are protected not only from weight gain but also from marked hepatic lipid accumulation in a mouse model of high-fat/high-carbohydrate feeding. It has further been shown in animal models of acute alcohol-induced liver steatosis and in in vitro experiments that PAI-1 is involved in mediating the pivotal effect of TNF-α on the liver.30 However, whether PAI-1 is also causally involved in the onset of fructose-induced hepatic steatosis in mice has not been clarified. Starting from this background, the aim of the present study was to assess in a mouse model whether PAI-1 is also a critical factor in the onset of fructose-induced hepatic steatosis and if so, what the underlying mechanisms involved are.

MATERIALS AND METHODS

Animals and Treatments

Mice were housed in a pathogen-free barrier facility accredited by the Association for Assessment and Accreditation of Laboratory Animal Care. Procedures were approved by the local Institutional Animal Care and Use Committee. The 6-week-old C57BL/6J and PAI-1−/− mice (Serpine1tm1Mlg ID: 1857287, Jackson, Bar Harbor, USA) were fed with plain water or water containing 30% fructose ad libitum for a period of 8 weeks. Water intake was assessed weekly twice, whereas food intake and body weight gain were assessed once a week. Animals were anesthetized with 80 mg ketamine and 6 mg xylazine/kg body weight intraperitoneally, and blood was drawn from the portal vein. Portions of liver tissue were frozen immediately in liquid nitrogen, whereas others were frozen fixed in OCT-mounting media (medite, Burgdorf, Germany) for later sectioning and mounting on microscope slides.

Cell Culture and Treatment

Alpha mouse liver 12 (AML-12) cells obtained from American type culture collection (ATCC, Manassas, USA) were grown in DMEM/F12 media (PAN-Biotech, Aidenbach, Germany) supplemented with 10% fetal bovine serum, 40 ng/ml dexamethasone, 0.005 mg/ml insulin, 5 ng/ml selenium and 0.005 mg/ml transferrin (PAA Laboratories, Colbe, Germany), as well as penicillin and streptomycin at 37°C. At 70% confluence, cells were serum starved for 18 h in starvation media supplemented with bovine serum albumin (BSA; 0.01%), insulin, selenium and transferring, as well as penicillin and streptomycin. Cells were then challenged with 10 ng/ml PAI-1 (Sigma, Steinheim, Germany) for 24 h. In addition, some cells were concomitantly treated with 20 ng/ml hepatocyte growth factor (HGF; Sigma).

Cells were rinsed twice with phosphate-buffered saline (PBS) and either used for triglyceride isolation or measurements of microsomal triglyceride transfer protein (MTTP) activity as detailed below.

Immunoblots

Liver tissue was homogenized in a lysis buffer (20 mmol/l MOPS, 150 mmol/l NaCl, 1 mmol/l EDTA, 1% Nonidet P-40, 1% sodium deoxycholate and 0.1% SDS) containing a protease and phosphatase inhibitors mix (Sigma) to prepare total protein lysates. Cytosolic protein lysates were prepared by homogenizing liver tissue in a lysis buffer (1 mol/l HEPES, 1 mol/l MgCl2, 2 mol/l KCl and 1 mol/L DTT) containing a protease and phosphatase inhibitors mix (Sigma). Proteins (30 or 60 μg protein/well) were separated in 8–15% SDS-polyacrylamide gels depending on the respective molecular weight of the target protein. Proteins were transferred onto Hybond-P polyvinylidene difluoride membranes (Amersham Biosciences, Freiburg, Germany) using a semidry electroblotter. The resulting blots were then probed with antibodies against phospho-c-Met, total c-Met, TNF-α, β-actin, phospho Akt or total Akt (all obtained from Cell Signaling Technology, Danvers, USA). Bands were visualized using the Super Signal Western Dura kit (Thermoscientific, Rockford, USA). To ensure equal loading, all blots were stained with Ponceau red. Protein bands were densitometrically analyzed using the Flurochem Software (Alpha Innotech, San Leandro, USA).

Hepatic Triglyceride Determination and Lipid Staining

Mouse liver tissue was homogenized in 2 × PBS. Tissue lipids were extracted with methanol:chloroform (1:2), dried in an evaporating centrifuge and resuspended in 5% fat-free BSA. Colorimetric assessment of triglyceride levels was carried out using a commercially available kit (Randox, Krefeld, Germany). Values were normalized to protein concentration in liver homogenate using the Bradford assay (Bio-Rad Laboratories, Munich, Germany). Triglyceride levels in cell cultures were determined by using a commercially available kit (Bio Vision, California, USA) following the instructions of the manufacturer. In addition, to determine hepatic lipid accumulation, frozen sections of liver (10 μm) were stained with Oil Red O (Sigma) for 15 min, washed and counterstained with hematoxylin for 45 s (Sigma).

RNA Isolation and Real-Time RT-PCR

Total RNA was extracted from liver tissue samples using Trifast (PEQLAB, Erlangen, Germany). RNA concentrations were determined spectrophotometrically and 1 μg of total RNA was subjected to DNase digestion, followed by a reverse transcription using a Moloney murine leukemia virus reverse transcriptase and oligo (dT) primers (Fermentas, St Leon-Rot, Germany). PCR primers were designed using the software Primer 3 (Whitehead Institute for Biomedical Research, Massachusetts, USA) and are summarized in Table 1. Sybr Green Universal PCR Master Mix (Applied Biosystems, Damstadt, Germany) was used to prepare the PCR mix. Primers were added to a final concentration of 300–400 nmol/l.

Table 1 Primer sequences used for real time PCR and CT values
Full size table

The amplification reactions were carried out in an iCycler (Bio-Rad) with initial hold steps (95°C for 2 min) and 50 cycles of a three-step PCR (95°C for 15 s, 60°C for 15 s and 72°C for 30 s). The fluorescence intensity of each sample was measured at each temperature change to monitor amplification of the target gene. The comparative CT method was used to determine the amount of target gene, normalized to an endogenous reference (18S) and relative to a calibrator (2---ΔΔCt).

Clinical Chemistry and Pathologic Evaluation

Plasma alanine aminotransferase (ALT) activity was determined using a commercially available kit (Randox, Krefeld, Germany). Paraffin sections of liver (5 μm) were stained with hematoxylin and eosin to assess liver histology.

Immunohistochemical staining for cluster of differentiation 3-ζ

To determine cluster of differentiation 3-ζ (CD3-ζ), paraffin-embedded liver sections (5 μm) were deparaffinized and rehydrated in ethanol solutions with decreasing concentrations. Liver tissue was stained for CD3-ζ by blocking slides with 8% BSA for 45 min, followed by washing with 2% BSA and an overnight incubation with a polyclonal primary antibody solution (Santa Cruz Biotechnology, Heidelberg, Germany) in a humidified chamber at 4°C. Liver sections were washed, and CD3-ζ was detected by incubating slides with a peroxidase-linked secondary antibody and diaminobenzidine (Peroxidase Envision Kit; DAKO, Hamburg, Germany). A negative control was carried along where PBS tween was added instead of CD3-ζ antibody.

Endotoxin Assay

Endotoxin was determined as described previously in detail.13 Briefly, plasma samples were incubated at 70°C for 20 min. Plasma endotoxin levels were determined using a commercially available endpoint Limulus Amebocyte Lysate assay (Charles River, L’Arbaesle, France) following the instructions of the manufacturer.

MTTP Assay in Liver Tissue and AML-12 Cells

Liver tissue was homogenized in a lysis buffer (1 mmol/l Tris, pH 7.6, 1 mmol/l EGTA, 1 mmol/l MgCl2 and protease inhibitors) and centrifuged at 7400 g for 30 min at 4°C. Protein concentration was determined using the Bradford assay, and 100 μg of protein was used to determine the MTTP activity using a commercially available kit according to the manufacturer's instructions (Chylos, New York, USA). To determine MTTP activity in AML-12 cells, cells were washed with ice-cold PBS, homogenized in a lysis buffer (1 mmol/l Tris, pH 7.6, 1 mmol/l EGTA, 1 mmol/l MgCl2 and protease inhibitors) and then centrifuged at full speed for 1 h at 4°C The supernatant was used to determine the MTTP activity using a commercially available kit according to the manufacturer's instructions (Chylos).

Statistical Analysis

Data are expressed as means+s.e.m. Student's t-test or one-way ANOVA with Tukey's post hoc test were used to determine the statistical differences between the treatment groups as appropriate. A P-value <0.05 was considered to be significant.

RESULTS

PAI-1 and Fructose-Induced Steatosis

As shown before by our own group,19, 21 chronic intake of fructose led to a significant increase of triglycerides in the livers of wild-type mice (plus ∼5.8-fold in comparison with controls; Figure 1c). This fructose-induced steatosis was associated with a significant ∼2.5-fold increase in PAI-1 mRNA expression in livers of fructose-fed mice in comparison with water controls (see Table 2). To test the hypothesis that PAI-1 is causally involved on the onset of fructose-induced steatosis mice, PAI-1−/− mice were fed with 30% fructose solution or water for 8 weeks. Lipid staining, pathological changes and triglyceride content did not differ between wild-type and PAI-1−/− mice fed with water (Figure 1). Contrary to the findings in fructose-fed wild-type mice, hepatic triglyceride accumulation was only increased by approximately twofold in livers of PAI-1−/− mice exposed to fructose in comparison with PAI-1−/− mice fed with water (Figure 1c). Similar results were also found when staining liver sections with hematoxylin and eosin and oil red O, respectively (Figure 1a and b). In line with these findings the liver to body weight ratio and ALT levels in plasma were also only found to be increased in wild-type animals fed fructose (Table 3). Weight gain was also significantly higher in fructose-fed mice than in water fed or only controls; however, weight gain did not differ between fructose-fed groups (Table 3). Despite the protective effects of the loss of PAI-1 on the onset of fructose-induced hepatic steatosis, both portal endotoxin and hepatic TNF-α protein concentration was significantly increased in fructose-fed mice regardless of strain in comparison with water controls (Figure 2).

Figure 1
figure 1

Effect of PAI-1 deletion on hepatic lipid accumulation and histology in fructose-fed mice. (a) Representative photomicrographs of hematoxylin and eosin (H & E) staining of liver sections. (b) Representative photomicrographs of oil red O staining of liver sections. (c) Quantitation of hepatic triglycerides content (n=5–7). Data are expressed as means±s.e.m. aP<0.05 compared with wild-type mice fed with plain water. bP<0.05 compared with PAI-1−/− mice fed with plain water. cP<0.05 compared with wild-type (WT) fructose-fed mice.

Full size image
Table 2 Effect of fructose feeding on the expression of PAI-1, CD1d, IL-15 and IFN-γ in livera
Full size table
Table 3 Effect of fructose feeding on body weight and liver, as well as BCL-XL mRNA expressiona
Full size table
Figure 2
figure 2

Effect of PAI-1 deletion on portal endotoxin and hepatic TNF-α protein levels in fructose-fed mice. (a) Endotoxin levels (n=5–6). (b) Representative photographs of western blots of TNF-α and β-actin (c) quantitative analysis of blots (n=5). Data are expressed as means±s.e.m. aP<0.05 compared with wild-type (WT) mice fed with plain water. bP<0.05 compared with PAI-1−/− mice fed with plain water.

Full size image

Enhanced c-Met Signaling, an Increased Lipoprotein Synthesis and Phosphorylation of Akt and mRNA Expression of B-cell Lymphoma Extra Large are Associated with the Protective Effect of PAI-1 Deletion

As it has been shown before that PAI-1 has a crucial role in the activation of the hepatocyte growth factor-c-Met-depending modulation of hepatic lipid metabolism in hepatic steatosis of various etiologies (eg, alcohol and endotoxin; for overview, see ref. 26), we determined the effect of fructose feeding on c-Met phosphorylation, expression of apolipoprotein B (ApoB) mRNA and activity of MTTP in livers of PAI-1−/− and wild-type mice. No differences in phosphorylation of c-Met were found between livers of water controls regardless of strain and fructose-fed wild-type mice. In contrast, in PAI-1−/− mice fed chronically with 30% fructose solution, phosphorylation was significantly increased in comparison with wild-type controls fed with water (plus approximately 9-fold in comparison with wild-type and plus approximately 4-fold in comparison with PAI-1−/− mice fed with water; Figure 3a). In line with these findings, expression of ApoB mRNA was only found to be increased in fructose-fed PAI-1−/− mice (plus ∼3.1-fold in comparison with wild-type and plus ∼2.3-fold in comparison with PAI-1−/− mice fed with water), whereas those of fructose-fed wild-type mice remained at the level of controls (Figure 3b). Furthermore, MTTP activity was also found to be increased only in PAI-1−/− mice fed with fructose (30% in comparison with wild-type mice (P<0.05) and 10% in comparison with PAI-1−/− mice fed with water (P is not significant; Figure 3c)).

Figure 3
figure 3

Effect of PAI-1 deletion on c-Met, ApoB mRNA expression and MTTP activity, as well as Akt protein levels in livers of fructose-fed mice. (a) Representative photographs of western blots of phospho-c-Met, total c-Met and quantitative analysis of blots (n=4) (b) ApoB mRNA levels normalized to β-actin expression (n=5) (c) MTTP activity (n=5) and (d) representative photographs of western blots of phospho Akt and total Akt, and quantitative analysis of blots (n=5). Data are expressed as means±s.e.m. aP<0.05 compared with wild-type (WT) mice fed with plain water. bP<0.05 compared with PAI-1−/− mice fed with plain water. cP<0.05 compared with wild-type fructose-fed mice.

Full size image

As it has been shown before by others that c-Met might also be involved in the regulation of Akt,31, 32 we determined Akt phosphorylation in livers of fructose and water-fed mice. Despite elevated TNF-α protein levels in both fructose-fed groups, phosphorylation levels of Akt were only found to be decreased in livers of fructose-fed wild-type mice (Figure 3d). In livers of fructose-fed PAI-1−/− mice, phosphorylation of Akt did not differ from that of water-fed controls (Figure 3d). In line with these findings, mRNA expression of B-cell lymphoma extra large (BCL-XL), which was shown to be regulated through HGF-c-Met-Akt-dependent pathways,33 was found to be significantly reduced in livers of wild-type mice fed with fructose. In contrast, in livers of PAI-1−/− mice fed with fructose, expression of BCL-XL was significantly higher than in livers of PAI-1−/− mice fed with water (Table 3).

Effect of PAI-1 and HGF on Fat Accumulation and MTTP Activity in AML-12 Cells

To further investigate the role of HGF and PAI-1 in modulating triglyceride export in the liver AML-12 cells, a model of murine hepatocytes was challenged with PAI-1 in the presence or absence of HGF. The treatment of AML-12 cells with HGF had no effect on triglyceride content and MTTP activity (Figure 4a and b). In contrast, when the cells were challenged with PAI-1, MTTP activity was significantly lower in comparison with naïve cells (∼81%). In line with these findings, triglyceride concentration was significantly increased in these cells (approximately +10-fold in comparison with naïve cells) too. Interestingly, when cells were treated with HGF while concomitantly being challenged with PAI-1, MTTP activity remained at the level of naïve cells and cells did not accumulate triglycerides (Figure 4a and b).

Figure 4
figure 4

Effect of PAI-1 on triglyceride accumulation and MTTP activity in AML-12 cells. (a) Triglyceride accumulation and (b) MTTP activity. Data are expressed as means±s.e.m. (n=3). aP<0.05 compared with naïve cells. bP<0.05 compared with HGF-treated cells. cP<0.05 compared with PAI-1-treated cells.

Full size image

Effect of PAI-1 Deletion on Hepatic Natural Killer T Cells

As it has been suggested by the results of other groups that MTTP regulates endogenous and exogenous antigen presentation by group 1 cluster of differentiation 1d (CD1d) molecules34 and also its function (eg, binding and activation of iNKT cells),35 we determined expression of CD1d, markers of iNKT cells, as well as iNKT cell activation in livers of PAI-1−/− and wild-type mice fed with water or fructose. Expression of CD1d was markedly higher in livers of PAI-1−/− mice in comparison with wild-type mice regardless of additional treatment; however, as expression varied considerably between animals in the water control group, differences did only reach a level of significance for fructose-fed PAI-1−/− mice (P<0.05 in comparison with wild-type controls and fructose-fed wild types; Table 2). In line with these findings, expressions of the iNKT cell markers CD3-ɛ and natural killer 1.1 (NK1.1) were also found to be significantly higher in livers of PAI-1−/− mice in comparison with wild-type mice regardless of additional treatments (Figure 5b and c). Similar results were also found for CD3-ζ staining in liver sections of PAI-1−/− mice regardless of treatment (representative pictures of staining are shown in Figure 5a). Interestingly, markers of iNKT cell activation (eg, interleukin 15 (IL-15) and interferon-γ (IFNγ)) were both only found to be significantly induced in livers of PAI-1−/− mice fed with fructose, whereas expression levels of IL-15 and IFN-γ did not differ between water controls regardless of the strain and wild-type mice fed with fructose (Table 2).

Figure 5
figure 5

Effect of PAI-1 deletion on NKT cell markers in livers of fructose-fed mice. (a) Representative photographs of CD3-ζ staining of liver sections; brown stained parts marked with arrows indicate CD3-ζ-positive cells. (b) CD3-ɛ mRNA levels normalized to 18S expression (n=5). (c) NK1.1 mRNA levels normalized to 18S expression (n=5). Data are expressed as means±s.e.m. aP<0.05 compared with wild-type (WT) mice fed with plain water. bP<0.05 compared with PAI-1−/− mice fed with plain water. cP<0.05 compared with wild-type fructose-fed mice.

Full size image

DISCUSSION

Results of several human and animal studies suggest that an increased dietary fructose intake is a risk factor in the development of NAFLD.13, 14, 16, 36 In line with these findings, results of pilot human studies suggest that a reduction of fructose intake may exert beneficial effects on the progression of NAFLD.14 We recently reported that the onset of fructose-induced hepatic steatosis is associated with markedly increased endotoxin levels in portal plasma, an increased formation of ROS and an induction of nuclear factor-κB activity as well as TNF-α mRNA expression in the liver of mice.19 In humans with NAFLD and also in livers of obese mice (eg, ob/ob) with NAFLD, PAI-1 expression was reported to be increased, too.13, 28 We further reported that in TNFR1−/− mice, the onset of fructose-induced liver steatosis is markedly attenuated and the loss of TNFR1 in these mice was associated with normalization of PAI-1 expression in the liver.25 However, whether PAI-1 is causally involved in mediating the effects of fructose on the liver has not yet been clarified. In the present study, the hypothesis that PAI-1 is critically involved in the onset of fructose-induced NAFLD was tested in a mouse model. Indeed, in livers of fructose-fed mice, PAI-1 expression was induced. Furthermore, despite similarly elevated endotoxin levels in portal plasma and TNF-α protein concentrations in the liver, fat accumulation in livers of PAI-1−/− mice resulting from chronic exposure to 30% fructose solution was markedly attenuated (minus ∼50%). Interestingly, the magnitude of protection against fructose-induced hepatic steatosis found in livers of PAI-1−/− mice fed with fructose was similar to that reported before by our group for livers of fructose-fed TLR-4-mutant mice and mice treated with non-resorbable antibiotics, as well as TNFR1−/− mice.19, 21, 25 Taken together, these data further suggest that PAI-1 has a casual role on the onset of NAFLD in the liver. These data also suggest that PAI-1 may be a critical factor in mediating the effect of the increased translocation of bacterial endotoxin and the subsequent activation of TLR-4-dependent signaling cascades found in livers of mice chronically exposed to fructose.

Results of studies using animal models of acute and chronic alcohol-induced liver damage have suggested that PAI-1 may modulate hepatic lipid export.30 Indeed, PAI-1 is an inhibitor of urokinase-type plasminogen activator (uPA), which has been shown before to be involved not only in fibrinolysis but also to activate prohepatocyte growth factor (pro-HGF) to HGF.37, 38 Results of in vitro studies have suggested that HGF can stimulate the MTTP-dependent lipoprotein secretion in hepatocytes through binding to its receptor c-Met.39 In support of these in vitro results, it has been shown before in livers of mice after acute alcohol treatment that c-Met phosphorylation and MTTP activity, as well as ApoB concentration were upregulated in PAI-1−/− and TNFR1−/− mice.30 Furthermore, it has been shown that blocking lipoprotein secretion in the liver may increase the susceptibility of the liver to certain toxin challenges (eg, endotoxin).40 In the present study, phosphorylation of c-Met as well as mRNA expression of ApoB and MTTP activity were markedly upregulated in livers of PAI-1−/− mice fed with fructose, whereas in wild-type mice fed with fructose, phosphorylation of c-Met, ApoB mRNA expression and MTTP activity remained at the level of controls. Furthermore, using a murine cell culture model of hepatocytes (eg, AML-12 cells), we were able to show that the challenging of hepatocytes with PAI-1 may result in a reduction of MTTP activity and accumulation of triglyceride in cells, whereas this effect of PAI-1 was totally abolished in the presence of HGF. Taken together, these data suggest that in the present study, the protective effect of the loss of PAI-1 may have resulted from a compensatory increase in very-low-density lipoprotein (VLDL) synthesized to excrete the greater amount of triglyceride synthesis in hepatocytes resulting from the metabolism of fructose. These results further suggest that PAI-1 through an inhibition of uPA may modulate the activation of pro-HGF to HGF. This lack of active HGF may in turn lead to a suppression of the MTTP activity, resulting in an accumulation of triglycerides that are no longer shuttled out of the cells via ApoB/VLDL.

It has been reported before by others that c-Met may be involved in the regulation of phosphoinositide 3-kinase (PI3K)/Akt signaling cascade.31, 32 In line with these findings, phosphorylation of Akt was only found to be altered in fructose-fed wild-type mice. Indeed, expression of antiapoptotic BCL-2 family protein BCL-XL, which was shown33 to be modulated by HGF-c-Met-Akt-dependent pathways, was induced in livers of fructose-fed PAI-1−/− mice, whereas a similar effect was not found in livers of fructose-fed wild-type mice. These findings are also supported by the results of in vitro cell culture models of other groups in which it was shown that PAI-1 may act as a regulator of cell survival through altering the PI3K/Akt pathway.41 Taken together, these results suggest that the loss of PAI-1 probably through c-Met-dependent signaling pathway may protect the liver from impaired PI3K/Akt signaling associated with the chronic intake of fructose and may even influence apoptosis found under these conditions. Exact mechanisms involved in this rescue of PI3K/ Akt signaling and their implications on apoptosis, as well as the liver and the development of NAFLD remain to be determined.

It has been suggested by the results of Dougan et al35 and Brozovic et al42 that the MTTP is a critical factor not only for the lipidation of ApoB and hence the generation of VLDL but also in the regulation of CD1d function. Indeed, in an in vitro reductionist system, MTTP was able to transfer phospholipids to CD1d,35, 43 further suggesting that MTTP lipidation of CD1d is critical for CD1d to present both endogenous (ER loaded) and exogenous (endosomal or surface loaded) antigens to CD1d-restricted NKT cells. It has further been suggested by the results of Dougan et al43 that MTTP activity is required for NKT cell activation and NKT cell development. Interestingly, it was also shown by other groups that MTTP inhibition can not only cause a modest decrease in surface CD1d but also a marked defect in CD1d antigen presentation.35, 42 In addition, in different mouse models of NAFLD (eg, ob/ob mice and diet-induced NAFLD), it has been shown that the development of the disease was associated with a markedly decreased number of CD1d-reactive NKT cells in the liver.44, 45, 46 Furthermore, it was recently reported by Kremer et al47 that in hepatosteatosis, number of hepatic NKT cells is reduced in a Kupffer cell-dependent manner. In the present study, expression of CD1d and markers of NKT cells (eg, CD3-ɛ and NK1.1) were markedly higher in livers of PAI-1−/− mice regardless of feeding in comparison with wild-type mice. However, when determining IL-15 and IFN-γ, both shown before by others to be indicative for NKT cell activation,48, 49 we found a significant induction of these cytokines only in livers of fructose-fed PAI-1−/− mice. Taken together, these data suggest that in line with earlier findings of other groups,35, 42 the increased activity of MTTP found in livers of PAI-1−/− mice was associated not only with an increase of CD1d-reactive NKT cells but also an activation of these cells. Furthermore, these data also suggest that the loss of PAI-1 may be associated with an increased release of cytokines such as IL-15 and IFN-γ from NKT cells, adding to the protective effects found in the mice. Indeed, it has been suggested by Suzuki et al50 that IL-15 may contribute to regeneration activity in damaged livers. However, further studies are necessary to determine the role of NKT cells and IL-15 on the onset of fructose-induced steatosis.

The results of the present study suggest that PAI-1 may be a critical factor on the onset of fructose-induced steatosis through mechanisms involving a suppression of the compensatory upregulation of lipid export necessary to protect the liver from the increased amount of lipids resulting from a chronic dietary intake of fructose (see Figure 6). In addition, our data suggest that PAI-1 may also be involved in modulating CD1-reactive NKT cells in the liver (see Figure 6). However, whether similar mechanisms are also involved in the development of NAFLD in humans remains to be determined.

Figure 6
figure 6

Possible molecular mechanism involved in the development of fructose-induced fatty liver. Chronic intake of fructose may lead to elevated bacterial endotoxin levels in the portal blood and an activation of Kupffer cells, subsequently leading to an increased formation of ROS and an nuclear factor-κB-dependent induction of TNF-α. TNF-α can then cause insulin resistance (IR) in hepatocytes and an induction of PAI-1, which in turn may, through the c-Met signaling pathway, modulate hepatic lipid export, CD1d lipidation and NKT cell activation.

Full size image