Upregulation of proinflammatory and proangiogenic cytokines by leptin in human hepatic stellate cells

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Leptin upregulates collagen expression in hepatic stellate cells (HSCs), but the possible modulation of other actions has not been elucidated. The aim of this study was to investigate the expression and function of leptin receptors (ObR) in human HSCs and the biological actions regulated by leptin. Exposure of HSCs to leptin resulted in upregulation of monocyte chemoattractant protein 1 (MCP-1) expression. Leptin also increased gene expression of the proangiogenic cytokines vascular endothelial growth factor (VEGF) and angiopoietin-1, and VEGF was also upregulated at the protein level. Activated HSCs express ObRb and possibly other ObR isoforms. Exposure to leptin increased the tyrosine kinase activity of ObR immunoprecipitates and resulted in activation of signal transducer and activator of transcription 3. Several signaling pathways were activated by leptin in HSCs, including extracellular-signal–regulated kinase, Akt, and nuclear factor κB, the latter being relevant for chemokine expression. Leptin also increased the abundance of hypoxia-inducible factor 1α, which regulates angiogenic gene expression, in an extracellular-signal–regulated kinase– and phoshatidylinositol 3-kinase–dependent fashion. In vivo, leptin administration induced higher MCP-1 expression and more severe inflammation in mice after acute liver injury. Conversely, in leptin-deficient mice, the increase in MCP-1 messenger RNA and mononuclear infiltration was less marked than in wild-type littermates. Finally, ObR expression colocalized with VEGF and α-smooth muscle actin after induction of fibrosis in rats. In conclusion, ObR activation in HSCs leads to increased expression of proinflammatory and proangiogenic cytokines, indicating a complex role for leptin in the regulation of the liver wound-healing response.(HEPATOLOGY 2005;42:1339–1348.)

Leptin, the product of the obese (ob) gene, is a circulating peptide hormone mainly produced by the adipose tissue in proportion to its mass.1 In the hypothalamus, leptin regulates body weight homeostasis and energy expenditure with negative feedback. Defects in leptin production, as in ob/ob mice, or in leptin receptors, such as in db/db mice and fa/fa rats, causes severe hereditary obesity in rodents, but defective leptin production is very rare in humans, where obesity is usually associated with elevated leptin levels. Leptin's actions are mediated by leptin receptors (ObR), which belong to the class I cytokine receptor family and share common features with the interleukin 6 receptor.1 At least 6 isoforms of ObR are generated by alternative messenger RNA splicing, but in humans and rodents two major forms of leptin receptor are expressed. ObRb, also called the “long” isoform, is believed to be the signaling-competent receptor isoform, essential in mediating most of the biologic effects of leptin. Its activation triggers a signaling cascade involving the Jak family of protein tyrosine kinases and the signal transducer and activator of transcription factors.1 On the other hand, the short isoform, ObRa, is predominantly expressed in most tissues but is considered to have more limited signaling capabilities.

Originally thought to be only a satiety factor, leptin is a pleiotropic molecule that plays important roles in immune function, fertility, bone formation, and wound healing.2, 3 The influence of leptin on the pathophysiology of different systems also extends to the liver, where it modulates the response to injury. Leptin deficiency is associated with increased hepatotoxicity and mortality following endotoxin administration,4 but with reduced liver damage in models of T-cell–mediated hepatitis.5 Recent in vivo and in vitro studies have shown that leptin has a profibrogenic action on the liver related (at least in part) to a direct effect on the biology of matrix-producing cells such as hepatic stellate cells (HSCs).6 HSCs are liver-specific pericytes involved in the coordination of the wound healing response and in the development of fibrosis in the setting of chronic damage.7 Following injury, HSCs undergo a process known as “activation” and acquire a myofibroblast-like phenotype characterized by the ability to proliferate, migrate, and secrete fibrillar collagen and other matrix components critical for the process of liver fibrogenesis.

In this study we report that leptin increases the expression of proinflammatory and angiogenic cytokines by human HSCs, demonstrating a complex regulation of the liver wound-healing response. These events are associated with activation of different signaling pathways downstream of the leptin receptor, including nuclear factor κB (NF-κB) and hypoxia-inducible factor 1α (HIF-1α).

Abbreviations

HSC, hepatic stellate cell; ObR, leptin receptor; MCP-1, monocyte chemoattractant protein 1; VEGF, vascular endothelial growth factor; NF-κB, nuclear factor κB; HIF-1α, hypoxia-inducible factor 1α; α-SMA, α-smooth muscle actin; Stat, signal transducer and activator of transcription; ERK, extracellular-regulated kinase; JNK, c-jun N-terminal kinase; MAPK, mitogen-activated protein kinase.

Materials and Methods

Materials.

Monoclonal antibodies against α-smooth muscle actin (α-SMA) (clone 1A4) and β-actin were purchased from Sigma Chemical Co. (St. Louis, MO). Polyclonal antibodies against the phosphorylated forms of signal transducer and activator of transcription (Stat) 3, extracellular-regulated kinase (ERK), c-jun N-terminal kinase (JNK), p38MAPK, HSP27, Akt (Ser473), and p65NF-κB were obtained from Cell Signaling Technology (Beverly, MA). Monoclonal antibodies against VEGF and polyclonal antibodies against ObR, angiopoietin-1, and HIF-1α were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Recombinant human leptin, platelet-derived growth factor–BB (PDGF-BB) and interleukin 1, and recombinant murine leptin were from Peprotech (Rocky Hill, NJ). According to the manufacturer's note, the endotoxin levels in recombinant murine leptin were below 0.1 ng/μg as determined by the Limulus Amebocyte Lysate method. All other reagents were of analytical grade.

Cell Culture.

Human HSCs were isolated from wedge sections of liver tissue unsuitable for transplantation by collagenase/pronase digestion and centrifugation on Stractan gradients as previously described.8 All experiments were performed on at least three lines of cells cultured on uncoated plastic dishes, showing an “activated” or “myofibroblast-like” phenotype.

ELISA for MCP-1.

Confluent HSCs were deprived of serum for 24 hours. After replacement with fresh serum-free medium, the cells were incubated in the presence or absence of leptin for various time points. At the end of the incubation, the medium was collected and stored at −20°C until assayed. MCP-1 concentration in the conditioned medium was analyzed via ELISA using a commercially available antibody kit (Bender MedSystem Diagnostics, San Bruno, CA).

Analysis of Gene Expression.

Isolation of total RNA was performed using silica membrane filters (Macherey-Nagel, Düren, Germany). For RNAse protection assay, 2-10 μg of total RNA were analyzed using a commercially available kit and MultiProbe templates (BD Pharmingen, Palo Alto, CA) as previously described.9 For quantitative reverse-transcriptase polymerase chain reaction, total RNA (400 ng) was reverse-transcribed using Taqman Reverse Trancriptase Reagents (Applied Biosystems, Foster City, CA) following the manufacturer's protocol. Twenty-five nanograms of complementary DNA for each sample were analyzed in duplicate via quantitative polymerase chain reaction in an ABI Prism 7700 Sequence Detection System using the following conditions: 2 minutes at 50°C and 10 minutes at 95°C, followed by 40 cycles at 95°C for 15 seconds and 60°C for 1 minute. FAM-labeled probes and primers specific for MCP-1 (Mm00441242-m1) or the housekeeping gene GAPDH (Mm99999915-g1) were obtained from Applied Biosystems. Results are expressed as 2−ΔCt(ΔCt = Ct of the target gene − Ct of GAPDH).

ObR Signaling.

Western blotting and analysis of nuclear proteins were performed as previously described.9 After immunoprecipitation with anti-ObR antibodies, immune complex kinase assay was performed as reported elsewhere.10

Leptin Secretion.

Subconfluent HSCs were cultured in 24-well dishes in the presence of 500 μL of serum-free medium for 24-48 hours. Leptin concentration in the cell-conditioned medium was analyzed via radioimmunoassay (DSL-23100, DSL Inc., Webster, TX).

Animal Experiments.

All animals received humane care, and experimental protocols were conducted according to national and local guidelines. Specific pathogen-free, 5-week-old male C57Bl/6 mice were purchased from Charles River (Calco, Italy). Mice (n ≥ 3 in each group) were administered CCl4 (0.1 μL/g body wt in olive oil) and/or recombinant murine leptin (1 μg/g body weight) simultaneously via intraperitoneal injection and sacrificed after 24 hours.11

Six-week-old ob/ob mice and their lean littermates were obtained from The Jackson Laboratory (Bar Harbor, ME). Animals (n ≥ 3 in each group) were treated with a single intraperitoneal injection of CCl4 (1 μL/kg body weight) and sacrificed after 48 hours.

Male adult Wistar rats (Harlan-Nossan, Correnzana, Italy), initial weight 200-220 g, were fed with a standard pelleted diet and water ad libitum. Fibrosis was induced by treatment with CCl4 administered by gavage twice a week for 9 weeks.12 Control animals received an equal volume of vehicle.

Histology.

Paraffin-embedded liver tissue was cut and stained with hematoxylin-eosin and examined by an expert in liver pathology (S.M.) blinded to the type of treatment received by the animals. A score of 0-3 was provided for necrosis, mononuclear infiltration, and steatosis.

Immunofluorescence.

Experiments with HIF-1α were performed as previously described.13 Anti–HIF-1α antibodies were used at a dilution of 1:100.

For angiopoietin-1 and α-SMA, HSCs cultured on coverglass slips were fixed and permeabilized for 10 minutes in 4% paraformaldehyde, 0.2% Triton X-100 in phosphate-buffered saline. Cells were then dehydrated with acetone at −20°C for 5 minutes and blocked for 1 hour at room temperature with 3% bovine serum albumin and 0.05% Igepal-40 in phosphate-buffered saline. Cells were then sequentially incubated with primary α-SMA (1:100 dilution) and angiopoietin-1 antibodies (1:50 dilution) and secondary Alexa Fluor 488-labeled antibodies (1:75 dilution; Molecular Probes, Leiden, The Netherlands) and Texas Red (1:200 dilution; Santa Cruz Biotechnology). Some slides (negative controls) were stained solely with the secondary antibodies.

Indirect immunofluorescence on liver cryostat sections (6 μm) was performed essentially as described elsewhere.14 Primary antibodies against α-SMA, VEGF, and ObR were used at a dilution of 1:250, 1:100, and 1:100, respectively. Immunopositivity was revealed by the appropriate Cy3- (dilution 1:1,000) or FITC-conjugated antibodies (dilution 1:200).

Statistical Analysis.

Autoradiograms and autoluminograms are representative of three or more experiments with comparable results. Barograms show the mean ± SD of at least three independent experiments. Statistical analysis was performed via Student t test or Mann-Whitney test as appropriate.

Results

Leptin Upregulates MCP-1 and VEGF Secretion.

We tested the effects of leptin on the secretion of MCP-1, a proinflammatory chemokine. Exposure of HSCs to leptin caused an increase in MCP-1 secretion, an effect evident after 24 hours and maintained at later time points (Fig. 1A). Concentrations of leptin of 50 ng/mL were sufficient to induce upregulation of MCP-1, and peak effects were obtained at 100 to 200 ng/mL (data not shown). We next evaluated the effects of leptin on MCP-1 messenger RNA levels. Compared with serum-deprived HSCs, a markedly increased expression of the MCP-1 gene was present after stimulation with leptin (Fig. 1B).

Figure 1.

Leptin induces protein and gene expression of MCP-1 in human HSCs. (A) Serum-deprived HSCs were incubated in the presence or absence of 200 ng/mL leptin for the indicated times. MCP-1 secretion in the supernatants was analyzed via ELISA. *P < .05. (B) Total RNA isolated from HSCs exposed to leptin for different periods was analyzed for the expression of the indicated genes by RNAse protection assay as described in Materials and Methods. Transfer RNA (lane 1) was used as a negative control for hybridization. The size of the protected fragments is shown on the left. MCP-1, monocyte chemoattractant protein 1; tRNA, transfer RNA; IL-8, interleukin 8.

Because HSCs have been shown to contribute to angiogenesis,15 we explored whether leptin modulates proangiogenic cytokines. VEGF and angiopoietin-1 were expressed in unstimulated conditions (Fig. 2A). In response to leptin, messenger RNA levels of VEGF were markedly increased, whereas more modest effects were observed on angiopoietin-1 (Fig. 2A). The expression of both cytokines was clearly upregulated by culturing HSCs in conditions of hypoxia. Because expression of angiopoietin-1 had not been previously reported in HSCs, we performed double immunofluorescence for α-SMA, a marker of activated HSCs, and angiopoietin-1, confirming the expression of this cytokine (Fig. 2B). To further investigate the effects of leptin at the protein level, we performed immunoblot analysis of cell lysates using specific antibodies against VEGF or angiopoietin-1. These experiments confirmed that the expression of VEGF by HSCs is upregulated by leptin (Fig. 2C-D). In contrast, the effects of leptin on angiopoietin-1 expression were very modest (Fig. 2E) and were not significant when three independent experiments were analyzed (data not shown).

Figure 2.

Effects of leptin on the expression of angiogenic molecules. (A) Serum-deprived HSCs were cultured in condition of hypoxia (3% O2 for 6 hours) or exposed to 200 ng/mL leptin for the indicated time points. Total RNA was analyzed for the expression of the indicated genes by RNAse protection assay as described in Materials and Methods. Transfer RNA (lane 1) was used as a negative control for hybridization. Lane 2 shows a positive control for VEGF. The size of the protected fragments is shown on the left. (B) Combined immunofluorescence for α-SMA (left panel, red) and angiopoietin-1 (middle panel, green). The right panel shows the overlaid pictures. Original magnification ×400. (C) HSCs were incubated in the presence or absence of 200 ng/mL leptin for 24 hours. Total cell lysates were analyzed via immunoblotting using anti-VEGF antibodies. The blot was stripped and reprobed with anti–β-actin antibodies to ensure equal loading. Migration of molecular weight markers is indicated on the left. (D) Densitometry data from three independent experiments performed as described in panel C are shown as the mean ± SD. *P < .05. (E) The experiment was performed exactly as described in panel C, blotting for angiopoietin-1. tRNA, transfer RNA; VEGF, vascular endothelial growth factor; α-SMA, α-smooth muscle actin.

Human HSCs Express Functional Leptin Receptors.

To correlate the biological activities of leptin with the expression of leptin receptors, we analyzed HSCs isolated from four different donors using antibodies recognizing all ObR isoforms. An approximately 190-kd band representing ObRb, as reported in other cell types or in stable HSC lines,16–18 was found in all isolates (Fig. 3A), together with faster migrating bands that are likely due to ObRa and/or other “short” isoforms. To assess ObR functionality, we performed an immune complex tyrosine kinase assay of the receptor. Leptin increased phosphorylation levels of ObRb (Fig. 3B), as indicated by a band of similar molecular weight to the one observed via Western blot analysis (Fig. 3A). In addition, other proteins appeared to be phosphorylated in response to leptin. The Jak/Stat signaling pathway plays a critical role in mediating the effects of leptin, and Stat3 phosphorylation has been associated with ObRb activation in different cells, including HSCs.17, 18 To further establish the functionality of ObRb, we analyzed the effects of leptin on Stat3 activation (Fig. 3C). Phosphorylation of Stat3 appeared as early as 5 minutes after addition of leptin, indicating that human-activated HSCs express a functional “long form” of leptin receptor.

Figure 3.

Human HSCs express functional ObRb. (A) Culture-activated HSCs from four different isolates were analyzed for expression of leptin receptors via immunoblotting using antibodies recognizing all ObR isoforms. An approximately 190-kd band representing ObRb was found in all cell lines. (B) Serum-deprived HSCs were exposed to 200 ng/mL leptin for the indicated time points. Total cell lysates were immunoprecipitated with anti-ObR antibodies and analyzed for tyrosine kinase activity, using immune complex kinase assay, as described in Materials and Methods. A major phosphorylated band migrating at approximately 190 kd is indicated by an arrow. Other bands with increased phosphorylation in response to leptin are indicated by arrowheads. (C) Serum-deprived HSCs were exposed to 200 ng/mL leptin for the indicated time points. Total protein lysates were analyzed via immunoblotting using antibodies directed against tyrosine-phosphorylated Stat3. The membrane was reprobed for α-SMA to ensure equal loading. In all panels, migration of molecular weight markers is indicated on the left. ObR, leptin receptor; IP, immunoprecipitation; STAT3; signal transducer and activator of transcription 3; α-SMA, α-smooth muscle actin; ICKA, Immune Complex Kinase Assay.

Effects of Leptin on Activation of Different Mitogen-Activated Protein Kinases.

We analyzed the ability of leptin to induce activation of three major members of the mitogen-activated protein kinase (MAPK) family (Supplementary Fig. 1). Exposure of HSCs to leptin resulted in increased ERK phosphorylation of both isoforms (Supplementary Fig. 1A). We next investigated whether exposure of HSCs to leptin was associated with increased activation of JNK or p38MAPK. Leptin resulted in long-lasting phosphorylation of JNK, particularly of the 54-kd isoform (Supplementary Fig. 1B), while p38MAPK was not significantly activated (Supplementary Fig. 1C). This latter result was confirmed by the inability of leptin to increase phosphorylation of heat-shock protein 27, a target of p38MAPK (Supplementary Fig. 1D).

Because leptin has been shown to be expressed by activated rat HSCs,19 it is possible that autocrine secretion influences the response to recombinant leptin. However, similar effects on ERK activation were observed when recombinant leptin was added to the same media in which HSCs were cultured for the previous 24 hours, or when media were changed to remove autocrine leptin (data not shown). Moreover, leptin concentrations in media conditioned by HSCs for as long as 48 hours were consistently lower than 0.1 ng/mL. Thus, it is unlikely that an autocrine loop influenced the responses to recombinant leptin.

Leptin Activates NF-κB and Akt.

We then explored the effects of leptin on three nuclear signaling pathways. Electrophoretic mobility shift assay using specific consensus oligonucleotides for activator protein 1, NF-κB, and Stat3 indicated activation of all three transcription factors (Fig. 4). Activation of activator protein 1 and Stat3 confirms the ability of leptin to activate MAPK pathways, such as ERK or JNK, and to increase Stat3 phosphorylation, as described in Fig. 3 and Supplementary Fig. 1. Because activation of NF-κB in response to leptin is a novel finding in this cell type, we also investigated the contribution of different NF-κB subunits. Exposure of HSCs to leptin resulted in increased phosphorylation of the p65 subunit with peak activation at 5-10 minutes after addition of leptin (Supplementary Fig. 2A). To establish the possible role of other Rel proteins, we performed ELISA of proteins binding NF-κB oligonucleotides.20 Leptin induced a two-fold increase in p50 binding (Supplementary Fig. 2B), whereas no effects were found on p52, RelB, or c-Rel (data not shown). Thus, p50 and p65 appear to be the major targets of the action of leptin on NF-κB.

Figure 4.

Activation of transcription factors by leptin in HSCs. Serum-deprived HSCs were exposed to 200 ng/mL leptin for the indicated time points, 10 ng/mL PDGF-BB, or 10 ng/mL interleukin 1 (for 30 minutes) as positive controls as indicated. Nuclear extracts were analyzed via electrophoretic mobility shift assay as described in Materials and Methods, using consensus oligonucleotides for activator protein 1 (lanes 1-7), NF-κB (lanes 8-13), or Stat3 (lanes 14-19). Position of the retarded complexes is indicated by an arrow. IL- 1, interleukin 1; AP-1, activator protein 1; NF-κB, nuclear factor κB; Stat3, signal transducer and activator of transcription; PDGF, platelet-derived growth factor.

Phosphatidylinositol 3-kinase and its downstream effector, Akt, may be involved in the expression of angiogenic cytokines.21 Treatment of human HSCs with leptin rapidly increased Akt phosphorylation on the activation-specific residue, Ser473 (Supplementary Fig. 3).

Leptin Activates HIF-1α, a Key Factor in Angiogenesis.

Expression of VEGF and other angiogenic pathways is critically dependent on the activation of HIF-1α.21 When compared with unstimulated cells, the intracellular levels of HIF-1α were markedly increased after exposure to leptin, as indicated by the two specific bands observed in Western blot analysis (Fig. 5). To better define the activation of this pathway, we evaluated HIF-1α expression and localization with immunofluorescence (Fig. 6). As expected, incubation of the cells in conditions of hypoxia led to a marked increase in HIF-1α–specific signal and predominant nuclear localization (Fig. 6B). Similar changes, although less marked, were observed after exposure to leptin (Fig. 6C). Preincubation of HSCs with either the MEK inhibitor PD98059 or with the phoshatidylinositol 3-kinase/Akt inhibitor LY294002 abrogated the effects of leptin on HIF-1α (Fig. 6D-E).

Figure 5.

Leptin increases the HIF-1α levels in HSCs. Serum-deprived HSCs were exposed to 200 ng/mL leptin for the indicated time points. Total protein lysates were analyzed by immunoblotting using antibodies directed against HIF-1α. The membrane was stripped and reprobed for β-actin to ensure equal loading. Migration of molecular weight markers is indicated on the left. HIF-1α, hypoxia-inducible factor 1α.

Figure 6.

Leptin-induced nuclear translocation of HIF-1α requires ERK and phosphatidylinositol 3-kinase. Serum-deprived HSCs were (A) left untreated, (B) incubated in conditions of hypoxia for 1 hour, or (C) exposed for 1 hour to 200 ng/mL leptin alone or (D) after preincubation with 30 μmol/L PD98059 or (E) 10 μmol/L LY294002. Expression of HIF-1α (red fluorescence) was analyzed via indirect immunofluorescence as described in Materials and Methods. Nuclear DNA was stained with DAPI (blue fluorescence). Original magnification ×1000.

Leptin Administration Increases MCP-1 Expression and Hepatic Inflammation During Acute Liver Damage In Vivo.

We next evaluated the effects of recombinant leptin on MCP-1 expression and liver necro-inflammation after CCl4 administration in mice. As expected, acute toxic injury resulted in a marked upregulation of hepatic MCP-1 expression (Fig. 7A). When leptin was injected at the same time as CCl4, a further, significant increase in the levels of MCP-1 transcripts was observed. Increased chemokine expression was also associated with significantly more severe necrosis and inflammation with respect to animals treated with the toxin alone (Fig. 7B), while steatosis was not different (data not shown). In addition, serum aminotransferase levels were significantly higher when toxic damage was coupled with leptin injection (Fig. 7C). Injection of leptin in the absence of CCl4 did not affect MCP-1 expression, liver histology, or enzyme levels.

Figure 7.

Effects of recombinant leptin on MCP-1 expression in mice with acute liver damage. C57Bl/6 mice (n ≥ 3 in each group) were administered CCl4 (0.1 μL/g) and/or recombinant murine leptin (Lpt, 1 μg/g body weight) via intraperitoneal injection and killed 24 hours later. (A) Total RNA was analyzed for MCP-1 expression using quantitative reverse-transcriptase polymerase chain reaction as described in Materials and Methods. (B) Histological scores for necrosis and inflammation as indicated. Each symbol indicates an individual animal. (C) Serum alanine aminotransferase levels at the time of killing. *P ≤ .05 versus CCl4 alone. MCP-1, monocyte chemoattractant protein 1; Cnt, control; Lpt, leptin; ALT, alanine aminotransferase.

Reduced MCP-1 Expression and Liver Inflammation in Leptin-Deficient Mice.

To provide additional evidence for the role played by leptin in chemokine expression and liver inflammation, we compared the effects of acute liver damage in leptin-deficient Ob mice with those induced in their wild-type, lean littermates. Also in this case, administration of CCl4 led to a dramatic increase in the intrahepatic levels of MCP-1 messenger RNA in lean controls (Fig. 8A). However, in leptin-deficient animals, the increase in MCP-1 was significantly less marked, confirming that leptin contributes to the expression of this chemokine during liver injury. Remarkably, diminished MCP-1 expression was accompanied by a less severe histological phenotype in Ob mice, as indicated by the lower scores for necrosis and inflammation (Fig. 8B). As expected, steatosis score was higher in Ob mice both in control conditions and after CCl4 administration (Fig. 8B). Higher aminotransferase levels were also observed in lean mice, although this difference did not reach statistical significance (data not shown).

Figure 8.

Effects of leptin deficiency on MCP-1 expression and liver histology. Leptin-deficient mice (Ob) and their wild-type littermates (Lean) were administered CCl4 (1 μL/g in olive oil) via intraperitoneal injection and killed 24 hours later. (A) Total RNA was analyzed for MCP-1 expression using quantitative reverse-transcriptase polymerase chain reaction as described in Materials and Methods. (B) Histological scores for necrosis, inflammation, and steatosis as indicated. Each symbol indicates an individual animal. *P ≤ .05 versus lean CCl4. **Mann-Whitney test not feasible. MCP-1, monocyte chemoattractant protein 1; Cnt, control.

Colocalization Between Expression of Leptin Receptors and VEGF During Chronic Injury.

The possible relationship between fibrogenesis, leptin, and angiogenesis in vivo was studied in a rat model of fibrosis. In normal rat liver, expression of ObR was barely detectable, and expression of α-SMA was limited to vascular structures in the portal tracts (Fig. 9A). After induction of fibrosis due to chronic CCl4administration, expression of ObR was increased, and α-SMA–positive cells appeared in fibrotic septa. The signal for α-SMA and ObR showed a partial colocalization, indicating that during chronic injury, expression of leptin receptors may be found in activated HSCs. To more directly link leptin and angiogenic cytokines, we performed double immunofluorescence for ObR and VEGF in the same liver samples (Fig. 9B). VEGF staining, detectable in the hepatic lobule of normal rats, increased in intensity during chronic injury. Importantly, VEGF was abundantly expressed in areas of significant fibrosis and partially colocalized with ObR, indicating that cells responsive to leptin also express VEGF.

Figure 9.

Colocalization between expression of leptin receptors and VEGF during chronic injury. Immunofluorescence was performed in control Wistar rats or after induction of fibrosis by 9 weeks of treatment with CCl4 administered via gavage twice a week. (A) Combined immunofluorescence for α-SMA (left panel, green) and ObR (middle panel, red). The right panels show overlaid pictures in which signal colocalization (yellow) is indicated by arrows. (B) Combined immunofluorescence for VEGF (left panel, green) and ObR (middle panel, red). The right panels show overlaid pictures in which signal colocalization (yellow) is indicated by arrows. Original magnification ×400. Cnt, control; hl, hepatic lobule; a, hepatic artery; pt, portal tract; s, fibrotic septum; α-SMA, α-smooth muscle actin; ObR, leptin receptor.

Discussion

Leptin has recently emerged as an important mediator in the development of liver fibrogenesis. Absence of leptin or leptin signaling results in a marked reduction of liver fibrosis induced by different conditions, including thioacetamide intoxication, chronic CCl4 administration or experimental nonalcoholic steatohepatitis.18, 22, 23 In addition, leptin has been shown to stimulate the expression of collagen and to induce proliferation and survival of cultured HSCs.16, 18, 24 HSCs also regulate inflammation and promote angiogenesis, via secretion of various cytokines, including chemokines, and VEGF.25, 26 In this study, we report for the first time that leptin modulates the ability of HSCs to express the proinflammatory cytokine MCP-1 and the proangiogenic factor VEGF, and provide evidence for an in vivo significance of these findings.

MCP-1 is a critical mediator of intrahepatic inflammation and is the principal monocyte chemotactic factor secreted by HSCs.25 In addition, HSCs express MCP-1 in vivo in animal models of liver injury and in patients with liver disease.27, 28 Exposure of HSCs to leptin also led to increased VEGF expression at both gene and protein levels. Angiogenesis has been recently identified as a mechanism contributing to liver fibrogenesis and to the progression of portal hypertension in chronic liver injury.29, 30 Moreover, expression of VEGF by HSCs has been observed in the areas surrounding hepatocellular carcinoma and in the stromal tissue adjacent to hepatic metastases.26, 31 Thus, through the induction of this cytokine, leptin may contribute to local angiogenesis and tumor progression, especially in the context of nonalcoholic steatohepatitis.32 The effects of leptin on angiogenic cytokines may also explain the observation that obesity associated with hyperleptinemia is a risk factor for the development of hepatocellular carcinoma.33 In the present study, we also report the ability of HSCs to express angiopoietin-1. By binding to its cognate receptor, Tie-2, angiopoietin-1 mediates neovascular maturation into more complex and larger vascular structures and maintains vessel integrity.34 Although leptin's effect on angiopoietin-1 was not significant, these data underscore the potential relevance of HSCs in the processes leading to neovascularization during fibrosis. Further studies elucidating the mechanisms that regulate expression of angiopoietin-1 in HSCs are warranted.

The results obtained in cultured cells have also been extended to the in vivo situation. A connection between leptin, HSCs, and the expression of angiogenic cytokines is indicated by the observation that in rats with chronic liver damage, expression of ObR was increased and colocalized with α-SMA–positive cells and cells expressing VEGF. Furthermore, we demonstrated an in vivo link between leptin and the proinflammatory chemokine MCP-1 via two complementary approaches. Simultaneous administration of CCl4 and recombinant leptin resulted in higher intrahepatic levels of MCP-1 than when the toxin was administered alone. On the other hand, in leptin-deficient mice the increase in MCP-1 expression produced by CCl4 was significantly lower than in lean littermates. Importantly, in both models modulation of chemokine expression was paralleled by concordant phenotypic changes, as shown by the fact that liver inflammation was more marked when leptin was injected, while it was blunted in Ob mice. Taken together, these data provide compelling evidence for a role of leptin in hepatic MCP-1 expression in vivo.

The molecular and cellular mechanisms underlying the effect of leptin on fibrogenesis are still debated. The “long” form of the receptor, ObRb, is capable of fully transducing leptin's signals, while the short forms, particularly the widely expressed ObRa, possess limited signaling capabilities. Controversial results regarding the type of ObR expressed by rodent primary HSCs or immortalized human lines have been reported,16, 18, 35, 36 and the present study is the first to provide information on expression and signaling of ObR in primary human HSCs. Our data are consistent with the observations of Cao et al.16 and Saxena et al.,18 indicating that HSC express ObRb, together with one or more of the short forms. Accordingly, in the present study leptin increased Stat3 phosphorylation, as shown in cells expressing ObRb.1

The results of the present study provide additional information on the signaling pathways downstream of leptin receptors in human HSCs, characterizing the activation of pathways related to the induction of proinflammatory and angiogenic cytokines. In fact, leptin led to a marked activation of the transcription factor NF-κB, as demonstrated by electrophoretic mobility shift assays, increased phosphorylation of the p65 subunit, and increased nuclear translocation of p50. The fact that NF-κB is a target of ObR has not been previously reported in HSCs but was recently described in colonic epithelial cells.37 NF-κB activation is critical for the transcriptional induction of several proinflammatory genes, including chemokines such as MCP-1.38 Accordingly, inhibition of NF-κB activation blocked the increase in MCP-1 in response to leptin (data not shown).

Another intriguing and novel aspect of the present study is related to the ability of leptin to activate HIF-1α, a central factor in the modulation of cellular responses to hypoxia.39 HIF-1α participates in transcriptional activation of genes mediating adaptive responses to reduced oxygen availability, including VEGF, the most prominent HIF-1α target gene involved in vascular biology.40 In response to hypoxia, inhibition of prolyl hydroxylation leads to dissociation of HIF-1α from the inhibitory factor, VHL.39 An alternative mechanism of activation is dependent on cytokine receptors, which lead to increased recruitment of HIF-1α.21 Cytokine-mediated activation of HIF-1α has been shown to be critically dependent on the ERK group of mitogen-activated protein kinases and/or on the phoshatidylinositol 3-kinase/Akt pathway, both of which mediate an early increase in cellular levels of this factor.21, 40, 41 In addition, ERK-mediated phosphorylation of HIF-1α enhances its transcriptional activity.42 Activation of ERK and Akt downstream of leptin receptors has been previously described24 and was confirmed by the present data, which provide a molecular link connecting leptin to upregulation of HIF-1α and VEGF expression. Interestingly, leptin has been shown to induce angiogenesis also through a direct effect on vascular endothelial cells.43, 44 Leptin's ability to upregulate VEGF expression provides an alternative mechanism underlying this relevant biological action.

Although these data confirm and expand the role of leptin in liver wound healing, several aspects deserve additional investigation. For instance, db/db mice, which lack functional ObRb but express ObRa, develop significant fibrosis in a dietary model of nonalcoholic steatohepatitis,45 indicating that expression of the long receptor may not be necessary to mediate fibrosis. Studies employing selective isoform silencing are required to dissect the relative role of different ObR in HSCs. In humans, it has been proposed that leptin may limit ectopic fat accumulation, and hyperleptinemia in patients with fatty liver could be related to leptin resistance in hepatocytes.46 Leptin administration has been shown to reduce steatosis in patients with lipodystrophy, and it has been proposed that high doses of leptin could circumvent leptin resistance in obese patients with fatty liver.47, 48 In contrast, human HSCs appear to be sensitive to leptin's actions, although this aspect is difficult to investigate in vivo. The profibrogenic and proangiogenic effects of leptin recommend some caution with respect to the possible therapeutic use of recombinant leptin. Further information on the mechanisms underlying hepatocyte resistance and cell-specific responses to leptin is required to elucidate this apparent paradox and to devise optimal therapeutic strategies.

In conclusion, expression of functional leptin receptors by human HSCs mediates the upregulation of proinflammatory and proangiogenic cytokines. These data argue for a complex role of leptin in the regulation of liver wound-healing response. Because inflammation and angiogenesis contribute significantly to the development of liver fibrosis, these data reveal a novel connection between leptin, liver fibrogenesis, and possibly liver cancer.

Acknowledgements

The expert technical help of Nadia Navari and Wanda Delogu is gratefully acknowledged. We are also indebted to Dr. Paola Parronchi for help with the NF-κB ELISAs and to Dr. Sergio Paladini for the leptin radioimmunoassay.

Ancillary