Adiponectin, an adipocytokine, has been identified in adipose tissue, and its receptors are widely distributed in many tissues, including the liver. The present study was performed to clarify the role of adiponectin in lipopolysaccharide (LPS)-induced liver injury using KK-Ay obese mice. We analyzed the effects of adiponectin pretreatment on liver injury induced by D-galactosamine/LPS (GalN/LPS) in KK-Ay obese mice. GalN/LPS treatment induced significant increases in aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels in the blood, apoptotic and necrotic changes in hepatocytes, and/or showed a high degree of lethality. The GalN/LPS-induced liver injury was more pronounced in KK-Ay obese mice than in lean controls. Pretreatment with adiponectin ameliorated the GalN/LPS-induced elevation of serum AST and ALT levels and the apoptotic and necrotic changes in hepatocytes, resulting in a reduction in lethality. In addition, pretreatment with adiponectin attenuated the GalN/LPS-induced increases in serum and hepatic tumor necrosis factor α (TNF-α) levels and increased peroxisome proliferator-activated receptor (PPAR) α messenger RNA expression in the liver. Furthermore, abdominal macrophages from KK-Ay obese mice pretreated with adiponectin in vitro exhibited decreased LPS-induced TNF-α production compared with controls. Finally, adiponectin pretreatment also ameliorated TNF-α-induced liver injury. In conclusion, these findings suggest that adiponectin prevents LPS-induced hepatic injury by inhibiting the synthesis and/or release of TNF-α of KK-Ay obese mice. (HEPATOLOGY 2004;40:177–184.)
Adiponectin, an adipocytokine, has been identified in adipose tissue as a result of screening for adipose-specific genes.1–4 Adiponectin messenger RNA (mRNA) is present exclusively in adipose tissue and blood,1–4 and its receptors are abundant in liver and muscle.5 Blood levels of adiponectin and its mRNA levels in adipose tissue are negatively correlated with body mass index.3, 4 In fact, levels of adiponectin and its mRNA have been shown to be lower in obese models, including agouti yellow obese mice.6 This suggests that a reduction in adiponectin expression may be involved in the development of obesity or obesity-related metabolic disorders, such as insulin resistance, hyperlipidemia, and atherosclerosis. The antidiabetic, antilipogenic, and antiatherogenic actions of adiponectin have been investigated in several previous studies.7–11 Adiponectin receptors that mediate antidiabetic metabolic effects have been identified in peripheral tissues; adiponectin R2 receptors are particularly abundant and predominant in the liver.5
In addition to metabolic disorders, various inflammatory diseases are common complications of obesity.12–14 Recent evidence has suggested a role of adiponectin in inflammation relative to its antiatherogenic action.10, 11, 15 Treatment with exogenous adiponectin has been shown to improve atherosclerosis through the inhibition of macrophage aggregation.11, 15 Adiponectin treatment also regulates immunocytokines, including tumor necrosis factor α (TNF-α), in vitro15 and improves nonalcoholic steatohepatitis via the inhibition of lipogenic factors and TNF-α.16 These previous findings suggest that reduced production of adiponectin may be related to the pathogenesis of immune dysfunction in obesity. However, little is known about the effects or detailed mechanisms of action of adiponectin in inflammatory diseases or related cytokine interactions.
Lipopolysaccharide (LPS)-induced liver injury has been used as an experimental model in which to analyze the mechanism of endotoxin-induced acute liver injury.17, 18 A number of inflammatory cytokines, including TNF-α, play pivotal roles in the development of LPS-induced liver injury.19–21 Compared with the effects of LPS alone, the administration of LPS together with or following a sublethal dose of D-galactosamine (GalN) induced more severe hepatic damage accompanied by apoptotic and necrotic changes in the liver.19, 21 TNF-α also causes hepatocyte apoptosis in this severe liver injury model.20, 21 Peroxisome proliferator activated receptor α (PPAR-α), a target for adiponectin,6 has been shown to suppress hepatocyte apoptosis and to regulate the production of TNF-α.22, 23 We hypothesized that exogenous adiponectin treatment may modulate the development of GalN/LPS-induced liver injury by affecting TNF-α. To clarify the protective role of adiponectin in acute fulminant liver injury, we analyzed the effects of adiponectin on the GalN/LPS-induced increase in lethality, the elevation of serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels, and apoptosis and necrosis of hepatocytes. We also analyzed the effects of adiponectin on the expression of PPAR-α mRNA in the liver and on the GalN/LPS-induced production of hepatic TNF-α in KK-Ay obese mice.
Male KK-Ay obese and lean control mice, 8 to 10 weeks old, were purchased from Seac Yoshitomi (Yoshitomi, Fukuoka, Japan). Mice were housed in a bacteria-free room at a temperature of 21 ± 1°C and with 55 ± 5% humidity and daily illumination between 0700 and 1900 (12 hours light, 12 hours dark). The mice were allowed access to standard powdered mouse food (CLEA Japan Ltd., Tokyo, Japan) and tap water ad libitum. All animals were treated in accordance with the Oita Medical University guidelines for the care and use of laboratory animals.
Mammalian full-length adiponectin (Oita Medical University, Oita, Japan) was dissolved in phosphate-buffered saline (PBS) at a concentration of 1 μg/μL; the solution was freshly prepared on the day of administration. Endotoxin levels were less than 0.1ng/mg protein as measured by an endotoxin testing kit (Seikagaku Corp., Tokyo, Japan). LPS (Sigma, St. Louis, MO) and TNF-α (Dainippon Pharmacy, Osaka, Japan) were dissolved in PBS at concentrations of 0.1 μg/μL and 2 ng/μL, respectively. The control solutions were also freshly prepared on the day of administration.
For the preparation of mice with GalN/LPS-induced fulminant hepatic failure, mice were given an intraperitoneal injection of GalN (800 μg/g body weight; Sigma), immediately followed by an intraperitoneal injection of LPS (100 ng/g body weight) or TNF-α (3 ng/g body weight). To determine the effects of adiponectin on mice with fulminant hepatic failure, adiponectin (1.5 μg/g body weight) was injected intraperitoneally at 2, 24, and 48 hours before GalN/LPS administration. The doses of LPS, TNF-α, and adiponectin were determined in previous studies6, 23 and in our preliminary study. In addition, the doses of adiponectin alone did not induce liver injury as determined by evaluating liver enzymes, cytokines, and liver histology. Serum and liver samples for TNF-α analysis or histology were obtained at 0, 1.5, 6, and 24 hours after the administration of GalN/LPS (n = 4-6 for each group).
Blood AST, ALT, TNF-α, and Adiponectin Levels.
Serum levels of liver enzymes, including AST and ALT, were determined using an automatic analyzer (SRL, Tokyo, Japan). Serum TNF-α (BioSource, Tokyo, Japan) and adiponectin (Otsuka Pharmacy, Tokyo, Japan) were determined by enzyme-linked immunosorbent assays (ELISA) according the manufacturer's protocol.
Small pieces of liver were removed, rinsed with saline, fixed with 10% formalin, and embedded in paraffin. Tissue sections were cut at a thickness of 5 μm and stained with hematoxylin-eosin (HE). To examine hepatocytes, HE-stained liver sections were analyzed with an image analysis system (Olympus, Tokyo, Japan).
Preparation of Peritoneal Macrophages, Peritoneal Fluid, and Culture Medium.
Mice were maintained free of specific pathogens in autoclaved cages in a laminar flow hood to minimize the spontaneous activation of macrophages. To produce responsive macrophages, 4 mL of sterile saline was injected intraperitoneally. The mouse peritoneal cavity was washed with 4 mL of saline. The exudate was centrifuged at 300g for 15 minutes at 4°C, and the supernatant was decanted and analyzed. The resultant peritoneal cells were obtained and plated at 5 × 105 cells/well in 96-well plates with RPM1 medium and 10% heat-inactivated PBS (Invitrogen Corp., Tokyo, Japan). After incubation for 30 minutes at 37°C in a 5% CO2 atmosphere, nonadherent cells were removed by washing with ice-cold saline, and adherent macrophages were used for the experiments. The adherent cells were greater than 95% macrophages. Resident peritoneal macrophages were prepared from untreated control mice in an LPS-free system. Attached macrophages were cultured in RPM1 and used for experiments. All reagents to which macrophages were exposed in culture were free of detectable LPS. Peritoneal macrophages cultured in RPM1 medium in 96-well plates were stimulated with LPS (10 ng/mL); the medium was replaced with 0.2 mL of fresh medium containing 10 ng/mL LPS at 0, 2, or 4 hours after the addition of LPS. After incubation for 0, 2, or 4 hours, the replaced medium was collected into microcentrifuge tubes and centrifuged at 900g for 5 minutes at 4°C. The supernatant was concentrated by centrifugal filtration using filter tubes (Millipore, Bedford, MA), and the TNF-α released from macrophages was analyzed by ELISA.
ELISA of TNF-α
Liver, serum, and medium TNF-α were determined with an ELISA kit (BioSource) and an optical density reader according to the manufacturer's instructions. Liver samples (100 mg) that had been frozen at −80°C were homogenized in 1 mL of PBS and centrifuged at 1500g. The supernatant was collected and stored at −80°C. Protein concentrations of the liver solutions were analyzed using the method of Bradford (BioRad Lab, Richmond,CA); OD readings of samples were converted to pg/mL using standard curves generated with the recombinant cytokine supplied with the kit. The limit of detection was 4.87 pg/mL for each assay.
Detection of Apoptosis Using In Situ Labeling.
Liver tissues were fixed in 10% phosphate-buffered formalin, embedded in paraffin, cut into serial sections 5 μm thick, and stained. Apoptotic hepatocytes were detected with the in situ terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) method using an apoptosis kit (Medical Biological Lab, Nagoya, Japan). Sections were treated with proteinase K, and endogenous peroxidase activity was blocked by treatment with 0.02% hydrogen peroxide. Tissue sections were treated with a mixture of terminal deoxynucleotidyl transferase, digoxigenin-labeled dUTP, and dATP at 37°C for 1 hour, followed by incubation with peroxidase-labeled antidigoxigenin antibody solution for 30 minutes. As a negative control, PBS was substituted for the mixture of deoxynucleotidyl transferase, digoxigenin-labeled dUTP, and dATP.
Liver TNF-α, PPAR-α, and PPAR-γ mRNAs were determined by polymerase chain reaction (PCR) amplification and quantified by real-time quantitative PCR. Total cellular RNA was prepared from selected mouse tissues using TRIzol (Lifetech, Tokyo, Japan) according to the manufacturer's protocol. Total RNA (20 μg) was electrophoresed on 1.2% formaldehyde-agarose gels. RNA quality and quantity were assessed by EtBr-agarose gel electrophoresis and by measuring the relative absorbance at 260 nm versus that at 280 nm. Complementary DNA was synthesized from 150 ng of total RNA in a volume of 20 μL with a ReverTra-Dash reverse transcriptase kit (Toyobo, Tokyo, Japan) using random hexamer primers. Reactions were diluted to 50 μL with sterile distilled H2O and stored at −20°C. Primers for mouse TNF-α, PPAR-α, and PPAR-γ were designed, synthesized, optimized, and provided as preoptimized kits: TNF-α (Cat. No. Mm00443258m1; GenBank accession No. NM013693), PPAR-α (Cat. No. Mm00440939m1; GenBank accession No. NM011144), and PPAR-γ (Cat. No. Mm 00440945m1; GenBank accession No. NM011146). Primers for ribosomal RNA as internal controls were also provided as a preoptimized kit (Cat. No. Hs99999901). Using an ABI PRISM 7000 sequence detector (ABI, Foster City, CA), PCR amplifications were performed in volumes of 50 μL containing 100 ng complementary DNA template in PCR Master Mix (Roche, Nutley, NJ) according to the following program: 50°C for 2 minutes; 95°C for 10 minutes; 40 cycles at 95°C for 15 seconds and 60°C for 1 minute. Samples were analyzed in duplicate. Results were analyzed with Sequence Detection Software (ABI), and the expression levels of TNF-α, PPAR-α, and PPAR-γ mRNAs were normalized to ribosomal RNA, as outlined in Perkin-Elmer's User Bulletin No. 2 (ABI).
Data are expressed as means ± SEM. Analysis of variance was used with Bonferroni's post hoc test for comparisons among the 3 groups. Student's t or Mann-Whitney test was used for comparisons between 2 groups where appropriate. Statistical significance for lethality rates was determined by Fisher exact test.
GalN/LPS or GalN/TNF-α-Induced Changes in Serum AST and ALT Levels and Lethality in KK-Ay Obese and Lean Control Mice.
Lethality was not observed in either lean or obese mice after PBS treatment. GalN/LPS-induced lethality was increased in KK-Ay obese animals as compared with lean controls (P < .05; Fig. 1A). In addition, the changes in serum AST (lean control, 1,245 ± 252 IU/L; Ay obese, 2,960 ± 474 IU/L) and ALT (lean control, 815 ± 82 IU/L; Ay obese, 1,912 ± 258 IU/L) levels by GalN/LPS treatment were more severe in KK-Ay obese animals than in lean controls (P < .01 for each; Fig. 1B). GalN/TNF-α-induced changes in serum AST (lean control, 1,150 ± 138 IU/L; Ay obese, 1,955 ± 291 IU/L) and ALT (lean control, 935 ± 187 IU/L; Ay obese, 1,675 ± 228 IU/L) were also increased in KK-Ay obese animals compared with lean controls (P < .05 for each).
GalN/LPS or GalN/TNF-α-Induced Hepatic Apoptosis and Necrosis in KK-Ay Obese and Lean Control Mice.
Figures 1C and 1D demonstrate that the apoptosis and necrosis of hepatocytes were induced by GalN/LPS treatment. TUNEL staining indicated that hepatic apoptosis 6 hours after GalN/LPS-treatment was more severe in KK-Ay obese animals than in lean controls (lean control, 7.6 ± 1.4; Ay obese, 22.5 ± 4.0; P < .01; Fig. 1D). Focal necrosis induced by accumulated macrophages and neutrophils and massive necrosis were also increased in KK-Ay obese mice compared with lean controls after GalN/LPS treatment (Fig. 1C). Similar results were observed in GalN/TNF-α-induced hepatic apoptosis and necrosis (data not shown).
Adiponectin Concentrations in KK-Ay Obese Mice.
The concentrations of adiponectin in the serum and white adipose tissue were decreased by 32% and 26%, respectively, in KK-Ay obese mice compared with those in normal, lean controls (P < .05 for each).
Effects of Adiponectin on GalN/LPS-Induced Lethality and on the Changes in Serum AST and ALT Levels in KK-Ay Obese Mice.
The administration of GalN/LPS increased lethality in treated animals compared with PBS-treated controls (P < .01; Fig. 2A). GalN/LPS treatment also significantly (P < .01) increased serum levels of AST (control, 64 ± 6 IU/L; LPS, 1,835 ± 368 IU/L) and ALT (control, 76 ± 4 IU/L; LPS, 2,470 ± 408 IU/L) compared with PBS-treated controls after the treatment (Fig. 2B). Pretreatment with adiponectin (AD) significantly ameliorated lethality (42%; P < .05) and serum levels of AST (LPS, 1,835 ± 368 IU/L; LPS AD, 937 ± 96 IU/L) and ALT (LPS, 2,470 ± 408 IU/L; LPS AD, 1,127 ± 254 IU/L) compared with pretreatment with PBS (Figs. 2A and 2B).
Effects of Adiponectin on GalN/LPS-Induced Hepatic Apoptosis and Necrosis in KK-Ay Obese Mice.
Administration of GalN/LPS induced apoptosis and necrosis of hepatocytes (Fig. 2C). TUNEL staining indicated significant hepatic apoptosis 6 hours after GalN/LPS-treatment (control, 1.8 ± 1.0; LPS, 16.7 ± 1.7; P < .01; Fig. 2D). Focal necrosis induced by accumulated macrophages and neutrophils and massive necrosis were observed in HE-stained liver sections 24 hours after GalN/LPS treatment (Fig. 2C). Pretreatment with adiponectin attenuated the percentage of TUNEL-positive staining (LPS, 16.7 ± 1.7; LPS AD, 8.2 ± 1.2; P < .01; Fig. 2D) as well as focal and massive necrosis in the liver.
Effects of Adiponectin on GalN/LPS-Induced Changes in TNF-α Levels and PPAR-α and PPAR-γ mRNA Expression in the Liver or Serum.
As shown in Fig. 3, the administration of GalN/LPS induced increases in TNF-α levels in serum (control, 14 ± 4 pg/mL; LPS, 8,810 ± 824 pg/mL [1.5 hours after the treatment]) and liver (control, 4.7 ± 0.9 pg/mg; LPS, 222 ± 32 pg/mg [6 hours after the treatment]) compared with PBS-treated controls (P < .01). Conversely, GalN/LPS induced decreases in mRNA levels of PPAR-α (control, 100.0 ± 9.8%; LPS, 52.5 ± 4.9%) and PPAR-γ (control, 100.0 ± 12.7%; LPS, 67.0 ± 5.8%) in the liver compared with the levels in the liver of PBS-treated controls (P < .01 for each). Adiponectin pretreatment attenuated the GalN/LPS-induced increases in serum TNF-α protein (LPS, 8,810 ± 824 pg/mL; LPS AD, 2,650 ± 47 pg/mL), hepatic TNF-α protein (LPS, 222 ± 32 pg/mg; LPS AD, 83 ± 9 pg/mg), and hepatic TNF-α mRNA (control, 100.0 ± 10.1%; LPS, 1817.5 ± 154.0%; LPS AD, 1407.5 ± 102.5%) compared with the effects of PBS pretreatment (P < 0.05 or P < .01; Figs. 3A, B, and C). Pretreatment with adiponectin also attenuated the GalN/LPS-induced decrease in PPAR-α mRNA (LPS, 52.5 ± 4.9%; LPS AD 69.5 ± 4.3%; P < .05), but not the decrease in PPAR-γ mRNA (LPS, 67.0 ± 5.8%; LPS AD, 70.5 ± 4.4%; P > .1), in the liver compared with the effects of PBS pretreatment (Figs. 4A and B).
Effects of Adiponectin on LPS-Induced TNF-α Levels in Abdominal Macrophages of KK-Ay Obese Mice in vitro.
The administration of LPS induced increases in supernatant TNF-α levels of abdominal macrophages from KK-Ay obese mice compared with the levels of nontreated macrophages (P < .01; Fig. 3D). Adiponectin treatment attenuated the LPS-induced increases in supernatant TNF-α levels, compared with control levels in vitro, 2 hours after treatment (P < .01; Fig. 3D).
Effects of adiponectin on TNF-α-induced changes in serum AST and ALT levels and apoptotic and necrotic changes in hepatocytes of KK-Ay obese mice.
Administration of TNF-α produced increases in serum levels of AST (control, 59 ± 6 IU/L; TNF-α, 1,770 ± 257 IU/L) and ALT (control, 67 ± 8 IU/L; TNF-α, 2,127 ± 247 IU/L), accompanied by apoptosis and necrosis of hepatocytes (Figs. 5A and 5B). TUNEL staining indicated significant hepatic apoptosis 6 hours after GalN/TNF-α treatment (control, 1.5 ± 1.0 IU/L; TNF-α, 10.6 ± 1.6 IU/L; P < .01; Fig. 5C). In HE-stained liver, focal necrosis—induced by accumulated macrophages and neutrophils—and massive necrosis were observed 24 hours after GalN/TNF-α treatment (Fig. 5A). Pretreatment with adiponectin attenuated the serum levels of AST (TNF-α, 1,770 ± 257 IU/L; TNF-α AD, 1,277 ± 192 IU/L) and ALT (TNF-α, 2,127 ± 247 IU/L; TNF-α AD, 1,481 ± 127 IU/L; Fig. 5B). Pretreatment with adiponectin also attenuated the percentage of TUNEL-positive staining (TNF-α, 10.6 ± 1.6; TNF-α AD, 7.5 ± 0.6; Fig. 5C) as well as the incidences of focal and massive necrosis.
Adiponectin secreted from adipose tissue has been shown to have antidiabetic, antilipogenic, and antiatherogenic actions.7–11 The present study demonstrates the protective role of adiponectin against endotoxin-induced lethal liver injury. It is well known that obese animals—in addition to humans—are highly susceptible to inflammatory disease.12–14 Previous studies demonstrated that obese animal models had altered sensitivity to endotoxin.13, 14 In accordance with previous reports,13, 14 KK-Ay obese mice were also more sensitive to endotoxin in the present study. As shown in the present study, serum adiponectin levels are low in obese mice, indicating the involvement of hypoadiponectinemia in the development of insulin resistance and atherosclerosis. The present study also demonstrated that KK-Ay obese mice pretreated with adiponectin were partially protected from endotoxin-induced liver injury and lethality. In addition, it has been demonstrated that the extent of the inhibitory effect of adiponectin was higher in KK-Ay obese mice than in lean control mice (T.M. and H.Y., unpublished data, 2003). These observations suggested that, in addition to its effects in metabolic disorders, adiponectin may regulate inflammation processes such as endotoxin-induced liver injury in KK-Ay obese mice.
In the present study, the administration of GalN/LPS induced elevation of serum AST and ALT levels, and apoptosis and necrosis of hepatocytes accompanied by marked increases in serum and hepatic TNF-α. GalN/LPS-induced liver injury develops in a stepwise fashion. The first step is apoptosis, which is caused by the initial humoral factors, including inflammatory cytokines.17, 18, 21 The second step is focal necrosis, which is induced by accumulated polymorphonuclear cells and lymphomononuclear cells.17, 18, 21 Finally, massive necrosis is induced by intrahepatic macrophages and neutrophils. We demonstrated that pretreatment with adiponectin prevented the apoptotic changes in hepatocytes induced by GalN/LPS. In addition, adiponectin treatment attenuated the GalN/LPS-induced increases in serum and hepatic TNF-α, a potent inducer of hepatocyte apoptosis. These results suggest that adiponectin may ameliorate the development of GalN/LPS-induced liver injury by affecting an early process, such as TNF-α-induced hepatocyte apoptosis. In the present study, abdominal macrophages from KK-Ay obese mice pretreated with adiponectin in vitro exhibited significantly decreased LPS-induced TNF-α production compared with controls. The result indicated direct effects of adiponectin on LPS-induced toxicity.
TNF-α has been implicated in the lethal liver injury induced by GalN/LPS.19 The administration of TNF-α has been shown to accelerate hepatic injury.21 In the present study, adiponectin pretreatment attenuated the TNF-α response to GalN/LPS and effectively attenuated TNF-α-induced liver injury. These results suggest that adiponectin may ameliorate GalN/LPS-induced liver injury by inhibiting the production and/or action of TNF-α. However, adiponectin appears to be more beneficial in LPS-induced hepatotoxicity than in TNF-α-induced hepatotoxicity because the magnitude of adiponectin's protective effects were more pronounced in LPS-induced hepatotoxicity. As part of the intracellular signaling mechanism of this phenomenon, adiponectin was demonstrated to suppress TNF-α-induced activation of nuclear factor-κ B (NF-κB).11 As the TNF-α/NF-κB signaling pathway plays a critical role in the inflammation process, further studies are needed to examine the effects of adiponectin on intracellular signaling, including the regulation of NF-κB.
PPAR-α, a nuclear receptor, is a likely candidate to explain the protective action of adiponectin against liver injury. PPAR-α is localized predominantly in hepatocytes and plays an important role as a target for adiponectin in lipid metabolism.6 The administration of adiponectin has been shown to increase PPAR-α ligand activity.6 The pathogenic significance of PPAR-α in inflammation and atherosclerosis has been proven.22, 24, 25 PPAR suppressed apoptosis in hepatocytes via PPAR-α.22, 23, 25 The administration of a PPAR-α agonist was shown to impair the production of TNF-α.22 Pretreatment of myocytes with PPAR-α activators decreased LPS-induced expression of TNF-α in the medium and TNF-α mRNA in myocytes.26 Conversely, PPAR-α-null mice were shown to be more sensitive to the effects of LPS.27 Thus, PPAR-α regulates the inflammation process by affecting the production of cytokines, such as TNF-α, and the actions of cytokine signaling pathways. In the present study, the low level of PPAR-α mRNA after treatment with GalN/LPS was partially restored by pretreatment with adiponectin, suggesting that adiponectin may prevent LPS-induced hepatic injury by increasing PPAR-α mRNA in the liver. In any case, the mechanisms for activating intracellular PPAR-α signaling by adiponectin needs further clarification and merits further study.
In summary, the present study demonstrated that adiponectin administration ameliorates liver injury and the high lethality induced by either GalN/LPS or TNF-α. The ability of adiponectin to inhibit inflammatory TNF-α production and its action may indicate a protective role of adiponectin against the development of severe inflammatory diseases. The present study also suggests that adiponectin may be a useful therapeutic tool for endotoxin-induced liver injury.