The nuclear receptor peroxisome proliferator-activated receptor-γ (PPAR-γ) is anti-inflammatory in a cell-based system and in animal models of endotoxemia. We have shown that PPAR-γ gene expression is downregulated in macrophages after lipopolysaccharide (LPS) stimulation. However, it remains unknown whether hepatic PPAR-γ is altered in sepsis and, if so, whether LPS directly downregulates PPAR-γ. To study this, rats were subjected to sepsis by cecal ligation and puncture (CLP). Hepatic tissues were harvested at 5, 10, and 20 h after CLP. PPAR-γ gene expression and protein levels were determined by RT-PCR and Western blot analysis, respectively. The results showed that PPAR-γ gene expression decreased at 10 and 20 h and that its proteins levels were reduced at 20 h after CLP. PPAR-γ levels were also decreased in animals that were administered LPS. To determine the direct effects of LPS on PPAR-γ downregulation, LPS binding agent polymyxin B (PMB) was administered intramuscularly after CLP. The administration of PMB significantly reduced plasma levels of endotoxin, but it did not prevent the downregulation of PPAR-γ expression. We found that circulating levels of TNF-α still remained significantly elevated in PMB-treated septic animals. We, therefore, hypothesize that the decrease of PPAR-γ expression is TNF-α dependent. To investigate this, Kupffer cells (KCs) were isolated from normal rats and stimulated with LPS or TNF-α. TNF-α significantly attenuated PPAR-γ gene expression in KCs. Although LPS decreased PPAR-γ in KCs, the downregulatory effect of LPS was blocked by the addition of TNF-α-neutralizing antibodies. Furthermore, the administration of TNF-α-neutralizing antibodies to animals before the onset of sepsis prevented the downregulation of PPAR-γ in sepsis. We, therefore, conclude that LPS downregulates PPAR-γ expression during sepsis via an increase in TNF-α release.
- peroxisome proliferator-activated receptor-γ
- Kupffer cells
- tumor necrosis factor-α
- cecal ligation and puncture
sepsis and septic shock, caused by systemic immune responses to severe bacterial infection, often induce multiple organ failure and continue to be significant causes of mortality despite the advances made in the management of afflicted patients (8, 26). Previous studies have shown that hepatic dysfunction occurs early after the onset of sepsis (38, 39). Sepsis-induced hepatocellular dysfunction and injury are associated with upregulated proinflammatory cytokines. It is well known that macrophages are the main source of TNF-α when stimulated by lipopolysaccharide (LPS) or other bacterial toxins. It has been previously reported that hepatic residential macrophages Kupffer cells (KCs) can release a large amount of proinflammatory cytokines (such as TNF-α and IL-1β) in sepsis, and they play a major role in hepatocyte (HC) injury in sepsis (40). Based on the observations that the liver is affected early after the onset of sepsis and that KCs are an important source for proinflammatory cytokines, we investigated an anti-inflammatory protein peroxisome proliferator-activated receptor-γ (PPAR-γ) in the liver in this study.
PPARs are members of the nuclear receptor superfamily. PPAR-γ is not only a key regulator of adipocyte differentiation, lipid metabolism, glucose homeostasis, and cell proliferation, it also plays an important role in regulating immunoresponses during inflammation (6, 25, 34, 37). It has been reported that the activation of PPAR-γ by its ligands, such as 15-deoxy-Δ12,14-prostaglandin J2 or synthetic agonists, inhibits the production of proinflammatory cytokines from macrophages and reduces organ injury in experimental animals (1, 9, 31, 35, 37, 50). Although studies have shown that PPAR-γ gene expression is downregulated in macrophages after LPS stimulation (29), it remains unknown whether PPAR-γ in the liver is altered during sepsis. We are also interested to know whether LPS itself or the downstream mediator (e.g., TNF-α) affects PPAR-γ expression. In this study, we examined hepatic PPAR-γ expression in the rat model of polymicrobial sepsis. Because hepatic macrophage KC is an important player in proinflammatory cytokines release and hepatic injury in sepsis, KCs were isolated from septic animals and PPAR-γ levels were assessed in those animals. To determine whether LPS directly downregulates PPAR-γ or through a TNF-α-dependent pathway, LPS binding agent polymyxin B (PMB) and TNF-α-neutralizing antibodies were used in vivo in septic animals. In addition, the effects of LPS and TNF-α on PPAR-γ expression were further studied in KC cultures. Thus the primary objective of this study was to determine whether PPAR-γ is reduced during sepsis and, if so, whether its downregulation is mediated via endotoxin or proinflammatory cytokine TNF-α.
MATERIALS AND METHODS
Polymicrobial sepsis model.
Male adult Sprague-Dawley rats (280–320 g), purchased from Charles River Laboratories (Wilmington, MA), were used in this study. All surgery was performed using aseptic procedures with the exception of the induction of sepsis by cecal ligation and puncture (CLP). Polymicrobial sepsis was induced by CLP as previously described (5, 42). Briefly, rats were fasted overnight before the induction of sepsis, although they were allowed water ad libitum. The animals were anesthetized with isoflurane inhalation, and a 2-cm ventral midline abdominal incision was made. The cecum was then exposed, ligated just distal to the ileocecal valve to avoid intestinal obstruction, and punctured twice with an 18-gauge needle. The punctured cecum was squeezed to expel a small amount of fecal material and then returned to the abdominal cavity. The incision was closed in layers, and the animals were resuscitated by 3 ml/100 g body wt normal saline subcutaneously immediately after CLP. Sham-operated animals underwent the same surgical procedure except that the cecum was neither ligated nor punctured. Studies were then conducted at 5, 10 (i.e., early sepsis), and 20 h (hypodynamic phase) after the induction of sepsis or sham operation. This model of sepsis is associated with a survival rate of >90% at 24 h, but ∼5% at 48 h after CLP without cecal excision (44).
The experiments described here adhered to the National Institutes of Health guidelines for the use of experimental animals. This project was approved by the Animal Care and Use Committee of The Feinstein Institute for Medical Research.
Intraperitoneal administration of LPS.
As previously reported (48), the animals were anesthetized with isoflurane inhalation, and a 2-cm midline incision was performed. An Alzet micro-osmotic pump (200 μl; Durect, Cupertino, CA) containing 2 ng/μl Escherichia coli LPS (055:B5; Sigma, St. Louis, MO) or sterile normal saline was implanted in the abdominal cavity followed by closure of the abdominal wall in layers. All rats received normal saline (3 ml/100 g body wt) subcutaneously immediately after the pump implantation to provide fluid resuscitation. The infusion rate of the pump was 8 μl/h for a duration of 20 h. Each rat received ∼1 μg/kg body wt LPS.
Administration of PMB and measurement of endotoxin and TNF-α.
In additional groups of septic animals, at 0.5 h before CLP and 9 h after CLP, PMB (Sigma) was administered intramuscularly at a total dose of 4,000 U/kg (2,000 U/kg each injection). This dosage was determined according to our previous experiment (48). Blood and hepatic tissues were collected at 20 h after CLP or sham operation (5 or 6 rats/group). PMB is an LPS binding antibiotic, which binds and detoxifies lipid A. In vitro studies have confirmed that PMB neutralizes E. coli LPS activity (21). Plasma levels of endotoxin were determined by using a Limulus amebocyte lysate kit (Associates of Cape Cod, Falmouth, MA), and plasma levels of TNF-α were assessed by using an ELISA kit (BD Pharmingen, San Diego, CA). The assays were performed according to the instructions provided by the manufacturer.
Administration of TNF-α-neutralizing antibodies.
TNF-α-neutralizing antibodies or control IgG (R&D Systems, Minneapolis, MN) were administrated intravenously at a dose of 1 mg/kg immediately after the onset of sepsis. Hepatic tissues were collected at 20 h after CLP (4 rats/group). RNA and protein were isolated from these samples as described below to determine PPAR-γ expression.
KC and HC isolation.
KCs and HCs were isolated from normal, CLP, and sham-operated animals by collagenase perfusion of liver, isopycnic sedimentation in a Percoll gradient, and selective adherence as previously reported (45, 46) with some modifications. Under isoflurane anesthesia, the inferior vena cava was cannulated after a midline incision was made, and the portal vein was severed. The liver was immediately perfused in situ with ∼60 ml of warm (37°C) HBSS (without Ca2+ and Mg2+) at a rate of 15 ml/min. This was followed by perfusion of 120 ml of HBSS with 0.5 mM CaCl2 containing 0.02% collagenase (type IV, 160 U/mg; Worthington, Lakewood, NJ) at the same perfusion rate. The liver was then removed en bloc, rinsed with HBSS, minced in a Petri dish containing 0.02% collagenase solution, and incubated for 20 min at 37°C to further dissociate the cells. The cell suspension was then passed through a sterile 150-mesh sieve and centrifuged (50 g for 2 min at 4°C) to sediment HCs. KCs remained in the supernatant of the cell suspension. The pellets containing HCs were washed with Williams medium E (Invitrogen, Carlsbad, CA), and the cell suspension was layered over 50% Percoll. HCs were collected as pellets in Percoll solution after centrifugation at 400 g for 15 min. The supernatant containing KCs was centrifuged at 450 g for 15 min. The cell pellets were resuspended in DMEM with 10% heat-inactivated FBS and gently layered on top of a two-step Percoll gradient (consisting of 25% Percoll on the top and 50% Percoll at the bottom). The gradients were centrifuged at 1,000 g for 20 min at 4°C. The KC fraction in the layer of 50% Percoll was collected. The viability of KCs was determined by Trypan blue exclusion, which was >95%. Approximately 90% nonparenchymal liver cells ingested India ink and more than 80% of those cells were positive on peroxidase staining. With the use of the purification procedure described above, the purity of KCs was found to be at least 90% with a yield of KCs of ∼10 × 106/liver. The isolated KCs respond to LPS and release proinflammatory cytokines in the culture (46). HCs isolated from rats were purified to >95% purity by repeated centrifugation at 50 g followed by further purification over 50% Percoll as described above. The viability of HCs was >90%. The isolated and purified HCs expressed liver enzyme aminotransferase and hepatic proteins such as cytochrome P-450 in the culture (45). The isolation procedure took ∼1.5 h for KCs and <1 h for HCs.
Stimulation of the isolated KCs and HCs.
The isolated KCs and HCs were cultured individually or cocultured in DMEM or Williams medium E containing 10% heat-inactivated FBS, 10 mM HEPES, 100 U/ml penicillin, and 100 μg/ml streptomycin. The isolated KCs were allowed to adhere to the bottom of the culture plate overnight before coculture with HC or stimulation. Both KCs and HCs were cultured at a density of 1 × 106/ml in 24-well cell culture plates. The HC and KC cocultures consisted of 0.5 × 106 attached KCs in 24-well cell culture plates with addition of 0.5 × 106 HCs. KCs or HCs were stimulated by LPS (10 and 100 ng/ml) or TNF-α (10 ng/ml) for 20 h. In an additional experiment, goat TNF-α-neutralizing antibody (10 μg/ml; R&D Systems, Minneapolis, MN), which neutralizes the bioactivity of TNF-α, or nonspecific goat IgG (10 μg/ml; R&D Systems) was added to KC culture treated with 100 ng/ml LPS. In the coculture system, HCs were collected by shaking and gentle blowing with a pipette to free the HCs (HCs did not attach to the plate well and could be shaken off). The cell lysates were collected for Western blot analysis.
PPAR-γ and TNF-α gene expression.
Total RNA was extracted from hepatic tissues, KCs, or HCs by using TRIzol reagent (Invitrogen), and 4 μg of RNA were reverse transcribed as previously described (43). The resulting cDNAs were amplified by PCR using primers for rat PPAR-γ (accession no. AB011365; forward: CGG TTG ATT TCT CCA GCA TT; reverse: AGC AAG GCA CTT CTG AAA CC) and rat TNF-α (accession no. L00981; forward: CCC AGA CCC TCA CAC TCA GA; reverse: GCC ACT ACT TCA GCA TCT CG). The PCR reaction was conducted at 35 cycles. For PPAR-γ, each cycle consisted of 30 s at 94°C, 30 s at 58°C, and 1 min at 72°C. For TNF-α, each cycle consisted of 45 s at 94°C, 45 s at 60°C, and 2 min at 72°C. Rat GAPDH (accession no. M17701; forward: TGA AGG TCG GTG TCA ACG GAT TTG GC; reverse: CAT GTA GGC CAT GAG GTC CAC CAC) served as a housekeeping gene. PCR cycles used above were carefully tested to avoid the plateau phase of amplification. After the RT-PCR procedure, the reaction products were separated in 1.6% Tris-borate-EDTA-agarose gel containing 0.22 μg/ml ethidium bromide by electrophoresis. The gel was then photographed, and the band density was determined by a digital image system (Bio-Rad, Hercules, CA).
PPAR-γ Western blot analysis.
Hepatic tissues collected from animals at 10 and 20 h after CLP or sham operation (5 or 6 rats/group) were homogenized in a lysis buffer, containing 10 mM Tris·HCl, pH 7.5 with 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 2 mM sodium orthovanadate, 0.2 mM PMSF, 2 μg/ml leupeptin, and 2 μg/ml aprotinin. After centrifugation at 16,000 g for 10 min, the supernatant was collected and the protein concentration was determined by using the Bio-Rad DC protein assay kit. Similarly, KCs and HCs were disrupted in the above lysis buffer to release protein, and the protein concentration was measured. Protein (50 μg) from hepatic tissue or protein (10 μg) from KCs or HCs was loaded on 4–12% Bis-Tris gel (Invitrogen) and electrophoretically fractionated in MOPS-SDS running buffer (Invitrogen). The protein on the gel was then transferred to a 0.45-μm nitrocellulose membrane and blocked with 5% nonfat dry milk in 10 mM Tris·HCl with 0.1% Tween 20, pH 7.5. The membrane was incubated with rabbit anti-PPAR-γ polyclonal antibody (1:1,000; Cayman, Ann Arbor, MI) overnight at 4°C followed by incubation in 1:20,000 horseradish peroxidase-linked anti-rabbit IgG for 1 h at room temperature. Mouse anti-β-actin monoclonal antibody (1:20,000; Sigma) was used as the loading control in this experiment. To reveal the reaction bands, the membrane was reacted with enhanced chemiluminescence Western blot detection system (Amersham, Piscataway, NJ) and exposed on X-ray film. The Bio-Rad GS-800 calibrated densitometer analysis system was used to quantitate the Western blots. This system can select the contour of the band, subtract the background, and calculate the density.
Hepatic tissues collected from animals at 20 h after CLP or sham operation (3 or 4 rats/group) were fixed in 10% buffered formalin and processed for paraffin sectioning with standard histology procedures. The paraffin sections were dewaxed and rehydrated; this was followed by a microwave antigen retrieval procedure as described below. Slides were soaked in 20% citric acid buffer, pH 6.0 (Vector Laboratories, Burlingame, CA), and heated in the microwave oven; temperature was maintained at 95°C for 15 min. The slides were cooled in room temperature for 5 min and then rinsed with Tris-buffered saline (pH 7.6). Endogenous peroxidase was blocked by 2% H2O2 in 60% methanol for 20 min. Normal goat serum (3%) was used to block the nonspecific binding sites. The sections were then incubated in 1:50 rabbit anti-PPAR-γ polyclonal antibody (Cayman) for 2.5 h at room temperature. After sections were washed with Tris-buffered saline, they were reacted in 1:200 biotinylated anti-rabbit IgG (Vector Laboratories) for 1 h. Vectastain ABC reagent and diaminobenzidine kit (Vector Laboratories) were used to reveal the immunohistochemical reaction. For the negative control, the primary antibody was substituted by normal rabbit IgG.
All data are expressed as means ± SE. The statistical analysis methods are one-way or two-way ANOVA with Student-Newman-Keuls test. Student's t-test was also used to analyze the data. Differences in values were considered significant if P < 0.05.
Alterations of PPAR-γ expression in sepsis.
As shown in Fig. 1, although PPAR-γ gene expression in the liver decreased by 19.9% at 5 h after CLP, such a reduction was not statistically different from sham-operated animals. In contrast, hepatic PPAR-γ gene expression was downregulated by 50.1% at 10 h and by 69.8% at 20 h after CLP, respectively (P < 0.05; Fig. 1). In addition, PPAR-γ protein levels in hepatic tissues did not change at 10 h but were significantly reduced at 20 h after CLP (P < 0.05; Fig. 2). Thus hepatic PPAR-γ gene and protein expressions are downregulated in sepsis.
On the other hand, KCs and HCs were isolated from animals at 20 h after CLP or sham operation. As indicated in Table 1, PPAR-γ gene expression decreased significantly in KCs at 20 h after CLP (P < 0.05), whereas PPAR-γ expression in HCs was not altered. Immunohistochemistry method showed that total PPAR-γ staining did not appear to be altered in HCs at 20 h after CLP (Fig. 3). These data suggest that KCs contribute mainly to the decrease of hepatic PPAR-γ at 20 h after CLP.
Effects of LPS on PPAR-γ expression in sepsis.
Similar to that after CLP, infusion of low-dose LPS significantly decreased hepatic PPAR-γ protein levels (Fig. 4). Such a decrease is associated with an increase in hepatic TNF-α gene expression (TNF-α-to-GAPDH ratio: 0.31 ± 0.03 in vehicle vs. 0.48 ± 0.04 in LPS rats; P < 0.05; n = 4 or 5 animals). To determine whether LPS directly downregulates PPAR-γ expression in sepsis, LPS-binding agent PMB was given 0.5 h before and 9 h after the onset of sepsis. Although PMB significantly reduced the plasma levels of endotoxin (from 0.64 to 0.24 EU/ml) in the animals at 20 h after CLP (P < 0.05; Fig. 5A), it did not prevent the downregulation of hepatic PPAR-γ gene expression in septic animals (Fig. 5B). Thus it does not appear that LPS directly downregulates hepatic PPAR-γ gene expression in sepsis. On the other hand, we observed that, although TNF-α gene expression in hepatic tissues was slightly decreased by PMB treatment (no statistical difference; Fig. 6A), plasma levels of TNF-α were still significantly elevated in these animals (Fig. 6B), suggesting that the proinflammatory cytokine TNF-α may play an important role in downregulating hepatic PPAR-γ expression in sepsis.
Effects of LPS and TNF-α on PPAR-γ expression in isolated KCs and HCs.
KCs and HCs from normal rats were stimulated with LPS (10 and 100 ng/ml) or TNF-α (10 ng/ml) for 20 h. As indicated in Fig. 7A, KC PPAR-γ protein decreased by 56.8% with 10 ng/ml LPS and by 62.2% with 100 ng/ml LPS (P < 0.05). Moreover, TNF-α (10 ng/ml) reduced KC PPAR-γ protein by 69.9% (P < 0.05; Fig. 7B). In contrast, LPS did not affect PPAR-γ protein levels in HCs either in monoculture (Fig. 8A) or in coculture with KCs (Fig. 8B). Similarly, TNF-α (10 ng/ml) did not alter PPAR-γ protein in HCs (data not shown). The above results suggest that the decrease of hepatic PPAR-γ in sepsis is primarily attributed to KCs and that TNF-α plays a major role in such a decrease.
To further confirm the effect of TNF-α on PPAR-γ expression, goat TNF-α-neutralizing antibody (10 μg/ml) was used with LPS (100 ng/ml) in KC culture. As shown in Fig. 9, TNF-α-neutralizing antibody completely blocked the downregulating effect of LPS on PPAR-γ protein expression. In contrast, nonspecific goat IgG did not significantly affect PPAR-γ protein expression. Thus LPS decreases PPAR-γ via the increase in proinflammatory cytokine TNF-α release.
Effects of TNF-α-neutralizing antibody on PPAR-γ expression in septic animals.
TNF-α-neutralizing antibody was administrated in the animals at the same time of CLP procedure, and hepatic tissue samples were collected at 20 h after CLP. PPAR-γ gene expression and protein levels were measured in those animals. The results showed that PPAR-γ gene expression was significantly downregulated in CLP animals after nonspecific IgG treatment, but the decrease was almost prevented after TNF-α-neutralizing antibody treatment (Fig. 10A). Similar results were also obtained in PPAR-γ protein levels in those animals (Fig. 10B).
The cardiovascular response to polymicrobial sepsis is characterized by an early, hyperdynamic phase (i.e., 2–10 h after CLP) followed by a late, hypodynamic phase (i.e., ≥16 h after CLP) (41, 47). The present study demonstrates that hepatic PPAR-γ gene expression was significantly downregulated at the early stage of sepsis (10 h after the onset of sepsis), and its protein levels decreased significantly at the hypodynamic phase of sepsis (20 h after CLP). The finding that the decrease of PPAR-γ gene occurred earlier than that of the protein levels suggests that the decreased levels of PPAR-γ protein in sepsis are caused by the downregulation of its gene expression. One would expect a delay in reduction of protein synthesis after the downregulation of the gene expression. It has been well established that macrophages are the primary source of TNF-α release under various inflammatory conditions. Our data indicate that the increase in TNF-α can attenuate PPAR-γ expression. Because PPAR-γ is an anti-inflammatory protein, its downregulation could further enhance the production of proinflammatory cytokines. It is well known that tissue macrophages such as KCs can release proinflammatory cytokines (e.g., TNF-α, IL-1β) and play a major role in tissue injury in sepsis (40). It is most likely that the downregulation of PPAR-γ and the increased proinflammatory cytokine release from KCs contribute to hepatic injury in sepsis.
PPARs are members of the nuclear receptor superfamily. They can heterodimerize with retinoid X receptor to regulate transcription of genes linked to lipid and glucose metabolism. The first PPAR cloned from the mouse liver was reported in 1990 (12), followed by the discovery of other PPAR homologs in other species (27). To date, three different PPARs (α, β, γ) have been identified (19). Nuclear receptor PPAR-γ was originally identified as a key regulator of adipocyte differentiation and lipid metabolism. In addition to its role in adipogenesis, PPAR-γ serves as a transcriptional regulator involved in glucose and lipid metabolism (30). More recently, various leukocyte populations, including monocytes, macrophages, lymphocytes, and dendritic cells, have also been shown to express PPAR-γ, suggesting a role for this molecule in the regulation of immune responses (7). In vitro studies have shown that PPAR-γ gene expression is downregulated in macrophages after LPS stimulation (29). It has also been reported that PPAR-γ expression is markedly reduced in lungs and thoracic aortas after the onset of sepsis (50).
Our present results demonstrated that hepatic PPAR-γ was downregulated in sepsis. Studies have shown that deprivation of PPAR-γ gene in animals is fatal. Mice that are genetically deficient in PPAR-γ die in utero (30). Heterozygous PPAR-γ-deficient mice (i.e., reduced PPAR-γ activity relative to wild-type animals) show greater intestinal injury than wild-type animals after gut ischemia-reperfusion (22), indicating the protective role of PPAR-γ in the injury. Activation of PPAR-γ by its ligands or agonists has shown to be anti-inflammatory in animal models of endotoxemia and sepsis (10, 17, 23, 29). Our recent study (29) showed that the anti-inflammatory property of phytochemical curcumin was mediated by upregulation of PPAR-γ. Administration of curcumin before the onset of sepsis attenuated tissue injury-reduced mortality and decreased the expression of TNF-α in septic animals. Treatment of mice with synthetic PPAR-γ agonists has also been shown to decrease mortality in an animal model of endotoxemia (23). In addition, endogenous ligands of PPAR-γ such as 15-deoxy-Δ12,14-prostaglandin J2 also reduce proinflammatory cytokine production in macrophages and prevent organ injury in a variety of experimental animal models of diseases (1, 9, 31). PPAR-γ agonists attenuate the inflammatory process of experimental colitis and adjuvant-induced arthritis (9, 31, 35). However, the above studies indicate that PPAR-γ ligands or agonists need to be given before or at an early stage of the injury to achieve their protective effects. For example, our recent study (29) showed that curcumin needs to be administrated to animals before or at 5 h after sepsis to prevent organ injury. Treatment at 10 h after CLP by curcumin cannot achieve protection similar to the treatment given at earlier times in sepsis (unpublished observations). This phenomenon can be explained by our present results. Because downregulation of PPAR-γ gene expression occurred at 10 h after CLP, the treatment provided before this time point is certainly more effective. Thus activation of PPAR-γ is an important therapeutic approach in the treatment of sepsis and other inflammatory diseases. The treatment should be given as early as possible after the injury to achieve the optimal protective effects.
PMB is an LPS-neutralizing agent and has been used in cases of sepsis and/or endotoxic shock. Several studies showed that PMB reduces endotoxin levels in sepsis (13, 36). Although our present results also indicated that injection of PMB significantly reduced the plasma levels of endotoxin and partially reduced hepatic TNF-α gene expression, the circulating levels of TNF-α remained elevated. In polymicrobial sepsis, target cells are exposed to a variety of microorganisms, including Gram-negative and -positive bacteria, their metabolites, and active compounds. Most of these pathogens have cytotoxicity and inflammatory potentials. Although PMB neutralizes the major proinflammatory component of Gram-negative bacterial LPS, Gram-positive bacteria and other bacterial components such as lipoproteins can induce TNF-α production. In addition, the residual of LPS may also trigger the production of proinflammatory cytokines in the animals. Tani et al. (33) reported interesting results regarding endotoxin-adsorbing therapy using PMB. They found that PMB reduced cytokine levels in septic patients who survived, but proinflammatory cytokine levels in the nonsurvivor were not significantly reduced, although plasma endotoxin levels were decreased by PMB. Ayala et al. (3) reported that, although endotoxin-tolerant C3H/HeJ mice were nonresponsive to endotoxin as a stimulant for cytokine release in vivo and in vitro, laparotomy injury increased KC and splenic macrophage responsiveness to endotoxin and promoted the release of TNF-α and other cytokines. Such a response is probably due to the traumatic injury that provided sufficient stimulus directly and/or indirectly through the release of agents such as interferon-γ, IL-2, IL-4, and granulocyte-monocyte colony-stimulating factor to prime macrophages to release inflammatory mediators (3). This phenomenon was also observed in hemorrhage injury because hemorrhage can induce an increase in circulating levels of TNF-α, which is not associated with elevated levels of endotoxin (2). In this regard, our recent studies have indicated that the increase of endogenous norepinephrine in sepsis can increase TNF-α production because norepinephrine mainly activates α2-adrenergic receptor at more pathophysiological concentrations (20 nM or lower) in sepsis (49). Moreover, pathophysiological levels of norepinephrine not only upregulate TNF-α expression and release from macrophages but also potentiate LPS-induced TNF-α and IL-6 release. Together, these results suggest that traumatic injury and the induction of mediators from the injury also contribute to the increase of TNF-α levels observed in sepsis.
Although studies have shown that PPAR-γ gene expression is downregulated in endotoxemia and sepsis (29, 50), the mechanism responsible for such downregulation has not been fully studied. Our present results indicated that the downregulation of PPAR-γ in sepsis is most likely due to the direct effect of TNF-α. Because both PPAR-γ mRNA and protein expression are suppressed in sepsis, it is possible that TNF-α downregulates PPAR-γ expression at the transcriptional level. This result has also been previously supported (15). TNF-α downregulated adipocyte PPAR-γ by suppressing its gene transcription, and such a suppression was mediated via CCAAT/enhancer-binding protein, since TNF-α significantly inhibited CCAAT/enhancer-binding protein expression and its DNA binding at the promoter region (15). Liu et al. (16) showed that reduction of pulmonary PPAR-γ correlated with the increased TNF-α in plasma and lungs in the model of LPS-induced lung injury. The suppression of PPAR-γ is also associated with activation of NF-κB (16). Activation of PPAR-γ by its ligands has been shown to inhibit NF-κB and activator protein-1 pathways and suppress proinflammatory mediator release (30, 50). In line with these findings, TNF-α-induced decrease of hepatic PPAR-γ is at the transcriptional level and likely mediated by the activation of NF-κB pathway and CCAAT/enhancer-binding protein. It should be pointed out that, although TNF-α plays a critical role in downregulating PPAR-γ expression in sepsis, other inflammatory cytokines or mediators may also directly or indirectly interact with TNF-α and thereby contribute to such suppression.
PPAR-γ is present in a variety of different cell populations. The finding that the spleen contained relatively high expression of PPAR-γ mRNA (4) leads to the discovery that monocytes and especially elicited macrophages contain PPAR-γ (4, 24). PPAR-γ is expressed in the adipose tissues, heart, lungs, thoracic aorta, vascular endothelial cells, smooth muscle cells, mesangial cells, and atherosclerotic lesions (4, 11, 32, 50). Our results indicated that PPAR-γ is expressed in the liver and that its expression is found in both KCs and HCs. However, these two cell populations respond to LPS and TNF-α differentially. Both LPS and TNF-α decreased PPAR-γ in KCs but not in HCs. Although HCs have been reported to directly respond to LPS to produce protective acute-phase proteins through the Toll-like receptor 4 signal transduction pathway (20), increased TNF-α induces apoptotic signals and produces hepatic injury in sepsis and other disease conditions (14, 18). Because studies have shown that cell-to-cell interaction is required to enhance the effect of LPS (28), we cultured HCs alone or cocultured with KCs to determine whether KC-HC interaction or mediators produced by KCs affect PPAR-γ expression in HCs. Our results showed that both HC monoculture and coculture with KCs did not change PPAR-γ expression in HCs after LPS stimulation. Thus the downregulated PPAR-γ expression observed in the liver after polymicrobial sepsis is most likely due to the decreased PPAR-γ in KCs but not in HCs.
Perspectives and Significance
The downregulation of hepatic PPAR-γ expression in sepsis cannot be prevented by endotoxin neutralization. Administration of anti-TNF-α antibodies, however, prevents the decrease of PPAR-γ. Moreover, endotoxin-induced reduction of PPAR-γ can be restored by TNF-α-neutralizing antibodies. Thus proinflammatory cytokine TNF-α plays a direct role in the downregulation of PPAR-γ in sepsis. The systemic inflammatory response is a major component in the pathogenesis of sepsis and septic shock syndrome. Recent studies have focused on the modulation of the immunologic response in sepsis. Understanding the intracellular mechanisms of such responses is important for the development of specific antisepsis therapies. To this end, the finding that downregulation of PPAR-γ in sepsis is directly mediated by TNF-α would provide new information for antisepsis therapy research and development. In addition, the activation of PPAR-γ by its ligands or agonists in combination with anticytokine therapy could be more effective in the treatment of sepsis.
This study was supported by the National Institutes of Health Grants R01 GM-053008 and R01 GM-057468 (P. Wang).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society