Objective: In order to characterize the regulation of resistin gene expression, we explore the effect of tumornecrosis factor-α (TNF-α) on resistin mRNA expression and its underlying mechanism in 3T3-L1 adipocytes.
Methods and Procedures: Differentiated 3T3-L1 adipocytes were treated for 24 h with 0–10 ng/ml of TNF-α or with 2.5 ng/ml of TNF-α for 0–24 h, and then resistin mRNA levels were measured by northern blotting. To further explore the involvement of nitric oxide (NO) in TNF-α–regulated resistin expression, the effect of the NO donor, sodium nitroprusside (SNP), on resistin mRNA levels in adipocytes and the effect of the nitric oxide synthase (NOS) inhibitors, NG-nitro-l-arginine methyl ester (l-NAME), and S, S′−1,3-phenylene-bis(1,2-ethanediyl)-bis-isothiourea·2HBr (PBITU), on the TNF-α effect in adipocytes were examined. The effects of TNF-α on inducible NOS (iNOS) protein expression in adipocytes were also measured by western blotting.
Results: Our results showed that TNF-α caused a dose-dependent reduction in resistin mRNA levels. This effect seemed to be associated with the TNF-α–induced expression of iNOS. The results showed that TNF-α induced iNOS expression and release of NO after 24-h treatment of differentiated 3T3-L1 adipocytes. Pretreatment with l-NAME and PBITU significantly reversed the TNF-α–induced downregulation of resistin expression, while treatment with SNP mimicked the inhibitory effect of TNF-α on resistin expression. In addition, pretreatment with protein tyrosine kinase (PTK) inhibitors, genistein and AG-1288, prevented TNF-α–induced iNOS expression and subsequent resistin downregulation.
Discussion: Our data suggest that TNF-α suppresses resistin expression by inducing iNOS expression, thus causing overproduction of NO, which downregulates resistin gene expression.
The discovery of fat-derived bioactive substances has had a great effect on adipose tissue research and has suggested an additional role of adipose tissue as an endocrine organ. Adipocytes secrete many bioactive substances, including leptin, plasminogen activator inhibitor-1, and free fatty acid (1). Obesity is an important risk factor for the development of insulin resistance, diabetes, and atherosclerosis (2), and recent studies have suggested that adipose tissue is an important endocrine organ secreting adipocytokines (3), which are thought to play a pathological role in the development of metabolic syndrome (2,4). However, the physiological functions of most fat-derived factors and how their expression is regulated have not been well documented.
Tumor necrosis factor-α (TNF-α) was initially shown to be produced predominantly by macrophages and to function as an inducible cytokine with a wide range of important proinflammatory and immunological actions (5). However, it is also expressed by adipose tissue and skeletal muscle and is highly expressed in human obesity (6,7). TNF-α mRNA levels in adipose tissue and skeletal muscle show a very good correlation with percent body fat, BMI, and levels of hyperinsulinemia (8,9). Increased concentrations of TNF-α might therefore play an important role in obesity-induced insulin resistance in humans. In animal studies, increased TNF-α mRNA levels in adipose tissue were observed in obese fa/fa rats, and neutralization of TNF-α in these obese rats caused a significant increase in peripheral uptake of glucose in response to insulin (10). In an in vitro study, TNF-α was shown to inhibit insulin-stimulated glucose uptake in adipocytes by decreasing phosphorylation of the insulin receptor (11). Treatment of cultured adipocytes with TNF-α induces serine phosphorylation of insulin receptor substrate-1 and converts it into an inhibitor of insulin receptor tyrosine kinase activity (12). Furthermore, Stephens and Pekala demonstrated that TNF-α inhibited the expression of GLUT4 protein, an insulin-responsive glucose transporter, and caused insulin resistance in 3T3-Ll adipocytes (13). The study by Ruan et al. also showed that 24-h exposure of 3T3-L1 adipocytes to TNF-α resulted in reduced protein levels of GLUT4 and several insulin signaling proteins, including the insulin receptor, insulin receptor substrate-1, and acutely transforming retrovirus AKT8 in rodent T cell lymphoma (14). Their findings further suggested that nuclear factor-κB played as an obligatory mediator in TNF-α–induced insulin resistance (14). These observations suggest that TNF-α may play a role in the pathogenesis of obesity-associated insulin resistance.
Resistin was identified as a fat-derived factor specifically secreted by adipose tissue (15). Resistin gene expression is significantly increased during the differentiation of 3T3-L1 cells and primary preadipocytes into adipocytes (15,16). Elevated serum resistin levels are seen in insulin-resistant and obese mice either fed a high-fat diet or with genetic susceptibility (15). Immunoneutralization of endogenous resistin significantly improves insulin sensitivity in diet-induced obese mice (15). In contrast, administration of recombinant resistin impairs glucose tolerance in C57B1/6J mice in vivo and suppresses insulin-stimulated glucose uptake in 3T3-L1 adipocytes in vitro (15). Moreover, resistin gene expression is downregulated in animals treated with the insulin-sensitizing agent, thiazolidinedione. In human studies, adipose resistin mRNA levels are increased in severely obese subjects when compared with lean controls (17). A genetic epidemiological study on white families indicated that single-nucleotide polymorphisms in the resistin gene promoter region may alter the interaction between the BMI and the insulin sensitivity index in nondiabetic family members of type 2 diabetics (18). Recent studies have revealed that recombinant resistin inhibits glucose uptake by the L6 skeletal muscle cell line (19) and acutely impairs glucose metabolism in isolated skeletal muscle (20). In terms of the long-term effects of resistin, impaired glucose metabolism is seen in resistin transgenic animals (20). We recently produced bioactive recombinant resistin in Escherichia coli which similarly suppressed insulin-stimulated glucose uptake in a dose-dependent manner (21). On the basis of these findings, resistin was suggested to be a significant modulator of insulin sensitivity.
In addition to its established role in the inflammatory and immune systems, TNF-α exerts complex regulatory actions on adipose tissue, including effects on adipocyte development, lipid metabolism, and energy expenditure (22,23). It has been shown to play a role in the regulation of some fat-derived bioactive factors. For example, it suppresses gene expression of adipocyte complement-related protein of 30 kDa (ACRP30) and induces gene expression of plasminogen activator inhibitor-1 (24). Resistin gene expression in response to TNF-α is controversial (25,26,27), and the precise underlying mechanism is still unknown. TNF-α causes increased inducible nitric oxide synthase (iNOS) expression and subsequent NO production in various tissues, including vascular smooth muscle cells (28) and adipose tissue (23,29,30). There is also evidence that protein tyrosine kinase (PTK) might be involved in TNF-α–induced iNOS expression in several cell types (28,31,32). The aim of this study was to explore the regulatory effect of TNF-α on resistin expression and the underlying mechanism in 3T3-L1 adipocytes. In addition, the role of cytokine-induced iNOS expression and NO production in TNF-α–regulated resistin expression was investigated.
Methods and Procedures
To explore the effects of TNF-α on resistin gene expression, differentiated 3T3-L1 adipocytes were treated for 24 h with 0–10 ng/ml of TNF-α or with 2.5 ng/ml of TNF-α for 0–24 h, and then resistin mRNA levels were measured by northern blotting. At the same time, nitrate/nitrite in the medium was measured to confirm the hypothesis that TNF-α induces NO production in 3T3-L1 adipocytes. To further clarify the role of NO in TNF-α–regulated resistin expression, the effect of the NO donor, sodium nitroprusside (SNP), on resistin mRNA levels in adipocytes and the effect of the general NOS inhibitor, NG-nitro-l-arginine methyl ester (l-NAME), and specific iNOS inhibitor, S, S′−1,3-phenylene-bis(1,2-ethanediyl)-bis-isothiourea·2HBr (PBITU), on the TNF-α effect in adipocytes were examined. We also investigated the relationship between TNF-α treatment and iNOS protein expression, measured by western blotting, and whether PTK activation was involved in TNF-α–mediated induction of iNOS protein expression using the PTK inhibitors, genistein and AG-1288.
3T3-L1 preadipocytes (American Type Culture Collection, Rockville, MD) were grown and maintained in Dulbecco's modified Eagle's high-glucose medium containing 50 units/ml penicillin, 50 μg/ml streptomycin (all from Gibco BRL, Gaithersburg, MD), and 10% fetal bovine serum (Biowest, Nuaillé, France) (growth medium) in a 10% CO2 environment. The cells were allowed to grow until 2 days postconfluency and were then induced to differentiate by adding growth medium containing isobutylmethylxanthine (500 μmol/l), dexamethasone (25 μmol/l), and insulin (4 μg/ml) for 3 days. The medium was then changed to growth medium containing insulin, which was replaced every 3 days until the cells were fully differentiated, typically after 6 days (33). Before each experiment, cells were placed for 6 h in serum-free medium (Dulbecco's modified Eagle's medium low-glucose medium (Gibco BRL, Gaithersburg, MD) containing 0.1% bovine serum albumin). This study protocol was used in all experiments.
Total RNA was extracted from treated 3T3-L1 adipocytes using a Tri Reagent Kit (Sigma-Aldrich, St. Louis, MO). The integrity of the extracted total RNA was examined by 1% agarose gel electrophoresis and its concentration determined by UV light absorbance at 260 nm (34).
Northern blot analysis
Northern blotting was performed as described previously (34). For hybridization, the resistin and glyceraldehyde-3-phosphate dehydrogenase cDNA probes were radiolabeled using a random primer labeling system (35) and reagents were purchased from Promega (Madison, WI). After hybridization, the membranes were washed and autoradiographed with an intensifying screen at −80 °C. The hybridization signals on the autoradiogram were expressed in arbitrary densitometric units relative to that of the corresponding glyceraldehyde-3-phosphate dehydrogenase mRNA band.
Determination of nitrite concentration
Nitrite in cell culture supernatants was determined using Griess reagent with sodium nitrite as the standard.
Western blot analysis
Whole cell lysates were prepared by sonication in lysis buffer (1% Triton X-100, 50 mmol/l KCl, 25 mmol/l (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), pH 7.8, 10 μg/ml leupeptin, 20 μg/ml aprotinin, 125 μmol/l dithiothreitol, 1 mmol/l phenylmethylsulfonyl fluoride, 1 mmol/l sodium orthovanadate). Samples (100 μg of total protein) in 50 μl of reducing sample buffer were denatured at 100 °C for 10 min and resolved by 7.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis for 90 min at 100 V and the proteins transferred to a polyvinylidene difluoride membrane at 21 V for 120 min. The membrane was then incubated for 30 min at room temperature with 5% skimmed milk in phosphate-buffered saline and immunoblotted with primary antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) at 4 °C overnight followed by horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) at room temperature for 1 h, both antibodies being diluted in 5% skimmed milk in phosphate-buffered saline. Bound antibody was detected using chemiluminescence reagent (Amersham Biosciences, Buckinghamshire, UK). To detect multiple signals from a single membrane, the membranes were treated with stripping buffer (59 mmol/l Tris-HCl, 2% sodium dodecyl sulfate, 0.75% 2-mercaptoethanol) for 20 min at 37 °C prior to re-blotting with a different antibody (36).
Experiments were repeated at least four times. The results are expressed as the mean ± s.d. of the number of observations. Statistical significance was assessed using one-way ANOVA or Student's t-test. P < 0.05 was considered statistically significant.
Effects of TNF-α on resistin expression
3T3-L1 adipocytes were incubated for 24 h with or without increasing concentrations (1.25–10 ng/ml) of TNF-α. As shown by northern blots (Figure 1a,b), incubation with 1.25 ng/ml of TNF-α resulted in a significant decrease in resistin mRNA levels compared to controls. This response was dose dependent, with an effector concentration for half-maximum response of 3.31 ng/ml, maximal inhibition being ∼80%. When cells were incubated with 2.5 ng/ml of TNF-α for different times (0–24 h), a significant decrease in resistin mRNA levels was seen after 4 h of incubation and maximal inhibition was seen at 24 h (Figure 1c).
Effects of TNF-α on iNOS expression and NO production
TNF-α has been reported to stimulate iNOS expression in many cell types. We also observed a stimulatory effect of TNF-α on iNOS expression in 3T3-L1 adipocytes. Western blot analysis (Figure 2) showed that, in untreated 3T3-L1 adipocytes, iNOS expression was very low and that TNF-α stimulation resulted in a time-dependent (Figure 2a) and dose-dependent (Figure 2b) increase in iNOS expression. We also examined whether this TNF-α–induced iNOS expression resulted in increased NO production. After 24 h of incubation with increasing concentrations of TNF-α (1.25–10 ng/ml), NO release by 3T3-L1 adipocytes, as estimated by nitrite accumulation, increased progressively from 0.25 to 2.25 μmol/l (Figure 3a).
Mediating role of NO in the TNF-α–induced downregulation of resistin expression
To examine whether NO was involved in the TNF-α–induced suppression of resistin expression, the general NOS inhibitor, l-NAME, and specific iNOS inhibitor, PBITU, were used to block NO release. We first performed a pilot study to determine the concentration of l-NAME which abolished NO production and found that TNF-α–stimulated NO production was successful blocked by 1 h preincubation of 3T3-L1 adipocytes with 10−4 mol/l l-NAME (Figure 3b). Northern blots showed that these pretreatment conditions significantly inhibited TNF-α–induced suppression of resistin expression in 3T3-L1 adipocytes (Figure 4a, compare lanes T and LT; Figure 4b, compare lanes T and PT). Since PTK has also been reported to be involved in TNF-α–induced NO production (26,29,30), we tested whether the PTK inhibitors, genistein and AG-1288, could prevent the TNF-α–induced downregulation of resistin mRNA and found that pretreatment with 10−5 mol/l genistein and 10−4 mol/l AG-1288 had a significant inhibitory effect (Figure 4, compare lanes T and GT; Figure 4b, compare lanes T and AT). Besides, Daidzein, an inactive negative control compound for genistein, showed no effect on TNF-α–downregulated resistin mRNA expression (Figure 4b, compare lanes T and DT).
We also evaluate the role of PTK and NO in the TNF-α–mediated induction of iNOS expression. As shown in Figure 5, western blots showed that the stimulatory effect of TNF-α on iNOS protein expression was completely suppressed by preincubation of 3T3-L1 adipocytes with 10−5 mol/l genistein (compare lanes T and GT) and 10−4 mol/l AG-1288 (compare lanes T and AT). Unexpectedly, pretreatment with 10−4 mol/l l-NAME and 10−4 mol/l PBITU partially suppressed the induction of iNOS expression by TNF-α (compare lanes T and LT in Figure 5a and lanes T and PT in Figure 5b).
Although the above results showed that TNF-α stimulated iNOS expression and subsequent NO production, it was unclear whether NO could inhibit resistin expression. To test this, we examined the effect of the NO donor, SNP, on adipocyte resistin expression. As shown by northern blots (Figure 6a,b), 24-h treatment with 0.5 mmol/l SNP caused a significant decrease (∼30%) in resistin mRNA levels, ∼70% inhibition being seen at 1 mmol/l, 80% at 2 mmol/l, and 90% at 4 mmol/l. To explore the time course of the effects of SNP on resistin gene expression, cells were exposed to 1 mmol/l SNP for different times (0–24 h). As shown in Figure 6c, a significant decrease (∼30%) in resistin expression was observed after 2-h incubation with SNP, ∼40% inhibition being seen at 4 h, 50% at 12 h, and 80% at 24 h. These findings suggest that NO plays a pivotal role in the regulation of adipocyte resistin expression.
This study was performed to explore the regulatory effects of TNF-α on resistin mRNA levels in 3T3-L1 adipocytes and to clarify the underlying signaling pathway. The main findings were that TNF-α, by activating a PTK-dependent pathway and inducing iNOS expression and subsequently NO production, decreased resistin mRNA levels in a dose-dependent and time-dependent manner. Our finding that TNF-α decreased resistin mRNA levels in 3T3-L1 adipocytes is compatible with those of Shojima et al. (25) and Fasshauer et al. (27), but not that of Kaser et al. (26). The major factor contributing to this difference could be the different cell models used. Kaser et al. suggested that, in humans, peripheral blood mononuclear cells may be the major source of resistin and used this cell model to explore the effect of TNF-α on resistin expression and found that resistin mRNA levels, measured by a fluorescence-based real-time polymerase chain reaction, were markedly increased by TNF-α.
Prior to our study, although findings on the inhibitory effect of TNF-α on resistin mRNA levels in 3T3-L1 adipocytes were highly consistent, the regulatory mechanism was unclear. Fasshauer's study (27) had excluded involvement of protein kinase A, p44/42, or p38 mitogen-activated protein kinase in the suppression of resistin gene expression by TNF-α. In our study, the PTK inhibitors, genistein and AG-1288, completely abolished TNF-α–mediated induction of iNOS protein expression (Figure 5) and prevented TNF-α–induced resistin downregulation (Figure 4). These data demonstrate that a PTK/iNOS/NO-dependent pathway was involved in TNF-α–decreased resistin mRNA levels in 3T3-L1 adipocytes. Cytokines, including TNF-α, activate PTK and stimulate iNOS expression and subsequently NO production in vascular smooth muscle cells and astrocytes (31,37). This PTK/iNOS/NO pathway was also observed in Poljakovic's study (38), which demonstrated that PTKs, such as Janus kinase 2, are involved in cytokine-mediated iNOS induction in human kidney epithelial cells. Furthermore, the Janus kinase/signal transducers and activators of transcription signaling pathway has also been implicated in iNOS induction in human lung and colon epithelial cells (39,40); its involvement was not investigated in this study.
In this study, the general NOS inhibitor l-NAME and specific iNOS inhibitor PBITU prevented TNF-α–induced resistin downregulation (Figure 4), but partially suppressed TNF-α–mediated induction of iNOS protein expression (Figure 5). It is because that l-NAME and PBITU competed with l-arginine, the substrate of NOS for NO production, for the binding site on iNOS. Hence, pretreatment with l-NAME and PBITU only blocked iNOS activity and subsequent NO production, but not iNOS expression in 3T3-L1 adipocytes. In addition, NO inhibition partially suppressed TNF-α–induced iNOS expression. This finding also suggested that a positive feedback loop between NO production and iNOS expression is possible. The precise mechanism involved in this feedback loop needs further investigation.
Regulation by NO of the expression of another adipocytokine, leptin, was reported by Unno's group (41), who found that treatment of 3T3-L1 adipocytes with an interferon-gamma/lipopolysaccharide mixture caused a significant reduction in leptin secretion and mRNA levels, but significant induction of iNOS protein and mRNA expression, and that the inhibitory effect on leptin expression was fully prevented by pretreatment with NOS inhibitors. In addition, Merial et al. (23) reported that TNF-α can directly downregulate uncoupling protein-2, a mitochondrial protein expressed in adipocytes and involved in the control of energy expenditure, via an iNOS/NO-dependent pathway. A similar regulatory pattern has also been observed in brown adipocytes, in which Uchida et al. (42) reported that TNF-α stimulates iNOS expression and NO production, and then suppressed lipoprotein lipase activity. Together, these findings suggest that NO plays a pivotal regulatory role in adipocyte physiology. Actually, several studies have demonstrated that NO directly regulates adipocyte functions. For example, NO, produced by bradykinin-stimulated endothelial NOS activation, enhances insulin sensitivity in isolated rat adipocytes (43). Furthermore, iNOS-mediated NO production modulates cytokine/lipopolysaccharide-mediated lipolysis in adipocytes (44). The study by Yan et al. (45) demonstrated that the NO-releasing reagent, hydroxylamine, or the NOS substrate, l-arginine, promotes adipocyte differentiation and suggested that NO has a modulatory effect on adipogenesis.
In adipocytes, TNF-α and resistin regulate each other's expression. A recent study showed that resistin significantly increases TNF-α secretion by adipocytes (46), while our present findings suggest that oversecretion of TNF-α will downregulate adipose resistin expression. The two processes thus result in homeostasis of adipocyte TNF-α and resistin expression and this homeostasis might be broken in obese or insulin-resistant individuals. This might explain why both TNF-α and resistin levels are highly increased in obesity (47,48). This hypothesis is partially supported by Degawa-Yamauchi et al.'s study (49), which showed that TNF-α inhibits adiponectin release in adipocytes isolated from lean subjects, but not those from obese subjects. It is possible that the efficacy of TNF-α on adiponectin release is reduced in adipocytes isolated from obese subjects. A change in the effect of TNF-α on adiponectin regulation might also apply to resistin regulation, i.e., in adipocytes of obese subjects, TNF-α may not have a suppressive effect on resistin expression, thus leading to resistin overexpression. However, additional experiments are necessary to clarify the mechanisms involved in the reduction of the efficacy of TNF-α in the modulation of resistin expression during the development of obesity or insulin resistance.
In conclusion, our data demonstrate that TNF-α–induced iNOS expression and subsequently NO production are involved in the TNF-α–suppressed resistin expression in 3T3-L1 adipocytes. Our findings also suggest a possible paracrine or autocrine interaction between adipocytokines, such as TNF-α and resistin. The results of this study clarify the regulatory role of TNF-α in resistin expression and its underlying mechanism under normal physiological conditions. Abnormal interactions between adipocytokines may contribute to the pathogenesis of metabolic disorders, including obesity, diabetes, and atherosclerosis.
This work was supported by research grant from the Veterans General Hospitals University System of Taiwan Joint Research Program, Tsou's Foundation (VGHUST94-P7-49).