The FASEB Journal • FJ Express Full-Length Article
TNF␣ up-regulates apelin expression in human and mouse adipose tissue Danie`le Daviaud,* Je´re´mie Boucher,* Ste´phane Gesta,* Ce´dric Dray,* Charlotte Guigne,* Didier Quilliot,† Ahmet Ayav,† Olivier Ziegler,† Christian Carpene,* Jean-Se´bastien Saulnier-Blache,* Philippe Valet,* and Isabelle Castan-Laurell*,1 *INSERM, U586, Unite´ de recherches sur les obe´site´s, Universite´ Paul Sabatier, Institut Louis Bugnard IFR31, Toulouse, France; and †Service de diabe´tologie, Maladies me´taboliques et nutrition, CHU de Nancy, Hoˆpital J. d’Arc, Nancy, France We have recently identified apelin as a novel adipokine up-regulated by insulin and obesity. Since obesity and insulin resistance are associated with chronically elevated levels of both insulin and TNF␣, the present study was performed to investigate a putative regulation of apelin expression in adipocytes by TNF␣. Herein, we report a tight correlation between apelin and TNF␣ expression in adipose tissue of lean and obese humans. Apelin regulation by TNF␣ was further studied in cultured explants of human adipose tissue. The endogenous expression of TNF␣ in adipocytes isolated from the explants was accompanied by a 6 –9 h subsequent increase of apelin expression in adipocytes. This increase was reversed by inhibiting TNF␣ expression with 100 M isobutylmethylxanthine. In different mouse models of obesity, expression of both TNF␣ and apelin was also significantly increased in adipocytes of obese mice. Furthermore, short-term exposure to an i.p. injection of TNF␣ in C57Bl6/J mice induced an increase of apelin expression in adipose tissue as well as apelin plasma levels. Finally, a direct positive effect of TNF␣ has been shown in differentiated 3T3F442A adipocytes on apelin expression and secretion. The signaling pathways of TNF␣ for the induction of apelin were dependent of PI3-kinase, c-Jun NH2-terminal kinase (JNK), and MAPK but not PKC activation. All together, these findings suggest that apelin might be a candidate to better understand potential links between obesity and associated disorders such as inflammation and insulin resistance.—Daviaud, D., Boucher, J., Gesta, S., Dray, C., Guigne, C., Quilliot, D., Ayav, A., Ziegler, O., Carpene, C., SaulnierBlache, J.-S., Valet, P., Castan-Laurell, I. TNF␣ upregulates apelin expression in human and mouse adipose tissue. FASEB J. 20, E796 –E802 (2006)
ABSTRACT
Key Words: obesity 䡠 adipocyte 䡠 adipokine
Apelin is a bioactive peptide originally identified from bovine stomach extracts as the endogenous ligand of the orphan G protein-coupled receptor, APJ (1). Apelin is derived from a prepropeptide of 77-amino acid that is cleaved into a 55-amino acid fragment and E796
then into shorter forms. The physiologically active form is thought to be apelin-36, the most biologically effective being the pyroglutamylated form of apelin-13 (1). In rodents, mRNA expression of both apelin and APJ has been described in the central nervous system and peripheral tissues (for review see 2, 3). The expression of apelin in the pituitary gland (4, 5), as well as the presence of apelin in the bloodstream (6, 7) is consistent with an endocrine function of this peptide. The origin of blood apelin and the regulations of its input to the bloodstream are of interest. Although little is known about the physiological roles of apelin, several studies have shown that apelin is involved in the regulation of cardiovascular function (8 –10) and fluid homeostasis (11, 12). Very recently, it was shown that apelin inhibited insulin secretion in mice (13). So far, few data are available concerning apelin production and regulation in apelin-expressing cells. Apelin has recently been described by our group as a novel adipokine, produced and secreted by human and mouse isolated mature adipocytes. Apelin mRNA levels found in isolated adipocytes were close to those of the so-called stroma vascular fraction (containing other cell types present in adipose tissue) or organs known to express apelin such as kidney and heart (14). We showed that a strong relationship exists between adipocyte-secreted apelin and insulin. Indeed, a direct action of insulin in the regulation of apelin expression in adipocytes was demonstrated in vivo and in vitro in 3T3F442A differentiated adipocytes. In addition, adipocyte apelin mRNA levels as well as plasma apelin concentration were increased in various mouse models of obesity associated with hyperinsulinemia (14). Since obesity and insulin resistance are associated with chronically elevated levels of both insulin and TNF␣, the present study was performed to investigate a putative regulation of apelin expression in human and mouse adipocytes by TNF␣. A tight correlation has been observed between apelin and TNF␣ mRNA levels in 1
Correspondence: FIR 31, Institut Louis Bugnard, BP 84 225 INSERM U586, Toulouse F-31432, Cedex 4, France. Email:
[email protected] doi: 10.1096/fj.05-5243fje 0892-6638/06/0020-0796 © FASEB
different anatomical location of adipose tissues from lean to severe obese subjects. To determine whether TNF␣ exerts a direct effect on apelin expression in adipocytes, further experiments were performed in C57Bl6/J mice and 3T3F442A mouse adipocyte cell line. Even though TNF␣ has been shown to inhibit insulin receptor signaling and insulin metabolic functions (15), the present work clearly describes a TNF␣dependent up-regulation of apelin expression in human and mouse adipose tissue.
MATERIALS AND METHODS Human samples Human adipose tissue was collected according to the guidelines of the Ethical Committee of Toulouse-Rangueil and Nancy J. d’Arc Hospitals. All subjects gave their informed consent to participate to the study, and investigations were performed in accordance with the declaration of Helsinki as revised in 2000 (http://www.wma.net/e/policy/b3.htm). Human subcutaneous (s.c.) adipose tissue samples were obtained from 20 patients [40.3⫾2.2 yr old, Body Mass Index (BMI): 27.3⫾1.2 kg/m2] undergoing abdominal lipectomy for plastic surgery. No clinical data from these patients were available. Human intra-abdominal adipose tissue samples were obtained from 23 morbidly (grade III) obese subjects (41.3⫾2.2 yr old, BMI: 45.7⫾2.9 kg/m2) before a bariatric surgery. Tissue samples were immediately frozen in liquid nitrogen and stored at – 80°C. Animals Mice were handled in accordance with the principles and guidelines established by the National Institute of Medical Research (INSERM). FVB/n and C57Bl6/J female mice were obtained from Charles River laboratory (l’Arbresle, France). Mice were housed conventionally in a constant temperature (20 –22°C) and humidity (50 – 60%) animal room and with a 12 h light– dark cycle (lights on at 8:00 am). All mice had free access to food and water throughout the experiment. For the in vivo effect of TNF␣, 15-wk-old fasted female C57Bl6/J mice received either an i.p. injection of 2 g/mouse TNF␣ (SIGMA) or PBS (in a same vol) at 8:00 a.m. They were sacrificed 8 h later; blood and different organs (heart, kidney and adipose tissue) were taken. Diets
Invitrogen, Paisley, UK) supplemented with 1 mg/ml collagenase (SIGMA) and 1% BSA for 30 min at 37°C under shaking. Digestion was followed by filtration through a 150 m screen, and the floating adipocytes were separated from the medium containing the stroma-vascular fraction (SVF). Adipocytes were washed twice in DMEM and further processed for RNA extraction using the RNeasy mini kit (Qiagen, GmbH, Hilden, Germany). Human s.c. adipose tissue explants were prepared and cultured as described previously (16), and total RNAs of isolated adipocytes were extracted. Cell culture The mouse preadipose cell line 3T3F442A was grown at 37°C in 7% CO2 in DMEM containing 25 mM glucose, 100 U/ml penicillin, 100 g/ml streptomycin (culture medium), and 10% donor calf serum. For differentiation, confluent preadipocytes were cultured in culture medium supplemented with 10% fetal calf serum and 50 nM insulin for 8 to 10 d, after which more than 90% of the cells had accumulated fat droplets. Total RNAs were prepared using the RNeasy mini kit (Qiagen). Real-time RT-polymerase chain reaction (PCR) Total RNAs (1 g) were reverse-transcribed using random hexamers and Superscript II reverse transcriptase (Invitrogen). The same reaction was performed without Superscript II (RT-) to estimate DNA contamination. Real-time PCR was performed starting with 12.5 ng cDNA, and both sense and antisense oligonucleotides in a final vol of 25 l using the SYBR green TaqMan Universal PCR Master Mix (Applied Biosystems, Warrington, UK). Fluorescence was monitored and analyzed in a GeneAmp 7000 detection system instrument (Applied Biosystems). Analysis of the 18S rRNA was performed in parallel using the rRNA control Taqman Assay Kit (Applied Biosystem, Warrington, UK) to normalize gene expression. Results are expressed as 2(Ct18S⫺Ctgene)(1–(1/ 2(Ctgene⫺CtRT–)) ), where Ct corresponds to the number of cycles needed to generate a fluorescent signal above a predefined threshold. Oligonucleotide primers were designed using the Primer Express software (Applied Biosystems, Warrington, UK). All primers used were validated for PCR efficiency. Apelin assay Apelin was quantified using the nonselective apelin immunoassay kit (Phoenix Pharmaceuticals Belmont, CA) following the manufacturer’s instructions. Statistical analysis
C57Bl6/J mice were assigned to chow, low-fat diet (LFD) or high-fat diet (HFD (SAFE, France). Energy contents of the specific diets were (% kcals): 20% protein, 70% carbohydrate, and 10% fat for LFD; 20% protein, 35% carbohydrate, and 45% fat for HFD. The main source of fat in HFD was lard (20 g/100g of food). C57Bl6/J (5 wk old) were fed a HFD or LFD for 8 wk. In 8-week-old FVB/n mice, obesity was induced by a single i.p. injection (0.5 g/kg) of gold thioglucose (GTG) (RDI, Flanders, NJ). All mice were sacrificed at 13 wk of age.
Results are expressed as means ⫾ sem. Statistical differences between two groups were evaluated using Student’s t test. Correlations were analyzed using the nonparametric Spearman rank test. The concentration of significance was set at P ⬍ 0.05.
Isolation of adipocytes from adipose tissue
The relationship between obesity, TNF␣, and apelin expression in humans
Mouse intra-abdominal and s.c. adipose tissues were dissected immediately after sacrifice, minced in 5 ml of Dulbecco’s modified Eagle’s medium (DMEM; Life Technologies, Inc.,
TNF␣ and apelin expression were investigated in s.c. abdominal fat pad from lean or moderately obese
TNF␣ UP-REGULATES APELIN IN ADIPOSE TISSUE
RESULTS
E797
humans (BMI ranged from 20.5 to 36.5 kg/m2, n⫽20). A linear correlation (r⫽0.790, slope 0.085, P⬍0.01) was found between TNF␣ and apelin mRNA levels (Fig. 1A). A similar correlation (r⫽0.994, slope 0.097, P ⬍ 0.01) was found in the intra-abdominal adipose tissue of massively obese subjects (BMI ranged from 39.2 to 59.3 kg/m2, n⫽23), (Fig. 1B). Thus, it clearly appears that a parallel increase of TNF␣ and apelin expression occurred in adipose tissue, whatever the anatomical location, the highest correlation being observed in morbid grade III obesity. No significant correlation with apelin has been found for other adipokines such as leptin, adiponectin or the plasminogen activator inhibitor type-1 factor (PAI-1) (not shown). To further study apelin regulation by TNF␣ in human adipose tissue in vitro, explants of s.c. adipose tissue were prepared and maintained in primary culture for 48h as described previously (16). Adipocytes were then isolated from explants and both TNF␣ and apelin mRNAs were quantified. As shown in Fig. 1C, the rise in TNF␣ expression started 3 h after the beginning of primary culture and was significantly different after 6 h culture from cells at time zero. Apelin expression in isolated adipocytes from cultured explants increased significantly only beyond the 12th h after the beginning of culture. Furthermore, explants were cultured in the presence of isobutylmethylxanthine (IBMX), a phosphodiesterase inhibitor known to inhibit TNF␣ secretion and expression in adipocytes from cultured explants (16). In the present study, 100 M IBMX, attenuated the rise of TNF␣ as well as apelin mRNA in adipocytes. Taken together, these data strongly suggest the involvement of TNF␣ produced during the culture in the up-regulation of apelin in human adipocytes.
Regulation of apelin by TNF␣ in mouse Since obesity is associated with a rise in adipose tissue TNF␣ content, we checked in mouse models of obesity the in vivo regulation of apelin by TNF␣. Thus, we studied apelin and TNF␣ expression in adipocytes isolated from C57Bl6/J mice either fed a HFD or LFD, and in mice rendered hyperphagic and obese by goldthioglucose (GTG) treatment. Both HFD-fed C57Bl6/J and GTG-treated FVB/n mice exhibited a significant increase in body mass (28.8⫾1.5 and 39.4⫾3.1 g, respectively) when compared with their control littermates (23.4⫾0.5 and 23.0⫾0.4 g, respectively). Apelin plasma levels were increased in both models of obese mice (14), while TNF␣ plasma levels were not significantly different from control mice (not shown). However, as shown in Fig. 2, both TNF␣ and apelin expression were significantly increased in isolated adipocytes of obese mice. Interestingly, in a model of obese transgenic mice (17) where TNF␣ expression was not significantly increased (18) in adipocytes, apelin expression (14) was also not increased. In addition, the expression of other pro-inflammatory adipokines such as PAI-1 or interleukin (IL)-6 (IL-6) known to be altered in obesity and/or diabetes (18) were not significantly increased (not shown) in HFD-fed C57Bl6/J and GTG-treated FVB/n mice. These results in mouse are in agreement with a coregulation between apelin and TNF␣ mRNA in adipocytes. To determine whether increased TNF␣ levels could directly up-regulate apelin expression as depicted in mouse, we investigated the direct effect of i.p. injection of TNF␣ into C57Bl6/J mice on apelin expression. Eight hours after injection, 2 g/ mouse TNF␣ elicited a significant increase of apelin expression in isolated
Figure 1. Correlation between apelin and TNF␣ expression in adipose tissue from both lean and obese humans. A) Human s.c. adipose tissue samples were obtained from 20 patients (40.3⫾2.2 yr old, BMI: 27.3⫾1.2 kg/m2) undergoing abdominal lipectomy for plastic surgery. B) Human intra-abdominal adipose tissue samples were obtained from 23 morbide grade III obeses (41.3⫾2.2 yr old, BMI: 45.7⫾2.9 kg/m2) before a bariatric surgery. C) Primary culture of explants was prepared from human s.c. tissue. Adipocytes were isolated from explants before (time 0) and after 3, 6, 24, and 48 h primary culture (left panel) without (–) or in the presence (⫹) of 100 nM IBMX (right panel). Results are mean values ⫾ sem from four separate experiments, *P ⬍ 0.05; **P ⬍ 0.01 when compared with controls. E798
Vol. 20
July 2006
The FASEB Journal
DAVIAUD ET AL.
transduction pathways in adipocytes, we tested the potential intracellular targets of TNF␣ by using specific inhibitors (Fig. 4B). Differentiated 3T3F442A cells were either preincubated with 20 M SP 600125 (an inhibitor of JNK) or 10 M LY294002 (a PI3K inhibitor). Both of them significantly decreased the TNF␣-induced apelin expression. Since TNF␣ has been shown to signal through ERK 1 and ERK 2 activator in 3T3F442A adipocytes (19), we also checked the effects of the MAPK inhibitor PD098059. TNF␣-induced increase of apelin expression was significantly inhibited by 50 M of PD098059. The potential involvement of PKC was studied by using 10 nM of GF109203X, a PKC inhibitor. The GF109203X did not inhibit the expression of apelin induced by TNF␣. Figure 2. Transcriptional coregulation of apelin and TNF␣ in two mouse models of obesity. Apelin and TNF␣ mRNA levels (black and white bars respectively) were quantified in isolated adipocytes from 13-week-old overweight C57Bl6/J mice fed a HFD, obese gold-thioglucose (GTG)-injected mice and controls mice (LFD fed mice and saline injected mice, respectively). Values are expressed as fold increase and compared with associated controls (dotted line). Results are mean values ⫾ sem from four mice per group. *P ⬍ 0.05, **P ⬍ 0.01 when compared with controls.
adipocytes compared with control mice (Fig. 3A). A similar increase was observed in the stroma-vascular fraction (2.34⫾0.16 in control and 11.90⫾3.82 arbitrary units in TNF␣-treated mice, P⬍0.05, n⫽6). No significant variations of apelin mRNA levels were observed in other tissues such as heart (9.1⫾0.91 in control and 10.96⫾1.54 arbitrary units in TNF␣-treated mice, n⫽6) and kidney (8.81⫾1.23 in control and 9.62⫾0.76 arbitrary units in TNF␣-treated mice, n⫽6) known to express apelin. In addition, a significant increase of apelin plasma levels (Fig. 3B) was observed in TNF␣-treated mice. All together these results suggest that 1) TNF␣ can up-regulate apelin expression in vivo specifically in adipose tissue; 2) the increased apelin expression in adipose tissue by TNF␣ could contribute to apelin plasma levels. Apelin expression is directly regulated by TNF␣ via the MAPK, JNK and PI3-K but not the PKC signaling pathway To test a direct effect of TNF␣ on fat cells, 10 ddifferentiated adipocytes from the murine 3T3F442A cell line were used. TNF␣ had no effect on apelin expression in undifferentiated 3T3F442A cells (not shown). The effect of TNF␣ on apelin expression in 3T3F442A cells was concentration-dependent (Fig. 4A). The increased expression of apelin was followed by an increased secretion of apelin in the medium (from 0.04 ng/ml in control cells to 0.35 ng/ml in treated cells). Thus, TNF␣ can promote directly both apelin expression and secretion in adipocytes. Since TNF␣ is known to signal through various TNF␣ UP-REGULATES APELIN IN ADIPOSE TISSUE
Figure 3. Effect of TNF␣ on apelin expression and secretion in mice. C57Bl6/J mice received either a single intraperitoneal (i.p.) injection of 2 ng/mouse TNF␣ in PBS or vehicle alone (control mice) and were sacrificed 8 h after injection. A) Apelin mRNA levels in isolated adipocytes and (B) apelin plasma levels from control and TNF␣-treated C57Bl6/J mice. Results are mean values ⫾ sem from 5– 6 mice in each group, *P ⬍ 0.05, **P ⬍ 0.001 when compared with control mice. E799
Figure 4. Regulation of apelin expression by TNF␣. in 3T3F442A adipocytes A) Dose response effect of 8 h TNF␣ treatment on apelin mRNA levels in 10-day differentiated 3T3F442A adipocytes. Results are mean values ⫾ sem from 3 separate experiments. B) Apelin mRNA levels in 10-day differentiated 3T3F442A adipocytes after 24 h serum deprivation and incubated 8 h or not with 10 ng/ml TNF␣ in the absence (–) or presence (⫹; added 1 h before TNF␣ treatment) of LY294002 (10 M), PD098059 (50 M), GF109203X (10 nM), or SP 600125 (20 M). Results are mean values ⫾ sem from four separate experiments, *P ⬍ 0.05 when compared with controls; §P ⬍ 0.05 when compared with TNF␣stimulated cells.
DISCUSSION A large number of studies have demonstrated that obese animals as well as humans have increased TNF␣ mRNA and protein levels in white adipose tissue but the correlation between adiposity and circulating concentrations of TNF␣ remains controversial (for review see 20). A variety of experimental and clinical studies suggest that TNF␣ may act as an important auto/ paracrine regulator of fat cell function which limits adipose tissue expansion, but which also promotes insulin resistance that may, in turn causes metabolic disturbances. The chronic and graduated elevation of TNF␣ in obesity may also promote modifications in the secretory function of the fat cell. It has already been described that increased PAI-1, TGF-, and IL-6 productions are linked with high TNF␣ expression (21). E800
Vol. 20
July 2006
Although clearly demonstrated to be directly regulated by TNF␣, such adipokines are over-expressed in massively obese patients and are associated with a cluster of abnormalities including hyperinsulinemia insulin resistance or hypertension (22). The present study clearly demonstrated a tight positive correlation between apelin and TNF␣ expression in adipose tissue of massively obese as well as in lean subjects. These results are in concordance with the results obtained by Heinonen et al. (23) showing that apelin plasma levels correlate positively with body mass index. Moreover, the time course of TNF␣ and apelin mRNA expression in isolated adipocytes from human cultured explants is likely to explain a paracrine effect of endogenously produced TNF␣ on apelin expression in the surrounding adipocytes. This hypothesis has been supported by the absence of up-regulation of apelin mRNA in IBMXtreated explants since this phosphodiesterase inhibitor has already been described to inhibit TNF␣ expression and secretion (16). The up-regulation of both TNF␣ and apelin expression has also been shown in adipocytes isolated from different mouse models of obesity. Adipocytes from obese mice (or obese subjects) are bigger and could content more mRNA than smaller ones when results are expressed as mg of total mRNA per cell. It is therefore possible that the increased mRNA levels measured might be greater than reported. To further study, the effect of TNF␣ on apelin expression in adipocytes, an i.p. injection of TNF␣ was performed and analyses were performed 8 h after injection, time already used to show increased levels of other adipokines in adipose tissue (24). TNF␣ induced a significant increase of apelin mRNA in both isolated adipocytes and the stroma-vascular fraction. However the expression of apelin in other tissues such as kidney and heart was not modified. Moreover, the rise in apelin mRNA observed in 3T3F442A adipocytes led us to conclude to a direct effect of TNF␣ on apelin expression. Furthermore, the signaling pathways by which TNF␣ induced apelin expression were depicted. TNF␣ has been shown to activate several signaling molecules such as PI3kinase, PKC, MAP kinase or JNK in adipocytes (25). By using specific inhibitors of these distinct TNF␣ signaling pathways, we found that TNF␣ (8 h) regulates apelin expression in adipocytes through different pathways involving the MAPK, JNK and PI3-kinase. We previously described that insulin induced apelin expression in adipocytes by using also the MAPK and PI3-kinase pathways (14). The expression of the forkhead transcription factor gene Foxc2 in 3T3-L1 adipocytes involved similar MAPK- and PI3K-dependent pathways (26). Common pathways, used by TNF␣ and insulin to induce gene expression in 3T3-L1 adipocytes have been already demonstrated (27). GFX 109203X, a broad-spectrum inhibitor of PKC isoforms, did not modified the TNF␣-induced expression of apelin. Interestingly, this signaling pathway differs from TNF␣mediated PAI-1 expression in 3T3-L1 adipocytes where PKC pathway plays a central role (27). So far, very little
The FASEB Journal
DAVIAUD ET AL.
is known about the signaling pathways mediating apelin expression in other cells type. The present work clearly demonstrates for the first time that TNF␣ may act as a direct regulator of apelin expression in both human and mouse adipocytes. In addition, i.p. injection of TNF␣ in mice induced increased both apelin expression in adipose tissue and blood plasma levels of apelin. In different mouse models of obesity, both apelin expression in adipocytes and apelin plasma levels were increased. Moreover, in obese humans, plasma apelin levels were significantly higher than in lean subjects (14). These findings were very recently strengthened by a study showing that apelin plasma levels were dramatically increased in morbid obese patients (23). Thus, when considering the total adipose tissue mass in the body, the endocrine role of adipocyte-produced apelin might be of importance. These results presented herein suggest that the increased apelin expression in adipose tissue could contribute to apelin plasma levels. In addition, the data from Xu et al. (28) showed that an increased macrophage and inflammation gene levels (such as TNF␣) seems to occur selectively in adipose tissue before the rise in circulating-insulin levels. Such an observation indicates that inflammation is observed before the appearance of insulin resistance. Thus, a mutual regulation of apelin production in adipocytes by both insulin and TNF␣ could be hypothesized, leading to the maintenance of sustained apelin expression and secretion in obesity. Taking in account the physiological roles of apelin in the cardiovascular system, the over-production of apelin in the obese could be one of the last protections before the emergence of obesityrelated disorder such as hypertension and cardiovascular dysfunctions. On the other hand, apelin up-regulation in obesity could contribute to endocrine or metabolic dysfunctions such as diabetic retinopathy since apelin has been proposed as a novel angiogenic factor in retinal endothelial cells (29). Thus, further studies to understand the contribution of apelin in associated regulations/dysregulations of obesity are of major interest. We gratefully acknowledge the animal facilities staff (Zootechnie, IFR31), the plastic surgery service of Rangueil Hospital, Bruno Se´guy (INSERM U466, IFR31) for help in TNF␣ assay and Max Lafontan for fruitful discussions.
4. 5. 6.
7.
8.
9.
10.
11.
12.
13. 14.
15.
16.
17.
18.
REFERENCES 19. 1.
Tatemoto, K., Hosoya, M., Habata, Y., Fujii, R., Kakegawa, T., Zou, M. X., Kawamata, Y., Fukusumi, S., Hinuma, S., Kitada, C., et al. (1998) Isolation and characterization of a novel endogenous peptide ligand for the human APJ receptor. Biochem. Biophys. Res. Commun. 251, 471– 476 2. Masri, B., Knibiehler, B., and Audigier, Y. (2005) Apelin signalling: a promising pathway from cloning to pharmacology. Cell Signal. 17, 415– 426 3. Kleinz, M. J., and Davenport, A. P. (2005) Emerging roles of apelin in biology and medicine. Pharmacol. Therapeutics 107, 198 –211
TNF␣ UP-REGULATES APELIN IN ADIPOSE TISSUE
20. 21. 22.
Brailoiu, G. C., Dun, S. L., Yang, J., Ohsawa, M., Chang, J. K., and Dun, N. J. (2002) Apelin-immunoreactivity in the rat hypothalamus and pituitary. Neurosci. Lett 327, 193–197 De Mota, N., Lenkei, Z., and Llorens-Cortes, C. (2000) Cloning, pharmacological characterization and brain distribution of the rat apelin receptor. Neuroendocrinology 72, 400 – 407 Foldes, G., Horkay, F., Szokodi, I., Vuolteenaho, O., Ilves, M., Lindstedt, K. A., Mayranpaa, M., Sarman, B., Seres, L., Skoumal, R., et al. (2003) Circulating and cardiac levels of apelin, the novel ligand of the orphan receptor APJ, in patients with heart failure. Biochem. Biophys. Res. Commun. 308, 480 – 485 Kleinz, M. J., and Davenport, A. P. (2004) Immunocytochemical localization of the endogenous vasoactive peptide apelin to human vascular and endocardial endothelial cells. Regul. Pept. 118, 119 –125 Lee, D. K., Cheng, R., Nguyen, T., Fan, T., Kariyawasam, A. P., Liu, Y., Osmond, D. H., George, S. R., and O’Dowd, B. F. (2000) Characterization of apelin, the ligand for the APJ receptor. J. Neurochem. 74, 34 – 41 Ishida, J., Hashimoto, T., Hashimoto, Y., Nishiwaki, S., Iguchi, T., Harada, S., Sugaya, T., Matsuzaki, H., Yamamoto, R., Shiota, N., et al. (2004) Regulatory roles for APJ, a seven-transmembrane receptor related to AT1, in blood pressure in vivo. J. Biol. Chem. 279, 26274 –26279 Ashley, E. A., Powers, J., Chen, M., Kundu, R., Finsterbach, T., Caffarelli, A., Deng, A., Eichhorn, J., Mahajan, R., Agrawal, R., et al. (2005) The endogenous peptide apelin potently improves cardiac contractility and reduces cardiac loading in vivo. Cardiovascular Res. 65, 73– 82 Taheri, S., Murphy, K., Cohen, M., Sujkovic, E., Kennedy, A., Dhillo, W., Dakin, C., Sajedi, A., Ghatei, M., and Bloom, S. (2002) The effects of centrally administered apelin-13 on food intake, water intake and pituitary hormone release in rats. Biochem. Biophys. Res. Commun. 291, 1208 –1212 De Mota, N., Reaux, A., Messari, S. E., Chartrel, N., Roesch, D., Dujardin, C., Kordon, C., Vaudry, H., Moos, F., and LlorensCortes, C. (2004) Apelin, a potent diuretic neuropeptide counteracting vasopressin actions through inhibition of vasopressin neuron activity and vasopressin. Proc. Natl. Acad. Sci. U.S.A. 101, 10464 –10469 Sorhede Winzell, M., Magnusson, C., and Ahren, B. (2005) The apj receptor is expressed in pancreatic islets and its ligand, apelin, inhibits insulin secretion in mice. Regul. Pept. 131, 12–17 Boucher, J., Masri, B., Daviaud, D., Gesta, S., Guigne´, C., Mazzucotelli, A., Castan-Laurell, I., Tack, I., Knibiehler, B., Carpene, C., Audigier, Y., Saulnier-Blache, J. S., and Valet, P. (2005) Apelin, a newly identified adipokine up-regulated by insulin and obesity. Endocrinology 146, 1764 –1771 Hotamisligil, G. S., Peraldi, P., Budavari, A., Ellis, R., White, M. F., and Spiegelman, B. M. (1996) IRS-1-mediated inhibition of insulin receptor tyrosine kinase activity in TNF-alpha- and obesity-induced insulin resistance. Science 271, 665– 668 Gesta, S., Lolmede, K., Daviaud, D., Berlan, M., Bouloumie, A., Lafontan, M., Valet, P., and Saulnier-Blache, J. S. (2003) Culture of human adipose tissue explants leads to profound alteration of adipocyte gene expression. Horm. Metab. Res. 35, 158 –163 Valet, P., Grujic, D., Wade, J., Ito, M., Zingaretti, M. C., Soloveva, V., Ross, S. R., Graves, R. A., Cinti, S., Lafontan, M., and Lowell, B. B. (2000) Expression of human alpha 2-adrenergic receptors in adipose tissue of beta 3-adrenergic receptor-deficient mice promotes diet-induced obesity. J. Biol. Chem. 275, 34797–34802 Boucher, J., Castan-Laurell, I., Daviaud, D., Guigne´, C., Bule´on, M., Carpe´ne´, C., Saulnier-Blache, J. S., and Valet, P. (2005) Adipokine expression profile in adipocytes of different mouse models of obesity. Horm. Metab. Res. 37, 761–76 Engelman, J. A., Berg, A. H., Lewis, R. Y., Lisanti, M. P., and Scherer, P. E. (2000) Tumor necrosis factor alpha-mediated insulin resistance, but not dedifferentiation, is abrogated by MEK1/2 inhibitors in 3T3–L1 adipocytes. Mol. Endocrinol. 14, 1557–1569 Ruan, H., and Lodish, H. F. (2003) Insulin resistance in adipose tissue: direct and indirect effects of tumor necrosis factor-␣. Cytokine Growth Factor Rew. 14, 447– 455 Coppack, S. W. (2001) Pro-inflammatory cytokines and adipose tissue. Proc. Nutr. Soc. 60, 349 –356 Kershaw, E. E., and Flier, J. S. (2004) Adipose tissue as an endocrine organ. J. Clin. Endocrinol. Metab. 89, 2548 –2556
E801
23.
Heinonen, M. V., Purhonen, A. K., Miettinen, P., Paakkonen, M., Pirinen, E., Alhava, E., Akerman, K., and Herzig, K. H. (2005) Apelin, orexin-A and leptin plasma levels in morbid obesity and effect of gastric banding. Regul. Pept. 130, 7–13 24. Samad, F., Yamamoto, K., and Loskutoff, D. J. (1996) Distribution and regulation of plasminogen activator inhibitor-1 in murine adipose tissue in vivo. J. Clin. Invest. 97, 37– 46 25. Jain, R. G., Phelps, K. D., and Pekala, P. H. (1999) Tumor necrosis factor-␣ initiated signal transduction in 3T3–L1 adipocytes. J. Cell. Physiol. 179, 58 – 66 26. Gronning, L. M., Cederberg, A., Miura, N., Enerback, S., and Tasken, K. (2002) Insulin and TNF alpha induce expression of the forkhead transcription factor gene Foxc2 in 3T3–L1 adipocytes via PI3K and ERK 1/2-dependent pathways. Mol. Endocrinol. 16, 873– 883
E802
Vol. 20
July 2006
27.
Pandey, M., Loskutoff, D. J., and Samad, F. (2005) Molecular mechanisms of tumor necrosis factor-␣-mediated plasminogen activator inhibitor-1 expression in adipocytes. FASEB J. 19, 1317–1319 28. Xu, H., Barnes, G. T., Yang, Q., Tan, G., Yang, D., Chou, C. J., Sole, J., Nichols, A., Ross, J. S., Tartaglia, L. A., and Chen, H. (2003) Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J. Clin. Invest. 112, 1821–1830 29. Kasai, A., Shintani, N., Oda, M., Kakuda, M., Hashimoto, H., Matsuda, T., Hinuma, S., and Baba, A. (2004) Apelin is a novel angiogenic factor in retinal endothelial cells. Biochem. Biophys. Res. Commun. 325, 395– 400
The FASEB Journal
Received for publication November 14, 2005. Accepted for publication February 16, 2006.
DAVIAUD ET AL.
The FASEB Journal • FJ Express Summary
TNF␣ up-regulates apelin expression in human and mouse adipose tissue Danie`le Daviaud,* Je´re´mie Boucher,* Ste´phane Gesta,* Ce´dric Dray,* Charlotte Guigne,* Didier Quilliot,† Ahmet Ayav,† Olivier Ziegler,† Christian Carpene,* Jean-Se´bastien Saulnier-Blache,* Philippe Valet,* and Isabelle Castan-Laurell*,1 *INSERM, U586, Unite´ de Recherches sur les Obe´site´s, Universite´ Paul Sabatier, Institut Louis Bugnard IFR31, Toulouse, France; and †Service de Diabe´tologie, Maladies Me´taboliques et Nutrition, CHU de Nancy, Hoˆpital J. d’Arc, Nancy, France To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.05-5243fje SPECIFIC AIMS Apelin is a novel peptide identified as the endogenous ligand for the orphan receptor APJ. Apelin mRNA has been detected in various tissues but little is known about apelin production and regulation in apelinexpressing cells. During the past decade, the secreted factors by adipose tissue (or adipokines) have gained much attention. Numerous adipokines are altered with obesity and could affect insulin-sensitivity or obesityassociated disorders. These adipokines represent possible drug targets for clinical investigations. Recently, we demonstrated that apelin is secreted by adipocytes, increased with obesity (plasma apelin levels were significantly higher in obese than in lean subjects) and regulated by insulin. Since obesity and insulin resistance are associated with chronically elevated levels of both insulin and TNF␣, the present study was performed to investigate a putative regulation of apelin expression by TNF␣ in human and mouse adipocytes. This relation was further studied in primary culture of human subcutaneous (s.c.) adipose tissue explants, in C57Bl6/J mice and the signaling pathways by which TNF␣ increased apelin expression was evaluated in 3T3F442A adipocytes.
PRINCIPAL FINDINGS 1. The expression of TNF␣ and apelin are tightly correlated in human adipose tissue from lean to morbide obese subjects In human adipose tissue, there is a very tight correlation between the expression of TNF␣ and apelin, whatever the anatomical location of fat pad: in s.c. adipose tissue, r ⫽ 0.790, slope 0.085, P ⬍ 0.01; in intra-abdominal adipose tissue: r ⫽ 0.994, slope 0.097, P ⬍ 0.01 (Fig. 1A, B). This correlation occurs in 1528
moderately obese subjects (BMI ranged from 20.5 to 36.5 kg/m2, n⫽20) and in morbidly obese subjects (BMI ranged from 39.2 to 59.3 kg/m2, n⫽23). This strong correlation between TNF␣ and apelin was not observed between apelin and other adipokines such as leptin, adiponectin, or PAI-1. To further study apelin regulation by TNF␣ in human adipose tissue in vitro, explants of s.c. adipose tissue were prepared and maintained in primary culture for 48 h. Adipocytes were then isolated from explants, and both TNF␣ and apelin mRNA were quantified. The rise in TNF␣ expression started 3 h after the beginning of primary culture and was significantly different from control cells after 6 h culture. During the same period, apelin expression dramatically increased in isolated adipocytes only beyond the 12th h after the beginning of culture. Furthermore, explants were cultured in the presence of 100 M isobutylmethylxanthine (IBMX), a phosphodiesterase inhibitor shown to inhibit TNF␣ production in adipocytes from cultured explants. By inhibiting the in situ expression of TNF␣ with IBMX, the up-regulation of apelin was prevented (Fig. 1C). This suggests a paracrine effect of endogenously produced TNF␣ on apelin expression in human adipose tissue. 2. Injection (i.p.) of TNF␣ into C57Bl6/J mice increases apelin expression in adipose tissue and apelin blood levels To determine whether increased TNF␣ levels could directly up-regulate apelin expression, we investigated the direct effect of i.p. injection of TNF␣ into C57Bl6/J mice on apelin expression in adipocytes. Eight hours after injection, TNF␣ (2 g/mouse) elicited a signifi1
Correspondence: IFR 31, Institut Louis Bugnard, BP 84 225 INSERM U586, Toulouse F-31432, Cedex 4, France. Email:
[email protected] doi: 10.1096/fj.05-5243fje 0892-6638/06/0020-1528 © FASEB
Figure 1. Correlation between apelin and TNF␣ expression in adipose tissue from both lean and obese humans. A) Human s.c. adipose tissue samples were obtained from 20 patients (40.3⫾2.2 yr old, BMI: 27.3⫾1.2 kg/m2) undergoing abdominal lipectomy for plastic surgery. B) Human intra-abdominal adipose tissue samples were obtained from 23 morbidly (grade III) obese subjects (41.3⫾2.2 yr old, BMI: 45.7⫾2.9 kg/m2) before a bariatric surgery. Correlations were analyzed using the nonparametric Spearman rank test. C) Primary culture of explants was prepared from human s.c. tissue. Adipocytes were isolated from explants before (time 0) and after 3, 6, 24, and 48 h primary culture (left panel) without (–) or in the presence (⫹) of 100 nM IBMX (right panel). Results are mean values ⫾ sem from four separate experiments, *P ⬍ 0.05; **P ⬍ 0.01 when compared with controls.
cant increase of apelin expression in isolated adipocytes (6.47 ⫾ 0.86 and 20.34 ⫾ 6.62 arbitrary units, n ⫽ 6) and in the stroma-vascular fraction (2.34⫾0.16 and 11.90⫾3.82 arbitrary units, P⬍0.05, n⫽6) compared with control mice. No significant variations of apelin mRNA levels were observed in other tissues such as heart (9.1⫾0.91 in control and 10.96⫾1.54 arbitrary units in TNF␣-treated mice, n⫽6) and kidney (8.81⫾1.23 in control and 9.62⫾0.76 arbitrary units in TNF␣-treated mice, n⫽6) known to express apelin. In addition, a significant increase of apelin plasma levels was observed in TNF␣-treated mice (1.27⫾0.06 ng/ml vs. 0.50⫾0.07 ng/ml in control mice, n⫽6). All together these results suggest that 1) TNF␣ can upregulate apelin expression in vivo specifically in adipose
tissue; 2) the increased apelin expression in adipose tissue by TNF␣ could contribute to apelin plasma levels. 3. Apelin mRNA induction in response to TNF␣ in adipocytes was mediated via the MAPK, c-Jun NH2terminal kinase (JNK) and PI3-K but not the PKC signaling pathway To test a direct effect of TNF␣ on fat cells, 10 ddifferentiated adipocytes from the murine 3T3F442A cell line were used. Short exposure (8 h) of TNF␣ induced apelin expression (Fig. 2A)and secretion in the medium (values ranging from 0.04 ng/ml in control cells to 0.35 ng /ml in treated cells). Since TNF␣ is Figure 2. Regulation of apelin expression by TNF␣⫾. in 3T3F442A adipocytes. A) Dose response effect of 8 h TNF␣ ⫾ treatment on apelin mRNA levels in 10-day-differentiated 3T3F442A adipocytes. Results are mean values ⫾ sem from 3 separate experiments. B) Apelin mRNA levels in 10-day-differentiated 3T3F442A adipocytes after 24 h serum deprivation and incubated 8 h or not with 10 ng/ml TNF␣⫾ in the absence (⫺) or presence (⫹; added 1 h before TNF␣⫾ treatment) of LY294002 (10 M), PD098059 (50 M), GF109203X (10 M), or SP 600125, (20 M). Results are mean values ⫾ sem from 4 separate experiments, *P ⬍ 0.05 when compared with controls; § P ⬍ 0.05 when compared with TNF␣⫾-stimulated cells.
TNF␣ UP-REGULATES APELIN IN ADIPOSE TISSUE
1529
Figure 3. Schematic diagram representing the putative roles of apelin secreted by adipose tissue in response to TNF␣.
known to signal through various transduction pathways in adipocytes, we tested the potential intracellular targets of TNF␣ by using specific inhibitors. We showed that the induction of apelin expression by TNF␣ on 3T3 F442A adipocytes was mediated by PI3-kinase, MAPK, JNK but not the PKC signaling (Fig. 2B). CONCLUSIONS AND SIGNIFICANCE TNF␣ has been shown to regulate expression of different adipokines like leptin or the plasminogen activator inhibitor type-1 factor (PAI-1). The present work clearly demonstrates for the first time that TNF␣ may act as a
1530
Vol. 20
July 2006
direct regulator of apelin expression in both human and mouse adipose tissue. In addition, i.p. injection of TNF␣ in mice induced increased both apelin expression in adipose tissue and blood plasma levels of apelin. Our data bring new open lines of investigation in this field and propose apelin as a potential target in obesity acting either as an endocrine or paracrine factor (Fig. 3). It therefore appears of interest to test whether the TNF␣-induced increase apelin expression in adipose tissue and secretion is implicated in the changes occurring in adipose tissue during the onset of obesity and the emergence of endocrine, metabolic, or cardiovascular dysregulations.
The FASEB Journal
DAVIAUD ET AL.