A Novel Role for Adipose Ephrin-B1 in Inflammatory Response

Aims Ephrin-B1 (EfnB1) was selected among genes of unknown function in adipocytes or adipose tissue and subjected to thorough analysis to understand its role in the development of obesity. Methods and Results EfnB1 mRNA and protein levels were significantly decreased in adipose tissues of obese mice and such reduction was mainly observed in mature adipocytes. Exposure of 3T3-L1 adipocytes to tumor necrosis factor-α (TNF-α) and their culture with RAW264.7 cells reduced EFNB1 levels. Knockdown of adipose EFNB1 increased monocyte chemoattractant protein-1 (Mcp-1) mRNA level and augmented the TNF-α-mediated THP-1 monocyte adhesion to adipocytes. Adenovirus-mediated adipose EFNB1-overexpression significantly reduced the increase in Mcp-1 mRNA level induced by coculture of 3T3-L1 adipocytes with RAW264.7 cells. Monocyte adherent assay showed that adipose EfnB1-overexpression significantly decreased the increase of monocyte adhesion by coculture with RAW264.7 cells. TNF-α-induced activation of extracellular signal-regulated kinase 1/2 (ERK1/2) was reduced by EFNB1-overexpression. Conclusions EFNB1 contributes to the suppression of adipose inflammatory response. In obesity, reduction of adipose EFNB1 may accelerate the vicious cycle involved in adipose tissue inflammation.


Introduction
Obesity, especially visceral fat obesity, is an important aspect of the metabolic syndrome and atherosclerosis [1,2]. In the Human Body Map project [3], our group provided evidence for the endocrine function of adipose tissue, in addition to serving as an energy storage organ [4]. Our group also discovered Adiponectin among human adipose tissue cDNAs [5]. Clinical evidence indicates that adipocytes produce various cytokines and chemokines, which we named adipocytokines, and that the obesity-related changes in adipocytokines contribute to the development of the metabolic syndrome [6]. Infiltration of immunocytes, such as macrophages, is observed in obese adipose tissues and these cells induce chronic lowgrade inflammation by producing a battery of inflammatory cytokines and chemokines [7][8][9]. Moreover, the infiltrated macrophages interact with adipocytes via inflammatory mediators, such as free fatty acids and adipocytokines, to generate a metabolic vicious cycle, which eventually lead to the clinical spectrum of the metabolic syndrome [10].
It is our view that the visceral fat status affects and reflects the gene expression profile in peripheral blood cells. In this regard, we reported recently the association between visceral fat adiposity and gene expression profile of peripheral blood cells in human subjects [11,12]. Furthermore, in a series of exploratory research studies, we searched for genes of unknown function in adipocytes and adipose tissues by comparing the cDNA microarray-based gene expression patterns of human peripheral blood cells and mouse adipose tissues. Our search identified various genes, among them ephrin-B1 (EfnB1). Historically, the first Eph receptor, EphA1 and its Eph receptor-interacting (ephrin) ligand, ephrin-A1, were cloned from cancer cells [13,14]. Increasing evidence indicate that Eph receptors and their ephrin ligands mainly serve as a cell communication system and their interactions play crucial roles in both normal steady conditions and pathological diseases [15]. A unique feature of the Eph-ephrin complexes is their capability to transduce bidirectional intracellular signals; such property closely regulates the physiological and pathological cellular events as well as developmental processes [16].
To our knowledge, there is virtually no information on the regulation and functions of EFNB1 in adipocytes. In this report, we describe the novel role of EFNB1 in the development of adipose inflammation.

Animals
Male C57BL/6N mice and ob/ob mice were obtained from Charles River Japan Inc. (Kanagawa, Japan) and maintained at 22°C under a 12:12-h light-dark cycle (lights on from 7:00 to 19:00). For analysis of tissue distribution, 12-week-old male C57BL/6N mice were euthanized by bleeding from the inferior vena cava under anesthesia after 12 hrs of fasting, and various tissue samples were excised. For the diet-induced obese (DIO) model study, 8-week-old male C57BL/6N mice were fed either regular chow diet (MF; Oriental Yeast, Osaka, Japan) or highfat and high-sucrose (HF/HS) diet (F2HFHSD; Oriental Yeast) for 8 weeks. At 16 weeks of age, the mice were euthanized under feeding condition, blood samples were collected from the inferior vena cava, and epididymal white adipose tissues (WAT) were excised. For obese model mice study, C57BL/6N and ob/ob mice were fed regular chow diet and sacrificed at 8 or 16 weeks of age. In all experiments, mice were anesthetized with an intraperitoneal injection of a mixture of medetomidine (0.3 mg/kg body weight), midazolam (4 mg/kg body weight) and butorphanol tartrate (5 mg/kg body weight). The experimental protocols were approved by the Ethics Review Committee for Animal Experimentation of Osaka University School of Medicine. This study also conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health.

Fractionation of WAT
WAT were minced in Krebs-Ringer buffer [composition: 120 mmol/L NaCl, 4 mmol/L KH 2 PO 4 , 1 mmol/L MgSO 4 , 1 mmol/L CaCl 2 , 10 mmol/L NaHCO 3 , 30 mmol/L HEPES, 20 mmol/L adenosine, and 4% (wt/vol) bovine serum albumin (Calbiochem, San Diego, CA)]. Tissue suspensions were centrifuged at 500 x g for 5 min to remove erythrocytes and free leukocytes. Collagenase was added to a final concentration of 2 mg/mL and suspensions were incubated at 37°C for 20 min under continuous shaking. The cell suspension was filtered through a 250 µm filter and then spun at 300 x g for 1 min to separate the floating mature adipocytes fraction (MAF) from the stromal vascular cell fraction (SVF) pellet. This fractionation and washing procedures were repeated twice with Krebs-Ringer buffer. Finally, both fractions were washed with phosphate buffered saline (PBS) and subjected to quantitative real-time polymerase chain reaction (RT-PCR).

Introduction of siRNA
On day 7 after the induction of differentiation, 3T3-L1 adipocytes were transfected with siRNA for EfnB1 (forward sequence 5'-CUA UGA AGA UGU UAU GAA TT-3' and reverse sequence 5'-UUC AUA ACG AUC UUC AUA GTG-3', Qiagen, Valencia, CA) using DeliverX Plus siRNA Transfection kit (Affymetrix, Santa Clara, CA) according to the protocol recommended by the manufacturer. The transfected cells were incubated for 24 hrs and then treated with the indicated concentrations of TNF-α. After treatment with TNF-α for 24 hrs, the cells were harvested and subjected to quantitative RT-PCR or western blotting. In these experiments, allstars negative control siRNA (Qiagen) was used as a control.

Construction and preparation of Ephrin-B1-expressing adenovirus
The full-length cDNA of EfnB1 from mouse colon was subjected to RT-PCR using Pfu DNA polymerase (Promega, Madison, WI) with primers containing a restriction enzyme cutting site at the end. The amplicons were cloned into pENTRTM1A vector (Life Technologies, Carlsbad, CA) using restriction enzyme sites. After confirming the correct sequences, the genes encoding ephrin-B1 in pENTRTM1A vector were transferred into the adenoviral expression vector (pAd/CMV/V5-DEST; Life Technologies) by recombination following the instructions provided by the manufacturer. The resultant pAd/CMV plasmids containing the target gene were linearized by PacI digestion and transfected into 293A cells by lipofectamine-2000 (Life Technologies) according to the protocol recommended by the manufacturer. On day 2 after transfection, the 293A cells were passaged and cultivated until 80% of cells became detached. The cell suspension was frozen then thawed three times. After centrifugation at 1,750 x g for 15 min, the supernatant was used as the gene expression adenoviral preparation (Ad-EfnB1). The titers for the adenoviral preparation were approximately 2×10 8 plaque forming units (pfu)/mL. In these experiments, the adenovirus expressing βgalactosidase (Ad-βgal) was used as a control.

Infection of adipocytes with the prepared adenovirus
For efficient transduction of the adenovirus, 3T3-L1 cells stably expressing Coxsackie-Adenovirus Receptor (CAR-3T3-L1) were used in the adenoviral study [18,19]. On day 7 after the induction of differentiation, CAR-3T3-L1 adipocytes were infected with Ad-EfnB1 or Ad-βgal at 2.0 multiplicity of infection (MOI) ( Figure S2). At 24 hrs after adenovirus infection, the medium was changed to remove uninfected adenovirus. At day 9, CAR-3T3-L1 adipocytes were cocultured with RAW264.7 cells using the transwell inserts, as described above. After coincubation for 24 hrs, CAR-3T3-L1 adipocytes were harvested and subjected to quantitative RT-PCR or immunoblotting.

Quantitative RT-PCR
Isolation of total RNA and production of cDNA were performed as described previously [17]. RT-PCR was performed on the ViiATM 7 real-time PCR system (Life Technologies) using the THUNDERBIRDTM qPCR Mix (TOYOBO, Osaka, Japan) according to the instructions provided by the manufacturer. For quantitative precision, the same amount of total RNA was consistently used for each expression analysis and the expression level of each gene was normalized by the mRNA level of a housekeeping gene, ribosomal protein, large, P0 (Rplp0/36B4). The following is a list of the primers used in this study: mouse Ephrin-B1,

Monocyte adherent assay
The adhesion of THP-1 human monocytic cell line to 3T3-L1 adipocytes was determined as described previously with minor modification [20]. Briefly, THP-1 cells were fluorescently labeled by incubation with calcein-AM (Dojin Chemical, Kumamoto, Japan) for 45 min, and washed twice in RPMI medium. 3T3-L1 adipocytes transfected with siRNA for EfnB1 or negative control were incubated with TNF-α (1 ng/mL) for 24 hrs and CAR-3T3-L1 adipocytes infected with Ad-EfnB1 or Adβgal were cocultured with RAW264.7 cells for 24 hrs prior to the adhesion assay. The labeled THP-1 cells were added to 3T3-L1 adipocytes and allowed to adhere for 90 min at 4°C. Cells were washed gently three times to remove non-adherent monocytes. Adherent cells were lysed with 50 mmol/L Tris (pH 8.4)/0.1% SDS, and fluorescence was measured with excitation at 485 nm wavelength and detection at 535 nm wavelength.

Immunoblotting
Preparation of protein extracts from tissues and cells was performed as described previously [21]. For measurement of extracellular signal-regulated kinase (ERK) 1/2 phosphorylation, CAR-3T3-L1 adipocytes were infected with Ad-EfnB1 or Ad-βgal before the assay. After 24-hr incubation, cells were incubated with or without TNF-α (1 ng/mL) for 5 min. At the end of incubation, the cells were harvested and subjected to immunoblotting. Twenty µg of protein was subjected to 4-20% gradient SDS-PAGE gel and then transferred to a nitrocellulose membrane (GE Healthcare, Little Chalfont, UK). For immunoblotting, the membrane was incubated with 1:1,000 dilution of goat anti-ephrin-B1 (R&D systems Inc., Minneapolis, MN), rabbit anti-α-tubulin, mouse anti-phospho-Erk1/2, or rabbit ant-Erk1/2 (Cell signaling technology, Danvers, MA). Detection was achieved using the enhanced chemiluminescence kit (GE Healthcare).

Statistical analysis
All values were expressed as mean±SD. Differences between groups were analyzed by one-factor ANOVA and unpaired Student's t-test. P values less than 0.05 were considered statistically significant.

Changes in Ephrin-B1 expression in obese adipose tissue
EFNB1 mRNA levels in peripheral blood cells correlated negatively with the estimated visceral fat area (eVFA) in human subjects (range of BMI, 25.4-51.2 kg/m 2 ; range of eVFA, 80-386 cm 2 ) ( Figure S3). The significance of EFNB1 in adipose tissue was analyzed in mice and cell experiments. Figure 1A shows the tissue distribution of EfnB1 mRNA in lean control (C57BL/6) mice. EfnB1 mRNA was highly expressed in the colon and lung, and also abundantly detected in WAT.
Next, we examined changes in EfnB1 mRNA level in obesity. The EfnB1 mRNA level in WAT was significantly lower in mice fed HF/HS diet than the normal chow group ( Figure 1B). Furthermore, adipose tissue EfnB1 mRNA level was significantly lower in obese ob/ob mice at 8 and 16 weeks of age, compared with that in lean control mice at the corresponding ages ( Figure 1C). The amount of EFNB1 protein was also lower in ob/ob mice compared to lean control mice at 16 weeks of age ( Figure 1D). Next, we examined EfnB1 mRNA level in MAF and SVF following fractionation of adipose tissue ( Figure 1E). EfnB1 mRNA was abundantly expressed in MAF compared to SVF and was significantly lower in obese MAF. These results indicate that under-expression of EfnB1 mRNA in adipocytes may account for the low EfnB1 mRNA level in obese WAT.
We also investigated the effects of nutritional changes on adipose tissue EfnB1 mRNA level. Fasting increased EfnB1 mRNA level in WAT ( Figure S4a), but streptozotocin (STZ)induced insulin-deficiency had no influences on the mRNA level ( Figure S4b), suggesting no direct effect for insulin on EfnB1 mRNA level in WAT.

Changes in Ephrin-B1 expression in 3T3-L1 adipocytes
Since EfnB1 was mainly expressed in mature adipocytes ( Figure 1E), we examined the regulation of EfnB1 using 3T3-L1 Ephrin-B1 and Adipose Inflammation PLOS ONE | www.plosone.org adipocytes. EfnB1 mRNA level decreased during differentiation of 3T3-L1 cells into adipocytes (Figure 2A). To investigate the regulation of EfnB1 in WAT of obese animals, 3T3-L1 adipocytes were stimulated with tumor necrosis factor-α (TNFα) and cocultured with RAW264.7 macrophages ( Figure S1). As shown in Figure 2B, TNF-α reduced EfnB1 mRNA level in 3T3-L1 adipocytes and the effect was dose-dependent. Interestingly, EfnB1 mRNA level was markedly reduced in 3T3-L1 adipocytes after coculture of these cells with RAW264.7 cells, while the level in RAW264.7 cells was low and not influenced by co-culture with 3T3-L1 adipocytes ( Figure 2C). EfnB1 mRNA level significantly decreased in 3T3-L1 adipocytes in proportion with the number of RAW264.7 cells ( Figure 2D). Furthermore, the amount of EFNB1 protein also decreased in 3T3-L1 adipocytes under co-culture with RAW264.7 cells ( Figure 2E). Collectively, the results of these in vitro experiments suggest that macrophage-derived factors, e.g., TNF-α, suppress adipose EFNB1 expression.

Effects of suppression and overexpression of Ephrin-B1 on adipose inflammatory response
To explore the role of EFNB1 in adipocytes, knockdown and overexpression of EFNB1 were conducted by using siRNA and adenovirus, respectively. The introduction of siRNA designed for EfnB1 (EfnB1-siRNA) successfully reduced EfnB1 mRNA ( Figure 3A) and EFNB1 protein ( Figure 3B) levels. Interestingly, Mcp-1 mRNA level was significantly increased with and without TNF-α, when 3T3-L1 adipocytes were transfected with EfnB1-siRNA ( Figure 3C). The introduction of EfnB1-siRNA caused a significant increase in Il-6 mRNA level in TNF-α-unstimulated 3T3-L1 adipocytes ( Figure 3C). However, EfnB1-siRNA had no effect on Adiponectin mRNA ( Figure 3C).
Experiments designed to examine the effects of overexpression of EFNB1 were performed using adenovirus expressing EfnB1 (Ad-EfnB1) as described in Materials and Methods section. Transfection with Ad-EfnB1 successfully and dose-dependently increased EfnB1 mRNA ( Figure 4A) and EFNB1 protein ( Figure 4B) levels in CAR-3T3-L1 preadipocytes, but did not in 3T3-L1 preadipocytes ( Figure 4A). Following transfection of CAR-3T3-L1 adipocytes with Ad-EfnB1, the cells were cocultured with RAW264.7 cells ( Figure  S2). EfnB1 mRNA ( Figure 4C) and EFNB1 protein ( Figure 4D) was substantially overexpressed in CAR-3T3-L1 adipocytes with or without co-culture with RAW264.7 cells. Figure 4E shows changes in EfnB1 mRNA related to adipose inflammation. High mRNA levels of Mcp-1 and Mcp-3 were noted in CAR-3T3-L1 adipocytes following coculture with RAW264.7 cells, while such increase was markedly suppressed following transfection with Ad-EfnB1. The coculture-related increase in Il-6 mRNA level tended to diminish in CAR-3T3-L1 adipocytes transfected with Ad-EfnB1. Similar to the siRNA study, transfection with Ad-EfnB1 did not alter Adiponectin mRNA level in CAR-3T3-L1 adipocytes.

Discussion
The main findings of the present study were: (1) expression of EFNB1 in WAT and the expression level was significantly low in obese mature adipocytes (2). EFNB1 expression in 3T3-L1 adipocytes was significantly reduced by TNF-α, and markedly suppressed by coculture of 3T3-L1 adipocytes with RAW264.7 cells(3). Adipose EFNB1 exhibited antiinflammatory properties and inhibited monocyte adhesion to adipocytes (4). EFNB1 in adipocytes suppressed TNF-αmediated ERK1/2 activation.
The present study demonstrated for the first time a significantly low level of EFNB1 expression in obese fat tissue. Regulation of EFNB1 has remained uncertain and has not been examined in adipocytes. TNF-α, which is increased in obesity, significantly reduced EfnB1 mRNA expression level ( Figure 2B), suggesting that this cytokine is one of factors accounting for the low expression of EFNB1 in obese adipose tissue. Interestingly, coculture with RAW264.7 macrophages cells definitely suppressed EFNB1 level in adipocytes ( Figure  2C-E). The coculture-related reduction of EFNB1 was larger than TNF-α-mediated decrease in EfnB1, indicating that macrophage-derived undetermined factors can modulate EFNB1 expression in adipocytes.
The Eph-ephrin system is important in neural tissue development, plasticity, and regeneration, immune function, bone homeostasis, and various types of cancers [15]. However, the role of Eph-ephrin system in various metabolic diseases (e.g., diabetes, atherosclerosis, and obesity) remains poorly understood. Konstantinova et al [22] demonstrated that β-cells communicated through EphA-ephrin-A and its signaling controls insulin secretion in response to glucose levels. In other studies, EFNB1 was detected in T-lymphocytes and macrophages in atherosclerotic lesions of the human aorta [23,24]. Interestingly, EFNB1 inhibited MCP-1-dependent monocyte migration, although the role of this phenomenon in atherosclerogenesis remains to be defined. In the present study, EFNB1-overexpression clearly suppressed the elevation of Mcp-1, Mcp-3, and Il-6 mRNA levels in adipocytes when they were cocultured with RAW 264.7 macrophages, while knockdown of EFNB1 increased Mcp-1 in adipocytes. These results suggest that EFNB1 under-expression in obese WAT accelerates the infiltration of monocytes/macrophages into WAT.
Ephrin-B1 and Adipose Inflammation PLOS ONE | www.plosone.org Chronic low-grade inflammation is observed in obese adipose tissue and is closely associated with the metabolic syndrome. Several metabolic stresses, such as stress of the endoplasmic reticulum [25], oxidative stress [26], and hypoxia [27] may induce adipose inflammation and adipocyte dysfunction, which can cause disorders of circulating fatty acids, accumulation of reactive oxygen species, and hyperadipocytokinemia, located upstream in the development of the metabolic syndrome and atherosclerosis [2,28,29]. Importantly, various adipocytes-derived chemokines increase monocyte recruitment into adipose tissue and the MCP-1/ CCR2 pathway is known to play a crucial role in monocyte/ macrophage infiltration into obese adipose tissue [30][31][32], suggesting the involvement of adipose MCP-1 in the prevention of the metabolic syndrome. Inflammatory cytokines, such as TNF-α, interleukin-1β (IL-1β), and transforming growth factor-β (TGF-β), increase MCP-1 through the activation of ERK, nuclear factor-κB (NFκB), and c-Jun N-terminal kinase (JNK) pathway [33][34][35]. In the course of adipocyte hypertrophy, ERK is activated while mitogen-activated protein kinase phosphatase-1 (MKP-1) is inactivated [18]. Inhibition of MKP-1 causes ERK activation and results in the elevation of MCP-1, indicating the significant role of MKP-1 in obese adipocytes [18]. Importantly, the EFNB1 signal suppresses ERK activation in several cells [36,37]. Taken together, our results suggest that adipose EFNB1 regulates MCP-1 expression through the ERK pathway, although further investigation is needed to confirm these findings. Figure 6 provides a model that summarizes the results of the present study. In obese adipose tissue, TNF-α and macrophage-derived undetermined factors suppress EFNB1 expression in adipocytes. Such under-expression results in augmentation of MCP-1 expression, which in turn increases the recruitment of monocytes into adipose tissues. Infiltrated After removal of nonadherent monocytes, the intensity of fluorescence was measured as described in the Materials and Methods section. A, Effect of Ephrin-B1-knockdown on monocyte adhesion to 3T3-L1 adipocytes. 3T3-L1 adipocytes were transfected with siRNA for Ephrin-B1 (EfnB1) or negative control (Cont) and then the cells were incubated in the absence or presence of tumor necrosis factor-α (TNF-α) at 1 ng/mL for 24 hrs prior to adherent assay. n=6 for each group. B, Effect of overexpression of Ephrin-B1 on monocyte adhesion to CAR-3T3-L1 adipocytes. CAR-3T3-L1 adipocytes were transfected with an adenovirus expressing Ephrin-B1 (Ad-EfnB1) or β-galactosidase (Ad-βgal). The transfected adipocytes were cocultured with RAW264.7 cells for 24 hrs before the adhesion assay. n=6 for each group. C, Representative blots of ERK signal in the adenovirus study. Western blotting was performed using antibodies against Ephrin-B1 (EFNB1), α-tubulin, phosphorylated extracellular signal-regulated kinase 1/2 (P-ERK1/2), and total extracellular signal-regulated kinase 1/2 (T-ERK1/2). After 24 hrs of transfection with Ad-EfnB1 or Ad-βgal, CAR-3T3-L1 adipocytes were stimulated with or without 1 ng/mL of TNF-α for 5 min and subjected to immunoblotting. n=3 for each group. D, Phosphorylation of ERK1/2. The relative signal intensity for P-ERK1/2 and T-ERK1/2 was calculated by densitometry and expressed as phosphorylation of ERK. n=3 for each group. EfnB1 and EFNB1, Ephrin-B1; CAR-3T3-L1, 3T3-L1 cells stably expressing Coxsackie-Adenovirus Receptor; TNF-α, tumor necrosis factor-α. Values are mean±SD. **P<0.01. #P<0.05; # #P<0.01; # # #P<0.001, compared to TNF-α (-) or coculture (-).