Disclosure: The authors declare no conflict of interest.
Version of Record online: 25 MAY 2013
Copyright © 2012 The Obesity Society
Volume 21, Issue 4, pages 731–736, April 2013
How to Cite
Miyata, Y., Fukuhara, A., Otsuki, M. and Shimomura, I. (2013), Expression of activating transcription factor 2 in inflammatory macrophages in obese adipose tissue. Obesity, 21: 731–736. doi: 10.1002/oby.20274
See the online ICMJE Conflict of Interest Forms for this article.
- Issue online: 25 MAY 2013
- Version of Record online: 25 MAY 2013
- Manuscript Accepted: 16 MAY 2012
- Manuscript Received: 9 APR 2011
White adipose tissue (WAT) of obesity is in the state of inflammation with progressive infiltration by macrophages and overproduction of reactive oxygen species (ROS), which can induce WAT dysfunction, including insulin resistance and adipocytokine dysregulation. Activating transcription factor 2 (ATF2) is a member of the ATF/cAMP response element binding family of transcription factors and known to be activated by cellular stressors, such as inflammatory cytokines, lipopolysaccharide (LPS), and ROS.
Design and Methods, Results:
Here, we show that ATF2 protein was significantly more induced in WAT of ob/ob mice compared with C57BL/6J mice. Total and phosphorylated ATF2 were highly expressed in infiltrated macrophages. Furthermore, flow cytometry analysis demonstrated that ATF2 expression was high in CD11c-positive/CD301-negative M1 macrophages. Phosphorylation of ATF2 was induced by treatment with either H2O2 or LPS in RAW264.7 macrophage cells, and suppression of ATF2 expression by small-interfering RNA induced mRNA levels of ATF3, an anti-inflammatory molecule in macrophages in WAT.
These results suggest that ATF2 is an important transcriptional factor relating to inflammation through the suppression of ATF3 in M1 macrophages of WAT.
The inflammatory state of white adipose tissue (WAT) in obesity has been highlighted in recent reports demonstrating extensive infiltration of macrophages in the tissue of obese human and rodent models (1,2). The massive accumulation of macrophages can potentially lead to a considerable amplification of the inflammatory state in WAT. Recent studies indicated that the macrophages in WAT consist of at least two different phenotypes; the classically activated M1 macrophages and alternatively activated M2 macrophages (3,4). M1 macrophages in WAT produce proinflammatory cytokines, thus contributing to the induction of insulin resistance (3,4). Conversely, M2 macrophages, representing the major resident macrophages in lean adipose tissue, are reported to have protective effects against inflammation and insulin resistance in WAT (4,5). Thus, it is conceivable that the obesity-associated infiltration of immunocytes, especially M1 macrophages, is one of the major causes of WAT dysfunction.
Adipose tissues of obese mice are exposed to oxidative and hypoxic stress. Systemic markers of oxidative stress increase with adiposity, consistent with the role of reactive oxygen species (ROS) in the development of obesity-induced insulin resistance (6). Furthermore, ROS is also known to induce the expression of various genes of proinflammatory cytokines including tumor necrosis factor-α, interleukin-6, and monocyte chemotactic protein–1. Experimental evidence also indicates that the adipose tissue of obese mice is persistently hypoxic and underperfused (7). Fat hypoxia induces endoplasmic reticulum stress and results in induction of proinflammatory cytokine gene expression, decrease in insulin sensitivity, and dysregulation of adipocytokine production in WAT (7,8). In WAT from obese subjects, hypoxia-inducible factor-1α (HIF-1α) is also activated and elicits fat fibrosis through lysyl oxidase gene expression (9). Moreover, hypoxia causes increased ROS production in adipocytes (10), whereas inflammation and ROS stabilize hypoxia-inducible factor-1α protein (11-13). Hence, a vicious cycle among fat inflammation, fat ROS, and fat hypoxia is present in WAT of obese subjects.
Activating transcription factor 2 (ATF2) is a member of the ATF/cAMP response element binding family of transcription factors and possess a bZIP DNA-binding domain (14,15). ATF2 forms homo and heterodimers with other members of the bZIP family and binds to the cyclic adenosine monophosphate response element. Transcriptional activation of ATF2 requires phosphorylation, which is mediated by p38, Jun N-terminal protein kinase (JNK), and extracellular signal-regulated kinase (ERK). In response to various stresses including inflammation, ROS, and hypoxia, Jun N-terminal protein kinase, extracellular signal-regulated kinase, and p38 phosphorylate ATF2 and enhance its trans-activating capacity (14,15).
ATF2 is expressed in WAT, and plays a role in adipocyte differentiation and fat storage (16,17). However, pathophysiological roles of ATF2 remain unknown in adipose tissue of obese subjects. This study identified for the first time that ATF2 was overexpressed and phosphorylated especially in infiltrated macrophages in adipose tissue of obese mice. Phosphorylation of ATF2 was induced by treatment with either H2O2 or lipopolysaccharide (LPS) in RAW264.7 macrophage cells, and suppression of ATF2 expression by small-interfering RNA (siRNA) induced mRNA levels of ATF3, an anti-inflammatory molecule in macrophages in WAT. Taken together, we suggested that ATF2 is suggested to be an important transcriptional factor regulating inflammation through ATF3 in M1 macrophages of WAT.
Materials and Methods
Antibodies and other reagents
Anti-ATF2 polyclonal antibody was purchased from Santa Cruz Biotechnology (sc-187; Santa Cruz, CA). Antiphospho-ATF2 polyclonal antibody and anti-ATF2 monoclonal antibody was purchased from Cell Signaling Technology (Beverly, MA). Other reagents were purchased from Sigma-Aldrich (St Louis, MO).
Male 9- to 10-week-old ob/ob mice, db/db mice, and 21 week diet-induced obese mice were purchased from Charles River Labs (Yokohama, Japan). Mice were housed in groups of two per cage, maintained in a room under controlled temperature (23 ± 1.5°C) and humidity (45 ± 15%) on a 12-h dark/12-h light cycle, and had free access to water and chow (MF; Oriental Yeast, Tokyo, Japan). The experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of Osaka University, Graduate School of Medicine.
Detection of ATF2 expression in adipose tissue and liver
Whole epidermal adipose tissue and liver were lysed in ice-cold lysis buffer (20 mmol/l Tris-HCl pH 7.2, 1 mmol/l EGTA, 1% Triton X-100, 150 mmol/l NaCl, 100 mmol/l NaF) containing both protease inhibitors (100 µg/ml phenylmethylsulfonyl fluoride, 2 µg/ml aprotinin, and 10 µg/ml leupeptin) and phosphatase inhibitor cocktail (Nacalai Tesque, Kyoto, Japan). The protein concentration of the lysates was determined by bicinchoninic acid assay (Pierce, Rockford, IL). Next, 1.5 mg (adipose tissue) or 3 mg (liver) of proteins was incubated with 5 µg of anti-ATF2 antibodies (sc-187; Santa Cruz Biotechnology) for 8-12 h at 4°C. Immunocomplexes were collected by protein G-Sepharose (GE Healthcare, Piscataway, NJ) and washed three times with lysis buffer, followed by SDS-PAGE and immunoblot analysis with anti-ATF2 antibody (sc-187; Santa Cruz Biotechnology).
Fractioning of adipose tissue
Adipocytes and stromal-vascular fractions were isolated as described earlier (18). Epididymal fat pads from male ob/ob mice were excised and minced in Krebs-Ringer-bicarbonate-HEPES (N-2-hydroxyethylpiperazine-N′-ethanesulfonic acid) buffer which contained 120 mmol/l NaCl, 4 mmol/l KH2PO4, 1 mmol/l MgSO4, 1 mmol/l CaCl2, 10 mmol/l NaHCO3, 30 mmol/l HEPES, 20 µmol/l adenosine, and 4% (wt/vol) bovine serum albumin (Calbiochem, San Diego, CA). Tissue suspensions were centrifuged for 2 min to remove erythrocytes and free leukocytes. Collagenase was added to the final concentration of 1 mg/ml and incubated at 37°C for 30 min with shaking. The cell suspension was filtered through a 110-µm filter and then spun for 1 min to separate floating mature adipocytes fraction from the stromal-vascular fractions pellet. This fractioning and washing procedure were repeated twice with Krebs-Ringer-bicarbonate-HEPES buffer. Finally, both fractions were washed with phosphate-buffered saline. Cells from each fraction were used for protein extraction or flow cytometry. The pooled WAT from three ob/ob mice was used to detect ATF2 protein in each fraction.
WAT was removed immediately, fixed in 10% neutral buffered formalin for 24-48 h, and then processed into paraffin blocks. After deparaffinization and hydration, the sections were autoclaved for 15 min at 121° in 10 mmol/l Citrate buffer (pH 6.0). Then, to inactivate endogenous peroxidases, the sections were treated with 1% H2O2 at room temperature (RT) for 30 min, followed by incubation with ATF2 antibody (1:50) or phospho-ATF2 antibody (1:40) for 60 min. To detect the primary antibody, the slides were reacted with the EnVision kit (Dako, Glostrup, Denmark). Next, the ATF2-stained sections were rinsed and incubated with Mac-2 antibody (1:3800) for 60 min at RT as described earlier (19). Then, the slides were also incubated with alkaline phosphatase-conjugated antirat IgG (Santa Cruz Biotechnology) (1:100) for 30 min at RT, and reacted with Fast Red reagent (Dako). Primary antibodies were diluted with Antibody Diluent (Dako) and alkaline phosphatase-conjugated antirat IgG was diluted with Antibody Diluent (Dako) mixed with normal mouse serum (Dako) (1:20).
Fluorescence-activated cell sorting analysis was carried out as described earlier (4,20,21). Briefly, cells in the stromal-vascular fractions were suspended in flow cytometry stain buffer (eBiosciences, San Diego, CA), and incubated with 2.4G2 (BD Pharmingen, San Diego, CA) for 15 min. Then, the cells were rinsed and resuspended in flow cytometry stain buffer and stained with FITC-conjugated anti-F4/80 (eBiosciences), phycoerythrin-Cy7-conjugated anti-CD11c (eBiosciences), and Alexa Fluor 647-conjugated anti-CD301 (AbD serotec, Oxford, UK) antibodies or the matching control isotypes for 30 min at 4°C. Next, the cells were fixed with 2% paraformaldehyde (Nacalai Tesque) in phosphate-buffered saline for 10 min at 37°C, and chilled on ice for 1 min. Paraformaldehyde was removed by centrifugation, and the pelletized cells were resuspended in ice-cold 90% methanol and incubated for 30 min at 4°C. Subsequently, the cells were rinsed and incubated with anti-ATF2 monoclonal antibody or normal rabbit IgG (Cell Signaling Technology) for 30 min at RT. Then, the cells were rinsed and incubated with phycoerythrin-conjugated antirabbit IgG antibodies (Imgenex, San Diego, CA) for 30 min at RT. After the reaction, the cells were rinsed and resuspended in flow cytometry stain buffer and analyzed using a FACSAria cell sorter (BD Biosciences).
Mouse peritoneal macrophages
Murine peritoneal macrophages from C57BL/6J (Charles River Labs) were prepared as described earlier (22), and treated with 10 ng/ml LPS or 500 µmol/l H2O2 for 30 min. Cell lysates were subjected to SDS-PAGE and immunoblot analysis with anti-ATF2 monoclonal antibody and phospho-ATF2 polyclonal antibody.
RAW264.7 cells (ATCC) were maintained in Dulbecco's modified Eagle medium (Nacalai Tesque) containing 10% fetal bovine serum and 3.75 × 105 cells per 1 well of a 6-well plate were transfected with ATF2 siRNA (Qiagen, Valencia, CA) (25 nmol/l) using 7.5 µl Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA). Twenty-four hours later, the cells were treated with 10 ng/ml LPS for 30 min, and subjected to SDS-PAGE and immunoblot analysis with the indicated Abs. To determine the effects of siRNA on SOCS3 and ATF3 gene expression, total RNA was extracted from the cells after 2-h treatment with LPS, and the mRNA expression was measured by real-time RT-PCR.
RNA extraction and quantitative real-time PCR
Total RNA was extracted from RAW264.7 cells using RNA STAT-60 Reagent (Tel-Test, Friendswood, TX). First-strand cDNA was synthesized from total RNA using the Transcriptor First Strand cDNA Synthesis Kit (Roche). Real-time PCR was performed on LightCycler system using the FastStart DNA Master SYBR Green I (Roche, Indianapolis, IN), and values are normalized to the level of 36B4 mRNA. Sequences of primers used for real-time PCR were the following: ATF2, 5′-CAG TGG ATT GGT TAG GAC TCA GTC A-3′ and 5′-GAG GAG TTG TGT GAG CTG GAG AAG-3′; SOCS3, 5′-CTG CAG GAG AGC GGA TTC TA-3′ and 5′-GGT CTT GAC GCT CAA CGT GA-3′; ATF3, 5′-GCC AAG TGT CGA AAC AAG AAA AAG-3′ and 5′-TCC TCG ATC TGG GCC TTC AG-3′; 36B4, 5′-GCT CCA AGC AGA TGC AGC A-3′ and 5′-CCG GAT GTG AGG CAG CAG-3′.
All data were expressed as mean ± s.d. Differences between groups were examined for statistical significance using the Student's t-test. A P value ≤ 0.05 denoted the presence of a statistically significant difference.
ATF2 protein expression in WAT of obese mice
The expression level of ATF2 protein was estimated in WAT and liver of control C57BL/6J or obese ob/ob mice. ATF2 protein and mRNA expression was significantly elevated in WAT from ob/ob mice, but not in liver (Figure 1a, Supplementary Figure S1 online), compared with the control mice. To further investigate ATF2 expression in WAT of ob/ob mice, adipose tissue was fractionated into the mature adipocyte fraction, and stromal-vascular fraction, containing preadipocytes, blood-derived cells, and endothelial cells. ATF2 protein expression was almost limited to the stromal-vascular fraction (Figure 1b).
ATF2 expression in macrophages of WAT of ob/ob mice
To determine the localization of cells expressing ATF2, WAT from ob/ob mice was subjected to immunohistochemical analysis. Cells positive for Mac-2, a surface marker of activated macrophages, formed crown-like structures (Figure 2a, red cells, upper panel) (19), and ATF2 was also stained in these cells (Figure 2a, brown staining in the nucleus, upper panel), suggesting the expression of ATF2 in macrophages of ob/ob mice. An earlier study reported that ATF2 was activated by phosphorylation (14). Immunostaining with antiphosphorylated ATF2 antibody demonstrated activation of ATF2 in these cells (Figure 2a, brown staining in the nucleus, bottom panel). Similar results were noted in db/db mice and diet-induced obese mice (Supplementary Figure S2 online). Adipose tissue macrophages consist of at least two different phenotypes; the classically activated M1 macrophages and alternatively activated M2 macrophages (3,4). The CD11c-positive M1 macrophages and the CD301-positive M2 macrophages in WAT were separated using flow cytometry as reported earlier (4), and expression level of ATF2 protein was measured by FACS analysis. Between these two macrophage subpopulations, the mean fluorescence intensity of ATF2 was significantly higher in CD11c-positive M1 macrophages than in CD301-positive M2 macrophages (Figure 2b, and Supplementary Figure S2b online).
Induction of ATF2 phosphorylation in RAW264.7 macrophage cells and mouse primary macrophages
Earlier report indicated that various stresses including inflammation, ROS, and hypoxia induce ATF2 phosphorylation (14,15), and WAT of obese mice is exposed to inflammation, ROS, and hypoxia (6,7,23). Thus, to determine the possible upstream inducer of ATF2 phosphorylation in macrophage cells, we treated RAW 264.7 cells with LPS and H2O2 or incubated under hypoxia. In accordance with earlier report (22), we confirmed that treatment with LPS enhanced ATF2 phosphorylation (Figure 3a). In addition, treatment with H2O2 also induced ATF2 phosphorylation, however, incubation under hypoxic condition (1% O2) did not (Figure 3a). Furthermore, also in mouse peritoneal macrophages, LPS and H2O2 promoted ATF2 phosphorylation (Figure 3b).
Effects of ATF2 knockdown in macrophages
Overexpression of ATF2 caused a significant decrease in cell survival (ref.24 and data not shown), therefore, we investigated the role of ATF2 in macrophage cells using experimental RNAi. To define the function of ATF2 in macrophages, endogenous ATF2 expression was suppressed using siRNA in RAW 264.7 cells. Treatment with LPS induced phosphorylation of ATF2 (Figures 3 and 4a), and ATF2-specific siRNA suppressed the expression levels of both total and phosphorylated ATF2 and ATF2 mRNA compared with the control siRNA (Figure 4a, and Supplementary Figure S3 online). Under these conditions, we measured the mRNA levels of SOCS3 gene, which is known as the direct target of ATF2 (22). SOCS3 mRNA levels were reduced in RAW264.7 cells transfected with ATF2 siRNA compared with the control siRNA (Supplementary Figure S3 online). We newly found that knockdown of ATF2 increased mRNA levels of ATF3 (Figure 4b). This effect was also observed in RAW264.7 cells treated with LPS (Figure 4b). These results suggest that ATF2 might downregulate ATF3 mRNA expression in macrophages.
ATF2 is ubiquitously expressed in tissues (25). Earlier reports showed that ATF2 expression levels were increased in early stage of adipocyte differentiation and decreased in the later stage (26). Moreover, Maekawa et al. generated the mice with mutations in both ATF2 and CRE-BPa, an ATF2-related gene, both of which belong to the ATF/cAMP response element binding family of transcription factors, and concluded that ATF2 and CRE-BPa are necessary for induction of PPARγ, a key transcription factor mediating adipocyte differentiation (16). However, pathophysiological roles of ATF2 remain unknown in WAT of obesity. In this study, for the first time, we found overexpression of ATF2 in adipose tissue of ob/ob mice compared with control mice. In WAT from ob/ob mice, ATF2 was stained mainly in macrophages, with localization in the crown-like structures, and FACS analysis demonstrated that the expression levels of ATF2 protein were higher in F4/80-positive macrophages, particularly in CD11c-positive cells, than in F4/80-negative cells (Figure 2b and data not shown). These results indicate that high-level expression of ATF2 protein results from infiltrating macrophages in WAT of ob/ob mice.
Our study showed for the first time that ATF2 was phosphorylated in adipose tissue of ob/ob mice, and that oxidative stress induced phosphorylation of ATF2 in RAW264.7 macrophages. With regard to oxidative stress, we reported earlier that fat accumulation correlated with systemic oxidative stress, and that ROS production was increased in adipose tissue of obese mice, accompanied by augmented expression of NADPH oxidase and decreased expression of antioxidative enzymes (6). Therefore, oxidative stress seems to be involved in ATF2 phosphorylation in adipose tissue of ob/ob mice. With regard to the LPS signaling, serum endotoxin levels were elevated in obese animals. For example, ob/ob and db/db mice showed lower intestinal resistance and higher portal endotoxemia compared with lean control mice (27), and in humans, high-fat diet increased plasma LPS concentration two to three times (28). Such metabolic endotoxemia was suggested to be involved in inflammation and insulin resistance (28). Another activator of LPS signaling pathway is endogenous saturated free fatty acids (29). Palmitic and stearic fatty acids activate toll-like receptor-4 similar to LPS (30). WAT in obesity is lipolytically active (31), and large quantities of free fatty acids should be released from adipocytes. Taken together, LPS signaling, including endogenous endotoxin and free fatty acids, and ROS should be involved in the induction of ATF2 activation in WAT of obese mice.
We demonstrated that suppression of ATF2 by siRNA resulted in upregulation of ATF3 expression in RAW264.7 cells, indicating that ATF2 is a negative regulator of ATF3. The transcriptionally suppressive effect of ATF2 was reported in thrombomodulin promoter. ATF2 binds to the thrombomodulin promoter, and recruits HDAC4 and form a transcriptional repression complex in the promoter (32). However, several studies indicated that ATF3 is one of the target genes of ATF2 (33-35). ATF2 binds to the promoter region of the ATF3 gene and enhances ATF3 gene expression in fibroblasts (33) and gonadotroph cells (35). The mechanism of this opposite effect of ATF2 on ATF3 transcription is unknown, however, a recent study indicated that ATF3 gene has two functional promoter regions containing the ATF2 binding motif, and these alternate promoters are differentially used at both transcriptional and translational levels (36). Accordingly, in macrophages, ATF2 might suppress ATF3 expression through usage of different promoter or recruitment of other transcriptional complexes.
Recently, Suganami et al. reported that ATF3 can transcriptionally repress the production of tumor necrosis factor-α in RAW264.7 macrophages, and transgenic overexpression of ATF3 specifically in macrophages attenuated proinflammatory M1 macrophage activation in the adipose tissue of obese mice (37). Thus, ATF3 acts as transcriptional repressor of LPS signal in macrophages in WAT. This study demonstrated that ATF2 is highly expressed and activated in CD11c-positive/CD301-negative macrophages in WAT of ob/ob (Figure 2), and that suppression of ATF2 by siRNA resulted in upregulation of ATF3 expression in RAW264.7 cells (Figure 4). Our data support Suganami's conclusion, and we suggest that ATF2 is one of the upstream molecules of ATF3 in macrophages. Collectively, we assume that the large amount of ATF2 protein in M1 macrophages of WAT should relate to suppressing ATF3 gene expression, which results in a proinflammatory state in obese adipose tissue.
We showed the functional role of ATF2 in macrophage cells through siRNA-mediated knockdown in RAW264.7 cells. Indeed, ATF2-overexpression experiments are the best way to mimic the macrophages in WAT of obese mice. However, earlier reports indicated that overexpression of ATF2 caused a decrease in cell survival in rat pheochromocytoma PC12 cells (24), and activation of ATF2 is implicated in apoptosis in leukemia cell lines (38). In accordance, our experimental data demonstrated that ATF2 overexpression caused cell damage in RAW264.7 cells (data not shown). In this article, our data indicated that ATF2 is associated with macrophage inflammation, and we speculate that RAW264.7 cells overexpressing ATF2 should induce excessive inflammation resulting in cell death. Taken together, we concluded the inflammatory role of ATF2 in macrophages through loss of function study.
In summary, this study demonstrated that ATF2 is induced and activated in M1 macrophages in adipose tissue of obese subjects. ATF2 suppressed ATF3 expression, and resulted in inflammatory state. ATF2 might be an important regulator of adipose tissue inflammation.
We thank Haruyo Sakamoto for technical help. This work was supported by a Grant-in-Aid for JSPS fellow.