Serum Amyloid A3 Gene Expression in Adipocytes is an Indicator of the Interaction with Macrophages

The infiltration of macrophages into adipose tissue and their interaction with adipocytes are essential for the chronic low-grade inflammation of obese adipose tissue. In this study, we identified the serum amyloid A3 (Saa3) gene as a key adipocyte-derived factor that is affected by interaction with macrophages. We showed that the Saa3 promoter in adipocytes actually responds to activated macrophages in a co-culture system. Decreasing C/EBPβ abundance in 3T3-L1 adipocytes or point mutation of C/EBPβ elements suppressed the increased promoter activity in response to activated macrophages, suggesting an essential role of C/EBPβ in Saa3 promoter activation. Bioluminescence based on Saa3 promoter activity in Saa3-luc mice was promoted in obese adipose tissue, showing that Saa3 promoter activity is most likely related to macrophage infiltration. This study suggests that the level of expression of the Saa3 gene could be utilized for the number of infiltrated macrophages in obese adipose tissue.

Obesity, in particular excess visceral adiposity, and obesity-related metabolic disorders have emerged as crucial health issues worldwide. These metabolic disorders are collectively referred to as metabolic syndrome and increase the risk of developing type 2 diabetes, cardiovascular disease, and cancers 1,2 . Overnutrition, involving excess nutrient intake that exceeds energy expenditure, and a sedentary lifestyle contribute to the aberrant accumulation of white adipose tissue through adipocyte hyperplasia and hypertrophy, leading to pathological changes of adipose functions 1,2 . Previous reports have clearly showed that expanding adipose tissue is well characterized as a chronic state of low-grade inflammation that is strongly linked to alterations of cellular composition and the abnormal production of cytokines, including tumor necrosis factor α (TNF-α ), interleukin (IL)-1, and monocyte chemoattractant protein-1 (MCP-1/CCL2) [2][3][4][5][6] .
More recent research has indicated that obesity induces the accumulation of immune cells, including macrophages, T cells, and B cells, in adipose tissue [4][5][6][7][8][9][10][11] . In particular, macrophage infiltration into adipose tissue is an essential event for fat inflammation and systemic metabolic disorders because the number of macrophages in adipose tissue correlates with the degree of obesity, and is linked to adipose inflammation and insulin resistance 4,5,7,11 . In fact, a lack of MCP-1/CCL2 or its receptor, CC motif chemokine receptor-2 (CCR-2), was shown to reduce macrophage accumulation in obese adipose tissue and fat inflammation, and to improve systemic insulin sensitivity in high-fat diet (HFD) fed mice compared with wild-type mice 12,13 . Moreover, weight loss induced by surgery or exercise and conditional ablation of macrophages using the CD11c promoter resulted in a reduction in the number of adipose tissue macrophages of obese mice in parallel to the decreased production of pro-inflammatory cytokines and suppressed systemic insulin resistance [14][15][16] . In particular, recent works have strongly suggested that crosstalk between adipocytes and infiltrated macrophages plays a crucial role in chronic adipose inflammation and adipose remodeling 10,[17][18][19][20][21][22] . Hence, determination of the number of infiltrated macrophages and in vivo evaluation of their interaction with adipocytes are essential for assessment of the chronic inflammatory state in obese adipose tissue.
Our recent study showed that, in mice supplemented with an anti-inflammatory vitamin, vitamin B6, macrophage marker gene expression and macrophage infiltration in obese adipose tissue were specifically down-regulated 20 . Furthermore, we performed a comparative analysis on the gene expression of obese adipose tissue of db/db mice and with vitamin B6 supplementation. Using this approach and by obtaining transcriptomic data, candidate genes related to the increase in infiltrated macrophages in adipose tissue in vivo were identified 20,22 . In this study, because activated macrophages could spread systemically in obese mice, we focused on adipocyte-derived genes that are differentially regulated by interaction with activated macrophages in order to establish a new non-invasive in vivo model for evaluation of the adipose inflammatory state associated with macrophage infiltration. We isolated the adipocyte-derived serum amyloid A3 (Saa3) gene for the monitoring of macrophage infiltration in adipose tissue. We characterized the mouse Saa3 promoter region upstream of the luciferase (Saa3-luc) reporter chimeric gene in order to monitor macrophage infiltration. Finally, we generated Saa3-luc transgenic mice (Saa3-luc mice) and subjected them to in vivo luminescence imaging analysis upon being fed an HFD. In vivo assessment of the chronic low-grade inflammatory state of obese fat by determining macrophage infiltration would be important to evaluate the preventive effect of pharmaceutical agents and functional food against obesity-related disorders.

Saa3 gene expression in adipocytes is up-regulated by interaction with infiltrated macrophages.
To analyze the changes in gene expression in adipose tissue during the development of obesity, this study identified a gene cluster including 402 genes whose expression was shown by DNA microarray analysis to be fivefold higher in the adipose tissue of db/db mice than in that of db/+ mice. Moreover, we utilized an in vitro co-culture system using 3T3-L1 adipocytes and RAW264.7 macrophages and performed gene expression analysis on differentiated 3T3-L1 adipocytes cultured with stimulated or unstimulated RAW264.7 macrophages. We found that 224 adipocyte genes were significantly up-regulated more than fivefold upon co-culture with macrophages activated by lipopolysaccharide (LPS). Then, we carried out a comparative analysis based on these in vivo and in vitro transcriptomic data in order to identify in vivo adipocyte-derived genes that were possibly affected by infiltrated macrophages in obese adipose tissue (Fig. 1A). In this study, as shown in Table 1, we isolated 15 genes whose expression was significantly up-regulated more than fivefold in both in vivo and in vitro gene expression profiles. We further characterized the expression of these 15 genes and showed that the mRNA expression levels of three genes (Saa3, Timp1, and Il1rn) were specifically increased in obese adipose tissue of db/db mice (Fig. 1B). These observations indicated that these three genes can be considered to be key adipocyte-derived factors whose expression is up-regulated by interaction with macrophages and potentially available as indicators of the number of macrophages in obese adipose tissue. Furthermore, we selected the Saa3 gene for the observation of macrophage infiltration into adipose tissue because Saa3 gene expression in adipocytes was increased upon co-culture with macrophages with the strongest transcriptional induction (by 72-fold) among the three genes. As shown in Fig. 2A, Saa3 gene expression in adipose tissue was significantly increased in db/db mice compared with that in db/+ mice as well as mRNA of macrophage marker genes such as Emr1 and Mpeg1. We also performed quantitative RT-PCR analysis of the expression level of the Saa3 gene in adipose tissue of mice with diet-induced obesity. In mice fed an HFD for 16 weeks, epididymal white fat mass was greater with the increased expression of the Emr1 and Mpeg1 mRNAs, whereas the Saa3 mRNA level in adipose tissue was significantly up-regulated (11-fold) in obese mice with an HFD compared with that in non-obese mice (Fig. 2B). To determine whether Saa3 expression levels in obese adipose tissue were correlated with macrophage infiltration, we measured Saa3 and Emr1 mRNA expression levels and F4/80-positive cell number in the epididymal white adipose tissue of individual mice fed an HFD and showed significant correlations between Saa3 mRNA level and the number of F4/80-positive cells (r = 0.565, p < 0.05) (Fig. 2C) and Emr1 mRNA level (r = 0.471, p < 0.05) ( Supplementary Fig. 1A) by using Pearson's correlation coefficient analyses. We further performed time course analyses of mRNA expression in adipose tissue of mice fed an HFD and observed that the upregulation of Saa3 mRNA expression is paralleled by an increase in the Emr1 mRNA expression ( Supplementary Fig. 1B). Next, to determine the adipose tissue cell types that were responsible for these mRNA changes, we isolated mature adipocytes from epididymal adipose tissue by collagenase digestion. The expression of Saa3 mRNA in the adipocyte fractions isolated from adipose tissue of db/db mice was up-regulated compared with that in db/+ mice (Fig. 2D). Similarly, we showed that Saa3 mRNA expression increased in the adipocyte fraction isolated from mice fed an HFD (Fig. 2D). We further examined Saa3 mRNA expression in adipocyte and macrophage cell lines and several types of cells in obese adipose tissue and compared mRNA expression profiles among the 15 candidate genes, suggesting that Saa3 mRNA is expressed in both adipocytes and macrophages (Supplementary Fig. 2A). As shown in Supplementary Fig. 2B, we observed that Saa3 mRNA is expressed in activated macrophage RAW264.7 cells and bone marrow-derived M1 macrophages. Further immunohistochemical analyses demonstrated that Saa3 protein is observed in F4/80-positive cells (macrophages) and is also seen within vesicle-like structures in adipocytes (Fig. 2E). Taken together, these observations suggest that Saa3 gene expression in mature adipocytes responds to macrophages and that M1 macrophages were partially responsible for the Saa3 mRNA in obese adipose tissue in vivo.
Saa3 promoter activity is up-regulated in the presence of activated macrophages. The levels of Saa3 mRNA in 3T3-L1 adipocytes were strikingly up-regulated upon co-culture with activated macrophages (Fig. 3A). The expression of Saa3 mRNA in adipocytes was not affected by the presence of unstimulated macrophages or of LPS alone. Moreover, when differentiated 3T3-L1 adipocytes were treated with conditioned medium of macrophages activated by LPS (MacCM), Saa3 mRNA expression level was significantly increased (Fig. 3B). Previous reports have shown that TNF-α secreted by activated macrophages is involved in the inflammatory response in obese adipose tissue, which is linked to obesity-related metabolic disorders including insulin The venn diagram shows genes that are upregulated in the db/db white adipose tissue and up-regulated in 3T3-L1 adipocytes by the interaction with activated RAW264.7 macrophages. Of a total 402 genes up-regulated in db/db adipose tissue, the expression of 15 genes was also increased in 3T3-L1 adipocytes by the interaction with activated macrophages (p < 0.05). (B) Total RNAs from several types of tissue of db/+ and db/db mice were isolated. The relative mRNA expression level of each gene was determined by quantitative PCR and normalized to L19 mRNA level. The data are representative of two independent experiments. resistance 11,17 . We showed here that exogenously added TNF-α increased Saa3 gene expression in mature 3T3-L1 adipocytes (Fig. 3B), suggesting that TNF-α is one of potent inflammatory mediators secreted by activated macrophages to up-regulate Saa3 mRNA expression in lipid-laden adipocytes.
Previously, Geurts et al. screened synovial genes whose expression is significantly up-regulated during the course of rheumatoid arthritis 23 . The Saa3 promoter (− 314/+ 50) showed strong transcriptional induction in synovial fibroblasts in rheumatoid arthritis, suggesting that this promoter region is involved in the response to the inflammatory process in this disease 23,24 . Here, to examine whether the Saa3 promoter activity can respond to activated macrophages, we constructed an Saa3 promoter-luciferase chimeric gene (Saa3-luc) (Fig. 3C) and infected 3T3-L1 adipocytes with retroviruses harboring this gene (3T3-L1/Saa3-luc adipocytes). Next, we constructed a co-culture system composed of 3T3-L1/Saa3-luc adipocytes and RAW264.7 macrophages, and tested the Saa3 promoter activity in the adipocytes. Luciferase activity in the 3T3-L1 adipocytes was significantly elevated upon co-culture with macrophages activated by LPS treatment (Fig. 3D). In order to examine the effect of adherent RAW264.7 macrophages on the Saa3 promoter activity in adipocytes, we performed the co-culture of 3T3-L1 adipocytes and RAW264.7 cells in a contact system. RAW264.7 macrophages were directly plated onto 3T3-L1 adipocytes; the SAA3 promoter activity in adipocytes was examined ( Supplementary Fig. 3), and was found to be up-regulated to the same extent as in the trans-well system, suggesting that the increase in the SAA3 promoter activity is mainly dependent on the secretion of factors such as TNF-α from macrophages. Moreover, we examined the effect of the medium conditioned with activated RAW264.7 macrophages on the Saa3 promoter activity in 3T3-L1 adipocytes. The addition of this conditioned medium markedly increased luciferase activity in 3T3-L1 cells (Fig. 3E). In this study, we further showed that the Saa3 promoter in mature 3T3-L1 adipocytes can respond to exogenously added TNF-α (Fig. 3F), suggesting that such promoter activity is partially up-regulated by TNF-α secreted by activated RAW264.7 macrophages. In addition, we analyzed TNF-α mRNA level in adipose tissue and tested its correlation with Saa3 mRNA. In fact, there was significant correlation between Saa3 mRNA level and TNF-α mRNA level (r = 0.637, p < 0.05) (Fig. 3G) by using Pearson's correlation coefficient analysis. Although previous reports have shown that Saa3 mRNA expression in adipocytes is up-regulated in response to saturated fatty acids 7,25 , the addition of palmitic acid conjugated with fatty acid-free BSA did not affect the Saa3 promoter activity ( Supplementary Fig. 4).
In this study, we examined paracrine effects of secreted Saa3 on adipocytes. A previous work by Hiratsuka et al. 26 demonstrated that Saa3 exerts its biological effects via its function as a ligand for the TLR4 receptor. We first analyzed the mRNA expression of the TLR4 receptor in 3T3-L1 cells and mature adipocytes derived from adipose tissues and RAW264.7 macrophages. The mRNA expression of TLR2 and TLR4 receptors in adipocytes was extremely low ( Supplementary Fig. 5A), suggesting that Saa3 can not affect adipocyte functions. We further examined the biological effects of secreted Saa3 on adipocytes. Mouse Saa3 was overexpressed in HEK293 cells, and the condition medium was prepared from these HEK293 cells. This condition medium enhanced the phosphorylation of p65 (NF-κ B) and extracellular signal-regulated kinase 1/2 proteins in RAW264.7 macrophages, which express TLR4 receptor; however, similar stimulation did not occur in differentiated adipocytes ( Supplementary Fig. 5B).

C/EBPβ is involved in the responsiveness of the Saa3 promoter to activated macrophages.
Three CCAAT/enhancer-binding protein β (C/EBPβ )-binding sites are present in the mouse Saa3 region (− 152, −107, and − 77) 24,27 . Because previous studies showed that C/EBPβ responds to inflammatory stimuli including IL-1  Table 1. Analysis of two transcriptomes to isolate genes whose expression is upregulated by the interaction with macrophages. DNA microarray analysis was repeated with the Cy3 and Cy5 dyes reversed (a dye swap). Fold change (Fold 1) represents the average of mRNA expression level in db/db mice relative to db/+ mice. Fold change (Fold 2) represents the average of mRNA expression level in the 3T3-L1 adipocytes upon co-culture with macrophages in the presence of LPS relative to with macrophages in the absence of LPS.
Scientific RepoRts | 6:38697 | DOI: 10.1038/srep38697  and TNF-α , we tested the role of C/EBPβ in the responsiveness of the Saa3 promoter to activated macrophages. In HEK293 cells, Saa3 promoter activity was dramatically up-regulated when C/EBPβ was overexpressed and enhanced by the addition of TNF-α (Fig. 4A). In this study, we generated three mutants of C/EBPβ -binding sites in the Saa3 promoter and infected 3T3-L1 adipocytes with retroviruses harboring these three mutant Saa3 promoters (mut1, mut2, and mut3) (Fig. 4B). We further tested the Saa3 promoter activity in the adipocytes using a co-culture system. Luciferase activity of wild-type Saa3 promoter in the 3T3-L1 adipocytes was significantly elevated upon co-culture with activated macrophages by nearly sixfold, whereas the promoter activities of the three mutants were up-regulated by 1.7-to 3.2-fold (Fig. 4C). Moreover, we examined the effect of MacCM and TNF-α on each of the Saa3 mutant promoters. Similarly, the three point mutations decreased the responsiveness to both MacCM and TNF-α (Fig. 4D,E). In particular, point mutation 3 (mut3) abolished the Saa3 promoter activation by TNF-α , suggesting the critical role of the C/EBPβ -binding site (− 152) for the inflammatory response of the Saa3 promoter in adipocytes. To confirm the involvement of C/EBPβ in Saa3 promoter activation by the interaction with macrophages, we performed siRNA-mediated C/EBPβ knock-down in 3T3-L1 cells. The reduction in C/EBPβ mRNA level was assessed by quantitative RT-PCR (Fig. 5A). C/EBPβ silencing resulted in decreases in both Saa3 mRNA and Saa3 promoter activity upon co-culture with macrophages activated by LPS (Fig. 5B,C).
In addition, C/EBPβ RNAi gene silencing was shown to suppress the Saa3 promoter activation by MacCM and TNF-α (Fig. 5D,E). Moreover, we showed that phosphorylation of C/EBPβ on Thr188 in 3T3-L1 adipocyte was induced by MacCM addition (Fig. 5F). These findings strongly suggest that C/EBPβ actually regulates Saa3 promoter activity in the presence of activated macrophages. Finally, the Saa3 promoter-luciferase chimeric gene was applied to in vivo bioluminescent imaging analysis to monitor adipose inflammation caused by macrophage infiltration during the development of obesity. In this study, we generated Saa3-luc transgenic mice harboring Saa3-luc and two poly-A sites. F3 mice were divided into two groups and fed either HFD or normal diet (ND) for 16 weeks, and subjected to in vivo bioluminescence analysis. At the end of these 16 weeks, in vivo bioluminescent imaging revealed that obese adipose tissue of Saa3-luc mice fed HFD60 could be visualized by luciferase activity (Fig. 6A). We isolated several types of tissue from Saa3-luc mice and confirmed that the luciferase activity based on Saa3 promoter activity was significantly increased in obese adipose tissue compared with that in non-obese adipose tissue (Fig. 6B). Moreover, LPS administered by intraperitoneal injection did not show any bioluminescence in the SAA3-luc mice (data not shown).

Discussion
The infiltration of macrophages into adipose tissue is associated with obesity; the crosstalk between adipocytes and infiltrated macrophages has been considered as an important pathological phenomenon during adipose tissue inflammation [10][11][12][13][14][15][16] . In order to evaluate the number of macrophages during adipose tissue inflammation induced by HFD, quantitative analysis of the mRNA of macrophage markers, such as Emr1 and Ccl2, and immunohistological analysis of the proportion of adipose tissue sections stained positively for F4/80 are most commonly used. However, since obese adipose tissue is excised from obese mice for such analyses, a large number of mice are needed in these investigations and there has been limited opportunity to study the detailed pathological processes in each individual animal. Recently, in vivo imaging technologies have been developed by using a bioluminescence reporter system, and extensively applied for studies of molecular mechanisms in living organisms because the luciferase gene is not present in mammalian physiology 28,29 . To date, luminescence assays of luciferase reporter systems produced using a transgene or by viral infection have been utilized for the real-time biomedical monitoring of tumor development, inflammatory disease, circadian oscillations, and viral infection in living cells or whole animals [30][31][32][33][34][35] . In this study, we isolated the Saa3 gene as an adipocyte-derived factor that is tightly associated with the increased infiltration of macrophages by a comparative analysis between in vivo and in vitro gene clusters. We have shown that the level of Saa3 mRNA was positively correlated with the number of macrophages in adipose tissue in vivo and also that the level of Saa3 mRNA was specifically up-regulated in adipose tissue of obese mice; then, we established transgenic mice carrying the luciferase gene under the control of the Saa3 promoter and subjected them to in vivo luminescence imaging analysis. Because the luciferase activity was detected in obese adipose tissue, it was shown that Saa3 promoter activity could be utilized for monitoring the adipose inflammatory state, possibly serving as an index of the number of infiltrated macrophages in adipose tissue for the following reasons. 1) Saa3 mRNA expression is correlated with the number of macrophages (F4/80-positive cells) and with the mRNA expression of Emr1, which is a marker of macrophages in adipose tissue. 2) Saa3 mRNA expression is down-regulated on administration of a diet rich in vitamin B6 (data not shown), which can decrease the number of macrophages in adipose tissue 20 .
3) The up-regulation of Saa3 mRNA expression is paralleled by an increase in the Emr1 mRNA expression in adipose tissue during obesity development. 4) A short-term HFD (for 1 week) possibly activated resident macrophages in adipose tissue, but, in this study, neither Saa3 mRNA expression nor Saa3 promoter activity (luciferase activity) in adipose tissue was altered after a 1-week HFD (Supplementary Fig. 6). On the other hands, in previous work on a transcription factor, nuclear factor (NF)-κ B, which is involved in numerous inflammatory responses, transgenic mice that carry the luciferase gene under the control of NF-κ B were developed 31,35 . TNF-α and IL-1β , both of which can activate NF-κ B signaling, are reportedly produced during adipose tissue inflammation, whereas recently obtained evidence has indicated that NF-κ B activation plays an important role in the onset of insulin resistance 4,5,36 . Calsen et al. investigated NF-κ B activation in vivo non-invasively in obese mice fed an HFD and showed that, interestingly, whole-body in vivo NF-κ B activity increased more in the HFD group than in a control diet group 35 . In particular, in this previous study, the greatest effect of high-fat feeding was observed in the thoracic region, suggesting that these mice are widely available for the in vivo exploration of NF-κ B activation; however, adipose tissue inflammation could not be visualized specifically by this bioluminescence imaging approach.  Here, we focused on the Saa3 gene as an adipocyte-derived factor involved during adipose tissue inflammation, which sharply responds to these cells' interaction with macrophages. The protein encoded by this gene is one of the members of the saa family, which are known as acute-phase proteins expressed in response to inflammation 37 , and has been studied in terms of the mRNA expression profile in chronic inflammatory diseases such as rheumatoid arthritis, atherosclerosis, and obesity 24,[37][38][39] . The other members of Saa, Saa1 and Saa2, are primarily produced in the liver during acute inflammation 37 , whereas Saa3 mRNA is highly expressed in extrahepatic tissue including lung, large intestine, and adipose tissue. Previous reports have shown that Saa3 mRNA expression in adipocytes is increased in response to saturated fatty acids, IL-1β , and LPS in vitro 25,40,41 . In particular, Saa3 expression in 3T3-L1 adipocytes is reportedly up-regulated in the presence of saturated fatty acid (palmitic acid) in a TLR4-dependent manner 7,25 . In this study, we examined whether the Saa3 promoter in mature 3T3-L1 adipocytes can respond to exogenously added palmitic acid, however, we could not observe any changes of Saa3 promoter activity. In addition, Saa3 protein which is also a ligand for TLR4 could not induce the phosphorylation of p65 (NF-κ B) and ERK proteins in 3T3-L1 cells. These observations suggest that TLR4 expression in 3T3-L1 cells used in our experiments is quite low. However, because nutritional fatty acids, whose blood levels are often increased in obesity, have been shown to activate the TLR4 signaling and induce inflammatory signaling in adipocytes in vivo 7 , a possibility could not be excluded that saturated fatty acids were involved in the enhanced bioluminescence in the obese adipose tissue of mice with HFD. On the other hands, Van de Loo et al. previously explored targeted genes that are differentially regulated in the synovial membrane during the course of rheumatoid arthritis; they focused on the Saa3 gene, whose transcriptional activity was shown to correlate most closely with disease severity 23,24 . Furthermore, the Saa3 promoter region (− 314/+ 50) was shown to confer its cytokine-inducible expression in murine and human synovial fibroblasts, with the strongest transcriptional induction and strength. In this study, we showed that this promoter of the mouse Saa3 region (− 314/+ 50) in adipocytes actually responds to activated macrophages in a co-culture system. Notably, an abundance of C/EBPβ -binding sites can be observed in the promoters of certain genes, including Saa3, whose expression correlates with the severity of collagen-induced arthritis 23 . Three C/EBPβ -binding sites are actually located in the mouse Saa3 region (− 152, − 107, and − 77) 23,27 . Several studies have shown that C/EBPβ is transcriptionally activated by inflammatory stimuli, including inflammatory cytokines such as IL-6, IL-1, and TNF-α , and the role of C/EBPβ in the inflammatory cascade during the development of obesity has been discussed [42][43][44] . In this study, C/EBPβ actually functioned in Saa3 promoter activation by the interaction with macrophages and exogenously added TNF-α . This study also showed that C/EBPβ can be post-translationally activated via the phosphorylation of Thr-188. C/EBPβ is a well-studied key transcription factor that is up-regulated in the early stages of adipocyte differentiation. Previous reports have showed that the phosphorylation of Thr-188 in C/EBPβ by extracellular signal-regulated kinase induces subsequent phosphorylation on Ser-184 and Thr-179, resulting in DNA-binding function acquisition and transactivation 45 . In this study, macrophage-conditioned medium induced the phosphorylation of Thr-188, suggesting a potential role of macrophages in the modulation of adipogenesis. However, TNFα secreted from macrophages is well-known to strongly inhibit adipogenesis by downregulating PPARγ expression via NF-κ B signaling 46 , which suggests that activated macrophages are involved in the inhibition of adipogenesis. Further studies are needed to reveal the molecular mechanism of C/EBPβ phosphorylation in adipocytes by the interaction with macrophages. A previous report showed that C/EBPβ deletion in obese db/db mice reduced adiposity, hepatic steatosis, and diabetes 42 . Furthermore, Rahman et al. demonstrated that bone marrow transplantation from mice lacking C/EBPβ reduced the levels of inflammatory markers and macrophages in adipose tissue, and maintained insulin sensitivity upon feeding on an HFD, suggesting the inflammatory roles of macrophage C/EBPβ in adipose tissue 44 . Meanwhile, C/EBPβ RNAi gene silencing was shown to inhibit NF-κ B DNA-binding activity induced by palmitic acid treatment in 3T3-L1 adipocytes, whereas adenoviral-mediated over-expression of C/EBPβ significantly increased NF-κ B binding activities in 3T3-L1 adipocytes 44 . These observations suggest a functional relationship between C/EBPβ and the transactivation potential of NF-κ B.
Taking these findings together, the increase in luciferase activity in obese adipose tissue in vivo might be mediated by locally produced TNF-α , and the determination of Saa3 promoter activity would be useful for monitoring the adipose inflammatory state associated with increased macrophage content in adipose tissue (Fig. 7). As mentioned above, this study suggested that Saa3 mRNA is also expressed in M1 macrophages in obese adipose tissue. Previous studies showed that HFD also increases the number of M1 macrophages in a variety of tissues such as livers 47 , pancreas 48,49 , skeletal muscles 50,51 , whereas, in this study, the luciferase activity was specifically increased in obese adipose tissue of SAA3-luciferase mice. Taken together, these observations suggest the enhanced bioluminescence in the obese adipose tissue is mainly dependent on an increased luciferase activity in the adipocytes, however, a possibility could not be excluded that Saa3 expression in M1 macrophages partially contributed to the enhanced bioluminescence in the obese adipose tissue in mice fed with HFD. On the other hand, a recent study revealed that Saa3 is an endogenous ligand of TLR4 26 as mentioned above, so it started to be discussed in terms of having an inflammatory role via activation of the TLR4 signaling pathway. Previous reports have demonstrated that Saa3 induced in pre-metastatic lungs by S100A8 and S100A9 has a critical role in the accumulation of myeloid cells as a positive-feedback regulator for chemoattractant secretion in a TLR4-dependent manner 26 . Furthermore, a recent report has shown that Saa3 activates the NLRP3 inflammasome and promotes Th17 allergic asthma in mice 52 . Thus, because mouse Saa3 reportedly plays important roles during the pathological process of various diseases, these findings suggest that the mice established in this study would be widely available for the in vivo non-invasive exploration of compounds or food factors targeting Saa3 gene expression.

Methods
Animals and diets. Male db/db (BKS.Cg-m +/+ Lepr db/J, 6 weeks old) and db/+ (BKS.Cg-Dock7m +/+ Lepr db/J, 6 weeks old) mice were obtained from Charles River Japan. db/db mice fed high-fat diet (HFD: 60% of total calories from fat) and db/+ mice fed AIN93G diet for 3weeks, respectively. Male C57BL/6j mice were purchased from Charles River Japan and they were divided into two groups. Seven-week old mice fed AIN93G (normal diet, ND) or HFD for 16 weeks. Male CD-1 (ICR) mice (4 weeks old, Charles River Japan) were housed in groups of 2 or 3 in metal cages. All mice were given free access to stock diet and deionized water, and housed in a room with controlled temperature (24 ± 1 °C) and a 12 h light/dark cycle; light from 0800 to 2000, daily. The animal study was approved by the Hiroshima University Animal Committee (Permit Number: C13-3), and the mice were maintained in accordance with the Hiroshima University Guidelines for the Care and Use of Laboratory Animals. siRNA, plasmid transfection and Cell culture. Mouse 3T3-L1 preadipocytes, mouse macrophage RAW264.7 cells and human HEK293 cells were cultured in a maintenance medium (10% fetal bovine serum, 100 units/ml penicillin and 100 μ g/ml streptomycin in Dulbecco's modified medium) at 37° in 5% CO 2 /95% humidified air. 3T3-L1 cells were differentiated as previously described 20,22 . Stimulation of RAW264.7 cells and co-cultivation of 3T3-L1 cells and RAW264.7 cells were performed based on previous protocol 20,22 . Differentiated 3T3-L1 adipocytes were stimulated with 10 ng/ml of mouse recombinant tumor necrosis factor-α (R&D system, Minneapolis, MN) for 24 hr. cDNA was amplified using a PCR primer set (5′ -CCGCGTTCATGCACCGCCTG-3′ and 5′ -ACCCGCGCCGCGCTAGCAGT-3′ ) designed according to the nucleotide sequences of mouse C/EBPβ and subcloned into pcDNA3.1, which generated pcDNA-C/EBPβ . DNA transfections were performed using GeneJuice Transfection Reagent (Merck Millipore) according to the manufacturer's instructions. Small interfering RNA (siRNA) duplex oligoribonucleotides against mouse C/EBPβ were synthesized by Sigma. The sequences were as follows: sense 5′ -CGGGUUUCGGGACUUGAUGTT-3′ , antisense 5′ -CAUCAAGUCCCGAAACCCGTT-3′ .
Universal negative control siRNA (Sigma) was used as a control siRNA. Differentiated 3T3-L1 cells were transfected with these siRNAs to a final concentration of 20 nM using LipofectAMINE RNAimax (Invitrogen). DNA microarray analyses. Total RNAs were isolated from epididymal white adipose tissue of male db/db mice and db/+ mice using RNeasy lipid tissue kit (Qiagen Sciences, Germantown, MD), and pooled RNAs were subjected to cRNA synthesis for a DNA microarray analysis (44 K whole mouse genome 60-mer oligo microarray, Agilent Technologies, Palo Alto, CA). All procedures of fluorescence labeling, hybridization, and image processing were performed according to the manufacturer's instructions. Gene expression data were obtained and statistically analyzed using Agilent Feature Extraction software (version 9.5). 3T3-L1 cells and RAW264.7 cells were co-cultured in maintenance medium using transwell system as described above and RAW264.7 cells were further stimulated with 1 μ g/ml LPS for 24 hr. Total RNAs were isolated from 3T3-L1 cells using RNeasy kit (Qiagen). DNA microarray analysis, based on a system containing 5693 gene probes (Affymetrix, Santa Clara, CA, USA), was used to compare transcriptional profiles between 3T3-L1 cells co-cultured with RAW264.7 cells in the presence or absence of LPS. This array contains a broad range of genes derived from publicly available, well-annotated mRNA sequences. Preparations were quantified and their purity was confirmed by standard spectrophotometric methods. The results were expressed as the ratio of fluorescent intensity of the genes expressed in the two groups. The microarray data are also deposited in the NCBI GEO data base (available on the World Wide Web at www. ncbi.nlm.nih.gov/geo) under accession number GSE70527.
Isolation of mature adipocytes. Epididymal white adipose tissue isolated from male db/db, db/+, ND and HFD mice were minced in phosphate-buffered saline and digested with 1 mg/ml collagenase Type I (Worthington Chemical Corporation) for 30 min at 37 °C. The resulting cell suspension was filtered through a 100-μ m filter and centrifuged at 233 × g for 1 min to separate mature adipocytes from stromal vascular fraction (SVF) cells.
Immunohistochemical analysis. The epididymal adipose tissue was isolated and fixed with neutral buffered formalin and embedded in paraffin. An immunohistochemical study was carried out using 4-μ m-thick paraffin-embedded sections and rat anti-mouse F4/80 antibody (AbD Serotec, Raleigh, NC). The number of F4/80-positive cells in more than 100 serial fields was counted in a blinded fashion through the microscope, and the data were obtained as the mean number/mm 2 .
The transgene construct carrying Saa3-Luc and two polyadenylation sites was excised from agarose gel, purified, and used for microinjection into BDF1 mouse eggs at Japan SLC Inc. (Hamamatsu, Japan). Among 89 mouse pups, Tg10-3 line was studied. The male chimera harboring the Saa3-luc transgene was mated with C57BL/6 J female mice to obtain F1 offspring. The heterozygous F1 male offspring from this breeding were then backcrossed with purebred C57BL/6 J females to obtain F2 offspring; this process was continued until the F3 generation of mice was obtained. After weaning, the male heterozygous Saa3-luc transgenic mice were divided into two groups and fed HFD and AIN93G, respectively.
Luciferase assay. For in vitro reporter studies, transfected 3T3-L1 (3T3-L1/Saa3-luc) cells were seeded at 1 × 10 5 cells per ml in 12 well culture plate. Confluent 3T3-L1/Saa3-luc cells were differentiated as described above. Using a transwell system, 5 × 10 4 RAW264.7 cells were cultured in the upper chamber and stimulated by LPS as mentioned above. 3T3-L1/Saa3-luc cells were stimulated with MacCM or 10 ng/ml of TNF-α for 24 hr. Subsequently, 3T3-L1/Saa3-luc cells were lysed in ice-cold lysis buffer (1% Triton-X 100, 2 mM DTT, 10% glycerol, and 25 mM Tris-HCl pH 7.5). After sonication, cell lysate was centrifuged at 12,000 rpm and supernatant was collected. Several tissues obtained from Saa3-luc mice were homogenized in the lysis buffer and centrifuged at 12,000 rpm. Luciferase activity was quantified using the luciferase assay kit (Toyo Inki, Japan) and a luminometer (Turner Model TD-20), expressed as relative light units normalized to total protein concentration of the cell or tissue extracts.
Western blot analysis. Mature 3T3-L1 adipocytes were stimulated by MacCM for 5, 10, 15 minutes. After stimulation, 3T3-L1 cells were washed with ice-cold PBS and scraped with RIPA buffer. Cells lysates were centrifuged at 12,000 rpm, 4 °C, 10 min for remove cell debris and supernatant were collected new tubes. Protein concentration of the supernatant was determined using the Bio-Rad protein assay kit (Bio-Rad) with BSA as a standard. 20 μ g (protein equivalents) of the supernatant were resolved by SDS-PAGE, transferred onto a polyvinylidene difluoride (PVDF) membranes and immunoblotted with anti-CEBP/β (Santa Cruz Biotechnology), anti-p-CEBP/β (abcam) antibodies. Uncropped scans of the blots were supplied in Supplementary Fig. 7.
In vivo bioluminescent imaging analysis. For in vivo bioluminescent imaging, male Saa3-luc mice were injected intraperitoneally with D-luciferin (150 mg/kg body weight, Promega) and 10 minutes later, anesthetized with pentobarbital sodium. After 5 minutes, Saa3-luc mice were placed supine position on the plate and imaged for 1 min with the camera set at the highest sensitivity by NightOWL II Imaging Systems LB983 (Berthold Technologies, Bad Wildbad, Germany). Photons emitted from tissues were analyzed using Indigo in vivo image software (Berthold). Signal intensity was quantified as the sum of all detected photon counts per second and presented as count/sec (cps). For any given analysis, all images were adjusted to the same scale of minimum and maximum luminescent intensity.

Statistical analyses.
Values are presented as means ± S.E. Statistical significance was determined by unpaired Student's t test.