Nitric Oxide Produced by Macrophages Inhibits Adipocyte Differentiation and Promotes Profibrogenic Responses in Preadipocytes to Induce Adipose Tissue Fibrosis

Low-grade inflammation in adipose tissue is a major factor in the development of obesity-associated insulin resistance (1). In addition to inflammation, adipose tissue fibrosis is exacerbated in obese human subjects and in high-fat diet (HFD)–fed rodents (2). Fibrosis limits the expandability of adipose tissue and contributes to ectopic fat accumulation and the development of insulin resistance (3). Inducible nitric oxide synthase (iNOS) is a member of the NOS family, which produces nitric oxide (NO). In addition to its role in the killing of certain pathogens by macrophages, iNOS is inducible in many cell types and has been implicated in insulin resistance and various metabolic diseases (4,5). iNOS activation is involved in the pathogenesis of inflammation and fibrosis (6). In particular, classically activated (M1) macrophages have increased iNOS expression and contribute to adipose tissue inflammation (7,8). HFD increases iNOS mRNA transcript levels in adipose tissue (5,9), but the role of iNOS in adipose tissue fibrosis is unknown.

Hypoxia is likely to be a major factor contributing to adipose tissue inflammation and fibrosis (2,10–12). Hypoxia stimulates the transcription of hypoxia-inducible factor (HIF)-1α and promotes stabilization of the HIF-1α protein by prolyl hydroxylase domain protein–dependent hydroxylation (13). Activation of HIF-1α inhibits preadipocyte differentiation and initiates adipose tissue fibrosis (2,3). HIF-1α is well known to upregulate the expression of iNOS (14). In addition, NO produced by iNOS can regulate HIF-1α stability (13). A low NO concentration promotes HIF-1α degradation under hypoxic conditions (15), whereas a high level of NO increases HIF-1α stability through the inhibition of prolyl hydroxylase domain activity under normoxic conditions (16). Furthermore, a high concentration of NO inhibits the mitochondrial respiratory chain (17), whereas a low NO level promotes mitochondrial biogenesis (18). Recent studies (19) have shown that mitochondrial dysfunction can also induce pseudo-hypoxic HIF-1α activation under normoxic conditions.

In the current study, we show that iNOS^−/− mice are protected from HFD-induced adipose tissue fibrosis. Interestingly, the expression of mitochondrial biogenesis factors was significantly decreased in adipose tissue of HFD-fed wild-type (WT) mice, and this alteration was reversed in iNOS^−/− mice. The accumulation of DNA damage has been linked to aging and the onset of age-related diseases (20,21). p53 is a key player in the intrinsic cellular response to DNA damage, and p53 activation leads to cell cycle arrest, apoptosis, and senescence (22). In particular, the p53 signaling pathway represses the expression of peroxisome proliferator–activated receptor γ coactivator 1α (PGC-1α) (23). We thus examined the possibility that NO produced by macrophages induces mitochondrial dysfunction in preadipocytes by activating p53 signaling and that this is responsible for the accumulation of HIF-1α and the fibrogenic response in preadipocytes.

Eight-week-old male WT (C57BL/6J) and iNOS^−/− (C57BL/6-NOS2^tm1Lau) mice (The Jackson Laboratory, Bar Harbor, ME) were fed either a normal chow diet (ND; 12 kcal% fat; Biopia, Gunpo, Korea) or an HFD (60 kcal% fat, 90% from lard and 10% from soybean oil; Research Diets, New Brunswick, NJ). Mice were housed at ambient temperature (22 ± 1°C) with a 12-h light-dark cycle and free access to water and food. After 16 weeks of feeding, mice were fasted for 5 h in the morning before they were euthanized. Blood samples were collected for biochemical analysis, and epididymal white adipose tissue (eWAT) was rapidly removed, weighed, and stored at −80°C. All animal experimental protocols were approved by the Institutional Animal Care and Use Committee of the Asan Institute for Life Sciences.

Plasma glucose levels were determined using a glucose and lactate analyzer (YSI2300; Yellow Springs Instruments, Yellow Springs, OH). Plasma insulin and adiponectin levels were measured using radioimmunoassay kits (Linco Research, St. Charles, MO).

The glucose tolerance test (GTT) and the insulin tolerance test (ITT) were performed at 14 and 15 weeks of diet feeding, respectively. For the GTT, mice were fasted overnight and then administered 1 g/kg glucose i.p. One week later, mice were fasted for 5 h in the morning and then injected with 0.75 mU/kg i.p. regular human insulin for the ITT. Blood was collected before injection, and at 15, 30, 60, 90, and 120 min after injection for blood glucose level measurements.

In a separate set of experiments, body composition was measured in WT mice and iNOS^−/− mice fed a ND or HFD for 15 weeks (n = 5 each). Mice were anesthetized with ketamine/xylazine (8 and 1.6 mg/kg i.p., respectively), and total fat mass and lean body mass were measured using a PIXI-Mus Small Animal Densitometer (Lunar Corp., Madison, WI).

Tissues were fixed in 10% formalin, embedded in paraffin, and sectioned. Tissue sections were stained with hematoxylin-eosin (H-E). Immunohistochemistry for F4/80 protein was performed by incubating tissue sections with anti-F4/80 antibody (ab6640; Abcam, Cambridge, MA). To detect connective tissue, additional sections were stained with Masson trichrome (MT) stain or 0.1% Sirius Red in saturated picric acid.

Bone marrow–derived macrophages (BMDMs) were prepared as described previously (24). Nonadherent cells were carefully removed, and fresh medium was added every 2 days. On day 8, the cells were collected for experiments.

eWAT from 4-week-old WT mice was digested with collagenase (2 mg/mL; Sigma-Aldrich, St. Louis, MO) in Hanks’ balanced salt solution, and adipose-derived stromal-vascular cells were isolated and cultured, as described previously (25). Fresh medium was added every 2 days. On day 8, the cells were collected for experiments.

RAW264.7 cells or BMDMs were treated with 10 ng/mL lipopolysaccharide (LPS) for 24 h, and the conditioned media were transferred to 3T3-L1 preadipocytes or primary preadipocytes for 8 h. To examine the role of iNOS activation, macrophages were cultured in media containing LPS with 30 μmol/L S-methylisothiourea (SMT; Sigma-Aldrich). SMT is an iNOS-selective inhibitor that, at concentrations up to 1 mmol/L, does not inhibit the activity of xanthine oxidase, diaphorase, lactate dehydrogenase, monoamine oxidase, catalase, cytochrome P450, or superoxide dismutase (26). In another set of experiments, 3T3-L1 preadipocytes were treated with either of the NO donors sodium nitroprusside (SNP) or Deta-NONOate (Deta-NO; Sigma-Aldrich) for 8 h.

3T3-L1 preadipocytes were differentiated into mature adipocytes, as described previously (27), in the presence or absence of SNP or macrophage-conditioned media with or without SMT. From day 0 of differentiation, 0.25 or 0.5 mmol/L SNP was added, whereas RAW cell-conditioned media (one-fifth dilution) were supplemented from day 2 of differentiation. Morphological changes in adipocytes were observed by phase contrast microscopy.

NO production was estimated by measuring the media nitrite produced by RAW264.7 cells or BMDMs using Griess reagent (Promega, Madison, WI).

Mature 3T3-L1 adipocytes treated with SNP were fixed with 10% formaldehyde for 1 h. After a wash with PBS, the cells were stained with oil red O solution (Sigma-Aldrich) for 30 min. The slides were then washed several times with water, and excess water was evaporated by heating the stained cultures to ∼32°C.

To measure DNA damage in the cells (28), cultured 3T3-L1 preadipocytes were fixed with 2% formaldehyde in PBS and stained with the primary histone H2AX (γH2AX) antibody (Millipore, Billerica, MA), followed by visualization with tetramethylrhodamine-conjugated anti-mouse IgG secondary antibody (Invitrogen, Carlsbad, CA). For nuclei staining, cells were incubated with 0.5 μg/mL DAPI in PBS. Immunofluorescence staining for α-smooth muscle actin (αSMA), a marker of myofibroblasts, was performed using an antibody from Abcam (ab5694). For the staining of intracellular lipid droplets, cells were also stained with CellTrace BODIPY TR Methyl Ester (C34556; Invitrogen).

NO can react with superoxide to form peroxynitrite (ONOO⁻), which can contribute to the fibrotic response (29). As a marker of oxidative damage mediated by peroxynitrite, we measured the 3-nitrotyrosine (3-NT) level using an ELISA kit (ab116691; Abcam).

Gene expression in tissues and cells was assessed using real-time PCR using gene-specific primers (Supplementary Table 1) and the 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA) using an SYBR Green PCR Kit (Applied Biosystems). Total RNA was isolated using TRIzol (Invitrogen), and 1 μg of each sample was reverse transcribed with random primers using a Reverse Aid M-MuLV Reverse Transcription Kit (Fermentas, Hanover, MD). The relative expression levels of each gene were normalized to those of 18S rRNA.

Western blotting was performed as described previously (27). Antibodies against iNOS (catalog #610432; BD Transduction Laboratories, Lexington, KY), HIF-1α (catalog #NB100–134; Novus Biologicals, Littleton, CO), total p53 (catalog #2624; Cell Signaling Technology, Danvers, MA), and phospho-p53 (catalog #9284; Cell Signaling Technology) were used. For housekeeping controls, we used α-tubulin (catalog #NB100–690; Novus Biologicals) for in vivo study and β-actin (catalog #A5441; Sigma-Aldrich) for in vitro study, as we had difficulty using β-actin as a housekeeping gene for in vivo samples.

A XF24 extracellular flux analyzer (Seahorse Bioscience, North Billerica, MA) was used to measure the oxygen consumption rate in 3T3-L1 preadipocytes, as described previously (30).

Data are presented as the mean ± SEM. Statistical significance was determined using an unpaired two-tailed t test or ANOVA. Data were analyzed using SPSS version 17 (SPSS Inc., Chicago, IL). P < 0.05 was used as the threshold for statistical significance.

In agreement with previous studies (5,9), HFD feeding for 16 weeks significantly increased iNOS mRNA expression in eWAT (Fig. 1A). This was associated with an increased level of 3-NT, a marker of oxidative damage mediated by peroxynitrite (29) (Fig. 1B). Compared with the ND, HFD feeding increased body weight in both WT and iNOS^−/− mice. However, the body weight and food intake of HFD-fed iNOS^−/− mice were significantly lower than those of WT mice (Fig. 1C). Fasting plasma glucose and insulin levels were reduced in HFD-fed iNOS^−/− mice compared with WT mice (Fig. 1D). Plasma adiponectin levels were significantly decreased by HFD feeding in WT mice, but not in iNOS^−/− mice (Fig. 1D). HFD-induced glucose intolerance and insulin resistance measured by GTT and ITT, respectively, were attenuated in iNOS^−/− mice (Fig. 1E). H-E, F4/80 protein, MT, and Sirius Red staining revealed adipose tissue inflammation and fibrosis in HFD-fed WT mice, whereas iNOS^−/− mice showed significantly less inflammation and fibrotic changes in the eWAT (Fig. 1F).

Interestingly, the weight of eWAT was significantly higher in HFD-fed iNOS^−/− mice than in HFD-fed WT mice (Fig. 2A), in contrast to the body weight changes (Fig. 1A). Similarly, densitometer measurement of body composition revealed a higher fat mass in iNOS^−/− mice than in WT mice (Fig. 2B). Consistent with the notion that adipose tissue fibrosis leads to ectopic fat accumulation (3), hepatic steatosis in HFD-fed mice was significantly reduced in iNOS^−/− mice (Fig. 2C).

Of the various immune cells involved in inflammation, adipose tissue macrophages play a central role in the genesis of adipose tissue inflammation (31). At least two distinct populations of macrophages infiltrate the adipose tissue: proinflammatory (M1) and anti-inflammatory (M2). In agreement with previous studies (8), an HFD for 16 weeks significantly increased the expression of M1 markers and some M2 markers in the eWAT (Fig. 3A and B). This was associated with increased expression of fibrosis markers (Fig. 3C). Changes in most of the markers of macrophage polarization and fibrosis were partially reversed in iNOS^−/− mice (Fig. 3A–C).

Expression levels of HIF-1α mRNA and protein were significantly increased in WT mice receiving an HFD (Fig. 4A and B) (12,32). Levels of HIF-1α mRNA transcripts were not different in iNOS^−/− mice compared with WT mice, but HIF-1α protein levels were significantly reduced in iNOS^−/− mice receiving an HFD (Fig. 4B). On the other hand, the expression levels of HIF-2α and HIF-3α significantly decreased in WT mice fed the HFD but significantly increased in iNOS^−/− mice fed the HFD (Fig. 4A), supporting the concept that HIF-2α or HIF-3α is compensatory to HIF-1α (33).

Recent studies have suggested that mitochondrial dysfunction is an important pathogenic factor of fibrosis (34). Accordingly, the expression levels of genes encoding mitochondrial biogenesis factors, adiponectin, and mitochondrial respiratory complex proteins were found to be significantly decreased in the eWAT of HFD-fed WT mice, and these changes were partially reversed in iNOS^−/− mice (Fig. 4C and D).

HFD feeding can cause changes in gut microbiota, resulting in increased plasma levels of LPS, which can exacerbate adipose tissue inflammation and obesity (35). Treatment with LPS (10 ng/mL) for 8 h significantly increased iNOS but not HIF-1α protein levels in RAW264.7 macrophages (Fig. 5A) and BMDMs isolated from WT mice (Fig. 5B), and accordingly increased the concentrations of nitrite in culture supernatants (Fig. 5A and B). As expected, BMDMs isolated from iNOS^−/− mice did not show increased expression of iNOS and nitrite in the culture supernatant after LPS treatment (Fig. 5B).

Factors secreted by macrophages can promote a profibrotic phenotype in preadipocytes and reduce adipocyte differentiation (36–39). Therefore, we tested the possibility that NO produced by activated macrophages could augment the HIF-1α protein level and decrease the expression of mitochondrial biogenesis factors in preadipocytes. Conditioned media from LPS-stimulated RAW264.7 macrophages increased the expression of HIF-1α protein and fibrogenic gene mRNA transcripts, and suppressed PGC-1α expression in 3T3-L1 preadipocytes (Fig. 5C). Treatment with LPS did not increase the HIF-1α protein level in 3T3-L1 preadipocytes (Fig. 5D), indicating that the LPS remaining in the macrophage culture supernatant is not responsible for the changes. Cotreatment of macrophages with LPS and SMT did not completely reverse conditioned media-induced changes in preadipocytes, but significantly ameliorated them (Fig. 5C).

Interestingly, preadipocytes treated with conditioned media from LPS-treated BMDMs from iNOS^−/− mice showed significantly less accumulation of HIF-1α protein (Fig. 5E) and higher levels of mitochondrial biogenesis factors (Fig. 5F) than those treated with conditioned media from BMDMs from WT mice.

In a further assessment of the role of NO, treatment with NO donors, SNP or Deta-NO, increased the protein levels of HIF-1α and the mRNA expression of profibrogenic genes, and suppressed PGC-1α expression in 3T3-L1 preadipocytes (Fig. 5G).

We next examined the effect of rosiglitazone on NO-mediated changes in the HIF-1α protein level. Rosiglitazone is a peroxisome proliferator–activated receptor γ (PPARγ) agonist that increases mitochondrial biogenesis in differentiated adipocytes (27). 3T3-L1 preadipocytes showed significantly lower expression of PPARγ than differentiated adipocytes, and treatment with rosiglitazone did not significantly increase PPARγ expression (data not shown). Nevertheless, rosiglitazone significantly increased mitochondrial respiration in SNP-treated 3T3-L1 preadipocytes (Fig. 6A). Treatment with rosiglitazone also reversed the SNP-mediated increase in the HIF-1α protein level (Fig. 6B), and changes in PGC-1α and fibrogenic gene transcription (Fig. 6C).

A growing body of evidence suggests that DNA damage is linked to adipose tissue inflammation and systemic insulin resistance (21), and that the p53 signaling pathway can repress PGC-1α expression (23). In particular, NO has been shown to increase p53 accumulation (40). We thus tested the possibility that NO decreases mitochondrial biogenesis via p53-dependent suppression of PGC-1α. The highest activation state of p53 occurs when it is phosphorylated on serine-15 (41). Conditioned media from LPS-treated RAW cells increased p53 phosphorylation in 3T3-L1 preadipocytes, whereas conditioned media from SMT cotreated cells did not show these effects (Fig. 7A). The mRNA expression levels of p21, a downstream effector of p53 that regulates many cellular processes (42), were similarly altered (Fig. 7B).

γH2AX phosphorylation is a rapid and sensitive cellular response to the presence of DNA double-stranded breaks (28). Treatment with NO donors significantly increased γH2AX staining in 3T3-L1 preadipocytes, indicating that NO induces DNA damage in these cells (Fig. 7C). NO donors also significantly increased phosphorylated p53 protein levels and p21 mRNA levels (Fig. 7D). Small interfering RNA (siRNA) directed against p53 reversed SNP-mediated changes in HIF-1α protein and phosphorylated p53 protein levels, and p21 and PGC-1α mRNA levels in 3T3-L1 preadipocytes (Fig. 7E and F).

In agreement with these in vitro findings, significantly higher levels of phosphorylated p53 were noted in the eWAT of WT mice, but not iNOS^−/− mice, fed an HFD (Fig. 7G). In addition, the mRNA transcript levels of p21 were significantly increased in the eWAT of WT mice receiving an HFD, a change that was not observed in HFD-fed iNOS^−/− mice (Fig. 7H). This suggests that NO can increase p53 signaling in the eWAT of HFD-fed mice as well as in preadipocytes cultured in vitro.

We next examined whether NO affects adipocyte differentiation in 3T3-L1 preadipocytes. SNP treatment significantly decreased the expression of adipogenesis markers and increased the expression of fibrosis markers (Fig. 8A and B). Phase-contrast imaging, oil red O and BODIPY staining, and immunofluorescent staining for αSMA in differentiated 3T3-L1 adipocytes revealed an increased frequency of cells with a profibrotic phenotype in SNP-treated cells, including a flattened fibroblast-like morphology, decreased lipid droplet content, and increased levels of αSMA expression (Fig. 8C). Treatment with conditioned media from LPS-treated RAW264.7 cells also decreased the expression of adipogenesis markers and oil red O staining of lipid droplets in 3T3-L1 adipocytes (Fig. 8D and E). Conditioned media from cells cotreated with SMT did not completely reverse these changes, but significantly ameliorated them (Fig. 8D and E).

In our current study, we found that iNOS^−/− mice were protected from HFD-induced adipose tissue fibrosis. The HFD significantly increased both the mRNA and protein expression levels of HIF-1α in the eWAT of the mice. This change may have occurred because of adipose tissue hypoxia (2,10–12). It has been suggested that HIF-1α induction by tissue hypoxia leads to an upregulation of “fibrotic response” genes, resulting in the local fibrosis and necrosis of adipocytes, which attracts classically activated proinflammatory M1 macrophages and leads to metabolic dysfunction (2). Intriguingly, our study showed that the HIF-1α protein levels, but not the transcript levels, were significantly reduced in iNOS^−/− versus WT mice receiving the HFD. Similarly, conditioned media from activated macrophages or treatment with NO donors significantly increased HIF-1α protein levels in preadipocytes. This finding suggested that the NO produced by activated macrophages amplifies and sustains the hypoxia-induced fibrogenic responses in adipose tissue.

Notably, we found that the expression of mitochondrial biogenesis factors was significantly decreased in the eWAT of WT mice, but not iNOS^−/− mice, on an HFD. NO produced by activated macrophages reduced mitochondrial respiration in preadipocytes and their differentiation to mature adipocytes in a paracrine manner. Among the distinct cell types that reside in the adipose tissue stromal compartment, preadipocytes represent progenitor cells that are more committed to the adipocyte lineage (36). Accordingly, the administration of SNP or Deta-NO decreased the expression of PGC-1α and increased profibrotic responses in preadipocytes. In agreement with previous studies (38,39), conditioned media from macrophages significantly increased profibrotic responses in preadipocytes and decreased the expression of mitochondrial biogenesis factors. Importantly, HIF-1α protein was decreased, and mitochondrial biogenesis factors were increased in preadipocytes treated with conditioned media from LPS-treated BMDMs from iNOS^−/− mice, suggesting an important role of NO. However, SMT significantly but partially blocked these responses in preadipocytes by conditioned media from LPS-treated macrophages. This may be because LPS-stimulated macrophages produce various proinflammatory cytokines besides NO. A recent study (43) reported that various proinflammatory cytokines produced by activated macrophages, such as tumor necrosis factor-α, interleukin (IL)-6, and IL-1β, decreased mitochondrial function in 3T3-L1 adipocytes. Taken together, NO produced by activated macrophages can promote a profibrotic phenotype in preadipocytes by decreasing mitochondrial function, in concert with other cytokines.

NO generated by iNOS has been shown to evoke p53 accumulation (40). In agreement, we found that SNP or Deta-NO significantly increased DNA damage and the levels of phosphorylated p53 in preadipocytes, and that siRNA targeting p53 reversed the SNP-induced downregulation of PGC-1α. In support of the idea that p53 signaling contributes to the reduced expression of mitochondrial biogenesis factors that occurs in the adipose tissue of HFD-fed mice, the levels of phosphorylated p53 protein and p21 transcript were significantly increased in the eWAT of WT mice, but not iNOS^−/− mice, on an HFD. Taken together, these data suggest that NO produced by activated macrophages leads to p53 accumulation in preadipocytes, which induces mitochondrial dysfunction and profibrotic changes.

Although the inflammatory response that occurs during tissue injury is necessary for proper tissue repair, it can lead to pathological tissue fibrosis if it becomes dysregulated (44). The fibrotic change is directly related to the degree of inflammation, but also with the appearance of M2 macrophages (45). In agreement, the expression of markers for M2 macrophages was significantly increased in long-term HFD-fed mice. iNOS is considered a marker of M1 macrophages, but our present study showed that the increase in M2 markers in HFD-fed mice was significantly decreased in HFD-fed iNOS^−/− mice. This may contribute to the decreased adipose tissue fibrosis in iNOS^−/− mice.

iNOS^−/− mice in our colony were resistant to the HFD-induced body weight gain. Even though the original article on iNOS and insulin resistance did not report differences in body weight (5), a later study (46) showed that the deletion of iNOS in ob/ob mice increased energy expenditure. iNOS^−/− mice in our colony consumed significantly less food than WT mice by a presently unknown mechanism. We did not perform pair-fed experiment to examine how much difference in the body weight results from decreases in food intake or increases in energy expenditure in HFD-fed iNOS^−/− mice. However, during the early period of HFD feeding, the body weight of HFD-fed iNOS^−/− mice was not significantly different from that of WT mice, even though iNOS^−/− mice consumed significantly less food than WT mice. This suggests that, during this early period of HFD feeding, decreased energy expenditure compensates for decreased food intake. Thus, increases in energy expenditure may not be the major cause of the body weight decrease in iNOS^−/− mice. The cause of this difference between studies is currently unknown. However, body weight reduction in iNOS^−/− mice may explain reduced insulin resistance in these mice.

Alternatively, reduced adipose fibrosis may also explain reduced insulin resistance in iNOS^−/− mice. Interestingly, the weight of the eWAT depot in HFD-fed iNOS^−/− mice was significantly higher than that of HFD-fed WT mice, even though the iNOS^−/− mice were resistant to the HFD-induced body weight gain. The measurement of body composition also showed that fat mass was higher in HFD-fed iNOS^−/− mice than in WT mice. These findings are unexpected but are consistent with the notion that adipose fibrosis limits adipose tissue expandability. Impaired adipose tissue expandability may lead to the development of ectopic fat accumulation and insulin resistance (3). Consistently, we found that HFD-induced steatosis in the liver, a representative site of ectopic fat accumulation, was ameliorated in iNOS^−/− mice.

Our present study contradicts a previous report (47) showing that iNOS deficiency in myeloid cells does not prevent diet-induced insulin resistance. In that study, iNOS deficiency in macrophages did not attenuate adipose tissue inflammation or the LPS-induced inflammatory response in macrophages. The cause of this discrepancy is presently unclear. However, the degree of tissue injury, inflammation, and fibrosis may not correlate with each other (6), and it would be interesting to see whether these mice with myeloid iNOS deficiency are protected from adipose tissue fibrosis.

It should also be noted that the role of iNOS in fibrosis is highly controversial and depends on the tissue and experimental protocol. For example, accelerated liver fibrosis was found in HFD-fed iNOS^−/− mice (48,49), whereas iNOS^−/− mice were protected from cholesterol-induced liver fibrosis (50). Interestingly, iNOS^−/− mice have shown increased hepatic injury but decreased fibrosis after long-term carbon tetrachloride administration (6).

Taken together, we suggest that NO produced by activated macrophages induces p53-dependent PGC-1α suppression and mitochondrial dysfunction in preadipocytes; and that this is responsible for the increased HIF-1α accumulation, defective adipocyte differentiation, and increased fibrogenesis (Fig. 8F). Interactions between macrophage iNOS and preadipocytes may therefore potentially represent a novel therapeutic target for restoring healthy adipose tissue function.

Acknowledgments. The authors thank Dr. Kevin Clayton (Boston BioEdit) for English proofreading.

Funding. This study was supported by the National Research Foundation of Korea (NRF), which is funded by the Ministry of Education, Science, and Technology (grants NRF-2006-2005412 and 2009-0091988 to K.-U.L.; grant 2012R1A1A3012626 to E.H.K.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. J.E.J. designed and conducted the study, performed experiments, analyzed and interpreted the data, and wrote the manuscript. M.S.K. and J.-Y.Y. performed in vivo and in vitro experiments and interpreted the results. M.-O.K. performed some parts of the in vitro experiments. J.H.K., H.S.P., A.-R.K., H.-J.K., and B.J.K. contributed to the data analysis or performed some parts of experiments. Y.E.A., J.S.O., W.J.L., and R.A.H. critically reviewed the manuscript, provided suggestions, and contributed to the discussion. E.H.K. and K.-U.L. conceptualized and designed the study and analyzed the data. E.H.K. and K.-U.L. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.