Insulin Regulates Lipolysis and Fat Mass by Upregulating Growth/Differentiation Factor 3 in Adipose Tissue Macrophages

The worldwide prevalence of obesity increases morbidity and mortality and imposes a growing public health burden. Most excess food intake is converted into fat, and specifically into triglycerides (TGs), which are stored in adipocytes of white adipose tissue (WAT). As adipocytes accumulate fat and increase in size, they start to secrete proinflammatory adipocytokines, recruit or polarize macrophages and other hematopoietic cells inside WAT, and cause chronic inflammation and obesity-related disorders (1). The TG content in adipocytes is determined by the balance between the synthesis and breakdown of TG. Although TG synthesis depends on the uptake of nutrients, the rate of lipid removal through lipolysis is proportional to the total fat mass as well as the activities of lipases, and is regulated by external factors, such as catecholamine and insulin. It is important to understand the mechanisms of fat accumulation to dissect the pathophysiology of obesity. Our previous genetic analyses using F2 progeny between the Tsumura, Suzuki, obese diabetes (TSOD) and control BALB/c mice revealed a naturally occurring mutation in Acvr1c encoding the type I transforming growth factor-β (TGF-β) receptor activin receptor–like kinase 7 (ALK7) in BALB/c mice (2–5). The mutation gives rise to a stop codon in the kinase domain of ALK7. The congenic strain T.B-Nidd5/3 is isogenic with TSOD mice except for the BALB/c-derived ALK7 mutation and exhibits decreased adiposity because of enhanced lipolysis. The activation of ALK7 downregulates the master regulators of adipogenesis, C/EBPα and peroxisome proliferator–activated receptor γ (PPARγ), in differentiated adipocytes, which leads to the suppression of lipolysis and to increases in adipocyte size and TG content.

To understand the regulatory mechanisms associated with ALK7, it is essential to determine its physiological ligand. TGF-β family members such as Nodal, inhibin-βB (activin B or activin AB), growth/differentiation factor (GDF) 3, and GDF11 bind ALK7 and mediate its signals under specific conditions (6–9). Among these ligands, GDF3 seems to function under nutrient-excess conditions, because both GDF3 and ALK7 knockout mice attenuate fat accumulation in the face of high-fat diet (HFD)–induced obesity (9,10). However, it has not been shown that GDF3 directly activates ALK7 in adipocytes. Besides, neither the producer nor the upstream regulator of GDF3 under nutrient-excess conditions is known. In the current study, we establish GDF3 as the physiological ligand that activates ALK7 in adipocytes, and CD11c⁺ adipose tissue macrophages (ATMs) as the main cell source of GDF3. We further demonstrate that insulin upregulates GDF3 in ATMs ex vivo and stimulates fat accumulation in vivo through the GDF3-ALK7 signaling pathway. Our findings reveal a novel mechanism by which insulin regulates adiposity through ATMs in addition to its classically defined direct effect on adipocytes.

Animal experiments were performed in accordance with the rules and regulations of the Animal Care and Experimentation Committee, Gunma University. The TSOD mouse was originally established from an outbred ddY strain as an inbred strain with obesity and urinary glucose (11). The congenic mouse strain T.B-Nidd5/3 was developed and characterized previously (3,4). The GDF3 knockout mouse with a genetic background of C57BL/6J was described previously (10). C57BL/6N and BALB/cA mice were purchased from CLEA Japan. Only male mice were phenotypically characterized in the current study. Mice had ad libitum access to water and standard laboratory chow (CE-2; CLEA Japan) in an air-conditioned room with 12-h light/dark cycles. An HFD (55% fat, 28% carbohydrate, and 17% protein in calorie percentage; Oriental Yeast Co., Ltd.) was given to mice from 4 weeks of age for the indicated duration. For macrophage depletion, liposomes containing 110 mg/kg body weight clodronate (ClodronateLiposomes.org) were injected intraperitoneally twice per week. For the in vivo insulin administration, saline or 0.75 units/kg body weight insulin (Humulin R; Lilly), the amount generally used for insulin tolerance tests, was injected intraperitoneally twice daily. For bone marrow (BM) transplantation, recipient C57BL/6N mice at 8–10 weeks of age were irradiated twice with an individual dose of 5.4 Gy with a 3-h interval, and subsequently received an intravenous injection of 2 × 10⁶ BM cells from donor wild-type or GDF3 knockout mice. Mice were sacrificed after anesthetization by isoflurane inhalation. Blood was collected from the inferior vena cava using 23-gauge needles and syringes. Serum nonesterified fatty acid (NEFA) levels were measured as a marker of lipolysis by NEFA C-test (Wako).

Epididymal WAT (epiWAT) was minced and digested with 1 mg/mL collagenase type I (Invitrogen) for 1 h at 37°C during shaking. The digested cells were filtered through a 250-μm nylon mesh (Kyoshin Rikoh) and centrifuged at 50g for 10 min. The floating adipocytes were washed with PBS twice. After dispersing the pellet containing the stromal-vascular fraction (SVF), the medium was filtered through a 40-μm nylon mesh and centrifuged at 300g for 10 min. The pellet was then incubated with erythrocyte-lysing buffer consisting of 155 mmol/L NH₄Cl, 5.7 mmol/L K₂HPO₄, and 0.1 mmol/L EDTA at room temperature for 1 min and washed twice with PBS.

The cells in the SVF were resuspended in PBS, 2 mmol/L EDTA, and 2% FBS, and were incubated with excess Fc block (anti-CD16/CD32 antibodies; BD Biosciences) to block Fc receptor–mediated, nonspecific antibody binding. Cell surface markers were stained on ice in the dark for 20 min using CD11b-phycoerythrin-Cy7, F4/80-allophycocyanin (Tonbo Biosciences), and CD11c-phycoerythrin (BD Biosciences) monoclonal antibodies. Some cells were stained as negative controls with fluorochrome-matched isotype control antibodies. After excluding dead cells by staining with 7-aminoactinomycin D, live cells were subjected to characterization of cell populations or to sorting of specific cell populations by a FACSVerse or a FACSAriaII Flow Cytometer (BD Biosciences).

RNA was extracted using Sepasol-RNA I Super (Nacalai Tesque). Total RNA (1 μg) was reverse transcribed using oligo-(dT)_12–18 primer and Superscript III (Invitrogen). Quantitative PCR was performed with SYBR Premix Ex Taq (TaKaRa Bio) using a LightCyler 480 System (Roche). The results were normalized against 36B4 mRNA expression. The primer sequences are listed in Supplementary Table 1.

Rabbit polyclonal anti-ALK7 antibody was described previously (4). Rabbit monoclonal antibodies toward Smad3, phospho-Smad3 (Ser 423/425), Akt, and phospho-Akt (Ser 473) were purchased from Cell Signaling Technology. Rat monoclonal anti-Cripto and goat polyclonal anti-GDF3 antibodies were purchased from R&D Systems. Mouse monoclonal antibodies toward β-actin and α-tubulin were purchased from Sigma-Aldrich. For immunoblotting, isolated adipocytes and the SVF were lysed with buffer (20 mmol/L HEPES, pH 7.4, 150 mmol/L NaCl, 1% Triton X-100, 0.2 mmol/L EDTA, and 1 mmol/L dithiothreitol) containing protease and phosphatase inhibitors. The protein extracts (8–10 μg for macrophages and 20 μg for other cells) were loaded onto polyacrylamide gels for electrophoresis. For imaging of whole-mount epiWAT, euthanized mice were perfused with 40 mL of fresh 1% formaldehyde in PBS via intracardiac injection over a few minutes. EpiWAT was subdivided into small pieces (∼0.1 cm³) by scissors, and was then fixed in 1% formaldehyde in PBS and blocked in 5% BSA in PBS at room temperature for 30 min. For immunostaining of CD11c⁺ ATMs, cells attached on slide glasses by Cytospin (Thermo Fisher Scientific) were fixed with 3.7% formaldehyde in PBS for 30 min at room temperature. With permeabilization by 0.1% Triton X-100, the tissues or the cells were incubated with 10 μg/mL anti-GDF3 antibody or control IgG overnight at 4°C followed by the Alexa Fluor 488–conjugated secondary antibody (diluted at 1:500; Invitrogen) for 1 h at room temperature, and were observed under a laser-scanning confocal microscope. The concentration of GDF3 in a medium was measured by mouse GDF3 ELISA kit (Elabscience).

The binding site of Smad3 and Smad4 (CAGA)₁₄ (12) was inserted into the pGL4.10[luc2] vector (Promega). Human embryonic kidney 293 T (HEK293T) cells cultured in DMEM containing 10% FBS and 1 mmol/L l-glutamine were transfected with 20 ng of the reporter plasmid, 10 ng of the control plasmid pGLA474[hRluc/TK] (Promega), 12.5 ng of plasmid containing ALK7 cDNA (4), and 6.25 ng of that containing Cripto cDNA derived from mouse embryo, using Lipofectamine 2000 reagent (Invitrogen). After 48 h, the recombinant proteins of human GDF3, bone morphogenetic protein 3 (BMP3), activin B, and TGF-β1 (R&D Systems) were added to the medium. After a further 24 h, the luciferase activities were measured by the Dual-Luciferase Reporter Assay System (Promega). The light units were normalized to Renilla luciferase activity.

Isolated mouse adipocytes (600 μL) were incubated at 37°C for 3 h in Krebs-Ringer HEPES buffer (20 mmol/L HEPES, pH 7.4, 120 mmol/L NaCl, 5 mmol/L KCl, 2 mmol/L CaCl₂, 1 mmol/L MgCl₂, and 1 mmol/L KH₂PO₄) containing 2 mmol/L glucose and 1% fatty acid–free BSA. Lipolysis was assessed by measuring the concentration of glycerol in the buffer using a Free Glycerol Determination Kit (Sigma-Aldrich).

All quantitative data were expressed as the mean ± SD. Data analysis used GraphPad Prism software. The P values were calculated using Student t test, one-way ANOVA with Tukey multiple-comparison test, or repeated-measures ANOVA with Bonferroni multiple-comparison test, as appropriate, to determine significant differences between group means.

Because ALK7 knockout mice show reduced fat accumulation when fed an HFD, but exhibit normal weight when fed regular chow (9), the ALK7 signal could be activated under nutrient-excess conditions. We thus screened TGF-β superfamily members that exhibit differential expressions depending on nutritional states and also between the absence or presence of functional ALK7. For this purpose, we isolated tissues potentially involved in nutritional metabolism from ALK7-intact C57BL/6 and ALK7-deficient BALB/c lean mouse strains fed either regular chow or an HFD. We also isolated these tissues from ALK7-intact TSOD and ALK7-deficient T.B-Nidd5/3 obese mouse strains, both of which have the same genetic background (3,4). Among the 33 members of the mammalian TGF-β superfamily (13), GDF3, BMP3, inhibin-βB, and TGF-β1 showed differential expression in WAT (Fig. 1A and Supplementary Fig. 1). Their expression in WAT was strongly upregulated in C57BL/6 mice fed an HFD compared with those fed regular chow. Some of them were also upregulated in obese TSOD mice fed regular chow and even in ALK7-deficient BALB/c mice fed an HFD. In contrast to the other three ligands, GDF3 showed a remarkably high and specific expression in epiWAT of TSOD and HFD-fed C57BL/6 mice, which is consistent with previous findings (4,10). We then examined the ligand activity through ALK7 in HEK293T cells expressing a luciferase reporter containing a Smad3/4 responsive element (12), which acts downstream of ALK7 in adipocytes (4). Consistent with a previous finding (9), GDF3 activated the reporter in a dose-dependent fashion only in the presence of exogenously expressed ALK7 and Cripto, a coreceptor that enhances signaling via the type I and type II receptor kinase complex (14) (Fig. 1B). In contrast, BMP3 did not show such enhancement. Activin B, a dimer of inhibin-βB, and TGF-β1 activated the reporter even in the absence of ALK7 and Cripto, although both induced slight activation with the receptor expression. These findings make GDF3 the most likely candidate ligand for ALK7.

ALK7-deficient T.B-Nidd5/3 mice at 7 weeks of age showed a significant reduction in epiWAT weight compared with control TSOD mice (Supplementary Fig. 2A).The mice exhibited increased levels of mRNA encoding the transcription factors PPARγ and C/EBPα, and their downstream genes encoding adipose TG lipase (ATGL) and hormone-sensitive lipase (HSL), as previously reported in older mice (4). Serum levels of NEFA reflecting enhanced lipolysis were also elevated relative to control TSOD mice. Therefore, the ALK7-deficient phenotypes become overt at 7 weeks of age. GDF3 inhibited lipolysis in adipocytes from TSOD mice at this age, whereas BMP3, activin B, or TGF-β1 did not (Fig. 2A), which is consistent with the findings from the luciferase assays (Fig. 1B). Importantly, GDF3 inhibited lipolysis and activated the downstream Smad3 by phosphorylation only in ALK7-intact adipocytes from TSOD mice, but not in ALK7-deficient adipocytes from T.B-Nidd5/3 mice (Fig. 2B and C). These findings establish that GDF3 can signal through ALK7 in adipocytes.

Because GDF3 is expressed in thymus, spleen, and BM as well as in WAT (Fig. 1A), as originally reported (15), it might be expressed in hematopoietic cells rather than adipocytes in WAT. To identify the cell source of GDF3, we first dissociated the epiWAT into the SVF and mature adipocytes, then further fractionated SVF cells by fluorescence-activated cell sorting using fluorochrome-conjugated antibodies targeting macrophage surface markers (16). GDF3 transcripts were enriched in the SVF, particularly in CD11b⁺ F4/80⁺ macrophages (defined as ATMs), with the greatest elevation seen in those expressing CD11c (Fig. 2D and Supplementary Fig. 2B). Immunostaining with anti-GDF3 antibody revealed that most of the CD11c⁺ ATMs express GDF3 (94.8 ± 2.2%; n = 3: ∼100 cells were examined in total). GDF3⁺ cells were located around individual adipocytes in WAT, consistent with localization in ATMs. In contrast, BMP3 and inhibin-βB were expressed mainly in mature adipocytes, whereas TGF-β1 was ubiquitously expressed in every cell fraction (Supplementary Fig. 2C). Concomitant increases in the CD11c and GDF3 transcripts were also found in the SVF of HFD-fed C57BL/6 mice (Supplementary Fig. 1D). Although inflammasome activation has recently been shown to induce GDF3 in ATMs from aged mice (17), the GDF3 induction in TSOD or HFD-treated C57BL/6 mice was not accompanied by the upregulation of inflammasome activation–related genes, such as tumor necrosis factor-α (TNF-α), monocyte chemotactic protein-1 (MCP-1), NLR family pyrin domain containing 3 (NLRP3), and caspase-1 (Supplementary Fig. 1B and D).

To evaluate the role of GDF3-producing ATMs in vivo, we intraperitoneally injected clodronate to deplete macrophages (18). Clodronate treatment partially but significantly decreased the percentage of ATMs, including that of CD11c⁺ ATMs, as well as the expression of F4/80, CD11c, and GDF3, in both TSOD and T.B-Nidd5/3 mice (Fig. 3A and Supplementary Fig. 3A and B). However, clodronate decreased total body weight, particularly epiWAT weight, only in TSOD mice, indicating that the effects of the drug depend on intact ALK7. Furthermore, clodronate increased the PPARγ, C/EBPα, ATGL, and HSL transcripts, and the serum NEFA concentration normalized to the epiWAT weight, in TSOD mice (Fig. 3B). Therefore, the effects of macrophage depletion from ALK7-intact TSOD mice are remarkably similar to the phenotypic changes in ALK7-deficient T.B-Nidd5/3 mice when compared with control TSOD mice (3,4), indicating that the GDF3-ALK7 axis represents a major link between macrophages and adipocytes in the regulation of whole-body lipid metabolism and fat accumulation.

We next explored the external factors that increase GDF3 production under nutrient-excess conditions. CD11c⁻ ATMs isolated from epiWAT of TSOD mice showed elevated levels of the GDF3 transcript during culture in FBS-containing medium (Supplementary Fig. 4A), suggesting that some FBS component converts CD11c⁻ to CD11c⁺ ATMs and concomitantly induces GDF3 expression. Because obesity is frequently coincident with hyperinsulinemia, we suspected that insulin might upregulate GDF3. Plasma insulin concentrations are ∼170 pmol/L in lean BALB/c mice and ∼1.7 nmol/L in obese TSOD mice (2). Ex vivo administration of 10 μU/mL insulin (61 pmol/L) increased expression of both CD11c and GDF3 after a 24-h culture in CD11c⁻ macrophages derived from epiWAT of TSOD mice, and wortmannin, an inhibitor of phosphatidylinositol 3-kinase, inhibited insulin-induced GDF3 upregulation (Fig. 4A). Insulin also increased the expression of the typical M2 markers arginase and chitinase-like 3, but not that of the M1 markers TNF-α and MCP-1. Although 61 pmol/L insulin induced GDF3 in ATMs, it increased GDF3 only weakly in macrophages derived from lung, peritoneum, or BM of TSOD mice (Supplementary Fig. 4B). This was evident in the low level of expression of the insulin receptor in these macrophages in contrast to that in ATMs. These findings indicate the tissue selectivity of insulin sensitivity in macrophages.

Although the above findings raise the possibility that insulin inhibits lipolysis and accumulates fat in adipocytes through the upregulation of GDF3 in ATMs, insulin is generally believed to do so by directly acting on adipocytes. We next investigated the effects of insulin on isolated adipocytes. We confirmed that insulin phosphorylates the downstream Akt kinase, but does not activate Smad3 by a noncanonical pathway, in adipocytes (Fig. 4B). However, the concentration of insulin (61 pmol/L) we administered ex vivo to ATMs (Fig. 4A) did not inhibit basal or catecholamine-induced lipolysis in adipocytes, although a higher concentration of insulin (25 nmol/L) did so (Fig. 4C). These findings indicate that a much higher dose of insulin is required to directly inhibit lipolysis in adipocytes than is required to upregulate GDF3 in ATMs. Although ALK7 deficiency has been reported to enhance catecholamine-induced lipolysis in adipocytes (19), we found that unstimulated lipolysis is already elevated and that the extent of stimulation by catecholamine is not changed in ALK7-deficient adipocytes (Fig. 4C). These findings confirm the previous finding that ALK7 deficiency elevates basal lipolysis by affecting the expression levels of adipose lipases (4).

To reinforce the functional significance of the activity of insulin through GDF3 production from ATMs, we performed reconstitution assays by incubating adipocytes with the supernatant of CD11c⁻ ATMs that had been treated with or without 61 pmol/L insulin. Note that this concentration of insulin does not directly inhibit lipolysis in isolated adipocytes (Fig. 4C). Insulin induced the secretion of GDF3 into their supernatants (1–2 ng/mL), which dose-dependently increased the phosphorylation of Smad3 and inhibited lipolysis in adipocytes of ALK7-intact TSOD mice, but not in those of ALK7-deficient T.B-Nidd5/3 mice (Fig. 4D and E). We confirmed that 61 pmol/L insulin did not change the expression levels of inhibin-βB and TGF-β1 in the ATMs (Supplementary Fig. 4C), both of which can induce Smad3 phosphorylation in adipocytes. These findings indicate that the insulin-stimulated release of GDF3 from ATMs successfully inhibits lipolysis in adipocytes ex vivo.

To clarify whether insulin functions through the GDF3-ALK7 signaling pathway in vivo, we intraperitoneally administered insulin twice a day for 2 weeks to TSOD and T.B-Nidd5/3 mice. This in vivo insulin treatment elevated the WAT weight and decreased the levels of the ATGL and HSL transcripts and serum NEFA in an ALK7-dependent manner (Fig. 5A), suggesting that insulin inhibits lipolysis and accumulates fat through the upregulation of GDF3 in ATMs.

In order to exclude the possibility that the effects of insulin via the GDF3-ALK7 axis are applicable only to the TSOD strain, for which the molecular pathogenesis of obesity and diabetes is unknown (2), we administered insulin to a commonly used C57BL/6 strain fed an HFD that indeed expressed ALK7 in WAT (Supplementary Fig. 5A). Insulin increased adiposity in parallel with reductions in the expression of adipose lipases in epiWAT and serum NEFA concentrations in C57BL/6 mice (Fig. 5B). However, no such effects were found in ALK7-deficient BALB/c mice fed an HFD. These findings indicate that the effects of insulin via ALK7 under nutrient-excess conditions continue irrespective of the mouse strain.

To further substantiate the role of the GDF3-ALK7 axis in insulin activity in vivo, we injected clodronate to deplete macrophages and then administered insulin to C57BL/6 mice fed an HFD. We confirmed that neither clodronate nor insulin treatment alters the food intake of mice (Supplementary Fig. 5B). Clodronate treatment markedly decreased ATMs, including CD11c⁺ ATMs, and concomitantly reduced GDF3 levels in the SVF (Fig. 6A and Supplementary Fig. 5B). Remarkably, it eliminated the in vivo effects of insulin to increase CD11c and GDF3 in the SVF and adiposity in the whole body, and to decrease adipose lipases and the serum concentration of NEFA. These findings demonstrate that insulin can regulate fat metabolism and mass through its effects on macrophages in vivo.

Finally, we performed BM transplantation experiments to directly prove the involvement of GDF3 in the insulin activity. We transplanted the BM of GDF3-deficient C57BL/6 mice (10) to wild-type C57BL/6 mice to evade the cell elimination by the immune system due to MHC mismatch. The recipient mice were then fed an HFD and treated with insulin. We confirmed that GDF3 deficiency in BM cells does not affect the number of ATMs (Supplementary Fig. 5C). In contrast to the mice harboring the wild-type BM, those harboring the GDF3-deficient BM, and thus losing GDF3 in the SVF failed to mediate the in vivo effects of insulin to inhibit lipolysis in the WAT (Fig. 6B). These findings demonstrate that GDF3 production is necessary for insulin to regulate fat metabolism and mass under nutrient-excess conditions.

We showed that GDF3 produced from CD11c⁺ ATMs acts as a ligand of ALK7 in adipocytes to inhibit lipolysis and accumulate fat under nutrient-excess conditions. The GDF3-ALK7 axis within WAT should represent the major interactive mechanism between macrophages and adipocytes in the regulation of adiposity, because nonselective macrophage depletion by clodronate highlights the ALK7-specific effects, such as decreases in body and epiWAT weights, and increases in the expressions of C/EBPα, PPARγ, ATGL, and HSL, as well as NEFA production in WAT, in ALK7-intact TSOD mice, but not in their ALK7-deficient counterparts. Although many studies have focused on the effects of macrophages in the formation of chronic inflammation associated with obesity, the current study demonstrates the role of ATMs in fat accumulation per se. Although CD11c⁺ macrophages are conventionally understood to be M1 macrophages that are recruited to and/or polarized in obese WAT to induce a chronic inflammatory state (20), the GDF3-producing cells express a substantial level of M2 markers. Similar to our findings, it has recently been shown that a prototypical M2 marker, CD301b, as well as arginase, are selectively expressed in CD11c⁺ mononuclear phagocytes including ATMs, and that depleting these cells leads to weight loss and increased insulin sensitivity in mice (21).

We found that a physiologically low concentration of insulin alters the properties of CD11c⁻ ATMs ex vivo by increasing the expressions of CD11c and GDF3. Moreover, in vivo insulin administration inhibits lipolysis and expands WAT in an ALK7-dependent manner, which indicates that insulin regulates fat metabolism and mass via the GDF3-ALK7 axis. Consistently, the in vivo effects of insulin on WAT are absent after the depletion of macrophages or the transplantation of GDF3-deficient BM. It is intriguing that ATMs appear to specifically express a high level of insulin receptor compared with macrophages in other tissues. Although insulin is generally thought to inhibit lipolysis directly in adipocytes by regulating the cAMP-mediated signaling pathway (22–25) and/or by suppressing the transcription of adipose lipases (26–28), these actions in adipocytes have been detected only at higher concentrations of insulin (1–100 nmol/L) compared with those applied to ATMs in the current study (61 pmol/L). In fact, we observed that 25 nmol/L insulin can directly inhibit both basal and catecholamine-stimulated lipolysis in adipocytes, whereas 61 pmol/L insulin cannot. Therefore, insulin can differentially regulate fat metabolism and mass depending on its local concentration in WAT. The novel pathway through ATMs might explain the exquisite sensitivity of lipolysis to suppression by insulin infused into human subjects (half-maximal suppression at 101 pmol/L) (29).

Given that GDF3 induction in ATMs requires a minimal concentration of insulin, the GDF3-ALK7 pathway should be active at the beginning of hyperinsulinemia under nutrient-excess conditions. This has a clinical implication for the importance of “early intervention” in adiposity before the manifestation of insulin resistance. Insulin resistance–related chronic hyperinsulinemia may accelerate fat accumulation even under the same energy balance via activation of the GDF3-ALK7 axis, which makes it much harder for obese individuals to reduce adiposity. Future research should focus on novel targeting strategies for this pathway, such as inhibitors of GDF3 and ALK7, specific depletion of ATMs, and macrophage-specific inhibition of insulin receptor expression.

In summary, we present a novel mechanism of obesity (Fig. 6C). Under nutrient-excess conditions, insulin efficiently activates insulin receptors expressed on CD11c⁻ ATMs and converts them to CD11c⁺ ATMs to produce GDF3. GDF3 locally stimulates ALK7 on adipocytes and activates Smads 2–4 to downregulate PPARγ, C/EBPα, and also adipose lipases to store excess nutrient as fat (4). However, persistent activation of this physiological pathway enlarges adipocytes and may change adipocytokine repertoires to cause chronic inflammation and insulin resistance. In fact, ALK7-intact, aged obese mice exhibit elevated levels of proinflammatory MCP-1 and TNF-α, a reduced level of insulin-sensitizing adiponectin, and greater glucose intolerance, compared with their ALK7-deficient counterparts (4,5). As such, the insulin-GDF3-ALK7 axis plays a pivotal role in both physiological and pathological fat accumulation in WAT.

Acknowledgments. The authors thank the members of the Laboratory of Molecular Endocrinology and Metabolism, Gunma University, particularly T. Nara, E. Kobayashi, and T. Ushigome for colony maintenance of mice and S. Shigoka for assistance in preparing the manuscript. The authors also thank the staffs at Bioresource Center, Gunma University for help in the breeding of the mice.

Funding. This work was supported by the Japan Society for the Promotion of Science KAKENHI grant JP25860739 to S.Y. and grants JP24659442 and JP25126702 to T.I. and grants from the Japan Diabetes Foundation and the Novo Nordisk Insulin Study Award (to T.I.).

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

Author Contributions. Y.B., S.Y., and K.M. performed experiments. K.O. performed experiments, designed experiments, and wrote the paper. M.J.I. and C.W.B. provided experimental reagents. T.I. designed experiments and wrote the paper. T.I. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.