Spred2 KO mice onto a C57BL/6J background were kindly provided by Dr. Akihiko Yoshimura (25, 26). The lack of Spred2 expression in the mesenteric white adipose tissue (mWAT) and the liver of Spred2 KO mice was confirmed by TaqMan real-time quantitative polymerase chain reaction (RT-qPCR) (data not shown). C57BL/6J mice were used as wild-type (WT) control. WT and Spred2 KO mice were bred and maintained under continuous 12-h light and 12-h dark cycle under specific pathogen-free conditions at the Department of Animal Resources, Okayama University, Okayama, Japan. Male mice were used in this study. All animal protocols were approved by the Animal Care and Use Committee of the Okayama University, and all experiments were performed in accordance with relevant guidelines and regulations.

Mice were weaned from parents at 4 weeks of age and given free access to water and either the standard diet (SD: 359 kcal/100 g; 12.8% energy as fat; MF, Oriental Yeast Co. Ltd., Tokyo, Japan), or high fat diet (HFD: 506 kcal/100 g; 60.0% energy as fat; HFD-60, Oriental Yeast Co. Ltd.). At different time intervals, mice were anesthetized, blood was collected by cardiac puncture, and mWAT and liver were excised. A part of the tissue was snap-frozen in liquid nitrogen and stored at −80°C for subsequent analyses. Another part of the tissue was fixed in 10% formalin, embedded in paraffin, and thin sections were prepared and histologically examined. Alternatively, livers were frozen in O.C.T compound and cryostat sections were stained by oil red O as per the standard techniques. In another set of experiments, a whole mWAT was resected and used for isolation of adipocytes and stromal vascular fraction as described below.

Mice (20 weeks of age) were anesthetized with isoflurane (Mylan Inc., Canonsburg, PA, USA) and scanned by micro-CT (LaTheta LCT-200; Hitachi Aloka Medical, Ltd., Tokyo, Japan). The abdominal adiposity was assessed using image-analyzing software (LaTheta v1.20).

Serum total cholesterol, triglyceride, and blood glucose concentrations were measured using standard protocols. The serum insulin concentrations were measured using an insulin assay kit (Morinaga Institute of Biological Science Inc., Tokyo, Japan). The serum FFA concentrations were measured using a LabAssay NEFA kit (Wako Pure Chemical Industries Inc., Osaka, Japan).

After 16 weeks of feeding (20 weeks of age), glucose and insulin tolerance tests were conducted by intraperitoneally injecting glucose (1 g/kg body weight, Sigma-Aldrich, St. Louis, MO, USA) and insulin (1 U/kg body weight, Eli Lilly Japan K.K., Hyogo, Japan), respectively, as previously described (27). Blood was sampled from the tail vein of mice, and blood glucose level was evaluated using a blood glucose test meter (Lab Gluco: ForaCare Suisse AG, St. Gallen, Switzerland).

mWAT was minced and digested using type II collagenase (Sigma-Aldrich) dissolved in PBS at 37°C for 20 min on an orbital shaker. The digested tissues were centrifuged at 500 × g for 10 min. The pelleted cells represented the SVF, and the floating cells represented the adipocyte fraction (28).

Bone marrow (BM) was harvested from the femur and tibia of WT or Spred2 KO mice. For generation of BMDMs, BM cells were cultured in RPMI-1640 medium, supplemented with 10% FBS and antibiotics, in the presence of M-CSF (50 ng/ml, PeproTech, Inc., Rocky Hill, NJ, USA) for 7 days. Adherent BMDMs were collected and plated onto 24-well plates at a density of 5 × 10⁵ cells per well. Cells were incubated overnight in serum-free RPMI-1640 medium, and then stimulated by bovine serum albumin (BSA)-conjugated palmitic acid (PA, 500 μM, Cayman chemical, Ann Arbor, MI, USA) for 24 h. BSA-conjugated PA was prepared as previously described (29) with modifications. In another set of experiments, BMDMs were pre-treated with U0126 (10 μM, Promega, Madison, WI, USA) or vehicle (DMSO) for 1 h and then cultured with BSA-conjugated PA for 24 h.

3T3-L1 preadipocytes (American Type Culture Collection) were cultured in DMEM containing 10% FBS and antibiotics. Differentiation of 3T3-L1 preadipocytes to mature adipocytes was performed, as described (30). Adipocyte differentiation was determined by oil red O staining.

A total of 1 × 10⁶ adipocytes differentiated from 3T3-L1 preadipocytes at day 8 were transfected with Spred2-specific or non-targeting control siRNAs (final concentration 5 pmol; Thermo Scientific, Yokohama, Japan) (17, 31) using a Lipofectamine RNAiMAX (Thermo Scientific) according to the manufacturer's instructions and plated in a 24-well plate. For plasmid transfections, BMDMs were seeded into 24-well plate at a density of 2 × 10⁵ cells per well and incubated overnight in a 5% CO₂ incubator at 37°C. For each transfection, 0.75 μL Lipofectamine 3000 (Life Technologies, Carlsbad, CA, USA) and 0.5 μg Spred2 expression plasmid (kindly provided by Dr. Masakiyo Sakaguchi, Okayama university, Japan) were added to 50 μL Opti-MEM (Life Technologies) and incubated for 10 min at room temperature before the mixtures were added to the cells. Cells that overexpressed Spred2 were grown in RPMI-1640 medium with 10% FBS for 24 h. Cells were then incubated overnight in serum-free RPMI-1640 medium, after which the cells were cultured with BSA-conjugated PA (500 μM) for 24 h.

Epididymal fat pads were minced and digested using type II collagenase (Sigma-Aldrich) dissolved in PBS at 37°C for 20 min on an orbital shaker. The digested tissues were centrifuged at 280 × g for 5 min. The infranatant containing the collagenase solution was removed, and the floating layer of adipocytes was washed three times with PBS. For lipolysis assay, aliquots of isolated adipocytes were incubated in the absence or presence of 100 ng/ml TNFα (R&D systems Inc., Minneapolis, MN, USA) for 3 h. Glycerol content in the infranatant medium was measured using a glycerol colorimetric assay kit (Cayman chemical). To investigate FA uptake, isolated adipocytes were incubated with 20 μM BODIPY FL C16 (Thermo Scientific) in DMEM containing 1%FA-free BSA (Sigma) in the absence or presence of 100 nM insulin (Takara Bio Inc., Shiga, Japan) for 1 h. After the transport period, cells were washed with PBS and then lysed in 0.05N NaOH (32). Fluorescence intensity was measured using FlexStation 3 multi-mode microplate reader (Molecular Devices, LLC. San Jose, CA, USA) at Ex 503 and Em 512 nm, and normalized to intra-droplet fluorescence intensity.

mWAT paraffin sections were stained with hematoxylin and eosin, and the adipocyte size was assessed using the NIH's ImageJ. For immunohistochemistry, sections were incubated with either 10 μg/ml of rat anti-F4/80 antibody (Bio-Rad Laboratories, Inc., Hercules, CA, USA), 2 μg/ml of rabbit anti-iNOS antibody (Abcam plc., Cambridge, MA, USA), or 10 μg/ml of rat anti-CD206 antibody (Bio-Rad) for 90 min, followed by incubation with either Polink-2 Plus HRP Rat-NM with DAB kit or Polink-2 Plus HRP Rabbit with DAB kit (GBI, Inc., Bothell, WA, USA) according to the manufacturer's instructions. Antigen was visualized using 3,3′-diaminobenzidine substrate (Sigma-Aldrich).

Total RNA was extracted from mWAT, adipocytes, or SVF using TRIsure reagent (Bioline Reagents Ltd., London, UK). cDNA was synthesized by the High Capacity cDNA Reverse Transcription Kit (Thermo Scientific) using 500 ng of total RNA. RT-qPCR was performed using StepOne Plus Real-Time PCR system with Taqman PCR master mix (Thermo Scientific). The primers used in this study were: Spred2 (Mn01223872_g1), tnfa (Mm00443258_m1), mcp-1 (Mm00441242_m1), adiponectin (Mm00456425_m1), leptin (Mm00434759_m1), pai1 (Mm00435858_m1), and rplp0 (Mm00725448_s1) (Thermo Scientific). The expression level of the gene of interest was normalized against rplp0 and expressed as fold-increases relative to the negative control for each treatment at each time point as previously described (17, 33).

Cells were lysed using a lysis buffer (Cell Signaling Technology Inc., Danvers, MA, USA). After centrifugation, clear supernatants were collected and stored at −80°C until further use. The protein concentration in the lysates was measured using protein-dye binding assay (Bio-Rad). Equal amounts (15 μg) of cell lysates were fractionated using sodium dodecyl sulfate–polyacrylamide (SDS) gel electrophoresis (Thermo Scientific), and then proteins were transferred to nitrocellulose membrane. After blocking with TBS-T (Tris-buffered saline containing 0.1% Tween-20) containing 5% BSA, the membranes were incubated with an antibody against either phosphorylated ERK or total ERK (Cell Signaling Technology Inc.) overnight. After washing with TBS-T, membranes were incubated with horseradish peroxidase-conjugated anti-rabbit IgG antibody (Cell Signaling Technology Inc.), and each protein present on the membrane was visualized using an enhanced chemiluminescence system (ImmunoStar LD; Wako Pure Chemical Industries Inc.). Blots were photographed and analyzed using the Image Studio software.

The concentration of TNFα or MCP-1 was measured using a standard sandwich ELISA, as previously described (34, 35). The primary antibody, detection antibody, and recombinant TNFα and MCP-1 were purchased from R&D Systems. The ELISAs used in this study did not cross-react with other known murine cytokines.

All data were expressed as the mean ± SEM. Data were analyzed using the GraphPad Prism software (GraphPad Software, San Diego, CA). Student's t-test and analysis of variance (ANOVA) were performed to determine the statistical significance. P-values smaller than 0.05 were considered significant.

We first compared the weights of WT and Spred2 KO mice fed with SD. Spred2 KO mice exhibited increased body weight compared to WT mice (Figure 1A). Since HFD leads to obesity, particularly in C57BL/6J background (36, 37), we next fed WT and Spred2 KO mice with HFD and examined their weights. As shown in Figure 1A, HFD led to increased body weight in both WT and Spred2 KO mice; however, Spred2 KO mice gained significantly more weight than WT mice (27% heavier) than WT mice after 20 weeks. No difference was found in food consumption between the groups (Figure 1B). There was also no significant difference in the body length between the two strains. CT images demonstrated increased mass of abdominal fat in Spred2 KO mice than in WT mice (Figure 1C). mWATs were subsequently weighed after 16 weeks of SD or HFD. mWATs of Spred2 KO mice were heavier than those of WT mice, when fed with either SD or HFD (Figure 1D). These results clearly indicated that Spred2-deficiency enhanced diet-induced obesity.

Obesity is associated with dynamic adipose tissue remodeling characterized by adipocyte hypertrophy (38, 39). As shown in Figures 2A,B, the size of adipocytes in mWAT increased with age (in weeks) in HFD-fed WT and Spred2 KO mice. The size of adipocytes in Spred2 KO-mWAT was significantly larger than that in WT-mWAT, especially after 16 weeks of HFD (Figure 2B). The distribution of adipocyte size shifted toward larger size in Spred2 KO-mWAT after 16 weeks of HFD (Figure 2C).

HFD may affect the level of Spred2 mRNA expression in mWAT. As shown in Figure 2D, Spred2 mRNA expression in WT-mWAT significantly decreased when fed with HFD compared to SD. Next, the serum lipid concentration was measured (Figure 2E). HFD caused an elevation in serum levels of total cholesterol and triglycerides at the age of 20 weeks. FFA levels in HFD-Spred2 KO mice were also significantly higher than those in WT mice after 8 weeks of HFD and remained elevated after 16 weeks of HFD (Figure 2F). Thus, over-nutrition caused by HFD exacerbated dyslipidemia in Spred2 KO mice.

Dyslipidemia may lead to low-grade inflammation, characterized by macrophage infiltration in adipose tissue (40). Immunohistochemical analyses revealed that the number of F4/80 positive macrophages (Figure 3A) in Spred2 KO-mWAT was significantly higher than that in WT-mWAT (Figure 3B, left). Crown-like structures, composed of macrophages surrounding dead or dying adipocytes (Figure 3A) (41), were significantly increased in Spred2 KO-mWAT (Figure 3B, right). Macrophages migrating in mWAT are known to have M1 phenotype (pro-inflammatory macrophages) (42). Consequently, we performed immunostaining using antibodies against either inducible nitric oxide synthase (iNOS) or CD206, known representative markers for M1 or M2, respectively, (43). The number of iNOS⁺ macrophages (Figure 3C) was significantly higher in Spred2 KO-mWAT than that in WT-mWAT (Figure 3D), whereas there were only a few CD206⁺ macrophages in both groups and M1:total macrophage ratio was similar between both groups (data not shown). These results suggested that Spred2 KO mice exhibited exacerbated infiltration of pro-inflammatory macrophages.

Obesity-induced adipose tissue inflammation leads to insulin resistance (4, 44, 45). As shown in Figure 4A, serum levels of glucose in Spred2 KO mice were significantly higher than those in WT mice, although the insulin levels were similar between the groups. Consequently, in vivo metabolic tests were performed, which showed that both glucose intolerance and insulin resistance were aggravated in Spred2 KO mice when compared with WT mice (Figures 4B,C). Thus, Spred2 KO mice showed impaired insulin resistance.

The liver plays a principal role in lipid metabolic pathways. Increased serum lipid levels and impaired insulin resistance may cause non-alcoholic fatty liver disease (NAFLD)/non-alcoholic steatohepatitis (NASH) (46) in Spred2 KO mice. The expression of Spred2 in WT-livers tended to increase when mice were fed with HFD (p = 0.12, Figure 5A). The data in Figure 5B demonstrated that HFD-Spred2 KO-livers were heavier than WT-livers. Histologically, in comparison with WT-livers, Spred2 KO-livers revealed moderate vesicular steatosis, and influx of neutrophils and F4/80-positive macrophages. Perivenular/pericellular fibrosis was not evident by Masson's trichrome staining (Figure 5C). The serum levels of ALT, a marker of liver injury, tended to increase in Spred2 KO mice as compared with WT mice (p = 0.14, after 16 weeks of HFD, Figure 5D). These data indicated that Spred2 KO mice developed NAFLD/NASH.

Macrophage infiltration in mWAT can be induced by cytokines, among which TNFα and MCP-1 appear to be crucial (47, 48). To examine the cytokine response, we examined the expression of TNFα and MCP-1 mRNA in mWATs obtained from mice fed with HFD. After 16 weeks of HFD, mRNA levels of TNFα and MCP-1 in Spred2 KO-mWAT tended to be higher than those in WT-mWAT (Figure 6A), although the data was not statistically significant (TNFα; p = 0.43, MCP-1; p = 0.10). Next, mWATs were divided into adipocytes and stromal vascular fraction (SVF) that contained preadipocytes, endothelial cells, and macrophages (49), and the cytokine responses of both fractions were examined. As shown in Figure 6B, expression levels of TNFα and MCP-1 were 2.2-fold and 3.1-fold higher in Spred2 KO-SVF, respectively, than those in WT-SVF. The expression level of TNFα was higher in Spred2 KO-adipocytes than that in WT-adipocytes, but its expression level was lower than that in the SVF. No difference was found in the expression levels of adipocyte-MCP-1 between the mouse groups (Figure 6B). Interestingly, expression of circulating MCP-1/CCL2 tended to be higher in Spred2 KO mice than that in WT mice (p = 0.09), despite no detection of TNFα in either group (Figure 6C). The levels of other adipokines, including adiponectin, leptin, and plasminogen activator inhibitor-1 (PAI-1) in the adipocytes were similar between the groups, although, a decreasing trend was observed in the levels of adiponectin (p = 0.14, 6 mice each, Supplementary Data 1). These results suggested that SVF preferentially plays an important role in the regulation of adipose tissue inflammation by generating TNFα and MCP-1.

Spred2 in adipocytes may regulate cytokine production and lipid metabolism. For cytokine production, we knocked down Spred2 expression (~80% reduction) in adipocytes differentiated from 3T3-L1 preadipocytes using short interfering (si)RNA and the cells were stimulated with TNFα (10 ng/ml) for 24 h. Spred2 knockdown did not alter the expression levels of de novo TNFα and MCP-1 as compared to the controls (data not shown), suggesting that the role of Spred2 in adipocyte cytokine production is minimal. For lipid metabolism, we compared glycerol release and fatty acid uptake between primary WT-and Spred2 KO-adipocytes. No difference was found in the glycerol release (WT vs. Spred2 KO, 3.12 ± 0.07 vs. 3.42 ± 0.24 mg/l/10⁶ cells, n = 4 each, not significant) and fatty acid uptake (WT vs. Spred2 KO, 1.20 ± 0.20 vs. 1.32 ± 0.26/normalized fluorescence, n = 4 each, not significant) between the groups after the stimulation with TNFα and insulin, respectively. To examine the involvement of Spred2 in adipocyte hypertrophy, we knocked down Spred2 expression (~80% reduction) in adipocytes differentiated from 3T3-L1 preadipocytes and cultured for 21 days. After oil red O staining, adipocyte hypertrophy was quantified by extracted oil red O level. Spred2 knockdown did not alter the amount of lipid contents (control vs. Spred2 knockdown, 0.52 ± 0.02 vs. 0.49 ± 0.07/absorbance at 490 nm, n = 3 each, not significant). These data suggest that Spred2 in adipocytes does not play a significant role in cytokine production and lipid metabolism.

We hypothesized that macrophages, a major structural component of SVF, are the main cells involved in the production of cytokines. To confirm the role of macrophages in mWAT, we employed bone marrow-derived macrophages (BMDMs), as majority of adipose tissue macrophages are derived from the bone marrow (50). Upon stimulation of macrophages with SFA (palmitic acid; PA), Spred2 KO-macrophages produced significantly higher levels of TNFα and MCP-1 than WT-BMDMs (Figure 7A). However, pretreatment with U0126, a selective MEK/ERK inhibitor, significantly reduced the production of TNFα and MCP-1 in both WT-and Spred2 KO-macrophages (Figure 7A), suggesting that the cytokine response was dependent on the ERK/MAPK pathway. In addition, the phosphorylation of ERK was significantly enhanced in Spred2 KO-macrophages compared to WT-macrophages after PA-stimulation (Figure 7B, left). The phosphorylation of ERK in WT-macrophages returned to the basal level by 4 h, whereas levels of ERK phosphorylation remained elevated in Spred2 KO-macrophages (Figure 7B, right). Thus, macrophages appear to be responsible for the enhanced production of TNFα and MCP-1 in absence of Spred2.

The above data raised the question of whether greater amounts of Spred2 in a cell could reduce cytokine responses. To address this, we overexpressed Spred2 in WT-macrophages using an overexpression plasmid that increased Spred2 expression by 90-fold, as determined by Taqman RT-qPCR (data not shown). Although no difference was found in the level of TNFα, Spred2 overexpression significantly reduced MCP-1 level after PA stimulation compared with the control cells (Figure 8). Thus, Spred2 overexpression effectively reduced PA-induced MCP-1, but not TNFα response. Endogenous Spred2 in macrophages may be sufficient to down-regulate TNFα response.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.