Mangiferin Ameliorates Obesity-Associated Inflammation and Autophagy in High-Fat-Diet-Fed Mice: In Silico and In Vivo Approaches

Obesity-induced insulin resistance is the fundamental cause of metabolic syndrome. Accordingly, we evaluated the effect of mangiferin (MGF) on obesity and glucose metabolism focusing on inflammatory response and autophagy. First, an in silico study was conducted to analyze the mechanism of MGF in insulin resistance. Second, an in vivo experiment was conducted by administering MGF to C57BL/6 mice with high-fat-diet (HFD)-induced metabolic disorders. The in silico analysis revealed that MGF showed a high binding affinity with macrophage-related inflammatory cytokines and autophagy proteins. In the in vivo study, mice were divided into three groups: normal chow, HFD, and HFD + MGF 150 mg/kg. MGF administration to obese mice significantly improved the body weight, insulin-sensitive organs weights, glucose and lipid metabolism, fat accumulation in the liver, and adipocyte size compared to HFD alone. MGF significantly reduced the macrophages in adipose tissue and Kupffer cells, inhibited the gene expression ratio of tumor necrosis factor-α and F4/80 in adipose tissue, reduced the necrosis factor kappa B gene, and elevated autophagy-related gene 7 and fibroblast growth factor 21 gene expressions in the liver. Thus, MGF exerted a therapeutic effect on metabolic diseases by improving glucose and lipid metabolism through inhibition of the macrophage-mediated inflammatory responses and activation of autophagy.


Introduction
The prevalence of metabolic diseases, such as obesity, arteriosclerosis, hyperlipidemia, and diabetes, is increasing worldwide [1]. Insulin resistance is a common pathogenic event in type 2 diabetes, obesity, non-alcoholic fatty liver disease, and cardiovascular disease. Fat accumulation in the liver and muscles, where high proportions of glucose and lipid metabolism occur, is the main reason for obesity-induced insulin resistance [2,3]. Additionally, obesity-induced chronic inflammation is strongly associated with the pathogenesis of insulin resistance [4]. An increased adipose tissue accumulation due to excessive energy consumption can cause inflammatory responses and insulin signaling disorders by increasing the secretion of inflammatory cytokines from macrophages [5,6]. Recently, studies on the relationship of autophagy with metabolic diseases have been increasing [7]. Suppression of autophagy-related gene 7 (ATG7) impairs insulin signaling, whereas restoration of ATG7 expression improves insulin resistance [8].
The rhizome of Anemarrhenae asphodeloides belongs to the family Liliaceae and is cultivated in Korea, Mongolia, China, and other Eastern countries. It is 0.5-1.5-cm wide and disc-shaped and diverges its surroundings. Anemarrhenae Rhizoma has been used to treat febrile diseases accompanied by high fever, thirst, dry cough, fever, and diabetes [9]. A review focused on pharmacological and biological effects of MGF on metabolic disorders showed that MGF, an Anemarrhenae Rhizoma polyphenol, alleviates obesity and fatty liver,

Effects of MGF on Weight Gain
First, food intake and body weight (BW) of each group were measured to assess the effects of MGF. A significant difference was found in comparing liver, body, and epididymal fat weights between the high-fat diet (HFD) and normal chow (NC) groups at week 14, as the former group was higher than the latter. (BW, 51.92 ± 1.32 g vs. 28.86 ± 1.29 g, p < 0.001; epididymal fat, 2.30 ± 0.18 g vs. 0.41 ± 0.04 g, p < 0.001; liver, 2.26 ± 0.18 g vs. 1.18 ± 0.05 g, p < 0.001). Compared to the HFD group, the MGF group had significantly lower BW gain, epididymal fat, and liver weight (BW change, 25.20 ± 1.04 vs. 31.62 ± 1.03 g, p < Binding energies are expressed on the middle top section of each sub-panel. The top three shortest interaction distances are displayed in pink, and the interacting amino acids are shown in yellow. MGF, mangiferin; TNF-α, tumor necrosis factor-α; IL-6, interleukin-6; NF-κB, nuclear factor kappa B; ATG7, autophagy-related gene 7; FGF21, fibroblast growth factor 21.

Effects of MGF on Weight Gain
First, food intake and body weight (BW) of each group were measured to assess the effects of MGF. A significant difference was found in comparing liver, body, and epididymal fat weights between the high-fat diet (HFD) and normal chow (NC) groups at week 14, as the former group was higher than the latter. (BW, 51.92 ± 1.32 g vs. 28.86 ± 1.29 g, p < 0.001; epididymal fat, 2.30 ± 0.18 g vs. 0.41 ± 0.04 g, p < 0.001; liver, 2.26 ± 0.18 g vs. 1.18 ± 0.05 g, p < 0.001). Compared to the HFD group, the MGF group had significantly lower BW gain, epididymal fat, and liver weight (BW change, 25.20 ± 1.04 vs. 31.62 ± 1.03 g, p < 0.01; epididymal fat, 1.74 ± 0.12 vs. 2.30 ± 0.18 g, p < 0.05; liver, 1.19 ± 0.14 vs. 2.26 ± 0.18 g, p < 0.01; Figure 3a-d). The HFD group showed significantly more calories per day than the NC group (24.19 ± 3.76 g vs. 9.21 ± 0.64 g, p < 0.001). The calories consumed by the MGF group did not differ significantly from the HFD group ( Figure 3).  Figure 3a-d). The HFD group showed significantly more calories per day than the NC group (24.19 ± 3.76 g vs. 9.21 ± 0.64 g, p < 0.001). The calories consumed by the MGF group did not differ significantly from the HFD group ( Figure 3).

Effects of MGF on Glucose Metabolism
To evaluate the effect of MGF on glucose metabolism, fasting blood glucose (FBG) and blood glucose levels at each timepoints were assessed with the oral glucose tolerance test (OGTT). The FBG level was significantly higher in both HFD and MGF groups than in the NC group (HFD, 225 ± 18.77 mg/dL vs. 106 ± 5.65 mg/dL, p < 0.001; MGF, 166 ± 12.24 mg/dL vs. 106 ± 5.65 mg/dL, p < 0.01). The MGF group showed a significantly lower FBG level than the HFD group (166 ± 12.24 mg/dL vs. 225.00 ± 18.77 mg/dL, p < 0.05; Figure 4a). In the case of the OGTT, blood glucose levels reached a peak at 30 min in all groups. The blood glucose levels at 30 and 60 min were significantly lower in the MGF group compared to the HFD group (30 min, 284 ± 14.70 mg/dL vs. 404 ± 12.96 mg/dL, p < 0.001; 60 min, 209 ± 18.87 mg/dL vs. 271 ± 12.14 mg/dL, p < 0.05; Figure 4b). The area under the curve (AUC) for the HFD group was significantly higher than that for the NC or MGF group (Figure 4c). The MGF group showed significantly lower incremental area under the curve (iAUC) than the NC and HFD group (Figure 4d)

Effects of MGF on Glucose Metabolism
To evaluate the effect of MGF on glucose metabolism, fasting blood glucose (FBG) and blood glucose levels at each timepoints were assessed with the oral glucose tolerance test (OGTT). The FBG level was significantly higher in both HFD and MGF groups than in the NC group (HFD, 225 ± 18.77 mg/dL vs. 106 ± 5.65 mg/dL, p < 0.001; MGF, 166 ± 12.24 mg/dL vs. 106 ± 5.65 mg/dL, p < 0.01). The MGF group showed a significantly lower FBG level than the HFD group (166 ± 12.24 mg/dL vs. 225.00 ± 18.77 mg/dL, p < 0.05; Figure 4a). In the case of the OGTT, blood glucose levels reached a peak at 30 min in all groups. The blood glucose levels at 30 and 60 min were significantly lower in the MGF group compared to the HFD group (30 min, 284 ± 14.70 mg/dL vs. 404 ± 12.96 mg/dL, p < 0.001; 60 min, 209 ± 18.87 mg/dL vs. 271 ± 12.14 mg/dL, p < 0.05; Figure 4b). The area under the curve (AUC) for the HFD group was significantly higher than that for the NC or MGF group (Figure 4c). The MGF group showed significantly lower incremental area under the curve (iAUC) than the NC and HFD group (Figure 4d). NC group (24.99 ± 5.43 vs. 8.92 ± 1.99, p < 0.05), but the fasting insulin level did not differ significantly (Figure 4e
In the oral fat tolerance test (OFTT), the triglyceride (TG) level of each group peaked at 120 min and then steadily declined. A significant difference was found in comparing the TG level between HFD and NC groups at every measurement time except 360 min, as the former group was higher than the latter (Figure 5f). The MGF group showed a lower TG level at every measurement time than the HFD group, and significantly lower at 120 min (303.2 ± 28.00 mg/dL vs. 395.0 ± 17.59 mg/dL, p < 0.05) and 240 min (191.4 ± 15.79 mg/dL vs. 287.8 ± 20.11 mg/dL, p < 0.01). The MGF group showed a higher TG level at 120 min than the NC group but without statistical significance (Figure 5f). When three groups were compared for TG AUC, a significantly high outcome was found in the HFD group than in the MGF group (76,440 ± 4193.22 mg•min/dL vs. 101,904 ± 4,454.11 mg•min/dL, p < 0.01; Figure 5g). The fasting insulin level was significantly elevated in the HFD group compared to the NC (3.74 ± 0.59 ng/dL vs. 1.17 ± 0.26 ng/dL, p < 0.01) and MGF (2.04 ± 0.35 ng/dL vs. 3.74 ± 0.59 ng/dL, p < 0.05) groups. The HFD group had significantly higher HOMA-IR compared to the NC (62.76 ± 15.00 vs. 8.92 ± 1.99, p < 0.01) and MGF (24.99 ± 5.43 vs. 62.76 ± 15.00, p < 0.05) groups. HOMA-IR increased significantly in the MGF group than in the NC group (24.99 ± 5.43 vs. 8.92 ± 1.99, p < 0.05), but the fasting insulin level did not differ significantly (Figure 4e,f).
F4/80 gene expression in adipose tissue was significantly lowered in both NC and MGF groups (NC, 0.99 ± 0.03 vs. 10.68 ± 1.18, p < 0.001; MGF 6.59 ± 0.86 vs. 10.68 ± 1.18, p < 0.05) compared to that in the HFD group ( Figure 6f). A significant decrease in TNF-α was found in the MGF group compared to its level in the HFD group (2.01 ± 0.23 vs. 9.20 ± 2.36, p < 0.05; Figure 6g). The IL-6 expression was significantly lowered in the NC group (HFD, 1.00 ± 0.00 vs. 2.03 ± 0.39, p < 0.05; MGF, 1.00 ± 0.00 vs. 1.92 ± 0.31, p < 0.05) than both HFD and MGF groups, but no significant difference was found between the HFD and MGF groups ( Figure 6h). In the oral fat tolerance test (OFTT), the triglyceride (TG) level of each group peaked at 120 min and then steadily declined. A significant difference was found in comparing the TG level between HFD and NC groups at every measurement time except 360 min, as the former group was higher than the latter (Figure 5f). The MGF group showed a lower TG level at every measurement time than the HFD group, and significantly lower at 120 min (303.2 ± 28.00 mg/dL vs. 395.0 ± 17.59 mg/dL, p < 0.05) and 240 min (191.4 ± 15.79 mg/dL vs. 287.8 ± 20.11 mg/dL, p < 0.01). The MGF group showed a higher TG level at 120 min than the NC group but without statistical significance (Figure 5f). When three groups were compared for TG AUC, a significantly high outcome was found in the HFD group than in the MGF group (76,440 ± 4193.22 mg·min/dL vs. 101,904 ± 4,454.11 mg·min/dL, p < 0.01; Figure 5g).

Effects of MGF on Kupffer Cells (KCs) and Gene Expressions
The total percentage of KCs was significantly higher in the HFD group compared to the NC group, while the MGF group showed a significantly decreased KC proportion compared to the HFD group (31.76 ± 1.47 vs. 44.61 ± 2.26, p < 0.001; Figure 7a,b).

Effects of MGF on Kupffer Cells (KCs) and Gene Expressions
The total percentage of KCs was significantly higher in the HFD group compared to the NC group, while the MGF group showed a significantly decreased KC proportion compared to the HFD group (31.76 ± 1.47 vs. 44.61 ± 2.26, p < 0.001; Figure 7a,b).
The NF-κB gene expression level in the liver was significantly higher in the MGF group (2.73 ± 0.37 vs. 5.20 ± 0.58, p < 0.05) than in the HFD group (Figure 7c). A significant difference was found in a comparison of AKT1 gene expression levels only between HFD and NC groups (Figure 7d). When the three groups were compared for ATG7 gene expression level, a significantly low outcome was discovered in the HFD group than in the NC group and MGF group (NC, 0.61 ± 0.10 vs. 1.01 ± 0.03; MGF, 0.61 ± 0.10 vs. 1.00 ± 0.13, p < 0.05; Figure 7e). FGF21 expression was significantly higher in the MGF group than in the HFD group (6.04 ± 0.73 vs. 2.92 ± 0.71, p < 0.05; Figure 7f).

Effects of MGF on Adipocyte and Hepatic Fat Area
When the three groups were compared for the size of adipocytes and fat area in the liver, a significantly high outcome was discovered in HFD group than of the NC group  The NF-κB gene expression level in the liver was significantly higher in the MGF group (2.73 ± 0.37 vs. 5.20 ± 0.58, p < 0.05) than in the HFD group (Figure 7c). A significant difference was found in a comparison of AKT1 gene expression levels only between HFD and NC groups (Figure 7d). When the three groups were compared for ATG7 gene expression level, a significantly low outcome was discovered in the HFD group than in the NC group and MGF group (NC, 0.61 ± 0.10 vs. 1.01 ± 0.03; MGF, 0.61 ± 0.10 vs. 1.00 ± 0.13, p < 0.05; Figure 7e). FGF21 expression was significantly higher in the MGF group than in the HFD group (6.04 ± 0.73 vs. 2.92 ± 0.71, p < 0.05; Figure 7f).

Effects of MGF on Adipocyte and Hepatic Fat Area
When the three groups were compared for the size of adipocytes and fat area in the liver, a significantly high outcome was discovered in HFD group than of the NC group (fat: 16

Discussion
This study aimed to investigate the underlying mechanisms of MGF in insulin resistance at molecular and genetic level focusing on the inflammatory response and autophagy, and new findings were obtained through in silico and in vivo experiments on MGF. First, the relationship between MGF and NF-κB signaling, MAPK signaling pathway, and inflammatory response explored by in silico experiments signified the possibility of improving insulin resistance by regulation of inflammation. Taking the results from the in silico, in vivo experiments specifically studied BW, ATMs, KCs, and mRNA expression of F4/80, TNF-α, IL-6, and NF-κB. Second, in silico experiments investigating the association between MGF and insulin signaling and the PI3K-AKT signaling pathways revealed the possibility of improving insulin resistance through a decrease in fat accumulation. The in vivo experiment demonstrated that MGF decreased the weight and fat areas of the fat pad and liver and increased the mRNA expressions of FGF21 compared to the HFD group. Finally, in silico experiments, it was revealed that MGF affected autophagyrelated mechanisms such as mTOR signaling pathways among the genes related to insulin resistance. Consistent with the in silico results, MGF modulated the mRNA expressions of ATG7 and FGF21 in the in vivo experiments. Our study suggested that the action mechanism of MGF on obesity-associated abnormal glucose and lipid metabolism might be inhibiting inflammation in adipose tissue, enhancing autophagy, and manipulating ATMs and KCs. Therefore, the potential of MGF as a therapeutic agent for metabolic disorders is highlighted through the results.
According to previous studies, the anti-inflammatory and anti-diabetic effects of MGF by inhibiting IRAK1 phosphorylation in NF-κB and MAPK pathways [12], tumor growth and inflammatory responses-related pathways [14], inhibition of JNK activation that can cause liver lipid deposition and insulin resistance [15], and renal fibrosis inhibition via PTEN/PI3K/AKT pathway have been reported [16]. Therefore, this study explored the MGF effect on the possible causes of insulin resistance, including increased fat mass

Discussion
This study aimed to investigate the underlying mechanisms of MGF in insulin resistance at molecular and genetic level focusing on the inflammatory response and autophagy, and new findings were obtained through in silico and in vivo experiments on MGF. First, the relationship between MGF and NF-κB signaling, MAPK signaling pathway, and inflammatory response explored by in silico experiments signified the possibility of improving insulin resistance by regulation of inflammation. Taking the results from the in silico, in vivo experiments specifically studied BW, ATMs, KCs, and mRNA expression of F4/80, TNF-α, IL-6, and NF-κB. Second, in silico experiments investigating the association between MGF and insulin signaling and the PI3K-AKT signaling pathways revealed the possibility of improving insulin resistance through a decrease in fat accumulation. The in vivo experiment demonstrated that MGF decreased the weight and fat areas of the fat pad and liver and increased the mRNA expressions of FGF21 compared to the HFD group. Finally, in silico experiments, it was revealed that MGF affected autophagy-related mechanisms such as mTOR signaling pathways among the genes related to insulin resistance. Consistent with the in silico results, MGF modulated the mRNA expressions of ATG7 and FGF21 in the in vivo experiments. Our study suggested that the action mechanism of MGF on obesityassociated abnormal glucose and lipid metabolism might be inhibiting inflammation in adipose tissue, enhancing autophagy, and manipulating ATMs and KCs. Therefore, the potential of MGF as a therapeutic agent for metabolic disorders is highlighted through the results.
According to previous studies, the anti-inflammatory and anti-diabetic effects of MGF by inhibiting IRAK1 phosphorylation in NF-κB and MAPK pathways [12], tumor growth and inflammatory responses-related pathways [14], inhibition of JNK activation that can cause liver lipid deposition and insulin resistance [15], and renal fibrosis inhibition via PTEN/PI3K/AKT pathway have been reported [16]. Therefore, this study explored the MGF effect on the possible causes of insulin resistance, including increased fat mass and hepatic lipid accumulation, inflammatory responses mediated by macrophages, and dysregulation of autophagy.
In silico experiments revealed hub pathways of overlapped genes between insulin resistance and MGF, and showed that the significant pathways were NF-κB, MAPK, and PI3K-AKT signaling [17]. Because of few previous studies about these signaling pathways, the interactions between MGF and ATG7 and FGF21 were identified by molecular docking simulation. Sequential in vivo experiments described the effects of MGF on weight, glucose and lipid metabolism, inflammatory response in adipose tissues, and autophagy in the liver of HFD-fed mice.
Molecular docking techniques indicated MGF-protein interactions, which were evaluated based on binding affinity. As TNF-α and NF-κB were among the top 10 hub genes, and the functional analysis of MGF was related to inflammation in macrophages and autophagy, several genes were selected for docking simulation. NF-κB, TNF-α, IL-6, and macrophage marker F4/80 are inflammation-related substances [18], and NF-κB is essential for an inflammatory response in the hepatic tissue [19]. The reduced activities of PI3K and AKT, which affects insulin resistance results from increased usage of free fatty acids in tissues, reduced glucose inflow from muscles, and lack of inhibition of hepatic glucose output [5]. Molecular interactions showed a high binding affinity of MGF with TNF-α, IL-6, NF-κB, AKT, ATG7, and FGF21. A high binding relationship between MGF and their target proteins demonstrates therapeutic efficacy [20]. Therefore, these molecular docking results suggest that the mechanisms of MGF in insulin resistance might be related to inflammation and autophagy.
In the current study, significant reduction in weight gain was observed in the MGF group, indicating the effectiveness of MGF in preventing obesity. The epididymal fat pads in mice are a major indicator for evaluating changes in white adipose tissue (WAT) because of the secretion of various adipokines and insulin resistance [21,22]. Liver weight also is an indirect indicator of obesity and inflammatory conditions, as fat accumulation in obesity due to high-fat diet accounts for liver weight [23]. The epididymal fat and liver weights were significantly lower in the MGF group than in the HFD group. The obesity preventive effects were not based on differences in food intake but on significant immune-modulating benefits, including decreased M1 ATM, and inhibition of TNF-α and NF-κB. The WAT increase is closely linked to obesity, and fat accumulation in the liver is affected by the increased NEFA from WAT lipolysis connected through the hepatic portal vein [6]. In obesity, lipids from high lipolysis of enlarged fat are accumulated in KCs before the number of KCs increased. Fat-laden KCs promote TG accumulation in liver and develop the severity of NAFLD by secreting pro-inflammatory cytokines, including TNF-α and IL-1β which suppress PPARα pathway via NF-κB activation [24]. We observed that MGF group showed significantly low fat weight, decreased adipocyte size and less fat accumulation in the liver along with the reduced number of KCs and down-regulation of NF-κB.
When hyperglycemia is maintained, a non-enzymatic reaction occurs between blood glucose and proteins or lipoproteins, resulting in advanced glycation end-products (AGEs). AGEs intensify the pathological aging of blood vessels; thus, controlling blood glucose is essential for people with diabetes and obesity [25,26]. The FBG levels were significantly decreased in the MGF group. The OGTT results showed significantly lower glucose levels in the MGF group than in the HFD group when the levels were measured at 30 and 60 min. In addition, fasting serum insulin levels and HOMA-IR in the MGF group were significantly lower than in the HFD group, indicating the effectiveness of MGF in improving glucose metabolism.
Although HFD leads to high TG levels, the enlarged adipose tissue induces high lipolysis, leading to a high level of circulating NEFA, which comes into the liver, resulting in high hepatic TG levels [27]. In diabetes, the serum lipoprotein concentration also increases due to the disorder of insulin-inhibiting lipoprotein production [28]. Obesity-induced dyslipidemia is characterized by elevated NEFA, TG, and LDL-C levels and reduced HDL-C levels [29]. In this study, the serum lipids of the HFD group consisted of elevated NEFA, TG, TC, LDL-C, and HDL-C levels, same as the typical lipid profile in obesity. However, the MGF administration significantly reduced NEFA and TG levels and elevated HDL-C levels. Moreover, the NEFA and TG levels in the MGF group were not statistically different from those in the NC group. In OFTT, TG levels in the MGF group were significantly lower at 120 and 240 min compared to those in the HFD group. These results demonstrate that MGF can improve obesity-induced dyslipidemia.
Obesity, a low-grade, chronic inflammatory condition, increases ATMs and abnormal cytokine secretion and impairs macrophage metabolism by increasing the M1 phenotype, which acts as an obstacle to insulin signaling [30]. F4/80 is a marker that allows the screening of mononuclear phagocytes and is known to be strongly correlated to body mass index and adipocyte size [6]. When the number of pro-inflammatory M1 ATMs is increased in obesity, TNF-α and IL-6 are highly secreted, which are involved in the insulin resistance of adipocytes by inhibiting insulin signaling and stimulating lipolysis of adipocytes. In the MGF group, total ATM, M1 ATM, F4/80, and TNF-α levels were significantly reduced compared to those in the HFD group.
Obesity-induced insulin resistance results in abnormal elevated circulating insulin, promotion of fatty acid synthesis, and increase in FGF21 ultimately causing a FGF21resistant state. Liver-derived FGF21 signaling in adipose tissue is essential for maintaining insulin sensitivity and lipolysis. Higher liver-derived FGF21 levels can lower BW, blood glucose and liver fat [31], and activate ATG7 expression which can improve the impairment of autophagy and hepatosteatosis [32]. ATG7, a vital autophagic element, contributes to insulin signaling and regulates hepatic insulin sensitivity by ameliorating systemic glucose homeostasis and endoplasmic reticulum stress [8]. Recently, it was suggested that the possibility of increased autophagy activity induced by obesity might act as a compensatory signal to reduce insulin resistance and inflammatory responses in adipose tissue [33]. The mRNA expression of NF-κB, AKT, ATG7, and FGF21 in the liver tissue was analyzed to confirm the effects of MGF on the linkage of autophagy and insulin resistance, and the current results explained that MGF decreased inflammation and regulated autophagy. In the MGF group, the mRNA expression of ATG7 and FGF21 was significantly increased compared to that in the HFD group, and NF-κB was significantly decreased. NF-κB is vital for inflammatory responses and stress signals. However, AKT expression increased with no significance in the MGF group compared to the HFD group. AKT plays a role in multiple pathways by inhibiting glucose production [34]. Reduced activation of PI3K and AKT decrease glucose inflow mediated by GLUT4 in muscles, and cause insulin resistance due to dysregulation of hepatic glucose output in liver [5]. However, unlike prior studies, our data revealed no significant difference in AKT between the MGF and HFD groups [35].
Controversial results between the current analysis and previous studies emphasize the need for additional research.

Search for MGF Target Genes
We searched the SMILES database of MGF from PubChem (pubchem.ncbi.nlm.nih. gov/ (accessed on 18 March 2021)). MGF target genes were collected using SwissTar-getPrediction (swisstargetprediction.ch) based on MGF's SMILES, and data with a zero probability were eliminated.

Search for Overlapping Target Genes between MGF and Insulin Resistance
We collected the target genes of insulin resistance from GeneCards (genecards.org/ (accessed on 18 March 2021)) and used the term "insulin resistance". Relevance scores > 10 from GeneCards were selected. Overlapping genes between insulin resistance and MGF were obtained, and VENNY 2.1 (bioinfogp.cnb.csic.es/tools/venny/ (accessed on 18 March 2021)) was used to represent the overlapping genes in a Venn diagram.

Analysis of Functional Enrichment
To obtain GO and KEGG data, overlapping genes were analyzed using DAVID (david. ncifcrf.gov/ (accessed on 18 March 2021)). Each CC, BP, and MF from GO terms were sorted by gene count, and the top 10 BP, top 5 CC, and MF were selected. Gene count sorted the KEGG pathways, and 17 pathways related to insulin resistance were determined. The role of the obtained data was analyzed using R software ver. 4.2. The results were visualized with dot plots indicating the enrichment terms.

Docking Simulation
The binding site of MGF to target proteins and its binding affinity were investigated to evaluate the interactions between target proteins and MGF. As the ligand, we chose MGF, TNF-α, and NF-κB, which are highly related to vital pathways of the functional enrichment analysis. IL-6, AKT1, and FGF21 were also selected as receptors to evaluate their effects on insulin resistance, and ATG7 was chosen as a receptor to assess its effects on autophagy. We obtained 3D data for MGF (CID 5281647) from PubChem (pubchem.ncbi.nlm.nih.gov/ (accessed on 18 March 2021)). The structures of TNF-α (2AZ5), IL-6 (1ALU), NF-κB (1NFI), AKT1 (1UNP), ATG7 (3T7H), and FGF21 (6M6E) were obtained from the PDB (rcsb.org). Target proteins were pretreated using the Biovia Discovery Studio Visualizer, deleting unnecessary domains and adding polar groups. Molecular docking and binding affinity analysis of MGF to TNF-α, IL-6, NF-κB, AKT1, ATG7, and FGF21 were performed using Pyrx and Autodock VINA. Visualization of the binding structures was performed using the Biovia Discovery Studio Visualizer.

Study Design and Drug Administration
A total of fifteen mice were simultaneously randomized to the three groups without considering any other variable: NC, HFD (control), and MGF 150 mg/kg groups. HFD and MGF groups were administered with HFD for 14 weeks to induce obesity. After 6 weeks of HFD, none of the groups showed a significant difference except for the BW in the NC group. Subsequently, MGF 150 mg/kg/day was orally administered to the MGF group, and normal saline was administered to the HFD and NC groups using a gavage needle for 8 weeks.

Measuring BW, Oral Intake, and Liver and Fat Tissues
We measured BW weekly using an electronic balance (CAS 2.5D; Seoul, Republic of Korea) in the morning at the same time. Each mouse was placed on a plastic bowl during BW recording session, and to minimize the error during the measurement, BW was recorded once the mice was stable. The quantity consumed by each group was computed every morning by the difference between the day before feed and the following day using an electronic balance (CAS 2.5D; Seoul, Republic of Korea). At the end of the study, liver and epididymal fat pads were weighed using an electronic balance.

OGTT and Insulin Resistance Measurement
OGTT was performed at week 11. After fasting for 14 h, the mice were orally fed 2 g/kg of glucose liquefied in distilled water. The blood glucose level was evaluated at 0, 30, 60, 90, 120, and 180 min using a strip-operated blood glucose sensor (AccuCheck Performa, Basel, Switzerland) with blood samples from tail vein.
After 14 weeks, blood was collected from 6-h fasting mouse's tail vein to measure blood glucose levels and insulin concentration. BD Microtainer was used to collect a blood sample at room temperature, followed by 20 min of centrifuging at 2000 G to isolate the serum. An ultrasensitive insulin ELISA kit (Crystal Chem Inc., Elk Grove Village, IL, USA) was used to measure the serum insulin levels. Insulin standards and samples were dispensed 5 µL each into antibody-coated microplates. After 2 h of incubation (at 4 • C), we washed the samples five times and conjugated them with the anti-insulin enzyme reacted with the enzyme substrate solution. After 10 min, the reaction stop solution was added, and the insulin level was measured at 450 nm.

OFTT and Lipid Analysis
At week 12 of the trial, OFTT was performed after 14 h of fasting. After the mice were administered with 2 mL/kg of olive oil, TG levels were assessed by collecting tail vein blood at 0, 120, 240, and 360 min. Accutrend Plus (Roche, Basel, Switzerland) was used to evaluate the TG concentrations. At week 14, we evaluated TC, phospholipids, NEFA, LDL cholesterol, and HDL cholesterol with blood samples from the heart.

Analysis of Gene Expression in Adipose Tissue
The mice were euthanized at week 14, and epididymal fat pads were dissected and wrapped with aluminum foil. Then, samples were stored in liquid nitrogen at −70 • C until RNA extraction. Mini RNA Isolation IITM (ZYMO RESEARCH, Irvine, CA, USA) was used to isolate RNA from adipose tissue. Defrosted adipose tissue was transferred to tubes with 300 µL aliquots of the ZR RNA buffer and pulverized by a homogenizer. After centrifugation at 1000 rpm, the supernatant was moved into a column, which was mixed with 350 µL of the RNA wash buffer. Fifty microliters of RNA-free water were added for centrifugation, and the extracted RNA was stored at −70 • C.

Histological Analysis of the Epididymal Fat Pad and Liver
We fixed the fat and liver samples in 10% neutral-buffered formalin and immersed them in 70, 80, 90, and 100% ethanol. The samples were then embedded in paraffin, sliced into 4-µm sections using a microtome, and arranged on gelatin-coated slides. To stain tissues, each slide was dewaxed in xylene, and then 70, 80, 95, and 100% ethanol and distilled water were used to rehydrate the tissues. Hematoxylin and eosin were used to stain the rehydrated tissue, and the histological image of the stained sample was visualized under a high-resolution camera-mounted optical microscope (Olympus BX-50, Olympus Optical, Tokyo, Japan). The fat areas of epididymal fat tissue and liver tissue were measured using ImageJ software.

Isolation of Stromal Vascular Cells (SVCs) and Liver Immune Cells
The dissected epididymal fat pads were soaked in PBS/2% BSA solution, cut out into 1-2-mm size, and mixed with collagenase (Sigma, St. Louis, MO, USA) and DNase I (Roche, Basel, Switzerland). After shaking and removing undigested tissue, PBS and 2% FBS were added to pellet, and the sample was filtered by a 100-µm filter to obtain the SVCs.
The liver sample without the gall bladder was prepared soaking in the RPMI 1640 medium with 100 mL/L fetal calf serum and squashed through a 200-G mesh filter. Subsequently, Percoll, PBS, and heparin were added before centrifugation. The supernatant was incubated in 1X ACK lysis buffer (Lonza) to dissolve red blood cells, and unnecessary tissue was discarded after centrifugation.

Statistical Analysis
All statistical analyses were performed using GraphPad PRISM 5 (GraphPad Software Inc., San Diego, CA, USA). Between-group differences were evaluated with the one-way analysis of variance and Tukey's post hoc test. All data are expressed as mean ± standard error. Two-tailed p-values < 0.05 were set to indicate statistical significance. Statistical significance compared with the NC and HFD groups is presented with number signs (#) and asterisks (*), respectively (# or * for p < 0.05; ## or ** for p < 0.01; and ### or *** for p < 0.001).

Conclusions
MGF improves metabolic phenotypes, including BW, decreased fat mass and hepatic lipid accumulation, and improved glucose and lipid metabolism. Moreover, MGF modulated macrophages in WAT and liver, and gene expressions related to inflammation and autophagy. Therefore, the study suggests that MGF improves glucose and lipid metabolism by inhibiting inflammation and enhancing autophagy in adipose tissue and the liver. These findings indicate that MGF is potentially applicable in metabolic syndrome, including obesity and type 2 diabetes.