Emodin Improves Glucose and Lipid Metabolism Disorders in Obese Mice via Activating Brown Adipose Tissue and Inducing Browning of White Adipose Tissue

Background Adipose tissue (e.g. white, brown and brite) plays a critical role in modulating energy metabolism. Activating brown adipose tissue (BAT) and inducing browning in white adipose tissue (WAT) has been proposed to be a potential molecular target for obesity treatment. Emodin is a natural anthraquinone derivative that exhibits variety of pharmacologic effects including lowering lipids and regulating glucose utilization. However, the underlying mechanism of action is still unclear. In the present study, we investigated whether emodin could alleviate obesity via promoting browning process in adipose tissue. Methods C57BL/6J mice were fed with high fat diet to induce obesity. Emodin at the doses of 40 and 80 mg/kg were orally given to obesity mice for consecutive 6 weeks. Parameters including fasting blood glucose, oral glucose tolerance, blood lipids, and the ratios of subcutaneous white adipose tissue (scWAT) or BAT mass to body weight, and morphology of adipose tissue were observed. Besides, the protein expression of uncoupling protein 1 (UCP1) and prohibitin in BAT and scWAT was determined by immunohistochemistry method. Relative mRNA expression of Cd137, transmembrane protein 26 (Tmem26) and Tbx1 in scWAT was analyzed using qRT-PCR. And the protein expression of UCP1, CD36, fatty acid transporter 4 (FATP4), peroxisome proliferator-activated receptor alpha (PPARα) and prohibitin of scWAT and BAT were analyzed using western blotting. In addition, ultra-high-performance liquid chromatography with electrospray ionization tandem mass spectrometry was utilized to detect the small lipid metabolites of scWAT and BAT. Results Emodin decreased the body weight and food intake in HFD-induced obesity mice, and it also improved the glucose tolerance and reduced the blood lipids. Emodin treatment induced beiging of WAT, and more multilocular lipid droplets were found in scWAT. Also, emodin significantly increased markers of beige adipocytes, e.g. Cd137, Tmem26 and Tbx1 mRNA in scWAT, and UCP1, CD36, FATP4, PPARα and prohibitin protein expression in scWAT and BAT. Furthermore, emodin perturbed the lipidomic profiles in scWAT and BAT of obese mice. Emodin increased total ceramides (Cers), lysophosphatidylcholines (LPCs), lyso-phosphatidylcholines oxygen (LPCs-O), and phosphatidylethanolamines oxygen (PEs-O) species concentration in scWAT. Specifically, emodin significantly up-regulated levels of Cer (34:1), LPC (18:2), LPC-(O-20:2), PC (O-40:7), PE (O-36:3), PE (O-38:6), PE (O-40:6), and sphingolipid (41:0) [SM (41:0)], and down-regulated PC (O-38:0), PE (O-40:4), PE (O-40:5) in scWAT of obesity mice. In terms of lipid matabolites of BAT, the emodin remarkably increased the total PCs levels, which was driven by significant increase of PC (30:0), PC (32:1), PC (32:2), PC (33:4) and PC (38:0) species. In addition, it also increased species of LPCs, e.g. LPC (20:0), LPC (20:1), LPC (22:0), LPC (22:1), LPC (24:0), and LPC (24:1). Especially, emodin treatment could reverse the ratio of PC/PE in HFD-induced obese mice. Conclusions These results indicated that emodin could ameliorate adiposity and improve metabolic disorders in obese mice. Also, emodin could promote browning in scWAT and activate the BAT activities. In addition, emodin treatment-induced changes to the scWAT and BAT lipidome were highly specific to certain molecular lipid species, indicating that changes in tissue lipid content reflects selective remodeling in scWAT and BAT of both glycerophospholipids and sphingolipids in response to emodin treatment.

Conclusions: These results indicated that emodin could ameliorate adiposity and improve metabolic disorders in obese mice. Also, emodin could promote browning in scWAT and activate the BAT activities. In addition, emodin treatment-induced changes to the scWAT and BAT lipidome were highly specific to certain molecular lipid species, indicating that changes in tissue lipid content reflects selective remodeling in scWAT and BAT of both glycerophospholipids and sphingolipids in response to emodin treatment.
Keywords: emodin, obesity, brown adipose tissue, white adipose tissue, browning of white adipose tissue, lipid metabolic signature BACKGROUND White adipose tissue (WAT) and brown adipose tissue (BAT) play a critical role in modulating energy metabolism (1). The adipocytes within WAT store large amounts of triglycerides as chemical energy in unilocular droplets, which are released into circulation as needed. WAT also functions to produce hormones and cytokines, regulates immune system and supports local tissue frame (2). Increases in WAT mass are directly associated with increased rates of metabolic diseases such as obesity and type 2 diabetes (3). BAT is specialized for energy expenditure, which is characterized by small multi-atrial lipid droplets, abundant mitochondria and expresses uncoupling protein 1 (UCP1) (4,5). BAT dissipates energy as heat and stores energy for use of non-shivering thermogenesis, which also plays a significant role in energy regulation (5,6). It has been confirmed that when the body is stimulated by cold exposure (7) or activated by b-adrenoceptors agonist (8), brown-like phenotypic adipocytes (e.g. beige adipocytes) can be detected in WAT, which are characterized by an increased number of mitochondria and increased expression of brown fat marker genes (e.g. Ucp1, Pgc-1a, Prdm16). The above process is called browning of WAT. Recent studies showed that increasing metabolic activity of brown and beige adipose tissue might be a novel way to ameliorate glucose and lipid metabolism in obese patients (9)(10)(11)(12). Furthermore, changes in tissue lipid content reflects selective remodeling in scWAT and BAT of both phospholipids and glycerol lipids in response to specific conditions such as exercise training (1).
Emodin (1,3,8-trihydroxy-6-methylanthraquinone) is a natural anthraquinone derivative, and is the main component of Rheum palmatum L (13). It has been reported to exhibit anti-inflammatory, anti-bacterial, anti-cancer, anti-diabetic, anti-ulcerogenic, immunosuppressive, pro-apoptotic and chemopreventive activities (14)(15)(16)(17)(18). It has been found that the emodin can regulate glucose utilization and lower lipids in epididymal WAT by activating AMP activated protein kinase (AMPK) pathway (19,20). In addition, the emodin could inhibit adipocyte differentiation and enhances osteoblast differentiation from bone marrow mesenchymal stem cells (BMSCs) (21). And it could improve the inactive glucocorticoid-induced adipose tissue dysfunction by selective inhibition on 11b-hydroxysteroid dehydrogenase type 1 (11b-HSD1) in 3T3-L1 adipocyte (22). Our previous research also showed that emodin could inhibit the accumulation of white adipocytes and inducing browning of WAT in apolipoprotein E knockout (ApoE -/-) mice (23). However, the effect of emodin on the lipid metabolic signature of scWAT and BAT has not been investigated. Here, we report the effect of emodin on adipose tissue in high fat diet-induced obese mice, as well as a comprehensive analysis of lipid composition in scWAT and BAT.
Mice were randomly divided into two groups according to weight and fed with normal control diet (n=8) or fed with high fat diet (HFD) for 8 weeks to induce hyperlipidemia. After 8 weeks of HFD, mice were randomly divided into four groups as follows (n=8/group): HFD group, emodin 40 mg/kg group, and emodin 80 mg/kg group, and CL 316243 1 mg/kg group. Mice in normal control group and HFD group were administrated with equal amount of 0.1% carboxymethyl cellulose-Na (CMC-Na). Mice in emodin 40 and 80 mg/kg groups were taken emodin (dissolved in 0.1% CMC-Na) by intragastric administration for consecutive 6 weeks. Before dissection, mice in CL 316243 1 mg/ kg treatment group were intraperitoneally injected with 1mg/kg/ day of CL 316243 disodium salt for 3 days. All the animal studies were in accordance with ethics standards of the Animal Care and Welfare Committee of Beijing University of Chinese Medicine (Certificate No. BUCM-04-2018070603-3015).

Oral Glucose Tolerance Test
After 6 weeks of intervention, the mice were fed by oral gavage with 50% D-glucose (2.0 g/kg) after overnight (12 h) fasting. Blood samples were taken from the tail 0, 30, 60, 90 and 120 min after oral gavage, and glucose levels were measured by the One Touch Ultra blood glucose monitoring system (ONETOUCH Ultra Easy).

Measurement of Lee's Index
At the end of the treatment, the body mass of the mice was accurately weighed, and the body length (the distance from the tip of the nose to the anus) was accurately measured, and then the Lee's index was calculated according to the reference (24).

Measurement of Adipose Tissue Mass/ Body Weights
Subcutaneous WAT (scWAT) mass and scapular brown adipose tissues (BAT) mass were accurately weighed. The ratios of scWAT mass to body weight (BW) and BAT mass to body weight (BW) were calculated.

Serum Biochemical Analysis
Serum TC, TG, HDL-c and LDL-c levels were measured with the method of biochemical kits (Nanjing Jiancheng, China). The levels of FFA were determined by the Mouse FFA ELISA kit (Kete, China). Serum leptin and adiponectin levels were analyzed using the mouse leptin and adiponectin ELISA kit respectively (Cloud-clone, China).

Histological and Immunohistochemical Analysis
BAT and scWAT were fixed with 10% formalin, dehydrated, embedded in paraffin and sectioned. For histological analysis, sections were deparaffinized and stained with hematoxylin and eosin. The expression of UCP1 (1:500, ab10983, Abcam) and prohibitin (1:500, ab75766, Abcam) in mouse BAT and scWAT was determined by immunohistochemistry. All images were acquired with the microscope (Leica, Germany). The expression level of UCP1 and prohibitin in BAT and scWAT of the mice was quantified by using Image Pro Plus 6.0.

Quantitative Real-Time PCR Analysis
Total RNA of scWAT was extracted with Trizol ® Reagent (Ambion, USA). Reverse transcription of total RNA (1 mg) was performed with Revert Aid First Stand cDNA Synthesis Kit (Thermo Scientific, USA). Real-time quantitative PCR (qRT-PCR) was performed with a SYBR Green Master Mix (Novoprotein, China). The PCR reaction was operated in triplicate for each sample using the Step One Real-Time PCR System (Applied Biosystems, USA). After standardizing the expression level of internal control actin in each sample, the data were expressed in arbitrary units. The sequences of primer in this study were shown in Table 1.

Western Blot Analysis
The homogenates of scWAT and BAT were dissolved in RIPA lysate and protease inhibition for protein extraction. Sample protein concentrations were measured using the bicinchoninic acid (BCA) method (Beyotime). Total protein (10mg/Lane) was separated on a 12% acrylamide/acrylamide gel using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) TABLE 1 | Primer sequences were used for quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR).

Gene name
Forward

Quality Control
Five QC samples of adipose tissue were continuously injected at the beginning of the sequence to monitor the UPLC-MS system stability by the Overlay Graphs method using Mass Lynx software. And QC samples were run at regular intervals (8 samples) throughout the entire sequence.

Mass Spectrometry
Electrospray ionization tandem mass spectrometry (XEVO TQ-S Micro, Waters, USA) was used for mass spectrometry. And the conduct conditions of ESI + and ESIshowed in Table 4.
Masslyxn4.1 was used for mass spectrometry data acquisition.

Statistical Analysis
The data of target metabolism group were operated by skyline 19.1. The parameters were set as follows: the quality extraction error was 5 ppm and the allowable retention time error was 15s.
Other data were statistically analyzed using SAS 8.2 software. All data were expressed as means ± SE. Two-way analysis of variance for repeated measures was used for body weight analysis (for the effects of treatment and time). Other statistics was performed using the one-way analysis of variance (ANOVA) followed by SNK-q test. P-value<0.05 was considered as statistically significant.

Emodin Can Inhibit Obesity and Appetite and Reduce Fat Mass in HFD Induced Obese Mice
The HFD treated mice showed characteristics of obesity. When compared with the control mice, parameters including body weight, food intake, Lee's index and scWAT/BW ratio were significantly increased in HFD treated mice (by 51.6%, 41.1%, 8.5% and 39.1%, respectively). Emodin at the dose of 40 mg/kg caused a significant reduction in body weight at week 5 and 6 (by 13.0% and 15.7%, respectively), and emodin 80 mg/kg caused a remarkable reduction in body weight at week 3, 4, 5, 6 (by 11.1%,   12.6%, 12.6% and 13.9%, respectively), when compared with HFD mice ( Figure 1A). Emodin (40 mg/kg, 80 mg/kg) significantly decreased the food intake by 9.8% and 7.3%, respectively, when compared with the obese mice (P <0.01 or P <0.05) ( Figure 1B).
Lee's index can be used as an indicator to evaluate the degree of obesity in adult obese model mice (25). Emodin at the doses of 40 and 80 mg/kg and CL316243 (1 mg/kg) treatment group could significantly reduce the Lee's index, when compared with HFD mice (P <0.01) ( Figure 1C).
The function of BAT is consuming glucose and lipids, mediating the thermogenic effects of non-shivering, thereby increasing energy expenditure (26). Interestingly, compared with HFD mice, emodin (80 mg/kg) and CL 316243 (1 mg/kg) treatment significantly decreased the ratio of BAT/BW (by 22.5% and 47.5%, respectively ( Figure 1E).

Emodin Ameliorates Abnormal Blood Glucose and Blood Lipid in Mice Fed With HFD
In this study, we investigated whether emodin improved glucose tolerance in obese mice. The results indicated an impaired glucose tolerance in HFD mice, and high fat diet significantly increased AUC index, when compared with control mice (P <0.01). Emodin at the doses of 40 and 80 mg/kg and CL316243 (1 mg/kg) treatment significantly decreased AUC value (31.1%, 35.3% and 45.1% respectively), when compared with HFD mice (P <0.01) (Figures 2A, B). These results suggested that emodin could ameliorate glucose metabolism in obese mice.
As an endogenous insulin sensitizer secreted by adipose tissue, reduction of adiponectin is an independent risk factor for hyperlipidemia and diabetes (28). When compared with control mice, serum adiponectin in HFD mice was significantly decreased (P <0.01). Emodin (40 mg/kg, 80 mg/kg) treatment could significantly increase the serum adiponectin levels in HFD mice (by 28.6% and 42.9%, respectively), when compared with HFD mice (P <0.01 or P <0.05) ( Figure 2I).

Emodin Induces Browning of scWAT in Mice Fed With HFD
We analyzed the morphology of scWAT and the expression of thermogenic protein UCP1 and mitochondrial membrane protein prohibitin in scWAT ( Figure 3A). When compared with the control mice, the diameter of fat cells in HFD mice increased and the number of cells per unit area decreased. When compared with HFD mice, the adipocytes of the mice in the emodin (40 mg/kg, 80 mg/kg) groups are small and tightly arranged, with obvious nuclei. As an important thermogenic protein, UCP1 is specifically expressed in BAT (6,29). Prohibitin, mainly located in the inner membrane of mitochondria, plays an important role in maintaining mitochondrial morphology, function and regulating energy metabolism (30). Therefore, we measured the expression of UCP1 and prohibitin protein in scWAT ( Figure 3B). The expression of UCP1 and prohibitin protein in scWAT of emodin 80 mg/kg-treated group was significantly increased (P <0.01), Compared with HFD mice. We also evaluated the mRNA expression of beige adipocyte marker genes, such as Cd137, Transmembrane protein 26 (Tmem26) and Tbx1. As expected, the expression of several beige adipocyte marker genes, including Cd137, Tmem26 and Tbx1, was significantly upregulated in scWAT after emodin (80 mg/kg) and CL316243 (1 mg/kg) treatment ( Figure 3C). In order to confirm whether emodin can induce the browning of scWAT, we measured thermogenic protein and fatty acid transporter ( Figure 3D). The expression of UCP1, prohibitin, CD36, FATP4 and PPARa protein in scWAT of emodin (80 mg/ kg) and CL316243 (1 mg/kg) treatment group increased significantly, when compared with HFD mice (P <0.05 or P <0.01) (Figures 3E-I). These results suggested that emodin could induce the browning of scWAT in HFD mice.

Emodin Activates Brown Adipose Tissue in Mice Fed With HFD
We also analyzed the morphology of BAT and the expression of thermogenic protein UCP1 and mitochondrial membrane  protein prohibitin in BAT (Figures 4A-C). When compared with control mice, the diameter of fat cells in BAT of HFD mice was significantly increased, the number of cells per unit area decreased, and the number of white fat cells increased. This indicates that long-term HFD feeding results in so-called 'whitening' of BAT. When compared with HFD mice, the adipocytes of the mice in the emodin (40 mg/kg, 80 mg/kg) groups are small and tightly arranged, with large and obvious nuclei. We also measured the expression of UCP1 and prolibitin protein in BAT. The expression of UCP1 and prohibitin protein in BAT of emodin and CL316243-treated group was significantly increased (P <0.05 or P <0.01), when compared with HFD mice.
To confirm that emodin can activate BAT of HFD mice, we measured thermogenic protein and fatty acid transporter ( Figure  4D). The expression of UCP1, prohibitin, CD36, FATP4 and PPARa protein in BAT of emodin and CL316243-treated group increased significantly, when compared with HFD mice (P <0.05 or P <0.01) (Figures 4E-I). These results suggested that emodin could activate BAT in HFD mice.

Phospholipid Metabolism Is Altered in scWAT With Emodin Treatment
Based on the above experimental results, we analyzed scWAT by targeted metabolomics. We selected the biomarkers that have changed and created a heat map (Figures 5A-E). The lipidomics data show that emodin treatment can perturb the lipidomics profile in HFD mice, and several phospholipid species

Phospholipid Metabolism Is Altered in BAT With Emodin Treatment
In addition, we also analyzed BAT by targeted metabolomics. We selected the biomarkers that have changed and created a heat map (Figures 6A-E). In BAT, the lipidomics indicated a significant reduction of PEs and PCs in BAT of HFD mice. And emodin treatment caused a significant increase of PEs and PCs in BAT, when compared with HFD mice. The increase in total PC was driven by significant increase of PC (30:0), PC

DISCUSSION
Emodin ameliorates adiposity and improves whole-body metabolic balance in obese mice. In this study, we demonstrated that emodin could decrease the body weight and food intake in high fat diet induced obese mice, also it could improve the glucose tolerance and reduce the blood lipids, which was consistent with the results of previous studies in vivo (22,31,32). The regulating of white adipose tissue and brown adipose tissue acts a critical role in combating obesity and metabolic disease. As a thermogenic tissue, BAT is innervated by both sympathetic and sensory nerves. The activity and metabolism of BAT could be influenced by cold exposure and exercise (33), as well as some natural product extracts or compounds, such as rose hip supplement (34), black raspberry (35), green tea extract (36), genistein (37), pentamethylquercetin (38), and etc. Here, we first proved that emodin could promote browning in scWAT. The white adipose tissue mass was decreased after emodin treatment. After treatment of emodin, more multilocular lipid droplets were found in scWAT, the mRNA expression of browning markers including Cd137, Tmem26 and Tbx1 was increased in scWAT, and the protein expression of browning markers including UCP1 and prohibitin was increased in scWAT of obese mice. Meanwhile, the protein expression of UCP1, prohibitin, PPARa was increased in BAT of obese mice after emodin treatment. PPAR a is the key factor of BAT thermogenesis, which can regulate lipid catabolism and thermogenic gene expression in coordination with Pgc-1a and Prdm16 (39). PPAR a can enhance the expression of PGC-1a and UCP-1 by increasing the activity of erythropoietin (EPO). PPAR a also plays a coordinating role with SIRT1 activated by EPO and jointly regulates the level of NAD + to heighten the metabolic activity (40). The mitochondria are involved in the metabolic control of brown adipocytes. Mitochondrial function is related to the endocrine function of adipocytes. In addition, brown adipocytes rely on mitochondria to maintain intracellular metabolism. Located in mitochondrial inner membrane, prohibitin plays a critical role in maintaining the shape and function of mitochondria and regulating energy metabolism (30,41). The results of western blotting demonstrated that emodin increased the protein expression of PPAR a and prohibtin of scWAT and BAT in obese mice. As a fatty acid translocase, CD36 acts a pivotal part in the uptake and transport of long-chain fatty acids (LCFA) in heart and adipose tissues (42,43). It was found that cold exposure drastically accelerated plasma clearance of triglycerides as a result of increased uptake into BAT, a process crucially dependent on local LPL activity and transmembrane receptor CD36 (44). Fatty acid transporter 4 (FATP 4) is a member of the fatty acid transport proteins (FATPs), which plays a significant role in the transport of longchain fatty acids with more effectively compared with FATP1. It was found that FAT/CD36 and FATP4 were the most effective fatty acid transporters (45,46). In this study, emodin accelerated the transport and consumption of fatty acids and improved the disorder of lipid metabolism by increasing the expression of CD36 and FATP4 protein in both scWAT and BAT of HFDinduced mice. White and brown adipocytes exhibit different lipid metabolic signature, which reflect their distinct organelle composition and cell functions. The neutral lipids in the lipid droplets core are surrounded by a monolayer of phospholipids (47). PCs, PEs and CLs make up 89% of the phospholipids in BAT, which are increased in response to cold exposure and exercise (48). It was confirmed that regulation and metabolism of PCs, PSs and PEs prevented inflammation of adipose tissue, hyperlipidemia and obesity (49). Exercise can increase specific molecular species of PCs and PEs in brown adipocytes. It has been reported that after exercise, the increase in total PC was driven by the significant increase of the highly abundant PC (36:2) species, as well as increases in numerous species of PC and PC-O. However, there was no overall change in abundance of PE after exercise in BAT (1). Our recent study manifested that the lipidomic profile of adipocytes was remolded with high fat diets, and emodin treatment could perturb the profile and reverse some small lipid metabolites of HFD mice.
Furthermore, the relative abundance of PCs and PEs on the surface of LDs is important for their dynamics (50). An increase in the relative amount of PEs on the surface of lipid droplets can promote fusion of smaller droplets into larger ones (51). Inhibition of PCs biosynthesis can promote TG storage increases the size of the lipid droplets presumably (52,53). Either abnormally high, or abnormally low cellular PC : PE molar ratios can influence energy metabolism in various organelles (50). It has been shown that both PC amount and PC : PE molar ratio tend to increase, and palmitate-and stearatecontaining LPC species were upregulated in 16-week-old Lep ob/ob adipose tissue macrophage, which related to WAT inflammation and contribute to the development of insulin resistance in obesity (54). It has been proved the different composition of Only metabolites with VIP >1 and P <0.05 were selected in heat map, and different shades of color present the concentration (red, white and blue presented the high, normal and low concentration). The red up arrow indicates up regulation or promotion, and the blue down arrow indicates down regulation or inhibition. HFD, high fat diet; Emo, emodin; PC, phosphatidylcholine; PE, phosphatidylethanolamine; Cer, ceramides; LPC, lyso-phosphatidylcholine. Data are expressed as mean ± SE. *P < 0.05, **P < 0.01 vs. HFD group.
phospholipids in white and brown adipocyte, and thermogenic adipocytes possess higher abundance of PCs and PEs, with longer (C >36) and more polyunsaturated species (55). Our results also indicated that the ratio of PC : PE was significantly increased in brown adipose tissue but not scWAT in HFD-induced mice compared with control mice. Interestingly, emodin 40 and 80 mg/kg treatment and CL 316243 could significantly increase abundance of PCs and PEs, and decrease the PC : PE ratio in BAT of obesity mice. Specifically, emodin significantly up-

CONCLUSION
These results indicated that emodin could ameliorate adiposity and improve metabolic disorders in obese mice. Also, emodin could promote browning in scWAT and activate the BAT activities. In addition, emodin treatment-induced changes to the scWAT and BAT lipidome were highly specific to certain molecular lipid species, indicating that changes in tissue lipid content reflects selective remodeling in scWAT and BAT of both glycerophospholipids and sphingolipids in response to emodin treatment.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.