MD001, a Novel Peroxisome Proliferator-activated Receptor α/γ Agonist, Improves Glucose and Lipid Metabolism

Peroxisome proliferator-activated receptor (PPAR)-α/γ dual agonists have been developed to treat metabolic diseases; however, most of them exhibit side effects such as body weight gain and oedema. Therefore, we developed a novel PPARα/γ dual agonist that modulates glucose and lipid metabolism without adverse effects. We synthesised novel compounds composed of coumarine and chalcone, determined their crystal structures, and then examined their binding affinity toward PPARα/γ. We investigated the expression of PPARα and PPARγ target genes by chemicals in HepG2, differentiated 3T3-L1, and C2C12 cells. We examined the effect of chemicals on glucose and lipid metabolism in db/db mice. Only MD001 functions as a PPARα/γ dual agonist in vitro. MD001 increased the transcriptional activity of PPARα and PPARγ, resulting in enhanced expression of genes related to β-oxidation and fatty acid and glucose uptake. MD001 significantly improved blood metabolic parameters, including triglycerides, free fatty acids, and glucose, in db/db mice. In addition, MD001 ameliorated hepatic steatosis by stimulating β-oxidation in vitro and in vivo. Our results demonstrated the beneficial effects of the novel compound MD001 on glucose and lipid metabolism as a PPARα/γ dual agonist. Consequently, MD001 may show potential as a novel drug candidate for the treatment of metabolic disorders.


Results
Chemical synthesis. Our synthetic procedure is summarised in Fig. 1. The synthesis commenced with a well-known Pechmann condensation of commercially available (1) 24 and ethyl benzoyl acetate to afford a roughly 1:1 mixture of 2a and 2b owing to the presence of two isomeric ortho and para hydroxyl groups to the methyl group. Compounds 2a and 2b proved to be labile during SiO 2 column chromatography and insoluble in various solvents probably due to the 1,3-dihydroxy group. Thus, the mixture of 2a and 2b was methylated in the presence of K 2 CO 3 and Me 2 SO 4 25,26 to give rise to 3a and 3b, pleasingly, which were readily dissociable and stable during the separation process. The structure of 3b was confirmed by the X-ray crystallography ( Supplementary Fig. S1a). Having a secured route to the two isomers 3a and 3b, we next turned our attention to the fundamental Friedel-Crafts acylation. Given that acetyl groups are easily condensed with the benzaldehyde to afford the cinnamoyl groups under basic conditions, we initially attempted Friedel-Crafts acetylation. However, despite our efforts to force the reaction with various Lewis acids such as BF 3 •OEt 2 , SnCl 4 , and TiCl 4 , all failed to yield any of the desired products. We then attempted to directly introduce a cinnamoyl group. Friedel-Crafts cinnamoylation of 3a in the presence of BF 3 •OEt 2 as a Lewis acid did not yield 4a. However, to our delight, the reaction of 3a with SnCl 4 gave rise to the desired product 4a even though 50% of the starting material 3a remained. Finally, the treatment of 3a with TiCl 4 in refluxing CH 2 Cl 2 furnished 4a at a good yield (68%) (Supplementary Table S2). The reaction of 3b with TiCl 4 in refluxing CH 2 Cl 2 did not provide 4b, but in refluxing dichloroethane gave rise to 4b at a moderate yield (38%, 70% conversion of 3a). The structure of 4a was verified from by X-ray crystallography ( Supplementary Fig. S1b). Finally, global deprotection by the action of BBr 3 in refluxing dichloroethane provided 5a (41%) and 5b (37%) respectively [27][28][29] .
Scientific RepoRts | (2019) 9:1656 | https://doi.org/10.1038/s41598-018-38281-0 all of which are target genes of PPARγ. The opposite effect was observed upon siRNA-mediated knockdown of PPARγ (Fig. 2E, Supplementary Fig. S3B). These results suggest that MD001 may act as a dual agonist and regulate metabolism via specific activation of PPARα and PPARγ. Studies have proven PPARγ to be a traditional molecular target for the development of anti-diabetic drugs that improve insulin sensitivity and glucose tolerance. Therefore, we compared the ability of MD001 to enhance glucose metabolism with that of rosiglitazone in HepG2, differentiated 3T3-L1, and C2C12 myotubes. MD001 significantly increased glucose consumption in a dose-dependent manner ( Fig. 3A-C). In addition, quantitative RT-PCR analysis showed that MD001 significantly increased the expression of glucose transporter GLUT2 (HepG2) and GLUT4 (3T3-L1 and C2C12), suggestive of its stimulatory effect on glucose metabolism at least in part through increased expression of glucose transporter (Fig. 3D-F).
To further examine whether MD001 could increase β-oxidation via PPARα activation, we analysed the expression of PPARα target genes related to β-oxidation in HepG2, differentiated 3T3-L1, and C2C12 myotubes. MD001 significantly increased the expression levels of ACOX, CPT, malonyl-CoA decarboxylase (MLYCD), and fatty acid transporter (FATP) in HepG2, and ACOX and CPT in both 3T3-L1 and C2C12 cells (Fig. 4A-C). Therefore, we evaluated whether MD001 could stimulate fatty acid oxidation and found that MD001 significantly increased the β-oxidation rate in HepG2 cells (Fig. 4D); similar results were observed in differentiated 3T3-L1 and C2C12 myotubes (Fig. 4E,F). We examined whether the increased expression of PPARα target genes and enhanced β-oxidation is dependent on PPARα by suppressing PPARα expression using PPARα-specific siRNA. The MD001-mediated increase in β-oxidation was abrogated following downregulation of PPARα expression, thereby confirming the stimulatory effect of MD001 on β-oxidation via PPARα activation (

MD001 improves metabolic profiles in db/db mice.
To investigate its effects in vivo, MD001 was administered once a day to wild type C57BL/6J and diabetic db/db mice. MD001 significantly decreased blood glucose levels in a dose-dependent manner in db/db mice (Fig. 5A). An oral glucose tolerance test (OGTT) revealed that MD001 significantly decreased blood glucose levels in db/db mice (Fig. 5B, Supplementary  Fig. S5A). In addition, an intraperitoneal insulin tolerance test (IPITT) showed that MD001 lowered blood glucose levels by increasing insulin sensitivity ( Supplementary Fig. S5B,C). In the livers of specimens in the diabetic animal model, the expression of genes related to gluconeogenesis is upregulated, contributing to hyperglycaemia 30 . Given that MD001 was shown to recover insulin sensitivity and improve blood glucose levels in db/db mice as shown above, the effect of MD001 on gene expression related to gluconeogenesis was examined. As shown in Supplementary Fig. S5D, MD001 decreased the expression of phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase (G6Pase) in the livers of db/db mice. TZDs, including rosiglitazone, pioglitazone, and troglitazone, are known to induce severe body weight gain as an adverse effect both in animals and humans 9,12,31,32 . Therefore, we evaluated the effect of MD001 on body weight. As shown in Fig. 5C, rosiglitazone significantly induced body weight gain in db/db mice compared to the vehicle control; on the contrary, MD001 and WY14643 showed no induction of body weight gain in both wild type C57BL/6J and db/db mice without changes in food ingestion ( Supplementary Fig. S6). As the agonist-mediated activation of PPARα or PPARγ has been known to lower lipid levels as well as blood glucose in diabetic patients 7,8 , we examined whether MD001 could lower lipid levels. Rosiglitazone and WY14643, used as positive controls, significantly decreased blood TG, FFA, and glucose levels in db/db mice (Fig. 5D,E,K), consistent with a previous study 9 . As a PPARα/γ dual agonist, MD001 also significantly decreased blood TG, FFA, insulin, and glucose levels in db/db mice (Fig. 5A,D,E,K,L). However, in wild type C57BL/6J mice, MD001 did not have any effect on the levels of plasma lipids and blood glucose when compared to the vehicle group ( Supplementary Fig. S7). In addition, although MD001 did not affect the total cholesterol levels ( Fig. 5F), it significantly decreased LDL and increased HDL levels (Fig. 5G,H), indicating that MD001 may improve cholesterol metabolism in diabetic animal models. While rosiglitazone significantly increased blood alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels, MD001 significantly reduced blood ALT and AST levels in db/db mice; no effect was observed on blood ALT and AST levels in wild type C57BL/6 mice, suggestive of the absence of toxic effects of MD001 on the liver, unlike pure PPARγ agonists (Fig. 5I,J, Supplementary Fig. S7F,G) 33,34 . In addition, MD001 increased the serum adiponectin (Acrp30) levels (Fig. 5M). Furthermore, rosiglitazone treatment resulted in significant increases in liver weight (40%) and fat mass (50%) as compared to the vehicle control in db/db mice, whereas MD001 treatment showed no increase in liver weight or fat mass ( Supplementary Fig. S8). These results strongly suggest that MD001 may efficiently decrease blood glucose and lipid levels without significant changes in body weight and hepatotoxicity. MD001 improves metabolic disorders in db/db mice. Fatty liver is a complication associated with insulin resistance, obesity, and type 2 diabetes. While PPARα activation has been known to alleviate fatty liver by stimulating β-oxidation and reducing lipogenesis, the effect of activated PPARγ on hepatic steatosis is controversial 35,36 . We assessed whether MD001, as a dual agonist of PPARα/γ, may alleviate fatty liver in db/db mice. As shown in Fig. 6A, rosiglitazone exacerbated hepatic steatosis, as reported in a previous studies 35,36 . However, MD001 treatment resulted in the reduction in size and number of hepatic lipid droplets in a dose-dependent manner, suggesting that MD001 alleviated fatty liver (Fig. 6A). In addition, MD001 significantly reduced hepatic TG and FFA, but not cholesterol; rosiglitazone, on the other hand, failed to reduce hepatic TG (Fig. 6B-D). Quantitative RT-PCR and immunoblot analyses showed that MD001 significantly increased the expression of target genes of PPARα (ACOX, CPT, and MLYCD), and PPARγ (GLUT2, GK, and CD36) (Fig. 6E, Supplementary  Fig. S9), but decreased the expression of genes associated with hepatic lipogenesis, including adipocyte determination and differentiation-dependent factor 1 (ADD1), acetyl-CoA carboxylase (ACC), and fatty acid synthase (FAS) (Supplementary Fig. S10). Thus, MD001 may reduce hepatic steatosis at least in part by stimulating β-oxidation or inhibiting hepatic lipogenesis. Next, we examined the effect of MD001 on fat-cell size in db/db mice, as the progression of obesity induces increases in cell size and adipocyte number. Adipocyte size is one of the indicators of metabolic stresses such as inflammation, insulin resistance, and hyperlipidaemia 37,38 . While rosiglitazone and WY14643 significantly increased the number of fat cells below 100 μm in diameter, they significantly reduced the number of fat cells In addition, qRT-PCR and immunoblot analyses showed that MD001 significantly increased the expression of ACOX, CPT, GLUT4, and CD36, indicating that MD001 may not only induce fatty acid and glucose uptake but also stimulate β-oxidation in adipose tissue at least in part (Fig. 7D, Supplementary Fig. S11). Furthermore, MD001 decreased the expression of inflammatory genes TNFα and MCP-1, as well as macrophage-marker genes CD11b and CD11c ( Supplementary Fig. S12).

Discussion
Type 2 diabetes and its related complications, including hypertension, arteriosclerosis, and diabetic retinopathy are recognised as serious problems in Westernized societies. The advantages of PPARα and PPARγ agonists in the treatment of metabolic syndrome have led to the development of PPARα/γ dual agonists. In this study, we synthesised derivatives of Int B and examined their potential roles as PPARα and PPARγ agonists. Of these, only MD001 enhanced the transcriptional activity of PPARα and PPARγ in vitro. Interaction with PPAR requires acidic hydrogen to act as a hydrogen-bonding donor on the interacting chemical compound. Compounds 4a and 4b lack acidic hydrogen, which may affect their stimulatory activities. Although MD002 bears acidic hydrogen (for hydrogen bonding) and a similar chemical structure, it failed to enhance the transcriptional activity of PPARα and PPARγ, which may be attributable to the different location of its methyl group. Further studies on the relevant structure-activity relationships (SARs) may reveal the mechanism involved.
In vitro binding assays revealed that MD001 exhibited a significantly higher binding affinity for PPARγ than PPARα (Supplementary Table S3). However, MD001 significantly lowered hyperlipidaemia, though the blood glucose lowering effect of MD001 was not as high as that of rosiglitazone, which may be attributable to the existence of a different regulatory mechanism that needs further elucidation. Therefore, the mechanism responsible for MD001-mediated lipid homeostasis is likely to be different from that of rosiglitazone. In vitro and in vivo experiments showed that MD001 induced glucose consumption and β-oxidation in the liver, adipose tissue, and skeletal muscle. In addition, MD001 increased the expression levels of PPARα and PPARγ target genes. These results further confirmed the effect of MD001 on blood glucose levels and hyperlipidaemia through the simultaneous activation of PPARα/γ. The currently-available PPARα/γ dual agonists are associated with PPARγ-related side-effects such as fluid retention and weight gain, limiting their application at higher doses for improved efficacy 39 . Although the blood glucose lowering effect of MD001 was lower than that of rosiglitazone, the MD001-mediated increase in fatty acid oxidation via PPARα activation suggests that MD001 may have favourable effects on hyperlipidaemia and obesity without inducing body weight gain, at least in part. Hepatic steatosis, a common complication in obesity and type 2 diabetes, is closely associated with insulin resistance 40 . MD001 alleviated fatty liver by reducing TG and FFA levels in db/db mice, which was associated with increased expression levels of ACOX, CPT, and MLYCD and decreased expression levels of ADD1, ACC, and FAS by PPARα activation. In addition, hepatomegaly is commonly associated with fatty infiltration of the liver and increased serum ALT 41 . Rosiglitazone induced significant increases in blood ALT and fatty liver, resulting in hepatomegaly (Figs 5I and 6A, Supplementary Fig. S8). MD001-treated obese db/db mice showed significant improvement in fatty liver, with no signs of hepatomegaly-a common PPARα agonist-associated adverse effect in rodents 3,42 . Several PPARα/γ dual agonists, including ragaglitazar, are known to alleviate fatty liver without hepatomegaly 43 , though the underlying molecular mechanisms remain to be elucidated.
About 40% of patients with type 2 diabetes eventually suffer from kidney failure; PPAR agonists are known to have renoprotective effect 44 . Examination of kidney showed that MD001 as well as WY14643 and rosiglitazone significantly reduced the diameter of the glomerular capsule ( Supplementary Fig. S15). In addition, reduction of haemoglobin (Hb) and haematocrit (HCT) levels in PPAR agonist-treated patients is often observed. The examination of RBC, Hb, and HCT showed that rosiglitazone significantly decreased RBC count, Hb, and HCT levels, whereas MD001 did not decrease RBC count, Hb, and HCT levels ( Supplementary Fig. S16), suggesting that MD001 does not induce negative effects on blood profile components. A toxicology study revealed no differences in the white blood cell counts and haemoglobin. Liver and kidney toxicities were not observed in wild type C57/ BL6 mice treated with 50 or 100 mg/kg MD001 for eight weeks (Supplementary Table S4). In addition, there were no significant differences in body weight change (data not shown).
In summary, we have demonstrated that MD001 improved glucose and lipid metabolism in db/db mice through PPARα and PPARγ activation. In addition, MD001 showed no severe adverse effects such as fatty liver, body weight gain, liver toxicity, and hepatomegaly commonly observed with previous PPAR agonists, thereby alleviating metabolic disorders. MD001 ameliorates abnormal lipid profiles through the consumption of excess lipids from peripheral tissues, instead of storage in adipose tissue. The development of MD001 as a PPARα/γ dual agonist targeting metabolic disease may help to overcome the limitations associated with previous PPAR agonists.
Transient transfection and luciferase activity assay. For    In vitro glucose consumption assay. The glucose consumption assay was performed as previously described 23 . Briefly, cells were treated with MD001 or rosiglitazone for three days. The conditioned medium from cultured cells was harvested and assayed for glucose content using a Glucose Colorimetric Assay Kit II (BioVision Inc., Milpitas, CA, USA).
In vivo experiment. The Ajou University Animal Care and Use Committee approved all animal studies (IACUC2015-0001), and the experiment conformed to the Guide for the Care and Use of Laboratory Animals published by the United States National Institutes of Health. All experiments were performed in accordance with relevant guidelines and regulations. Six-week-old C57BLKS/J-Lepr db /Lepr db or wild type C57BL/6J male mice were purchased from Orient Bio, Inc. (Seongnam, Korea) and acclimatised for one week. Mice were randomly grouped and orally administered with vehicle, WY14643 (20 mg/kg), rosiglitazone (20 mg/kg), or MD001 (5 mg/ kg or 20 mg/kg) once a day for two months. For the oral glucose tolerance test (OGTT) and intraperitoneal insulin tolerance test (IPITT), mice were fasted for 12 h and were treated with sterile glucose (1 g/kg, Sigma-Aldrich/ Millipore) or human insulin (1 unit/kg, Eli Lilly and Company, Indianapolis, IN, USA). Blood glucose levels were measured at the indicated time point using an OneTouch Ultra Blood Glucose Monitoring System (LifeScan, Inc., Milpitas, CA, USA). For the toxicity study, wild type C57BL/6J male mice were randomly grouped and orally administered with vehicle and MD001 (50 mg/kg or 100 mg/kg) once a day for two month. Blood cells, body weight change, and blood and urine metabolites were analysed for toxicity.

Biochemical analysis.
Livers and blood samples were collected from mice in each group. Serum and liver TG, FFA, and total cholesterol were measured using WAKO reagents (WAKO chemicals USA, Inc., Richmond, VA, USA) using Hitachi Clinical Analyzer 7180 (Hitachi High-Technologies GLOBAL, Tokyo, Japan). Serum concentrations of low-density lipoprotein (LDL) and high-density lipoprotein (HDL) were determined using Sekisui reagents (Sekisui Medical Co., Ltd., Tokyo, Japan) 47 . Plasma alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels using WAKO reagents (Wako Chemicals USA, Inc) were measured for analysis of hepatotoxicity.
Tissue section and staining. Tissues specimens including liver, perigonadal adipose tissue, skeletal muscle, spleen, kidney, and heart, were collected from mice in each group, fixed in 10% formalin, and embedded into paraffin. Tissue sections (5 μm) were stained with haematoxylin and eosin (H&E).
Statistical analysis. All data were analysed using GraphPad Prism 5.0 software (La Jolla, CA, USA). The results are expressed as mean ± SD or mean ± SEM. Statistical significance was calculated using one-way or two-way analysis of variance (ANOVA) with a post hoc Bonferroni multiple comparison test to compare the differences between groups. P < 0.05 was considered statistically significant.

Data Availability
All data generated or analysed during this study are included in this published article (and its Supplementary  Information files).