Xanthine-based KMUP-1 improves HDL via PPARγ/SR-B1, LDL via LDLRs, and HSL via PKA/PKG for hepatic fat loss[S]

The phosphodiesterase inhibitor (PDEI)/eNOS enhancer KMUP-1, targeting G-protein coupled receptors (GPCRs), improves dyslipidemia. We compared its lipid-lowering effects with simvastatin and explored hormone-sensitive lipase (HSL) translocation in hepatic fat loss. KMUP-1 HCl (1, 2.5, and 5 mg/kg/day) and simvastatin (5 mg/kg/day) were administered in C57BL/6J male mice fed a high-fat diet (HFD) by gavage for 8 weeks. KMUP-1 inhibited HFD-induced plasma/liver TG, total cholesterol, and LDL; increased HDL/3-hydroxy-3-methylglutaryl-CoA reductase (HMGR)/Rho kinase II (ROCK II)/PPARγ/ABCA1; and decreased liver and body weight. KMUP-1 HCl in drinking water (2.5 mg/200 ml tap water) for 1–14 or 8–14 weeks decreased HFD-induced liver and body weight and scavenger receptor class B type I expression and increased protein kinase A (PKA)/PKG/LDLRs/HSL expression and immunoreactivity. In HepG2 cells incubated with serum or exogenous mevalonate, KMUP-1 (10−7∼10−5 M) reversed HMGR expression by feedback regulation, colocalized expression of ABCA1/apolipoprotein A-I/LXRα/PPARγ, and reduced exogenous geranylgeranyl pyrophosphate/farnesyl pyrophosphate (FPP)-induced RhoA/ROCK II expression. A guanosine 3′,5′-cyclic monophosphate (cGMP) antagonist reversed KMUP-1-induced ROCK II reduction, indicating cGMP/eNOS involvement. KMUP-1 inceased PKG and LDLRs surrounded by LDL and restored oxidized LDL-induced PKA expresion. Unlike simvastatin, KMUP-1 could not inhibit 14C mevalonate formation. KMUP-1 could, but simvastatin could not, decrease ROCK II expression by exogenous FPP/CGPP. KMUP-1 improves HDL via PPARγ/LXRα/ABCA1/Apo-I expression and increases LDLRs/PKA/PKG/HSL expression and immunoreactivity, leading to TG hydrolysis to lower hepatic fat and body weight.

The pleiotropic effects of statins, including inhibition of RhoA geranylgeranylation, have been documented ( 22 ). The benefi ts of statins may not only be due to their cholesterol-lowering effects but also due to inhibiting isoprenoid synthesis, the products of which are important lipid attachments for intracellular signaling ( 23 ). We explored whether the contributions of eNOS/cGMP and RhoA/ ROCK II expression to lipid metabolism are affected by KMUP-1.
Application of mevalonate to liver cells results in biosynthesis of isoprenoids, including farnesyl pyrophosphate (FPP) and geranylgeranyl pyrophosphate (GGPP), the levels of which are reduced by HMGR inhibitor statins. GGPP-activated geranylgeranylation of RhoA/ROCK II and the downstream of PPAR ␥ are involved in the regulation of HDL ( 22 ). ATP-binding cassette transporter ABCA1 (member 1 of the human transporter subfamily ABCA) and apolipoprotein A-I (ApoA-I) are involved in the regulation of cholesterol effl ux and are the major protein components of HDL ( 24 ). In this study, we explored whether the nonstatin/non-HMGR inhibitor KMUP-1 can increase HDL by improving PPAR ␥ /SR-B1 in livers. Cellular cholesterol homeostasis is accomplished, in part, by PPARs and Liver X receptor (LXR ␣ ) ( 25 ). Statin-induced RhoA/ROCK inactivation contributes to LXR ␣ / PPARs activation and pleiotropic effects ( 24,25 ). Isoprenoid intermediates affect PPARs and LXR ␣ activation ( 26 ). Activation of isoprenoids produces FPP and GGPP, which inhibit ABCA1 directly by antagonizing LXR ␣ and indirectly through RhoA by activating geranylgeranylation ( 26 ). PPAR ␥ is expressed in fat storage and associated infl ammation ( 27 ). We explored whether KMUP-1 inhibits infl ammatory signaling via the RhoA/ROCK/PPAR ␥ pathway like statins ( 28,29 ).
Increased LDL receptors (LDLRs) by simvastatin can lower the plasma LDL ( 28,29 ). The xanthine analog KMUP-1 is suggested to increase LDLRs/PPAR ␥ and inhibit SR-B1 expression, thus lowering fat. On the other hand, antiatherosclerosis requires activation of eNOS and PPAR ␥ in a LDL receptor (LDLR)-defi cient model (30)(31)(32). KMUP-1, a PDEI, inhibits GPCRs and activates cGMP/PKG via eNOS expression in the cardiovascular system (5)(6)(7)(8). Here we explored its HMGR activity/expression and LDL-lowering and HDL-increasing effects in HFD mice. Immunoblots of HMGR and liver expression of SR-B1/PKA/PKG, proteins associated with the PPAR ␥ pathway, and immunoreactivity or expression of PKA/PKG/hormone-sensitive lipase (HSL)/ LDLRs in livers or HePG2 cells were measured to evaluate their links to lipid metabolism for TG hydrolysis and body/ liver weight lowering.

Animals
In the 8-week experiment, C57BL/6J male mice (21 ‫ف‬ 22 g) were fed a HFD as a model of hyperlipidemia for 8 weeks. Mice were fasted for one night before the experiment and then changed from a standard diet (STD) to a HFD and randomly divided into fi ve groups, including two control and three treatment groups. Six mice were used in each group. The control mice received either STD or HFD, and the treatment group was fed a HFD with KMUP-1 HCl (2.5 and 5 mg/kg/day) or simvastatin (5 mg/kg/day) administered by gavage to assess weight gain, followed by biochemical analysis.
In the 14-week experiment, mice were fed a HFD from week 1 to week 14. KMUP-1 HCl (2.5 mg) was added to 200 ml tap water, and mice had free access to drinking water from week 1 to week 14 (protective group) or from week 8 to week 14 (treatment group). Tap water was used to normalize mineral nutrition. Animals were housed in the animal center with a day-night cycle system at Kaohsiung Medical University. All procedures and protocols were approved by the Animal Care and Use Committee at Kaohsiung Medical University and complied with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health.

Biochemical analysis of serum
In 2 months, the 3 day food intake averaged for each animal was measured. The weight gain and plasma lipid levels of each group were determined and compared with the nontreatment control group. Mouse blood was collected in daytime by cardiac puncture followed by centrifugation at 90 g (Benchtop Centrifuge, Eppendorf, Westbury, NY ) to separate serum, and frozen at Ϫ 80°C for biochemical analysis using a Hitachi Clinical Analyzer 7070 (Hitachi High-Technologies Co., Tokyo, Japan). Agents used in the assays were obtained from Merck and Co. (Kenilworth, NJ). TG) total cholesterol, HDL cholesterol, and LDL cholesterol in mouse serum were measured by methods used in the clinic. To measure the hepatic TG, isolated livers were cut into small chips.

Cell culture
The HepG2 hepatoma cell line was purchased from the American Type Culture Collection (ATCC; Manassas, VA). Cells were cultured in DMEM. Culture media was supplemented with 5% heat-inactivated FBS, penicillin (100 U/ml), and streptomycin (100 g/ml). Cells were grown in a humidifi ed atmosphere containing 5% CO 2 at 37°C, in which the oxygen tension in the incubator was held at 140 mm Hg (20% O 2 , v/v; normoxic conditions). KMUP-1 HCl dissolved in distilled water or simvastatin in vehicle (propylene glycol) was incubated with the cells for 24 h, followed by protein extraction. The fi nal concentration of propylene glycol in medium never exceeded 0.1%.

Western blotting analysis of protein expression in HepG2 cells and livers
HepG2 cells were treated with various concentrations of drugs for 24 h. Reactions were terminated by washing twice with cold PBS, and the cells were then harvested. Proteins in the whole-cell lysate were homogenized in ice-cold lysis buffer and protease inhibitor (Sigma-Aldrich, St. Louis, MO). The homogenate was centrifuged at 20,000 g for 15 min at 4°C, and supernatant was recovered as the total cellular protein. Cytosolic and membrane fractions of HepG2 cells were prepared using a CNM (Cytosol Nuclear Membrane) compartment protein extraction kit (Bio-Chain Institute Inc., Hayward, CA) according to the manufacturer's instructions. All of the fractionated protein solutions were stored at Ϫ 80°C until analysis. To measure the expression levels of proteins by drugs, the total cell protein was extracted after incubation with treatments for 24 h, and then Western blotting analyses were performed as described previously ( 7,8 ). For the [ 14 C]HMG-CoA for another 20 min incubation period and terminated by further addition of 1 N HCl. An aliquot was removed by column and counted to determine the amount of [ 14 C]mevalonate formed (Ricerca Co. Ltd., Taipei, Taiwan).

cGMP pathway and RhoA/ROCK II expression
To confi rm that RhoA antagonist C3 exoenzyme (5 g/ml) and ROCK antagonist Y27632 (10 M), dissolved in 10% propylene glycol, can inactivate ROCK II, they were added to cells in culture for 24 h to measure the expression of ROCK II and related expression of PPAR ␥ and ABCA1 in HepG2 cells. To confi rm that the cGMP antagonist Rp-8-pCPT-cGMPS (10 M), dissolved in 10% propylene glycol, can increase ROCK II and that KMUP-1 can reduce Rp-8-pCPT-cGMPS-induced activation of ROCK II, HepG2 cells were preincubated with Rp-8-pCPT-cGMPS for 30 min as control and then in combination with KMUP-1 (10 M) for 24 h.

Immunohistochemistry staining of LDLRs in livers
Liver tissues were fi xed in 10% buffered formalin for 24 h and then embedded in paraffi n. The paraffi n-embedded liver tissue sections (4 m thick) were fi rst heat immobilized and deparaffi nized using xylene and then rehydrated in a graded ethanol series, followed by a fi nal wash in distilled water. Finally, tissue sections were stained with PAS and Mayer's hematoxylin solution. For IHC of hepatic LDLRs in animals after drinking KMUP-1 expression of SR-B1, HMGR, PPAR ␥ , and ROCK II, isolated liver tissues cut into small chips were placed into extraction buffer (Tris 10 mM [pH 7.0], NaCl 140 mM, PMSF 2 mM, DTT 5 mM, NP-40 0.5%, pepstatin A 0.05 mM, and leupeptin 0.2 mM) for hepatic protein extraction and centrifuged at 20,000 g for 30 min. The obtained protein extract was boiled to a ratio of 4:1 with sample buffer (Tris 100 mM [pH 6.8], glycerol 20%, SDS 4%, and bromophenol blue 0.2%). Electrophoresis was performed using 10% SDS-PAGE (1 h, 100 V, 40 mA, 20 g protein per lane). Separated proteins, after three repeated centrifugations to discard up-layer tissue lipid inpurity, were transferred to PVDF membranes treated with 5% fat-free milk powder to block the nonspecifi c IgGs (90 min, 100 V) and incubated for 1 h with specifi c protein antibody. The blot was then incubated with antimouse or anti-goat IgG linked to alkaline phosphatase (1:1,000) for 1 h.  PKG by Western blotting or PKA/PKG and HSL immunoreactivity by fl uorescence staining combined with image scanning in the absence or presence of oxidized LDL (200 g/ml).

Statistical evaluation
The experimental results from KMUP-1 and simvastatin were expressed as means ± SE. Statistical differences were determined by independent and paired Student's t -test in unpaired and paired samples, respectively. Whenever a control group was HCl (2.5 mg/200 ml for 1-14 weeks or 8-14 weeks), antigen retrieval of deparaffi nated sections was performed in Dako target retrieval solution (pH 9.0) in a vegetable steamer followed by quenching of endogenous peroxidase activity with 3.0% H 2 O 2 in methanol. Sections were then incubated with specifi c primary antibodies overnight at 4°C in a humidifi ed chamber. The sections were then examined using a REAL EnVision TM Detection System kit (DAKO, Carpinteria, CA) and counterstained with hematoxylin. Images were obtained through a Nikon Eclipse TE200-S microscope.

Expression and fl uorescence staining of LDLRs in the presence of exogenous LDL
HepG2 cells were used to determine the cellular protein expression of LDLRs in the presence of exogenous LDL (500 g/ ml). Bodipy-493/503 (green) and LDLRs on HepG2 cells were detected with a secondary antibody conjugated to Cy3 (red) overnight at 4°C followed by merger of obtained BODIPY and LDL images to analyze the location of LDLRs. All images were collected and analyzed by scanning with a Nikon Eclipse TE200-S microscope (Tokyo, Japan).

Expression of PKA/PKG and immunoreactivity of PKA/HSL
To determine that KMUP-1 can affect PKA, we incubated KMUP-1 (10 STD group. HFD-induced hypercholesterolemia was significantly improved by KMUP-1 supplementation. In particular, the HDL cholesterol level was signifi cantly increased by KMUP-1 and simvastatin. When the food intake in animals fed a HFD after maturity (8 weeks) slowed down, the feeding period was prolonged to 14 weeks. Some factors that affected food intake remained unclear.

Weight changes and gross liver morphology
Drinking KMUP-1 HCl (2.5 mg/200 ml water) by mice fed a HFD decreased the body weight in both the protection and treatment groups ( Fig. 1A ). Fatty tissues were characteristically found on the surface of HFD livers ( Fig.  1B ). Fatty liver was markedly decreased in the protective group, and this effect was more prominent than in the treatment group ( Fig. 1B ).

HFD-induced SR-B1, HMGR, ROCK II, PPAR ␥ , and ABCA1 liver expression
In terms of the effects of KMUP-1 on increased HDL, drinking KMUP-1 was observed to inhibit HFD-induced hepatic SR-B1 expression and to promote PKA/PKG compared with more than one treated group, one-way ANOVA or two-way repeated measures ANOVA was used. When the ANOVA showed a statistical difference, the Dunnett's or Student-Newman-Keuls test was applied. A P value <0.05 was considered signifi cant in all experiments. Analysis of the data and plotting of the fi gures were done using SigmaPlot software (Version 8.0, Chicago, IL) and SigmaStat (Version 2.03, Chicago, IL) run on an IBM compatible computer. Table 1 shows the 8-week body-weight gain of animals fed with a STD or HFD. Consumption of HFD for 8 weeks signifi cantly increased body weight compared with the STD group ( p < 0.05). KMUP-1 HCl (2.5, 5 mg/kg p.o.) and simvastatin supplementation (5 mg/kg p.o.) reduced body weight gain compared with the control HFD group ( p < 0.05). HFD caused dramatic increases in serum TG, total cholesterol, and LDL cholesterol compared with the end-product feedback regulation phenomenon of HMGR ( Fig. 3D ).

Decreased RhoA/ROCK II and enhanced eNOS expression
KMUP-1 concentration-dependently inhibited the translocation of RhoA from cytosol to membrane in HepG2 cells ( Fig. 4A ). ROCK II is the downstream effector of RhoA in hepatic cellular signaling. KMUP-1 or simvastatin (10 Ϫ 9 -10 Ϫ 5 M) concentration-dependently reduced ROCK II protein expression due to inhibition of RhoA translocation ( Fig. 4B, C ). KMUP-1 concentration-dependently increased the expression of eNOS and accordingly resulted in decreased RhoA/ROCK II expression in HepG2 cells ( Fig. 4D ).

Increased RhoA/ROCK II expression in the presence of GGPP and FPP
Application of exogenous GGPP and FPP increased RhoA/ROCK II expression, and KMUP-1 (10 Ϫ 9 -10 Ϫ 5 M) attenuated this phenomenon in HepG2 cells ( Fig. 8A , C, D ). In contrast, simvastatin did not decrease ROCK II expression in the presence of exogenous GGPP and FPP ( Fig. 8B ).

IHC of LDLRs and PKG/PKA expression
HFD-induced LDLRs expression in livers was estimated using IHC staining methods. Notably, drinking KMUP-1 HCl increased the hepatic LDLRs of HFD animals in both could be near the viability range of HepG2 cells ( Fig. 11A ). PKA and HSL (green fl uorescence) also showed increased immunoreactivity in HepG2 cells treated with KMUP-1 or simvastatin ( Fig. 11B and C ). However, PKG immunoreactivity was not signifi cantly affected by KMUP-1 (data not shown).

Fluorescent staining of cellular LDLRs/PKA/HSL
HepG2 cells were stained with fl uorescence and treated with different concentrations of KMUP-1 or simvastatin for 24 h. Results showed increased LDLR (green fl uorescence) expression in HepG2 cells with different concentrations of KMUP-1 (10 Ϫ 6 , 10 fi ndings suggest that RhoA/ROCK II is the target protein to elevate cGMP ( 7 ). KMUP-1 decreases ROCK II expression by enhancing the eNOS/cGMP pathway as simvastatin does. The cGMP-dependent action of KMUP-1 was made evident by pretreatment with a cGMP antagonist, Rp-8-pCPT-cGMPS. Simvastin inhibited HMGR activity and the geranylgeranylation of RhoA/ROCK II. Notably, KMUP-1 inhibited GGPP-or FPP-activated geranylgeranylation of RhoA/ROCK II, which is dependent on the cGMP pathway, independently of the inhibition of HMGR activity. GGPP or FPP was shown to increase cell permeability using liposome preparation techniques ( 28,29 ). A large concentration of GGPP or FPP was added to culture medium to obtain similar effects and to prevent the undesired side effects of the use of liposomes. PPAR ␥ , downstream of ROCK II signaling, has an important role modulating HDL ( 23 ). Unlike statin's inhibition of RhoA geranylgeranylation, KMUP-1 enhances the cGMP pathway to inactivate RhoA and reverses PPAR ␥associated ABCA1 expression to improve HDL, even in the presence of isoprenoids. Cholesterol effl ux to ApoA-I is processed in ABCA1-expressing liver cells, a major housekeeping mechanism for cellular cholesterol homeostasis. Both ABCA1 and ApoA-I play critical roles in the formation of HDL ( 24,25 ). KMUP-1 increased the expression of ABCA1 and ApoA-I, which might contribute to the elevation of plasma HDL concentrations.
KMUP-1 and simvastatin have been shown to increase eNOS/cGMP and inhibit the ROCK II pathway in the cardiovascular system, potentially inhibiting atherosclerosis (5)(6)(7)33 ). Elevated cGMP/PKG in livers also potentially affects the lipid catabolism of hepatocytes by lipolysis of intracellular oil globulets through HSL ( 15 ). Inhibition of ROCK II by PKA has been shown to enhance adipogenesis and thus has no antiobesity benefi ts ( 26 ). We confi rmed that KMUP-1's increase of liver PKG is similar to adipocytes activated via inducible NOS and released NO ( 27 ). KMUP-1 can increase PKA immunoreactivity and inhibit ROCK II expression but not the immunoreactivity of PKG in HepG2 cells. KMUP-1 and simvastatin enhance HSL activities through increased expression of PKG/PKA in livers or PKA immunoreactivity in HepG2 cells ( 6,7,30 ). However, whether eNOS enhancement by KMUP-1 increases HSL immunoreactivity through activated cGMP/PKG remains to be further investigated.
In this study, we demonstrated that KMUP-1 attenuates HFD-induced hepatic SR-B1 expression. Taken together, we suggest that the increase in HDL by KMUP-1 administration could be attributed to inhibition of SR-B1 and activation of PPAR ␥ -associated signaling cascades ( Fig. 12 ). levels of LDL. We thus used fl uorescent staining of HepG2 cells to compare LDLRs/PKA/HSL immunoreactivity in this study.
In the hyperlipidemic state caused by HFD, HMGR expression is lower than with a STD in a mouse model. Expression of HMGR was also suppressed in HepG2 cells in a serum medium ( 23 ). Administration of either simvastatin or KMUP-1 reversed HFD-induced downregulation of HMGR and RhoA/ROCK II expression via geranylgeranylation. HMGR is subject to feedback control through multiple mechanisms, mediated by end-products of mevalonate metabolism ( 18,23 ). Interestingly, simvastatin could but KMUP-1 could not inhibit the production of 14 C mevalonate.
Application of GGPP induces RhoA translocation and GTP binding to RhoA (24)(25)(26)(27), which results in ROCK II activation. KMUP-1 decreases ROCK II expression via a cGMP-dependent pathway ( 7 ). Treatment of HepG2 cells with C3 exoenzyme, a RhoA antagonist, inactivated ROCK II by inhibiting the translocation of RhoA, whereas Y27632, a ROCK antagonist, can directly inactivate ROCK II. These  Statins reduce the formation of isoprenoids, which are responsible for posttranslational modifi cation of proteins. Simvastatin enhanced eNOS/PPAR ␥ expression, inhibited geranylgeranylation activity, and increased PKA immunoactivity, in contrast to previous negative expression by Western blotting in HepG2 cells (38)(39)(40)(41). KMUP-1 enhances eNOS/PPAR ␥ expression without HMGR activity, suggesting that inhibition of HMGR activity is not required for KMUP-1 to improve lipid accumulation. KMUP-1 lacks HMGR activity, but it increases LDLRs, eNOS, and PPAR ␥ expression; inhibits SR-B1 expression like simvastatin; and maintains regulatory activity on geranylgeranylation and its feedback system from RhoA ( 38,42 ). Thus, KMUP-1 can inhibit lipid accumulation and HFD-driven infl ammation in livers. Previously, PPAR ␥ agonist activity was shown to affect weight gain in adipose tissues, the storage sites related to lipid accumulation and mobilization ( 43 ). The reduction of weight gain by KMUP-1 is parallel to changes in liver weight, accompanied by increase of hepatic PPAR ␥ expression and lowering of TG in liver/serum. These facts indicate that HDF-induced accumulation/mobilization of TG and infl ammation in fatty livers were inhibited by KMUP-1 via improved PPAR ␥ /SR-B1 expression and elevated PKG/ PKA/HSL expression or immunoreactivity. PKA and PKG are increased by nonspecifi c phosphodiesterase inhibitors and/or eNOS activators ( 1,2,7 ). Elevated cGMP is accompanied by increased HSL with antiobesity effects ( 7,44 ). In the present study, simvastatin and KMUP-1 quantitively increased HSL/PKA/LDLRs immunoreactivity in HepG2 cells. Increased TG hydrolysis via PKA/PKG/ HSL in lipolysis and inhibition of adipogenesis in peripheral adipocytes are crucial for antiobesity effects, besides inhibiting biosynthesis of cholesterol via HMGR activity ( 45 ). KMUP-1 may decrease LDL-associated lipid metabolism or remove plasma LDL via increasing LDLRs, leading to circulation and hepatic fat loss via HSL around the lipid droplets of adipocytes in the body and at the sites of lipid storage in hepatic cells.
Evidence from liver IHC and fl uorescent staining of LDLRs in HepG2 cells suggests that most LDLRs are expressed on cell membranes, which allows LDL-cholesterol to be bound and internalized via an endocytosis mechanism and prevents LDL from diffusing around the membrane surface. KMUP-1 removed plasma LDL by activating hepatic LDLRs, increased HDL via PPAR ␥ activation and SR-B1 inhibition, attenuated RhoA geranylgeranylation via eNOS/cGMP, and caused fat loss via translocation of HSL through PKA. However, LDL is oxidized in infl ammatory fatty livers, and PKA expression is decreased. oxLDL results in the increase of fatty acid synthesis. KMUP-1 reverses oxLDL-reduced PKA expression in HepG2 cells, suggesting that it would protect circulating LDL against oxidization and decrease fatty acid accumulation ( 46,47 ).
In conclusion, decreases in weight gain and liver/ serumTG, increased HDL, and enhanced LDLRs/HSL expression suggest that hepatic fat loss can be achieved by administering the nonstatin xanthine analog KMUP-1, making it a hopeful treatment for obesity and infl ammatory fatty liver. KMUP-1, a PDEI and eNOS enhancer, affects multiple signaling cascades, including expression of PPAR ␥ /SR-B1/LDLRs/PKA/PKG/HSL, involved in hepatic fat loss and body-weight lowering effects ( Fig. 12 ).