Identification of Picrasidine C as a Subtype-Selective PPARα Agonist

Dec 13, 2016 - *Tel (Y. Kanno): +81 47 472 2539. Fax: +81 47 472 2539. E-mail: [email protected]., *Tel (W. Li): +81 47 472 1161. Fax: +81 47 4...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/jnp

Identification of Picrasidine C as a Subtype-Selective PPARα Agonist Shuai Zhao,† Yuichiro Kanno,*,‡ Wei Li,*,§ Tatsunori Sasaki,§ Xiangyu Zhang,⊥ Jian Wang,⊥ Maosheng Cheng,⊥ Kazuo Koike,§ Kiyomitsu Nemoto,‡ and Huicheng Li*,† †

College of Life Science, Northeast Forestry University, Harbin 150040, People’s Republic of China Department of Molecular Toxicology, Faculty of Pharmaceutical Sciences, Toho University, Miyama 2-2-1, Funabashi, Chiba 274-8510, Japan § Faculty of Pharmaceutical Sciences, Toho University, Miyama 2-2-1, Funabashi, Chiba 274-8510, Japan ⊥ Key Laboratory of Structure-Based Drug Design and Discovery of Ministry of Education, Shenyang Pharmaceutical University, Shenyang 110016, People’s Republic of China ‡

ABSTRACT: Picrasidine C (1), a dimeric β-carboline-type alkaloid isolated from the root of Picrasma quassioides, was identified to have PPARα agonistic activity by a mammalian one-hybrid assay from a compound library. Among the PPAR subtypes, 1 selectively activated PPARα in a concentrationdependent manner. Remarkably, 1 also promoted PPARα transcriptional activity by a peroxisome proliferator response element-driven luciferase reporter assay. Furthermore, 1 induced the expression of PPARα-regulated genes involved in lipid, glucose, and cholesterol metabolism, such as CPT-1, PPARα, PDK4, and ABCA1, which was abrogated by the PPARα antagonist MK-886, indicating that the effect of 1 was dependent on PPARα activation. This is the first report to demonstrate 1 to be a subtype-selective PPARα agonist with potential application in treating metabolic diseases, such as hyperlipidemia, atherosclerosis, and hypercholesterolemia. fibrate drugs, such as fenofibrate, ciprofibrate, and bezafibrate, showed only modest selectivity for PPARα over the other PPAR subtypes,9 resulting in side effects and safety issues that were identified during clinical trials to be associated with the induction of other PPAR subtypes.10 Efforts to improve the fibrate class or nonfibrate compounds’ pharmacological profiles led to the synthesis of selective PPARα agonists. However, only a few of such compounds were identified, such as GW764711 and WY14643,12 to have varying degrees of PPARα selectivity. Therefore, identifying more potent subtype-selective PPARα agonists is an important goal. Plants used in traditional medicine have been attracting attention for their ability to ameliorate metabolic diseases, including diabetes mellitus and hyperlipidemia- and hypercholesterolemia-associated abnormal lipid metabolism. Both extracts and natural compounds have been reported to exert the hypolipidemic effects through activating PPARα.13−19 With the aim of identifying selective PPAR subtype agonists, we screened a compound library extracted from natural products using a mammalian one-hybrid assay. We identified picrasidine C (1), a dimeric β-carboline-type alkaloid, which was isolated from the root of Picrasma quassioides, as a subtypeselective PPARα agonist that could induce mRNA expression of PPARα-regulated genes, such as CPT-1, PPARα, PDK4, and ABCA1.

P

eroxisome proliferator activated receptors (PPARs) are ligand-activated transcription factors that belong to the nuclear receptor superfamily.1,2 PPARs contain three known subtypes, PPARα, -γ, and -β/δ, which share high sequence and structural homology, but have unique tissue expression and physiological functions.3,4 When activated by their ligand, PPARs heterodimerize with the retinoid X receptor (RXR), bind to PPREs (peroxisome proliferator hormone response elements), and regulate the transcription of various genes (such as CPT-1, PPARα, PDK4, and ABCA1), thereby influencing various cellular functions, including the regulation of glucose, lipid, and cholesterol metabolism. They have consequently been identified as ideal targets for treating metabolic diseases, including diabetes, obesity, hypertension, and dyslipidemia. The identification of novel PPAR agonists for drug development has attracted an increasing amount of attention. In rodents and humans, PPARα is expressed in tissues with high rates of fatty acid oxidation, including the liver, kidney, heart, skeletal muscle, and brown adipose tissue;5,6 it is also expressed in a range of vascular cells, such as endothelial cells, vascular smooth muscle cells (VSMCs), and monocytes/ macrophages.7 Previous investigations have demonstrated the critical role that PPARα plays in regulating lipid homeostasis, and it has also been an attractive molecular target for drug development. For example, fibrates, which predominantly target PPARα, have become a well-established class of drugs for treating hypertriglyceridemia and mixed dyslipidemia over the last several decades.8 However, these synthetic hypolipidemic © XXXX American Chemical Society and American Society of Pharmacognosy

Received: September 28, 2016

A

DOI: 10.1021/acs.jnatprod.6b00883 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

this further, we cotreated 1 (10 μM) with GW0742 (10 μM), a well-known PPARβ/δ-selective agonist. However, 1 could not reverse GW0742-induced PPARβ/δ activation, indicating that 1 is not an antagonist of PPARβ/δ (Figure 1D). Furthermore, we performed in silico docking calculations. We docked compound 1 to the proteins to evaluate the affinity and selectivity of 1 to PPARα. Compound 1 fit well within the PPARα ligand binding pocket (Figure 2A) and formed more interactions, including hydrogen bonds with residues Cys276 and Thr279 and hydrophobic interactions with Ala333, His440, Val444, Phe273, Cys275, Val 332, and Ile 272 (Figure 2D), compared to PPARβ/δ (Figure 2B and E) and PPARγ (fewer interactions, Figure 2C and F), probably because PPARα offered a more suitable pocket for ligand binding than other subtypes. These data suggest that 1 is a subtype-selective PPARα agonist. To examine the effect of 1 on PPARα-mediated PPRE transcriptional activation, a PPRE-luc reporter assay was conducted using a PPRE-driven luciferase reporter-gene (PPRE-luc) plasmid and PPARα and RXRα gene expression plasmids in HepG2 cells. Both 1 (10 μM) and WY14643 (50 μM) significantly increased PPRE-luc activity in cells that were cotransfected with the PPARα and RXRα gene expression plasmids (Figure 3), indicating that 1 not only activated PPARα but also promoted PPARα-mediated PPRE transcriptional activation. Next, the effect of 1 on PPARα-regulated gene expression was further examined. The human monocytic leukemia cell line THP-1 expresses endogenous PPARα26 and differentiates into macrophage-like cells, mimicking native monocyte-derived macrophages, after treatment with phorbol esters.27 This cell line was used as a tool to investigate the effect of 1 on PPARαregulated gene expression. THP-1 cells were first treated with phorbol 12-myristate 13-acetate (PMA) (200 nM) for 48 h to induce differentiation into macrophage-like cells. Then, cells were treated with 1 (1, 10, 20 μM), WY14643 (50 μM), or solvent (0.1% DMSO) for 24 h. The mRNA expressions of the PPARα-regulated genes carnitine palmitoyltransferase-1 (CPT-1),28 peroxisome proliferator-activated receptor α (PPARα),29 pyruvate dehydrogenase kinase 4 (PDK4),30 and ATP binding cassette transporter A1 (ABCA1)31 were measured by real-time quantitative-PCR (RT-qPCR). Compared with WY14643 (50 μM), 1 (1, 10, 20 μM) concentration-dependently increased PPARα-regulated gene expression, resulting in increased mRNA levels for CPT-1, PPARα, PDK4, and ABCA1 (Figure 4A−D). The induction of PPARα target genes by 1 was variable among gene species. As shown in Figure 3, CPT-1 and PDK4 were moderately induced by 1, while PPARα and ABCA1 were strongly induced by 1. In particular, 1 markedly increased PPARα mRNA expression at the relatively low concentration of 1 μM to levels comparable to those of WY14643 at 50 μM. These data indicate that 1 might be a gene-selective PPARα agonist, which is distinct from that of the well-known PPARα agonist WY14643. A structural specificity of 1 may be responsible for this regulation. Further investigation on the structure−activity relationship as well as the protein expression is needed. To further confirm this PPARα agonistic effect on target gene mRNA induction, the cells were pretreated with MK-886, a selective PPARα antagonist.32 The induction of CPT-1, PPARα, PDK4, and ABCA1 mRNA expression induced by 1 and WY14643 was reversed by MK-886 treatment (Figure 5A−D), demonstrating that the agonistic activity of 1 on mRNA induction of PPARα-regulated genes was dependent on PPARα activation.



RESULTS AND DISCUSSION The compound library consisted of 76 natural alkaloid compounds from Ailanthus altissima, Picrasma javanica, Simaba cuspidata, Quassia amara, Simarouba amara, Eurycoma longifolia, and Picrasma quassioides, as well as their structurally related chemical synthetic analogues.20 With the aim of identifying more selective PPAR subtype agonists, a mammalian onehybrid assay was performed as previously reported.21 Briefly, a GAL4-DNA binding domain (DBD)-fused human PPARαligand binding domain (PPARα-LBD), PPARβ/δ-LBD, PPARγ-LBD protein expression plasmid (GAL4/PPARαLBD, GAL4/PPARβ/δ-LBD, GAL4/PPARγ-LBD), and GAL4responsive reporter gene (GALRE-luc) expression plasmid were constructed (Chart 1A, B, C) and used in mammalian Chart 1. Plasmid Structural Representations

one-hybrid assays in HepG2 cells, a human hepatocarcinomaderived cell line. In this study, GAL4/PPARα-LBD and GALRE-luc were used to identify candidates from the compound library described above with PPARα agonistic activity. Picrasidine C (1) was shown to have remarkable PPARα agonistic activity at a concentration of 10 μM, which is superior to the positive control WY14643, a well-known PPARα-selective agonist.12,22,23 However, structurally related derivatives did not show significant activity in the same assay at a concentration of 10 μM. Picrasidine C, which was isolated from the root of Picrasma quassioides, is a dimeric β-carboline-type alkaloid. Since the three identified PPAR subtypes, α, β/δ and γ, share a high sequence and structural homology, the selectivity of PPARα induction by 1 was confirmed using the mammalian one-hybrid assay of all three PPAR subtypes described above with various concentrations (0.1−20 μM) of 1 and positive control PPAR subtypes: WY14643 (PPARα agonist),12 GW0742 (PPARβ/δ agonist),24 and troglitazone (PPARγ agonist).25 Compound 1 showed a significant concentrationdependent increase of luciferase activity on GAL4/PPARαLBD, superior to the positive control (WY14643), at a concentration of 0.5 μM (Figure 1A). However, no increases on PPARβ/δ and PPARγ were observed (Figure 1B,C). There was a constant decrease of PPARβ/δ activity, indicating the possibility that 1 is an antagonist of PPARβ/δ. To investigate B

DOI: 10.1021/acs.jnatprod.6b00883 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Figure 1. Selectivity of 1 for PPAR subtypes. HepG2 cells were transfected with pGL4.74 (0.01 μg) and the GALRE-luc reporter plasmid (0.1 μg), together with the expression vector for (A) GAL4/PPARα-LBD (0.05 μg), (B) GAL4/PPARβ/δ-LBD (0.05 μg), and (C) GAL4/PPARγ-LBD (0.05 μg). Cells were treated with solvent (0.1% DMSO), a positive control for PPAR subtypes, or increasing concentrations of 1 (0.1−20 μM). Positive controls for PPAR subtypes were WY14643 for GAL4/PPARα-LBD, GW0742 for GAL4/PPARβ/δ-LBD, and troglitazone for GAL4/ PPARγ-LBD. (D) HepG2 cells were transfected with pGL4.74 (0.01 μg) and the GALRE-luc reporter plasmid (0.1 μg), together with the expression vector for GAL4/PPARβ/δ-LBD (0.05 μg). Cells were treated with solvent (0.1% DMSO), a positive control for PPAR β/δ (10 μM), 1 (10 μM), and both of them. After 24 h, the luciferase activity was measured using the dual-luciferase reporter assay system. The results are shown as the average fold activation over the solvent control after Renilla normalization of four individual transfection experiments (mean ± SD, n = 4); *P < 0.05. **P < 0.01.



It has been well established that fibrates reduce plasma triglycerides largely by inducing hepatic PPARα-mediated transcription of key proteins in hepatic long-chain fatty acid β-oxidation, such as CPT1.33 Fibrates rapidly induce the PPARα-regulated gene PDK4 in various tissues to decrease the serum levels of triglycerides and promote energy metabolism.34 Fibrates enhance the expression of ABCA1 by activating PPARα, contributing to an increase in high-density lipoprotein biogenesis and inducing cholesterol removal.26,35 In this study, 1 also induced PPARα-regulated genes induced by fibrates, as mentioned above. Considering the selective PPARα agonistic activity and novelty of the structure of 1, which is different from all other marketed PPARα-activated drugs, including fibrates, 1 could be a PPAR subtype-selective PPARα agonist with potential for application in treating metabolic diseases, such as hyperlipidemia, atherosclerosis, and hypercholesterolemia. Our knowledge of the biofunctions of 1 is currently quite limited. To develop 1 into a drug for clinical use, further investigation of its potential off-target effects, toxicity, pharmacokinetics, bioavailability, and ADME (absorption, distribution, metabolism, excretion) is needed.

EXPERIMENTAL SECTION

Cell Culture. HepG2 cells were cultured in Dulbecco’s modified Eagle’s medium (Wako, Osaka, Japan) containing 10% fetal bovine serum, penicillin (100 units/mL), and streptomycin (100 μg/mL). THP-1 cells were cultured in RPMI-1640 medium (Wako, Osaka, Japan) containing 10% fetal bovine serum, penicillin (100 units/mL), and streptomycin (100 μg/mL). Cells were incubated in a humidified atmosphere with 5% CO2 at 37 °C. PMA (200 nM) was used to differentiate THP-1 cells into cells with a macrophage-like appearance. Chemicals. Troglitazone and WY14643 were purchased from Sigma-Aldrich (St. Louis, MO, USA). GW0742 was purchased from Sigma-Aldrich (Steinheim, Germany). MK-886 was purchased from Wako (Osaka, Japan). PMA was purchased from CALBIOCHEM (Calbiochem-EMD Biosciences; San Diego, CA, USA). The natural dimeric derivate of β-carboline compound picrasidine C (P1) used in this study was isolated from the root of Picrasma quassioides BENNET (Simaroubaceae, Japanese name “Nigaki”) according to methods set forth in a previous study.36 The purity of the compound was confirmed to be >98% by 1H NMR and HPLC analyses. Plasmid Construction. Expression plasmids containing the GAL4 DNA-binding domain (GAL4) fused to the PPARα-ligand-binding domain (LBD) (159−468 a.a.), PPARβ/δ-LBD (136−441 a.a.), and PPARγ-LBD (157−475 a.a.), as well as the PPRE-driven luciferase reporter plasmid (pPPRE-Luc), were constructed as previously reported.21 C

DOI: 10.1021/acs.jnatprod.6b00883 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Figure 2. Binding mode of 1 with PPARα. Protein of PPARα (A), PPARβ/δ (B), and PPARγ (C) is displayed in ribbon style, and 1 is displayed in green CPK style. Interaction pattern of 1 with PPARα (D), PPARβ/δ (E), and PPARγ (F) is shown as a 2D diagram, with hydrogen bonds represented as green dot lines and hydrophobic interactions represented as pink dotted lines. HepG2 cells and inserted into pcDNA5/TO (Invitrogen, Carlsbad, CA, USA) with an N-terminal myc tag and pCMV-3tag6 (Stratagene, Santa Clara, CA, USA), respectively. The following oligonucleotide primers were used to amplify insertion sequences: for PPARα, FW-5′GCGGGATCCGTGGACACGGAAAGCCCACTCTGCCCC-3′ and Rev-5′-ATCTCAGTACATGTCCCTGTAGATCTCCTGCA-3′; for RXRα, FW-5′ ATCATGGACACCAAACATTTCC-3′ and Rev-5′CTAAGTCATTTGGTGCGG-3′. Luciferase Reporter Analysis. HepG2 cells were transfected with the appropriate expression and reporter plasmids, as well as pGL4.74 (Promega) as an internal standard, using the PEI Max reagent (Polysciences, Warrington, PA, USA). The cells were treated with individual test compounds for 24 h after incubation overnight. Using the dual-luciferase reporter assay system (Promega), the luciferase activities were measured. The experimental firefly luciferase activities were normalized against Renilla luciferase activities. Quantitative Reverse-Transcription Polymerase Chain Reaction. Total RNA was isolated from whole-cell lysates using ISOGEN II (Nippon Gene, Toyama, Japan), and cDNA was synthesized using a ReverTraAce qPCR RT kit (Toyobo, Osaka, Japan). Quantitative polymerase chain reaction was conducted (40 cycles of 98 °C for 5 s, 55 °C for 20 s, and 72 °C for 30 s) using the KOD SYBR qPCR Mix (Toyobo) and the 7500 fast system SDS software (Applied Biosystems, Life Technologies, Carlsbad, CA, USA) according to the manufacturers’ protocol. Specific primers were used to target ABCA1 (5′-TGTCCAGTCCAGTAATGGTTCTGT-3′ and 5′-AAGCGAGATATGGTCCGGATT-3′), CPT-1 (5′ACAGTCGGTGAGGCCTCTTAT-3′ and 5′-TCTTGCTGCCTGAATGTGAGT-3′), PPARα (5′-CTATCATTTGCTGTGGAGATCG-3′ and 5′-AAGATATCGTCCGGGTGGTT-3′), PDK4 (5′-AGAGCCTGATGGATTTGGTG-3′ and 5′-GCTTGGGTTTCCTGTCTGTG-3′), and β-actin (5′-TCCTCCTGAGCGCAAGTACTC-3′ and 5′-CTGCTTGCTGATCCACATCTG-3′).

Figure 3. Effect of 1 on the transcriptional activity of PPARα. HepG2 cells were cotransfected with pGL4.74 (0.01 μg), the PPRE-luc reporter plasmid (0.1 μg), the PPARα expression vector (0.05 μg), and/or the RXRα expression vector (0.05 μg). Cells were treated with solvent (0.1% DMSO), WY14643 (50 μM) as a positive control, or 1 (10 μM). After 24 h, luciferase activity was measured using the dual-luciferase reporter assay system. The results are shown as the average fold activation over the solvent control after Renilla normalization of four individual transfection experiments (mean ± SD, n = 4); *P < 0.05. **P < 0.01. The GAL4-responsive reporter gene (pG5-luc) was obtained from Promega (Madison, WI, USA). The coding sequences for full-length human PPARα and RXRα were amplified from cDNA extracts of D

DOI: 10.1021/acs.jnatprod.6b00883 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Figure 5. Effects of MK-886 on 1-induced mRNA expression on PPARα-regulated genes in THP-1 cells. THP-1 cells were first treated with PMA (200 nM) for 48 h to induce differentiation into a macrophage-like appearance. Cells were first pretreated with MK-886 (200 μM) for 1 h; then, cells were treated with solvent (0.1% DMSO), WY14643 (50 μM), or 1 (20 μM). After 24 h, cells were harvested, total RNA was extracted, and the levels of (A) CPT-1, (B) PPARα, (C) PDK4, and (D) ABCA1 mRNA were measured by RT-qPCR. Results were normalized against the β-actin mRNA levels and expressed as the average of the fold induction over the solvent control of four individual experiments (mean ± SD, n = 4); *P < 0.05. **P < 0.01.

Figure 4. Effect of 1 on mRNA expression of PPARα target genes in THP-1 cells. THP-1 cells were first treated with PMA (200 nM) for 48 h to induce differentiation into a macrophage-like appearance. Then, cells were treated with solvent (0.1% DMSO), WY14643 (50 μM), or increasing concentrations of 1 (1 μM, 10 μM, 20 μM). After 24 h, cells were harvested, total RNA was extracted, and the levels of (A) CPT-1, (B) PPARα, (C) PDK4, and (D) ABCA1 mRNA were measured by RT-qPCR. Results were normalized against the β-actin mRNA levels and expressed as the average of the fold induction over the solvent control of four individual experiments (mean ± SD, n = 4); *P < 0.05. **P < 0.01. E

DOI: 10.1021/acs.jnatprod.6b00883 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products



Molecular Docking. AutoDock 4 was used to perform docking calculations.37 The crystal structures of PPARα (PDB code: 3KDT),38 PPARβ/δ (PDB code: 3OZ0),39 and PPARγ (PDB code: 2ATH)40 were used in docking calculations. A grid box with a grid spacing of 0.375 Å was generated to define the binding pocket. Affinity grid fields were generated using the auxiliary program AutoGrid 4. Structure of compound 1 was built and minimized with the Accelrys Discovery Studio 3.0 software package (Accelrys Inc., San Diego, CA, USA), with flexible torsions assigned, and all dihedral angles were allowed to freely rotate. The Lamarckian genetic algorithm was used to determine the appropriate binding positions, orientations, and conformations of ligands. The optimized parameters were as follows: the maximum number of energy evaluations was increased to 25 000 000 per run, the iterations of Solis &Wets local search were 3000, the number of individuals in population was 300, and the number of generations was 100. Results differing by