Identification of bioactive compounds from flowers of black elder (sambucus nigra l.) that activate the human peroxisome proliferatoractivated receptor (ppar)

PHYTOTHERAPY RESEARCH
Phytother. Res. 24: S129–S132 (2010)
Published online 11 March 2010 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/ptr.3005
Identifi cation of Bioactive Compounds from
Flowers of Black Elder (Sambucus nigra L.) that
Activate the Human Peroxisome Proliferator-
activated Receptor (PPAR) γ
Kathrine B. Christensen1,2*, Rasmus K. Petersen3, Karsten Kristiansen4 and
Lars P. Christensen2
1Department of Food Science, Aarhus University, Kirstinebjergvej 10, 5792 Aarslev, Denmark
2Institute of Chemical Engineering, Biotechnology and Environmental Technology, University of Southern Denmark, Niels Bohrs
Allé 1, 5230 Odense M, Denmark
3BioLigands ApS, International Science Park Odense, 5230 Odense M, Denmark
4Department of Biology, University of Copenhagen, Ole Maaløes Vej 5, 2200 Copenhagen N, Denmark
Obesity is one of the predisposing factors for the development of overt Type 2 diabetes (T2D). T2D is caused
by a combination of insulin resistance and β-cell failure and can be treated with insulin sensitizing drugs that
target the nuclear receptor peroxisome proliferator-activated receptor (PPAR) γ. Extracts of elderfl owers (Sam-
bucus nigra) have been found to activate PPARγ and to stimulate insulin-dependent glucose uptake suggesting
that they have a potential use in the prevention and/or treatment of insulin resistance. Bioassay-guided chro-
matographic fractionation of a methanol extract of elderfl owers resulted in the identifi cation of two well-known
PPARγ agonists; α-linolenic acid and linoleic acid as well as the fl avanone naringenin. Naringenin was found to
activate PPARγ without stimulating adipocyte differentiation. However, the bioactivities of these three metabo-
lites were not able to fully account for the observed PPARγ activation of the crude elderfl ower extracts and
further studies are needed to determine whether this is due to synergistic effects and/or other ligand-independent
mechanisms. Elderfl ower metabolites such as quercetin-3-O-rutinoside, quercetin-3-O-glucoside, kaempferol-3-
O-rutinoside, isorhamnetin-3-O-rutinoside, isorhamnetin-3-O-glucoside, and 5-O-caffeoylquinic acid were
unable to activate PPARγ. These fi ndings suggest that fl avonoid glycosides cannot activate PPARγ, whereas
some of their aglycones are potential agonists of PPARγ. Copyright 2010 John Wiley & Sons, Ltd.
Keywords: Sambucus nigra; type 2 diabetes; peroxisome proliferator-activated receptor (PPAR) γ; adipocyte differentiation; naringenin; fatty acids.
release of free fatty acids that all promote a less insulin INTRODUCTION
resistant state (Tripathi and Srivastava, 2006).
Plants contain many health-promoting compounds Overt type 2 diabetes (T2D) is characterized by hyper- and more than a thousand different species have been glycaemia caused by a combination of insulin resistance used in traditional medicine for the treatment of diabe- and pancreatic β-cell failure. Together with dyslipidae- tes (Marles and Farnsworth, 1995). Besides the overall mia, hypertension, and visceral obesity, these clinical benefi cial effects of these species due to a high content traits are the core characteristics of the metabolic syn- of fi ber, vitamins, and minerals, several plant metabo- drome. WHO has estimated that 20–30% of the adult lites also have specifi c effects on many mechanisms population in all countries meet the criteria of the related to insulin secretion and gluconeogenesis that are metabolic syndrome and this is primarily due to excess important in the treatment of diabetes and hence, plants energy intake and a minimum of physical activity. are a likely source of natural products with potential Therapeutic strategies for insulin resistance and T2D include lifestyle interventions as well as administration Preparations of black elder (Sambucus nigra L.) are of antidiabetic drugs. The thiazolidinediones (TZDs), a used in traditional medicine as diuretics and to treat class of insulin-sensitizing drugs, target the peroxisome colds, infl uenza, infl ammation, and diabetes (Swanston- proliferator-activated receptor (PPAR) γ. Activation of Flatt et al., 1991; Kültür, 2007). A study on aqueous the PPARγ promotes a multitude of processes such as extracts of elderfl owers showed that they exhibit insu- re-distribution of fat from visceral to subcutaneous fat lin-like and insulin-releasing actions in vitro (Gray et al., tissue, increased release of adiponectin, and decreased 2000). The bioactive metabolites were not identifi ed and major metabolites such as quercetin-3-O-rutino- * Correspondence to: Kathrine B. Christensen. University of Southern side, lupeol, and β-sitosterol did not individually stimu- Denmark, Faculty of Engineering, Institute of Chemical Engineering, Bio- late insulin secretion. Extracts of elderfl owers have also technology and Environmental Technology, Niels Bohrs Allé 1, 5230 recently been shown to activate PPARα, δ, and γ without stimulating adipocyte differentiation. In addition, elderfl ower extracts had a benefi cial effect Grant sponsor: EU Interreg IIIA and the Danish Council for Strategic Research (project No. 2101-01-0065).
on insulin-stimulated glucose uptake suggesting that Copyright 2010 John Wiley & Sons, Ltd. elderfl owers produce compounds with bioactivities which the active fractions I-3 (63.5 mg) and J-3 (30.4 mg) similar to those of partial PPARγ agonists (Christensen et al., 2009). Partial PPARγ agonists are believed not to promote the same magnitude of undesirable side effects Preparation of samples for bioassays. All fractions, iso-
as full agonists such as the thiazolidinediones (TZDs). lated compounds, and standards were dissolved in Hence, they might be used as an alternative to the tra- dimethyl sulfoxide (DMSO) before they were tested for ditional antidiabetic drugs used in the prevention and/ bioactivity in the bioassays. Standards of naringenin, quercetin-3-O-rutinoside, quercetin-3-O-glucoside, Most studies on the health-promoting effects of elder kaempferol, kaempferol-3-O-rutinoside, isorhamnetin- focus on the berries, although elderfl owers are also a 3-O-rutinoside, isorhamnetin-3-O-glucoside, 5-O- rich source of bioactive metabolites such as triterpe- caffeoylquinic acid, and caffeic acid were purchased noids, fl avonoids, and phenolic acids (Gray et al., 2000; from Sigma Aldrich Chemie (Steinheim, Germany) or Extrasynthese (Genay Cedex, France).
The aim of this study was to identify bioactive PPARγ transactivation bioassay.
compounds responsible for the activation of PPARγ PPARγ-mediated transactivation, a mouse embryonic broblast cell line was transiently transfected at fractionation. PPARγ transactivation and adipocyte differentiation bioassays were used to characterize (Christensen et al., 2009). For each well a total of 0.05 μg the bioactivity of fractions and isolated compounds DNA (0.0025 μg normalization vector pRL-CVM + from the fractionation of a crude elderfl ower methanol 0.03 μg of the Gal4-responsive Luciferase reporter + 0.015 μg pM-hPPARγ-LBD) were used. The media in the plates were changed 6 h after transfection to 200 μL Dulbecco’s modifi ed Eagle’s media (DMEM) with MATERIALS AND METHODS
streptomycin) containing either vehicle (0.1% DMSO), General experimental procedures. Reverse phase
positive control (1 μM rosiglitazone (Rosi), Novo (RP)-18 aluminium sheets (Merck, Darmstadt, Nordisk® A/S), or test sample dissolved in DMSO Germany) were used for TLC with 50% acetonitrile (1 : 103, 1 : 104, and 1 : 105 dilutions of the stocks were (MeCN) as eluent and UV light followed by visuali- used). After 18 h, the cells were washed with phos- zation with vanillin-sulfuric acid reagent were used phate-buffered saline (PBS) (200 μL/well) and lysed for inspection of all plates. Octadecyl-functionalized with lysis buffer (20 μL/well). Assay for Photinus and silica gel (RP-18, Sigma Aldrich Chemie, Steinheim, Renilla activities were measured directly in the plate Germany) was used for fl ash column chromatography using a Dual-GloTM Luciferase assay system (Promega). (CC). Semi-preparative high performance liquid All transient transfection experiments were done in chromatography (HPLC) was carried out using a triplicate and double determination for each triplicate Dionex P680 HPLC pump equipped with a Dionex was carried out. Photinus activities were normalized to UVD340U detector and a Develosil ODS-HG-5 RP-18 the corresponding Renilla activities to compensate for column (5 μm; 250 × 20 mm, Nomura Chemical Co., differences in transfection effi ciency. Positive control Seto, Japan) using a fl ow rate of 5 mL/min. Liquid chro- (Rosi) gave 100–150-fold activation compared to vehicle matography with photodiode array and mass spectro- metric detection (LC–PAD–MS) was performed with LC, PAD, and MS settings as previously described Adipocyte differentiation bioassay. 3T3-L1 cells were
(Christensen et al., 2008). Luciferase measurements grown in DMEM supplemented with 10% calf serum were performed on a LUMIstar BMG luminometer.
and antibiotics (62.5 μg/mL penicillin and 100 μg/mL streptomycin). The cells were seeded in 24-well Preparation of plant extract. Elderfl owers (cultivar
plates and grown to confl uence. At two days post- Haschberg) were obtained from Research Farm confl uence (designated day 0) the cells were induced Lindhof, Germany, in June 2007. Elderfl owers were to differentiate with DMEM supplemented with 10% frozen at –22 °C immediately after harvest. The frozen fetal bovine serum (FBS), antibiotics, vehicle or fl owers (3 kg) were homogenized and extracted twice with methanol (MeOH) (8 L) for 24 h in the dark at (DEX protocol) or 1 μM dexamethasone, 0.5 mM 5 °C with periodical shaking. The extracts were com- isobutylmethylxanthine and 1 μg/mL insulin (MDI pro- bined, fi ltered, and dried under vacuum.
tocol) (Christensen et al., 2009). After 48 h, the cells Fractionation of crude extract. A part of the elder-
were re-fed with DMEM supplemented with 10% FBS fl ower extract (15 g) was separated by fl ash CC (100 mm and the positive control, vehicle or the test sample i.d., 400 g RP-18 silica gel) using the following solvent (DEX protocol) or including 1 μg/mL insulin (MDI gradient: 100% H2O (500 mL), 10% MeCN in H2O protocol). From day 4, medium contained DMEM with (900 mL), 20–90% MeCN in H2O in 10% steps (1000 mL 10% FBS and antibiotics and were changed every each), 100% MeCN (1000 mL). One hundred and ten second day until day 8. All test samples were tested in fractions (100 mL each) were collected and combined three concentrations as indicated. All tests were per- into 13 fractions (A to M) based on TLC. The active formed in triplicate. At day 8, cells were washed in PBS fractions I (177 mg) and J (258 mg) were further sepa- and fi xed in 3.7% formaldehyde for 1 h. Cellular triac- rated by semi-preparative HPLC (solvent A = H2O with ylglycerols were stained with Oil-Red-O (0.5 g of Oil- 500 ppm TFA, solvent B = 100% MeCN; gradient 20% Red-O in 100 mL of isopropanol and diluted 6 : 4 with B at 0 min; 70% B at 60 min; 100% B at 75 min; 20% water) for 1 h. After staining, plates were washed twice B at 85 min) yielding four fractions each (I/J-1 to 4) of Copyright 2010 John Wiley & Sons, Ltd. Phytother. Res. 24: S129–S132 (2010)
POTENTIAL ANTIDIABETIC PROPERTIES OF ELDERFLOWERS was then tested for activation of PPARγ in several dif- ferent concentrations. Weak but signifi cant and dose-dependent activation of 3.0, 3.6, and 6.0 fold were Major metabolites in the crude MeOH extract of observed at concentrations of 40, 100, and 200 μM, quercetin-3-O-rutinoside, kaempferol-3-O-rutinoside, Naringenin and the two isolated fatty acids from frac- isorhamnetin-3-O-rutinoside, 5-O-caffeoylquinic acid, tions K and L did not stimulate adipocyte differentia- 3,5-di-O-caffeoylquinic acid, 5-O-p-coumaroylquinic tion as determined by the adipocyte differentiation acid, naringenin, α-linolenic, and linoleic acid In addition to the metabolites isolated by bioassay- Initial separation of the elderfl ower MeOH extract by guided fractionation, a number of known elderfl ower RP fl ash CC resulted in 13 fractions (A to M). Fractions metabolites and related compounds were tested for activ- B and C contained in addition to the phenolic acids ation of PPARγ. These were 5-O-caffeoylquinic acid, mentioned above also 3-O-caffeoylquinic acid, 3-O-p- caffeic acid, quercetin-3-O-rutinoside, quercetin-3-O- coumaroylquinic acid and 1,5-di-O-caffeoylquinic acid glucoside, kaempferol-3-O-rutinoside, isorhamnetin-3- as minor constituents. Fractions D to H contained in O-rutinoside, isorhamnetin-3-O-glucoside, quercetin, addition to the fl avonol glycosides mentioned above and kaempferol. None of these compounds, except also small amounts of quercetin-3-O-glucoside and isor- kaempferol, was able to activate PPARγ signifi cantly in hamnetin-3-O-glucoside. The polyphenols in fractions concentrations of 50–100 μM. Kaempferol at 100 μM B to H were identifi ed by LC–PAD–MS (Christensen was able to elicit a 10-fold activation of PPARγ.
et al., 2008). Fractions I and J contained fl avonoid derivatives but primarily the fl avanone naringenin and were further separated by semi-preparative HPLC to give four fractions each. Fractions I-3 and J-3 were pure naringenin. Fractions K and L contained α-linolenic and linoleic acid, respectively, as shown by LC–PAD–MS and were not further separated.
All obtained fractions and pure compounds were dis- solved in DMSO and tested for activation of PPARγ in several different concentrations. Testing fractions A to M in the PPARγ transactivation bioassay gave four positive results as fractions I, J, K, and L were able to activate PPARγ (Fig. 1). Fractions I and J at 50 μg/mL signifi cantly activated PPARγ by 14- and 19-fold, respectively. Fractions K and L at 20 μg/mL signifi - cantly activated PPARγ by 10- and 22-fold, respectively. Figure 2. Activation of PPARγ by naringenin tested at three dif-
Further testing of the fractions I/J-1 to 4 showed that ferent concentrations; 40, 100, and 200 μM. Bioactivity is given fractions I/J-3 were able to activate PPARγ as well (data as fold activation relative to DMSO (set to 1.0). Rosiglitazone not shown). Fractions I/J-3 contained naringenin, which (Rosi) was used as positive control.
Figure 1. Activation of PPARγ by fractions I, J, K, and L of the MeOH extract of elderfl owers. Each column represents one experiment
performed in triplicate. Bioactivity is given as fold activation relative to DMSO (set to 1.0). Rosiglitazone (Rosi) was used as positive
control.
Copyright 2010 John Wiley & Sons, Ltd. Phytother. Res. 24: S129–S132 (2010)
MeOH extract and the resulting fractions. Hence, syn- DISCUSSION
ergistic effects and/or other ligand-independent mecha-nisms may underlie the observed pronounced activation LC–PAD–MS analysis of the crude MeOH elderfl ower of PPARγ as well as the positive effect on insulin- extract used in this work confi rmed the presence of stimulated glucose uptake by elderfl ower extracts most of the major metabolites previously reported for extracts of elderfl owers (Christensen et al., 2008). Several compounds with structures similar to that of However, this is the fi rst report on the presence of naringenin have been identifi ed as activators and/or agonists of PPARγ. These include among others the Bioassay-guided fractionation of the elderfl ower isofl avones genistein and daidzein primarily found in extract gave four bioactive fractions (I, J, K, and L) able soybeans and soy-products (Dang et al., 2003; Mezei et to activate PPARγ. Fractions K and L contained pri- al., 2003) as well as widespread fl avone and fl avonol marily α-linolenic and linoleic acid, respectively. These aglycones (Huang et al., 2005; Liu et al., 2008). Clearly, fatty acids are commonly found in plants and are well many fl avone, fl avanone, fl avonol, and isofl avone agly- known activators of PPARs, including PPARγ (Chou cones have marked infl uence on PPARγ indicating that the size and the shape of the fl avonoid aglycones make From fractions I and J, the fl avanone naringenin was them suitable ligands for PPARγ in contrast to their found to be the bioactive constituent able to signifi - respective glycosides. This was illustrated by the testing cantly activate PPARγ, although at rather high concen- of standards of major elderfl ower metabolites for trations. A few investigations link naringenin and PPARγ activity in this study. Neither the fl avonol gly- PPARγ activation but published results are contradic- cosides nor the phenolic acids tested here, showed any tory. Liang et al. (2001) found naringenin to be unable effect on PPARγ confi rming the results from the bioas- to activate PPARγ, whereas Liu et al. (2008) showed say-guided fractionation. Only the aglycone kaempferol that naringenin is able to up-regulate transcription of was able to signifi cantly activate PPARγ, which is in adiponectin and induce dose-dependent expression of accordance with previous results (Liang et al., 2001). PPARγ-controlled luciferase reporter in a concentra- The present results therefore clearly shows that fl avonoid glycosides are unable to activate PPARγ in dependent activation of PPARγ by naringenin contrast to most of their aglycones as demonstrated corresponds with our fi ndings although different cell both here and in other studies for, e.g., kaempferol, lines were used. The fact that Liang et al. (2001) found chrysin, apigenin, and naringenin (Huang et al., 2005; naringenin to be inactive as a PPARγ activator could be due to the application of a different bioassay set-up using RAW264.7 cells as ligand-activation of PPARs is both promoter and cell-type specifi c to a certain degree.
Acknowledgments
In conclusion, the degree of activation of PPARγ by the isolated compounds naringenin, α-linolenic, and Support from the EU-Interreg III A programme in the Fyn−K.E.R.N Region and the Danish Council for Strategic Research (Project No. linoleic acid was insuffi cient to explain the effi cient 2101-01-0065) is gratefully acknowledged. The authors are grateful to activation observed for both the crude elderfl ower Dr Mirja Kaemper for supplying the plant material.
REFERENCES
Chou Y-C, Prakash E, Huang C-F, Lien T-W, Chen X, Su I-J, Chao tors for therapy of metabolic syndrome. Basic Clin Pharma- Y-S, Hsieh H-P, Hsu JT-A. 2008. Bioassay-guided purifi ca- col Toxicol 96: 3–14.
tion and identifi cation of PPARα/γ agonists from Chlorella Kültür S. 2007. Medicinal plants used in Kırklareli Province sorokiniana. Phytother Res 22: 605–613.
(Turkey). J Ethnopharmacol 111: 341–364.
Christensen KB, Minet A, Svenstrup H, Grevsen K, Zhang H, Liang Y-C, Tsai S-H, Tsai D-C, Lin-Shiau S-Y, Lin J-K. 2001. Sup- Schrader E, Rimbach G, Wein S, Wolffram S, Kristiansen K, pression of inducible cyclooxygenase and nitric oxide syn- Christensen LP. 2009. Identifi cation of plant extracts with thase through activation of peroxisome proliferator-activated potential antidiabetic properties: Effect on human peroxi- receptor-γ by fl avonoids in mouse macrophages. FEBS Lett some proliferator-activated receptor (PPAR), adipocyte 496: 12−18.
differentiation, and insulin-stimulated glucose uptake. Liu L, Shan S, Zhang K, Ning Z-Q, Lu, X-P, Cheng Y-Y. 2008. Phytother Res 23: 1316–1325.
Naringenin and hesperetin, two fl avonoids derived from Christensen LP, Kaack K, Fretté XC. 2008. Selection of elderberry Citrus aurantium up-regulate transcription of adiponectin. (Sambucus nigra L.) genotypes best suited for the prepara- Phytother Res 22: 1400–1403.
tion of elderfl ower extracts rich in fl avonoids and phenolic Marles RJ, Farnsworth NR. 1995. Antidiabetic plants and their acids. Eur Food Res Technol 227: 293–305.
active constituents. Phytomed 2: 137–189.
Dang ZC, Audinot V, Papapoulos SE, Boutin JA, Löwik CWGM. Mezei O, Banz WJ, Steger RW, Peluso MR, Winters TA, 2003. Peroxisome proliferator-activated receptor γ (PPARγ) Shay N. 2003. Soy isofl avones exert antidiabetic and as a molecular target for the soy phytoestrogen genistein. hypolipidemic effects through the PPAR pathways in J Biol Chem 278: 962–967.
obese Zucker rats and murine RAW 264.7 cells. J Nutr 133:
Gray AM, Abdel-Wahab YHA, Flatt PR. 2000. The traditional plant treatment, Sambucus nigra (elder), exhibits insulin- Swanston-Flatt SK, Flatt PR, Day C, Bailey CJ. 1991. Traditional like and insulin-releasing cctions in vitro. J Nutr 130: 15–20.
dietary adjuncts for the treatment of diabetes mellitus. Proc Huang TH-W, Kota BP, Razmovski V, Roufogalis BD. 2005. Nutr Soc 50: 641–651.
Herbal or natural medicines as modulators of peroxisome Tripathi BK, Srivastava, AK. 2006. Diabetes mellitus: Complica- proliferator-activated receptors and related nuclear recep- tions and therapeutics. Med Sci Monit 12: RA130–147.
Copyright 2010 John Wiley & Sons, Ltd. Phytother. Res. 24: S129–S132 (2010)

Source: http://sundhedsfremmeogplanter.dk/docs/file/ChristensenKB_et-al_2010_Black_elder-PPARgamma.pdf