The Herbal Drug Melampyrum pratense L. (Koch): Isolation and Identification of Its Bioactive Compounds Targeting Mediators of Inflammation

Melampyrum pratense L. (Koch) is used in traditional Austrian medicine for the treatment of different inflammation-related conditions. In this work, we show that the extracts of M. pratense stimulated peroxisome proliferator-activated receptors- (PPARs-)α and -γ that are well recognized for their anti-inflammatory activities. Furthermore, the extract inhibited the activation of the proinflammatory transcription factor NF-κB and induction of its target genes interleukin-8 (IL-8) and E-selectin in vitro. Bioassay-guided fractionation identified several active flavonoids and iridoids including melampyroside and mussaenoside and the phenolic compound lunularin that were identified in this species for the first time. The flavonoids apigenin and luteolin were distinguished as the main components accountable for the anti-inflammatory properties. Apigenin and luteolin effectively inhibited tumor necrosis factor α (TNF-α)-induced NF-κB-mediated transactivation of a luciferase reporter gene. Furthermore, the two compounds dose-dependently reduced IL-8 and E-selectin protein expression after stimulation with lipopolysaccharide (LPS) or TNF-α in endothelial cells (ECs). The iridoids melampyroside and mussaenoside prevented the elevation of E-selectin in LPS-stimulated ECs. Lunularin was found to reduce the protein levels of the proinflammatory mediators E-selectin and IL-8 in ECs in response to LPS. These data validate the ethnomedical use of M. pratense for the treatment of inflammatory conditions and point to the constituents accountable for its anti-inflammatory activity.


Introduction
Melampyrum pratense L. (Koch) or common cow-wheat belongs to the family of Orobanchaceae and is a herbaceous flowering plant. In Austrian ethnomedicine, the dried herbal drug has been used externally in a pillow or as tea to treat gout and rheumatism [1]. In Romanian traditional medicine, Melampyrum species are used to treat rheumatic disorders and skin infections [2]. In England, the plant was recommended as decoction for kidney ailments [3]. Altogether, the applications of this plant in traditional medicine strongly suggest anti-inflammatory activity.
This work, therefore, aims at validating the anti-inflammatory activity of M. pratense and identifying the compounds responsible for this effect by means of bioassay-guided fractionation. To this end, we applied a panel of functional and target-oriented cell models in order to identify constituents of M. pratense that abolish TNF-or LPS-induced expression of proinflammatory adhesion molecules (E-selectin) or chemokines (IL-8), inhibit the nuclear factor -light-chain 2 Evidence-Based Complementary and Alternative Medicine enhancer of activated B cells (NF-B) pathway, or activate peroxisome-proliferator-activated receptors (PPARs), thus interfering with different steps of the complex inflammatory process.
Although there are various molecular mediators involved in the development of the inflammatory response, many proinflammatory pathways converge at the point of the activation of NF-B [4][5][6]. It regulates the expression of an array of inflammatory response genes such as cytokines (e.g., TNF-, IL-1 , IL-6/10/12, and IFN-), chemokines (e.g., monocyte chemotactic protein-1 (MCP-1), IL-8), enzymes producing inflammatory mediators (e.g., COX-2, 5-lipoxygenase, and inducible nitric oxide synthase), adhesion molecules (e.g., E-selectin, ICAM-1), and enzymes degrading extracellular matrix (e.g., matrix metallopeptidase 9). Hence, targeting different steps of the NF-B signaling cascade is a promising approach to combat different inflammatory conditions [7]. E-selectin belongs to the endothelial adhesion molecules and plays an important role in recruiting leukocytes to the site of inflammation as it recognizes and binds to sialylated carbohydrates present on their surface proteins. Interleukin-8 (IL-8), a prototype of CXC chemokines, mainly acts as a neutrophil chemoattractant; its production is induced by TNF-, LPS, and IL-1 [8,9]. Another ubiquitously expressed class of transcription factors that are relevant regulators of the inflammatory response is the one of the nuclear receptors [10]. Recently, much attention has been drawn to the PPARs, which control the expression of inflammatory and metabolic genes in macrophages and lymphocytes [11][12][13]. PPARs are ligand-activated transcription factors exerting important functions in lipid and glucose metabolism but also regulating inflammation. PPAR , which is highly expressed in the vascular wall, skeletal muscle, heart, liver, and kidney, is mainly involved in the regulation of lipid catabolism [14]. Its activation was also shown to inhibit proliferation [15,16] and to modulate the inflammatory response of vascular smooth muscle cells (VSMCs) [17]. The activation of the other PPAR subtype studied in this work, PPAR , increases insulin sensitivity and contributes to the regulation of genes involved in inflammation, hypertension, and atherosclerosis [10,[18][19][20].
The chemical composition of M. pratense is very scarcely described in the literature. The aucubin content was determined as 5.11% [21]. 23 phenolic compounds, including 17 flavonoids and 6 phenol carboxylic acids, have been identified by two-dimensional paper chromatography [22,23]. Suomi et al. (2001) quantified iridoid glycosides in M. pratense using optimal on-line combination of partial filling micellar electrokinetic chromatography and electrospray ionization mass spectrometry (ESI-MS) and found aucubin at a concentration of ca. 100 mg/L and traces of catalpol [24].
In this work, we performed detailed chemical analysis of M. pratense and tested a number of plant constituents in cell-based in vitro assays for PPAR and PPAR activation, NF-B inhibition, and downregulation of TNF--or LPSstimulated expression of E-selectin and IL-8. Our data clearly demonstrate in vitro anti-inflammatory activity of M. pratense extracts and their individual components.
For bioactivity evaluation, all indicated compounds or dried extracts were reconstituted in dimethyl sulfoxide (DMSO), aliquoted, and stored at −20 ∘ C until use.

Plant
Material. The aerial parts of Melampyrum pratense were field collected at Neustift am Walde, Vienna, Austria. A voucher specimen (MP16072008) is deposited at the Department of Pharmacognosy, University of Vienna, Austria.

Extraction and Fractionation
. 183 g of dried and powdered plant material was first processed with the nonpolar solvent DCM yielding 3.7 g (2.0% yield) extract and subsequently with the polar solvent MeOH yielding 28.4 g (15.5% yield) using an accelerated solvent extractor ASE200 (Thermo Scientific Austria GmbH, Vienna, Austria). The instrument was equipped with 22 mL stainless steel extraction cells and 60 mL glass collection bottles. The extraction was performed at 40 ∘ C and 150 bar using 3 extraction cycles, 5 min heat-up time, 2 min static time, 10% flush volume, and 60 sec nitrogen purge.
Subsequently, tannins and chlorophylls were excluded from polar and nonpolar extracts, respectively, in order to avoid possible interferences with the assays and to increase the relative amount of the active compounds. The tannins were removed using liquid-liquid partitions between CHCl 3 and mixtures of MeOH/H 2 O [25], leading to a loss of 95% of the original MeOH extract yielding 1.4 g (4.9% of the MeOH extract, 0.7% of the crude drug). Chlorophyll was removed from the DCM extract by liquid-liquid-partition between DCM and MeOH : H 2 O (1 : 1) yielding 647.5 mg (17.5% of the DCM extract, 0.3% of the crude drug). In detail, 1 g extract was dissolved in 150 mL DCM, 150 mL MeOH : H 2 O (1 : 1) was added, and DCM was evaporated under reduced pressure (800-600 mbar) at 40 ∘ C. Consequently, insoluble chlorophyll precipitated in the methanol-water phase and could be filtered.

HPLC
Analysis. An HPLC instrument from Shimadzu (Kyoto, Japan) equipped with a CBM-20A system controller, a DGU-20A5 membrane degasser, an LC-20AD solvent delivery unit, an SIL-20AC HT autosampler, a CTO-20AC column oven, an SPD-M20A photodiode array detector, and a lowtemperature evaporative light scattering detector (ELSD-LT) was used for HPLC-DAD/ELSD experiments. Data analysis was conducted with the chromatography software LCsolution Ver.1.2 (Shimadzu). Chromatographic separation was performed on an Acclaim 120 C 18 reversed-phase column (150 mm × 2.1 mm i.d., 3 m) equipped with an Acclaim 120 C 18 guard column (10 mm × 2.1 mm i.d., 5 m) from Dionex (Germering, Germany), at 35 ∘ C and a flow rate of 0.4 mL/min. Water modified with formic acid (pH = 2.6) and MeCN were used as mobile phase A and B, respectively. For the methanolic extract, gradient elution was performed as follows: start isocratic at 2% of B for 5 min, followed by 2%-10% of B in 11 min, 10%-20% of B in 24 min, 20%-50% of B in 20 min, and 50%-95% of B in 5 min. For all other extracts, the following gradient was used: start isocratic at 2% of B for 5 min, followed by 2%-15% of B in 5 min, 15%-21% of B in 3 min, 21%-35% of B in 27 min, and 35%-95% of B in 10 min. The injection volume was 5 L for all samples. The DAD collected data from 190 to 400 nm.

LC-MS Parameters.
The tentative identification of the main flavonoids and iridoid glycosides was facilitated by HPLC-DAD-MS. These analyses were performed on an UltiMate 3000 RSLC-series system (Dionex) coupled to an HCT 3D quadrupole ion trap mass spectrometer equipped with an orthogonal ESI source (Bruker Daltonics, Bremen, Germany). HPLC separation was carried out as described above. The eluent flow was split roughly 1 : 8 before the ESI ion source, which was operated as follows: capillary voltage: +3.5/−3.7 kV, nebulizer: 26 psi (N 2 ), dry gas flow: 9 L/min (N 2 ), and dry temperature: 340 ∘ C. The mass spectrometer was operated in an automated data-dependent acquisition (DDA) mode to obtain MS 2 , MS 3 , and MS 4 spectra (collision gas: He, isolation window: 4 Th, and fragmentation amplitude: 1.0 V).
The identity of the main compounds in the extracts was confirmed either by the comparison of the retention times, UV-and MS -spectra, with reference compounds (apigenin, luteolin, chrysoeriol, diosmetin, aucubin, and catalpol), or by isolation and structural characterization by 1D and 2D NMR and MS (melampyroside, mussaenoside, and lunularin).

NF-B Transactivation
Assay. The transactivation of a NF-B-driven luciferase reporter gene was quantified in HEK293/NF-B-luc cells (Panomics, RC0014) as previously described [29]. The cells were maintained at 37 ∘ C and 5% CO 2 cell culture incubators in Dulbecco's modified Eagle's medium (DMEM; Lonza, Basel, Switzerland) supplemented with 2 mM glutamine, 100 g/mL hygromycin B, 100 U/mL benzylpenicillin, 100 g/mL streptomycin, and 10% fetal bovine serum (FBS). On the day before the experiment, the cells were stained by incubation for 1 h in serum-free medium supplemented with 2 M Cell Tracker Green CMFDA (C2925; Invitrogen). The cells were, then, reseeded in 96-well plates at a density of 4 × 10 4 cells/well in phenol redfree and FBS-free DMEM overnight. Cells were pretreated as indicated for 30 min prior to stimulation with 2 ng/mL TNFfor 4 h. The final concentration of DMSO in the experiments was 0.1% or lower. An equal amount of DMSO was always tested in each experiment to assure that the solvent vehicle does not influence the results. After cell lysis, the luminescence of the firefly luciferase and the fluorescence of the Cell Tracker Green CMFDA were quantified on a Genios Pro plate reader (Tecan, Grödig, Austria). The luciferasederived signal from the NF-B reporter was normalized by the Cell Tracker Green CMFDA-derived fluorescence to account for differences in the cell number. The known NF-B inhibitor parthenolide (Sigma-Aldrich, Vienna, Austria) was used as a positive control.

PPAR Luciferase Reporter Gene Assay.
To evaluate PPAR or PPAR activation, transient transfection of HEK-293 cells with the respective PPAR expression plasmid and a luciferase reporter plasmid containing PPAR response element (PPRE) was used [30]. HEK-293 cells (ATCC, USA) were cultured in DMEM with phenol red, supplemented with 100 U/mL benzylpenicillin, 100 g/mL streptomycin, 2 mM glutamine, and 10% FBS. The cells were seeded in 10 cm dishes at a density of 6 × 10 6 cells/dish, incubated for 18 h, and transfected by the calcium phosphate precipitation method with 4 g of the reporter plasmid (tk-PPREx3-luc), 4 g PPAR or PPAR receptor expression plasmid, and 2 g green fluorescent protein plasmid (pEGFP-N1, Clontech, CA, USA) as internal control. After 6 h, the transfected cells were harvested and reseeded (5 × 10 4 cells/well) in 96-well plates containing DMEM without phenol red, supplemented with 100 U/mL benzylpenicillin, 100 g/mL streptomycin, 2 mM glutamine, and 5% charcoal-stripped FBS. The cells were further treated with the indicated extracts or compounds or the solvent vehicle and incubated for 18 h. The medium was, then, discarded, the cells were lysed, and luciferase activity and fluorescence were measured on a Genios Pro plate reader (Tecan, Grödig, Austria). The luminescence signals obtained from the luciferase activity measurements were normalized to the EGFP-derived fluorescence, to account for differences in the transfection efficiency or cell number. The specific PPAR agonist GW7647 (Cayman, Missouri, USA) and the PPAR agonist pioglitazone (Molekula Ltd., Shaftesbury, UK) were used to verify the subtype specificity of the performed measurements.

Statistical Analysis.
Statistical analyses were performed using Prism Software (ver. 4.03; GraphPad Software Inc., San Diego, CA). For NF-B transactivation and PPAR luciferase reporter gene assay, data were normalized to DMSO treated control the mean value of which was set as 1.0, and for E-selectin and IL-8 expression, the data were analyzed as Evidence-Based Complementary and Alternative Medicine 5 percent inhibition in comparison to the respective LPS-or TNF--treatment controls. The experimental data are presented as means ± standard error (SE). Statistical significance was determined by ANOVA using Bonferroni post hoc test.

Results and Discussion
The upregulation of E-selectin and IL-8 production in HUVECtert cells was induced by two diverse stimuli, the proinflammatory cytokine TNF-or the bacterial product LPS. TNF-and LPS activate distinct but partially overlapping signaling pathways implemented in acute and chronic inflammation. They were used as chemically distinct inflammatory agonists interacting with different receptors to enhance the reliance of the screen. Since, next to endothelial cells, the LPS receptor, toll-like receptor 4, and the tumor necrosis factor receptor play a key role in the activation of other cell types, for example, leukocytes, results of the screen are likely to be relevant to other cell types as well.
The investigations of extracts of Melampyrum pratense herb with respect to downregulation of LPS-or TNF-induced IL-8 and E-selectin expression, PPAR activation, or NF-B inhibition showed equal or even better results for chlorophyll-depleted and detannified extracts compared to crude DCM and MeOH extracts (Table 1). Further fractionation of these extracts was performed using solid phase extraction.
The SPE subfractions were further tested for their antiinflammatory activity in vitro. The strongest downregulation of IL-8 and E-selectin induced by LPS or TNF-was found in the 70% SPE fraction of the chlorophyll-free DCM extract as well as in the 30% and 70% SPE subfraction of the detannified MeOH extract (Table 2). Investigations regarding PPAR activation and NF-B inhibition showed highest activities in the 70% SPE subfractions ( Table 2). The high activity of the 100% SPE fractions of the chlorophyll-free DCM extract could be explained by the presence of fatty acids in these fractions that were identified by TLC and GC-MS (data not shown) using reference compounds [31].
HPLC analysis revealed that besides mussaenoside 4, and melampyroside 5, the crude methanolic extract contained several other iridoid glycosides, identified as catalpol 1 (shoulder of 2nd peak), aucubin 2, and mussaenosidic/loganic acid 3 (see Figure 1). The identification of these compounds was done with commercial reference compounds via the comparison of the UV-spectra and retention time as well as by LC-MS analysis.
On the other hand, the detannified MeOH extract mainly contained flavonoids of which luteolin 6, apigenin 7, chrysoeriol 8, and traces of diosmetin could be identified (Figure 2) by comparison of the UV-spectra and retention time with commercial reference standards as well as a comparison with an MS flavonoid database [32].
The chlorophyll-free DCM extract mainly contained mussaenoside, next to lunularin 9 and melampyroside ( Figure 3). Mussaenoside and melampyroside were enriched  in the 30% SPE subfraction. The main component of the 70% SPE subfraction was lunularin. The identification was validated via the comparison of retention times as well as UV-spectra with those of the isolated purified reference compounds (structure confirmed by NMR and LC-MS).
The main compounds from M. pratense were tested together with their initial extract for PPAR activation, NF-B inhibition, and the downregulation of LPS-or TNF-induced production of IL-8 and E-selectin. The iridoid glycosides aucubin, catalpol, melampyroside, and mussaenoside reduced E-selectin expression in response to LPS stimulation ( Figure 4). The effect of iridoidglycosides was specific for Eselectin as they did not downregulate IL-8 ( Figure 5) nor 6 Evidence-Based Complementary and Alternative Medicine    activate PPARs (data not shown) or inhibit NF-B ( Figure 6). The iridoids catalpol and aucubin can be found in many plant families and have already been described for M. pratense and other Melampyrum species [24,33]. Jeong et al. (2002) showed that aucubin inhibited the expression of TNF-and IL-6 in Ag-stimulated rat basophilic leukemia-(RBL-) 2H3 mast cells in a dose-dependent manner with an IC 50 of 0.10 and 0.19 g/mL, respectively. In that study, aucubin was also determined to inhibit antigen-induced nuclear translocation of the p65 subunit of NF-B. Since activator protein-1 binding activity was not affected, its results suggested that aucubin is a specific inhibitor of NF-B activation in mast cells, which might explain its beneficial effect in the treatment of chronic inflammatory diseases [34]. However, in our experimental setup, aucubin did not inhibit NF-B at 30 M.
The flavonoids apigenin and luteolin, on the other hand, showed strong inhibition of NF-B ( Figure 6) as well as a strong downregulation of both IL-8 and E-selectin proteins (Figures 4 and 5). Since luteolin and apigenin displayed the highest activity on all anti-inflammatory targets, doseresponse studies were performed. Luteolin downregulated Eselectin with an IC 50 Figures 4 and 5). A number of previous studies have also demonstrated the NF-B inhibitory action of these two flavonoids utilizing a number of different methods and cell models [35][36][37][38][39][40][41]. Depending on little changes in the flavone backbone, flavonoids can play a modulating, biphasic, and regulatory action on immunity and inflammation [42]. Studies regarding the effect of flavonols (kaempferol, quercetin, and myricetin) and flavones (flavone, chrysin, apigenin, luteolin, baicalein, and baicalin) on TNF--stimulated ICAM-1 expression, which is like E-selectin a proinflammatory adhesion molecule, revealed kaempferol, chrysin, apigenin, and luteolin as active inhibitors of ICAM-1 expression. Additionally, apigenin and luteolin were shown to inhibit the NF-B signaling at the level of I B kinase (IKK) activity, with consequent effect on I B degradation, NF-B binding to the DNA, and NF-B transactivation activity [35]. Several flavonoids were investigated for the inhibitory effect on LPS-induced TNF-production from macrophages. Positive results were provided for the flavones luteolin, apigenin, and chrysin, the flavonols quercetin and myricetin, the flavanonol taxifolin, and the anthocyanidin cyanidin chloride in vitro. Nevertheless, serum TNF-production in vivo was inhibited only by luteolin or apigenin, and only luteolin or quercetin inhibited 12-O-tetradecanoylphorbol-13-acetate (TPA) induced ear edema. These results suggest that the structure of luteolin is the most suitable for the oral antiinflammatory activity and that the presence of absence of a hydroxy group may cause a loss of efficiency [43]. Moreover, luteolin was shown to display excellent radical scavenging and cytoprotective properties, especially when tested in complex biological systems where it can interact with other antioxidants like vitamins. Nevertheless, its antiinflammatory effects at micromolar concentrations can only be partly explained by its antioxidant capacities. They include the activation of antioxidative enzymes, suppression of the NF-B pathway, and inhibition of proinflammatory substances. The glycosidic form of luteolin often present in plants is cleaved in vivo, and the aglycones are conjugated and metabolized after nutritional uptake which has to be considered when evaluating in vitro studies [44]. Besides the inhibition of the induction of inflammatory cytokines such as TNF-, IL-8, IL-6 and granulocyte-macrophage colonystimulating factor (GM-CSF), luteolin was found to attenuate cyclooxygenase (COX)-2 expression and rise in intracellular Ca 2+ levels [45]. In our study, the stilbene-like compound lunularin only had an inhibitory effect on IL-8 and E-selectin after stimulation with LPS but not with TNF-. This compound is mainly known from liverworts [46,47]. Besides, in higher plants, it was found in celery, Apium graveolens [48], and Morus species [49]. Activity studies have only been conducted concerning antibacterial, antifungal, and antioxidant characteristics [50], but the anti-inflammatory effects of lunularin have not yet been reported. Thus, by showing its ability to downregulate E-selectin and IL-8 in endothelial cells stimulated with LPS, the present study gives first insight into a possible anti-inflammatory action of this compound.