Comparative metabolomics approach coupled with cell- and gene-based assays
for species classification and anti-inflammatory bioactivity validation
of Echinacea plants
Chia-Chung Houa
, Chun-Houh Chenb
, Ning-Sun Yanga
, Yi-Ping Chena
, Chiu-Ping Loa
, Sheng-Yang Wangc
,
Yin-Jing Tiend
, Pi-Wen Tsaia
, Lie-Fen Shyura,e,

a
Agricultural Biotechnology Research Center, Academia Sinica, Taipei 115, Taiwan, ROC
b
Institute of Statistical Sciences, Academia Sinica, Taipei 115, Taiwan, ROC
c
Department of Forestry, National Chung-Hsing University, Taichung 402, Taiwan, ROC
d
Institute of Statistics, National Central University, Taoyuan County 320, Taiwan, ROC
e
Department of Biochemical Science and Technology, National Taiwan University, Taipei 106, Taiwan, ROC
Received 26 March 2009; received in revised form 7 August 2009; accepted 26 August 2009
Abstract
Echinacea preparations were the top-selling herbal supplements or medicines in the past decade; however, there is still frequent misidentification or
substitution of the Echinacea plant species in the commercial Echinacea products with not well chemically defined compositions in a specific preparation. In this
report, a comparative metabolomics study, integrating supercritical fluid extraction, gas chromatography/mass spectrometry and data mining, demonstrates
that the three most used medicinal Echinacea species, Echinacea purpurea, E. pallida, and E. angustifolia, can be easily classified by the distribution and relative
content of metabolites. A mitogen-induced murine skin inflammation study suggested that alkamides were the active anti-inflammatory components present in
Echinacea plants. Mixed alkamides and the major component, dodeca-2E,4E,8Z,10Z(E)-tetraenoic acid isobutylamides (8/9), were then isolated from E. purpurea
root extracts for further bioactivity elucidation. In macrophages, the alkamides significantly inhibited cyclooxygenase 2 (COX-2) activity and the
lipopolysaccharide-induced expression of COX-2, inducible nitric oxide synthase and specific cytokines or chemokines [i.e., TNF-, interleukin (IL)-1, IL-6,
MCP-1, MIP-1] but elevated heme oxygenase-1 protein expression. Cichoric acid, however, exhibited little or no effect. The results of high-performance liquid
chromatography/electron spray ionization/mass spectrometry metabolite profiling of alkamides and phenolic compounds in E. purpurea roots showed that
specific phytocompound (i.e., alkamides, cichoric acid and rutin) contents were subject to change under certain post-harvest or abiotic treatment. This study
provides new insight in using the emerging metabolomics approach coupled with bioactivity assays for medicinal/nutritional plant species classification, quality
control and the identification of novel botanical agents for inflammatory disorders.
 2010 Elsevier Inc. All rights reserved.
Keywords: Echinacea; Anti-inflammation; Metabolomics; Alkamides
1. Introduction
Metabolomics is one of the more-recently introduced "-omics"
technologies, joining genomics, transcriptomics and proteomics as
tools in global systems biology [1,2]. Metabolomics refers to the highthroughput
analysis, either by a targeted or a global (unbiased)
strategy, of the metabolic state of a biological system based on the
simultaneous measurement of the distribution of small molecules in
the cellular biomass [3]. Metabolomics could provide the needed links
between the complex metabolite mixtures in traditional herbal
medicines and molecular pharmacology [4] and has great potential
in the development of active secondary metabolites from medicinal
plants as novel or improved phytotherapeutic agents [3].
Echinacea spp. (Asteraceae), endemic to North America, are widely
used as herbal medicines and food supplements in Europe and North
America. Previous studies have reported that Echinacea plants contain
a variety of components including caffeic acid derivatives, alkamides,
and polysaccharides [5], and the three Echinacea species, E. purpurea,
E. pallida, and E. angustifolia, have some common or different
phytochemical profiles in their roots, leaves or flowers [6­12].
Although a number of analytical methods for analyzing the chemical
constituents of Echinacea plants have been reported, there is still
frequent misidentification or substitution of the three species in the
commercial Echinacea products [13,14]. This study provides an
Available online at www.sciencedirect.com
Journal of Nutritional Biochemistry xx (2010) xxx­xxx

This work was supported by institutional grant of Academia Sinica and
national research program for genomic medicine (NSC 97-3112-B-001-020)
of National Science Council of Taiwan, R.O.C.
 Corresponding author. Agricultural Biotechnology Research Center,
Academia Sinica, Taipei 115, Taiwan, ROC. Tel.: +886 2 26515028; fax: +886
2 26515028.
E-mail address: lfshyur@ccvax.sinica.edu.tw (L.-F. Shyur).
0955-2863/$ - see front matter  2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jnutbio.2009.08.010
ARTICLE IN PRESS
efficient protocol to distinguish the three medicinal Echinacea species
by a metabolomics approach, using supercritical fluid extraction (SFE)
coupled with gas chromatography (GC)/mass spectrometry (MS) to
generate the metabolite profiles of the three medicinal Echinacea
species, grown under organic farming conditions, and then comparing
them by various approaches, including principal component analysis
(PCA) and generalized association plots (GAP).
Echinacea plant preparations are claimed to have several pharmacological
effects, such as immune enhancement, anti-inflammation,
and prevention of common cold [15]. Cyclooxygenase-2 (COX-2)
catalyzes the synthesis of prostaglandins from the substrate arachidonic
acid and has been also reported to be expressed within human
tumor neovasculature as well as in neoplastic cells in human prostate,
colon, breast, and lung cancer tissues [16,17]. Numerous reports have
demonstrated that Echinacea constituents have inflammation-related
bioactivities such as stimulation of cytokine production [18­21] or
inhibition of COX-1 or COX-2 activity in vitro [22­25]. Recently,
Echinacea alkamides were reported to be a new class of cannabinomimetics
modulating tumor necrosis factor  (TNF-) gene expression
via the cannabinoid type 2 (CB2) receptor [26,27]. Inducible nitric
oxide synthase (iNOS) is one of the key enzymes producing nitric
oxide (NO) and is expressed in response to a variety of proinflammatory
cytokines and inflammatory agents such as lipopolysaccharide
(LPS) at inflammatory sites [28]. Heme oxygenase-1 (HO-
1), the rate-limiting enzyme in the breakdown of heme into carbon
monoxide, iron and bilirubin, can be induced in various cell types such
as endothelial cells, monocytes/macrophages, neutrophils, and
fibroblasts by different stresses during the inflammatory response
[29]. Specific metabolites displayed by the metabolomics study
results in Echinacea plant extracts in this study were being further
investigated for their inflammation-modulating activity on these
important proinflammatory mediators, stress protein and/or cytokine/chemokine
profiles in murine macrophages or skin system.
2. Materials and methods
2.1. Plant materials
Echinacea purpurea, E. pallida, and E. angustifolia were grown in Nantou Country,
Taiwan, using organic farming conditions. Voucher specimens of the three Echinacea
species, namely, Epu-1, Epa-1, and Ean-1, have been deposited at the Agricultural
Biotechnology Research Center, Academia Sinica, Taipei, Taiwan. Echinacea plants at
the flowering stage were harvested and separated into two groups (aerial parts and
root parts) and then lyophilized.
2.2. Reagents and materials
3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), 12-O-tetradecanoylphorbol-13-acetate
(TPA), LPS, 1,4-dithiothreitol (DTT), indomethacin, and
aspirin were purchased from Sigma Chemical (St. Louis, MO, USA). Celecoxib
(Celebrex) was from Pharmacia (Northumberland, UK). Silica gel (230­400 mesh),
silica gel 60 F254 thin-layer chromatography (TLC) and RP-18 F254s TLC plates were
from Merck (Darmstadt, Germany), and RP-18 silica gel (75C18-OPN) was from
Cosmosil (Kyoto, Japan). All other chemicals and solvents were reagent or highperformance
liquid chromatography (HPLC) grade. Antibody sources are given below
in the appropriate section.
2.3. Supercritical fluid extraction
Supercritical fluid extraction of Echinacea root or aerial tissues were performed
using an ISCO SFE System 1220 (Lincoln, NE), with supercritical CO2 as extraction fluid.
One-gram samples of lyophilized tissue were extracted at 60°C for 30 min under a
pressure of 3500 psi with an average CO2 flow of 1.5 ml/min. The extracts were collected
and air-dried over nitrogen gas, and dissolved in ethyl acetate for further analysis.
2.4. GC/MS analysis
GC/MS analyses were performed using a ThermoFinniganTrace GC 2000 equipped
with a Polaris Q mass detector and the Xcalibur software system. A Supelcowax-10
fused silica capillary column (30 m×0.25 mm, film thickness 0.25 m) was used. The
oven temperature was initiated at 50°C for 5 min then increased at a rate of 5°C/min to
280°C and held for 5 min. The injector and ion source temperatures were 230°C and
200°C, respectively. Helium was used as the carrier gas at 1 ml/min. The mass
spectrometer was operated in electron impact mode (70 eV). Data acquisition was
performed in the full scan mode from m/z 50­650 with a scan time of 0.58 s.
2.5. Clustering and visualization
Two multivariate statistical analysis and visualization methods, namely, biplot
[30,31] and generalized association plots GAP [32,33], were used to analyze the GC/MS
data of 15 extracts from E. purpurea, E. pallida, and E. angustifolia roots (5 independent
root extracts from each species). A base 10 logarithm transformation was applied to
peak area data for scaling range of peak areas from 0­108
down to 0­8. Seventy-nine
major compound peaks from each GC chromatogram (a 15-by-79 data matrix) were
analyzed and displayed by GAP with two proximity matrices and two hierarchical
clustering trees (dendrograms) for illustrating plant clusters and compound groups.
Biplot is then used to display originally high-dimensional relative between-plant and
between-compound structures with the interaction patterns of compound groups on
plant clusters in a low (two) dimensional space.
2.6. High-performance liquid chromatography­electron spray ionization/mass
spectrometry analysis of metabolite profiles of E. purpurea root extracts with various
post-harvest or abiotic treatments
E. purpurea plants harvested at the same flowering stage were randomly divided into
six experimental groups. Plant samples with leaves struck with a brush of metal needles
one day before harvest were defined as the physical stress group. The drought stress group
did not receive watering one week before harvest. For post-harvest treatment groups, the
Echinacea plants were harvested and then shade-dried, sun-dried, or incubated at 42°C for
2 h before being lyophilized to dryness. Plant root or aerial tissue extracts were prepared
by ultrasonic extraction. A ThermoFinnigan LCQ Advantage ion trap mass spectrometer
(San Jose, CA, USA), coupled with an Agilent 1100 series liquid chromatograph system was
employed for HPLC/electron spray ionization/mass spectrometry (ESI/MS) analysis at the
Metabolomics Core Facility of the Agricultural Biotechnology Research Center, Academia
Sinica, Taiwan. One-gram samples of lyophilized root tissue were ground and extracted
with 10 ml MeOH for 15 min ultrasonically and then filtered through 0.22 m
polytetrafluoroethylene (PTFE) membranes prior to injection into a RP-18 column
[Phenomenex Luna 3 m C18(2), 150 mm×2.0 mm] at a flow rate of 0.2 ml/min. Eluting
peaks were monitored byphotodiodearray(DAD) detector simultaneouslyat210 and 254
nm for alkamides and at 254 and 330 nm for phenolic constituents before MS injection.
Chromatography of alkamides used a gradient of MeOH in water: 40­40% from 0 to 5 min,
40­60% from 5 to 15 min and 60­100% from 15 to 60 min. The phenolic constituents were
analyzed with gradient elution by 0.05% trifluoroacetic acid (TFA)/water (solvent A) and
0.05% TFA/MeOH (solvent B): 5­5% from 0 to 5 min, 5­20% from 5 to 30 min, 20­45% from
30-40 min and 45­50% from 40 to 50 min. Ionization was performed in the positive ion
mode for all analyses.
2.7. Cell lines and culture conditions
Murine macrophage RAW 264.7 cell and human breast adenocarcinoma MCF-7 cell
were obtained from the American Type Culture Collection (Rockville, MD, USA). RAW
264.7 cells were grown in Dulbecco's modified Eagle medium (DMEM) (Gibco/BRL,
Grand Island, NY, USA), and MCF-7 cells were grown in RPMI 1640 medium (Gibco/
BRL), both media supplemented with 10% heat-inactivated fetal bovine serum (Gibco/
BRL), 100 U/ml penicillin and 100 g/ml streptomycin, in a 5% CO2 incubator at 37°C in
a humidified atmosphere.
2.8. Isolation of alkamide (8/9) and cichoric acid from E. purpurea
Approximately 1.5 kg of fresh E. purpurea roots were extracted by 100% methanol
(1.6 L×2) at room temperature. The total methanolic extract was evaporated in
vacuum, suspended in 1.0 L of water and partitioned with ethyl acetate (1.0 l×3) to
yield two fractions designated as ethyl acetate (EA) and water fractions. The water
fraction was further partitioned with n-BuOH to obtain BuOH fraction. Each fraction
was dried using rotary evaporator and lyophilizer. The EA fraction was chromatographed
on a silica gel column with a hexane-ethyl acetate gradient to yield Fractions
1­5. Of these, Fraction 3 contained an alkamides mixture (Alk mix). Fraction 3 was
subsequently separated on a RP-18 silica gel (75C18-OPN) column eluted with a H2O/
MeOH gradient to give dodeca-2E,4E,8Z,10Z(E)-tetraenoic acid isobutylamides (8/9).
The butanol fraction was chromatographed on Diaion gel with a H2O/MeOH solvent
system, and further purified on a Sephadex LH-20 column eluted with 60% MeOH to
give cichoric acid. The compound purity of alkamides 8/9 and cichoric acid were of
N95%. Chemical structures were identified by nuclear magnetic resonance (NMR) and
mass spectral analyses and agreed with the data published previously [34].
2.9. Luciferase reporter assay
Chimeric luciferase reporter gene containing the full-length 5-flanking promoter
region (-1334/-1 bp) of the human COX-2 promoter was employed for the study,
following a method published elsewhere [35]. The promoterless pPGL-3-basic vector
2 C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
Fig. 1. Typical chromatography and metabolite profiling of root extracts of E. purpurea (A), E. pallida (B), and E. angustifolia (C) from supercritical fluid extraction. Five samples of each
species were analyzed, with similar results.
3C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
Fig. 2. GC/MS spectrometric analysis of the root extracts of E. purpurea (A), E. pallida (B), and E. angustifolia (C). Peaks were identified by comparison of their retention times and MS
spectrum with the previous data in the literature.
4 C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
was used as a negative control plasmid for luciferase assay. Briefly, MCF-7 cells were
seeded in 24-well plates at a density of 1.2×105
cells per well and cotransfected with
pCOX-2-Luc and pPL-TK-Luc (Promega, Madison, WI, USA) plasmids using
LipofectAMINE reagent (Invitrogen), as suggested by the manufacturer's protocol.
The transfected cells were then treated with vehicle (0.001% dimethyl sulfoxide
[DMSO]), 50 ng/ml TPA alone, or TPA mixed with test extracts or compounds at
indicated concentrations for 6 h. Cell lysates were prepared using Passive Lysis Buffer
(Promega) and then employed for luciferase activity assays. The promoter activity in
arbitrary units was obtained from the ratio of firefly luciferase activity (pCOX-2-Luc) to
that of Renilla luciferase (pPL-TK-Luc).
2.10. Immunohistochemical study of COX-2 protein expression in mouse skin
Female ICR mice were supplied from the Laboratory Animal Center (College of
Medicine, National Taiwan University, Taipei, Taiwan). Mice were supplied water feed
ad libitum and kept on a 12-h light/dark cycle at 222°C. Mice were topically treated
on their shaven backs with vehicle (acetone, 200 l per site) or TPA (10 nmol/200 l
per site) for 4 h. For the extract or compound treatments, mice were pretreated with
extract or compound for 30 min first, further treated with TPA for 4 h and finally killed
by cervical dislocation. The formalin-fixed, paraffin-embedded skin tissues were then
stained according to the method of Chiang et al. [35].
2.11. Measurement of NO production and cell viability
RAW 264.7 cells were seeded in 96-well plates at 2×105
cells per well. The cells
were treated with test compounds for 1 h and then incubated for 24 h in fresh DMEM
with or without 1.0 g/ml LPS. Nitrite levels in RAW 264.7 cell culture medium were
determined using the Griess reaction. Aliquots (100 l) of cell culture supernatants
were reacted with 100 l of Griess reagent (1:1 mixture of 0.1% N-(1-naphthyl)
ethylenediamine in H2O and 1% sulfanilamide in 5% phosphoric acid) in a 96-well plate,
and absorbance at 540 nm was recorded using a plate reader (Labsystems Multiskan
MS). Results were expressed as total M nitrite produced in the three test groups:
untreated cells, cells treated with LPS only, or cells treated with LPS plus phyto-extract
or compound. RAW 264.7 cells (1×104
) treated for 24 h with test extracts or
compounds were examined for cell viability using the MTT based colorimetric assay
[36]. Survival of macrophage cells after treatment was calculated by the following
Fig. 3. Chemical structures of the 34 metabolites from three Echinacea species.
5C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
formula: viable cell number (%)=OD570 (treated cell culture)/OD570 (vehicle
control)×100.
2.12. COX activity assays
COX-2 and COX-1 enzymatic activity was measured using the chemiluminescent
COX inhibitor screening assay kit (Cayman Chemical, Ann Arbor, MI, USA) according to
the manufacturer's instructions.
2.13. Quantification of cytokine profile with multicytokine detection system
Cytokine production profile in RAW 264.7 cells was analyzed using Beadlyte mouse
22-plex multicytokine detection system (Upstate) according to the manufacturer's
instructions. Briefly, RAW 264.7 cells (2×105
cells per well) seeded in 96-well plates
were treated with test compounds for 1 h and then incubated for 4 and 20 h,
respectively, with or without cotreatment with 1.0 g/ml LPS; 50 l of cultured media
were incubated with 25 l of bead solution overnight at 4°C and then incubated with
biotin for 1.5 h. Finally, streptavidin-phycoerythrin was added and incubated in
darkness for 30 min. Multi-wavelength fluorescence and cytokine concentrations were
determined by laser detector (Luminex 100, Luminex, Austin, TX, USA) and Bio-Rad
software (Bio-Rad, Hercules, CA, USA). Assays were performed in triplicate, and the
obtained mean values were used for data analysis.
2.14. Western blot analysis
Total cellular protein extracts were prepared according to Lo et al. [37]. The protein
content was measured by the Bradford method (Bio-Rad Protein Assay Kit). Protein (20
g) was separated by 5­20% gradient sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (SDS-PAGE) and then electrotransferred onto polyvinylidene difluoride
membrane (Immobilon, Millipore, Bedford, MA, USA). The blots were incubated in
blocking buffer (3% w/v skim milk in Tris-buffered saline) for 30 min and then
incubated with monoclonal antibodies against G3PDH (Biogenesis, Poole, UK), COX-2
(Cayman Chemical), or iNOS (Santa Cruz Biotechnology, Santa Cruz, CA, USA) overnight
at 4°C. After incubation with the appropriate antirabbit or antimouse IgG conjugated to
horseradish peroxidase for 3 h at room temperature, the immunoreactive bands were
visualized using enhanced chemiluminescence reagents (ECL, Amersham).
2.15. Statistical analysis
Data are expressed as meanS.D. Statistical significance of differences between
treatments was determined by analysis of variance (ANOVA) with Fisher's post hoc
test. Pb.05 was considered to be statistically significant.
3. Results
3.1. GC/MS analysis and compound identification of plant extracts from
three Echinacea species
Randomly selected root tissues harvested from flowering stage of
the three medicinal species E. purpurea, E. pallida, and E. angustifolia
(5 roots from each species) were collected and extracts prepared
using SFE. The total yield of E. purpurea, E. pallida, and E. angustifolia
extracts were 1.07%, 1.22% and 0.86%, respectively, by dry weight of
root tissues. Equal amounts of extracts were dissolved in ethyl acetate
and analyzed by GC/MS (Fig. 1). Very similar peak profiles were seen
in root extracts from the same Echinacea species, while distinct
chromatograms were observed between different species, suggesting
an efficient and reproducible protocol of extraction and GC/MS
analysis. The majority of the specific compounds in the spectra were
verified on the basis of retention time and mass spectra and agreed
with previously published spectral data [34,38­40] and the NIST and
Wiley databases (Figs. 2 and 3). Alkamides were identified as the
major lipophilic constituents in E. purpurea and E. angustifolia root
extracts (Figs. 2A, C, and 3), particularly undeca- and dodecanoic acids
with different degrees of unsaturation and bond configurations.
Among them, isobutylamide was the major structural type, and some
2-methylbutylamides were also observed (Fig. 3). E. pallida extract
contained far less alkamides than other two species, while a
significant amount of ketoalkenes and ketoalkynes were detected
(Figs. 2B and 3).
The metabolite distribution and relative content in the three
Echinacea root extracts were calculated on the basis of the ratio of
individual compound peak area to the total peak areas in each
chromatogram (Table 1). E. purpurea and E. angustifolia root
extracts contained approximately 74.06% and 90.73% alkamides,
respectively, while E. pallida root extract mainly composed of
polyacetylenes (80.81%) and only 6.12% alkamides. Isomeric
dodeca-2E,4E,8Z,10Z(E)-tetraenoic acid isobutylamides (8/9) were
the major alkamides, comprising approximately 13.99% and 22.38%,
respectively, of the E. purpurea and E. angustifolia root extracts,
determined using a standard calibration curve of alkamide. 2Monoene-8,10-diynoic
acid isobutylamide was only detected in E.
angustifolia and some monoterpene constituents were mainly
detected in E. purpurea (Fig. 3). These results demonstrate that
SFE coupled with GC/MS analysis can be effectively applied for
profiling and differentiating the chemical constituents existed in
Table 1
A summary of metabolite distribution in the three medicinal Echinacea plants
Metabolites E. purpurea E. pallida E. angustifolia
Alkamides Relative content (%)
undeca-2Z(E),4E(Z)-dien-8,10-diynoic acid
isobutylamide (1/2)
15.47 3.53
trideca-2E,7Z-dien-10,12-diynoic acid
isobutylamide (3)
5.19
dodeca-2Z,4E-dien-8,10-diynoic acid
2-methylbutylamide (4)
12.89 1.92
dodeca-2E,4Z-dien-8,10-diynoic acid
2-methylbutylamide (5)
16.36 0.85 1.10
dodeca-2Z(E),4E(Z)-en-8,10-diynoic acid
2-methylbutylamide (6/7)
1.92
dodeca-2E,4E,8Z,10E(Z)-tetraenoic acid
isobutylamide (8/9)
11.74 20.64
dodeca-2E,4E,8Z,10Z(E)-tetraenoic acid
isobutylamide (9/8)
2.25 1.74
dodeca-2E,4E,8Z-trienoic acid
isobutylamide (10)
0.74 5.02
dodeca-2E,4E,10E-trien-8-ynoic acid
isobutylamide (11)
6.07 5.29
dodeca-2E,4E-dienoic acid
isobutylamide (12)
1.43 14.85
undeca-2Z(E)-en-8,10-diynoic acid
isobutylamide (13)
16.66
undeca-2E(Z)-en-8,10-diynoic acid
isobutylamide (14)
12.05
dodeca-2E-en-8,10-diynoic acid
isobutylamide (15)
4.35
undeca-2Z-en-8,10-diynoic acid
2-methylbutylamide (16)
4.94
dodeca-2E-en-8,10-diynoic acid
2-methylbutylamide (17)
1.64
pentadeca-2E,9Z-dien-12,14-diynoic acid
isobutylamide (18)
2.45
Polyacetylenes
pentadeca-1,8-diene (19) 1.85 6.15
pentadeca-1,8,11-trien (20) 0.95
heptadeca-1,8,11-triene (21) 0.60
pentadeca-8Z-en-2-one (22) 24.02
pentadeca-8Z,11Z-dien-2-ol (23) 11.52
tetradeca-8Z-en-11,13-diyn-2-one (24) 16.44
pentadeca-8Z,13Z-dien-11-yn-2-ol (25) 12.96
heptadeca-8Z,11-dien-2-one (26) 1.76
pentadeca-8Z,11,13-trien-2-ol (27) 0.79
pentadeca-8Z-en-11,13-diyn-2-one (28) 5.62
Terpenes
aromadendren (29) 0.44
germacrene-D (30) 0.87
-gurjunene (31) 0.27
p-menth-1-ene-6-ol (32) 0.32
Peak numbering corresponds to Fig. 2.a
a
The relative content of each peaks were calculated on the basis of the ratio of
individual compound peak area to the total peak areas in each chromatogram.
6 C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
root extracts of these three Echinacea species. The aerial parts of the
three Echinacea species were also analyzed using the same
protocols; however, very few detectable peaks were seen, suggesting
that the aerial portion of Echinacea plants contain very little
alkamides (data not shown).
3.2. Metabolite clustering and visualization of the three medicinal
Echinacea species using biplot and generalized association plots GAP
GAP and biplot analyses were used to analyze the metabolites
distribution and content in E. purpurea, E. pallida, and E. angustifolia
Fig. 4. Metabolite clustering and visualization of 15 plants from three Echinacea species [E. purpurea (blue triangles), E. pallida (red triangles), and E. angustifolia (green triangles)] with
79 phytocompounds: observed in E. purpurea only (blue lines or dashed circles), E. pallida only (red lines or dashed circles), and E. angustifolia only (green lines or dashed circles); E.
purpurea/angustifolia shared compounds (cyan lines or dashed circles), E. purpurea/pallida shared compounds (orange lines or dashed circles), and E. angustifolia/pallida shared
compounds (magenta lines or dashed circles); E. purpurea/angustifolia/pallida shared compounds (grey lines or dashed circles) using two statistical and visualization methods. (A)
Generalized Association Plots (GAP). (B) Biplot display.
7C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
root tissues with base 10 logarithm transformed peak area of seventynine
compounds. We have developed a GAP method to analyze plant
metabolites. GAP is capable of providing five levels of information: (1)
a raw score for every single plant/compound combination (15-by-79
matrix); (2) individual plant score-vectors across all compounds (15
rows) and individual compound-vectors of all plants (79 columns);
(3) an association score for every plant-plant (15-by-15 matrix) and
compound-compound relationship (79-by-79 matrix); 4) a grouping
structure for compounds (dendrogram for compounds) and clustering
effect for plants (dendrogram for plants); and 5) an interaction
pattern of plant-clusters on compound-groups. Seventy-nine highly
reproducible compound peaks detected by GC/MS in each chromatogram
from individual root extract (5 root extracts from each species
for 15 root extracts in total) were analyzed and expressed as log10
(peak intensity). The GAP display in Fig. 4A has three matrix maps
permuted by two dendrograms. Two hierarchical clustering trees
(HCTs) with dendrograms are built on the between-plant Euclidean
distance matrix (15-by-15) and between-compound Pearson correlation
matrix (79-by-79), respectively for permuting these two
corresponding proximity matrices and the raw data (log10 peakFig.
5. Characterization of the effect of three Echinacea species on transcriptional activities of the cox-2 promoter. (A) Human breast adenocarcinoma MCF-7 cells co-transfected with
the full-length pCOX-2-Luc and internal control plasmid pRL-TK-Luc plasmid constructs were stimulated with TPA, in the absence or presence of the indicated concentrations of
extracts. Control cells were not exposed to TPA. Aliquots (10 g) of total protein lysate were subjected to dual luciferase reporter assay. The relative induction of cox-2 promoter
activity is shown, normalized to TPA treatment alone (100%). The significance of differences between means of different treatments were analyzed by Student's t-test. Significant
inhibition is indicated by *, **, and ***, with a P value b.1, b.05, and b.01, respectively. Indo, indomethacin treatment. (B) Immunohistochemical study of the inhibitory effect of
Echinacea roots extracts on TPA-induced COX-2 expression in mouse skin. Dorsal skin of female ICR mice was treated topically with acetone (vehicle control) or TPA (10 nmol in
acetone) for 4 h or with the indicated concentrations of Echinacea extracts or celecoxib for 30 min before TPA treatment for 4 h. Immunohistograms were taken with an Olympus DP-
70 camera on a Nikon ECLIPSE E800 microscope (magnification: ×200). The arrows indicate positive COX-2 staining (brown product).
8 C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
area) matrix (15-by-79). In the construction of the two HCTs, average
linkage distance option was used to compute the between-branch
distances with flipping mechanic of intermediate nodes guided by
rank-two elliptical seriation in Tien et al. [41].
The upper right distance matrix in rainbow scale clearly shows the
distinct clusters of the three Echinacea species analyzed based on the
Euclidean distance (although E. purpurea and E. angustifolia plants are
closer to each other than to E. pallida plants) assisted by an HCT with
color branches denoting three Echinacea species (blue for E. purpurea,
green for E. angustifolia and red for E. pallida); the lower left
correlation matrix in blue-red spectrum together with the colored
branches also suggest roughly seven compound groups; the upper left
plant-by-compound data matrix in rainbow spectrum then correlates
the interaction pattern of the seven compound groups on the three
plant clusters shown in Fig. 4B. Three compound groups were each
expressed in only one Echinacea species (blue for E. purpurea, green
for E. angustifolia and red for E. pallida); the other three compound
groups were shared by two Echinacea species (cyan for E. purpurea/
angustifolia, orange for E. purpurea/pallida, and magenta for E.
angustifolia/pallida), and one compound group was seen in all three
Echinacea species (grey).
A biplot can simultaneously display originally high-dimensional
relative between-plant and between-compound structures and their
interaction patterns in a low dimensional space. The between-plant
part is similar to a principal component analysis [42,43] for exploring
the between-sample relationship while the between-compound part
is related to a factor analysis [44] for understanding the betweencompound
structure. The biplot displayed in Fig. 4B contains two sets
of symbols namely fifteen triangles one color groups for each plant
with seventy-nine circles one for each compound. Color coding used
for the biplot in Fig. 4B is identical to that used in Fig. 4A for GAP.
Similar to the pattern in the GAP display for proximity matrix and
data matrix in Fig. 4A, seven compound clusters can be easily
identified in the biplot display of Fig. 4B. The red, green, and blue
circled compound clusters contain those compounds only detected in
E. pallida, E. angustifolia, and E. purpurea species, respectively. The
cyan, magenta, and yellow clusters contain compounds shared by
pairs of species. The middle grey cluster present compounds
identified as common to all three Echinacea species.
Fig. 4A and B suggest that both GAP and biplot analyses can clearly
differentiate the distribution of unique or common compounds in
different medicinal plant species. While both GAP and biplot are
capable in illustrating plant-clusters, compound-groups, and their
interaction patterns, the exact compound-profile shared by each plant
and the plant-profile affiliated to each compound are only available in
the GAP display.
3.3. Comparative study of the extracts from three Echinacea species on
COX-2 inhibition
COX-2 is an important inducible enzyme responsible for the
production of a large amount of proinflammatory prostaglandins at
an inflammatory site. We were interested in any correlation between
the metabolite profile and major metabolites of specific root extracts
with their potential anti-inflammatory bioactivity. Luciferase reporter
gene activity driven by the 12-O-tetradecanoylphorbol-13-acetate
(TPA)-induced COX-2 promoter in MCF-7 cells was employed as a
primary system to compare the bioactivity of the root extracts from
the three Echinacea species. Fig. 5A shows that E. angustifolia extract
at 10 and 50 g/ml suppressed TPA-induced full-length COX-2
promoter activity (100%) in MCF-7 cells to 67% (Pb.05) and 34%
(Pb1) of control, respectively. Aspirin (45 g/ml) and indomethacin
Fig. 6. Effect of alkamides on COX-2 enzymatic activity and expression, nitric oxide (NO)
production and iNOS protein expression in RAW 264.7 cells. (A) Test compounds at the
indicated concentrations were incubated with COX-2 enzyme and substrates, and COX-
2 activity measured. Celecoxib was used as a positive control. The data are
representative of three experiments and expressed as meanS.D. Significant inhibition
is indicated by *, **, and ***, with a P value b.05, b.01, and b.005, respectively. (B) Cells
were treated with the indicated concentrations of test compounds for 1 h, then LPS
(1 g/ml) was added for 18 h. Total cellular protein (20 g) was resolved by SDS-PAGE
then transferred to PVDF membrane and detected with antibody against COX-2 protein.
COX-2 protein levels were normalized to G3PDH. (C) Effect of alkamide mixtures and
8/9 on LPS-induced NO production. Cells were treated with or without the indicated
concentrations of test compound in 0.5% DMSO for 1 h, then LPS (1 g/ml) was added
and incubated for a further 24 h. Vehicle control was treatment with 0.5% DMSO only.
NO in the culture medium was measured as described in Materials and methods. Cells
were treated with or without the indicated concentration of test compounds for 24
h and percentage of cell viability was calculated. The data are expressed as meanSD
(n=3). Student's t-test was used to analyze significance of differences, where *
indicates P-value b.001. (D) Effect of test compounds on LPS-induced iNOS and HO-1
protein expression in RAW264.7 cells. Total cellular protein (20 g) was resolved by
SDS-PAGE then transferred to PVDF membrane and detected with specific antibodies.
iNOS and HO-1 protein expression were normalized to G3PDH levels (internal control).
9C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
(3.5 g/ml), used as reference controls in this experiment, suppressed
COX-2 promoter activity to 30-50% of the TPA control (Pb1).
Approximately 72% promoter activity was detected when transfected
MCF-7 cells were treated with 50 g/ml E. purpurea root extract
(Pb.1), while the E. pallida root extract did not show any suppressive
effect on COX-2 promoter activity at the same test concentration. The
major differences in suppression of the TPA-induced cox-2 promoter
activity between the root extracts of E. purpurea and E angustifolia are
likely to be because the total content of alkamides is greater in root
tissues of E angustifolia than in E. purpurea (90.73% vs. 74.06%,
Table 1). Additionally, Alk 8/9 and Alk 12 exist a larger amounts in E.
angustifolia than E. purpurea (22.38% vs. 13.99% and 14.85% vs. 1.43%,
respectively, shown in Table 1). When both compounds were tested
at the same concentration 100 M, Alk 12 was able to suppress a much
more cox-2 promoter activity induced by TPA than Alk 8/9 (60% vs.
24%, shown in Supplementary Fig. 1), indicating that at least Alk 12
makes a significant contribution to the detected difference in the cox-
2 promoter activity observed between the root extracts.
A TPA-induced mouse skin inflammation system and immunohistochemical
analysis were employed to further evaluate the ability of the
three Echinacea root extracts to inhibit COX-2 protein expression in
vivo. After treatment with TPA (10 nmol/200 l/site) for 4 h, COX-2
protein levels increased significantly in the epidermal layer of mouse
skin compared to the vehicle control (Fig. 5B). At a dose of 1 mg/200 l/
site on mouse skin, the root extract of E. angustifolia significantly
attenuated TPA-induced COX-2 protein expression with effect comparable
to that of celecoxib, a nonsteroidal anti-inflammatory agent and
selective COX-2 inhibitor, at the same test concentration. Little or no
inhibitory effect were observed for E. purpurea and E. pallida extracts,
respectively (Fig. 5B). As shown in Table 1, alkamides were the major
constituent components in the E. purpurea and E. angustifolia root
supercritical fluid extracts, suggesting that alkamides could bethe major
active compounds for the anti-inflammatory activity in both Echinacea
plants while polyacetylenes, the major constituent compounds in E.
pallida roots, might not have any inhibitory effect on COX-2 expression
at transcriptional and translational levels.
3.4. Echinacea alkamides inhibit COX-2 activity
In order to further characterize the anti-inflammatory activity of
the specific alkamides, we prepared an alkamide-enriched fraction
(designated Alk mix) and purified the major constituent 8/9 from the
E. purpurea roots by liquid chromatography. Alk mix at 5 and 25 g/ml
inhibited 37% and 63% COX-2 enzymatic activity, Alk 8/9 (at 5 and
100 M) inhibited COX-2 activity by 20-42%, while celecoxib at 0.1-
0.5 M inhibited activity by 40-62% (Fig. 6A). No detectable inhibition
of COX-1 activity was observed with the Alk mix or Alk 8/9 (data not
shown). Immunoblotting analysis of COX-2 protein expression in LPSstimulated
macrophages showed better suppression by Alk mix than
Alk 8/9 (40-53% vs. 0-19%) at the same concentrations (Fig. 6B).
Table 2
Effect of alkamides and chichoric acid isolated from E. purpurea on cytokine expression profile in murine macrophage RAW 264.7
10 C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
These results suggest that other alkamide(s) present in Alk mix, apart
from 8/9, also contributed to the inhibition of COX-2 activity and
protein expression.
3.5. Echinacea alkamides significantly inhibit NO production and iNOS
protein expression in LPS-stimulated macrophage
The anti-inflammatory effect of alkamides on LPS-induced nitric
oxide (NO) production and iNOS expression in RAW 264.7 cells was
studied. Alk mix and Alk 8/9 induced a dose-dependent inhibition
(Pb001) of nitrite production with IC50 of 18.11 g/ml and 36.43 M,
respectively (Fig. 6C). Western blot analysis also demonstrated
significant inhibition of LPS-induced iNOS protein expression by
alkamides. Following incubation with either 5 g/ml Alk mix or 5 M
Alk 8/9, the level of iNOS protein was reduced by approximately 50%
and 70%, and complete inhibition was observed with 25 g/ml Alk mix
and 100 M Alk 8/9 (Fig. 6D). Thus, the inhibition of NO production by
alkamides in LPS-stimulated macrophages is partly mediated by
suppressing the expression of iNOS. The levels of HO-1 were increased
by the higher concentrations of both alkamide treatments.
3.6. Effects of alkamides on cytokine/chemokine profiles in murine
macrophage with or without LPS stimulation
The effects of alkamides on cytokines or chemokines secretion in
RAW 264.7 cells with or without LPS stimulation were evaluated
Table 3
Effect of alkamides and cichoric acid isolated from E. purpurea on cytokine expression profile in LPS-stimulated macrophage RAW 264.7 cells
11C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
using a Beadlyte mouse cytokine array assay. Cichoric acid was tested
in parallel as a reference in this study. Twenty-two cytokines/
chemokines [interleukin (IL)-1, IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7,
IL-8, IL-10, IL-12p40, IL-12p70, IL-13, IL-15, TNF-, GM-CSF, IFN,
eotaxin, MCP-1, RANTES, MIP-1, and KC] on a single array were
analyzed. Those molecules responsive to phytocompound treatment
are summarized in Tables 2 and 3. An MTT viability assay confirmed
that Alk mix, Alk 8/9, and cichoric acid did not have detectable
cytotoxicity on macrophages at the tested concentrations (data not
shown). The Alk mix (at 5 or 25 g/ml) and Alk 8/9 (at 100 M)
suppressed the secretion of TNF-, IL-1, IL-6, IL-10, IL-12p70, IL-13,
vascular endothelial growth factor (VEGF), MCP-1, MIP-1, and
RANTES in RAW 264.7 cells, with a more significant effect after a 20h
treatment than after a 4-h treatment (Table 2). Cichoric acid (100
M) down-regulated TNF-, IL-6, IL-10, IL-12p70, IL-13, VEGF, and
MIP-1 with 4- or 20-h treatments, with a near tenfold reduction in
VEGF levels. These results suggest that alkamides and cichoric acid
may modulate macrophage activity, mainly by suppression of
cytokine and chemokine levels.
RAW 264.7 cells were stimulated with the bacterial endotoxin LPS
for 4 h or 20 h. After 4 h, the levels in the culture medium of 7
cytokines and 3 chemokines (TNF-, IL-1, IL-1, IL-2, IL-6, IL-10, IL-
12p70, MCP-1, MIP-1, and RANTES) were significantly elevated (1.5to
26-fold) (Table 3A), while after a longer time LPS stimulation (20
h), IL-13, VEGF, GM-CSF and KC were also increased (1.75- to 2.3-fold)
(Table 3B) compared to cells treated with vehicle only. When LPSchallenged
cells were treated with Alk mix (25 g/ml) or Alk 8/9 (100
M), the LPS-induced expression of all tested cytokines/chemokines
was reduced, with a stronger effect from Alk 8/9 (Table 3B). Cichoric
acid (100 M treatment for 4 h or 20 h), however, only inhibited LPSinduced
IL-10 (4.3-fold), TNF- (1.9-fold), and VEGF (3.8-fold)
expression, with a mild elevation (1.5-fold) of LPS-induced IL-2
level and a 6.4-fold increase in MIP-1 level. These results indicate
that compound alkamides are the key anti-inflammatory constituents
in Echinacea plants.
3.7. HPLC-ESI/MS metabolite profiling of alkamides and phenolic
compounds in E. purpurea roots under various post-harvest or abiotic
treatments
We examined whether various treatments before or after plant
harvesting can alter the specific alkamide or phenolic contents in E.
purpurea roots. E. purpurea plants at the flowering stage were divided
into 6 groups: untreated, physical stress (striking the leaves with a
metal brush), simulated drought, 42°C treatment, shade-dried, and
sun-dried groups, and the metabolite profiles of the root extracts,
prepared using ultrasonic extraction, were analyzed using HPLC-ESI/
MS. Table 4 summarizes the changes in content of alkamides 1-12,
cichoric acid and rutin (see also Fig. 7) quantified against a standard
calibration curve of the respective compound and analyzed by oneway
ANOVA. Total alkamides content in E. purpurea roots were
decreased by all five treatments as compared to the untreated root
extract; interestingly, the physical stress had the most dramatic effect,
with a decrease in total alkamide content of up to 44%; decreases of
37.8, 34.7, 26.7, and 22.7% were observed in shade-dried, sun-dried,
drought stress, and 42°C treatment, respectively (Table 4). Among the
12 alkamides, the changes of major alkamides 8/9 were statistically
significant (Pb5) in the physical stress, shade-dried and 42°C
treatment groups (Fig. 7A).
The results of quantitative analysis of cichoric acid and rutin under
various treatments show that shade-dried and sun-dried treatments
drastically attenuated both compounds contents (Pb5) in E. purpurea
roots (73-92% decrease, Pb5) (Table 4 and Fig. 7B). Surprisingly, the
physical stress resulted in approximately 65% increase in rutin
content (Pb5). Together, these results suggest that environmental
factors or post-harvest treatment can have some substantial effects on
metabolites distribution in medicinal plants.
4. Discussion
Echinacea preparations were the top-selling herbal supplements
or medicines in the past decade and represent the most common
herbal immunomodulators (http://nccam.nih.gov/). Current commercial
preparations sold under the name Echinacea may in fact be
completely different preparations, using different extraction methods
and plant parts or a mixture of plant varieties. It is clear, therefore,
that there are still many challenges facing the rational use of Echinacea,
such as determination of the most effective parts of the plant,
and the differences in bioefficay of the three most abundantly used
species (http://nccam.nih.gov/). We therefore established here an
efficient and highly reproducible extraction and analysis protocol for
better quality control and species classification of Echinacea plant
extracts using a metabolomics approach. The three Echinacea plants
were grown under controlled and home-defined organic farming
conditions and subjected to supercritical fluid extraction couple with
GC/MS for enrichment of and analyzing lipophilic constituents in
Echinacea roots. By employing PCA and bioplot analyses, the
distribution of unique and common compounds within the three
medicinal plant species could be clearly differentiated (Fig. 4).
Analysis of the bioefficacy of the metabolites identified by the
Table 4
Effect of specific post-harvest or abiotic treatments on metabolites in E. purpurea root extractsa
Compound (mg/g) Untreated Physical Drought 42°C treatment Shade-dried Sun-dried
Alkamides
1 1.890.33 1.060.11 1.370.23 1.480.30 1.240.23 0.870.07
2 0.690.31 0.210.05 0.310.05 0.430.15 0.280.02 0.380.06
3 0.690.08 0.470.10 0.490.11 0.650.07 0.450.13 0.310.05
4 1.050.26 0.660.18 0.640.16 0.800.10 0.600.19 0.460.06
5 1.570.35 1.070.12 1.300.39 1.690.25 1.030.32 0.880.26
6 0.100.03 0.070.03 0.070.02 0.080.02 0.070.02 0.050.01
7 0.120.01 0.090.02 0.100.02 0.130.02 0.090.02 0.060.02
8/9 3.920.28 1.830.24 3.040.65 2.330.28 2.460.46 3.570.38
10 0.230.09 0.090.02 0.100.03 0.140.05 0.100.01 0.110.02
11 0.260.09 0.360.20 0.290.07 0.340.09 0.210.06 0.170.04
12 0.100.03 0.090.03 0.070.02 0.120.03 0.080.01 0.080.03
Total alkamides 10.621.20 5.990.54 7.780.62 8.211.00 6.610.75 6.940.33
Cichoric acid 29.307.82 32.3011.92 26.393.64 23.0910.12 7.813.53 2.351.41
Rutin 109.4627.04 180.6944.53 94.3414.33 74.7350.29 8.483.84 11.663.13
Total 138.7526.18 212.9950.05 120.7314.38 97.8259.29 16.295.21 14.012.72
a
Values are means (S.D.). Mean levels of the compounds were measured as mg/g in dry weight (n=5).
12 C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
metabolomics profiling by cell- and gene-based assays demonstrated
that the order of efficacy of the three plant root extracts was E.
angustifolia N E. purpurea N E. pallida (Fig. 5). The abundant alkamides
in both E. angustifolia and E. purpurea were then correlated with the
observed anti-inflammatory activity in the extracts, in further
bioactivity studies using an enriched alkamides mixture or major
alkamides 8/9 (Fig. 6).
LPS from the outer membrane of Gram-negative bacteria can
trigger potent inflammatory responses during endotoxemia, including
induction of iNOS and COX-2 gene and/or protein expression and
cytokines released [45,46]. Our results indicate that alkamides
mixture (Alk mix) and its major constituents Alk 8/9 from E. purpurea
can inhibit expression of both iNOS and COX-2 proteins, and NO
production in LPS-stimulated murine macrophages. In comparison,
the purified Alk 8/9 had less effect than the Alk mix, indicating that
other alkamides present in Alk mix also contribute to these antiinflammatory
activities. Indeed, our preliminary data for a more
recent study shows that Alk 12 has a higher effect than Alk 8/9 on
suppression of transient COX-2 promoter activity induced by TPA in
MCF-7 cells (60% vs. 24%, shown in Supplementary Fig. 1) and NO
production (IC50=74.2 vs. 43.7 M, data not shown) in LPSstimulated
macrophages. Previous studies have reported that E.
purpurea aerial parts preparations induced IL-1, IL-10, and TNFsecretion
in human macrophages [21], and water extracts from E.
purpurea stimulated non-adherent mouse splenocyte populations to
produce TNF-, IL-10 and MIP-1 [19]. Likewise, Echinacea preparations,
either enriched in polysaccharides or in alkamide and phenolic
components, can reverse the expression changes of some inflammatory
cytokines or chemokines in rhinovirus-infected epithelial cells,
though treatment with either extract alone was also observed to
induce some proinflammatory cytokines or chemokines expression in
un-infected cells [18]. In this study, we observed that 4-h or 20h
treatment of murine macrophages with an alkamide mixture from
E. purpurea roots or with Alk 8/9 alone suppressed most proinflammtory
cytokines or chemokines, while only moderate effect was
seen with a 20-h cichoric acid treatment (Table 2). Similar alkamide
suppression of LPS-induced cytokines/chemokines activity in macrophages
was detected (Table 3), while cichoric acid only exhibited
moderate suppression of TNF- and VEGF secretion, and induction of
MIP-1 (Table 3B). Alkamides 8/9 and 12 have been reported
elsewhere to inhibit expression of TNF- induced by LPS in human
monocytes/macrophages via the CB2 receptor [19,27], which supports
part of our observed effects in murine macrophages. Together,
these data clearly indicate that the alkamides are the anti-inflammatory
constituents in Echinacea extracts, and suggest some potential
for the therapeutic use of alkamides in endotoxemia or septic shock
induced by LPS. Although cichoric acid has moderate inhibitory effects
on proinflammatory cytokine/chemokine secretion, little effect was
seen on COX-2 and iNOS expression (data not shown) in the same
macrophages with LPS stimulation, implying cichoric acid may not
play a role in LPS-related inflammatory disorders, but some other cellmediated
immunity.
Carbon monoxide, one the three byproducts of the catabolism of
heme by HO, can inhibit the expression of LPS-induced proinflammatory
cytokines both in vivo and in vitro [47,48]. The alkamide
promotion of HO-1 expression in macrophages (Fig. 6D) may
contribute to their inhibitory effect on the LPS-induced production of
proinflammatory cytokines TNF-, IL-1, and IL-6. Natural compound
suppression of iNOS expression by enhancing HO-1 expression has
been reported in previous studies [49­51]. Direct evidence to support
and link such a cascade mode of action of alkamides on HO-1 and
proinflammatory mediators awaits further studies.
Plant developmental stage, growing conditions, and post-harvest
treatments may significantly influence the plant secondary
Fig. 7. Alkamide, cichoric acid and rutin levels in roots of E. purpurea with or without specific post-harvest abiotic treatments. Bars with a different letter are significantly different from
each other (Pb5) as determined by ANOVA. Values are meansSD from five extracts (n=5) in two independent determinations.
13C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
metabolite status in plants. These changes may also be directly
correlated to the nutritional and functional values of vegetables or
medicinal plants. We attempted in this study to investigate whether
several simple treatments could affect the alkamide and phenolics
(i.e., cichoric acid and rutin) contents in E. purpurea roots (Table 4).
In summary, the total alkamide contents fell by 22% to 43% among
the five different treatments. Although some reduction was seen in
diynoic acid isobutylamide derivatives (1-4), the largest effects
were on the most abundant compound, Alk 8/9, with statistically
significant reductions following physical stress, heat treatment
(42°C for 2 h) and shade-drying. Conversely, the physical stressing
by striking the leaves with a brush of metal needles before harvest
significantly increased the rutin content, by 65% (Table 4).
Wounding by fresh-cutting has been reported elsewhere to increase
the total phenolic content in lettuce leaf tissue [52,53] and in
purple-flesh potatoes [54], and was correlated to an increase of
phenylalanine ammonia lyase activity and antioxidant activity.
Rutin is known to possess antioxidant activity, and a similar
increase in antioxidant activity can be expected in the E. purpurea
root tissue in the physical stress group. However, contradictory
effects were observed for alkamides and rutin in E. purpurea. It is
possible that an appropriate abiotic stress or post-harvest treatment
of the plant may be able to produce a product tailored to a specific
pharmacological profile or other desired outcome. It will be
interesting in future to investigate the signaling molecules,
enzymes, and pathway(s) that regulate the biosynthesis of the
alkamides or the specific phenolic compound in terms that translate
to their medical functions.
Acknowledgments
The authors are grateful to the High Field Nuclear Magnetic
Resonance Center, National Science and Technology Program for
Medical Genomics, Taiwan, for measurement of the NMR spectra. We
thank Dr Harry Wilson, Academia Sinica, for his careful reading of the
manuscript, and Miss Su-Shiang Lin for her technical assistant on
Echinacea plants cultivation.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jnutbio.2009.08.010.
References
[1] Rochfort S. Metabolomics reviewed: a new "omics" platform technology for
systems biology and implications for natural products research. J Nat Prod
2005;68:1813­20.
[2] Sumner LW, Mendes P, Dixon RA. Plant metabolomics: large-scale phytochemistry
in the functional genomics era. Phytochemistry 2003;62:817­36.
[3] Shyur LF, Yang NS. Metabolomics for phytomedicine research and drug
development. Curr Opin Chem Biol 2008;12:66­71.
[4] Wang M, Lamers RJ, Korthout HA, van Nesselrooij JH, Witkamp RF, van der
Heijden R. Metabolomics in the context of systems biology: bridging traditional
chinese medicine and molecular pharmacology. Phytother Res 2005;19:173­82.
[5] Bauer R. Chemistry, analysis and immunological investigations of Echinacea
phytopharmaceuticals. In: Wagner H, editor. Immunomodulatory agents from
plants. Boston (Mass): Birkhäuser Inc.; 1999. p. 41­88.
[6] Perry NB, van Klink JW, Burgess EJ, Parmenter GA. Alkamide levels in Echinacea
purpurea: a rapid analytical method revealing differences among roots, rhizomes,
stems, leaves and flowers. Planta Med 1997;63:58­62.
[7] He XG, Lin LZ, Bernart MW, Lian LZ. Analysis of alkamides in roots and achenes of
Echinacea purpurea by liquid chromatography-electrospray mass spectrometry. J
Chromatog A 1998;815:205­11.
[8] Sioley BD, Urichuk LJ, Tywin C, Coutts RT, Pang PKT, Shan JJ. Comparison of
chemical components and antioxidant capacity of different Echinacea species. J
Pharm Pharmacol 2001;53:849­57.
[9] Cech NB, Eleazer MS, Shoffner LT, Crosswhite MR, Davis AC, Mortenson AM. High
performance liquid chromatography/electrospray ionization mass spectrometry
for simultaneous analysis of alkamides and caffeic acid derivatives from Echinacea
purpurea extracts. J Chromato A 2006;1103:219­28.
[10] Perry NB, Burgess EJ, Glennie VI. Echinacea standardization: analytical methods
for phenolic compounds and typical levels in medicinal species. J Agric Food Chem
2001;49:1702­6.
[11] Gotti R, Pomponio R, Bertucci C, Cavrini V. Simultaneous analysis of the lipophilic
and hydrophilic markers of Echinacea plant extracts by capillary electrophoresis. J
Sep Sci 2002;25:1079­86.
[12] Bergeron C, Livesey JF, Awang DVC, Arnason JT, Rana J, Baum BR, et al. Quantitative
HPLC method for the quality assurance of Echinacea products on the north
American market. Phytochem Anal 2000;11:207­15.
[13] Gilroy CM, Steiner JF, Byers T, Shapiro H, Georgian W. Echinacea and truth in
labeling. Arch Intern Med 2003;163:699­704.
[14] Raloff J. Herbal lottery: what's on a dietary supplement's label may not be what's
in the bottle. Sci News 2003:163.
[15] Bauer R, Wagner H. Echinacea species as potential immunostimulatory drugs. In:
Wagner H, Farnsworth NR, editors. Economic and Medicinal Plant Research, vol. 5.
San Diego (Calif): Academic Press Inc.; 1991. p. 253­321.
[16] Masferrer JL, Leahy KM, Koki AT, Zweifel BS, Settle SL, Woerner BM, et al.
Antiangiogenic and antitumor activities of cyclooxygenase-2 inhibitors. Cancer
Res 2000;60:1306­11.
[17] Williams CS, Mann M, DuBois RN. The role of cyclooxygenases in inflammation,
cancer, and development. Oncogene 1999;18:7908­16.
[18] Sharma M, Arnason JT, Burt A, Hudson JB. Echinacea extracts modulate the pattern
of chemokine and cytokine secretion in rhinovirus-infected and uninfected
epithelial cells. Phytother Res 2006;20:147­52.
[19] Hwang SA, Dasgupta A, Actor JK. Cytokine production by non-adherent mouse
splenocyte cultures to Echinacea extracts. Clin Chimi Acta 2004;343:161­6.
[20] Barrett B. Medicinal properties of Echinacea: a critical review. Phytomedicine
2003;10:66­86.
[21] Burger RA, Torres AR, Warren RP, Caldwell VD, Hughes BG. Echinacea-induced
cytokine production by human macrophages. Int J Immunopharmacol 1997;19:
371­9.
[22] Clifford LJ, Nair MG, Rana J, Dewitt DL. Bioactivity of Alkamides Isolated from
Echinacea purpurea (L.) Moench. Phytomedicine 2002;9:249­53.
[23] Muller-Jakic B, Breu W, Probstle A, Redl K, Greger H, Bauer R. In vitro inhibition of
cyclooxygenase and 5-lipoxygenase by alkamides from Echinacea and Achillea
species. Planta Med 1994;60:37­40.
[24] Raso GM, Pacilio M, Di Carlo G, Esposito E, Pinto L, Meli R. In-vivo and In-vitro antiinflammatory
effect of Echinacea purpurea and Hypericum perforatum. J Pharm
Pharmacol 2002;54:1379­83.
[25] Hinz B, Woelkart K, Bauer R. Alkamide from Echinacea inhibit cyclooxygenase-2
activity in human neuroglioma cells. Biochem Biophy Res Comm 2007;360:441­6.
[26] Gertsch J, Schoop R, Kuenzle U, Suter A. Echinacea alkylamides modulate TNFgene
expression via cannabinoid receptor CB2 and multiple signal transduction
pathways. FEBS Lett 2004;577:563­9.
[27] Raduner S, Majewska A, Chen JZ, Xie XQ, Hamon J, Faller B, et al. Alkylamides from
Echinacea are a new class of cannabinomimetics. J Biol Chem 2006;281:
14192­206.
[28] MacMicking J, Xie QW, Nathan C. Nitric oxide and macrophage function. Annu Rev
Immunol 1997;15:323­50.
[29] Maines MD. Heme oxygenase: function, multiplicity, regulatory mechanisms, and
clinical application. FASEB J 1988;2:2557­68.
[30] Gabriel KR. The biplot graphical display of matrices with application to principal
component analysis. Biometrika 1971;58:453­67.
[31] Gower JC, Hand DJ. Biplots. London: Chapman & Hall; 1996.
[32] Chen CH. Generalized association plots for information visualization: the
applications of the convergence of iteratively formed correlation matrices. Stat
Sin 2002;12:1­23.
[33] Wu HM, Tien YJ, Chen CH. GAP: a graphical environment for matrix
visualization and cluster analysis. Computational Stat Data Analysis 2008, in
press (doi:10.1016/j.csda.2008.09.029).
[34] Bauer R, Remiger P, Wagner H. Alkamides from roots of Echinacea purpurea.
Phytochemistry 1988;27:2339­42.
[35] Chiang YM, Lo CP, Chen YP, Wang SY, Yang NS, Kuo YH, et al. Ethyl caffeate
suppresses NF-B activation and its downstream inflammatory mediators, iNOS,
COX-2, and PGE2 in vitro or in mouse skin. Br J Pharmacol 2005;146:352­63.
[36] Scudiero DA, Shoemaker RH, Paull KD, Monks A, Tierney S, Nofziger TH, et al.
Evaluation of a soluble tetrazolium/formazan assay for cell growth and drug
sensitivity in culture using human and other tumor cell lines. Cancer Res 1988;48:
4827­33.
[37] Lo AH, Liang YC, Lin-Shiau SY, Ho CT, Lin JK. Carnosol, an antioxidant in rosemary,
suppresses inducible nitric oxide synthase through down-regulating nuclear
factor-B in mouse macrophages. Carcinogenesis 2002;23:983­91.
[38] Bauer R, Remiger P, Wagner H. Alkamides from roots of Echinacea angustifolia.
Phytochemistry 1989;28:505­8.
[39] Bauer R, Khan IA, Wray V, Wagner H. Two acetylenic compounds from Echinacea
pallida roots. Phytochemistry 1987;26:1198­200.
[40] Bauer R, Remiger P, Wray V, Wagner HA. Germacrene alcohol from fresh aerial
parts of Echinacea purpurea. Plant Med 1988;54:478­9.
[41] Tien YJ, Lee YS, Wu HM, Chen CH. Methods for simultaneously identifying
coherent local clusters with smooth global patterns in gene expression profiles.
BMC Bioinformatics 2008;9:155.
[42] Hotelling H. Analysis of a complex of statistical variables into principal
components. J Educ Psychol 1931;24:417-441, 498-520.
[43] Hotelling H. Simplified calculation of principal components. Psychometrika
1936;1:27­35.
14 C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS
[44] Harman HH. Modern Factor Analysis. Chicago: Univ. of Chicago Press; 1967.
[45] Malazdrewich C, Thumbikat P, Abrahamsen MS, Maheswaran SK. Pharmacological
inhibition of Mannheimia haemolytica lipopolysaccharide and leukotoxin-induced
cytokine expression in bovine alveolar macrophages. Microb Pathog 2004;36:
159­69.
[46] Barrios-Rodiles M, Tiraloche G, Chadee K. Lipopolysaccharide modulates
cyclooxygenase-2 transcriptionally and posttranscriptionally in human macrophages
independently from Endogenous IL-1 and TNF-. J Immunol 1999;163:
963­9.
[47] Otterbein LE, Bach FH, Alam J, Soares M, Lu HT, Wysk M, et al. Carbon monoxide
has anti-inflammatory effects involoving the mitogen-activated protein kinase
pathway. Nat Med 2000;6:422­8.
[48] Morse D, Pischke SE, Zhou Z, Davis RJ, Flavell RA, Loop T, et al. Suppression of
inflammatory cytokine production by carbon monoxide involves the JNK pathway
and AP-1. J Bio Chem 2003;278:36993­8.
[49] Oh GS, Pae HO, Choi BM, Chae SC, Lee HS, Ryu DG, et al. 3-Hydroxyanthranilic
acid, one of metabolites of tryptophan via indoleamine 2,3-dioxygenase
pathway, suppresses inducible nitric oxide synthase expression by enhancing
heme oxygenase-1 expression. Biochem Biophys Res Commun 2004;320:
1156­62.
[50] Abuarqoub H, Foresti R, Green CJ, Motterlini R. Heme oxygenase-1 mediates the
anti-inflammatory actions of 2-hydroxychalcone in RAW 264.7 murine macrophages.
Am J Physiol Cell Physiol 2005;290:1092­9.
[51] Lee SH, Seo GS, Kim JY, Jin XY, Kim HD, Sohn DH. Heme oxygenase-1 mediates
anti-inflammatory effects of 2,4,6-Tris (methoxymethoxy) chalcone. Eur J
Pharmacol 2006;532:178­86.
[52] Kang HM, Saltveit ME. Antioxidant capacity of lettuce leaf tissue increases after
wounding. J Agric Food Chem 2002;50:7536­41.
[53] Choi YJ, Tomás-Barberán FA, Saltveit ME. Wound-induced phenolic accumulation
and browning in lettuce (Lactuca sativa L.) leaf tissue is reduced by exposure to nalcohol.
Postharvest Biol Tech 2005;37:47­55.
[54] Reyes LF, Cisneros-Zevallos L. Wounding stress increases the phenolic content and
antioxidant capacity of purple-flesh potatos (Solanum tuberosum L.). J Agric Food
Chem 2003;51:5296­300.
15C.-C. Hou et al. / Journal of Nutritional Biochemistry xx (2010) xxx­xxx
ARTICLE IN PRESS