Abstract!

Rhodiolae Crenulatae Radix et Rhizoma (Rhodio-la), the root and rhizome of Rhodiola crenulata(Hook. f. et Thoms.) H. Ohba, has been used as atraditional Chinese medicine (TCM) to increasethe body resistance against hypoxia in mountainsickness. The mechanism of this adaptogenicproperty deriving from Rhodiola, however, hasnot been revealed. Erythropoietin (EPO) is anerythrocyte-specific hematopoietic hormone thatincreases the production of red blood cells: thishormone is a crucial factor in regulating the bodybalance in responding to hypoxia. In cultured kid-ney fibroblasts (HEK293T), application of waterextract deriving from Rhodiola induced the ex-pression of EPO both in mRNA and protein levels.The activation of the hypoxia response element

(HRE) located on the promoter region of the EPOgene is one of the mechanisms accounting fortranscriptional activation. In addition, the Rhodio-la-induced EPO expression was triggered by anincrease of hypoxia-inducible factor-1α (HIF-1α)protein, via the reduction of HIF-1α degradationbut not the induction of HIF-1αmRNA. Moreover,the same EPO induction effect by Rhodiolawas al-so observed in cultured liver cells since liver is an-other vital organ to provide EPO regulation apartfrom the kidney. These results therefore elucidateone of the molecular mechanisms of this herb inmediating the anti-hypoxia function.

Supporting information available online athttp://www.thieme-connect.de/ejournals/toc/plantamedica

The Extract of Rhodiolae Crenulatae Radixet Rhizoma Induces the Accumulationof HIF-1α via Blocking the Degradation Pathwayin Cultured Kidney Fibroblasts

Authors Ken Y.Z. Zheng, Ava J.Y. Guo, Cathy W.C. Bi, Kevin Y. Zhu, Gallant K.L. Chan, Qiang Fu, Sherry L. Xu,Janis Y.Z. Zhan, David T.W. Lau, Tina T.X. Dong, Roy C.Y. Choi, Karl W.K. Tsim

Affiliation Section of Marine Ecology and Biotechnology, Division of Life Science and Center for Chinese Medicine,The Hong Kong University of Science and Technology, Hong Kong, China

Key wordsl" Rhodiola crenulata (Hook. f.

et Thoms.) H. Ohbal" Crassulaceael" hypoxia‑inducible factorl" erythropoietin

received October 22, 2010revised Nov. 17, 2010accepted Nov. 19, 2010

BibliographyDOI http://dx.doi.org/10.1055/s-0030-1250627Published online December 14,2010Planta Med 2011; 77: 894–899© Georg Thieme Verlag KGStuttgart · New York ·ISSN 0032‑0943

CorrespondenceDr. Karl W.K. TsimSection of Marine Ecologyand BiotechnologyDivision of Life ScienceThe Hong Kong Universityof Science and TechnologyClear Water Bay RoadHong KongChinaPhone: + 85223587332Fax: + 85223581559botsim@ust.hk

894

Zheng KYZ et al. The Extract of… P

Original Papers

Introduction!

Rhodiolae Crenulatae Radix et Rhizoma (Rhodio-la) is the root and rhizome of Rhodiola crenulata(Hook. f. et Thoms.) H. Ohba. (Crassulaceae),which is a plant growing at high altitudes in Asiaand Europe. Rhodiola, known as “arctic root” or“golden root”, has been used in the traditionalmedical system, in particular for those people liv-ing in high altitude. The therapeutic effects of thisspecialized herb include the increase of resistanceagainst different chemical, biological, and physi-cal stresses, as well as prevention of mountainsickness [1,2]. In previous studies, the intake ofRhodiola in animals was found to possess a potentadaptogenic activity in response to the exposureto multiple stress conditions [3]. The adaptogenicactivity of Rhodiola was found to be enhanced inthe water extract and the water soluble chemicalsto function so as to improve the anabolic state [4],i.e., increase the oxygen uptake in the body. How-

lanta Med 2011; 77: 894–899

ever, the mechanism of this adaptogenic activityof Rhodiola has not been fully revealed yet.Erythropoietin (EPO), an erythrocyte-specifichematopoietic hormone used to improve staminaby increasing the production of red blood cells, ismajorly produced by the kidney and the liver [5].Failure to increase the amount of circulatingblood EPO under hypoxic stress could lead to ane-mia [6]. Here, we hypothesize that the productionof EPO, induced by Rhodiola, could be a critical fac-tor to account for the adaptogenic activity againsthypoxia stress. The transcriptional regulation ofEPO is controlled by the hypoxia-inducible factor(HIF). Upon hypoxic stimulation, the activated HIFbinds onto the hypoxia response element (HRE)located on the promoter region of the EPO gene:the binding of HIF onto HRE, subsequently, trig-gers the transcription of the EPO gene. HIF is aheterodimeric transcription factor containing aninducible HIF-1α subunit and a constitutive HIF-1β subunit: this is an oxygen sensor that plays acentral role in themaintenance of oxygen homeo-

895Original Papers

stasis in various tissues [7]. In addition, the regulation of HIF-1αunder several pathologic conditions such as hypoxia or ischemiacould be mediated by transcriptional and/or translational con-trols [8,9]. Under normoxic conditions, HIF-1α is proline-hydrox-ylated leading to a conformational change that promotes thebinding to von Hippel Lindau protein (VLH) E3 ligase complexand subsequently causes ubiquitination and proteasomal degra-dation [10]. Under hypoxic conditions, or in the presence ofchemical hydroxylase inhibitors (desferrioxamine; DFO), thedegradation of HIF-1α could be reduced. As a result, the hypoxiaor DFO treatment could upregulate the expressions of HIF-1α andEPO. Up to now, the participation of EPO and HIF in the adapto-genic effect of Rhodiola has not been investigated. Here, we pro-vided different lines of evidence to support the notion that Rho-diola could induce the expression of EPO in cultured kidney cellsor in liver cells via the reduction of HIF degradation.

Materials and Methods!

Plant materials and preparation of RhodiolaThe roots and rhizomes of R. crenulatawere collected from Tibet,China, in September 2008. The roots were authenticated by oneof the authors, Dr. Tina Dong, during the field collection. A vouch-er specimen (# RHO-08-9-1) was deposited in the Center for Chi-nese Medicine, Hong Kong University of Science Technology,Hong Kong. The dried roots (300 g) were extracted twice withdistilled water at 100°C for 2 hours. The water extract was em-ployed here because the previous study had shown that theadaptogenic activity deriving from Rhodiola was water soluble[4]. The extract was collected and centrifuged at 3000 × g for10min. The supernatant was freeze-dried (yield, 51.4 g) and keptat −20°C. The dried extract was dissolved in Hankʼs balanced saltsolution (HBSS) and filtered through a 0.22 µm filter before use.

Standardization of extract by HPLCSalidroside was purchased from the National Institute for theControl of Pharmaceutical and Biological Products (Beijing, Chi-na), and the purity was higher than 98.0%. AR- and HPLC-gradereagents were purchased from Merck. A Waters HPLC systemconsisting of a 600 pump, a 717 auto-sampler, and a UV/VIS Pho-todiode Array 2996 Detector was used for all analyses. Chromato-graphic separations were carried out on an Agilent EclipseXDB‑C18 column (4.6mm× 250mm, 5 µm) with 0.1% phosphor-ic acid (as solvent A) and acetonitrile (as solvent B) in the mobilephase at a flow rate of 1.0mL/min at room temperature. A linearelution was 82% A and 18% B at 0–60min for the HPLC finger-print of Rhodiola extract. Ten µL (after a 0.45 µm Millipore filter)of the sample was injected, and the signals were detected at225 nm with UV detection. The HPLC fingerprint of Rhodiolawater extract revealed the major peaks (including salidroside, amajor ingredient of Rhodiola) at 275 nm. In addition, the amountof salidroside in the standardized extract of Rhodiola was deter-mined at 225 nm.

Cell cultureHuman embryonic kidney fibroblast HEK293T cells (ATCC CRL-11268) and human hepatocellular carcinoma HepG2 cells (ATCCHB-8065) were obtained from the American Type Culture Collec-tion (ATCC). Cells were cultured in Dulbeccoʼs modified Eagleʼsmedium (DMEM), supplemented with 10% fetal bovine serum,100 units/mL penicillin, and 100 units/mL streptomycin in a hu-

midified CO2 (5%) incubator at 37°C. All culture reagents werepurchased from Invitrogen Technologies.

Quantitative real-time PCRFor the analyses of EPO and HIF-1αmRNA expression in HEK293Tcells, the cultures were treated with various drugs. Total RNAwasisolated by TRIzol reagent and reverse transcribed into cDNAs ac-cording to the manufacturerʼs instructions (Invitrogen). Real-time PCR was performed by using SYBR Green Master mix andROX reference dye according to the manufacturerʼs instructions(Applied Bioscience). The PCR primers were: 5′-ACT TTC CGCAAA CTC TTC CG-3′ and 5′-TGA ATG CTT CCT GCT CTG GG-3′ forhuman EPO (330 bp; NM_000799.2) [11]; 5′-GCT TTA ACT TTGCTG GCC CCA GC-3′ and 5′-GCA GGG TCA GCA CTA CTT CGAAG‑3′ for human HIF-1α (221 bp; NM_001530.3); 5′-TGT GATGCC CTT AGA TGT CC-3′ and 5′-GAT AGT CAA GTT CGA CCG TC-3′for 18S rRNA (320 bp; NR_003286). The 18S rRNAwas used as aninternal control in all cases. The SYBR green signal was detectedby an Mx3000ptm multiplex quantitative PCR machine (Strata-gene). Transcript levels were quantified by using the ΔΔCt valuemethod, where the values of target genes were normalized by18S rRNA in the same sample before the comparison. The PCRproducts were analyzed by gel electrophoresis, and the meltingcurve analysis was to confirm the specific amplification.

DNA construction and transfectionThe HRE (5′-TCG AGG CCC TAC GTG CTG TCT CAC ACA GCC TGTCTG ACG-3′) derived from the human EPO gene contains a highlyconserved HIF-1 binding site (5′-TAC GTG-3′) and other uniquecis-acting sequences (5′-CAC AG-3′) that are functionally essen-tial for hypoxic induction [12]. Six HREs were synthesized, con-catemerized and then cloned in tandem (head-to-tail orienta-tion) into pBI‑GL vectors (BD Biosciences Clontech) that had adownstream reporter of firefly luciferase gene: this vector wasnamed pHRE‑Luc. Cultured HEK293T or HepG2 cells were tran-siently transfectedwith pHRE‑Luc by the calcium phosphate pre-cipitation method [13]. The transfection efficiency was about80%, as determined by another control plasmid having a gene ofβ-galactosidase under a cytomergalovirus enhancer promoter.

Western blot and antibodiesCultured HEK239T or HepG2 cells were treated with differentdrugs and then collected in high salt lysis buffer (10mM HEPES,pH7.5, 1mM EDTA, 1mM EGTA, 1MNaCl, and 0.5% Triton X-100)containing 10 µg/mL leupeptin, 10 µg/mL aprotinin, and 5mMbenzamidine HCl. The lysate was added with gel electrophoresisbuffer (0.125M Tris-HCl, pH 6.8, 4% SDS, 20% glycerol, and 2%2‑mercaptoethanol), which was denatured at 100°C for 5min.The lysate was subjected to 10% SDS-PAGE andWestern blotting.The expressions of HIF-1α, hydroxylated HIF-1α (HIF-1α-OH),EPO, and α-tubulin (a loading control) were identified by West-ern blot analysis. The antibodies (in 1: 1000 to 2000) to HIF-1α,HIF-1α-OH, and HIF-2α (1 :1000; Cell Signaling); EPO (1:200;Santa Cruz Biotechnology); α-tubulin (1:50000; Sigma); HRP-conjugated anti-rabbit secondary (1 :5000), and anti-mouse sec-ondary (1:10000; Zymed Laboratories) were employed here.The immune complexes were visualized using the enhancedchemiluminescence (ECL) method (Amersham Biosciences). Theintensities of the bands in the control and different samples, runon the same gel and under strictly standardized ECL conditions,were compared on an image analyzer, using a calibration plot

Zheng KYZ et al. The Extract of… Planta Med 2011; 77: 894–899

Fig. 1 Rhodiola extractinduces the expressionsof EPOmRNA and pro-tein in culturedHEK293T cells. A Cul-tured HEK293T cellswere treatedwith Rho-diola extract at differentdoses for 24 hours. TotalRNAs were isolatedfrom cultured HEK293Tcells and then reversedtranscribed (RT) intocDNAs. The level of EPOmRNAwas revealed byreal-time PCR, while 18SrRNA (not shown) wasused as an internal con-trol for normalization.B Rhodiola extract in-duced the amount ofEPO protein. HEK293Tcultures were changedwith fresh serum for 3hours before the addi-tion of Rhodiola extract

896 Original Papers

constructed from a parallel gel with serial dilutions of one of thesamples.

Other assaysThe luciferase assay was performed by a commercial kit (Tropix,Inc.). In brief, cultures were lysed by a buffer containing 100mMpotassium phosphate buffer (pH 7.8), 0.2% Triton X-100, and1mM dithiothreitol. The luminescent reaction was quantified ina Tropix TR717 microplate luminometer, and the activity was ex-pressed as absorbance (up to 560 nm) per mg of protein. The lu-ciferase activity was normalized by the fluorescent intensity ofEGFP in the same amount of protein in each sample. The proteinconcentrations were measured routinely by Bradfordʼs method.Statistical testswere done using one-wayANOVA. Statistically sig-nificant changes were classed as [*] for p < 0.05, [**] for p < 0.01,and [***] for p < 0.001.

Supporting informationHPLC fingerprint and determination of salidroside from thewaterextract of Rhodiola, assessment of cytotoxicity on culturedHEK293T and HepG2 cells, and data on analysis of accumulationof HIF-2a in HEK293T cells are available as Supporting Informa-tion.

(1mg/mL) for 6–24hours. DFO (100 µM)was used as a positivecontrol. EPO (~37 kDa)and α-tubulin(~55 kDa) were re-vealed by using the cor-responding antibodies.α-Tubulin served as aninternal control (upperpanel). The lower panelshows the quantitationof the blots by a densi-tometer. Values are ex-pressed as the percent-age of increase to basalreading (control culture;DMSO at 0.2%) and asmean ± SDwith n = 4,each with triplicatesamples.* P < 0.05,** p < 0.01, and*** p < 0.001, as com-pared to the control.

Results!

Thewater extract of Rhodiola has been used to promote longevityand to increase the bodyʼs resistance to physical and mentalstresses [1–2]. Here, the water extract of Rhodiola was preparedand calibrated by HPLC fingerprint, and the amount of salidrosidewas determined (see Fig. 1S, Supporting Information). After sev-eral calibrations, a standardized Rhodiola extract was achieved,which should contain not less than 12.34 µg of salidroside per1mg of the dried extract. From the calculations of extraction effi-ciency, the yield of Rhodiola water extract was about17.13 ± 1.06% (n = 3). These parameters established the chemicalstandards of Rhodiolawater extract for the subsequent studies onbiological activities.In order to reveal the effect of Rhodiola on EPOmRNA expression,a standardized water extract of Rhodiola was applied onto cul-tured HEK293T for 24 hours. The treatment of Rhodiola extractup to 1mg/mL did not change the cell morphology and the cellnumber, as revealed by microscopic examination and cell viabil-ity assay, respectively (see Fig. 2S, Supporting Information). Theextract therefore had no cytotoxicity on the cultures. However,the expression of EPO mRNA was induced by the treatment ofRhodiola extract. Total RNA was collected from the treated cellsand subjected to real-time quantitative PCR analysis by usingspecific primers flanking EPO mRNA. The induction was in adose-dependent manner, with a peaked response by ~ 120% at~ 0.1mg/mL of Rhodiola extract (l" Fig. 1A). The amount of Rho-diola-induced EPO protein was tested by Western blotting. Theapplication of Rhodiola extract induced the expression of EPOprotein at ~ 37 kDa in cultured HEK293T cells (l" Fig. 1B, upperpanel): the maximal induction was a ~ 150% increase after 24hours of treatment (l" Fig. 1B, lower panel). DFO (> 95% purity;Sigma) served as a control here with over a 200% increase ofEPO protein after 12 hours of treatment.The activation of HRE, located on the EPO promoter, by Rhodiolavia HIF signaling pathway could be one of the molecular mecha-nisms to account for Rhodiola-induced EPO expression in

Zheng KYZ et al. The Extract of… Planta Med 2011; 77: 894–899

HEK293T cells. To reveal the transcriptional activity of HRE, a lu-ciferase-reporter construct (pHRE‑Luc; l" Fig. 2A, upper panel)containing six HREs derived from the EPO gene and tagged up-stream of a luciferase gene [12] was transfected into culturedHEK293T cells. DFO inhibits prolyl hydroxylases by chelating theFe atom that is present within the active site of these enzymes[14]. Thus, DFO was used here to validate the pHRE‑Luc express-ing fibroblasts. As expected, application of DFO in the cultured fi-broblasts caused a strong induction in the reporter activity undernormoxic conditions: the activation was in a time-dependentmanner, and the maximal induction was at ~ 12 hours of treat-ment (l" Fig. 2A). In parallel, Rhodiola extract induced pHRE‑Lucactivity in a time-dependent manner, but the induction magni-tude was much lower than that of DFO (l" Fig. 2A). In addition,the Rhodiola-induced luciferase activity was in a dose-dependent

Fig. 3 Rhodiola extract regulates HIF-1α and HIF-1α-OH, but not thetranscription of HIF-1αmRNA in cultured HEK293T cells. A Total RNAs wereisolated from Rhodiola-treated (12 hours) HEK293T cells and then reversedtranscribed (RT) into cDNAs for the detection of mRNAs encoding for HIF-1α (221 bp) by real-time PCR analysis. The 18S rRNA (320 bp; not shown)served as an internal control. An over-layering of mineral oil was used as apositive control. B HEK293T cultures were changed with fresh serum 3hours before the addition of DFO (100 µM) for 5–60minutes. HIF-1α andHIF-1α-OH (both at ~ 120 kDa) were revealed by specific antibodies. α-tu-bulin (~ 55 kDa) served as a loading control. C HEK293T cultures weretreated with different drugs as in B for a longer time. HIF-1α and HIF-1α-OHwere revealed by specific antibodies. D HEK293T cultures were treatedwith 0 to 2mg/mL of Rhodiola extract for 6 hours, and HIF-1α and HIF-1α-OH were analyzed by specific antibodies (upper panel). The lower panelshows the quantitation of the blots by a densitometer. Values are expressedas the percentage of increase/decrease to basal reading (control culture;DMSO at 0.2%), and as mean ± SD with n = 4, each with triplicate samples.* P < 0.05, ** p < 0.01, and *** p < 0.001, as compared to the control.

Fig. 2 The stimulationof HRE-mediated tran-scriptional activity byRhodiola extract andDFO in culturedHEK293T cells. A A lu-ciferase-reporter con-taining six HREs, namelyas pHRE‑Luc, was usedas a study tool (upperpanel). CulturedHEK293T cells, trans-fected with pHRE‑Luc,were treatedwith Rho-diola extract (1mg/mL)or DFO (100 µM) for 0 to24 hours. The cell ly-sates were subjected toluciferase assay tomea-sure the transcriptionalactivity driven by HRE(lower panel). B Rhodio-la extract (0–1mg/mL)was applied onto pHRE-Luc-expressed fibro-blasts for 12 hours. Thecell lysates were sub-jected to luciferase as-say. Values are ex-pressed as the percent-age of increase to basalreading (control culture;DMSO at 0.2%) and asmean ± SDwith n = 4,each with triplicatesamples.* P < 0.05,** p < 0.01, and*** p < 0.001, as com-pared to the control.

897Original Papers

manner, and the peak was reached at ~ 0.3mg/mL of Rhodiola ex-tract (l" Fig. 2B).To account for the regulatory mechanisms of Rhodiola-inducedHRE activation, the HIF-1α-mediated signals could be activated.The transcriptional control of HIF-1α was analyzed; however,the mRNA encoding HIF-1αwas not altered in the Rhodiola-treat-ed cultures (l" Fig. 3A). The induction of HIF-1α protein was de-termined in cultured HEK293T. In order to test the responsive-ness of HIF signaling in HEK293T cells, DFO was used to treat thecultures for 0 to 60min. DFO increased the level of HIF-1α (an ac-tive form) but decreased the level of HIF-1α-OH (a degradedform) (l" Fig. 3B). Thus, cultured HEK293T cells constitute a sen-sitive model for testing the HIF signaling.In Rhodiola-treated cultures, the amount of HIF-1αwas increasedin a time-dependent manner; however, the amount of HIF-1α-OH was decreased significantly (l" Fig. 3C). Serving as a positivecontrol, DFO robustly accumulated HIF-1α and reduced markedlythe expression of HIF-1α-OH under a longer time of the treat-ment (l" Fig. 3C). Further validations were confirmed by the doseresponses of Rhodiola extract in regulating both HIF-1α and HIF-1α-OH levels (l" Fig. 3D, upper panel). At 1mg/mL of Rhodiola ex-tract, a maximal induction of HIF-1α at ~ 800% and a significantreduction of HIF-1α-OH at ~ 70% were revealed (l" Fig. 3D, lowerpanel).

The EPO induction was also tested in cultured HepG2 liver cells,another cell model commonly used for hypoxia analysis. InpHRE‑Luc transfected HepG2 cells, Rhodiola extract induced theluciferase activity in a dose-dependent manner (l" Fig. 4A). Theinduction of HIF-1α protein was determined in cultured HepG2cells. Here, DFO was used to treat the cultures for 6 hours, whichincreased the level of HIF-1α but decreased the level of HIF-1α-OH (l" Fig. 4B). In parallel, the amount of HIF-1α was increasedwhile the amount of HIF-1α-OH was decreased significantly inthe Rhodiola-treated liver cells (l" Fig. 4C). These results forHepG2 cells were very similar to those for HEK293T cells.

Discussion!

The genus Rhodiola L. (Crassulaceae) comprises about 96 speciesfound in the alpine regions of Asia and Europe. In China there is atotal of 73 species, 2 subspecies, and 7 varieties [15]. Most nota-

Zheng KYZ et al. The Extract of… Planta Med 2011; 77: 894–899

Fig. 4 Rhodiola extractregulates the expres-sions of HIF-1α and HIF-1α-OH in culturedHepG2 cells. A Cul-tured HepG2 cells,transfected withpHRE‑Luc, were treatedwith the DFO (100 µM)and Rhodiola extract(0–1mg/mL) for 24hours. The cell lysateswere subjected to lucif-erase assay. B CulturedHepG2 cells were treat-ed with DFO (0–100 µM) for 6 hours.HIF-1α and HIF-1α-OHwere analyzed by spe-cific antibodies. α-Tubu-lin (~ 55 kDa) served asa loading control.C Cultured HepG2 cellswere treated with Rho-diola extract (0–1mg/mL) for 6 hours. HIF-1αand HIF-1α-OH were an-alyzed by specific anti-bodies (upper panel).The lower panel showsthe quantitation of theblots by a densitometer.Values are expressed asthe percentage of in-crease/decrease to ba-sal reading (control cul-ture; DMSO at 0.2%)and as mean ± SD withn = 4, each with tripli-cate samples.* p < 0.05, ** p < 0.01,and *** p < 0.001, ascompared to the con-trol.

898 Original Papers

bly, the root and rhizome of R. crenulata has the best clinical ap-plication and therefore has been accepted by the Pharmacopoeiaof China [16]. In recent years, different pharmacological activitiesand efficacies of clinical usages of the genus Rhodiola have beenreported, e.g. in adaptogenic stress and cardioprotection, and forits antidepressive and anxiolytic, anti-anoxia, antifatigue, anti-aging, anti-oxidant, antitumor, and antiradiation effects as wellas its enhancement of learning and memory [17]. Amongst theseknown functions, the treatment ofmountain sickness, or anti-an-oxia, is the most well-described function for Rhodiola [1,2]. In-deed, people living in high altitude, for example Tibetan in China,have taken Rhodiola as a food supplement to improve the anoxicproblem, as well as to improve their sleep. In line with the pre-vious functional studies of Rhodiola, we have demonstrated herethe anti-anoxic effect of Rhodiola in stimulating EPO expression,which therefore could explain the adaptogenic effect of Rhodiola.The induction of EPO expression mediated by Rhodiola is mainlya result of reduced degradation of HIF-1α. Here, we favor the no-tion that Rhodiola, similarly to the action of DFO, prevents the hy-drolyation of HIF-1α so as to increase the accumulation of HIF-1α

Zheng KYZ et al. The Extract of… Planta Med 2011; 77: 894–899

protein. This molecular explanation of a “blood” stimulating func-tion is in excellent agreement with the ancient Chinesemedicinaltheory for functional usages of Rhodiola. Therefore, the hemato-poietic function of Rhodiola could be considered as a new direc-tion in developing health food supplements to prevent anemiaor altitude stress.Over 30 chemicals have been isolated and identified from the ex-tracts of Rhodiola [18]. Flavonoids, phenylpropanoids, and organ-ic acids are considered as themajor active constituents in Rhodio-la, which include salidroside, tyrosol, pyrogallol, gallic acid, β-si-tosterol, crenulatin, kaempferol, and ellagic acid [19]. Amongstthese chemicals, salidroside has been reported to possess variouspharmacological properties including resisting anoxia and hav-ing anti-aging, anticancer, anti-inflammation, anti-oxidation,hepatoprotective, and cardioprotective effects [20]. In line withthis notion, salidroside has been shown to have a protective ef-fect against hypoxia-induced apoptosis of cardiomyocytes by in-creasing HIF-1α and VEGF expressions [20]. Thus, salidroside inthe Rhodiola extract might account for the EPO induction in thecultured kidney fibroblasts. Currently, we are testing this hy-pothesis in our cultures.The transcriptional regulation of EPO expression is being con-trolled mainly by HIF. In humans, the α subunit has three iso-forms, HIF-1α to HIF-3α. HIF-1α and HIF-2α are synthesized con-stitutively and play similar roles in the regulation of oxygen ho-meostasis, while HIF-3α is the more distantly related isoform[21–23]. However, the antihypoxia effect of HIF-1α and HIF-2αcould be varied in different cell types. For instance, Rhodiola ex-tract induced HIF-1α accumulation but decreased HIF-2α in ourcultured fibroblasts (see Fig. 3S, Supporting Information). Thus,the Rhodiola-induced EPO expression could be mainly mediatedby HIF-1α but not HIF-2α.

Acknowledgements!

This research was supported by grants from the Research GrantsCouncil of Hong Kong (HKUST 6419/06M, N_HKUST629/07,662608, 661110) and Croucher Foundation (CAS-CF07/08.SC03)to KWKT and DTWL.

References1 Brown RP, Gerbarg PL, Ramazanov Z. Rhodiola rosea: a phytomedicinaloverview. Herbalgram 2002; 56: 40–52

2 Kelly GS. Rhodiola rosea: a possible plant adaptogen. Altern Med Rev2001; 6: 293–302

3 Gupta V, Saggu S, Tulsawani RK, Sawhney RC, Kumar R. A dose depen-dent adaptogenic and safety evaluation of Rhodiola imbricata Edgew,a high altitude rhizome. Food Chem Toxicol 2008; 46: 1645–1652

4 Gupta V, Lahiri SS, Sultana S, Kumar R.Mechanism of action of Rhodiolaimbricata Edgew during exposure to cold, hypoxia and restraint (C–H–R) stress induced hypothermia and post stress recovery in rats. FoodChem Toxicol 2009; 47: 1239–1245

5 Dame C, Kirschner KM, Bartz KV, Wallach T, Hussels CS, Scholz H. Wilmstumor suppressor, Wt1, is a transcriptional activator of the erythropoi-etin gene. Blood 2006; 107: 4282–4290

6 Jelkmann W. Erythropoietin: structure, control of production, andfunction. Physiol Rev 1992; 72: 449–489

7 Semenza GL. Regulation of mammalian O2 homeostasis by hypoxia-in-ducible factor 1. Annu Rev Cell Dev Biol 1999; 15: 551–578

8 Jewell UR, Kvietikova I, Scheid A, Bauer C, Wenger RH, Gassmann M. In-duction of HIF-1α in response to hypoxia is instantaneous. FASEB J2001; 15: 1312–1314

9 Mu D, Jiang X, Sheldon RA, Fox CK, Hamrick SE, Vexler ZS, Ferriero DM.Regulation of hypoxia-inducible factor 1α and induction of vascular

899Original Papers

endothelial growth factor in a rat neonatal stroke model. Neurobiol Dis2003; 14: 524–534

10 Jaakkola P, Mole DR, Tian YM, Wilson MI, Gielbert J, Gaskell SJ, Krieg-sheim Av, Hebestreit HF, Mukherji M, Schofield CJ, Maxwell PH, PughCW, Ratcliffe PJ. Targeting of HIF-α to the von Hippel-Lindau ubiquity-lation complex by O2-regulated prolyl hydroxylation. Science 2001;292: 449–451

11 Gao QT, Cheung JK, Choi RC, Cheung AW, Li J, Jiang ZY, Duan R, Zhao KJ,Ding AW, Dong TT, Tsim KWK. A Chinese herbal decoction preparedfrom Radix Astragali and Radix Angelicae Sinensis induces the expres-sion of erythropoietin in cultured Hep3B cells. Planta Med 2008; 74:392–395

12 Post DE, Van-Meir EG. Generation of bidirectional hypoxia/HIF-respon-sive expression vectors to target gene expression to hypoxic cells. GeneTher 2001; 8: 1801–1807

13 Xie HQ, Choi RC, Leung KW, Siow NL, Kong LW, Lau FT, Peng HB, TsimKWK. Regulation of a transcript encoding the proline-rich membraneanchor of globular muscle acetylcholinesterase: the suppressive rolesof myogenesis and innervating nerves. J Biol Chem 2007; 282:11765–11775

14 Wang XH, Cavell BE, Syed Alwi SS, Packham G. Inhibition of hypoxia in-ducible factor by phenethyl isothiocyanate. Biochem Pharmacol 2009;78: 261–272

15 Li T, Zhang H. Identification and comparative determination of rhodio-nin in traditional tibetan medicinal plants of fourteen Rhodiola speciesby high-performance liquid chromatography-photodiode array detec-

tion and electrospray ionization-mass spectrometry. Chem Pharm Bull2008; 56: 807–814

16 Shao ML. Pharmacopoeia of the Peopleʼs Republic of China (Part I). Bei-jing: Chemical Industry Press; 2010: 144

17 Panossian A, Wikman G, Sarris J. Rosenroot (Rhodiola rosea): traditionaluse, chemical composition, pharmacology and clinical efficacy. Phyto-medicine 2010; 17: 481–493

18 Nakamura S, Li XZ, Matsuda H, Yoshikawa M. Bioactive constituentsfrom Chinese natural medicines. XXVIII. Chemical structures of acyclicalcohol glycosides from the roots of Rhodiola crenulata. Chem PharmBull 2008; 56: 536–540

19 Kucinskaite A, Pobłocka-Olech L, Krauze-Baranowska M, Sznitowska M,Savickas A, Briedis V. Evaluation of biologically active compounds inroots and rhizomes of Rhodiola rosea L. cultivated in Lithuania. Medic-ina 2007; 43: 487–494

20 Zhang J, Liu A, Hou R, Zhang J, Jia X, Jiang W, Chen J. Salidroside protectscardiomyocyte against hypoxia-induced death: a HIF-1α-activated andVEGF-mediated pathway. Eur J Pharmacol 2009; 607: 6–14

21 Yeo, EJ, Cho YS, Kim MS, Park JW. Contribution of HIF-1α or HIF-2α toerythropoietin expression: in vivo evidence based on chromatin im-munoprecipitation. Ann Hematol 2008; 87: 11–17

22 Schofield CJ, Ratcliffe PJ. Oxygen sensing by HIF hydroxylases. Nat RevMol Cell Biol 2004; 5: 343–354

23 Rankin EB, Biju MP, Liu Q, Unger TL, Rha J, Johnson RS, SimonMC, Keith B,Haase VH. Hypoxia-inducible factor-2 (HIF-2) regulates hepatic eryth-ropoietin in vivo. J Clin Invest 2007; 117: 1068–1077

Zheng KYZ et al. The Extract of… Planta Med 2011; 77: 894–899