Digoxin and ouabain induce P-glycoprotein by activating calmodulin kinase II and hypoxia-inducible factor-1α in human colon cancer cells

Toxicology and Applied Pharmacology xxx (2009) xxx–xxx

YTAAP-11593; No. of pages: 8; 4C:

Contents lists available at ScienceDirect

Toxicology and Applied Pharmacology

j ourna l homepage: www.e lsev ie r.com/ locate /ytaap

ARTICLE IN PRESS

Digoxin and ouabain induce P-glycoprotein by activating calmodulin kinase II andhypoxia-inducible factor-1α in human colon cancer cells

Chiara Riganti a,b,⁎,1, Ivana Campia a,1, Manuela Polimeni a, Gianpiero Pescarmona a,b,Dario Ghigo a,b, Amalia Bosia a,b

a Department of Genetics, Biology and Biochemistry, University of Torino, Via Santena, 5/bis, 10126, Torino, Italyb Research Center on Experimental Medicine (CeRMS), Via Santena, 5/bis, 10126, Torino, Italy

Abbreviations: Pgp, P-glycoprotein; HIF-1 α, hypoxicalmodulin kinase II; BAPTA, 1,2-bis(2-aminophenoxacid; KB-R7943, 2-[2-[4-(4-nitrobenzyloxy) phenyl]ethhydroxyethyl)]-N-(4-methoxybenzenesulfonyl)]amino-N-benzylamine; FURA-AM, 1-[2-(5-carboxyoxazol-2-yl)-6-aamino-5′-methylphenoxy)-ethane-N,N,N′,N′-tetraaceticfoetal bovine serum; BSA, bovine serum albumin.⁎ Corresponding author. Department of Genetics

University of Torino, Via Santena, 5/bis, 10126, Torino,E-mail address: [email protected] (C. Riganti).

1 The first two authors equally contributed to the wo

0041-008X/$ – see front matter © 2009 Elsevier Inc. Adoi:10.1016/j.taap.2009.07.026

Please cite this article as: Riganti, C., et alinducible factor-1α in human colon cancer

a b s t r a c t

Article history:Received 30 April 2009Revised 10 July 2009Accepted 23 July 2009Available online xxxx

Keywords:DigoxinOuabainCalmodulin kinase IIHypoxia-inducible factor-1αP-glycoproteinDoxorubicin

Digoxin and ouabain are cardioactive glycosides, which inhibit the Na+/K+-ATPase pump and in this waythey increase the intracellular concentration of cytosolic calcium ([Ca++]i). They are also strong inducers ofthe P-glycoprotein (Pgp), a transmembrane transporter which extrudes several drugs, including anticanceragents like doxorubicin. An increased amount of Pgp limits the absorption of drugs through epithelial cells,thus inducing resistance to chemotherapy. The mechanism by which cardioactive glycosides increase Pgp isnot known and in this work we investigated whether digoxin and ouabain elicited the expression of Pgp witha calcium-driven mechanism. In human colon cancer HT29 cells both glycosides increased the [Ca++]i andthis event was dependent on the calcium influx via the Na+/Ca++ exchanger. The increased [Ca++]ienhanced the activity of the calmodulin kinase II enzyme, which in turn activated the transcription factorhypoxia-inducible factor-1α. This one was responsible for the increased expression of Pgp, which activelyextruded doxorubicin from the cells and significantly reduced the pro-apoptotic effect of the drug. All theeffects of glycosides were prevented by inhibiting the Na+/Ca++ exchanger or the calmodulin kinase II. Thiswork clarified the molecular mechanisms by which digoxin and oubain induce Pgp and pointed out that theadministration of cardioactive glycosides may widely affect the absorption of drugs in colon epithelia.Moreover, our results suggest that the efficacy of chemotherapeutic agent substrates of Pgp may be stronglyreduced in patients taking digoxin.

© 2009 Elsevier Inc. All rights reserved.

Introduction

P-glycoprotein (Pgp), a transmembrane ATP-dependent effluxpump, which is encoded by themdr-1 gene, is physiologically presentin the apical membrane of different epithelial cells, such as colonmucosa, hepatic bile ducts and renal proximal tubules (Takara et al.,2006). Pgp recognizes endogenous metabolites and xenobiotics assubstrates, including anticancer drugs such as anthracyclines (likedoxorubicin), epipodophyllotoxins and Vinca alkaloids; antibiotics;antiretroviral drugs; analgesic and anti-inflammatory drugs; immuno-suppressive drugs; and anti-arrhythmic and cardioactive drugs, such

a-inducible factor-1α; CaMKII,y)ethane-N,N,N′,N′-tetraaceticyl]isothiourea; KN93, 2-[N-(2-(4-chlorocinnamyl)-N-methyl-minobenzofuran-5-oxy]-2-(2′-acid acetoxymethylester; FBS,

, Biology and Biochemistry,Italy. Fax: +39 11 670 5845.

rk.

ll rights reserved.

., Digoxin and ouabain inducells, Toxicol. Appl. Pharma

as digoxin and verapamil (de Lange, 2004). Due to the broadspectrum of substrates, the amount of Pgp in critical tissues, asintestine, liver and kidney, may strongly affect the drug absorptionand excretion, influencing the efficacy of therapy (Bodo et al., 2003;Geick et al., 2001). Moreover, the overexpression of Pgp in tumorcells is the main mechanism of the multidrug resistance (MDR), anintrinsic or acquired cross-resistance toward different chemothera-peutic agents, which represents the major obstacle to a successfulcancer therapy (Takara et al., 2006).

Interestingly, many substrates of Pgp also act as Pgp inducers, thusenhancing their own efflux and limiting the intracellular accumulation.For instance, rifampin induces Pgp in human colon cells, increasing themdr-1 gene transcription via the activation of the pregnane-X receptor(Geick et al., 2001; Haslam et al., 2008). Also digoxin is both substrateand inducer of Pgp, but its effect is independent from the pregnane-Xreceptor (Haslam et al., 2008). Notably, other two glycosides, ouabainand palitoxin, share with digoxin the property to induce Pgp(Brouillard et al., 2001). Cardioactive glycosides are strong inhibitorsof the Na+/K+-ATPase pump and increase the [Ca++]i concentration,by forcing the Na+/Ca++ exchanger to extrude Na+ in exchange withCa++ (Kaplan, 2002). It has been suggested that the enhancedexpression of mdr-1 gene elicited by digoxin and ouabain depends on

ce P-glycoprotein by activating calmodulin kinase II and hypoxia-col. (2009), doi:10.1016/j.taap.2009.07.026

2 C. Riganti et al. / Toxicology and Applied Pharmacology xxx (2009) xxx–xxx

ARTICLE IN PRESS

the increased [Ca++]i levels: indeed the calcium chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) abrogatesthe Pgp induction exerted by ouabain, whereas thapsigargin, whichincreases the [Ca++]i by inhibiting the sarcoplasmic/endoplasmicreticulum Ca++-ATPase, enhances the ouabain effect (Baudouin-Legros et al., 2003; Brouillard et al., 2001). However the mechanismsby which the increased intracellular Ca++ levels modulate thetranscription of mdr-1 gene are not yet clarified.

An increase of [Ca++]imay activate the transcription factor hypoxia-inducible factor-1 (HIF-1) (Hui et al., 2006; Yuan et al., 2005), whichcontrols several genes involved in cell growth, angiogenesis, matrixremodeling, glucose and iron metabolism, intracellular pH modulation(O'Donnell et al., 2006), and in Pgp up-regulation (Comerford et al.,2002). HIF-1 is composed of two subunits: the β subunit is constitutive,whereas the α subunit is rapidly degraded in normoxia, but becomesstable in hypoxia (O'Donnell et al., 2006).

In this work we investigated whether digoxin and ouabaininduce the transcription of mdr-1 gene in human colon cancer cellsby a [Ca++]i-driven activation of HIF-1α, thus increasing the Pgpactivity and the cell resistance to doxorubicin.

Materials and methods

Materials. Foetal bovine serum (FBS) and RPMI 1640 medium weresupplied by BioWhittaker (Verviers, Belgium); plasticware for cellculture was from Falcon (Becton Dickinson, Bedford, MA). 2-[2-[4-(4-nitrobenzyloxy) phenyl]ethyl]isothiourea (KB-R7943), 2-[N-(2-hydroxyethyl)]-N-(4-methoxybenzenesulfonyl)]amino-N-(4-chloro-cinnamyl)-N-methylbenzylamine (KN93) and YC-1 were purchasedfrom Calbiochem (La Jolla, CA). Electrophoresis reagents wereobtained from Bio-Rad (Hercules, CA). When not otherwisespecified, the other reagents were purchased from Sigma ChemicalCo (St. Louis, MO).

Cells. Human colon cancer HT29 cells and human liver cancerHepG2 cells were cultured in RPMI 1640 supplemented with 10% FBS,1% penicillin/streptomycin, 1% L-glutamine and maintained in ahumidified atmosphere at 37 °C, 5% CO2 and 20% O2. To culture themin hypoxic conditions, cells were grown for 24 h in a humidifiedatmosphere at 37 °C, 5% CO2 and 3% O2.

[Ca++]i measurement. Cells were grown on sterile glass coverslipsunder the experimental conditions indicated in the Results section,washed twice with PBS and incubated for 10 min at 37 °C in Hepes–Cabuffer (10 mM Hepes, 145 mM NaCl, 1 mM CaCl2, 5 mM KCl, 1 mMMgSO4, and 10 mM glucose, pH 7.4), with 10 μM of calcium-sensitivefluorescent probe 1-[2-(5-carboxyoxazol-2-yl)-6-aminobenzofuran-5-oxy]-2-(2′-amino-5′-methylphenoxy)-ethane-N,N,N′,N′-tetra-acetic acid acetoxymethylester (FURA-AM). After FURA-AM loading,coverslips were washed with Hepes–Ca buffer and firmly positionedin a quartz cuvette (1 cm) containing 1 ml of Hepes–Ca. The cuvetteholder was thermostated at 37 °C and the fluorescence of coverslipswas measured in a Perkin Elmer LS-5 spectrofluorimeter (PerkinElmer) for 30min. Excitation and emissionwavelengths were 490 and530 nm, respectively. Calculation of [Ca++]i levels was performed aspreviously described (Hallam et al., 1984). The fluorescence of Ca++-saturated dye (Fmax), obtained by treating cells with 10 μMionomycin in Hepes–Ca buffer, was taken as the maximal emission.2 mM MnCl2 was then added to displace Ca++ from FURA and toobtain the value of FURA autofluorescence (Fmin) alone. To preventthe [Ca++]i increase, HT29 cells were pre-incubated for 3 h with10 μM BAPTA acetoxymethylester (BAPTA-AM) in order to load themwith the Ca++ chelator BAPTA, then washed with PBS, and subjectedto the same procedure of the other experimental points. To evaluatewhether the [Ca++]i increase may induce a pro-apoptotic response,the cytosolic release of cytochrome c was assessed: mitochondrial

Please cite this article as: Riganti, C., et al., Digoxin and ouabain induinducible factor-1α in human colon cancer cells, Toxicol. Appl. Pharma

and cytosolic fractions were separated as described (Wibom et al.,2002), then 10 μg from each fraction was subjected to SDS-PAGE andprobed with an anti-cytochrome c antibody (diluted 1:1,000 in PBS–BSA 1%, from Becton Dickinson).

Calmodulin kinase II (CaMKII) activity. CaMKII activity was mea-sured using the CycLex Calmodulin kinase II Assay Kit (CycLex Co.,Nagano, Japan), a single site binding immunoassay. Cells werecultured in 35-mm diameter Petri dishes, washed with ice-cold PBSand lysed in 0.2 ml lysis buffer (50 mM Tris–HCl, 0.1% Triton X-100,1 mM EDTA, 1 mM EGTA, 2 mM NaF, 2 mM NaVO4, 0.5 μg/mlleupeptin, 1 μg/ml pepstatin, 0.2 mM phenylmethylsulfonyl fluoridePMSF and 10 mM β-mercaptoethanol, pH 8.0). Samples weresonicated on crushed ice with two 10 s bursts and centrifuged at13,000 ×g for 5 min at 4 °C. Supernatants were treated following themanufacturer's instructions and the protein content was measured.Samples were diluted 1:10 in the kinase buffer provided with the kit,containing 20 mM ATP, and incubated for 60 min at 30 °C in 96-wellplates, pre-coated with the recombinant peptide Syntide-2, whichcontains a unique target sequence for CaMKII. Wells were washed 5times with 2% (v/v) Tween-20, and 100 μl of the horseradishperoxidase-conjugated anti-phospho Syntide-2 monoclonal antibodywas added. After 60 min of incubation at room temperature, sampleswere washed again, and 100 μl of the chromogenic substrate tetra-methylbenzidine was added. After 15 min of incubation at roomtemperature, the reaction was stopped with 100 μl of 0.5 N H2SO4 andthe absorbance was read at 450 nm, using a Packard EL340microplatereader (Bio-Tek Instruments, Winooski, VT). For each set ofexperiments, a titration curve was prepared, using serial dilution ofrecombinant CaMKII enzyme (CyClex Co.) in kinase buffer. Data wereexpressed as mU absorbance/mg cell proteins.

Western blot analysis. Cells were solubilized in boiling lysis buffer(125 mM Tris, 135 mM NaCl, 20% SDS and 10% glycerol, pH 7.4) andwhole-cell extracts were supplemented with the protease inhibitorcocktail set III (Calbiochem), 2 mM PMSF, 2.5 mM NaF and 1 mMNaVO4. 30 μg of proteins was separated by SDS-PAGE, transferred toPVDF membrane sheets (Immobilon-P, Millipore, Bedford, MA) andprobed with the following antibodies: anti-HIF-1α (from mouse,diluted 1:500 in PBS–BSA 1%, Becton Dickinson), anti-Pgp (fromrabbit, diluted 1:250 in PBS–BSA 1%, Santa Cruz Biotechnology, SantaCruz, CA), and anti-glyceraldehyde 3-phosphate dehydrogenase(anti-GAPDH, from rabbit, diluted 1:500 in PBS–BSA 1%, Santa CruzBiotechnology). After 1 h of incubation, the membrane was washedwith PBS–Tween 0.1% and subjected for 1 h to a peroxidase-conjugated anti-mouse or anti-rabbit antibody (diluted 1:3000 inPBS–Tween with Blocker Not-Fat Dry Milk 5%, Bio-Rad). Themembrane was washed again with PBS–Tween, and proteins weredetected by enhanced chemiluminescence (Immun-Star, Bio-Rad). Toassess HIF-1α phosphorylation, the whole-cellular lysate was immu-noprecipitated overnight with the rabbit polyclonal anti-HIF-1αantibody (diluted 1: 250, Santa Cruz Biotechnology) and theimmunoprecipitated proteins were separated by SDS-PAGE, trans-ferred to PVDFmembrane sheets and probedwith a biotin-conjugatedanti-phosphoserine antibody (diluted 1:1000 in TBS–BSA 3%, SigmaChemical Co.) for 1 h. The membrane was washed in TBS–Tween 0.1%and subjected for 1 h to a streptavidin- and horseradish peroxidase-conjugated polimer (diluted 1:10,000 in TBS–BSA 3%, Sigma ChemicalCo.). The membrane was washed again with TBS–Tween and proteinswere detected by enhanced chemiluminescence, as reported above.

Electrophoretic mobility shift assay (EMSA). Cells were plated in 60-mm diameter dishes at confluence and all the procedures for nuclearprotein extraction were performed at 4 °C using ice-cold reagents.Cells were mechanically scraped in PBS, washed and re-suspended inlysis buffer A (10 mM Hepes, 15 mM KCl, 2 mMMgCl2, 0.1 mM EDTA,

ce P-glycoprotein by activating calmodulin kinase II and hypoxia-col. (2009), doi:10.1016/j.taap.2009.07.026

Fig. 1. Effect of digoxin and ouabain on [Ca++]i in HT29 cells. (A) Cells were incubatedfor 24 h in the absence (0 μM) or presence of various concentrations (0.1 μM, 1 μM or10 μM) of digoxin (dgx) or ouabain (oua). (B) Cells were incubated for 3, 6, 24 or 48 h inthe presence (1 μM) of digoxin (dgx) or ouabain (oua). Cells were grown on sterile glasscoverslips, washed with PBS and incubated for 10 min in Hepes–Ca buffer containing10 μM FURA-AM. Fluorescence was detected throughout the first 30 min, as describedunderMaterials andmethods. Measurements were performed in duplicate and data arepresented as means±SE (n=4). Significance of each drug vs CTRL: ⁎pb0.005.

3C. Riganti et al. / Toxicology and Applied Pharmacology xxx (2009) xxx–xxx

ARTICLE IN PRESS

1 mM PMSF, 1 mM DTT, 10 μg/ml aprotinin, 2 μg/ml leupeptin, and0.1% NP-40, pH 7.6). This suspension was incubated for 10 min on icewith occasional vortexing, and centrifuged for 30 s at 13,000 ×g topellet nuclei, which were rinsed with 0.2 ml of wash buffer B (25 mMHepes, 2 mM KCl, 0.1 mM EDTA, 1 mM PMSF, 1 mM DTT, 10 μg/mlaprotinin, and 2 μg/ml leupeptin, pH 7.6) and incubated at 4 °C for20min. An equal volume of buffer C (25mMHepes, 0.1mMEDTA, and20% glycerol, pH 7.6) was added, themixwas centrifuged at 20,000 ×gand the supernatant was stored at −80 °C until used for EMSA. Thedetection of HIF-1α was performed on 10 μg nuclear extracts aspreviously described (Riganti et al., 2008).

Real time polymerase chain reaction (RT-PCR). Total RNA wasobtained as previously described (Chomczynski and Sacchi, 1987).5 μg of RNA was reverse-transcribed by 200 U of M-MLV reversetranscriptase (Invitrogen, Milan, Italy), in the presence of 40 U/μlRNAseOUT (Invitrogen). Quantitative RT-PCR was carried out usingIQ™ SYBR Green Supermix (Bio-Rad), according to themanufacturer'sinstructions. The same cDNA preparation was used for the quantita-tion of Pgp and GAPDH, used as a housekeeping gene. The sequencesof Pgp primers for quantitative RT-PCRwere: 5′-TGCTGGAGCGGTTCT-ACG-3′, 5′-ATAGGCAATGTTCTCAGCAATG-3′ (Invitrogen). Cycling forPgp was: 1 cycle at 94 °C for 2 min, followed by 45 cycles at 94 °C for30 s, annealing at 55 °C for 30 s, and extension at 72 °C for 30 s. Thesequences of GAPDH primers were 5′-GAAGGTGAAGGTCGGAGT-3′,5′-CATGGTGGAATCATATTGGAA-3′ (Invitrogen). Cycling for GAPDHwas: 1 cycle at 94 °C for 2 min, followed by 40 cycles at 94 °C for 30 s,annealing at 58 °C for 30 s, and extension at 72 °C for 30 s. The relativequantification of each sample was performed comparing the Pgp PCRproduct with the GAPDH product, using the Bio-Rad Software GeneExpression Quantitation (Bio-Rad).

Doxorubicin accumulation. Intracellular doxorubicin accumulationwas measured by a fluorimetric assay as described (Riganti et al.,2005). Excitation and emission wavelengths were 475 and 553 nm,respectively. A blankwas prepared in the absence of cells in each set ofexperiments and its fluorescence was subtracted from that measuredin the samples. Fluorescence was converted in ng doxorubicin/mg cellproteins, using a calibration curve prepared previously.

Annexin V assay. Cells were incubated in the experimentalconditions described under the Results section, then they were rinsedtwice with fresh PBS, detached with the Cell Dissociation Solution(Sigma Chemical Co.) and incubated for 10 min at room temperaturein 1 ml of binding buffer (100 mM Hepes, 140 mM NaCl, and 25 mMCaCl2, pH 7.5) containing 10 μM annexin V-fluorescein isothiocyanateconjugate (FITC). The cell suspensions were washed three times withfresh PBS and rinsed with 1 ml of binding buffer. The fluorescence ofeach sample was recorded using a FACSCalibur system (BectonDickinson). For each analysis 10,000 events were collected; the greenfluorescence (for annexin V-FITC) was selected using a 530 nm bandpass filter, the red fluorescence (for doxorubicin) was obtained with a640 nm longpass filter. The percentage of cells positive for annexin V,in the absence or presence of doxorubicin, was calculated by the CellQuest software (Becton Dickinson).

Statistical analysis. All data in text and figures are provided asmeans±SE. The results were analysed by a one-way Analysis ofVariance (ANOVA) and Tukey's test. pb0.05was considered significant.

Results

Digoxin and ouabain increase [Ca++]i in HT29 cells

After 24 h of incubation, digoxin and ouabain (0.1–10 μM)increased the [Ca++]i in HT29 cells as a function of dose (Fig. 1A).

Please cite this article as: Riganti, C., et al., Digoxin and ouabain induinducible factor-1α in human colon cancer cells, Toxicol. Appl. Pharma

When cardioactive glycosides were used at 1 μM, the raise of [Ca++]iwas not detectable after 3 h, increased after 6 h and was even moreevident after 24 h, whereas slightly decreased at 48 h (Fig. 1B). Nochanges in [Ca++]i levels were detected in cells pre-incubated for 3 hwith 10 μMBAPTA-AM before digoxin or ouabain (data not shown). Toassess whether the increase of [Ca++]i elicited by cardioactive glyco-sides induced the activation of an apoptotic process in HT29 cells, wemeasured the release of cytochrome c from mitochondria into thecytosol: after 24 h of incubation with 1 μM digoxin or ouabain, thecytochrome cwasnot released frommitochondria (data not shown). Atthe lights of these results, 1 μM digoxin or ouabain for 24 h was theminimal dose able to induce a significant increase of the [Ca++]i inHT29 cells, without exerting any pro-apoptotic effect. Therefore wechose these experimental conditions in all the subsequent experiments.

In many mammalian tissues the increase of [Ca++]i exerted bydigoxin or ouabain was due to the enhanced Ca++ influx through theNa+/Ca++ exchanger, driven by the increased [Na+]i (Kaplan, 2002).When theNa+/Ca++exchanger inhibitor KB-R7943was co-incubatedwith digoxin or ouabain, the cardioactive glycosides were not able toelicit any [Ca++]i increase (Fig. 2). This result suggested that the peakof [Ca++]i observed in HT29 cells treated with digoxin and ouabainwas due to the reverse activity of the Na+/Ca++ exchanger.

ce P-glycoprotein by activating calmodulin kinase II and hypoxia-col. (2009), doi:10.1016/j.taap.2009.07.026

Fig. 2. Effect of digoxin, ouabain and KB-R7943 on [Ca++]i. HT29 cells were cultured onsterile glass coverslips for 24 h in the absence (CTRL) or in the presence of digoxin (1 μM,DGX), ouabain (1 μM, OUA) and KB-R7943 (1 μM, KB), in different combinations.Subsequently the [Ca++]i was measured with FURA-AM in duplicate as reported underMaterials and methods. Data are presented as means±SE (n=4) vs CTRL: ⁎pb0.002;vs DGX or OUA: °pb0.005.

4 C. Riganti et al. / Toxicology and Applied Pharmacology xxx (2009) xxx–xxx

ARTICLE IN PRESS

Digoxin and ouabain enhance the CaMKII activity in HT29 cells, whereasKB-R7943 reverts their effect

Digoxin and ouabain increased the activity of CaMKII (Fig. 3) andthis effect was prevented by 3 h of pre-incubation with 10 μM BAPTA-AM (data not shown). When co-incubated with digoxin or ouabain,KB-R7943, which abolished the [Ca++]i peak induced by thecardioactive glycosides (Fig. 2), also prevented the increase of CaMKIIactivity. The KB-R7943 effect was superimposable to the effect of theCaMKII inhibitor KN93 (Fig. 3).

Digoxin and ouabain induce HIF-1α phosphorylation and activity in aNa+/Ca++ exchanger- and CaMKII-dependent way

The increase of [Ca++]i and CaMKII activity have been related tothe activation of the transcription factor HIF-1α in several cell lines(Hui et al., 2006; Yuan et al., 2005). Untreated HT29 cells showed anundetectable amount of HIF-1α protein (Fig. 4A) and activity

Fig. 3. Effect of digoxin, ouabain, KB-R7943 andKN93 onCaMKII activity. HT29 cellswerecultured for 24h in the absence (CTRL) or presence of digoxin (1 μM,DGX), ouabain (1 μM,OUA), KB-R7943 (1 μM, KB) and KN93 (10 μM, KN), differently combined. The cells werelysed and an ELISA assay for the activity of CaMKII was performed, as described underMaterials andmethods. Measurements were performed in duplicate. Data are presentedas means±SE (n=3) vs CTRL: ⁎pb0.05; vs DGX or OUA: °pb0.05.

Please cite this article as: Riganti, C., et al., Digoxin and ouabain induinducible factor-1α in human colon cancer cells, Toxicol. Appl. Pharma

(Fig. 4B). Both digoxin and ouabain induced the phosphorylation onserine of HIF-1α (Fig. 4A), which was accompanied by an increasedamount (Fig. 4A) and a clear nuclear translocation of the protein(Fig. 4B). When used alone, the Na+/Ca++ exchanger inhibitor KB-R7943 and the CaMKII inhibitor KN93 did not affect the HIF-1αphosphorylation, amount and activity. However, they both preventedthe phosphorylation and the nuclear translocation of HIF-1α elicitedby digoxin and ouabain, and reduced the total amount of HIF-1α tocontrol level (Fig. 4).

Fig. 4. Effect of cardioactive glycosides, KB-R7943 and KN93 on HIF-1α phosphorylationand activity. HT29 cells were incubated for 24 h in the absence (CTRL) or in the presenceof either digoxin (1 μM, DGX) or ouabain (1 μM, OUA), alone or together with KB-R7943 (1 μM, KB) and KN93 (10 μM, KN), then the following investigations wereperformed. (A) Western blotting detection of phospho(Ser)-HIF-1α (pHIF-1α) andtotal HIF-1α. To assess the pHIF-1α protein, whole-cell lysates were incubated with ananti-HIF-1α antibody, subsequently the immunoprecipitated proteins were separatedby SDS-PAGE and probed with an anti-phosphoserine antibody, as described under theMaterials and methods section.Western blot analysis for total HIF-1α was performedon the whole-cell lysates, using an anti-HIF-1α antibody. The expression of GAPDH, theproduct of a housekeeping gene, was used as a control of equal protein loading. Thefigure is representative of three experiments with similar results. (B) EMSA detection ofHIF-1α nuclear translocation. EMSA was performed on nuclear extracts as detailedunder Materials and methods. The lane marked with “+” was loaded with nuclearextracts obtained from HT29 cells incubated for 24 h in a humidified hypoxicatmosphere (3% O2, 5% CO2, 37 °C). This experimental condition was chosen as apositive control for maximal HIF-1α activation. In each experiment one lanewas loadedwith bidistilled water (-) in place of cellular extracts. The figure is representative ofthree experiments with similar results.

ce P-glycoprotein by activating calmodulin kinase II and hypoxia-col. (2009), doi:10.1016/j.taap.2009.07.026

Fig. 5. Effects of digoxin, ouabain, KB-R7943 and KN93 on Pgp expression. HT29 cellswere incubated for 24 h (for RT-PCR experiments) or 48 h (for Western blottinganalysis) in the absence (CTRL) or presence of either digoxin (1 μM, DGX) or ouabain(1 μM, OUA), alone or together with KB-R7943 (1 μM, KB) or KN93 (10 μM, KN), thenthe following investigations were performed. (A) Total RNA was extracted and reverse-transcribed by RT-PCR, as indicated under the Materials and methods section.Measurements were performed in triplicate and data are presented as means±SE(n=3) vs CTRL: ⁎pb0.02 vs DGX or OUA: °pb0.05 respectively. (B) To detect the Pgpprotein, cells were lysed and the whole-cellular lysate was subjected to Westernblotting (see Materials and methods). The expression of GAPDH, the product of ahousekeeping gene, was used as a control of equal protein loading. The figure isrepresentative of three experiments with similar results.

Fig. 6. Effects of digoxin, ouabain and YC-1 on Pgp expression. HT29 cells were grownfor 24 h (EMSA and RT-PCR) or 48 h (Western blotting) in the absence (CTRL) orpresence of digoxin (1 μM, DGX) or ouabain (1 μM, OUA), alone or together with theHIF-1α inhibitor YC-1 (5 μM, YC); then cells were subjected to the followinginvestigations. (A) EMSA was performed on nuclear extracts as reported underMaterials and methods. The lane marked with “+” was loaded with nuclear extractsderived from HT29 cells incubated for 24 h in a humidified hypoxic atmosphere (3%O2, 5% CO2, 37 °C), a condition under which the activation of HIF-1α was maximal. Ineach experiment one lane was loaded with bidistilled water (-) in place of cellularextracts. The figure is representative of three experiments with similar results. (B)Total RNA was extracted and reverse-transcribed by RT-PCR, as indicated underMaterials and methods. Measurements were performed in triplicate and data arepresented as means±SE (n=3) vs CTRL: ⁎pb0.02 vs DGX or OUA: °pb0.02respectively. (C) Cells were lysed and subjected to Western blotting to detect Pgpprotein (see Materials and methods). The expression of GAPDH, the product of ahousekeeping gene, was used as a control of equal protein loading. The figure isrepresentative of three experiments with similar results.

5C. Riganti et al. / Toxicology and Applied Pharmacology xxx (2009) xxx–xxx

ARTICLE IN PRESS

Digoxin and ouabain increase the Pgp expression and induce theresistance to doxorubicin with Na+/Ca++ exchanger- andCaMKII-dependent mechanism

HIF-1α is an effective activator of mdr-1 gene transcription(O'Donnell et al., 2006). In untreated HT29 cells, where HIF-1α wasnot active (Fig. 4B), Pgp was absent (Fig. 5). To the contrary, digoxinand ouabain strongly increased the Pgp mRNA (Fig. 5A) and protein(Fig. 5B), as they did with HIF-1α activity (Fig. 4B). When we addedthe HIF-1α inhibitor YC-1, the nuclear translocation of HIF-1α wasabolished in the presence of digoxin and ouabain (Fig. 6A); in parallel,YC-1 also prevented the up-regulation of Pgp mRNA (Fig. 6B) andprotein (Fig. 6C) elicited by the cardioactive glycosides.

By inducing Pgp expression, digoxin and ouabain reduced theintracellular accumulation of doxorubicin (Fig. 7A) and its cytotoxicity,as shown by the lower number of cells positive to annexin V in thepresence of doxorubicin (Fig. 7B). Again the inhibitor of Na+/Ca++

exchanger KB-R7943 and the inhibitor of CaMKII KN93 prevented theincrease of Pgp expression exerted by digoxin and ouabain (Figs. 5Aand B). Also the reduction of doxorubicin's accumulation and toxicityinduced by the cardioactive glycosides were abolished by KB-R7943and KN93 (Figs. 7A and B). None of the drugs induced significantcytotoxicity in the absence of doxorubicin (Fig. 7B).

Discussion

Cardioactive glycosides, such as digoxin and ouabain, which inhibitthe Na+/K+-ATPase pump and increase the intracellular [Na+]iconcentration, may enhance the Ca++ influx through the Na+/Ca++

Please cite this article as: Riganti, C., et al., Digoxin and ouabain induinducible factor-1α in human colon cancer cells, Toxicol. Appl. Pharma

exchanger. By doing so, they increase the [Ca++]i and exert a positiveinotropic effect on cardiac muscle (Kaplan, 2002). Ouabain anddigoxin have a similar structure and share most of the biologicalproperties; however the former has no current clinical employment,whereas the latter is one of the most widely used drug in heartdiseases (Adorisio et al., 2006).

Besides inhibiting the Na+/K+-ATPase, cardioactive glycosideshave many other effects: for instance, they affect the membranefluidity, modulate the transcription of several genes by activating NF-κB and AP1, inhibit the glycolysis, and increase the cellular synthesis

ce P-glycoprotein by activating calmodulin kinase II and hypoxia-col. (2009), doi:10.1016/j.taap.2009.07.026

Fig. 7. Effects of cardioactive glycosides, KB-R7943 and KN93 on doxorubicinaccumulation and cytotoxicity, detected as percentage of annexin V-positive cells.HT29 cells were cultured for 48 h in the absence (CTRL) or presence of either digoxin(1 μM, DGX) or ouabain (1 μM, OUA), alone or together with KB-R7943 (1 μM, KB) orKN93 (10 μM, KN). 5 μM doxorubicin was added to the medium for the last 24 h, thencells were subjected to the following investigations. (A) Doxorubicin intracellularaccumulation wasmeasured as described underMaterials andmethods. Measurementswere performed in duplicate and data are presented as means±SE (n=5) vs CTRL:⁎pb0.001 vs DGX or OUA:°pb0.005. (B) The positivity for annexin V-FITC wascalculated by FACS analysis (see Materials and methods for details) in cells incubatedwith digoxin, ouabain, KB-R7943 or KN93 plus doxorubicin (hatched bars), asdescribed above. In a parallel set of experiments, the analysis was performed underthe same experimental conditions in the absence of doxorubicin (open bars).Measurements were performed in duplicate and data are presented as means±SE(n=3) vs CTRL: ⁎pb0.01 vs DGX or OUA: °pb0.05.

6 C. Riganti et al. / Toxicology and Applied Pharmacology xxx (2009) xxx–xxx

ARTICLE IN PRESS

of reactive oxygen species (Prassas and Diamandis, 2008). The Srckinase, some phospholipase C isoforms, the Ras/Raf/MAP kinasepathway, and the ERK1-2 kinases are downstream effectors in thedigoxin and ouabain cell signaling (Prassas and Diamandis, 2008;Schoner and Scheiner-Bobis, 2007). Interestingly, cardioactive glyco-sides have been reported to increase the expression of pro-apoptoticmolecules, such as Fas ligand (Raghavendra et al., 2007) and caspase 3(Winnicka et al., 2008), and to induce the cell cycle arrest (Kometianiet al., 2005; Prassas and Diamandis, 2008). Therefore they haverecently been proposed as potential anticancer drugs (Prassas andDiamandis, 2008; Winnicka et al., 2008).

In mammalian tissues digoxin, ouabain and other inhibitors ofNa+/K+-ATPase pump are known inducers of Pgp (Baudouin-Legroset al., 2003; Brouillard et al., 2001), an ATP-binding cassettetransmembrane protein, which actively extrudes several endogenousmetabolites, toxic compounds and drugs (de Lange, 2004). Anticancerdrugs, such as anthracyclines, epipodophyllotoxins andVinca alkaloids,

Please cite this article as: Riganti, C., et al., Digoxin and ouabain induinducible factor-1α in human colon cancer cells, Toxicol. Appl. Pharma

are substrates of Pgp, whose expression induces a MDR phenotype intumors and reduces the efficacy of chemotherapy (Takara et al., 2006).Pgpmay be constitutively present in solid tissues ormay be induced bythe substrates themselves, like digoxin (Haslam et al., 2008). Themechanism by which cardioactive glycosides increase the Pgp is still amatter of debate. When [Ca++]i is reduced with BAPTA, digoxin andouabain lose their ability to increase Pgp in lung, liver and colon cancercells (Baudouin-Legros et al., 2003; Brouillard et al., 2001), suggestingthat the increased [Ca++]i may be responsible for the enhanced Pgpexpression elicited by cardioactive glycosides.

In our work we investigated how the [Ca++]i peak induced bydigoxin and ouabain increased Pgp in human colon cancer HT29 cellsand whether the cardioactive glycosides affected the efficacy of theanticancer drug doxorubicin, which is a Pgp substrate.

Our dose-dependence experiments showed that digoxin andouabain were able to increase the [Ca++]i at micromolar concentra-tions in HT29 cells. Notably, the intraluminal concentration of digoxinafter an oral administration is approximately in the micromolar rangeand at this concentration the drug was reported to induce the Pgpexpression and activity in human colon epithelia (Haslam et al., 2008).In HT29 cells treated with 1 μM digoxin or ouabain, the [Ca++]i levelsgradually raised in the time period of 24 h, and slightly decreased at48 h. Such a kineticsmay be in keepingwith the indirectmechanism bywhich cardioactive glycosides increase the [Ca++]i: it is reasonable tothink that also in colon, aswell as in cardiomyocytes, the raising [Ca++]ilevels are due to the calcium influx through the Na+/Ca++ exchanger,secondary to the Na+/K+-ATPase inhibition.

Since the increase of [Ca++]i may induce cells to enter inapoptosis, we measured the release of cytochrome c from mitochon-dria into the cytosol, taken as a sensitive index of the activation of theapoptosis intrinsic pathway. Neither digoxin nor ouabain induced therelease of cytochrome c after 24 h, suggesting that the higher [Ca++]ilevels observed in our experimental conditions were not sufficient toelicit apoptosis in HT29 cells or that the mechanisms of Ca++

extrusion and sequestration limited the calcium-induced damage inthese cells. Indeed the [Ca++]i did not further increase at 48 h in thepresence of digoxin and ouabain, suggesting that at this time point thecalcium extrusion and sequestration mechanisms counteracted theglycoside effect. When the Na+/Ca++ exchanger was blocked by KB-R7943, which does not allow any inward Ca++

flux (Liang et al.,2008), cardioactive glycosides were not able to increase the [Ca++]i inHT29 cells. Na+/Ca++ exchanger is expressed in several tissues(Annunziato et al., 2004), and our results suggested that it mayoperate also in HT29 cells. Therefore it is conceivable that theintracellular peak of calcium elicited by digoxin and ouabain wasdependent on the Na+/Ca++ exchanger activity.

Interestingly, two sesquiterpene lactones, artemisinin and parthe-nolide, also augmented the expression of Pgp in a calcium-dependentway in human colon cancer cells (Riganti et al., 2009). The calciumincrease was obtained through a different mechanism (i.e. theinhibition of the sarcoplasmic/endoplasmic reticulum Ca2+-ATPase)andwith a different kinetics, due to the faster andmore transient peakof [Ca++]i induced (Riganti et al., 2009). Sesquiterpene lactones havea chemical structure unrelated to cardioactive glycosides and have adifferent cellular target; the only common effect exerted by both thedrug families is the raising of intracellular calcium, followed by theup-regulation of mdr-1 gene. Therefore the property of inducing Pgpseems not exclusive of cardioactive glycosides, but rather shared withmany other compounds increasing the intracellular calcium.

An intracellular increase of Ca++ has pleiotropic effects on humancells. For instance, calcium may activate calmodulin and CaMKII,which is a fine sensor of the [Ca++]i increase and a polyedric mediatorof calcium signaling (Anderson, 2005). Interestingly, digoxin andouabain enhanced the activity of the CaMKII enzyme in HT29 cells. Onthe contrary, KB-R7943, which abrogated the [Ca++]i increase,completely prevented the cardioactive glycoside effect, and was as

ce P-glycoprotein by activating calmodulin kinase II and hypoxia-col. (2009), doi:10.1016/j.taap.2009.07.026

7C. Riganti et al. / Toxicology and Applied Pharmacology xxx (2009) xxx–xxx

ARTICLE IN PRESS

effective as the CaMKII inhibitor KN93. The inibition of Na+/K+-ATPase pump by ouabain has been reported to modulate the genetranscription in cardiomyocytes in a Ca++- and CaMK-dependentmanner (Huang et al., 1997), but to our knowledge the molecularbasis of this event is still unknown. Acting as a serine–threoninekinase, CaMKII can modulate the activity of several enzymes andtranscription factors, including HIF-1α (Yuan et al., 2005; Zhu et al.,2003). It has been suggested that CAMKII increases the HIF-1αamount with an indirect mechanism, i.e. by activating calcineurin,which is critical to prevent the degradation of HIF-1α (Liu et al.,2007). In adddition, when phosphorylated on serine by differentkinases, HIF-1α is more stabile (O'Donnell et al., 2006; Sodhi et al.,2001). We thus hypothesize that CaMKII may directly phosphorylateand activate HIF-1α in HT29 cells, as occurred in a rat pheochromo-cytoma cell line (Yuan et al., 2005).

HIF-1α was undetectable and not phosphorylated in untreatedHT29 cells, whereas became phosphorylated on serine in the presenceof digoxin and ouabain, which enhanced the CaMKII activity. Inparallel the amount and the nuclear translocation of HIF-1α wereincreased. Again these events were abolished by KB-R7943, whichprevented the CaMKII activation by calcium, and by KN93, whichdirectly inhibited the CaMKII enzyme. Our results suggest thatcardioactive glycosides activate HIF-1α in human colon cells with aNa+/Ca++ exchanger- and CaMKII-dependent mechanism, and thatthe phosphorylation by CaMKII stabilizes HIF-1α and promotes itsnuclear translocation. In human hepatoblastoma Hep3B cells and inhuman prostate cancer PC3 cells, which had a constitutively activeHIF-1α also under normoxia, nanomolar concentrations of digoxinand ouabain reduced the amounts of HIF-1α, by affecting the mRNAtranslation; this effect was evident in hypoxia, but less clear innormoxia and in cells transfected with an HIF-1α expression vector,probably because under these conditions tumor cells had a differentregulation of HIF-1α synthesis (Zhang et al., 2008). The differences inthe drugs concentration, in the experimental conditions (normoxia vshypoxia) and in the cell lines (HT29 cells are devoid of a constitutiveexpression of HIF-1α) may account for the discrepancies with ourresults. It is a quite common evidence that cardiac glycosides displayopposite effects in different experimental models, as a consequence oftheir pleiotropic effects on cell metabolism (Prassas and Diamandis,2008; Schoner and Scheiner-Bobis, 2007). Moreover HIF-1α may bevariably regulated at several levels and by several factors (O'Donnellet al., 2006; Sodhi et al., 2001): the prevalence of either factoraccounts for the cells amount and activity of HIF-1α. In ourexperimental model, digoxin and ouabain control the intracellularcalcium levels and the CaMKII activity, two critical factors whichpromote HIF-1α activation in HT29 cells (Riganti et al., 2009).

A hypoxia responsive element, which binds HIF-1α, has beenidentified in the promoter ofmdr-1 gene (Comerford et al., 2002). Theactivation of HIF-1α may explain why the inner core of solid tumors,which are frequently hypoxic and contain high amounts of HIF-1α,often express Pgp and are resistant to chemotherapy (O'Donnell et al.,2006). In HT29 cells the enhanced expression of Pgp induced bydigoxin and ouabain correlated with the activation of HIF-1α. Indeedin the presence of the HIF-1α inhibitor YC-1, the cardioactiveglycosides lose their ability to up-regulate the mdr-1 gene. Alsowhen the nuclear translocation of HIF-1α was prevented by the co-incubation with KB-R7943 or KN93, neither Pgp mRNA nor Pgpprotein expression were detectable. When highly expressed, Pgpextrudes more actively its substrates, limiting their intracellularaccumulation. Digoxin itself is more actively transported out of thehuman colon T84 cells, as a consequence of the increase of Pgp(Haslam et al., 2008). As an index of Pgp activity we measured theintracellular accumulation of doxorubicin, an anthracycline which iswidely used in clinical protocols and often represents the first-linetherapy against solid and hematological malignancies (Cortes-Funesand Coronado, 2007). Digoxin and ouabain strongly reduced the

Please cite this article as: Riganti, C., et al., Digoxin and ouabain induinducible factor-1α in human colon cancer cells, Toxicol. Appl. Pharma

doxorubicin intracellular content, and transformed the HT29 cells,which are sensitive to the pro-apoptotic effect of doxorubicin, in cellsresistant to the drug. However the resistance was corrected by KB-R7943 or KN93, which restored both the intracellular accumulationand the cytotoxicity of doxorubicin.

In summary, our work puts light on the molecular mechanism bywhich cardioactive glycosides increase Pgp expression, suggesting thatthe raising levels of [Ca++]i, due to the inward flux through the Na+/Ca++ exchanger, activate the CaMKII enzyme. CaMKII in turn mayfavor the activation of HIF-1α, which is responsible for the increasingexpression and activity of Pgp. We obtained superimposable results inthe human liver cancer HepG2 cell line (data not shown), suggestingthat digoxin and ouabain induce Pgp expression in human colon andliver cells with a common mechanism.

Our findings might have physiopathological consequences. Thenumber of patients taking digoxin for heart disease is high, andclinicians should take into account that the absorption and systemicdistribution of other drugs may be impaired in these patients, becauseof the digoxin-induced expression of Pgp in colon and liver. Moreover,cardioactive glycosides have pro-apoptotic effects in several solidtumor cell lines (Raghavendra et al., 2007; Winnicka et al., 2008) andenhance the damages induced by radiotherapy in transformed cells(Nasu et al., 2002). For these reasons they have recently proposed aspotential adjuvant drugs in cancer therapy (Prassas and Diamandis,2008; Zhang et al., 2008). Unfortunately, many anticancer drugs usedin chemotherapy protocols are substrates of Pgp. At the light of ourresults we hypothesize that the co-administration of digoxin mightincrease the efflux of chemotherapeutic agents from tumor cells. Theeffects of the interaction between doxorubicin and cardioactiveglycosides on cardiomyocytes have been sometimes contradictory:for example, the chronic administration of doxorubicin plus digoxin inNew Zealand rabbits reduces their overal survival when compared tothe administration of doxorubicin alone (Reeves et al., 1990); on thecontrary other works have shown that cardioactive glycosidesprevent the acute myocardial damage in animals and humans(Somberg et al., 1978; Whittaker and Al-Ismail, 1984). Interestinglysuch a protective effect of digoxin owes to the reduced doxorubicinuptake in cardiomyocytes (Somberg et al., 1978), a finding in keepingwith our results on colon and liver cancer cells. This variability ofeffects might depend on the different species evaluated, on theparameters considered to assess the cell damage and on the differentprotocols of drugs administration. Therefore future in vitro and in vivostudies seem mandatory to better elucidate the effects of cardioactiveglycosides on chemotherapy efficacy in cancer cells, as well as onchemotherapy-dependent side effects in non transformed tissues.

Conflict of interest statementThe authors declare that there are no conflicts of interest.

Acknowledgments

We are grateful to Costanzo Costamagna for the technicalassistance provided.

Funding source: This work was supported by grants from theFondazione Internazionale Ricerche Medicina Sperimentale (FIRMS),the Compagnia di San Paolo, the Regione Piemonte (Ricerca SanitariaFinalizzata 2007 and 2008, Turin, Italy) and the Ministero dell'Uni-versità e della Ricerca (Rome, Italy).

References

Adorisio, R., De Luca, L., Rossi, J., Gheorghiade, M., 2006. Pharmacological treatment ofchronic heart failure. Heart Fail. Rev. 11, 109–123.

Anderson, M.E., 2005. Calmodulin kinase signaling in heart: an intriguing candidatetarget for therapy of myocardial dysfunction and arrhythmias. Pharmacol. Ther.106, 39–55.

ce P-glycoprotein by activating calmodulin kinase II and hypoxia-col. (2009), doi:10.1016/j.taap.2009.07.026

8 C. Riganti et al. / Toxicology and Applied Pharmacology xxx (2009) xxx–xxx

ARTICLE IN PRESS

Annunziato, L., Pignataro, G., Di Renzo, F., 2004. Pharmacology of brain Na+/Ca2+

exchanger: frommolecular biology to therapeutic perspectives. Pharmacol. Rev. 56,633–654.

Baudouin-Legros, M., Brouillard, F., Tondelier, D., Hinzpeter, A., Edelman, A., 2003. Effectof ouabain on CFTR gene expression in human Calu-3 cells. Am. J. Physiol. CellPhysiol. 284, C620–C626.

Bodo, A., Bakos, E., Szeri, F., Varadi, A., Sarkadi, B., 2003. The role ofmultidrug transportersin drug availability, metabolism and toxicity. Toxicol. Lett. 140–141, 133–143.

Brouillard, F., Tondelier, D., Edelman, A., Baudouin-Legros, M., 2001. Drug resistanceinduced by ouabain via the stimulation of MDR1 gene expression in humancarcinomatous pulmonary cells. Cancer Res. 61, 1693–1698.

Chomczynski, P., Sacchi, N., 1987. Single-step method of RNA isolation by acidguanidinium thiocyanate–phenol–chloroform extraction. Anal. Biochem. 162,156–159.

Comerford, K.M., Wallace, T.J., Karhausen, J., Louis, N.A., Montalto, M.C., Colgan, S.P.,2002. Hypoxia-inducible factor-1-dependent regulation of themultidrug resistance(MDR1) gene. Cancer Res. 62, 3387–3394.

Cortes-Funes, H., Coronado, C., 2007. Role of anthracyclines in the era of targetedtherapy. Cardiovasc. Toxicol. 7, 56–60.

de Lange, E.C., 2004. Potential role of ABC transporters as a detoxification system at theblood-CSF barrier. Adv. Drug Deliv. Rev. 56, 1793–1809.

Geick, A., Eichelbaum, M., Burk, O., 2001. Nuclear receptor response elements mediateinduction of intestinal MDR1 by rifampin. J. Biol. Chem. 276, 14581–14587.

Hallam, T.J., Sanchez, A., Rink, T.J., 1984. Stimulus–response coupling in humanplatelets. Changes evoked by platelet-activating factor in cytoplasmic free calciummonitored with the fluorescent calcium indicator quin2. Biochem. J. 218, 819–827.

Haslam, I.S., Jones, K., Coleman, T., Simmons, N.L., 2008. Rifampin and digoxin inductionof MDR1 expression and function in human intestinal (T84) epithelial Cells. Br. J.Pharmacol. 154, 246–255.

Huang, L., Li, H., Xie, Z., 1997. Ouabain-inducedhypertrophy in cultured cardiacmyocytesis accompanied by changes in expression of several late response genes. J. Mol. Cell.Cardiol. 29, 429–437.

Hui, A.S., Bauer, A.L., Striet, J.B., Schnell, P.O., Czyzyk-Krzeska, M.F., 2006. Calciumsignaling stimulates translation of HIF-1alpha during hypoxia. FASEB J. 20, 466–475.

Kaplan, J., 2002. H. Biochemistry of Na,K-ATPase. Annu. Rev. Biochem. 71, 511–535.Kometiani, P., Liu, L., Ascari, A., 2005. Digitalis-induced signaling by Na+/K+-ATPase in

human breast cancer cells. Mol. Pharmacol. 67, 929–936.Liang, G.H., Kim, J.A., Seol, G.H., Choi, S., Suh, S.H., 2008. The Na+/Ca2+ exchanger

inhibitor KB-R7943 activates large-conductance Ca2+-activated K+ channels inendothelial and vascular smooth muscle cells. Eur. J. Pharmacol. 582, 35–41.

Liu, Y.V., Hubbi, M.E., Pan, F., McDonald, K.R., Mansharamani, M., Cole, R.N., Liu, J.O.,Semenza, G.L., 2007. Calcineurin promotes hypoxia-inducible factor 1α expressionby dephosphorylating RACK1 and blocking RACK1 dimerization. J. Biol. Chem. 282,37064–37073.

Nasu, S., Milas, L., Kawabe, S., Raju, U., Newman, R.A., 2002. Enhancement ofradiotherapy by oleandrin is a caspase-3 dependent process. Cancer Lett. 185,145–151.

O'Donnell, J.L., Joyce, M.R., Shannon, A.M., Harmey, J., Geraghty, J., Bouchier-Hayes, D.,2006. Oncological implications of hypoxia inducible factor-1alpha (HIF-1alpha)expression. Cancer Treat. Rev. 32, 407–416.

Please cite this article as: Riganti, C., et al., Digoxin and ouabain induinducible factor-1α in human colon cancer cells, Toxicol. Appl. Pharma

Prassas, I., Diamandis, E.P., 2008. Novel therapeutic applications of cardiac glycosides.Nat. Rev. Drug Discov. 7, 926–935.

Raghavendra, P.B., Sreenivasan, Y., Ramesh, GT., Manna, S.K., 2007. Cardiac glycosideinduces cell death via FasL by activating calcineurin and NF-AT, but apoptosisinitially proceeds through activation of caspases. Apoptosis 12, 307–318.

Reeves, W.C., Griffith, J.W., Wood, M.A., Whitesell, L., 1990. Exacerbation of doxorubicincardiotoxicity by digoxin administration in an experimental rabbit model. Int. J.Cancer 45, 731–736.

Riganti, C., Miraglia, E., Viarisio, D., Costamagna, C., Pescarmona, G.P., Ghigo, D., Bosia, A.,2005. Nitric oxide reverts the resistance to doxorubicin in human colon cancer cellsby inhibiting the drug efflux. Cancer Res. 65, 516–525.

Riganti, C., Doublier, S., Aldieri, E., Orecchia, S., Betta, P.G., Gazzano, E., Ghigo, D., Bosia,A., 2008. Asbestos induces doxorubicin resistance in MM98 mesothelioma cells viaHIF-1α. Eur. Respir. J. 32, 443–451.

Riganti, C., Doublier, S., Viarisio, D., Miraglia, E., Pescarmona, G., Ghigo, D., Bosia, A.,2009. Artemisinin induces doxorubicin resistance in human colon cancer cells viacalcium-dependent activation of HIF-1alpha and P-glycoprotein overexpression.Br. J. Pharmacol. 156, 1054–1066.

Schoner, W., Scheiner-Bobis, G., 2007. Endogenous and exogenous cardiac glycosides:their roles in hypertension, salt metabolism, and cell growth. Am. J. Physiol. CellPhysiol. 293, 509–536.

Sodhi, A., Montaner, S., Miyazaki, H., Gutkind, J.S., 2001. MAPK and Akt act cooperativelybut independently on hypoxia inducible factor-1alpha in rasV12 upregulation ofVEGF. Biochem. Biophys. Res. Commun. 287, 292–300.

Somberg, J., Cagin, N., Levitt, B., Bounous, H., Ready, P., Leonard, D., Anagnostopoulos, C.,1978. Blockade of tissue uptake of the antineoplastic agent, doxorubicin. J. Pharmacol.Exp. Ther. 204, 226–231.

Takara, K., Sakaeda, T., Okumura, K., 2006. An update on overcoming MDR1-mediatedmultidrug resistance in cancer chemotherapy. Curr. Pharm. Des. 12, 273–286.

Whittaker, J.A., Al-Ismail, S.A.D., 1984. Effect of digoxin and vitamin E in preventingcardiac damage caused by doxorubicin in acute myeloid leukemia. Br. Med. J. 288,283–284.

Wibom, R., Hagenfeldt, L., von Dobeln, U., 2002. Measurement of ATP production andrespiratory chain enzyme activities in mitochondria isolated from small musclebiopsy samples. Anal. Biochem. 311, 139–151.

Winnicka, K., Bielawski, K., Bilewaska, A., Surazynski, A., 2008. Antiproliferative activityof derivatives of ouabain, digoxin and proscillaridin A in human MCF-7 and MDA-MB-231 breast cancer cells. Biol. Pharm. Bull. 31, 1131–1140.

Yuan, G., Nanduri, J., Bhasker, C.R., Semenza, G.L., Prabhakar, N.R., 2005. Ca2++/calmodulin kinase-dependent activation of hypoxia inducible factor 1 transcrip-tional activity in cells subjected to intermittent hypoxia. J. Biol. Chem. 280,4321–4328.

Zhang, H., Qian, D.Z., Tan, Y.S., Lee, K.A., Gao, P., Ren, Y.R., Rey, S., Hammers, H., Chang, D.,Pili, R., Dang, C.V., Liu, J.O., Semenza, G.L., 2008. Digoxin and other cardiac glycosidesinhibit HIF-1α synthesis and block tumor growth. Proc. Natl. Acad. Sci. U. S. A. 105,19579–19586.

Zhu, W-H., Wang, S-Q., Chakir, K., Yang, D., Zhang, T., Brown, J.H., Devic, E., Kobilka, B.K.,Cheng, H., Xiao, R-P., 2003. Linkage of β1-adrenergic stimulation to apoptotic heartcell death through protein kinase A-independent activation of Ca2+/calmodulinkinase II. J Clin. Invest. 111, 617–625.

ce P-glycoprotein by activating calmodulin kinase II and hypoxia-col. (2009), doi:10.1016/j.taap.2009.07.026