cardiac glycosides induced toxicity in human cells ... · cardiac glycosides induced toxicity in...

Cardiac glycosides induced toxicity in human cells expressing �1-, �2-, or�3-isoforms of Na-K-ATPase

Marina Cherniavsky Lev, Steven J. D. Karlish, and Haim Garty†

Department of Biological Chemistry, Weizmann Institute of Science, Rehovot, Israel

Submitted 26 March 2015; accepted in final form 12 May 2015

Cherniavsky Lev M, Karlish SJ, Garty H. Cardiac glycosidesinduced toxicity in human cells expressing �1-, �2-, or �3-isoformsof Na-K-ATPase. Am J Physiol Cell Physiol 309: C126–C135, 2015.First published May 20, 2015; doi:10.1152/ajpcell.00089.2015.—TheNa�-K�-ATPase is specifically inhibited by cardiac glycosides, someof which may also function as endogenous mammalian hormones.Previous studies using Xenopus oocytes, yeast cells, or purifiedisoforms demonstrated that affinities of various cardiac glycosides forthree isoforms of the Na�-K�-ATPase (�1-�3�1) may differ, afinding with potential clinical implication. The present study investi-gates isoform selectivity and effects of cardiac glycosides on culturedmammalian cells under more physiological conditions. H1299 cells(non-small cell lung carcinoma) were engineered to express only one�-isoform (�1, �2, or �3) by combining stable transfection of iso-forms and silencing endogenous �1. Cardiac glycoside binding wasmeasured by displacement of bound 3H-ouabain. The experimentsconfirm moderate �1/�3:�2 selectivity of ouabain, moderate �2:�1selectivity of digoxin, and enhanced �2:�1 selectivity of syntheticderivatives (Katz A, Tal DM, Heller D, Haviv H, Rabah B, BarkanaY, Marcovich AL, Karlish SJD. J Biol Chem 289: 21153–21162,2014). Relative �2:�1 selectivity of digoxin vs. ouabain was alsomanifested by enhanced internalization of �2 in response to digoxin.Cellular proliferation assays of H1299 cells confirmed the patterns of�2:�1 selectivity for ouabain, digoxin, and a synthetic derivative andreveal a crucial role of surface pump density on sensitivity to cardiacglycosides. Because cardiac glycosides are being considered as drugsfor treatment of cancer, effects of ouabain on proliferation of 12cancer and noncancer cell lines, with variable plasma membraneexpression of �1, have been tested. These demonstrated that sensitiv-ity to ouabain indeed depends linearly on the plasma membranesurface density of Na�-K�-ATPase irrespective of status, malignantor nonmalignant.

cardiac glycosides; cell death; isoforms; Na-K-ATPase

THE NA�-K�-ATPASE (the Na� pump) is a ubiquitously expressedP-type ATPase that actively transports 3Na� ions out of cellsin exchange for 2K� ions per one molecule of ATP hydrolyzed(17). The pump consists of a catalytic �-subunit, �-subunit thatmodulates activity and stabilizes the protein, and an auxiliaryFXYD polypeptide. Four �-isoforms (�1- �4), three �-iso-forms (�1- �3), and seven FXYD isoforms exist in mammalianorganisms, resulting in considerable tissue-specific variabilityin pump structure and kinetics (21).

As can be expected from plasma membrane protein withsuch a vital function, the Na�-K�-ATPase is regulated by anumber of independent mechanisms (25). One such mode ofregulation involves inhibition by a family of inhibitors termedcardiac glycosides (CGs). Most CGs are derived from plants.

However, several so-called endogenous CGs, includingouabain, marinobufagenin, and digoxin, have been identified inmammalian tissues and are apparently synthesized by theadrenal gland (1). CGs such as digoxin are also clinicallyrelevant specific inhibitors of the Na�-K�-ATPase, used clas-sically to treat heart failure (9).

Relatively limited information exists regarding the specific-ity of different CGs for the 4 �-isoforms and 3 �-isoforms (�1,�2, �3, and �4 and �1, �2, and �3) or the 12 possiblecombinations of ��-complexes. The problem is that mostmammalian cells express more than a single �-isoform, anddiscrimination between pumps of different subunit composi-tions is not achieved easily. Mammalian kidney express only�1, allowing sensitivity to �1 to be assessed (20). Also �1-subunits of rodents are uniquely insensitive to ouabain,whereas �2 and �3 are sensitive. This allows one to differen-tiate between activity of rodent �1 and �2/�3 (but not between�2 and �3) in, for example, brain microsomal membranes,using micromolar and millimolar concentrations of ouabain(36). However, to assess isoform selectivity of CGs systemat-ically, it is necessary to use the individual recombinant pro-teins. This issue has been studied before using Xenopusoocytes expressing the different human isoform complexes (6)and more recently in this laboratory using plasma membraneand purified human isoform proteins obtained from the yeastPichia pastoris (P. pastoris) (14, 15). In Xenopus oocytes,activity of expressed pumps is assessed after subtraction of themuch lower activity of endogenous pumps. Yeast do notexpress endogenous Na�-K�-ATPase at all. Although bothsystems are suitable for expressing and characterizing individ-ual isoforms of choice, they are both quite different from themammalian cells expressing Na�-K�-ATPase. Regulatory el-ements involving lipid structures or microdomains and intra-cellular anchoring proteins may be lacking and might affectCG binding. To overcome this potential caveat, we havedeveloped cell clones of human non-small cell lung carcinoma(H1299), which express either human �1, �2, or �3 in com-bination with �1. This was done by transfecting cells with theappropriate �2- and �3-plasmids and silencing endogenous �1with appropriate siRNA. CG specificity was measured in intactcells as displacement of 3H ouabain by the CG of choice.

Previous experiments with recombinant human �1�1-, �2�1-,and �3�1-isoforms expressed individually in the yeast P.pastoris showed that ouabain, the CG used most commonlyin research, shows moderate selectivity for �1 and �3 over�2 (6), whereas the clinically used digoxin is partially �2selective (15). Most recently, synthetic derivatives ofdigoxin have been found to show enhanced selectivity for�2 over �1 and �3 both in the yeast expression system andin vivo, in an animal model (16).

† Deceased 30 November 2014.Address for reprint requests and other correspondence: M. Cherniavsky Lev,

Dept. of Biological Chemistry, Weizmann Institute of Science, Rehovot,7610001, Israel (e-mail [email protected]).

Am J Physiol Cell Physiol 309: C126–C135, 2015.First published May 20, 2015; doi:10.1152/ajpcell.00089.2015.

0363-6143/15 Copyright © 2015 the American Physiological Society http://www.ajpcell.orgC126

by 10.220.33.1 on Novem

ber 6, 2017http://ajpcell.physiology.org/

Dow

nloaded from

mailto:[email protected]

http://ajpcell.physiology.org/

A number of studies have suggested CGs as potential anti-cancer drugs, which can limit tumor growth (26, 29). It wasalso reported that, in colon cancer cells, �3 is expressed inpreference to �1 (33). By contrast, in a recent study (5), it wasobserved that malignant or oncogene-transfected cells are lesssensitive than nontumor breast cells to ouabain-mediated inhi-bition of proliferation and apoptosis. To further explore thisissue, we have measured cell proliferation and effects of CGsin cells expressing �1- or �2-isoforms and in several malignantand nontumor lines expressing only �1 and have relatedsensitivity to CGs to the abundance of plasma membraneNa�-K�-ATPase in these cells. The conclusion is that prolif-eration and survival in the presence of ouabain are determinedonly by the cell surface abundance of Na�-K�-ATPase.

MATERIALS AND METHODS

Cells expressing a single �-Na�-K�-ATPase isoform. Humannon-small cell lung carcinoma cell line (H1299) was used to generatecells that express individual �-isoforms. For construction of thevectors, the gfp coding sequence was inserted in the pIRES-purovector using EcoRI and NotI digestion enzymes. The coding se-quences of �1-, �2-, and �3-isoforms of the �-subunit of Na�-K�-ATPase were ligated into pIRES-puro in frame with green fluorescentprotein (GFP) between the first and the second amino acid of theirNH2 terminus. Cells were stably transfected with the GFP-tagged�1/�2/�3-vectors using JetPEI transfection reagent (PolyPlus Trans-

fection, Illkirch, France). Positive GFP-fluorescent clones were se-lected by flow cytometry, and individual clones were expanded.GFP-tagged �2- and �3-clones with the highest level of expressionunderwent endogenous �1-silencing by transfection with specificshRNA constructs (Qiagen, Valencia, CA) and selection with hygro-mycin (200 �g/ml). Successful expression of tagged isoform andsilencing of endogenous �1 were assayed by Western blotting usinganti-GFP and anti-�1 NH2-terminal antibody (6H). Plasma membranelocation of the transfected � was determined by fluorescence micros-copy and surface biotinylation. For generation of cell lines withdifferent levels of �1-expression, �1-GFP-expressing cells underwent�1-silencing with specific shRNA constructs. Several stable cloneswere selected, and the level of �1-expression was measured as thelevel of maximal ouabain-binding capacity of the cells as describedbelow.

Surface biotinylation and Western blotting. Biotinylation of cellsurface proteins was done as reported previously (4). In brief, cellmonolayers at �90% confluency were incubated for 15 min at 4°C ona rocking platform with 1 mg/ml EZ-Link N-hydroxy-sulfo-succin-imido-biotin (sulfo-NHS-SS-biotin; Pierce, Rockford, IL). Cells werewashed twice in PBS plus Ca2� and Mg2� (PBSCM), and freeunreacted sulfo-NHS-SS-biotin was quenched by 20-min incubationwith 0.1% BSA in PBSCM (wt/vol). In experiments testing Na�-K�-ATPase internalization, the biotinylated cells were incubated in cellculture medium tested or not for CGs for 6 h. At the end of this period,cells were washed four times in a reducing solution (100 nM Tris, pH8.6, 100 mM NaCl, 2.5 mM EDTA, 50 mM sodium-2-mercaptoeth-anesulfonate) and further incubated for 10 min at 4°C to allowcomplete cleavage of surface-exposed S-S bonds. Effective removalof surface-exposed biotin was verified by reducing one plate imme-diately after the biotinylation. After two rinses in PBS, cells werelysed in a buffer composed of the following: 50 mM Tris·HCl, pH 7.4,150 mM NaCl, 1% Triton X-100, 0.1% SDS, 5 mM EDTA, andprotease inhibitors [20 �g/ml leupeptin, 0.01 trypsin inhibitor units(TIU)/ml aprotinin, and 2 mM PMSF]. Cell lysates were centrifugedfor 10 min (14,000 g, 4°C), and the supernatants were incubated withstreptavidin-agarose beads (Immunopure immobilized streptavidin;Pierce) diluted in 50 mM Tris·HCl, pH 7.4, 100 mM NaCl, and 5 mMEDTA plus protease inhibitors. Following an overnight incubation at4°C, the beads were washed twice with rinsing solution (20 mMTris·HCl, pH 7.4, and 500 mM NaCl) and once with 10 mM Tris·HCl,pH 7.4. Washed bead pellets were resuspended in Laemmli bufferplus 50 mM dithiothreitol, and biotinylated internalized proteins wereeluted by incubation in 55°C for 30 min.

For Western blotting, biotinylated proteins and whole-cell lysatesamples were dissolved in 4� Laemmli sample buffer and resolved on7.5% or 10.0% Tris-Tricine SDS-PAGE. The resolved proteins weretransferred to PVDF membrane cut into several regions and incubatedwith different primary antibodies by standard methods. Expression

Fig. 1. Western blot analysis of �-isoform expression profile in selected clones.H1299 cell clones expressing green fluorescence protein (GFP)-tagged �-iso-forms were either transfected with �1-shRNA to silence the endogenous � ornot. Cells were surface biotinylated, and surface-expressed proteins wereisolated using streptavidin beads and subjected to Western blotting usinganti-�1 (6H) and anti-GFP antibodies. Anti-tubulin was used as loadingcontrol. Different background intensities are a result of different exposuretimes of the presented blots. Left panel depicts nontransfected H1299 cells.

Fig. 2. Confocal images of H1299 cells ex-pressing GFP-tagged �1-, �2-, and �3-iso-forms of Na�-K�-ATPase. Live �1-GFP(left)-, �2-GFP (middle)-, and �3-GFP(right)-expressing cells were visualized forGFP expression (excitation 488 nM, emis-sion 509 nM) using a scanning confocalmicroscope (Olympus FV1000) with a �60oil-immersion objective.

C127ISOFORMS, CARDIAC GLYCOSIDE SELECTIVITY, AND CELL DEATH

AJP-Cell Physiol • doi:10.1152/ajpcell.00089.2015 • www.ajpcell.org

by 10.220.33.1 on Novem


Dow

nloaded from


was detected by enhanced chemiluminescence and quantified usingImageQuant LAS 4000 mini (GE Healthcare, Piscataway, NJ).

3H-ouabain-binding experiments. Binding constants of differentCGs were determined by measuring the displacement of 5 nM 3H-ouabain by increasing concentrations of a competing CG as describedpreviously (30). Briefly, medium was removed from confluent mono-layers of the tested cells seeded in six-well plates, and the plates werewashed twice with RM buffer (20 mM Hepes-NaOH, 130 mM NaCl,0.5 mM MgCl2, 0.2 mM Na2HPO4, 0.4 mM NaH2PO4 and 0.2 mg/mlphenol red, pH 7.4) preheated to 37°C. For the displacement reaction,cells were incubated for 1 h at 37°C with 1 ml of RM buffercontaining 1.5 mM sodium vanadate, 5 nM of 3H-ouabain, andincreasing concentrations of the tested CG. In several experiments, 5mM of KCl was added to the RM buffer. Following the incubationperiod, 3H-ouabain solution was removed, and the cells were washedthree times with RM buffer (10 min, 4°C each). Cells were lysed with

500 �l of RIPA buffer [20 mM Tris, pH 7.4, 137 mM NaCl, 10%(vol/vol) glycerol, 0.1% (wt/vol) SDS, 0.5% (wt/vol) deoxycholate,1% (vol/vol) Triton X-100, and 2 mM EDTA] for 20 min. A sampleof the lysate was used for protein quantification, and the radioactivityof the remaining lysate was quantified. Measured radioactivity wasnormalized to the protein concentration of the sample. Fitted dissoci-ation constants, KD, were obtained using Kaleidagraph software. K0.5

was calculated using a one-site inhibition model: b/bCG�0 � K0.5/([CG] � K0.5), where b refers to the 3H-ouabain bound at a particularconcentration of the CG, and bCG�0 refers to the 3H-ouabain bound at5 nM 3H-ouabain in the absence of other CG. KD was calculated fromK0.5 by taking into account ouabain-CG competition as KDCG �K0.5/(1 � [Oubf]/KD Oub). The values of (1 � [Oubf]/KD Oub) were1.35, 1.09, and 1.35, for �1, �2, and �3, respectively. The KD valuereported for each CG is an average SE of three to four independentexperiments. The selectivity ratios KD�1/KD�2 and KD�1/KD�3 werecalculated as the quotient of the individual KD values.

For quantification of surface expression of different �-isoforms, thespecific binding capacity of 3H-ouabain was measured. The protocolwas similar to that described above with the following changes: cellswere incubated for the indicated time with 30 nM 3H-ouabain pluseither 270 nM (test_oub) or 500 �M (bg_oub) of nonradioactiveouabain. Specific ouabain binding capacity (bmax) (pmol/mg) wascalculated by the following equation:

bmax �

�CPM�test_oub��protein�

�CPM�bg_oub�

�protein� �CPM�total�

� �test_oub�

Cell proliferation assay. Effects of ouabain on cell proliferationwere determined using 2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2h-tetrazolium-5-carboxanilide (XTT)-based cell proliferation assay kit(Biological Industries, Beit Haemek, Israel) according to the manu-facturer’s instructions. Briefly, 5,000 mid-log-phase cells were seededin a 96-well tissue culture plate. The next day cells were incubatedwith indicated concentrations of CGs for specified time periodsfollowed by addition of the XTT solution (3 h, 37°C). Absorbancewas measured at 450 nm and 630 nm (nonspecific), and relativeproliferation was calculated as the ratio of specific absorbance inCG-treated and -untreated cells at the same time point. IC50 valueswere calculated by fitting the data into a four-parameter nonlinear-regression curve fit.

Fig. 3. Ouabain-binding capacities of various cell clones. Maximal ouabain-binding capacities of the �1-GFP-, �2-GFP-, and �3-GFP-expressing cells aswell as of cells with partially silenced �1-isoform to �50 and 30% of itsoriginal level (si-�1–50 and si-�1–30, respectively) were measured as de-scribed in MATERIALS AND METHODS. Means SE of 3 independent experi-ments are depicted.

Fig. 4. 3H-ouabain displacement by different concentrations of ouabain and digoxin. Representative curve of 5 nM 3H-ouabain displacement in �1-GFP- (�),�2-GFP- (�), and �3-GFP-expressing (Œ) cells by ouabain (left) or digoxin (right) with addition of 5 mM K�. Solid and dotted lines are the fitted curves fora 1-site competitive displacement model.

C128 ISOFORMS, CARDIAC GLYCOSIDE SELECTIVITY, AND CELL DEATH


by 10.220.33.1 on Novem


Dow

nloaded from


Acridine orange/ethidium bromide staining. For establishment ofthe cell death mechanism, following ouabain treatment, cells werestained with 8 �g/ml acridine orange/ethidium bromide (AO/EB) dyemix by a procedure described previously (32). Briefly, 5,000 cellswere seeded in a noncoated eight-well microslide (ibidi, Martinsried,Germany). After 24 h, 30 nM of ouabain was added, and, after anadditional 48 h, slides were centrifuged (200 revolution/min, 3 min)and stained by addition of AO/EB dye mix. Cells were visualizedusing Olympus FV10i-LIV microscope with a �60 water-immersionlens. For each cell line and each treatment, at least 300 cells werecounted, and staining was scored as live (AO stained), early apoptotic(AO stained, nucleus condensed or fragmented), late apoptotic (AOand EB stained, nucleus condensed or fragmented), or necrotic (AOand EB stained, nucleus similar to live cells).

Preparation of crude membrane fraction. Crude membrane frac-tions were prepared as described previously (12). Briefly, confluent10-cm plates were washed twice with PBS, collected, and sedimentedby centrifugation (1,000 revolution/min, 3 min, 4°C), suspended in500 �l in hypotonic lysis buffer [10 mM Tris·HCl, pH 7.4, 10 mMNaCl, 1.5 mM MgCl2, 1 mM dithiothreitol, and protease inhibitors(20 �g/ml leupeptin, 0.01 TIU/ml aprotinin, and 2 mM PMSF)], andincubated for 15 min on ice. The swollen cells were disrupted by15–20 passages through a 27-gauge syringe and clarified by centrif-ugation at 400 g (10 min, 4°C). Cell homogenate was spun at 22,000g for 30 min; pelleted crude membranes were resuspended in 100 �llysis buffer.

Antibodies and materials. A mouse monoclonal antibody to theNH2-terminal sequence of the �1-subunit of Na�-K�-ATPase (6H)was kindly provided by Dr. M. J. Caplan, Yale University School ofMedicine. A rabbit polyclonal antibody to the �2-subunit of Na-K-ATPase was from Millipore (Darmstadt, Germany). Mouse monoclo-nal antibodies to the �3-subunit of Na�-K�-ATPase (XVIF9-G10)and anti-GFP (B-2) were purchased from Santa Cruz Biotechnology(Santa Cruz, CA); monoclonal anti-�-tubulin was from Sigma-Al-drich (St. Louis, MO). Horseradish peroxidase-coupled secondaryantibodies were from Bio-Rad (Hercules, CA)

3H-ouabain was from Perkin-Elmer (Waltham, MA). Ouabain,digoxin, leupeptin, aprotinin, and PMSF were from Sigma-Aldrich.Digoxin-glycine-amide (DGlyN) and digoxin-methyl-amine (DMe)were synthesized as described previously (16).

Cell culture. H1299, MDA-MB-231, A549, LNCaP, SKMEL-5,and OVCAR-3 cells were cultured in RPMI 1640 medium (BiologicalIndustries, Beit Haemek, Israel) supplemented with 10% FCS, 1 mMNa pyruvate, penicillin, and streptomycin. LLC-PK1, MCF7, HeLa,HEK, PANC1, and HFF cells were cultured in DMEM medium(Biological Industries) supplemented with 10% FCS, penicillin, andstreptomycin.

RESULTS

Generation of isoform-specific cell clones. To study interac-tion of different CGs with the Na�-K�-ATPase in intact cells,we have generated H1299 cell clones that express differentindividual �-isoforms using stable transfection followed bysilencing of the endogenous �1. Figure 1 depicts a typicalexperiment in which H1299 cells were transfected with GFP-

Table 1. Binding constants of various CGs to H1299 cells expressing solely �1-, �2-, or �3-isoforms of Na�-K�-ATPase

Cardiac Glycoside

KD SE, nM Selectivity Ratio SE

�1 �2 �3 �1/�2 �1/�3

Ouabain 10.93 0.38 22.25 2.96 11.00 0.05 0.49 0.07 0.99 0.04Ouabain �5 mM K� 120.36 1.12 97.52 26.61 58.11 13.39 1.23 0.34 2.07 0.48Digoxin 37.03 4.12 20.23 1.32 19.46 2.62 1.83 0.23 1.90 0.33Digoxin �5 mM K� 438.97 53.5 91.39 19.30 159.00 19.37 4.80 1.17 2.76 0.48DGlyN 68.74 5.76 24.07 3.28 50.72 5.61 2.86 0.46 1.36 0.19DGlyN �5 mM K� 506.93 74.42 86.62 6.78 306.24 22.31 5.99 1.00 1.65 0.27DMe 41.54 5.70 13.78 2.84 33.28 5.25 3.02 0.75 1.25 0.26DMe �5 mM K� 173.99 1.01 40.94 1.71 162.06 55.75 4.25 0.18 1.07 0.37K0.5K�, mM 1.38 0.03 3.26 0.50 1.43 0.09 0.42 0.07 0.96 0.07

Values are means SE of 3–4 independent experiments. Values were calculated by simulating experimental data to the equation b/bCG � 0 � K0.5/([CG] �K0.5) using Kaleidagraph software; b refers to the 3H-ouabain bound at a particular concentration of the cardiac glycoside (CG), and bCG � 0 refers to the3H-ouabain bound at 5 nm 3H-ouabain in the absence of other CGs. DGlyN, digoxin-glycine-amide; DMe, digoxin-methyl-amine.

Fig. 5. Internalization of Na�-K�-ATPase �1- and �2-isoforms. Confluentmonolayers of H1299 cells expressing GFP-tagged �2- and native �1-isoforms were surface biotinylated. Different plates received either theindicated concentrations of cardiac glycosides (CGs) or diluent and wereincubated for an additional 6 h. At the end of this period, surface-expressedbiotin was cleaved by reduction with mercaptoethanesulfonate-Na. Cellswere lysed, and biotinylated proteins were purified and resolved on SDS-PAGE gel as described in MATERIALS AND METHODS. Upper blots wereprobed simultaneously with antibodies to �1-Na�-K�-ATPase (6H) andGFP. Anti-tubulin was used as a loading control. Intracellular biotinylated� was quantified and expressed as a fold increase relative to the diluent-treated sample. Means SE of 3 independent experiments are depicted. Inall cases, the difference between the fold internalization of the �1- and�2-isoforms was statistically significant (*P � 0.05).



by 10.220.33.1 on Novem


Dow

nloaded from


tagged �2 and �3 with and without silencing endogenous �1.The transfected cells were biotinylated, and surface-expressedproteins were pulled down and analyzed by Western blots.Interestingly, as long as �1 was present, �2 and �3 wereexpressed at a relatively low level. However, the subsequentsilencing of �1 resulted in a large increase in the abundance of�2 and �3, by a factor of 9.5- and 16.3-fold, respectively (lanes4 and 6 in Fig. 1). This may suggest regulation of translation/degradation or trafficking to match the amount of pumps in theplasma membrane required to maintain Na� and K� gradients.Because �1 silencing elevated the total levels of �2 and �3 andnot only its membrane expression (data not shown), the firstoption is more likely. Surface expression and correct cellularlocalization of the transfected GFP-tagged proteins could alsobe demonstrated by confocal imaging (Fig. 2).

To accurately quantify the level of expression of the iso-forms, we measured ouabain-binding capacity (bmax) of thedifferent cell clones. These values correspond to the number ofouabain-binding sites on the cell surface and provide an esti-mate of a number of surface-expressed pumps. As summarizedin Fig. 3, the values of maximal ouabain-binding capacities inthe �2- and �3-expressing cells are 0.68 0.05 and 0.36 0.02 (pmol ouabain/mg protein), which are 54.5% and 31.5%of the level in the �1-expressing cells (1.15 0.04 pmolouabain/mg protein), respectively. To eliminate, in subsequentexperiments, effects attributable to different expression levelsof Na�-K�-ATPase isoforms, we also generated additionalclones of �1-expressing cells in which the �1-isoform waspartly silenced. These clones denoted si-�1–50 and si-�1–30and expresses �52.1% and 30.2% of the amount of �1 similarto that of the �2- and �3-expressing cells, with maximalouabain-binding capacity values of 0.60 0.01 and 0.35 0.02 (pmol ouabain/mg protein), respectively. This permitscomparison of cell behavior with comparable surface abun-dance of different �-isoforms. All cell lines were shown to

have similar proliferation rates with a doubling time of �30 hin the absence of ouabain.

Characterization of CG specificity in intact cells. We haveused the H1299 cell clones expressing �1, �2, and �3 to testwhether the selectivity order for native CGs and chemicallymodified digoxin derivatives observed for human isoformproteins expressed in the yeast P. pastoris (15, 16) holds alsoin the intact mammalian cells. Figure 4 presents comparison ofdisplacement of 3H-ouabain for ouabain and digoxin. Table 1compares the KD values of �1, �2, and �3 for ouabain,digoxin, and digoxin derivatives (DGlyN and DMe) withenhanced selectivity for �2 (16). The KD values for ouabainbinding to the intact H1299 cells, in the absence of extracel-lular K ions, obtained by fitting the data to a one-site model areas follows: �1, 10.93 0.38 nM; �2, 22.25 2.96 nM; and�3, 11.00 0.05 nM. These values are virtually the same asobtained in the yeast membranes (15). Thus ouabain shows amoderate selectivity for �1 over �2 (KD�1/�2 � 0.49) with noselectivity for �3 over �1 (KD�1/�3 � 0.99). Furthermore,digoxin shows a moderate selectivity for �2 and �3 over �1,again quite similar to the finding in the yeast membranes (15).Thus the dissociation constant, KD, of CGs is an intrinsicproperty of the Na�-K�-ATPase, irrespective of the origin orlipid content of the membrane or potential regulatory factorspresent only in mammalian cells. Interestingly, in the presenceof 5 mM extracellular K�, the selectivity ratios for bothouabain and digoxin were substantially higher for �1/�2 andalso, but less so, for �1/�3. K� and CG binding are known tobe antagonistic toward each other at the extracellular bindingsurface (6). The K0.5 K for displacing 3H-ouabain is signifi-cantly higher for �2 compared with �1, whereas that for �3 iscloser to �1 (bottom line of Table 1). Thus the origin of thiseffect on the selectivity ratio is that at the physiologicalconcentration of 5 mM K�; the antagonism between the CGand K� is lower for �2 than �1 or �3. Another important result

Fig. 6. Correlation between ouabain-binding capacity and proliferation. A: relative cell proliferation of �1-GFP (o) and �2-GFP (Œ) cells was measured using2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2h-tetrazolium-5-carboxanilide (XTT)-based cell proliferation assay and expressed as a percentage of the correspond-ing value for nontreated cells (NT) at the same time point. Cells were treated for 48 h with different concentrations of the indicated CGs. Means SE of 4independent experiments are depicted. Solid and dashed lines are the fitted curves for the 4-parameter nonlinear regression-curve fit. B: relative proliferationvalues of �1-GFP cells and their partially silenced derivatives were measured following 72-h incubation with 30 nM ouabain and expressed as a percentage ofthe corresponding value for nontreated cells at the same time point. The values obtained were plotted against maximal ouabain-binding capacity of these cellsand fitted to a linear line (y � 94.7x � 20.1; R2 � 0.99).



by 10.220.33.1 on Novem


Dow

nloaded from


in Table 1 is that the selectivity ratio �1/�2 for DGlyN andDMe with enhanced selectivity for �2 (16) is higher than fordigoxin itself, at least in the absence of K� ions (16). Overall,it appears that the trends for CG selectivity features observedwith the H1299 cells are similar to those observed with theyeast membranes expressing �1�1, �2�1, and �3�1 or thepurified isoform proteins.

CG induced internalization of �1 vs. �2. It is known thatCGs, not only inhibit the Na�-K�-ATPase activity, but alsoinduce internalization of its subunits (3, 4, 22). It was ofinterest to compare internalization of �1 and �2. Both isoformsare expressed in cardiac myocytes, the cells that are the targetsof the therapeutic action of CG in congestive heart failure (7).Na�-K�-ATPase internalization is one of the responses to CG,and the internalization level of the two isoforms could havedifferent functional implications. To quantify internalization of�1 and �2 under conditions that are expressed in the same cell(similar to the situation in myocytes), we have used �1-expressing H1299 cells, which coexpressed �2-GFP, and stud-ied their response to sublethal concentrations of ouabain anddigoxin (30 and 100 nM). As can be seen from Fig. 5, �1, with

a KD for ouabain somewhat lower than that of �2, indeedundergoes a higher level of internalization than �2 in responseto ouabain, particularly at the lower concentration. By contrast,�2, which is somewhat more sensitive to digoxin (Table 1), isbetter internalized than �1 in response to sublethal concentra-tions of digoxin. These data rule out a major difference in theresponse of the two isoforms to CGs as well as major differ-ences in the ability of CGs to induce internalization. Becausethe degree of internalization of �1 and �2 reflects their relativeaffinities for ouabain and digoxin, respectively, we conclude that theinternalization reflects the degree of occupation of the differentisoforms by the different CGs.

Effect of CGs on cell proliferation. One might predict thatkilling of cells expressing �1 or �2 would show the sameisoform selectivity pattern for the different CGs as found in theouabain-displacement assays. However, in initial experimentsthat compared proliferation of �1-GFP cells, in which theendogenous pumps were not silenced, with the �2-GFP cells,the �1-GFP cells were similarly sensitive to ouabain comparedwith �2-GFP cells (Fig. 6A). Nevertheless, there is a differencebetween the two cell types in that the �1-GFP cells express a

Fig. 7. Establishment of cell death mechanism in H1299 cells with different level of Na�-K�-ATPase expression. H1299 cells with different level of Na�-K�-ATPaseexpression were subjected to acridine orange/ethidium bromide (AO/EB) staining assay as described in MATERIALS AND METHODS. The appearance of treated cells is comparedwith that of nontreated cells. AO intercalates between DNA of both live and dead cells; however, the nuclear staining of early apoptotic cells (blue arrows) is distinctly differentfrom that of live cells (white arrows) because of chromatin condensation and fragmentation. EB intercalates between DNA of cells with lost plasma membrane integrity; thusthe nuclei of late apoptotic cells appear with condensed and fragmented red chromatin (red arrows), and necrotic nuclei appear red and structurally intact. A: appearance ofAO/EB-stained H1299-derived cell lines either nontreated or following 48-h incubation with 30 nM ouabain. B: relative quantification of live (open bars), early apoptotic (hashedbars), and late apoptotic (solid bars) cells with different levels of Na�-K�-ATPase expression.



by 10.220.33.1 on Novem


Dow

nloaded from


higher surface density of pumps compared with �2-GFP, asjudged by ouabain-binding capacity (Fig. 3). This raised thequestion whether cells that express higher levels of Na�-K�-ATPase also exhibit higher resistance to CGs and survivebetter in the presence of the Na�-K�-ATPase inhibitors. Oneapproach used to address this question was to compare theproliferation of a number of H1299 cell clones in which �1 waspartially silenced compared with the parental cells. Figure 6Bcompares proliferation of these cell clones, measured as theratio between XTT readings of 30 nM ouabain and diluent-treated cells, following a 72-h incubation, and the number ofsurface-expressed pump units, measured as bmax for ouabainbinding. A clear linear relationship was obtained between thesurface density of active pumps and cell proliferation. Thissuggests that nearly all the surface-expressed pumps are essen-tial for proliferation, and any reduction in their number alsoreduces cell proliferation. Extrapolating the line to zero prolif-eration with 30 nM ouabain suggests that more than 20% of theoriginal number of pumps is required to support cell prolifer-ation.

Evaluation of the mode of cell death by AO/EB stainingprocedure (Fig. 7A) showed that the main mode of cell deathfollowing ouabain treatment is apoptosis and that the ratio ofapoptotic cells is a function of surface-expressed Na�-K�-ATPase molecules (Fig. 7B). This was also verified by mea-suring the activation of the proapoptotic enzyme caspase 3(data not shown).

With these insights, we have compared the proliferation of�2-GFP cells and the partly silenced �1-cells (si-�1–50 cells)that demonstrate an essentially equal surface abundance asjudged by ouabain-binding capacity, whereas the si-�1–50cells express only about 50% of active Na�-K�-ATPase pro-tein units compared with the nonsilenced �1-GFP cells (Fig.3). As depicted in Fig. 8, in response to incubation withdifferent CGs, �2-GFP cells are less sensitive to ouabain thanthe si-�1–50 cells, but �2-GFP cells are more sensitive todigoxin and DMe. The approximate values of the IC50 are asfollows: 43 nM and 119 nM for ouabain, 153 nM and 86 nMfor digoxin, and 62 nM and 22 nM for DMe for si-�1–50 cellsand �2-GFP cells, respectively. In other words, for �2-GFPcells and the si-�1–50 cells expressing comparable numbers ofactive pumps, the experiments indeed validate the differential

selectivity of �2- and �1-isoforms for the different CGs aspredicted.

As a more systematic test of the notion that cell surfacedensity of active pumps is a crucial factor in killing by CGs, wehave compared proliferation of different cell lines that expressdifferent amounts of Na�-K�-ATPase. In light of the reportssuggesting different sensitivities of tumor and nontumor cellsto CGs (5, 8, 27, 35, 37), it was of special interest to comparetumor-derived cell lines with nontransformed cells as well asthe isoform expression pattern because some cancer cellsexpress �3 rather than �1 (33, 38). Table 2 lists the 12 celllines tested and their origin (transformed vs. nontransformed).In Western blots, all cell lines were found to express �1 butinsignificant amounts �2 or �3 (Fig. 9). There was a very goodlinear relationship of cell proliferation in the presence ofouabain with the number of surface ouabain-binding sites (Fig.10). Thus the clear conclusion is that cell proliferation upontreatment with CGs depends crucially on the abundance ofactive surface-expressed pumps but not on the state of the cells,transformed or nontransformed.

DISCUSSION

This work addressed two main issues: 1) determining CGspecificities of different human �-isoforms in intact cells underphysiological conditions and their comparison to the previ-ously determined values in isolated yeast membranes and thepurified isoform proteins, and 2) determination of the relation-ship between isoform expression, surface density of the pumps,and cell proliferation.

The first aim was achieved by expressing �2 and �3 inH1299 cells and silencing endogenous �1. CG selectivity incell clones expressing single �-subunit together with �1 wasthen determined as displacement of specifically bound 3H-ouabain by the CG of choice. The data obtained suggested thatthe KD values for known CGs as well as �1/�2 and �1/�3selectivity are similar to those reported recently for P. pastorismembranes expressing individual �-subunits and the purifiedisoform proteins (15).

The main difficulty in determining the differential responsesof Na�-K�-ATPase isoforms to CGs in native human cells andtissues is the fact that all human cells express the �1-isoform,

Fig. 8. Comparison of proliferation of �2-GFP cells and si-�1–50 cells following CG treatment. Relative cell proliferation of si-�1–50 cells (o) and �2-GFP cells (Œ) wasmeasured using XTT-based cell proliferation assay and expressed as a percentage of proliferation of nontreated cells at the same time point. Cells were treated for 48 h withdifferent concentrations of the indicated CGs. Means SE of 4 independent experiments are depicted. Solid and dashed lines are the fitted curves for the 4-parameter nonlinearregression-curve fit.



by 10.220.33.1 on Novem


Dow

nloaded from


whereas the �2- and �3-isoform expression pattern is cell andtissue specific, and, of course, human tissues for research arenot readily available. Previously, individual human isoforms ofNa�-K�-ATPase have been expressed in systems that either donot express Na�-K�-ATPase or express very little Na�-K�-ATPase. These include P. pastoris (15), Saccharomycescerevisiae (13), baculovirus-infected insect cells (2), and Xe-nopus oocytes (10). As discussed previously (8), the highlevels of expression in P. pastoris membranes and purificationof the isoform proteins are crucial for reliable determination ofisoform selectivity of CGs. However, to the best of our knowl-edge, expression of individual human isoforms of Na�-K�-ATPase alone in human cells, and their use for determinationof isoform selectivity of CGs, has not yet been achieved. Byusing our expression system, we have now been able to isolatethe effects of different CGs (both natural and synthetic) on theindividual isoform proteins in a cellular environment. More-over, we have now been able to observe the fate of theindividual isoforms as they are internalized following inhibi-tion by specific CGs as well as proliferation of cells expressingonly �1- or �2-isoforms upon exposure to different CGs.

The present results with human cells expressing humanisoforms confirm that ouabain is moderately �1 selective (36),digoxin is moderately �2 selective (15), and the syntheticderivatives DGlyN and DMe shows enhanced �2 selectivity(16). The presence of K� ions in the reaction medium poisesthe affinities further toward the �2-isoform, attributable to areduced K-CG antagonism for �2, a factor potentially ofphysiological significance. The degree of �2 selectivity of bothdigoxin and its synthetic derivatives appears somewhat lowerthan observed for yeast membranes and purified isoform pro-

teins (15, 16). This may be due to the fact that the conditionsfor ouabain binding in whole cells are not identical to those forNa�-K�-ATPase activity of purified isoforms. The findings onisoform selectivity of the inhibitors with H1299 cell validate, atleast qualitatively, the findings using yeast cells and purifiedisoform proteins (15, 16). This suggests that the apparentaffinities at steady state of CG binding are an inherent propertyof the pump isoforms that is not affected by membrane lipidcomposition or other intracellular regulatory factors. Internal-ization analysis and cell proliferation measurements also sup-port the notion that �1 is somewhat ouabain selective, whereas�2 is somewhat digoxin selective, with enhanced selectivityfor DMe.

Partial silencing of �1-subunit of Na�-K�-ATPase enabledus to demonstrate a strong linear correlation between thespecific ouabain-binding capacity of the cells and their generalsensitivity to ouabain, as measured by cellular proliferationassay. Upon inhibition of Na�-K�-ATPase with CGs, one ofthe immediate consequences is inhibition of its ion-pumpingactivity and as a result dissipation of ion gradients, most likelya scenario of triggering of apoptotic cell death (18, 31, 34).This conclusion also fits well with a recent finding that the K0.5

for inhibition of growth of various cell lines by different CGsis linearly related to the Ki for inhibition of Na�-K�-ATPaseactivity of the purified human �1�1-isoform (28). In parallel,internalization of the pump, which is the cellular method ofremoving inhibited pumps from the plasma membrane, takesplace.

Previously, work by Xie and colleagues (19), showed estab-lishment of a viable derivative of the LLCPK1 cell line withknockdown of �1-isoform activity to �20% of the activity inthe parental cell line. The knockdown of �1 in the work by Xieet al. was correlated with an increase in basal Src activity andactivation of consequent signal transduction pathways. Al-though our present work does not address this issue directly,our previously published study (4) showed no relationship

Table 2. Description of the tested cell lines

Cell Line Origin

A549 Human non-small cell lung carcinomaH1299 Human non-small cell lung carcinomaHeLa Human cervical cancerHEK Human embryonic kidney (noncancer)HFF Human foreskin fibroblast (noncancer)LLC-PK1 Pig kidney proximal tube (noncancer)LNCap Human prostate adenocarcinomaMCF7 Human breast cancerMDAMB231 Human breast cancerOVCAR3 Human ovarian cancerPANC1 Human pancreatic carcinomaSKMEL5 Human melanoma

Fig. 9. Isoform expression profile in various cell lines. SDS-PAGE of 10 �g ofenriched membrane fraction was blotted and probed with specific antibodiesfor Na�-K�-ATPase �-isoforms. Right: 50 ng of Na�-K�-ATPase �-isoformspurified from Pichia pastoris.

Fig. 10. Correlation between ouabain-binding capacity and cell proliferation invarious cell lines. Proliferation ratio of the tested cell lines following 72-hincubation with either 30 nM ouabain or with diluent was plotted againstmaximal ouabain-binding capacity of these cells and fitted to a linear trend line(y � 69.3x � 11.2; R2 � 0.97).



by 10.220.33.1 on Novem


Dow

nloaded from


between Src/phosphoinositide-3-kinase signaling pathways toouabain-induced inhibition and subsequent internalization ofNa�-K�-ATPase in our cellular system, as opposed to resultsdescribed previously by Xie et al. (11, 22–24).

Several studies reported that different cell lines vary in theirresponse to CGs. For example, in a recent publication (5), itwas suggested that ouabain causes greater inhibition of prolif-eration and more extensive apoptosis in nontumor breast cellscompared with malignant or oncogene-transfected cells, con-cluding that CGs are not good candidates as cancer therapies.On the contrary, earlier studies found that Na�-K�-ATPaseexpression is downregulated in canine prostate cancer (27) andthat expression of Na�-K�-ATPase �- and �-subunits werefound to be decreased in human bladder cancer (8) as well.Reviewing multiple studies, Newmann et al. (29) tried toestablish the potential of using CGs as anticancer drugs and themechanism by which their antiproliferative effect is differentbetween normal and tumor cell lines. The suggestion is thatthere is a difference of isoform expression profile betweennormal and tumor cell lines isolated from human colon, with ashift of expression toward the �3-isoform in the tumor cells.Mijatovic et al. (26), on the other hand, suggested that it is the�1-subunit of Na�-K�-ATPase that could represent a novelanticancer target, by presenting evidence that human lungcancer cell lines overexpressing the �1-subunit were sensitiveto several CGs.

To test the hypothesis that the variation in response toouabain is a result only of different pump density in differentcells, or �3 expression, we studied isoform expression andtried to establish whether the linear correlation between pro-liferation and ouabain-binding capacity still exists when thestudy is expanded to include a variety of other cell lines. The datapresented in Fig. 10, which also includes data from the four celllines used in Fig. 6 to give a total of 16 cell lines, confirm thatthis hypothesis is indeed correct. The 16 cell lines testedexpress only �1 and show a strong linear correlation betweenplasma membrane Na�-K�-ATPase surface density, as judgedby ouabain-binding capacity, and cell proliferation at 30 nM ofouabain, regardless of their isoform content, tissue origin, orthe status of the cells as malignant or nonmalignant. Thus thepresent data indicate that inhibition of cell proliferation by CGsis inversely related to the plasma membrane pump density and,as such, is not necessarily a promising approach in treatingcancer. This conclusion is in accord with that in the recentpublication (5). As an example, the MCF7 and MDAMB231breast cancer cells, with the lowest sensitivity to ouabain,are also the cells that display highest levels of surfaceouabain-binding capacity (Fig. 10). The same cells wereshown (5) to be particularly insensitive to ouabain, digoxin,and bufalin. Thus the present study also provides a plausibleexplanation for the lower sensitivity of the cancer vs.noncancer cells to CGs (5).

GRANTS

This work was supported by a grant from the Israel Science Foundation789/12 to S. Karlish.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: M.C.L., S.J.D.K., and H.G. conception and design ofresearch; M.C.L. performed experiments; M.C.L. analyzed data; M.C.L. pre-pared figures; M.C.L., S.J.D.K., and H.G. edited and revised manuscript;M.C.L. and S.J.D.K. approved final version of manuscript; H.G. interpretedresults of experiments; H.G. drafted manuscript.

REFERENCES

1. Bagrov AY, Shapiro JI, Fedorova OV. Endogenous cardiotonic steroids:physiology, pharmacology, and novel therapeutic targets. Pharmacol Rev61: 9–38, 2009.

2. Blanco G, Hatfield WR, Minor NT, Sánchez G, Koster JC, DeTomasoAW, Mercer RW. Studies of Na,K-ATPase structure and function usingbaculovirus. Ann NY Acad Sci 834: 88–96, 1997.

3. Cain CC, Sipe DM, Murphy RF. Regulation of endocytic pH by theNa�,K�-ATPase in living cells. Proc Natl Acad Sci USA 86: 544–548,1989.

4. Cherniavsky-Lev M, Golani O, Karlish SJD, Garty H. Ouabain-induced internalization and lysosomal degradation of the Na�/K�-AT-Pase. J Biol Chem 289: 1049–1059, 2014.

5. Clifford RJ, Kaplan JH. Human breast tumor cells are more resistant tocardiac glycoside toxicity than non-tumorigenic breast cells. PLoS One 8:e84306, 2013.

6. Crambert G, Hasler U, Beggah AT, Yu C, Modyanov NN, Horis-berger JD, Lelièvre L, Geering K. Transport and pharmacologicalproperties of nine different human Na, K-ATPase isozymes. J Biol Chem275: 1976–1986, 2000.

7. Despa S, Lingrel JB, Bers DM. Na/K-ATPase �2-isoform preferentiallymodulates Ca transients and sarcoplasmic reticulum Ca release in cardiacmyocytes. Cardiovasc Res 95: 480–486, 2012

8. Espineda C, Seligson DB, James Ball W Jr, Rao J, Palotie A, HorvathS, Huang Y, Shi T, Rajasekaran AK. Analysis of the Na,K-ATPasealpha- and beta-subunit expression profiles of bladder cancer using tissuemicroarrays. Cancer 97: 1859–1868, 2003.

9. Francis GS. The contemporary use of digoxin for the treatment of heartfailure. Circ Heart Fail 1: 208–209, 2008.

10. Geering K, Theulaz I, Verrey F, Häuptle MT, Rossier BC. A role forthe �-subunit in the expression of functional Na�-K�-ATPase in Xenopusoocytes. Am J Physiol Cell Physiol 257: C851–C858, 1989.

11. Gupta S, Yan Y, Malhotra D, Liu J, Xie Z, Najjar SM, Shapiro JI.Ouabain and insulin induce sodium pump endocytosis in renal epithelium.Hypertension 59: 665–672, 2012.

12. Hamada H, Tsuruo T. Purification of the 170- to 180-kiloDalton mem-brane glycoprotein associated with multidrug resistance. 170- to 180-kiloDalton membrane glycoprotein is an ATPase. J Biol Chem 263:1454–1458, 1988.

13. Hauck C, Potter T, Bartz M, Wittwer T, Wahlers T, Mehlhorn U,Scheiner-Bobis G, McDonough AA, Bloch W, Schwinger RHG, Mül-ler-Ehmsen J. Isoform specificity of cardiac glycosides binding to humanNa�,K�-ATPase alpha1beta1, alpha2beta1 and alpha3beta1. Eur J Phar-macol 622: 7–14, 2009.

14. Kapri-Pardes E, Katz A, Haviv H, Mahmmoud Y, Ilan M, Khalfin-Penigel I, Carmeli S, Yarden O, Karlish SJD. Stabilization of the �2isoform of Na,K-ATPase by mutations in a phospholipid binding pocket.J Biol Chem 286: 42888–42899, 2011.

15. Katz A, Lifshitz Y, Bab-Dinitz E, Kapri-Pardes E, Goldshleger R, TalDM, Karlish SJD. Selectivity of digitalis glycosides for isoforms ofhuman Na,K-ATPase. J Biol Chem 285: 19582–19592, 2010.

16. Katz A, Tal DM, Heller D, Haviv H, Rabah B, Barkana Y, MarcovichAL, Karlish SJD. Digoxin derivatives with enhanced selectivity for the�2 isoform of Na,K-ATPase: effects on intraocular pressure in rabbits. JBiol Chem 289: 21153–21162, 2014.

17. Kühlbrandt W. Biology, structure and mechanism of P-type ATPases.Nat Rev Mol Cell Biol 5: 282–295, 2004.

18. Lang H, Schulte BA, Schmiedt RA. Ouabain induces apoptotic cell deathin type I spiral ganglion neurons, but not type II neurons. J Assoc ResOtolaryngol 6: 63–74, 2005.

19. Liang M, Cai T, Tian J, Qu W, Xie ZJ. Functional characterization ofSrc-interacting Na/K-ATPase using RNA interference assay. J Biol Chem281: 19709–19719, 2006.

20. Lingrel JB. Na,K-ATPase: isoform structure, function, expression. JBioenerg Biomembr 24: 263–270, 1992.



by 10.220.33.1 on Novem


Dow

nloaded from


21. Lingrel JB. The physiological significance of the cardiotonic steroid/ouabain-binding site of the Na,K-ATPase. Annu Rev Physiol 72: 395–412,2010.

22. Liu J, Kesiry R, Periyasamy SM, Malhotra D, Xie Z, Shapiro JI.Ouabain induces endocytosis of plasmalemmal Na/K-ATPase in LLC-PK1 cells by a clathrin-dependent mechanism. Kidney Int 66: 227–241,2004.

23. Liu J, Kesiry R, Periyasamy SM, Malhotra D, Xie Z, Shapiro JI.Ouabain induces endocytosis of plasmalemmal Na/K-ATPase in LLC-PK1 cells by a clathrin-dependent mechanism. Kidney Int 66: 227–241,2004.

24. Liu J, Liang M, Liu L, Malhotra D, Xie Z, Shapiro JI. Ouabain-induced endocytosis of the plasmalemmal Na/K-ATPase in LLC-PK1cells requires caveolin-1. Kidney Int 67: 1844–1854, 2005.

25. McDonough AA, Farley RA. Regulation of Na,K-ATPase activity. CurrOpin Nephrol Hypertens 2: 725–734, 1993.

26. Mijatovic T, Van Quaquebeke E, Delest B, Debeir O, Darro F, Kiss R.Cardiotonic steroids on the road to anti-cancer therapy. Biochim BiophysActa 1776: 32–57, 2007.

27. Mobasheri A, Fox R, Evans I, Cullingham F, Martín-Vasallo P, FosterCS. Epithelial Na, K-ATPase expression is down-regulated in canineprostate cancer; a possible consequence of metabolic transformation in theprocess of prostate malignancy. Cancer Cell Int 3: 8, 2003.

28. Moreno Banuls LY, Katz A, Miklos W, Cimmino A, Tal DM, Ain-binder E, Zehl M, Urban E, Evidente A, Kopp B, Berger W, Feron O,Karlish S, Kiss R. Hellebrin and its aglycone form hellebrigenin displaysimilar in vitro growth inhibitory effects in cancer cells and bindingprofiles to the alpha subunits of the Na�/K�-ATPase. Mol Cancer 12: 33,2013.

29. Newman RA, Yang P, Pawlus AD, Block KI. Cardiac glycosides asnovel cancer therapeutic agents. Mol Interv 8: 36–49, 2008.

30. Orlowski J, Lingrel JB. Differential expression of the Na,K-ATPasealpha 1 and alpha 2 subunit genes in a murine myogenic cell line.Induction of the alpha 2 isozyme during myocyte differentiation. J BiolChem 263: 17817–17821, 1988.

31. Panagiotidis M, Cidlowski JA. Inhibition of Na�/K�-ATPase byouabain potentiates apoptosis by inducing perturbations in cell calciumhomeostasis: a protective role selective for Bcl-2. FASEB J 20: A119,2006.

32. Ribble D, Goldstein NB, Norris DA, Shellman YG. A simple techniquefor quantifying apoptosis in 96-well plates. BMC Biotechnol 5: 12, 2005.

33. Sakai H, Suzuki T, Maeda M, Takahashi Y, Horikawa N, Mi-namimura T, Tsukada K, Takeguchi N. Up-regulation of Na�,K�-ATPase alpha 3-isoform and down-regulation of the alpha1-isoform inhuman colorectal cancer. FEBS Lett 563: 151–154, 2004.

34. Sapia L, Palomeque J, Mattiazzi A, Petroff MV. Na�/K�-ATPaseinhibition by ouabain induces CaMKII-dependent apoptosis in adult ratcardiac myocytes. J Mol Cell Cardiol 49: 459–468, 2010.

35. Tagliaferri P, Yanagihara K, Ciardiello F, Talbot N, Flatow U,Benade L, Bassin RH. Effects of ouabain on NIH/3T3 cells transformedwith retroviral oncogenes and on human tumor cell lines. Int J Cancer 40:653–658, 1987.

36. Tao QF, Hollenberg NK, Price DA, Graves SW. Sodium pump isoformspecificity for the digitalis-like factor isolated from human peritonealdialysate. Hypertension 29: 815–821, 1997.

37. Wang Z, Zheng M, Li Z, Li R, Jia L, Xiong X, Southall N, Wang S,Xia M, Austin CP, Zheng W, Xie Z, Sun Y. Cardiac glycosides inhibitp53 synthesis by a mechanism relieved by Src or MAPK inhibition.Cancer Res 69: 6556–6564, 2009.

38. Yang P, Cartwright C, Efuet E, Hamilton SR, Wistuba II, Menter D,Addington C, Shureiqi I, Newman RA. Cellular location and expressionof Na�, K� -ATPase � subunits affect the anti-proliferative activity ofoleandrin. Mol Carcinog 53: 253–263, 2014.



by 10.220.33.1 on Novem


Dow

nloaded from


cardiac glycosides induced toxicity in human cells ... · cardiac glycosides induced toxicity in...

Documents