Cardioactive glycosides exert positive inotropic effects on cardiomyocytes through the inhibition of Na+/K+-ATPase. We showed previously that in human hepatoma cells, digoxin and ouabain increase the rate of the mevalonate cascade and therefore have Na+/K+-ATPase-independent effects. In the present study we found that they increase the expression and activity of 3-hydroxy-3 methylglutaryl-CoA reductase and the synthesis of cholesterol in cardiomyocytes, their main target cells. Surprisingly this did not promote intracellular cholesterol accumulation. The glycosides activated the liver X receptor transcription factor and increased the expression of ABCA1 (ATP-binding cassette protein A1) transporter, which mediates the efflux of cholesterol and its delivery to apolipoprotein A-I. By increasing the synthesis of ubiquinone, another derivative of the mevalonate cascade, digoxin and ouabain simultaneously enhanced the rate of electron transport in the mitochondrial respiratory chain and the synthesis of ATP. Mice treated with digoxin showed lower cholesterol and higher ubiquinone content in their hearts, and a small increase in their serum HDL (high-density lipoprotein) cholesterol. The results of the present study suggest that cardioactive glycosides may have a role in the reverse transport of cholesterol and in the energy metabolism of cardiomyocytes.
- ATP-binding cassette protein A1 (ABCA1)
- liver X receptor
Cardioactive glycosides are used in the treatment of atrial fibrillation and chronic heart failure. Owing to their inhibition of Na+/K+-ATPase (EC 18.104.22.168), they increase the intracellular concentration of Na+, decrease the extrusion of Ca2+ via the Na+/Ca2+ exchanger and exert a positive inotropic effect in cardiomyocytes . They also exert effects independently of Na+/K+-ATPase inhibition [2,3], in that they control the transcription of specific genes  and modulate the synthesis of steroid hormones [5,6]. In human hepatoma cells, we observed that digoxin and ouabain increased the synthesis of cholesterol by up-regulating both the activity and the expression of HMGCR (3-hydroxy-3-methylglutaryl-CoA reductase) (EC 22.214.171.124), which is the rate-limiting enzyme of the mevalonate pathway. This event was mediated by the activation of the SREBP (sterol-regulatory-element-binding protein)-2 and of the SCAP (SREBP-cleavage-activating protein) .
The mevalonate pathway mediates the synthesis of sterols, such as cholesterol, and non-sterol isoprenoid metabolites, which are incorporated in the tail of ubiquinone. Both cholesterol and ubiquinone have critical roles in cardiovascular disease [8,9]. Ubiquinone (or CoQ10), the electron shuttle between NADH dehydrogenase (Complex I, EC 126.96.36.199), or succinateCoQ reductase (Complex II, EC 188.8.131.52), and ubiquinol-cytochrome c reductase (Complex III, EC 184.108.40.206) , is necessary for mitochondrial respiration and ATP synthesis. Ubiquinone is lower in the cardiomyocytes of patients with ventricular dysfunction  and low serum levels of ubiquinone are predictive of higher mortality in chronic heart failure . Dietary supplementation with CoQ10 has improved systolic function and haemodynamic parameters in those patients .
Modulating the synthesis of cholesterol and ubiquinone in cardiomyocytes might thus have important implications on cardiac metabolism and function. Since digoxin and ouabain activate the mevalonate pathway in liver cells , we set out to investigate whether they may also affect the levels of cholesterol and ubiquinone in cardiomyocytes, the main targets of this class of drugs.
FBS (fetal bovine serum) and culture medium were supplied by BioWhittaker; plasticware for cell cultures was from Falcon. Mevastatin was purchased from Calbiochem. Electrophoresis reagents were obtained from Bio-Rad Laboratories. Digoxin, ouabain, zaragozic acid and the BCA (bicinchoninic acid) kit for protein measurement were from Sigma. When not otherwise specified, all of the other reagents were purchased from Sigma.
H9c2 cells, a rat cardiomyocyte cell line which expresses α1, α2 and α3 subunits of Na+/K+-ATPase (Supplementary Figure S1 at http://www.BiochemJ.org/bj/447/bj4470301add.htm), was cultured in Dulbecco's modified Eagle's medium containing 10% (v/v) FBS (Sigma), 1% (v/v) penicillin/streptomycin and 1% (v/v) L-glutamine. The culture medium contained 32 μg/ml cholesterol, 8 μg/ml HDL (high-density lipoprotein) cholesterol, 13 μg/ml LDL (low-density lipoprotein) cholesterol and 55 μg/ml triacylglycerols. Total cholesterol, HDL cholesterol and triacylglycerols were measured with the OSR6516, OSR6187 and OSR61118 kits respectively (Olympus System Reagent), using an Olympus Analyzer spectrophotometer (Olympus Europa Holding). The results were calculated according to the titration curve of each kit. LDL cholesterol was calculated with the Friedewald formula.
Measurement of cholesterol and ubiquinone de novo synthesis
The de novo synthesis of cholesterol and ubiquinone was measured by radiolabelling cells with 1 μCi/ml [3H]acetate (3600 mCi/mmol; Amersham GE Healthcare, Piscataway, NJ), or with 1 μCi/ml [14C]mevalonic acid (67 mCi/mmol; Amersham GE Healthcare), as reported in . Cholesterol and ubiquinone synthesis was expressed as fmol of [3H]cholesterol or [14C]cholesterol/106 cells and as fmol of [3H]ubiquinone/106 cells, according to the titration curves. For the cholesterol titration curve, a stock solution of [3H]cholesterol (1 mCi/ml, 7 Ci/mmol, PerkinElmer) or [14C]cholesterol (0.04 mCi/ml, 60 mCi/mmol) was serially diluted and counted by liquid scintillation. The radioactivity (c.p.m.), subtracted from the radioactivity of the blank (distilled water), was related to the concentration of each solution, yielding a specific activity of 5852 c.p.m./pmol for [3H]cholesterol and 130 c.p.m./pmol for [14C]cholesterol. In the absence of commercially available [3H]ubiquinone, the titration curve was performed by incubating 106 cells for 24 h with 0.25, 0.5, 0.75 or 1 μCi/ml [3H]acetate. The lipids were extracted and separated by TLC. The spot corresponding to ubiquinone was isolated and the amount of [3H]ubiquinone was quantified by liquid scintillation. The assay was linear between 0.25 and 1 μCi/ml (Supplementary Figure S2 at http://www.BiochemJ.org/bj/447/bj4470301add.htm), which falls in the range of [3H]ubiquinone recovered from H9c2 cells radiolabelled with [3H]acetate.
The release of LDH (lactate dehydrogenase) in the extracellular medium, a marker of cell damage, and the positivity to annexin V–FITC, a marker of apoptosis, were measured as reported previously .
Measurement of HMGCR activity and expression
The activity and expression of HMGCR were measured in microsomal fractions collected by centrifugation of cell lysates at 40000 rev./min in an optima L-90K Beckman Coulter ultracentrifuge using a 70.1 Ti rotor for 1 h (4°C), as described previously .
RT-PCR (real-time PCR)
Total RNA was extracted with TRIzol® (Invitrogen); a 5 μg portion was reverse-transcribed by 200 units of M-MLV (Moloney murine leukaemia virus) enzyme (Invitrogen). Quantitative RT-PCR was carried out using IQ™ SYBR Green Supermix (Bio-Rad Laboratories), according to the manufacturer's instructions. The relative quantification of each sample was performed comparing the PCR products of each gene with the housekeeping gene product (actin), with the iQ™ 5 Optical System Software (Bio-Rad Laboratories). Results were expressed in arbitrary units. For each gene, the expression in untreated cells was considered to be ‘1’. Primers sequences are shown in Supplementary Table S1 at http://www.BiochemJ.org/bj/447/bj4470301add.htm.
Isolation of mitochondria and measurement of Complex I–III activity
The activity of ubiquinone-dependent and ubiquinone-independent Complex I–III was measured on non-sonicated extracted mitochondria as previously reported .
ATP detection assay
The ATP level in mitochondria was measured with the ATP Bioluminescent Assay Kit (Sigma), using a Magic Lite Analyzer (Ciba Corning Diagnostic). ATP was quantified as arbitrary light units and converted into nmol of ATP per mg of mitochondrial proteins, according to a calibration curve set earlier.
Cholesterol loading and spectrophotometric measurement of intracellular cholesterol
The cholesterol loading was performed by preparing cholesterol–β-methyl-cyclodextrin complexes as described previously . The amount of intracellular cholesterol was measured in cell lysates with the OSR6516 kit. The results were expressed as μg of cholesterol per mg of cell proteins.
Measurement of extracellular cholesterol
The amount of cholesterol in the cell culture medium after cholesterol loading was measured spectrophotometrically by means of the OSR6516 kit.
To evaluate cholesterol efflux, cells were incubated with 1 μCi/ml [3H]cholesterol for 1 h, washed five times with PBS and grown in fresh medium for 24 h. The cell culture medium was collected, cholesterol was extracted in methanol/hexane and resolved by TLC . The amount of [3H]cholesterol recovered was quantified by liquid scintillation. In parallel, intracellular [3H]cholesterol was measured in cell lysates. The ratio between extracellular [3H]cholesterol and total [3H]cholesterol was calculated as c.p.m. in medium/(c.p.m. in medium+c.p.m. in cells) .
To measure the efflux of cholesterol derived from endogenous synthesis, cells were incubated for 24 h with 1 μCi/ml [3H]acetate, then the cholesterol in the cell culture medium was extracted and the amount of [3H]cholesterol was quantified as reported above. For both labelling assays, the [3H]cholesterol was expressed as pmol/106 cells, according to the titration curve.
Western blot analysis
Cells were lysed in sample buffer [25 mM Hepes/KOH, pH 7.5, 135 mM NaCl, 1% (v/v) Nonidet P40, 5 mM EDTA, 1 mM EGTA, 1 mM ZnCl2, 50 mM NaF and 10% (v/v) glycerol] supplemented with protease inhibitor cocktail set III (Calbiochem), 2 mM PMSF and 1 mM Na3VO4. A 25 μg portion of proteins from whole-cell lysates was separated by SDS/PAGE and probed with the following antibodies: anti-ABCA1 [ABC (ATP-binding cassette) protein A1] (Abcam), anti-ABCG1 (Santa Cruz Biotechnology), anti- SR-BI (scavenger receptor class B, type I) (Novus Biologicals), anti-LXR (liver X receptor) β (Abcam), anti-LXRα (Abcam) and anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (Santa Cruz Biotechnology), followed by the secondary peroxidase-conjugated antibody (Bio-Rad Laboratories). Proteins were detected by enhanced chemiluminescence (PerkinElmer). Densitometric analysis of Western blots was performed using ImageJ software (http://rsb.info.nih.gov/ij/).
Flow cytometry analysis
Cells were washed with PBS, detached with cell dissociation solution (Sigma), fixed with 0.2 ml of 2% (w/v) paraformaldehyde, rinsed with 0.25% PBS-BSA and incubated for 45 min at 4°C with the anti-ABCA1 antibody (Abcam). After two washes in PBS-BSA, cells were incubated for 30 min at 4°C with an anti-(rabbit FITC)-conjugated antibody, then washed again and resuspended in PBS-BSA. Control experiments included incubation with non-immune isotypic antibody, followed by the secondary antibody. Samples were analysed with a FACS-Calibur flow cytometer (Becton Dickinson), using a 530 nm band-pass filter. For each analysis, 100000 events were collected and processed with Cell Quest software (Becton Dickinson).
EMSA (electrophoretic mobility-shift assay)
The probe containing the DR-4 (direct repeat response element-4) for LXR, validated with Genomatix Software (http://www.genomatix.de/; Munich, Germany), was: 5′-GCGACCCCAGTGATATCCCGTCGTC-3′. The putative DR-4 site is underlined. The probe was labelled with 50 μCi of [γ-32P]ATP (3000 Ci/mmol; Amersham Bioscience), using a T4 polynucleotide kinase (Roche). A total of 10 μg of proteins from nuclear extracts were incubated for 20 min at 4°C with the 32P-labelled probe, as described previously . As a control of specificity, nuclear extracts from cells stimulated with the LXR activator TO-901317 were incubated with a mutated labelled probe (5′-GCGACCTATTGTATATCTGTTCGTC-3′) or with the wild-type labelled probe plus a 100-fold excess of wild-type unlabelled probe. The DNA–protein complex was separated on a non-denaturing 4% polyacrylamide gel, then the gel was dried and autoradiographed after 72 h.
Cells were rinsed with fixation buffer [500 mM Hepes/KOH, pH 7.9, 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA and 11% (v/v) formaldehyde], washed twice with PBS, centrifuged for 5 min at 1000 g at 4°C and resuspended in lysis buffer [50 mM Tris/HCl, pH 8.0, 5 mM EDTA and 1% (w/v) SDS]. After sonication (five pulses of 10 s at a power setting of 10, using a SONOPULS Bandelin instrument), a 200 μl volume of each sample was taken as input. The remaining lysates were pre-treated for 2 h at 4°C with Protein G–Sepharose magnetic beads (Invitrogen), then divided into three portions (incubated with anti-LXRβ or anti-LXRα, with generic IgG or without any antibody respectively) and immunoprecipitated overnight at 4°C. The recovered DNA was washed, eluted with the elution buffer [0.1 M NaHCO3 and 1% (w/v) SDS], heated at 65°C for 6 h and incubated with proteinase K for 1 h at 55°C. Samples were cleaned by Qiaquick columns (Qiagen) and analysed by RT-PCR. The putative LXRE (LXR response element) sites on the ABCA1 rat promoter, as well as the non-specific sequences upstream of the promoter, used as negative internal controls, were validated with the Genomatix Software (http://www.genomatix.de/). Primer sequences are reported in Supplementary Table S1.
Treatment of mice and in vivo assays
The in vivo experiments were performed in conformity with the Public Health Service Policy on Human Care and Use of Laboratory Animals and were approved by the Institutional Review Board. Mice were housed in an environmentally controlled room (24°C, 12 h light/dark cycle) and provided with food and fresh water ad libitum. A standard diet for mouse maintenance was used (Mucedola srl, 4RF21 Certificate PF1610). FVB male mice (1 month old) were injected daily intraperitoneally with 1 mg of digoxin/kg, diluted in 0.9% NaCl solution, for 5 or 8 days. Before killing, mice were heparinized and anaesthetized with 250 mg of tribromoethanol/kg. Blood was collected through decapitation.
Hearts were excised and rinsed in ice-cold Dulbecco's PBS. Atria were removed, ventricular tissue and apex were collected in RNAlater solution (Ambion) or ice-cold 1 M NaCl and frozen at −80°C. RNA was extracted and used for RT-PCR experiments. Lipids were extracted from 50 mg of ventricular homogenates as described previously . Cholesterol was measured spectrophotometrically with the OSR6516 kit (Olympus System Reagent) and the results were expressed as μg of cholesterol/mg of heart tissue. The amount of ubiquinone was determined spectrophotometrically  and the absorbance was converted into μmol of ubiquinone/mg of heart tissue, using a calibration curve prepared previously.
Plasma total cholesterol and HDL cholesterol were determined with OSR6516 and OSR6187 kits respectively (Olympus System Reagent) and the results were expressed as mg/100 ml. Plasma digoxin was quantified with the Access Digoxin Kit (Beckman Coulter) and the results were expressed as ng/ml.
Intracellular triacylglycerol measurement
The amount of triacylglycerol was measured in mouse heart homogenates and in H9c2 cell lysates using the Triglyceride Quantification Kit (Abcam). Results were expressed in nmol/mg of cell or tissue proteins, according to the calibration curve set previously.
All of the data in the text and Figures are provided as means±S.D. The results were analysed by a one-way ANOVA and Tukey's test. P<0.05 was considered significant.
Digoxin and ouabain increase the synthesis of cholesterol and ubiquinone in cardiomyocytes by up-regulating the expression of HMGCR
H9c2 cells were treated with digoxin or ouabain and labelled with [3H]acetate. Both glycosides increased the de novo synthesis of cholesterol and ubiquinone in a dose- (Figure 1a) and time (Figure 1b) dependent manner. Since a 24 h incubation with 10 nM digoxin or ouabain elicited a detectable increase in the mevalonate cascade, but no cytotoxicity, as determined by LDH release and annexin V–FITC positivity (Supplementary Figure S3 at http://www.BiochemJ.org/bj/447/bj4470301add.htm), this experimental condition was chosen for all of our assays.
To ascertain whether the effects of digoxin and ouabain were limited to HMGCR, we studied other genes of the mevalonate pathway, under the transcriptional control of SREBP-2 . Both drugs up-regulated HMGCS (3-hydroxy-3-methylglutaryl-CoA synthase) (Figure 2c), although to a lower extent than HMGCR. We did not detect other effects exerted by either digoxin and ouabain in the analysed genes (Figure 2c).
To confirm that glycosides specifically modulate the first steps of the mevalonate pathway, we measured the synthesis of cholesterol in cells labelled with [3H]acetate (Figure 2d), a substrate upstream of the HMGCR enzyme, or with [14C]mevalonic acid (Figure 2e), the product of HMGCR, in the absence or presence of the HMGCR inhibitor mevastatin. As expected, mevastatin decreased the synthesis of cholesterol in cells labelled with [3H]acetate, but not in cells labelled with [14C]mevalonic acid. Mevastatin prevented the increase of cholesterol triggered by digoxin and ouabain in cells labelled with [3H]acetate (Figure 2d) and had no effect in cells labelled with [14C]mevalonic acid (Figure 2e). The squalene synthase inhibitor zaragozic acid , which blocks the mevalonate cascade downstream of HMGCR, also prevented the glycoside-elicited cholesterol increase, without a change in the increase in ubiquinone (Supplementary Figure S4 at http://www.BiochemJ.org/bj/447/bj4470301add.htm).
Overall, these data suggest that digoxin and ouabain up-regulate the mevalonate pathway probably by acting on HMGCS and HMGCR.
Digoxin and ouabain enhance the efficiency of the mitochondrial electron transport chain and ATP synthesis
Besides cholesterol, digoxin and ouabain also increased the synthesis of another mevalonate pathway derivative, ubiquinone (Figure 1), the only electron carrier between mitochondrial Complexes I or II and Complex III. Therefore glycosides enhanced the electron flow in mitochondria of H9c2 cells (Figure 3a); this increase was found both in the presence of the Complex I inhibitor rotenone (when the electron transport relies entirely on ubiquinone) and in its absence (i.e. when the electron flow depends on Complex I plus ubiquinone). The enhanced electron flow resulted in increased levels of ATP, an effect prevented by the F1Fo-ATP synthase inhibitor oligomycin (Figure 3b).
Digoxin and ouabain prevent the intracellular accumulation of cholesterol
When cholesterol levels rise in mammalian cells, HMGCR activity is down-regulated . When loaded with exogenous cholesterol (Figure 4a), H9c2 cells had a low activity of HMGCR (Figure 4b). Unexpectedly, in spite of the increased endogenous cholesterol synthesis (Figures 1 and 2), we detected low levels of intracellular cholesterol after treatment with digoxin and ouabain (Figure 4a). Moreover, when cells were treated with glycosides plus cholesterol, they showed lower amounts of intracellular cholesterol than cells treated with cholesterol alone (Figure 4a). HMGCR activity remained high in the presence of digoxin and ouabain (Figure 4b), which was consistent with the low level of intracellular cholesterol.
Of note, we observed a progressive cholesterol accumulation in the culture medium from cells treated with digoxin and ouabain (Figure 4c).
Digoxin and ouabain promote the efflux of cholesterol via the ABCA1 transporter
To investigate whether digoxin and ouabain stimulate the efflux of intracellular cholesterol, we pulsed H9c2 cells with [3H]cholesterol and we measured the amount of [3H]cholesterol recovered after 24 h in the culture medium (Figure 5a): both digoxin and ouabain increased the efflux of [3H]cholesterol from H9c2 cells. This increase was further enhanced by apoA-I (apolipoprotein A-I), a constituent of HDL particles, and was prevented by glyburide (Figure 5a), an inhibitor of the ABCA1 transporter . The amount of intracellular [3H]cholesterol in H9c2 cells did not change significantly under all of the experimental conditions (Supplementary Figure S5a at http://www.BiochemJ.org/bj/447/bj4470301add.htm). The relative amount of cholesterol effluxed by H9c2 cells was very small compared with the total labelling (Supplementary Figure S5b), which is consistent with mammalian cells in general [21,22].
In a parallel set of experiments, we labelled cells with [3H]acetate and measured [3H]cholesterol, as an index of de novo synthesized cholesterol, in the medium after 24 h (Figure 5b). Under these circumstances, digoxin and ouabain also promoted cholesterol efflux. Moreover, apoA-I protein further enhanced the efflux and glyburide decreased it.
ABCA1 was expressed in H9c2 cells at baseline and was increased by digoxin and ouabain, both in whole-cell lysates (Figure 5c) and on the cell surface (Figure 5d). ABCG1 and SR-BI, two other transporters involved in cholesterol efflux, were either only slightly increased by digoxin and ouabain (ABCG1) or not affected at all (SR-BI) (Supplementary Figure S6 at http://www.BiochemJ.org/bj/447/bj4470301add.htm).
Digoxin and ouabain increase the transcription of ABCA1 by activating LXR
To find out how glycosides increased the expression of ABCA1 and ABCG1, we investigated which transcription factors may up-regulate these transporters.
The promoters of ABCA1 [23,24] and ABCG1  contain sterol-responsive elements, which are activated by the heterodimers LXRα–RXR (retinoid X receptor) and LXRβ–RXR. The total amount of LXRβ and LXRα was not affected by either glycosides or mevastatin (Figure 6a). LXR did not bind to its DNA target sequence in untreated H9c2 cells (Figure 6b). Conversely, digoxin and ouabain promoted the binding of LXR to DNA, an event prevented by mevastatin and mimicked by the LXR activator TO-901317  (Figure 6b). As shown by chromatin immunoprecipitation assays, glycosides induced the binding of either LXRα and LXRβ to the LXRE sequence of the ABCA1 promoter (Figure 6c).
Mice treated with digoxin have a lower amount of cholesterol and a higher amount of ubiquinone in their hearts
The effects of digoxin were studied in mice treated for 5 and 8 days with 1 mg of the drug/kg per day. At the end of these time periods, blood digoxin levels were significantly higher when compared with untreated mice (Table 1) and within the accepted human therapeutic range [26,27]. The drug increased HMGCR mRNA in a time-dependent fashion (Figure 7a) and decreased cholesterol content (Figure 7b) in heart tissue. We did not detect significant differences in the serum levels of cholesterol in treated and untreated mice; however, HDL cholesterol was slightly higher in the treated group (Table 1). A high degree of variability was observed in the levels of cardiac ubiquinone, in digoxin-treated mice and in controls. Individual variability notwithstanding, the hearts of animals treated with digoxin in general showed higher levels of ubiquinone.
Since LXR may increase the expression of several genes involved in the synthesis of fatty acids and triacylglycerols , we set out to see whether these genes were activated in the hearts of digoxin-treated animals. We did not observe any changes in the expression of lipogenic genes (acetyl-CoA carboxylase, fatty acid synthase, glycerol-3-phosphate acyltransferase and stearoyl-CoA desaturase 1); only SREBP-1c was decreased after digoxin treatment (Figure 7d). Intracellular triacylglycerol levels were highly variable, but did not differ in each experimental group (Figure 7e). Similar findings were obtained in vitro in H9c2 cells after a 24 h incubation with digoxin or ouabain (Supplementary Figure S7 at http://www.BiochemJ.org/bj/447/bj4470301add.htm).
Cardioactive glycosides are specific inhibitors of Na+/K+-ATPase and positive inotropic drugs . The impact of cardioactive glycosides on cholesterol metabolism in cardiac tissue has not been investigated. We used H9c2 cells, which share some ultrastructural features with skeletal muscle [29–31], but retain the key electrophysiological and biochemical properties of cardiomyocytes [29,32], making them a useful tool for the in vitro investigation of cardioactive drugs.
Digoxin and ouabain enhanced the activity of the mevalonate pathway in a dose- and time-dependent manner in these cells. This increase was accompanied by the up-regulation of the rate-limiting enzyme HMGCR, which is under the transcriptional control of SREBP-2. Although SREBP-2 induces a transcriptional activation of several enzymes involved in cholesterol biosynthesis , in H9c2 cells only the HMGCR and HMGCS genes were up-regulated by both digoxin and ouabain. Such specificity suggests that glycosides may act not only by simply activating SREBP-2 , but also by recruiting other transcription factors and/or co-activators, which in turn co-operate with SREBP-2 on HMGCR and HMGCS promoters.
The effects of glycosides on the HMGCR step, or on one of the first steps in the mevalonate cascade, were confirmed by functional assays: the increase in cholesterol was indeed prevented by the HMGCR inhibitor mevastatin in cells labelled with [3H]acetate (i.e. the metabolite upstream of HMGCR), but not in cells labelled with [14C]mevalonic acid (i.e. the product of HMGCR) or treated with the inhibitor of squalene synthase, zaragozic acid.
Glycosides increased the synthesis of ubiquinone in H9c2 cells, thus enhancing the efficiency of mitochondrial electron transport and ATP synthesis. Both adequate intracellular Ca2+ levels and a constant supply of ATP are necessary to support an efficient contractile performance. The increase of intracellular Ca2+ is a well-documented effect of digoxin [1–3]; the increase in aerobic metabolism and ATP synthesis observed in the present study may represent an additional mechanism by which digoxin exerts its positive inotropic effects.
Cholesterol levels also affect other crucial cellular functions in cardiomyocytes: increased plasma membrane cholesterol modifies the affinity of Na+/K+-ATPase for inhibitors and alters the activity of Ca2+-ATPase, Na+/Ca2+-exchanger and Na+/K+-ATPase, three pumps that critically affect intracellular Ca2+ levels . As digoxin and ouabain increase the synthesis of cholesterol, one might expect that these cells ‘fill up’ with cholesterol. Such occurrence, however, has never been described in patients treated with digoxin. Indeed, H9c2 cells treated with digoxin and ouabain had low levels of intracellular cholesterol and showed a time-dependent cholesterol accumulation in the culture medium. The cholesterol transport assays that use [3H]cholesterol or the cholesterol precursor [3H]acetate suggest that digoxin and ouabain stimulate the efflux of both exogenously added or endogenously synthesized cholesterol, thereby preventing any accumulation of cholesterol in cardiomyocytes. From this perspective, one might consider this ‘cholesterol-buffering’ effect of glycosides to be a type of cardioprotective action.
A crucial mechanism of cholesterol efflux in mammalian tissues is its active transport via the ABC transporter ABCA1, which transfers cholesterol from the cytosol to apoA-I and mediates the first step of HDL assembly [23,33,34]. Interestingly, both digoxin and ouabain increased the amount of ABCA1 in H9c2 cells. In addition, ABCG1, a second transporter which mediates the efflux of cholesterol from the cells, was slightly increased. As we worked with a culture medium that contains HDLs, we hypothesized that glycosides might promote the delivery of cholesterol to the apoA-I present in the culture medium through ABCA1. When we used the ABCA1 inhibitor glyburide, cholesterol efflux decreased to control levels, even under maximal stimulation (i.e. when digoxin and ouabain were co-incubated with apoA-I): this suggests that in H9c2 cells, most of the glycoside-induced cholesterol efflux is mediated by ABCA1, whereas ABCG1 plays a secondary role.
Multiple sterol-responsive elements, activated by LXRα–RXR and LXRβ–RXR heterodimers in response to oxysterols, are contained in the promoter of the ABCA1 gene [23,24]. Giving the cell a cholesterol load induces the transcriptional activity of LXRs, suggesting that free cholesterol, or a cholesterol-derived metabolite produced within cells, can up-regulate ABCA1 . We hypothesize that a similar mechanism might also work with glycosides, which in our hands increased the synthesis of cholesterol, the binding of LXR on DNA target sequences and the activation of the ABCA1 promoter.
To our knowledge this is the first report that shows that cardioactive glycosides stimulate the reverse transport of cholesterol by increasing the expression of ABCA1 and the delivery of cholesterol to apoA-I, and therefore facilitate one of the first steps in the assembly of HDLs. This event may have physiological relevance, since an efficient reverse cholesterol transport and high levels of HDLs are protective factors against cardiovascular disease.
The concentration of digoxin used in our in vitro experiments (10 nM) falls within the clinically accepted therapeutic range [26,27]. At these concentrations, the well-known inotropic effects of glycosides are mainly mediated by their binding to the α2 subunit of Na+/K+-ATPase , which in rats has a higher affinity for ouabain than α1 subunit . Since α1 and α2 subunit levels were equal in H9c2 cells, it is likely that in our model the Na+/K+-ATPase-dependent effects of glycosides were mediated by the interaction with the α2 subunit. H9c2 cells mirror rat tissues, which have a mix of ouabain-insensitive (i.e. containing solely the α1 isoform) Na+/K+-ATPases and ouabain-sensitive (i.e. containing the α2 and α3 isoforms) ones . Rodents are more resistant than humans to the effects of digoxin on Na+/K+-ATPase [37,38]. Nonetheless, after a few days of treatment, in our animal models, digoxin reached the same levels as the accepted clinically therapeutic ones in humans. As such, it began to exert its effects on the mevalonate pathway in hearts, as the increased expression of HMGCR, the progressive decrease in cholesterol and increase in ubiquinone synthesis would indicate. Since in the liver, LXR up-regulates SREBP-1c and co-operates with it to promote the synthesis of fatty acids and triacylglycerols , we set out to investigate whether these events occurred in the heart as well. Differently from what happens in the liver, in our model digoxin decreased the expression of SREBP-1c. Such a decrease may prevent the up-regulation of lipogenic genes and the accumulation of triacylglycerols in cardiac myocytes, making the onset of a lipid-induced cardiomyopathy an unlikely event.
The present study sheds some new light on the metabolic effects of glycosides in cardiac myocytes. By enhancing the activity of the mevalonate pathway, glycosides increase the cellular levels of ubiquinone, the aerobic metabolism and the synthesis of ATP of the cells, three events critical in the control of contractile performance. In addition, glycosides finely control the levels of cholesterol in the cell. On one hand, they provide cells with a metabolite (cholesterol) necessary for the maintenance of membrane homoeostasis and the proper conformation and activity of transporters, channels and receptors. On the other, they prevent the intracellular accumulation of the same metabolite and facilitate its delivery to HDLs. Beside being used as therapeutic agents, ouabain- and digoxin-like substances are also endogenously produced; acting at nanomolar/picomolar concentrations, these endogenous glycosides have been implicated in the control of blood pressure and in the cardiovascular remodelling of chronic heart failure . One could surmise that the same effects observed with exogenously administered glycosides might be exerted by the ones that are endogenously produced. From this perspective, glycosides might indeed be novel regulators of cholesterol homoeostasis.
Ivana Campia performed the in vitro experiments, analysed the data and wrote the paper; Valentina Sala and Christian Leo performed the animal treatment and the in vivo assays; Joanna Kopecka, Nico Mitro and Costanzo Costamagna performed the in vitro experiments; Donatella Caruso and Gianpiero Pescarmona supervised the in vitro experiments and analysed the results; Tiziana Crepaldi supervised the in vivo experiments and analysed the results; Dario Ghigo and Amalia Bosia analysed the data and reviewed the paper prior to submission; Chiara Riganti conceived the study and wrote and reviewed the paper prior to submission.
This work was supported by the Fondazione Internazionale Ricerche Medicina Sperimentale (FIRMS), the Compagnia di San Paolo [grant number 3856 SD/CV], the Regione Piemonte Ricerca Sanitaria Finalizzata 2009 [grant number 204/30.04.2009] and the Giovanni Armenise-Harvard Foundation.
Abbreviations: ABC, ATP-binding cassette; ABCA1, ABC protein A1; apoA-I, apolipoprotein A-I; DR-4, direct repeat response element-4; EMSA, electrophoretic mobility-shift assay; FBS, fetal bovine serum; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HDL, high-density lipoprotein; HMGCR, 3-hydroxy-3-methylglutaryl-CoA reductase; HMGCS, 3-hydroxy-3-methylglutaryl-coA synthase; LDH, lactate dehydrogenase; LDL, low-density lipoprotein; LXR, liver X receptor; LXRE, LXR response element; RT-PCR, real-time PCR; RXR, retinoid X receptor; SR-BI, scavenger receptor class B, type I; SREBP, sterol-regulatory-element-binding protein
- © The Authors Journal compilation © 2012 Biochemical Society