FASEB J. Thermo Fisher Scientific
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published as doi: 10.1096/fj.07-8357com.
 (The FASEB Journal. 2007;21:3133-3141.)
© 2007 FASEB
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
fj.07-8357comv1
21/12/3133    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sainte-Marie, Y.
Right arrow Articles by Jaisser, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sainte-Marie, Y.
Right arrow Articles by Jaisser, F.

Conditional glucocorticoid receptor expression in the heart induces atrio-ventricular block

Yannis Sainte-Marie*,{dagger},{ddagger}, Aurelie Nguyen Dinh Cat*,{dagger},{ddagger},1, Romain Perrier§,||,1, Laurence Mangin,#,**, Christelle Soukaseum*,{dagger},{ddagger}, Michel Peuchmaur{dagger}{dagger},{ddagger}{ddagger}, François Tronche{dagger},§§, Nicolette Farman*,{dagger},{ddagger}, Brigitte Escoubet*,{dagger},{ddagger},#, Jean-Pierre Benitah§,|| and Frederic Jaisser*,{dagger},{ddagger},2

* Inserm, U772, Paris, France;

{dagger} Collège de France, Paris, France;

{ddagger} Université Paris Descartes, Paris, France;

§ Inserm, U637 IFR3, Montpellier, France;

|| Université Montpellier I, Montpellier, France;

Inserm, U698, Paris, France;

# Assistance Publique-Hôpitaux de Paris, Hôpital Bichat, Paris, France;

** Université Paris 7, Faculté de Médecine X. Bichat, Paris, France;

{dagger}{dagger} Assistance Publique-Hôpitaux de Paris, Hôpital R. Debré, Service de Pathologie, Paris, France;

{ddagger}{ddagger} Université Paris 7, EA3102, Paris, France; and

§§ CNRS FRE2401, Paris, France

2Correspondence: INSERM U772; Collège de France, 11 place Marcelin Berthelot, 75231 Paris, France. E-mail: frederic.jaisser{at}college-de-france.fr


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Corticosteroid hormones (aldosterone and glucocorticoids) and their receptors are now recognized as major modulators of cardiovascular pathophysiology, but their specific roles remain elusive. Glucocorticoid hormones (GCs), which are widely used to treat acute and chronic diseases, often have adverse cardiovascular effects such as heart failure, hypertension, atherosclerosis, or metabolic alterations. The direct effects of GC on the heart are difficult to evaluate, as changes in plasma GC concentrations have multiple consequences due to the ubiquitous expression of the glucocorticoid receptor (GR), resulting in secondary effects on cardiac function. We evaluated the effects of GR on the heart in a conditional mouse model in which the GR was overexpressed solely in cardiomyocytes. The transgenic mice displayed electrocardiogram (ECG) abnormalities: a long PQ interval, increased QRS and QTc duration as well as chronic atrio-ventricular block, without cardiac hypertrophy or fibrosis. The ECG alterations were reversible on GR expression shutoff. Isolated ventricular cardiomyocytes showed major ion channel remodeling, with decreases in INa, Ito, and IKslow activity and changes in cell calcium homeostasis (increase in Cal, in Ca2+ transients and in sarcoplasmic reticulum Ca2+ load). This phenotype differs from that observed in mice overexpressing the mineralocorticoid receptor in the heart, which displayed ventricular arrhythmia. Our mouse model highlights novel effects of GR activation in the heart indicating that GR has direct and specific cardiac effects in the mouse.—Sainte-Marie, Y., Cat, A. N. D., Perrier, R., Mangin, L., Soukaseum, C., Peuchmaur, M., Tronche, F., Farman, N., Escoubet, B., Benitah, J-P., Jaisser, F. Conditional glucocorticoid receptor expression in the heart induces atrio-ventricular block.


Key Words: corticosteroids • electrical remodeling • ion channel • conditional model


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
THE PATHOPHYSIOLOGICAL INFLUENCE of corticosteroid hormones (aldosterone, glucocorticoid hormones) on heart function has been highlighted in recent years. An impressive amount of experimental and clinical data has accumulated on deleterious effects of abnormal or excessive corticosteroid signaling, but their mechanisms are far from being fully understood. In particular, the role of cardiac mineralocorticoid versus glucocorticoid receptors remains elusive. Glucocorticoids (GCs) are among the most frequently used drugs for a wide range of disorders involving inflammation and/or immune activation such as asthma, chronic obstructive pulmonary disease, acute and chronic autoimmune diseases, and for treatment of organ transplant recipients. The chronic therapeutic use of GCs is associated with several side effects, including adverse cardiovascular events (1) , such as hypertension and left ventricular hypertrophy (2 , 3) . GCs also induce metabolic syndrome with hyperglycemia, dyslipidemia, and obesity, and are associated with early and progressive atherosclerosis (1) . GC treatment therefore contributes to a cluster of cardiovascular risk factors (4) , including heart failure (5) . Conversely, several observations suggest that GCs help maintain normal cardiac contraction. Dexamethasone (Dex) treatment enhances the development of contractile tension and increases contraction and relaxation velocities in cardiac muscle (6) . The decrease in contractile force of rat papillary muscle induced by adrenalectomy is prevented by Dex treatment (7) by modulating membrane Ca2+ transport (8 , 9) and activity of K+ channels (6 , 7 , 10) . GC may therefore contribute to cardiac electrical activity. Short-term Dex treatment has also been shown to decrease resting heart rate in healthy human volunteers (11) .

The direct and specific effects of GC on the heart thus remain somewhat unclear. They may result from confounding local and systemic effects, and from interference with the closely related mineralocorticosteroid system. Indeed, corticosteroids bind to nuclear receptors—glucocorticoid receptor (GR) and mineralocorticoid receptor (MR)—both of which are transcription factors expressed in cardiomyocytes (12) . The ligand/receptor interactions are complex, as both aldosterone and GC can activate cardiac MR, thereby directly affecting heart function (13 , 14) . Some of the cardiac or peripheral effects of GC may be mediated at least in part by MR activation (12 , 15) .

To determine the specific cardiovascular actions of corticosteroid receptors, we have designed a general experimental strategy for modulating steroid hormone receptors specifically in the heart, independent of systemic effects and without affecting circulating ligand levels (16) . In this study we generated a transgenic mouse model with conditional inducible cardiac-specific expression of human GR (hGR). This targeted approach precluded secondary effects due to general GC-induced alterations, allowing us to investigate the specific role of GR in cardiomyocytes. The observed effects of GR cardiac overexpression differ from those reported with our previous model based on conditional MR overexpression in the heart (14) . No major structural cardiac remodeling or early death was observed. Isolated adult cardiomyocytes showed ionic current and cell Ca2+ homeostasis remodeling, consistent with the previously reported effects of GCs. Electrophysiological phenotyping indicated that cardiac hGR overexpression resulted in conduction defects, with high-degree atrio-ventricular block (AVB) illustrating unknown effects of GR activation in the heart.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
This study conforms to European Union Council Directives (86/609/EEC) regarding the care and use of laboratory animals. The full-length human GR{alpha} coding sequence and the BGH polyadenylation signal excised from pcDNAhGR (kindly provided by M-E. Rafestin-Oblin) was inserted in a modified pBI3 vector (17) (see supplementary methods). One of the tetO-hGR founders obtained after pronuclear injection of the tetO-hGR construct into FVBN fertilized eggs was extensively characterized and used for phenotypic analyses. Double-transgenic {alpha}MHC-tTA/tetO-hGR mice (DT), with cardiac-specific overexpression of hGR, were obtained by crossing the tetO-hGR founder with the previously characterized {alpha}MHC-tTA mouse strain (18) , kindly provided by G. I. Fishman and maintained in a B6D2 genetic background. Genotypes were determined by PCR amplification from tail genomic DNA. All studies were done on age- and gender-matched littermate mice. Phenotypic analysis indicated that monotransgenic tetO-hGR, {alpha}MHC-tTA, and wild-type littermates were similar and were therefore used as controls. Survival of DT was normal up to 6 months of age. All mice were fed standard chow (A04, Scientific Animal Food Engineering, Epinay sur Orge, France) and drank tap water ad libitum. When required, food containing 1g/Kg doxycycline (Dox) was given to mothers from breeding until weaning and to their offspring or to adult animals. All experiments were performed in adrenal-intact mice, except for dexamethasone binding, which requires adrenalectomy (adx) prior to tissue sampling and binding assay (see below). Most mice were studied at the age of 2 months to avoid complexity related to long-term adaptations/compensations of GR expression.

Western blot analysis were done on heart apex homogenized in SDS lysis buffer (1% SDS, 10 mM Tris HCl pH=7.4, 1 mM orthovanadate) containing protease cocktail inhibitor (Roche Diagnostics, Meylan, France). Tissue lysates were then centrifuged at 12000 g for 5 min and protein concentration in supernatants was determined by the Bradford method (Bio-Rad, Marnes-la-Coquette, France). Proteins (10 to 15 µg) were separated by SDS-PAGE and transferred to nitrocellulose membrane (Amersham, Biosciences Europe GMBH, Orsay, France). Membranes were stained with Ponceau red, incubated for 2 h in 5% fat-free milk in TBS-Tween, then overnight at 4°C with appropriate antibodies: specific anti-GR against hGR (sc1003, E20) or against mGR (sc1004, M20) (both 1/1000, SantaCruz, Tebu-Bio, Paris, France), SERCA2a (1/500, SantaCruz), or PLN (1/5000, AffinityBioReagent Inc., St. Quentin en Yvelines, France). Signals were visualized by using horseradish peroxidase-conjugated secondary antibody (Amersham) and ECL Plus reagents (Amersham) and acquired in a Las3000 DarkBox (Fuji Photo Film Europe GMBH, Düsseldorf, Germany). ß-Actin was used as internal control for protein loading.

Dexamethasone binding was analyzed on cytosol extracts from the pooled ventricles of five control and five DT adrenalectomized mice. Adrenalectomy is required for such experiments to allow full access of 3H-Dex to the GR. Mice underwent adrenalectomy 4 days before heart sampling and were supplied with 0.9% NaCl in drinking water. The complete removal of the adrenal glands was confirmed by plasma corticosterone measurement, using a radioimmunoassay (MPBiomedicals, Illkirch, France). Cytosol extracts were used for 3H-Dex (Amersham) binding experiments as described previously (19) .

Cellular electrophysiology and [Ca2+]i analyses were conducted at room temperature (20~24°C) on 2-month-old female matched littermates (6 DT and 7 WT mice). Ventricular myocytes were isolated as described previously (20) . Action potentials (AP) and membrane currents were recorded at a frequency of 0.1 Hz, using the whole-cell configuration of the patch-clamp technique, as described previously (14 , 21) . Pipette tips had resistances of 1–2 M{Omega}. After membrane rupture, cell capacitance (Cm) was measured to estimate cell size before 30–60% series resistance compensation. APs were elicited in standard external (mM: NaCl 140, MgCl2 1.1, CaCl2 1.8, KCl 4, glucose 10, and HEPES 10, with the pH adjusted to 7.4 with LiOH) and internal (mM: KCl 135, MgCl2 4, EGTA 5, glucose 10, HEPES 10, Na2ATP 5, and Na2 creatine phosphate 3, with the pH adjusted to 7.2 with LiOH) Tyrode’s solutions by 1.5-fold excitation threshold current pulses of 2.5 ms in duration. For whole-cell outward K+ currents, internal solution was the same as for AP recordings while external solution was modified by equimolar replacement of NaCl with 138 mM cholineCl, 1 mM CoCl2 to block Na+ and Ca2+ currents and with 1 mM BaCl2 to block other K+ currents. The total outward K+ currents were evoked during 300 ms depolarizing voltage steps to potentials between –70 and +60 mV (10 mV increments) from a holding potential of –80 mV. Ito were obtained by subtracting corresponding current records with and without a 100 ms inactivating prepulse to +30 mV. The remaining currents after Ito inactivation are then recorded in the absence and presence of 250 µM 4-aminopyridine (4-AP). The 4-AP-resistant outward K+ current is defined as Iss, IKslow being obtained by subtracting corresponding current records (with the inactivating prepulse) with and without 4-AP. ICa was recorded with standard Tyrode’s solution with KCl replaced by CsCl in order to inactivate K+ current. Ca2+ currents were evoked by 300 ms depolarizing steps to voltages ranging between –50 and +60 mV in +10 mV increments, using a holding potential of –80 mV. A 500 ms voltage ramp to –40 mV was applied before each pulse to inactivate INa. INa were recorded in 0K+ solutions with 20 mM NaCl by equimolar substitution with 118 mM cholineCl, 1 mM CoCl2, and 1 mM BaCl2 in the external solution. INa were elicited by 30 ms steps from –100 mV to the –80 to –5 mV range in 5 mV increments. Currents amplitudes, measured as the difference between the peak current and the steady-state current at the end of the voltage steps, were normalized to the membrane capacitance (Cm). For ISS, current amplitude corresponds to the current at the end of pulse.

[Ca2+]i transients were investigated on isolated cardiomyocytes loaded with Ca2+-fluorescent probe Fluo3-AM (Molecular Probes, Carlsbad, CA, USA), using high-resolution confocal imaging (Zeiss LSM510), as described previously (14 , 21) . Fluorescence images were analyzed with IDL (Research Systems, Boulder, CO, USA). Ca2+ levels are expressed as F/F0, where F is fluorescence and F0 is diastolic fluorescence.

Electrocardiograms (ECGs) were recorded over a 10 min period in 2-month-old male and female mice anesthetized with isoflurane. Twenty-four hour telemetry recordings were obtained in freely moving mice following abdominal implantation of a DSI transmitter (TA10EA-F20, Data Science International, St. Paul, MN, USA). ECGs were analyzed with ECG-Auto software (ECG-Auto 1.12.36, EMKA France, Bourre, France). ECG intervals were measured from short recordings on average beats constructed from 200 consecutive QRST complexes. Intervals were determined semiautomatically from a library of QRST waveforms sampled from the tracing. Every average beat was manually checked and intervals were set using the first derivative of the tracing obtained from ECG-Auto software. As shown on typical examples in Fig. S1, PQ was measured from the onset of P wave to the onset of the QRS wave. QRS duration was measured from the onset of the Q wave to the end of high-amplitude electrical events, as detected on the first derivative. The duration of the QT interval was measured from the onset of the Q wave to the last detectable electrical event on the first derivative. QT interval was corrected for heart rate by drawing the regression line from individual beats for each mouse and was expressed as the value for RR at 150 ms. Arrhythmia was detected by analysis of the tachogram of the recording (10 min or 24 h). Tachograms were constructed from automatic R wave detection (ECG-Auto EMKA) and RR plotted against time. All abnormal RR intervals were manually checked for validation and labeling. AVB2 was defined as one nonconducting P-wave or Wenckebach period; advanced second degree and complete AVB were defined as AVB3. AVB3 duration over 24 h was calculated from 24 h telemetry recordings as the sum of the durations of the AVB3 events occurring over the 24 h period.

Details on the methods used for blood pressure, echocardiography and histological examination are given in the online supplementary methods.

Data are expressed as mean ±SE. Student’s t test was used to compare unpaired data between two groups. ANOVA was used for comparison when more than two groups were analyzed. If the global test was significant, pairwise comparisons were performed with a Tuckey-Kramer test. P < 0.05 was considered significant.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Conditional expression of hGR in the heart of transgenic mice
The tetracycline system (22) was used for conditional cardiac hGR expression (Fig. 1 A). hGR was overexpressed solely in cardiomyocytes by crossing the tetO-hGR strain we generated with the previously described transactivator {alpha}MHC-tTA strain (18) . Double-transgenic {alpha}MHC-tTA/tetO-hGR (referred to as DT) were born in the expected Mendelian ratio. Conditional cardiac GR overexpression was achieved using the tetO-hGR strain (i.e., cardiac-specific and Dox-dependent expression of hGR in DT mice only; see Fig. 1B, C ). Cardiac mouse GR expression was not modified (Fig. 1B, C ). Scatchard plot analyses of Dex binding (Fig. 1D ) indicated that the maximum number of glucocorticoid binding sites in DT mice (396±80 femtomol/mg protein) was 3-fold that in controls (122±11 femtomol/mg protein) with a similar affinity (Kd: DT: 3.4±0.2 nM; control: 5±0.3 nM). Plasma corticosterone (in ng/ml, control females: 194±19, DT females: 174±20, control males: 67±11, DT males: 67±9, n=4–6 in each group) and blood pressure (in mmHg, control: 136±3, DT: 126±6, n=5) did not differ between control and DT mice. Thus this conditional GR overexpression model can be used to study the primary role of cardiac GR activation in the absence of confounding systemic effects that might have secondary effects on heart function.


Figure 1
View larger version (13K):
[in this window]
[in a new window]

  		Figure 1. Conditional cardiac hGR overexpression. A) The tetracycline-inducible system (tet system) is a bigenic system allowing inducible expression of the gene of interest. To obtain cardiac-specific, doxycycline-controlled transgene expression, the {alpha}MHC-tTA transgenic mouse line (transactivator strain, {alpha}MHC-tTA) expressing the tTA transactivator under the control of the cardiac-specific {alpha}MHC promoter is mated with hGR transgenic mice (acceptor strain, tetO-hGR). This leads to wild-type and monotransgenic {alpha}MHC-tTA and tetO-hGR mice, which do not express hGR, and to the double-transgenic (DT) conditional and cardiac-specific model in which hGR expression is Dox dependent. B) hGR expression was demonstrated by Western blot using a specific antibody for human GR. Human kidney cell (HEK 293) extracts were used as a positive control. Only DT mice express hGR, and not monotransgenic tetO-hGR littermates, indicating that control of the transgene expression is not leaky. Expression of endogenous mGR was not modified. Actin was used as internal control for protein loading. C) hGR protein levels estimated in doxycycline-treated (+Dox) or untreated (–Dox) control and DT mice. The expression of hGR in DT mice was prevented by Dox treatment, indicating that the tetO-hGR mouse model was under Dox control. Actin was used as internal control for protein loading. D) Scatchard plot analysis of dexamethasone binding in the heart cytosol of adrenalectomized control (ctrl) and DT mice. Conditional hGR expression results in a 3-fold increase in Dex binding.

Cardiac overexpression of hGR leads to changes in ionic currents and [Ca2+]i homeostasis in isolated ventricular cardiomyocytes
Action potential (AP) measurement takes into account changes in the electrical properties of myocytes. No difference in myocyte size, evaluated by cell capacitance, was observed at 2 months of age (Cm in pF, controls: 149.3±5.5 (n=65); DT: 146.6±5.8 (n=64)), indicating the absence of myocyte hypertrophy. Ventricular cardiomyocytes from DT mice displayed slowing of both the depolarization and repolarization phases of AP (Fig. 2 A). Analysis of AP duration (APD) showed a significant lengthening (Fig. 2B , upper panel) from 20% of repolarization (1.3±0.2 vs.7.8±2.8 ms at 20%; 4.6±0.7 vs. 22.6±6.6 ms at 50%; and 51.9±5.8 vs. 87.9±11.5 ms at 90%; 31 WT vs. 27 DT cells, respectively) whereas neither the resting membrane potential (in mV, –80.4±0.3 and –81.4±0.5, in control and DT cells, respectively) nor the maximum AP amplitude (in mV, 119.0±1.1 and 116.3±1.5, in control and DT cells, respectively) were different. Moreover, Dox administration during pregnancy and to the offspring prevented APD increase in DTs (Fig. 2B, DT +Dox). At 20%, 50%, and 90% repolarization, mean APD in DT + Dox were 1.3 ± 0.1, 3.8 ± 0.2, and 46.8 ± 4.0 ms, respectively; n = 19 and no more different from controls. Maximum depolarization velocity (dV/dtmax) was significantly lower (19% lower) in DT mice than in their control littermates (Fig. 2B , lower panel). Dox treatment prevented this alteration (Fig. 2B, DT +Dox). Since Na+ current (INa) drives the AP upstroke phase, we next examined INa in WT and DT mice. INa amplitudes were substantially lower in DT than in WT myocytes (Fig. 2C , insets). Normalized to Cm, the decrease in INa densities was statistically significant between WT and DT mice at all potentials (Fig. 2C ). This 25% decrease parallels to the 20% decrease in dV/dtmax and was not associated with modification either in kinetic properties or voltage-dependent inactivation of the currents (Fig. S2).


Figure 2
View larger version (27K):
[in this window]
[in a new window]

  		Figure 2. Specific cardiac hGR overexpression remodels ventricular action potential (AP) and ionic currents, as measured by patch-clamp techniques. A) Representative APs recorded from control (dark line) and DT (gray line) cardiomyocytes. B) Bar graphs of pooled AP duration at 50% of repolarization (upper panel) and maximum upstroke velocity (dV/dtmax) (lower panel) (control, open bar; DT, gray bar; DT+Dox, hatched bar); n = number of cardiomyocytes. Cardiac hGR overexpression slowed both depolarization and repolarization phases of AP. These effects are prevented by Dox administration (DT+Dox) (i.e., prevention of hGR expression). C–F) Current density-voltage relationships of Na+ current (INa, C), transient outward K+ currents (Ito, D), slowing inactivating K+ current (IKslow, E), and L-type Ca2+ current (ICa, F) with respect to control (open symbols) and DT ventricle myocytes (filled symbols). Current density was obtained by normalizing the current amplitude by cell capacitance. Insets: representative current recordings from control (black lines) and DT (gray lines) cardiomyocytes. Data are means + SE, n = 13 to 31 cardiomyocytes from 6 to 7 mice, *P < 0.05, **P < 0.005 and ***P < 0.0005, vs. control.

We next investigated the outward and inward currents that may contribute to the increase in APD. In adult mouse ventricle, the outward K+ currents involved in repolarization consist of three components, differing in terms of specific time and voltage dependence and sensitivity to pharmacological agents: 1) a transient current (Ito); 2) a low-concentration 4-aminopyridine-sensitive current (IKslow, previously referred to as IKur); and 3) a non-inactivating steady-state component (Iss). Iss density did not differ significantly between control and DT myocytes (at +30 mV, pA/pF, 6.4±0.3 vs. 5.8±0.3, DT (n=17) vs. control (n=13), respectively). In contrast, the magnitudes of Ito (Fig. 2D ) and IKslow (Fig. 2E ) were lower in DT than in control myocytes. Neither the time-dependent nor the voltage-dependent properties of the currents varied significantly (data not shown).

L-type Ca2+ current (ICa) recorded in myocytes from DT were greater than those recorded from control littermates (Fig. 2F ). Current densities were significantly higher (in the –20 to +40 mV voltage range) (Fig. 2F ) in DT than in controls. Neither Ca2+ channel availability as a function of voltage nor activation kinetics differed significantly between DT and control myocytes (data not shown).

It is generally considered that electrophysiological analysis is more sensitive than biochemical ones and that small but highly significant variations in current densities are not always correlated with protein and RNA levels. No clear correlation between the above-mentioned currents and transcript levels of some ion channel subunits (Cacna1C, Scn5a, Kcna5, Kcnb1, and Kcnd2) were noticed (Fig. S3), suggesting that additional regulation steps, such as protein trafficking and/or isoform switch, might also be altered in this model.

The ICa traces (Fig. 2F , inset) suggest that the time course of ICa inactivation was modified in DT myocytes. Decay kinetics of elicited ICa indicated a significant increase in the fast component of ICa inactivation, with no change in the slow component. At 0 mV, the fast component decreased from 8.0 ± 0.3 to 6.8 ± 0.3 ms, P < 0.01; whereas the slow component was similar (71.0±2.1 vs. 74.3±5.5 ms; in 29 control and 19 DT myocytes, respectively). The faster inactivation of ICa presumably reflects the Ca2+-dependent inactivation of ICa and suggests changes in sarcoplasmic reticulum (SR) Ca2+ handling. Figure 3 A shows confocal line-scan images of the steady-state [Ca2+]i transients. In DT myocytes, the peak amplitude was greater (+53%; Fig. 3A, B ), the fluorescence signal was faster, and cell shortening increased (+91%, Fig. 3A, C ) compared with control myocytes. The [Ca2+]i transient declined much more rapidly in DT myocytes, with a 32% lower time constant of [Ca2+]i (monoexponential fit to rate of decay of [Ca2+]i transient) (Fig. 3D ) indicative of a faster rate of Ca2+ uptake into the SR, possibly resulting in increased SR Ca2+ load. The SR Ca2+ content, as estimated after caffeine application, was 50% higher in DT than in control myocytes (Fig. 3E, F ). This may be due to changes in the function of several SR proteins, including SR calcium ATPase (SERCA2a) or its regulator phospholamban (PLN). SERCA2a protein levels were unaltered in DT mice (Fig. 3G ), whereas PLN content was about half that in control mice (Fig. 3G ), leading to a significant decreased PLN/SERCA ratio (Fig. 3H ), which may result in enhanced SERCA activity.


Figure 3
View larger version (34K):
[in this window]
[in a new window]

  		Figure 3. [Ca2+]i signaling in hGR-overexpressing ventricle myocytes. A) Representative line-scan images of fluo-3 fluorescence, recorded in one control (Ctrl) (top) and one DT (bottom) myocyte field stimulated at 1 Hz. B–D) Bar graphs corresponding to B) the amplitude [Ca2+]i transient (measured as peak F/F0); C) cell shortening; D) [Ca2+]i transient rate of decay (monoexponential fit) in control (open bars) and DT mice (filled bars). Positive inotropism was observed in myocytes from hGR-overexpressing mice. E, F) Representative example (E) and pooled amplitude histogram (F) of [Ca2+]i transients induced by 10 mM caffeine in control (open bar) and DT (filled bar) mice. SR Ca2+ load is higher in DT than in control mice. G, H) Western blot analyses of SERCA2a and PLN levels in the hearts of control and DT mice. G) Representative Western blots of SERCA2a and PLN levels in 2-month-old control and DT mice, indicating similar SERCA2a levels and lower PLN levels in the DT mice. Actin was used as internal control for protein loading. H) Relative quantification, indicating a lower PLN/SERCA2a ratio in DT (filled bars) mice than in controls (open bar). Data are means ± SE, n = number of myocytes (B–F) or mice (H) studied; *P < 0.05, **P < 0.005, ***P < 0.0005.

hGR transgenic mice display atrio-ventricular block
The most striking cardiac feature of mice with cardiac hGR overexpression was bradycardia and high-degree AVB (Fig. 4 A, B). When measured over 24 h in freely moving mice, heart rate (HR) was lower in DT than in control mice (beats/min, DT: 553±5.6; control: 619±18.9, n=4 in each group, P < 0.05), with a greater mean RR interval (Fig. 4C ). AVB was observed during rest and activity in DT mice, whereas control mice presented AVB only occasionally (Fig. 4B ). 60% of the DT mice showed AVBs vs. 2% in monotransgenic littermates. AVB events over the 24 h period were more frequent in DT mice than in monotransgenic littermates (Fig. 4D ).


Figure 4
View larger version (26K):
[in this window]
[in a new window]

  		Figure 4. Cardiac hGR overexpression induced AV block. A) Representative ECG tracing in a DT mouse showing AVB3 episode. B) Representative 24 h tachogram (RR vs. time) for control (Ctrl) and DT mice. 24 h recordings were started at 7 AM. AVB are depicted as long RR Intervals. AVB events occurred all along the 24 h period in DT mice, whereas AVB occurred occasionally at rest in control mice. C) The mean 24 h RR interval was greater in DT mice (filled bars) than in controls (open bars), indicative of lower HR over the 24 h period. *P < 0.05, DT vs. controls. D) The number of AVB3 events over the 24 h period was greater in DT (filled bars) than in control (Ctrl, open bars) mice. P < 0.05, DT vs. controls.

In 2-month-old DT mice, AVB was associated with a long PQ interval and increased QRS duration and QTc interval (Table 1 ). The increase in intervals observed in DT mice was dependent on hGR expression since Dox administration (from embryonic development and thereafter) prevented PQ, QRS, and QTc enlargement (Table 1) . No ventricular arrhythmia was detected by short-term or 24 h telemetry electrocardiogram (ECG) recordings. To test whether the phenotype could be reversed in adult DTs, hGR expression was turned down by Dox administration to 3-month-old mice. The enlarged PQ observed in DT mice progressively decreased over 4 wk (Fig. 5 A). Even more impressive was the complete disappearance of AVB events within 1 month in Dox-treated DT mice (Fig. 5B ).


View this table:
[in this window]
[in a new window]

  		Table 1. Short-term ECG recordings in anesthetized micea


Figure 5
View larger version (17K):
[in this window]
[in a new window]

  		Figure 5. Conduction defects induced by conditional GR expression in cardiomyocytes are reversible. A) The enlargement of PQ interval in DT (filled bars) compared with control (open bars) was largely reversible on GR overexpression shutoff with Dox administration in adult DT mice. B) AVB events progressively disappeared in DT (filled bars) on GR overexpression shutoff with Dox administration in adult DT mice. *P < 0.05, DT vs. controls; $P < 0.05, DT + 30 days Dox vs. DT without Dox.

ECG alterations were not associated with major cardiac changes, as assessed with echocardiography. A 6% decrease in ejection fraction (EF), which may be due to decreased HR, was noticed in the absence of heart remodeling (Table 2 ). In particular, the evaluation of myocardial contractility by tissue Doppler, such as Sa (maximal systolic velocity of mitral annulus), showed no difference between control and DT mice. No hypertrophy was observed by echocardiography (LV mass/BW) or by gravimetry (HW/BW ratio (mg/g), controls: 5.2±0.2; DT: 5.4±0.3; n=19). Histological examination showed no cardiac remodeling in DT mice (see supplementary results).


View this table:
[in this window]
[in a new window]

  		Table 2. Echocardiographic studya

Therefore, ion channel alterations and conduction defects were not associated with structural heart disease, and might instead be related primarily to the functional consequences of cardiac hGR expression.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
To elucidate the role of the GR in the heart, we have generated a conditional mouse model that allows selective expression of GR in cardiomyocytes only. This mouse model can be used to evaluate the integrated pathophysiological consequences of increasing GR activation in the heart in the absence of the confounding secondary events that occur with pharmacological approaches. Unexpectedly, we found that cardiac GR overexpression reduced heart rate and depressed cardiac conduction with AVB. One advantage of this conditional model is to manipulate transgene expression over time. Inhibition of transgene expression by Dox administration prevents the appearance of the electrical abnormalities; this indicates that our findings are not dependent on transgene insertion into the genome or on cardiac tTA expression. When GR overexpression was switched off in 3-month-old animals, the altered conduction phenotype was reverted, suggesting a direct involvement of GR in the adult heart rather than a consequence of long-term expression or impaired development. Thus, enhanced GR activity may play an important role in regulating heart conduction and sinus node function. Noticeably, this occurs in the absence of inflammatory or remodeling/degenerative processes, as assessed with histopathological analysis. This situation clearly differs from the well-known beneficial effects of GCs on the AVB observed in some diseases, most probably resulting from their anti-inflammatory action.

Cardiac overexpression of hGR reproduces most of the known cellular effects of glucocorticoids on ventricular cardiomyocytes Adrenal glucocorticoids and synthetic Dex can promote prolonged cardiac AP repolarization and ionic current alterations (8 , 10) . Dex has been shown to increase ventricular ICa density (8 , 10) . Moreover, the up-regulation of dihydropyridine receptors (corresponding to L-type Ca2+ channel expression) has been reported after in vivo GC treatment (23) . Ito density decreases in mice treated with Dex (10) . In contrast to the decrease in IKslow density reported in our study, IKslow was unaffected in neonatal mice treated with Dex for 5 days (10) . Kcna5 mRNA up-regulation in the heart has been observed after in vivo Dex administration (24) . These differences in IKslow and Kcna5 regulation may reflect acute GR activation, a situation clearly different from that in our model in which the number of GC binding sites increases moderately, but chronically, in animals with normal plasma ligand concentrations. Some studies have suggested that, in addition to their electrophysiological effects, GCs may increase cardiac contractile function (6 , 7) . The mechanisms underlying this positive inotropic effect may involve an increase in intracellular [Ca2+]i transients (8) and ATP-dependent Ca2+ accumulation in the SR (9 , 25) . Our study also shows enhanced cell contractility and an increase in Ca2+ transient amplitudes, with faster decay kinetics in hGR-overexpressing cardiomyocytes attributed to increased Ca2+ influx during prolonged AP and SR Ca2+ content. This increase in SR Ca2+ content may result from a GC-mediated increase in SERCA activity, as previously shown in cardiomyocytes isolated from Dex-treated rats (9 , 25) . PLN expression decreases SR Ca2+-pump activity (26) . Indeed, the mice in our model displayed no change in SERCA2a expression but produced smaller amounts of total PLN, resulting in a decrease in PLN/SERCA2a ratio. This may lead to increased SERCA activity, resulting in enhanced SR Ca2+ load.

Electrophysiology measurements in ventricular cardiomyocytes isolated from these mice provided insight into cellular alterations underlying the observed phenotype. Caution should be taken when extrapolating to electophysiological remodeling occurring in the conductive tissue in the absence of direct measurements; the MHC promoter has also been shown to be active in the conduction tissue (27) . Down-regulation of Na+ and K+ currents may play an important role in conduction and sinus node defects. The weaker AP upstroke observed in our mouse model is associated with a decrease in INa, which has been recognized as a determinant of sinus rhythm and conduction (28 , 29) . Increases in the occurrence of AVB and prolonged PQ have been associated with Ito down-regulation in a mouse model (30) . Such increases in AVB and PQ duration were not observed in a mouse Ikslow down-regulation model obtained by overexpression of a dominant negative Kcnd2 subunit driven by the same MHC promoter (31) . However, the combined down-regulation of Ito and Ikslow favors PQ lengthening, as previously reported in another transgenic mouse model with combined Ito and Ikslow abolition generated by the cardiac-specific expression of dominant-negative mutants of Kcna5 and Kcnd2 (31) . Of note, cardiac hGR overexpression did not modify connexin 43 expression (data not shown), which has been involved in cardiac conduction (32) . The AV conduction defects reported here may therefore result from a combined decreased in INa and Ito/Ikslow currents.

Increasing attention has been devoted to analysis of the pathophysiological role of corticosteroid hormones in the cardiovascular field. Cardiomyocytes express the GR and the MR (12) , and many studies have reported deleterious effects of aldosterone in heart failure (13) . Recent clinical trials based on pharmacological MR antagonism in heart failure have highlighted the beneficial effects of MR antagonisms for reducing morbidity and mortality (33 , 34) . GCs were identified as a risk factor for heart failure in a large cohort of treated patients (5) and were associated with decreased heart rate in healthy volunteers (11) . The relative levels of MR and GR expression may be altered in cardiovascular diseases (35 , 36) . However, the respective specificities of the MR and GR signaling pathways are difficult to analyze, as both receptors can bind either aldosterone or GC, depending on plasma concentrations (12) .

Comparison of the phenotypes of cardiac GR- and MR-overexpressing mice should increase our understanding of the contribution of each receptor to cardiac pathophysiology. We recently investigated the consequences of MR overexpression in the heart using an experimental strategy similar to that described here (14) . At variance with the cardiac GR overexpression model, cardiac MR overexpression led to embryonic and perinatal death and to the occurrence of severe ventricular arrhythmia, without high-degree AVB (14) . Despite the overlapping regulation of some ion currents in the MR and GR conditional models, currents specific to each overexpression model may account for the striking differences in ECG phenotype (Table 2) . Cardiac MR overexpression has a major impact on Ca2+ current (14) , whereas the main targets in the GR model are Na+ and K+ currents. INa can be estimated based on the AP upstroke phase and maximum depolarization velocity (dV/dtmax). Indeed, re-examination of our previous model (14) indicated that dV/dtmax was not affected by cardiac-specific hMR overexpression, in contrast to what was observed for hGR overexpression. Ito was similarly affected in both mouse models, but additional IKslow down-regulation in the GR model may protect against arrhythmia (31) . Another major difference between MR- and GR-overexpressing cardiomyocytes is the degree of SR Ca2+ load: a high SR Ca2+ load was evidenced in cardiomyocytes from GR-overexpressing mice, but not in MR-overexpressing mice (Table 3 ). Thus, althoughclosely related, the MR and GR appear to have different specific cardiac effects.


View this table:
[in this window]
[in a new window]

  		Table 3. Comparison of the phenotype induced by cardiac-specific hMR (14) or hGR overexpression in conditional mouse models


   ACKNOWLEDGMENTS
 
We thank M. Muffat-Joly for assistance with blood pressure analysis, M-E. Rafestin-Oblin for advice concerning binding experiments, and the CEFI (Centre d’Explorations Fonctionnnelles Intégrées), IFR 02, Paris, for help in phenotypic analyses. We also thank A. Ouvrard-Pascaud for his help in the early steps of this work and Ana Maria Gomez for critical reading of the manuscript. Y.S-M. was supported by the CIFRE program of ANVAR (Lyon, France). A.N.D.C. and R.P. held Ph.D. fellowships from the Region Ile de France and the Ministère de la Recherche, respectively. This work was supported by the AVENIR INSERM grant, the ANR (ANR05-PCOD005) and ACI programs of the Ministère de la Recherche, and grants from the Genzyme Research Innovation Program, the Fondation de France, and the CRIT (Centre de Recherche Industrielle et Technique).


   FOOTNOTES
 
1 These authors contributed equally to this work.

Received for publication February 13, 2007. Accepted for publication April 12, 2007.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Girod, J. P., Brotman, D. J. (2004) Does altered glucocorticoid homeostasis increase cardiovascular risk?. Cardiovasc. Res. 64,217-226[Abstract/Free Full Text]
  2. Israel, B. A., Sherman, F. S., Guthrie, R. D. (1993) Hypertrophic cardiomyopathy associated with dexamethasone therapy for chronic lung disease in preterm infants. Am. J. Perinatol. 10,307-310[Medline]
  3. Evans, N. (1994) Cardiovascular effects of dexamethasone in the preterm infant. Arch. Dis. Child. Fetal. Neonatal. Ed. 70,F25-F30[Abstract]
  4. Wei, L., MacDonald, T. M., Walker, B. R. (2004) Taking glucocorticoids by prescription is associated with subsequent cardiovascular disease. Ann. Intern. Med. 141,764-770[Abstract/Free Full Text]
  5. Souverein, P. C., Berard, A., Van Staa, T. P., Cooper, C., Egberts, A. C., Leufkens, H. G., Walker, B. R. (2004) Use of oral glucocorticoids and risk of cardiovascular and cerebrovascular disease in a population-based case control study. Heart 90,859-865[Abstract/Free Full Text]
  6. Penefsky, Z. J., Kahn, M. (1971) Inotropic effects of dexamethasone in mammalian heart muscle. Eur. J. Pharmacol. 15,259-266[CrossRef][Medline]
  7. Lefer, A. M. (1968) Influence of corticosteroids on mechanical performance of isolated rat papillary muscles. Am. J. Physiol. 214,518-524[Free Full Text]
  8. Whitehurst, R. M., Jr, Zhang, M., Bhattacharjee, A., Li, M. (1999) Dexamethasone-induced hypertrophy in rat neonatal cardiac myocytes involves an elevated L-type Ca2+current. J. Mol. Cell. Cardiol. 31,1551-1558[CrossRef][Medline]
  9. Narayanan, N., Yang, C., Xu, A. (2004) Dexamethasone treatment improves sarcoplasmic reticulum function and contractile performance in aged myocardium. Mol. Cell. Biochem. 266,31-36[CrossRef][Medline]
  10. Wang, L., Feng, Z. P., Duff, H. J. (1999) Glucocorticoid regulation of cardiac K+ currents and L-type Ca2+ current in neonatal mice. Circ. Res. 85,168-173[Abstract/Free Full Text]
  11. Brotman, D. J., Girod, J. P., Garcia, M. J., Patel, J. V., Gupta, M., Posch, A., Saunders, S., Lip, G. Y., Worley, S., Reddy, S. (2005) Effects of short-term glucocorticoids on cardiovascular biomarkers. J. Clin. Endocrinol. Metab. 90,3202-3208[Abstract/Free Full Text]
  12. Sheppard, K. E. (2003) Corticosteroid receptors, 11 beta-hydroxysteroid dehydrogenase, and the heart. Vitam. Horm. 66,77-112[Medline]
  13. Struthers, A. D. (2002) Aldosterone: cardiovascular assault. Am. Heart. J. 144,S2-S7[CrossRef][Medline]
  14. Ouvrard-Pascaud, A., Sainte-Marie, Y., Benitah, J. P., Perrier, R., Soukaseum, C., Cat, A. N., Royer, A., Le Quang, K., Charpentier, F., Demolombe, S., et al (2005) Conditional mineralocorticoid receptor expression in the heart leads to life-threatening arrhythmias. Circulation 111,3025-3033[Abstract/Free Full Text]
  15. Funder, J. W. (2004) Is aldosterone bad for the heart?. Trends Endocrinol. Metab. 15,139-142[CrossRef][Medline]
  16. Ouvrard-Pascaud, A., Jaisser, F. (2003) Pathophysiological role of the mineralocorticoid receptor in heart: analysis of conditional transgenic models. Plügers Arch. 445,477-481[Medline]
  17. Baron, U., Freundlieb, S., Gossen, M., Bujard, H. (1995) Co-regulation of two gene activities by tetracycline via a bidirectional promoter. Nucleic Acids Res. 23,3605-3606[Free Full Text]
  18. Yu, Z., Redfern, C. S., Fishman, G. I. (1996) Conditional transgene expression in the heart. Circ. Res. 79,691-697[Abstract/Free Full Text]
  19. Couette, B., Fagart, J., Jalaguier, S., Lombes, M., Souque, A., Rafestin-Oblin, M. E. (1996) Ligand-induced conformational change in the human mineralocorticoid receptor occurs within its hetero-oligomeric structure. Biochem. J. 315,421-427[Medline]
  20. Benitah, J. P., Perrier, E., Gomez, A. M., Vassort, G. (2001) Effects of aldosterone on transient outward K+ current density in rat ventricular myocytes. J. Physiol. 537,151-160[Abstract/Free Full Text]
  21. Perrier, R., Richard, S., Sainte-Marie, Y., Rossier, B. C., Jaisser, F., Hummler, E., Benitah, J. P. (2005) A direct relationship between plasma aldosterone and cardiac L-type Ca2+ current in mice. J. Physiol. 569,153-162[Abstract/Free Full Text]
  22. Bujard, H. (1999) Controlling genes with tetracyclines. J. Gene Med. 1,372-374[CrossRef][Medline]
  23. Takimoto, K., Li, D., Nerbonne, J. M., Levitan, E. S. (1997) Distribution, splicing and glucocorticoid-induced expression of cardiac alpha 1C and alpha 1D voltage-gated Ca2+ channel mRNAs. J. Mol. Cell. Cardiol. 29,3035-3042[CrossRef][Medline]
  24. Takimoto, K., Levitan, E. S. (1994) Glucocorticoid induction of Kv1.5 K+ channel gene expression in ventricle of rat heart. Circ. Res. 75,1006-1013[Abstract]
  25. Rao, M. K., Xu, A., Narayanan, N. (2001) Glucocorticoid modulation of protein phosphorylation and sarcoplasmic reticulum function in rat myocardium. Am. J. Physiol. 281,H325-H333
  26. Luo, W., Wolska, B. M., Grupp, I. L., Harrer, J. M., Haghighi, K., Ferguson, D. G., Slack, J. P., Grupp, G., Doetschman, T., Solaro, R. J., Kranias, E. G. (1996) Phospholamban gene dosage effects in the mammalian heart. Circ. Res. 78,839-847[Abstract/Free Full Text]
  27. Dobrzynski, H., Billeter, R., Greener, I. D., Tellez, J. O., Chandler, N. J., Flagg, T. P., Nichols, C. G., Lopatin, A. N., Boyett, M. R. (2006) Expression of Kir2.1 and Kir6.2 transgenes under the control of the alpha-MHC promoter in the sinoatrial and atrioventricular nodes in transgenic mice. J. Mol. Cell. Cardiol. 41,855-867[CrossRef][Medline]
  28. Papadatos, G. A., Wallerstein, P. M., Head, C. E., Ratcliff, R., Brady, P. A., Benndorf, K., Saumarez, R. C., Trezise, A. E., Huang, C. L., Vandenberg, J. I., et al (2002) Slowed conduction and ventricular tachycardia after targeted disruption of the cardiac sodium channel gene Scn5a. Proc. Natl. Acad. Sci. U. S. A. 99,6210-6215[Abstract/Free Full Text]
  29. Guo, J., Zhan, S., Somers, J., Westenbroek, R. E., Catterall, W. A., Roach, D. E., Sheldon, R. S., Lees-Miller, J. P., Li, P., Shimoni, Y., Duff, H. J. (2006) Decrease in density of INa is in the common final pathway to heart block in murine hearts overexpressing calcineurin. Am. J. Physiol. 291,H2669-H2679
  30. Guo, W., Li, H., London, B., Nerbonne, J. M. (2000) Functional consequences of elimination of i(to,f) and i(to,s): early afterdepolarizations, atrioventricular block, and ventricular arrhythmias in mice lacking Kv1.4 and expressing a dominant-negative Kv4 alpha subunit. Circ. Res. 87,73-79[Abstract/Free Full Text]
  31. Brunner, M., Guo, W., Mitchell, G. F., Buckett, P. D., Nerbonne, J. M., Koren, G. (2001) Characterization of mice with a combined suppression of I(to) and I(K,slow). Am. J. Physiol. 281,H1201-H1209
  32. Gutstein, D. E., Morley, G. E., Fishman, G. I. (2001) Conditional gene targeting of connexin43: exploring the consequences of gap junction remodeling in the heart. Cell. Commun. Adhes. 8,345-348[Medline]
  33. Pitt, B., Remme, W., Zannad, F., Neaton, J., Martinez, F., Roniker, B., Bittman, R., Hurley, S., Kleiman, J., Gatlin, M. (2003) Eplerenone, a selective aldosterone blocker, in patients with left ventricular dysfunction after myocardial infarction. N. Engl. J. Med. 348,1309-1321[Abstract/Free Full Text]
  34. Pitt, B., Zannad, F., Remme, W. J., Cody, R., Castaigne, A., Perez, A., Palensky, J., Wittes, J. (1999) The effect of spironolactone on morbidity and mortality in patients with severe heart failure. Randomized Aldactone Evaluation Study Investigators. N. Engl. J. Med. 341,709-717[Abstract/Free Full Text]
  35. DeLano, F. A., Schmid-Schonbein, G. W. (2004) Enhancement of glucocorticoid and mineralocorticoid receptor density in the microcirculation of the spontaneously hypertensive rat. Microcirculation 11,69-78[CrossRef][Medline]
  36. Yoshida, M., Ma, J., Tomita, T., Morikawa, N., Tanaka, N., Masamura, K., Kawai, Y., Miyamori, I. (2005) Mineralocorticoid receptor is overexpressed in cardiomyocytes of patients with congestive heart failure. Congest. Heart Fail. 11,12-16[CrossRef][Medline]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
fj.07-8357comv1
21/12/3133    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sainte-Marie, Y.
Right arrow Articles by Jaisser, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sainte-Marie, Y.
Right arrow Articles by Jaisser, F.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS