INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

LONG-TERM VENTRICULAR PACING at a rapid rate (240/min for 4-6 wk) induces in canines and other mammals low-output biventricular failure associated with hemodynamic and neurohumoral manifestations that are thought to be similar to those seen in humans (4). Among the many factors putatively involved in impairment of ventricular function in this model, alterations have been found to occur at the level of the contractile protein content (32), excitation-contraction (E-C) coupling (1, 27, 28, 36), efferent cardiac sympathetic nerves (12), and myocardial -adrenergic receptor (-AR) signaling (11, 18, 20, 23).

Variables related to both contraction and -AR signaling have been measured in several studies considering this model (3, 11, 18, 20, 23, 29). In papillary muscles excised from healthy or failing canine hearts, Juneau et al. (18) found that, although basal contractile activity was depressed, isoproterenol caused a similar increase in tension generation (isometric contraction) and an actually greater increase in shortening (isotonic contraction) in the failing myocardial preparations, despite the fact that -AR density and cAMP generation measured in crude membranes derived from whole tissue homogenates were reduced. Vatner et al. (36) provided another example of possible dissociation between responses to agonists and alterations in -AR signaling by reporting that, after only 1 day of rapid pacing, peak left ventricular +dP/dt increments in response to isoproterenol were depressed by 50%, a marked functional deficit that they deemed to be out of proportion to the modest alteration in -AR signaling mechanisms (20). In contrast, Marzo et al. (23) found that, in the in situ situation, there was a depression of the dose-response curve for agonist-induced increases in peak left ventricular +dP/dt, along with reductions in -AR density and cAMP generation measured in membranes from whole tissue homogenates. In all previous studies investigating responses to nonselective -AR stimulation, -AR density and -AR-mediated cAMP generation measurements were made only in membranes derived from whole tissue homogenates.

Therefore, we investigated whether depression of the contractile responses to isoproterenol can be detected at the level of the isolated failing cardiomyocytes and, if present, whether the depression could be related to alterations in -AR signaling assessed in membranes extracted from isolated cardiomyocytes. We thus measured contractile and L-type Ca²⁺ current responses to isoproterenol in cardiomyocytes isolated from healthy and failing hearts, along with the -AR density and -AR-stimulated cAMP generation determined in membranes derived from the isolated cardiomyocytes.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Canine preparations of tachycardia-induced heart failure. All experimental procedures were performed in accordance with the guidelines of the Canadian Council for Animal Care and monitored by an institutional animal care committee. Forty-five mongrel canines (18-25 kg) of either sex were anesthetized (Na thiopental, 25 mg/kg, maintenance: isoflurane 1.25%) and mechanically ventilated. A Swan-Ganz catheter inserted via the external jugular vein was introduced into the pulmonary artery. After stabilization, pulmonary capillary wedge pressure and cardiac output (thermodilution technique) were measured. Two-dimensional echocardiography (model 77020AC, Hewlett-Packard) was performed, applying the probe onto the left parasternal area; the video images were recorded on VHS tape (model AG-6300, Panasonic) and subsequently analyzed to estimate the left ventricular end-diastolic dimensions. Under sterile conditions, a bipolar pacing electrode introduced through a neck incision into the right external jugular vein was positioned under fluoroscopy so that its tip lay at the right ventricular apex. The electrode was connected to a pacemaker (model SX 5984, Medtronic, Minneapolis, MN) placed in a subcutaneous pocket and activated after recovery from surgery (3 days later) at a rate of 240 beats/min. Pacing was maintained (ascertained daily with a stethoscope) until the development of overt heart failure (12). Each dog was closely monitored to detect clinical signs of cardiac failure (ascites, dyspnea, fatigue, lack of appetite, weight gain), which became apparent after 4-6 wk.

Plasma catecholamine. Venous blood samples (7 ml) were drawn from the conscious animals in the basal state and again before the terminal study. Samples obtained were then placed in heparinized tubes (0.1% EDTA, 0.2% glutathione) and centrifuged at 4,000 g for 15 min. The plasma was collected and stored at 80°C. Norepinephrine levels were measured by HPLC (12).

Terminal study (failing hearts and healthy controls). Once clinical signs of overt heart failure (particularly, signs of respiratory distress) were apparent, the terminal study was promptly scheduled. Twenty-two dogs served as healthy controls (the innocuous character of pacemaker insertion was controlled in 3 sham-operated dogs). The dogs were anesthetized (meperidine 50 mg iv and carefully titrated Na thiopental intravenous injection; maintained with isoflurane 1%) and mechanically ventilated. Hemodynamic and functional measurements (wedge pressure, cardiac output, left ventricular dimensions) were repeated for comparison with prepacing values. The hearts were then exposed through a left thoracotomy, excised, and immediately placed in cold (4°C) Tyrode solution (in mM: 128 NaCl, 1 MgSO₄, 0.47 NaH₂PO₄, 11 dextrose, 4.5 KCl, 2 CaCl₂, and 20 NaHCO₃, pH 7.4). All chemicals were obtained from Sigma Chemical, St. Louis, MO, unless specified otherwise. Tissue blocks (0.5-1.0 g) were dissected in cold Tyrode solution from the left ventricular anterior wall and used to prepare homogenates from which crude sarcolemmal membranes were extracted. The remainder of the anterior wall was excised and kept in cold Tyrode solution in which a large diagonal branch of the left anterior descending coronary artery was cannulated for the cardiomyocyte isolation procedure.

Preparation of isolated cardiomyocytes. Perfusion was first performed using Tyrode solution (0.3 mM Ca²⁺, gassed with a 95% O₂-5% CO₂ mixture) to remove blood. Afterward, perfusion was instituted with Ca²⁺-free HEPES buffer (in mM: 115 NaCl, 5 KCl, 35 sucrose, 10 HEPES, pH 7.0, 10 dextrose, and 4 taurine) supplemented with 5 mM nitriloacetic acid (5 min). Perfusion was changed to HEPES buffer solution containing 0.3 mM Ca²⁺ and then to HEPES containing 0.05% collagenase (type A, Boehringer Mannheim, Laval, Canada), 0.02% trypsin inhibitor (type II-s), and 0.28 mg/ml protease (type XIV) for 25 min, all solutions being oxygenated and maintained at 37°C. After the collagenase-protease digestion, the perfused region was dissected and transferred to a 50-ml Erlenmeyer flask containing 10 ml of 0.3 mM Ca²⁺ HEPES-collagenase and incubated at 37°C under a stream of O₂ for 20 min. This procedure was repeated four times using fresh HEPES-collagenase solution, and the supernatant was collected and filtered (200-µm nylon filter) after each incubation period. The collected aliquots, which contained the isolated cardiomyocytes, were centrifuged (50 g) for 1 min, and the pellets were resuspended in 0.3 mM Ca²⁺-HEPES solution (pH 7.4). Cardiomyocyte enrichment and their morphological integrity were verified with an inverted microscope (Diaphot, Nikon, Tokyo, Japan).

Contractile responses. Cardiomyocytes were placed in a 200-µl chamber on the plate of an inverted microscope equipped with phase-contrast optics (Diaphot, Nikon, Tokyo, Japan). The cells were perfused with standard Tyrode solution (2 mM Ca²⁺) at a rate of 1 ml/min and field stimulated at 0.5 Hz (pulses of 5-ms duration and 1.5 threshold current intensity). Bath temperature was maintained at 37°C with a custom-made Peltier-effect device. A selected cardiomyocyte was visualized under magnification (20×) using a charge-coupled device camera (model KP-M1U, Hitachi Demshi, Tokyo, Japan) and television video display. Contraction was measured by tracking motion of the edges on either side of the cell along its longitudinal axis with the use of a video edge motion detector (model VED 105, Crescent Electronics, Sandy, UT; described in Steadman et al., Ref. 34). The analog output of the edge detector was converted to a digital format by a data-acquisition system (Digidata 1200, Axon Instruments, Foster City, CA) and transferred to a personal computer hard disk under the control of Axotape software. Analysis was performed using the Clampfit program of the pClamp 6 software package, extracting the resting length, magnitude of contraction (resting length  minimum length), and the maximum time derivative of the cell length during contraction and relaxation (negative and positive dL/dt, respectively). Measurements were made on a signal averaged from 10 consecutive contractions.

L-type Ca²+ currents. Current-voltage relationships were determined in patch clamp experiments. Cardiomyocytes were transferred into a small (0.2 ml) tissue bath placed on the stage of an inverted microscope and superfused with Na⁺- and K⁺-free solution containing (in mM) 140 tetraethylammonium chloride (TEA-Cl), 0.5 MgCl₂, 10 HEPES (pH 7.3-7.4 adjusted with TEA-OH), 10 dextrose, 2 4-aminopyridine, and 5 CaCl₂. The L-type inward calcium current (I_Ca,L) was recorded in the whole cell configuration of the patch clamp technique using a voltage clamp amplifier (List Medical, EPC-7) and suction pipettes filled with a solution containing (in mM) 125 CsCl, 20 TEA-Cl, 10 HEPES, 10 EGTA, and 5 Mg₂-ATP. Currents were monitored on an oscilloscope and stored on a microcomputer hard disk under the control of computer software (pClamp 6, Axon Instruments) that was also used to generate the voltage clamp protocols and for data analysis. The voltage dependence of I_Ca,L activation was determined by delivering depolarizing voltage-clamp pulses (1-s duration) in 10-mV increments every 10 s from a holding potential of 50 mV.

Crude membrane preparations. Whole tissue blocks, isolated cardiomyocytes, or the noncardiomyocytic cell fraction were placed in lysis solution (5 mM Tris · HCl, pH 7.4, 2 mM EDTA, 5 µg/ml leupeptin, 5 µg/ml soybean trypsin inhibitor, and 10 µg/ml benzamidine) and homogenized with a Polytron (3 bursts of 10 s at maximum speed; Brinkmann Instruments, Westbury, NY). The homogenate was centrifuged at 1,000 g for 5 min at 4°C (whole tissue homogenate was filtered through 3 layers of cheesecloth before the centrifugation step). The supernatant was removed and centrifuged at 45,000 g at 4°C for 20 min. The supernatant was discarded, and the pellet was resuspended in 10 ml of the lysis solution and centrifuged at 45,000 g. This step was repeated twice, and the final pellet was resuspended in an ice-cold buffer containing 75 mM Tris · HCl (pH 7.4), 5 mM MgCl₂, and 2 mM EDTA to a final concentration of 0.5 µg/µl. Protein content was determined by the method of Bradford (5).

Adenylyl cyclase. Measurements were made after Salomon et al. (30) using 10 µg of membrane protein from tissue or cell homogenates in a total volume of 50 µl. The incubation medium included 0.12 mM ATP, 0.5 µCi [-³²P]ATP, 0.1 mM cAMP, 0.053 mM GTP, 2.8 mM phosphoenolpyruvate, 0.2 U pyruvate kinase, 1 U of myokinase, 30 mM Tris · HCl (pH 7.4), 5 mM MgCl₂, 1 mM 3-isobutyl-1-methyl-xanthine, and 0.8 mM EDTA. Reactions were initiated by adding the membranes to the incubation medium (37°C) and lasted for 30 min before being stopped by the addition of 1 ml of ice-cold solution containing 0.4 mM ATP, 0.3 mM cAMP, and [³H]cAMP (25,000 cpm), the latter being used to assess the efficiency of the isolation procedure. The cAMP was isolated by sequential chromatography on a Dowex cation-exchange resin and aluminum oxide. Enzyme activity stimulated by isoproterenol (10⁸-10⁴ M), NaF (10 mM NaF), and forskolin (100 µM) were determined in duplicate. Isoproterenol concentration-response curves were fitted to a sigmoid using the Allfit computer software.

-AR binding studies in tissue and cell membrane preparations. Binding assays were conducted using [¹²⁵I]iodocyanopindolol ([¹²⁵I]CYP) in crude membranes derived from whole tissue or isolated myocytes (prepared as described above). In preliminary experiments, membrane preparations (10 µg proteins) were incubated in duplicate assay tubes at room temperature for 90 min in a final volume of 500 µl containing 75 mM Tris · HCl (pH 7.4), 5 mM MgCl₂, 2 mM EDTA, and [¹²⁵I]CYP at concentrations of 10-400 pM. Nonspecific binding was defined as the one not being displaced by 100 µM isoproterenol. Because there was no difference between dissociation constant (K_d) values for [¹²⁵I]CYP measured in healthy and failing heart membrane preparations (data not shown), the subsequent experiments were done in triplicate using a single saturating concentration of [¹²⁵I]CYP (300 pM), nonspecific binding being determined using 10 µM alprenolol. Binding reactions were stopped by rapid filtration over Whatman GF/C fiberglass filters using a cold buffer solution containing 25 mM Tris · HCl (pH 7.4), 5 mM MgCl₂, and 2 mM EDTA. To determine the proportion of ₂-AR in the total -AR, competition experiments were done using [¹²⁵I]CYP (50 pM) and the selective ₂-AR antagonist ICI-118,551 (10¹²-10⁴ M). It is possible that, at 50 pM [¹²⁵I]CYP, the ₂-AR sites might be slightly overestimated in competition experiments assessing the ₁/₂-AR ratio (25), but 50 pM [¹²⁵I]CYP has been used by other investigators for this purpose (8). Data were subjected to nonlinear least-squares regression analysis (15).

-AR binding studies in tissue slices. Frozen tissue blocks obtained from six healthy and six failing hearts were placed in ice-cold Dulbecco's PBS (DPBS) and thawed. Micropunctures (2 mm diameter and 350 µm thickness) were prepared using a McIlwain tissue chopper and micropunch (37, 38) and placed in separate wells of tissue culture plates, one per well containing 500 µl of ice-cold DPBS. Saturation experiments were done at concentrations of 0.25-3.5 nM of the hydrophillic -AR antagonist [³H]()-CGP-12177 (specific activity: 47-53 Ci/mmol). Groups of six wells (replicates) were used to determine total binding, and three replicates were used for nonspecific binding (15% at K_d as measured in the presence of (±)timolol, 10⁵ M). After incubating at 4°C for 3-5 h, tissues were washed in 500 µl ice-cold DPBS (2 × 5 min) and prepared for scintillation counting. Receptor density was determined by Headie-Hofstee analysis of saturation curves (40).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cardiac output was significantly reduced by 61 ± 3.4% after rapid ventricular pacing for 4-6 wk compared with prepacing measurements performed in the same animals (from 2.9 ± 0.1 to 1.0 ± 0.1 l/min, P < 0.001). There were also statistically significant increases in the norepinephrine plasma levels (from 364 ± 70 to 1,336 ± 186 pg/ml) and left ventricular end-diastolic volume (from 62 ± 2.7 to 102 ± 3.8 ml).

Cardiomyocyte contraction and relaxation. As illustrated in Fig. 1, the cardiomyocytes isolated from failing hearts displayed an elongated resting length (L_max) compared with the ones that were isolated from healthy hearts (Table 1), in accordance with previous work from other laboratories (33). The magnitude of contraction dL measured under basal conditions was significantly depressed in failing cardiomyocytes, whether this variable was expressed in micrometer units or as a percentage of resting cell length (Table 1). The rate of cell shortening (negative dL/dt) during electrically induced contraction and the lengthening rate during relaxation (positive dL/dt) were similar between failing and healthy cardiomyocytes when expressed in micrometer units, but they were significantly smaller when the contraction and relaxation were expressed with reference to the resting cell length (Table 1).

View larger version (134K): [in this window] [in a new window]   Fig. 1.   Representative phase-contrast micrographs of cardiomyocytes isolated from a healthy (A) and a failing (B) heart. Failing cardiomyocytes displayed significantly longer resting length compared with the healthy ones. Note also that the failing cell contours were wavy.

Effect of isoproterenol on cardiomyocyte contraction. In both the failing and the healthy cardiomyocytes, isoproterenol induced significant augmentation of the electrically induced contractions at all concentrations tested (comparing the basal measurements with those made in the presence of isoproterenol). Inotropic responses to isoproterenol (_Iso dL/L_max) were concentration dependent in the healthy group but not among the failing cardiomyocytes (Fig. 2); the response to isoproterenol 10⁸ M was significantly greater in the healthy group than among the failing cells. Similar results were obtained whether the responses were expressed in micrometer units (not shown) or as a percentage of resting cell length (Fig. 2: _Iso dL/L_max). A 10⁸ M concentration was the maximum that could be used, beyond which the failing cardiomyocytes as well as the healthy ones developed rapid and irregular spontaneous activity, preventing electrical stimulation.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Depression of the isoproterenol-induced increments (Iso) in the magnitude of contraction (dL/Lmax, normalized with respect to maximum cell length) in failing cardiomyocytes compared with healthy ones.

Effects of isoproterenol on the rates of contraction and relaxation. Isoproterenol induced increases [_Iso (dL/dt)/L_max] in the rate of shortening (Fig. 3A) and relaxation (Fig. 3B). The increases in (dL/dt)/L_max of the electrically induced contractions were statistically significant at all concentrations in both the healthy and failing cardiomyocytes. However, significant concentration dependence could be demonstrated only in the healthy cell group, and the maximum responses induced in this group were significantly greater than among the failing cells (Fig. 3, A and B). The results were similar whether considering the responses normalized to resting cell length [Fig. 3; _Iso (dL/dt)/L_max] or expressed in micrometers per second units (not shown).

View larger version (10K): [in this window] [in a new window]   Fig. 3.   Depression of the Iso in the rate of shortening (A) and relaxation (B) in failing versus healthy cardiomyocytes. Rates of shortening and subsequent relaxation are expressed as %changes with reference to the maximum cell length (Lmax) on a per second basis. Isoproterenol exerted significant effects in both cell types at all concentrations (i.e., significantly different from zero). The responses to isoproterenol 108 M (and 5 × 109 in the case of relaxation) were significantly depressed (*) in failing cardiomyocytes versus healthy ones.

Effects of isoproterenol on L-type Ca²+ currents. The cell capacitance was significantly increased in the cardiomyocytes isolated from the failing hearts (261 ± 5.2 pF, n = 29 in 6 hearts) in comparison with those isolated from healthy hearts (170 ± 7.4 pF, n = 28 in 8 hearts, P < 0.05). Statistical analysis of the current-voltage curves (using the raw, peak current values) indicated that the currents were significantly voltage dependent (as expected) and that they were significantly increased under isoproterenol (Fig. 4A). There was a tendency for the isoproterenol effect to be greater among the healthy than among the failing cells (interaction: P = 0.068). When the peak current values were normalized with respect to capacitance (Fig. 4B), the isoproterenol effect emerged as being significantly greater among the healthy than among the failing cells. Interestingly, the normalized currents measured under basal conditions were similar between the two groups. When the slope conductances extracted from the curves determined in each cell were analyzed, we found that this variable was not different between the two groups under basal conditions and that it was increased under isoproterenol, this effect being much greater in the healthy than in the failing cells.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Isoproterenol-induced stimulation of the L-type inward calcium current (ICa,L) in failing versus healthy hearts (A: raw measurements, B: values normalized with reference to cell capacitance). In A, there was a tendency for the isoproterenol effect to be smaller among the failing cells than among the healthy ones (group treatment interaction: P = 0.068); this difference was statistically significant (P = 0.002) when the normalized current values were analyzed.

Adenylyl cyclase and -AR density: sarcolemmal membranes derived from isolated cardiomyocyte homogenate. The basal adenylyl cyclase activity as well as its maximum isoproterenol-stimulated activity were found to be similar in the healthy and failing hearts (Table 2). The concentration-response curves for isoproterenol-stimulated activity in healthy and failing cardiomyocytes were superimposable (Fig. 5A) (EC₅₀ failing: 5.0 ± 0.6 × 10⁷ M; healthy: 6.0 ± 0.2 × 10⁷ M). The maximal adenylyl cyclase responses to NaF and forskolin were also similar in the two groups. The total -AR numbers were, in fact, found to be significantly higher in the failing (558 ± 45 fmol/mg protein, n = 45) than in the healthy cardiomyocytes (386 ± 47 fmol/mg protein, n = 18, P = 0.04), a surprising finding that contrasted with the data obtained in membranes derived from whole tissue homogenate (see below). The ₁/₂ ratios measured in competition binding experiments were similar in the preparations from the healthy and failing hearts (Fig. 5B and Table 3). Regression analysis of -AR numbers as a function of cardiac output reduction and the increase in left ventricular end-diastolic dimensions did not indicate any clear relationship, but there was a significant trend for the -AR numbers to decrease as the left ventricular filling pressure increased (y = 706  15.6x, r = 0.46, P = 0.03, n = 21 data points). This result would suggest the possibility that -AR density might actually be increased at a moderate degree of functional alteration while being reduced with more severe ventricular dysfunction.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Preserved isoproterenol-stimulated adenylyl cyclase activity (A) and 1/2-adrenergic receptor ratio (B) in membranes derived from homogenates of cardiomyocytes isolated from failing hearts. A: the data are expressed as increments beyond the basal adenylyl cyclase activity (taken as zero) and compared between healthy and failing hearts. Mean ± SE; n, number of experiments per group. The isoproterenol-stimulated adenylyl cyclase activities were similar when measured in membranes extracted from cells isolated from either healthy or failing ventricles. B: competition of [125I]iodocyanopindolol (CYP) binding with the 2-selective antagonist ICI-118,551. Points are averages of triplicate determinations. Curves were best fitted to a 2-site model using a nonlinear least-squares regression program. Data are expressed as the fraction of total [125I]CYP binding. The 1/ 2 ratios were found to be similar (healthy: 79/21, K1 = 2.2 M, K2 = 1.7 nM vs. failing: 75/25, K1 = 0.9 M, K2 = 1.4 nM).

View this table: [in this window] [in a new window]   Table 3.   1/2-adrenergic receptor ratio measured in membranes from healthy or failing hearts

Adenylyl cyclase and -AR density: sarcolemmal membranes derived from whole tissue homogenate. In contrast to the data obtained in membranes derived from isolated cardiomyocytes, the basal values of adenylyl cyclase activity were significantly lower in preparations from failing than in those from healthy hearts (Table 2, basal). The concentration-response curve for isoproterenol-stimulated activity was markedly depressed in failing hearts (Fig. 6A) with a significantly lower maximum isoproterenol-stimulated activity (Fig. 6A and Table 2, isoproterenol), and higher EC₅₀ (failing: 1.2 ± 0.4 × 10⁶ M, healthy: 2.0 ± 0.9 × 10⁷ M). The responses to maximally stimulating concentrations of NaF and forskolin were similar in the two groups (Table 2). Total -AR numbers were similar in failing (548 ± 67 fmol/mg protein, n = 41) and in healthy hearts (530 ± 86 fmol/mg protein, n = 17). Competition binding curves (using the ₂-AR selective antagonist ICI-118,551 and the nonselective [¹²⁵I]CYP) could best be fitted to a two-site model (Fig. 6B) with the high-affinity site (K₂) and lower-affinity site (K₁) corresponding to ₂-AR and ₁-AR binding, respectively. The ₁/₂ ratio was significantly lower in the failing than in the healthy ventricles (Table 3).

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Depression of isoproterenol-stimulated adenylyl cyclase activity (A) and modification of the 1/2-adrenergic receptor ratio (B) in membranes derived from failing ventricular tissue homogenates. Same format as in Fig. 5. A: isoproterenol-stimulated adenylyl cyclase activity was significantly depressed when measured in membranes extracted from whole tissue homogenates of failing hearts (*P < 0.05). B: the 1/2 ratio was significantly lower in the membranes extracted from failing ventricular homogenates (45/55, K1 = 6 M, K2 = 2 nM) than in membranes from healthy ones (83/17, K1 = 0.3 M, K2 = 0.5 nM).

-AR density: whole tissue slices. No statistically significant difference was found between healthy (113 ± 15) and failing hearts (90 ± 11 fmol/mg protein, P = 0.26) in [³H]CGP-12177 binding studies performed in whole tissue slices.

-AR in membranes derived from the homogenate of a noncardiomyocytic cell fraction. Only weak basal adenylyl cyclase activity was detected (healthy: 3.2 ± 1.5, failing: 4.2 ± 1.0 pmol cAMP · mg protein¹ · min¹, not significant), and the responses to isoproterenol stimulation were irregular. However, relatively high adrenergic receptor numbers were detected, without any statistical difference between preparations extracted from failing (337 ± 47 fmol/mg protein, n = 14) and healthy hearts (196 ± 53, n = 5, P = 0.18). Neither was there any statistically significant difference in the ₁/₂ ratios between the two groups (Table 3). (Note, however, that the ₁/₂ ratio measured in membranes derived from whole tissue homogenates of failing hearts was similar to the one measured in membranes from the noncardiomyocytic cell fraction.)

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We found that the magnitude and rate of cell shortening as well as the rate of relaxation were decreased in failing versus healthy cardiomyocytes under basal conditions. These changes are consistent with a reduced myofibrillar content (32) and altered mechanisms of E-C coupling (27, 28, 36) in this model. The latter include reductions in L-type Ca²⁺ current (24) and dihydropyridine receptor density (Ref. 24, but not 36), reduction in ryanodine receptor density (14, 36), and reduction in sarcoplasmic reticulum Ca²⁺ ATPase (SERCA2) activity (26). We also found that the isoproterenol-induced changes in the magnitude and rate of cell shortening as well as the rate of relaxation were depressed in the failing cardiomyocytes at agonist concentrations of 5 × 10⁹-10⁸ M. Still higher isoproterenol concentrations (10⁶ M) could be achieved in the bath during the patch clamp experiments because of the presence of EGTA in the pipette and, again, the isoproterenol-induced increments in L-type Ca²⁺ current were found to be reduced, in agreement with Kääb et al. (19). The reduction was apparent when the raw measurements were analyzed and were statistically significant when normalized to cell membrane capacity. Such normalization is a common practice in the study of ionic currents (19, 24, 27), and its rationale is that the cell capacity is proportional to the total surface of the cell membrane (10).

The responses to isoproterenol were depressed despite the fact that there was no reduction in -AR density and -AR-mediated cAMP generation in crude membranes derived from isolated failing cardiomyocyte homogenates compared with healthy ones. The simplest interpretation of these results would be that the reduced contractile responses to isoproterenol are related to alterations in the mechanisms of E-C coupling and the contractile machinery in the face of a preserved cAMP generating capacity.

The reduced basal rate of relaxation is consistent with diminished SERCA2 expression and activity as well as reduced Ca²⁺ uptake in preparations extracted from hearts with tachycardia-induced failure (26, 27). However, Na⁺/Ca²⁺ exchange activity, the other major Ca²⁺ removal system of the heart, was found to be increased and appeared to fully compensate for the reduction in sarcoplasmic reticular Ca²⁺ uptake (27), suggesting that intracellular Ca²⁺ was extruded outside from the cell instead of being stored into the sarcoplasmic reticulum. Interestingly, isoproterenol's capacity to stimulate sarcoplasmic reticulum Ca²⁺ uptake was preserved (although with a longer time constant than in healthy hearts) (27).

Thus the present study suggests that the alterations in contractile responses to isoproterenol observed in the failing cardiomyocytes isolated from our preparations might have been due to the alterations in the contractile machinery that were already apparent in the functional depression observed under basal conditions. However, data obtained in other animal preparations of heart failure (2, 3) and in human heart failure (6, 7, 31) indicate that reductions in -AR density and -AR-mediated cAMP generation may indeed be a mechanism limiting the responses to isoproterenol in other pathophysiological situations.

It is important to note that, although many studies have been devoted to various physiological and biochemical measurements made in isolated cardiomyocytes (1, 16, 24, 27, 28), this is the first one to report measurements of -AR density and -AR-mediated cAMP generation made in crude membranes extracted from isolated cardiomyocyte homogenate. When we carried out our measurements in membranes extracted from whole tissue homogenate, total -AR density was found to be unchanged (in agreement with our measurements made in tissue slices). Yet we found that the ₁/₂-AR ratio was reduced, suggesting that the ₁-AR number may have been reduced. This result, together with the reduction in adenylyl cyclase activity that we found in membranes extracted from whole tissue homogenate, is in agreement with the data reported by Kiuchi et al. (20).

The presence of a noncardiomyocyte fraction could be a factor explaining the fact that data concerning -AR density and -AR-mediated cAMP generation differed between whole tissue membrane preparations and those from isolated cardiomyocytes. There was a relative increase in the ₂-AR number, which happened to be predominantly expressed in cardiofibroblasts (21). Their number could be increased as a consequence of changes in the cellular composition of the cardiac tissue as, for instance, an increase in the proportion of interstitial cells versus cardiomyocytes (22). This possibility is in agreement with our observation that the ₁/₂-AR ratio in whole tissue membranes of failing hearts was similar to that in the membranes derived from the noncardiomyocytic cell fraction (Table 3). Accordingly, an increase in noncardiomyocytic cells with low basal and -AR-mediated adenylyl cyclase activity could explain the blunting of isoproterenol-stimulated enzyme activity that we observed in membranes from whole tissue homogenate but not those from isolated cardiomyocytes. The fact that biochemical analysis of membranes derived from isolated myocytes and whole tissue may not necessarily yield the same answer deserves further consideration.

When in the course of our study we became aware of the trend that -AR density and -AR-mediated cAMP generation were not reduced in crude membranes derived from failing cardiomyocytes, we paid much attention to ensuring that a sufficiently severe degree of failure was achieved. Several indexes were used toward this end. At the whole organ and circulatory level, cardiac output was reduced by 61%, pulmonary capillary wedge pressure was increased fourfold, left ventricular end-diastolic volume was increased by 69%, and there was a 3.6-fold increase in plasma norepinephrine levels. At the level of the cardiomyocytes, they were found to be elongated and distorted and their basal contractile activity (as well as responses to isoproterenol) was depressed. The terminal study was scheduled once clear clinical signs of heart failure had developed (ascites, fatigue, lack of appetite). Close attention was paid to the development of signs of respiratory distress, which was the signal for prompt scheduling of the terminal study. Thus the canine preparations that were studied presented a clear profile of cardiac failure. To further investigate whether alterations of -AR signaling might have been related to the severity of heart failure, we investigated correlations between the various circulatory variables (reduction in cardiac output, etc.) and either -AR density or -AR-mediated cAMP generation. Pulmonary capillary wedge pressure was the only variable showing some degree of correlation (negative) with -AR density, and even then it was a weak one.

Not only was total -AR density not reduced in crude membranes obtained from failing cardiomyocytes, but, unexpectedly, we found it to be significantly increased (without significant change in the ₁/₂ ratio and adenylyl cyclase activity) compared with the healthy state. These results should be interpreted in the light of the fact that the canine preparations studied herein had been in the failing state for a short period of time; therefore, modifications of the receptor density observed in long-standing heart failure might not have had time to develop. It is also possible that the reduction in tissue norepinephrine levels and a reduced capacity of the cardiac sympathetic neurons to release norepinephrine (previously shown in our preparations; Ref. 12) might have caused a reactive increase in receptor density. Yet one other explanation might be an increase in -AR density associated with impaired energy metabolism (9, 35), because a decrease in intracellular high-energy phosphates is intimately involved in the development of tachycardia-induced heart failure (13, 26). Interestingly, the -AR internalization process appears to be ATP dependent (9, 35). Thus chronic energy starvation might be a common factor explaining the postulated alteration of proteins involved in E-C coupling (27) as well as an increase in sarcolemmal -AR density (when it occurs). On the other hand, sustained rapid pacing is a complex situation that has also been shown, in isolated neonatal rat cardiomyocytes, to induce -AR internalization without modification of total -AR numbers (39). Thus various mechanisms with divergent outcomes may be involved in the presence of sustained rapid pacing.

Perspectives