Atrial Natriuretic Factor Reduces Cell Coupling in the Failing Heart, an Effect Mediated by Cyclic GMP

Abstract

The influence of the atrial natriuretic factor (ANF) on heart-cell communication was investigated in cell pairs isolated from the ventricle of cardiomyopathic hamsters (BIO TO2; 11 months old), and the results were compared with controls (F1B) of same age. The results indicated that ANF (10⁻⁸M) added to the bath caused a decline in junctional conductance (gj) of 48 ± 2% (n = 15) within 90 s. The effect of ANF was suppressed by HS-142-1, a specific antagonist of guanylyl cyclase ANF receptor. Moreover, the decline in gj elicited by ANF was related to the synthesis of cyclic guanosine monophosphate (cGMP). Indeed, dibutyryl-cGMP (10⁻⁴M) decreased gj by 80 ± 3.5% (n = 15) within 90 s, and zaprinast, a selective inhibitor of cGMP phosphodiesterase, enhanced the effect of ANF on gj. The possible relationship between ischemia, ANF release, and impairment of cell coupling is discussed.

Atrial natriuretic factor (ANF) is a 28-amino acid peptide that was originally found to be secreted from atrial myocytes after a preliminary storage in granules ^(1-4). Evidence was presented that ventricular myocytes also are involved in the synthesis of ANF. The intraventricular conduction system of the rat seems to be rich in ANF ⁽⁵⁾, and ventricular transcription of ANF has been found in hypertrophied rat ventricle, as well as in the cardiomyopathic hamster ^(6,7). According to Franch et al. ⁽⁷⁾, the myocytes of cardiomyopathic ventricular muscle represent an appreciable source of ANF during the process of congestive heart failure (CHF). The increase of ventricular ANF messenger RNA (mRNA) expression was found in adult cardiomyopathic hamsters (90 days old) in which necrotic foci are the cause of severe destruction of the ventricular structure ⁽⁸⁾.

In addition, an enhanced ventricular ANF gene expression has been described in patients with hypertrophic cardiomyopathy without clinical or hemodynamic signs of CHF ⁽⁹⁾. It is known that ANF has a powerful natriuretic activity, regulates extracellular fluid volume, causes relaxation of vascular smooth muscle, inhibits the activity of the renin-angiotensin system, and controls vasopressin release ⁽¹⁰⁾.

In cardiomyopathic hamsters with CHF, ANF-specific granules were found in ∼20% of ventricular myocytes, whereas in the myopathic animals without heart failure, the granules were rarely found ⁽⁸⁾. No information is available about whether ANF alters the electrical properties of heart cells or the process of cell-to-cell communication. Considering that an increased plasma level of ANF has been found in some patients that develop supraventricular tachycardia, as well as in patients undergoing atrial pacing ^(11,12), it is important to investigate the influence of this hormone on the spread of electrical information between cardiac myocytes. In our study, this problem was investigated in cell pairs isolated from the ventricles of cardiomyopathic hamsters (TO-2), which represents a good model for cardiomyopathy and hypertrophy in humans ^(13,14).

METHODS

Male TO-2 cardiomyopathic Syrian hamsters (11 months old) (Biobreeders, Fitchburg, MA, U.S.A.) and healthy male F1B control hamsters of the same age were used. Both the control and cardiomyopathic animals were kept in air-conditioned facilities at the animal house with a normal laboratory animal diet and tap water ad libitum. The animals were anesthetized with sodium pentobarbital (50 mg/kg, i.p.), and the heart was removed with the animals under deep anesthesia.

Cell pairs were obtained by enzymatic dispersion of hamster ventricle by the method of Powell and Twist ⁽¹⁵⁾ and Tanigushi et al. ⁽¹⁶⁾.

The heart was removed and immediately perfused with normal Krebs solution containing (in mM): NaCl, 136.5; KCl, 5.4; CaCl₂, 1.8; MgCl₂, 0.53; NaH₂PO₄, 0.3; NAHCO₃, 11.9; glucose, 5.5; and HEPES, 5; with pH adjusted to 7.4. After 20 min, a calcium-free solution containing collagenase (0.4%; Worthington Biochemical Corp., Lakewood, NJ, U.S.A.) was recirculated through the heart for 1 h. The collagenase solution was washed out with 100 ml of recovery solution containing (in mM) taurine, 10; oxalic acid, 10; glutamic acid, 70; KCl, 25; KH₂PO₄, 10; glucose, 11; and EGTA, 0.5; with pH adjusted to 7.4. All solutions were oxygenated with 100% O₂.

The ventricles were minced (1- to 2-mm thick slices), and the resulting solution was agitated gently with a Pasteur pipette. The suspension was filtered through nylon gauze, and the filtrate centrifuged for 4 min at 22 g. The cell pellets were then resuspended in normal Krebs solution. All experiments were done at 36°C.

Suction pipettes were pulled from microhematocrit tubing (Clark Electromedial Instruments) by means of a controlled puller (Narishige, Tokyo, Japan), and their tips were polished with a microforge (Narishige). The pipettes, which were prepared immediately before the experiment, were filled with the following solution (in mM): KCl, 125; NaCl, 10; MgCl₂, 3; TEA, 20; Na₂-ATP, 5; EGTA, 10; and HEPES, 5; with pH adjusted to 7.3. In some experiments, KCl was replaced by Cs aspartate. The resistance of the filled pipettes varied from 2.5 to 3 MΩ.

Drugs

ANF, dibutyryl-cyclic guanosine monophosphate (cGMP), and zaprinast were from Sigma Chemical Co. (St. Louis, MO, U.S.A.). HS-142-1, an antagonist of ANF receptor, was kindly provided by Kyowa Hakko Kigyo Co., Tokyo, Japan.

Experimental procedures

All experiments were performed in a small chamber mounted on the stage of an inverted phase-contrast microscope (Diaphot; Nikon, Tokyo, Japan). The junctional resistance was determined in cell pairs with the use of two separate voltage-clamp amplifiers. Gigaohm sealing was achieved in each cell, and then the surface cell membrane of both cells was broken by application of a stronger suction (−30 to −65 cm H₂O), and a whole-cell clamp configuration was produced. Each pipette was connected to a separate voltage-clamp amplifier (Dagan Corp., Minneapolis, MN, U.S.A.) that made possible the control of the nonjunctional membrane potential in each cell, as well as the voltage gradient across the intercellular junction.

The experimental procedure consisted of holding the membrane potential of both cells at −40 mV. Cell 1 was then pulsed to 0 mV while the membrane potential of cell 2 was maintained unchanged. A voltage was created across the junctional membrane (V₁), and a compensating current of opposite polarity recorded from pipette 2 (1₂) represents the current flowing through the gap junction. As I₂ equals V₁/rj, the junctional resistance (rj) can be easily estimated. Data acquisition and command potentials were controlled with PCLAMP software (Axon Instruments, Foster City, CA, U.S.A.).

Series resistances (Rs₁ and Rs₂) originating from the tips of the micropipettes were compensated electronically before the experiment and checked periodically during the experiment. When necessary, the gap junctional conductance (gj) was corrected by taking into consideration the changes in series resistance. For this, the following equation was used: Equation 1

Voltage and current signals were displayed simultaneously on an oscilloscope (Tektronix 5113, Beaverton, OR, U.S.A.) and chart recorder (Gould 2400, Valley View, OH, U.S.A.).

Statistical analysis

Numeric data are expressed as mean ± SEM. Student's t test was used to estimate statistical significance, defined at a value of p < 05.

RESULTS

To investigate the influence of ANF on junctional conductance (gj) of myopathic heart cell pairs, the hormone was added to the bath while gj was continuously monitored. As shown in Fig. 1, the gj was reduced by ANF (10⁻⁸M). Average results from 15 cell pairs indicated a decline of gj from 10 ± 2.1 nS (n = 15) to 5.2 ± 1.98 nS (n = 15), which represents a decrease of gj of 48% within 90 s (Fig. 1). Studies of the voltage-current curve performed in a single cell pair from the cardiomyopathic cell population (see ¹⁴) indicated a linear relationship between the junctional current (I₂) and the transjunctional voltage (V₁). The mean slope of the straight line obtained by pulsing initially cell 1 and then cell 2 was estimated by linear-regression analysis and found to be 100 mV/nA (r² = 0.99; n = 6) in this particular cell pair (see Fig. 2). The slope corresponds to a junctional resistance of 100 MΩ. ANF (10⁻⁸M) changed the slope of the current-voltage relationship to 142 mV/nA (r² = 0.98; n = 5), which corresponds to a junctional resistance of 142 MΩ. This finding indicates a decline in gj (or an increase in junctional resistance) for different values of transjunctional voltage. A similar effect of ANF was found in a cardiomyopathic cell pair that has a low gj (Fig. 1, left).

FIG. 1: Left: Effect of atrial natriuretic factor (ANF; 10⁻⁸ M) on gap junctional conductance (gj) of single myopathic cell pair with a low gj. a: control; b: 90 s after ANF was added to the bath. I₂, junctional current; I₁, sum of current flowing across the nonjunctional membrane of cell 1 and current flowing through the gap junction. V₁, transjunctional voltage. Calibration at I₂, 0.1 nA, and at V₁, 40 mV. Pulse duration, 1 s. Middle: Dose-dependent effect of ANF on gj. Each bar is the average of 15 experiments. Vertical line at each bar, SEM. Right: Average effect of ANF (10⁻⁸ M) on gj of 10 cell pairs, with vertical line at the bar indicating SEM. At right, suppression of the effect of ANF caused by previous administration of HS-142-1 (25 μg/ml) to the bath.

FIG. 2: r ² = 0.99; n = 6) was found for the control (A). B, Curve obtained from the same cell pair after 3 min of atrial natriuretic factor (ANF; 10⁻⁸ M) administration, showing a slope of 142 mV/nA (r ² = 0.98; n = 6).

The effect of ANF on gj was dose dependent (Fig. 1) and was completely reversible. Moreover, no time dependence of gj was found in the control or in the cell pairs exposed to ANF for 4 min. Measurements of the time constant (t_m) of cell membrane performed in single myocytes by using electrotonic potentials recorded under current-clamp and exposed to ANF (10⁻⁸M) indicated no significant change in t_m and consequently no alteration in membrane resistance.

The series resistance, which was compensated electronically at the beginning of the experiment, remained unchanged, with the exception of one experiment in which the changes in gj elicited by ANF were corrected for the alteration in series resistance (see Methods).

A major question is how ANF reduces the exchange of electrical messages between apposing heart cells. Recently it was reported that the compound HS-142-1 is a specific antagonist of the guanylyl cyclase ANF receptor ⁽¹⁷⁾. It is then important to investigate whether the effect of ANF on gj is related to the activation of this receptor. For this, cell pairs of cardiomyopathic hamsters were exposed to HS-142-1 (25 μg/ml) for 5 min before the administration of ANF (10⁻⁸M) to the bath solution containing the antagonist. As shown in Fig. 1, the effect of ANF on gj was completely suppressed in 10 cell pairs, whereas in other experiments, the effect of ANF was reduced by 75 ± 5.8% (n = 6). HS-142-1, by itself, at least at this concentration, had no effect on gj.

Because it is known that cGMP is produced by ANF in several systems (see ¹⁸), it is conceivable that the decline in gj found with ANF in the cardiomyopathic hamster might be related to the synthesis of cGMP. Because no information is available on whether cGMP influences gj in the failing heart, we performed several studies on the effect of dibutyryl-cGMP on gj. As shown in Fig. 3, dibutyryl-cGMP (10⁻⁴M) reduced gj by 80 ± 3.5% (n = 15) within 90 s after bath application. The effect of the compound, which was dose dependent (Fig. 3), started in ∼40 s, increasing gradually to reach a maximal and steady value 2 min later. In control hamsters, dibutyryl-cGMP also reduced gj (see Fig. 3).

FIG. 3: A: Effect of different doses of dibutyryl-cyclic guanosine monophosphate (cGMP) on gap junctional conductance (gj) of cardiomyopathic cell pairs. Each bar is the average of 15 cell pairs. Vertical line at each bar, SEM. B: Effect of different doses of dibutyryl-cGMP on gj of normal controls. Each bar is the average of 13 cell pairs. Vertical line at each bar, SEM.

To investigate the role of cGMP on the effect of ANF, cell pairs were incubated with zaprinast (100 μM), a selective inhibitor of cGMP phosphodiesterase ⁽¹⁹⁾ for 5-7 min; the gj was measured, and then ANF (10⁻⁸M) was added to the bath containing zaprinast. As shown in Fig. 4 the effect of ANF on gj was appreciably increased when compared with controls in the absence of the phosphodiesterase inhibitor. Zaprinast alone, at least at this concentration, did not change gj. In the control hamsters, zaprinast also enhanced the effect of ANF on gj by 20 ± 1.9% (n = 8; data not shown).

FIG. 4: Increment of the effect of atrial natriuretic factor (ANF; 10⁻⁸ M) on gap junctional conductance (gj) of myopathic cell pairs elicited by zaprinast (100 μM). Bar at left shows effect of ANF alone. Bar at right, ANF was added to the bath containing zaprinast. Each bar is the average of 15 cell pairs. Vertical line at each bar, SEM.

Because ventricular myocytes represent a major source of circulating ANF during CHF ⁽⁷⁾, the question remains whether the ventricular cells of normal hamsters also respond to ANF. ANF (10⁻⁸M) administered to the bath caused a decline of gj in control animals by 40 ± 2.2% (n = 6; data not shown).

DISCUSSION

Our results indicate that ANF reduces appreciably the gj in cardiomyopathic ventricular cells.

The effect of ANF on cell communication is probably mediated by an increase in intracellular cGMP, as a consequence of stimulation of particulate guanylyl cyclase, which was found to be part of the ANF-receptor molecule ^(20,21). Two kinds of observations support this notion: (a) gj was reduced by dibutyryl-cGMP in myopathic cell pairs, and (b) the effect of ANF on gj was incremented by zaprinast, a cGMP phosphodiesterase inhibitor.

A decrease in gj also was described with cGMP in normal rat cardiac myocytes ⁽²²⁾. In neonatal rat heart myocytes, 8-bromo-cGMP reduces gj, probably by phosphorylating the gap-junction proteins ⁽²³⁾. Our results not only support the notion that the decline in gj elicited by ANF in the failing heart was related to formation of cGMP but also indicate that cGMP is a second messenger involved in the control of gj. Because the effect of ANF was blocked by HS-142-1, an ANF-receptor antagonist, the conclusion is that the effect of the peptide on cell communication was mediated by the activation of specific ANF receptors.

The effect of ANF on the gj of normal hamster seems to indicate that ANF receptors also are present in the ventricles of control hamsters. In normal cardiac muscle, there is evidence that the peptide reduces the action-potential duration and causes inhibition of slow inward Ca channel activity ⁽³¹⁾.

Although in our experiments, the possible role of changes in Ca_i or pH_i on the effect of ANF seems unlikely, considering the concentrations of EGTA and HEPES used in the internal solution, the possibility that ANF changes free Ca_i or pH_i in vivo cannot be discarded, and further studies will be needed to clarify this point.

Concerning the effect of ANF on the failing heart, it is known that the peptide reduces preload and afterload ⁽²⁴⁾. Moreover, infusion of exogenous ANF increases cardiac index in patients with CHF and promotes renal excretion of water and salt ⁽²⁵⁾. Although ANF plays a role as a compensatory mechanism during the initial phase of heart failure, the development of the disease leads to a reduced effect of ANF despite the presence of enhanced levels of the peptide in plasma ⁽²⁶⁾. The attenuation of the effect of ANF seems to be determined by the activation of the renin-angiotensin system, which leads to sodium retention ⁽²⁷⁾.

These results indicate that ANF has direct effect on cardiac muscle cells. Because there is release of ANF from the left ventricle of patients with hypertrophic cardiomyopathy ⁽²⁸⁾, from cultured heart cells under hypoxia ⁽²⁹⁾, and a massive atrial natriuretic peptide release during ischemia reperfusion ⁽³⁰⁾, the question remains whether cardiac arrhythmias seen during myocardial ischemia or in the failing heart are, in part, related to the effect of ANF on gj. Subsequent studies will provide information on this important issue. Moreover, the high plasma levels of ANF found in some patients with supraventricular tachycardia might indicate that the decline in cell coupling caused by the peptide is involved in the generation of this arrhythmia.

Acknowledgment: This work was supported by the American Heart Association and NIH (HL-34148, 2SO6GM, and RR03651).