Evaluation of Atrial Natriuretic Peptide and Brain Natriuretic Peptide in Atrial Granules of Rats with Experimental Congestive Heart Failure

Abstract

The natriuretic peptides are believed to play an important role in the pathophysiology of congestive heart failure (CHF). We utilized a quantitative cytomorphometric method, using double immunocytochemical labeling, to assess the characteristics of atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP) in atrial granules in an experimental model of rats with CHF induced by aortocaval fistula. Rats with CHF were further divided into decompensated (sodium-retaining) and compensated (sodium-excreting) subgroups and compared with a sham-operated control group. A total of 947 granules in myocytes in the right atrium were analyzed, using electron microscopy and a computerized analysis system. Decompensated CHF was associated with alterations in the modal nature of granule content packing, as depicted by moving bin analysis, and in the granule density of both peptides. In control rats, the mean density of gold particles attached to both peptides was 347.0 ± 103.6 and 306.3 ± 89.9 gold particles/μm² for ANP and BNP, respectively. Similar mean density was revealed in the compensated rats (390.6 ± 81.0 and 351.3 ± 62.1 gold particles/μm² for ANP and BNP, respectively). However, in rats with decompensated CHF, a significant decrease in the mean density of gold particles was observed (141.6 ± 67.3 and 158.0 ± 71.2 gold particles/μm² for ANP and BNP, respectively; p < 0.05 compared with compensated rats, for both ANP and BNP). The ANP:BNP ratio did not differ between groups. These findings indicate that the development of decompensated CHF in rats with aortocaval fistula is associated with a marked decrease in the density of both peptides in atrial granules, as well as in alterations in the quantal nature of granule formation. The data further suggest that both peptides, ANP and BNP, may be regulated in the atrium by a common secretory mechanism in CHF.

The natriuretic peptide family includes three peptides, atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP), and C-type natriuretic peptide (CNP), of which the first two, ANP and BNP, are expressed in the heart (Atlas et al. 1992). In the normal physiological state, ANP is produced by and secreted primarily from the cardiac atria in response to stretch forces applied to the atrial wall (Ledsome et al. 1985). In contrast, the main production and secretion of BNP is from the heart ventricles (Mukoyama et al. 1991), primarily the left (Yasue et al. 1994), although this peptide is expressed in small amounts in the atrium. Different studies have shown that both ANP and BNP are stored in the atrium in the same granules and that both are secreted into the circulation under similar conditions (Nakamura et al. 1991; Thibault et al. 1992; Takemura et al. 1998).

Congestive heart failure (CHF) is a syndrome characterized by structural and functional impairment of the myocardium and a severe disturbance of body fluid balance. A variety of neurohormonal systems are activated in CHF (Pettersson et al. 1989; Brodsky et al. 1998). These include the vasoconstrictor and anti-natriuretic systems, such as the sympathetic nervous system and the renin–angiotensin–aldosterone axis, as well as the upregulation of the natriuretic peptides, as delineated by the high synthesis and release of ANP and BNP (Moe et al. 1993). Indeed, many studies have demonstrated that plasma levels of these peptides rise significantly in clinical and experimental CHF (Asano et al. 1999; Friedl et al. 1999).

Most studies conducted on experimental models of CHF have focused on ANP, whereas less attention has been given to the role of BNP in the pathogenesis of CHF. BNP is considered to be a more reliable prognostic marker than ANP in patients with CHF (Friedl et al. 1999), and BNP levels found in the atrium are proportionally related to the severity of the disease (Mukoyama et al. 1991; Langenickel et al. 2000). Doyama et al. (1998) observed that BNP levels rise in response to increased atrial pressure and that atrial expression of BNP mRNA is correlated with that of ANP, assuming that these two genes may share some common regulatory mechanism, of which atrial pressure was suggested to be of major importance. In contrast, Moe et al. (1996) demonstrated that both ANP and BNP are activated in heart failure. However, the mechanisms for the increased plasma levels of the two peptides appear to be different. The latter authors also suggested that increased atrial pressure most likely plays a lesser role in BNP than that supposed for ANP. At present, there is no conclusive evidence to favor either one of these concepts. Most recently, Langenickel et al. (2000) demonstrated that compensated CHF was associated with overexpression of ventricular and atrial ANP but not BNP, whereas in overt heart failure the expression of both ANP and BNP was enhanced. These data suggest that, in contrast to ANP, cardiac BNP expression induced specifically in overt heart failure may be a useful marker for the transition from compensated to overt heart failure.

In recent years, the actions of the natriuretic peptides in CHF have been investigated in our laboratory in an experimental model of CHF in the rat. In this model, CHF is induced by an arteriovenous (A–V) fistula between the abdominal aorta and the inferior vena cava (Hoffman et al. 1988; Winaver et al. 1988, 1995), which overloads the right atrium and results in an elevation of ANP levels in the circulation. In addition, this model is characterized by downregulation of the receptors for ANP in the kidney and by disturbances in the intracellular signaling of the second messenger cGMP (Abassi et al. 1991). In a previous study, the morphometric pattern of the granules containing ANP in atriocytes of rats with CHF was examined. The changes in the morphometric pattern of the granules were compatible with an increased rate of synthesis and secretion of natriuretic peptides in these experimental animals (Avramovitch et al. 1995). The demonstration by immunocytochemical methods that ANP and BNP are stored side by side in the same granules provided the opportunity to use a double-labeling technique as a quantitative estimate of ANP and BNP content in the granules. We further assumed that if a different regulation of the two natriuretic hormones prevails in CHF, it would be expressed in the ratio between ANP and BNP in the formed granules. Therefore, in the present study we used a quantitative cytomorphometric approach, using double immunocytochemical labeling, to examine the regulation of ANP and BNP in CHF compared with controls.

Materials and Methods

Studies were performed in male Wistar rats (250–300 g; Harlan Laboratories, Jerusalem, Israel). The animals were kept in individual metabolic cages and were maintained on normal rat diet containing 0.5% NaCl and tapwater ad libitum. All animal handling was according to the institutional guidelines.

CHF was induced by surgical creation of an A–V fistula between the abdominal aorta and the inferior vena cava, according to the method of Stumpe et al. (1973) as adapted in our laboratory. Briefly, the rats were anesthetized IP by sodium pentobarbital (40 mg/kg) and the abdominal aorta and the inferior vena cava were exposed through a midabdominal incision. Two clamps were placed distal to the origin of the renal arteries on both vessels, and a small incision (1.0–1.2 mm OD) was made in the common wall of both vessels. After the surgical procedure, the rats were allowed to recover and then returned to their metabolic cages for daily monitoring of urine output and sodium excretion. Seven days after the operation, the rats with A–V fistula (n = 8) were further subdivided, on the basis of their daily urine excretion, into rats with compensated CHF [urinary sodium excretion (UNaV) >1200 μEq/24 hr; n = 6)] and sodium-retaining rats with decompensated CHF (UNaV <100 μEq/24 hr; n = 2) (Winaver et al. 1988). A sham operation was performed in four rats of the same strain. These rats served as a control group.

Immunocytochemistry

On the seventh postoperative day, control and CHF rats were anesthetized and the right internal jugular vein and left internal carotid artery were cannulated with polyethylene tubes (PE-50). The aorta was severed proximal to the shunt and a perfusion of 3% glutaraldehyde in 0.1 mM sodium cacodylate, pH 7.4, was introduced simultaneously through the internal jugular vein and the internal carotid artery. The perfusion lasted until the effluent fluid of the abdominal aorta became clear. After perfusion, small specimens of anterior right atrial tissue, including the auricle, were immersed and fixed overnight in a solution of 3% glutaraldehyde in 0.1 mM sodium cacodylate. After a wash in 10% sucrose in 0.1 mM cacodylate buffer, pH 7.4, the samples were postfixed with 1% osmium tetroxide, dehydrated in graded ethanols, cleared in propylene oxide, and embedded in Epon. Three blocks from the right atrium of each rat were selected randomly for sectioning (LKB; Nova Ultratome, Bromma, Sweden).

Sections (80–100 nm) were mounted on nickel grids and were immunostained by a double-labeling immunocytochemical method modified from Horisberger (1981), in two steps. After blocking of unspecific binding sites by 0.1% bovine serum albumin (BSA) in PBS for 15 min, a polyclonal antibody against human α-ANP (1:500) (Phoenix Pharmaceuticals; Mountain View, CA) that has no crossreactivity with BNP as determined in an RIA system (personal communication) was administered to the grid and incubated overnight in a humidified chamber at 4C. This was followed by 10-nm colloidal gold-tagged goat anti-rabbit second antibody (1:10) (Zymed Laboratories; San Francisco, CA) and a thorough wash in 0.1% BSA in PBS. Next, a polyclonal antibody against rat BNP (1:500) (Peninsula Laboratories; Belmont, CA), which has no crossreactivity with ANP as determined in an RIA system (personal communication), was administered to the other side of the grid, following the same procedure, with 15-nm gold-tagged goat anti-rabbit second antibody (1:10) (Zymed).

The double-labeled sections were then scanned without further staining in a JEOL 100 SX transmission electron microscope (Tokyo, Japan), and the most technically adequate and clearly stained sections were selected. Secretory granules were photographed randomly at a magnification of ×30,000 and were printed at a final magnification of ×60,000 from at least five different cells from each block (i.e., at least 50 prints/treatment group). Both types of gold particles in each individual granular profile (10- and 15-nm) were counted separately, directly from the final printed micrographs. Using an HP-9111A graphic tablet connected to a Power Macintosh 7100/60 AV, the profile area of each granule was measured by a program developed by Hammel et al. (1983, 1988, 1989), then analyzed using Microsoft Excel 98 (Mac version). The cross-sectional area (A_i) of each individual granule was converted into the equivalent volume using the simple transformation v = (4π/3)(A_i/π)^3/2. The resulting volume equivalents were plotted as a histogram. The multimodal histogram was analyzed by the moving-bin technique to reveal true peaks, as explained elsewhere in detail (Hammel et al. 1983). The periodicity of the distribution was extrapolated as the mean of the intermodal spaces. The ANP (or BNP) density was calculated from the ratio of number of gold beads of each antigen (N_i) on the respective granular profile (i.e., N_i/A_i). In addition, all granules were sorted by size, and the corresponding average density for each group of 25 granules was calculated as:

$\begin{array}{lll}\text{i}=\text{n}÷24\hfill & \hfill & \text{i}=\text{n}+24\hfill \\ \sum {\text{N}}_{\text{i}}\hfill & /\hfill & \sum {\text{A}}_{\text{i}}\hfill \\ \text{i}=\text{n}\hfill & \hfill & \text{i}=\text{n}\hfill \end{array}$

The data were correlated with the respective mean granule equivalent volume [A = ΣA_i/25, equivalent volume = (4π/3)(A/π)^3/2].

Statistical Analysis

Data are presented as mean ± SE. Statistical significance was evaluated by the nonparametric Kolmogorov–Smirnov test (Sokal and Rohlf 1969). p < 0.05 was considered statistically significant.

Results

Representative electron micrographs of granules in the right atrium in each of the three groups (compensated and decompensated rats with CHF and controls), double-labeled by the two different gold particle sizes (10 and 15 nm), are shown in Figure 1. A total of 947 granules were counted, of which 226 were from the control group, 370 from the decompensated CHF subgroup, and 351 from the compensated CHF subgroup. As shown, most of the labeling was concentrated within the limits of the granules, with minimal background activity. Moreover, the density of both sizes of gold particles was higher in the granules of normal and compensated rats (Figure 1A and 1B) compared with granules of decompensated rats (Figure 1C). In control rats, the mean densities of gold particles were 347.0 ± 103.6 and 306.3 ± 89.9 gold particles/μm² for ANP and BNP, respectively. Similar mean densities were revealed in the compensated rats (390.6 ± 81.0 and 351.3 ± 62.1 gold particles/μm² for ANP and BNP, respectively). However, in rats with decompensated CHF, a significant decrease in the mean density of gold particles was observed (141.6 ± 67.3 and 158.0 ± 71.2 gold particles/μm² for ANP and BNP, respectively; p < 0.05 for both ANP and BNP vs controls and compensated rats). When non-immune rabbit serum replaced both primary antibodies, no gold particles of any size were observed in the atrial granules (Figure 1D).

To evaluate the characteristics of granule formation and release, we applied the technique of moving-bin analysis (Hammel et al. 1983, 1988, 1989). In moving-bin analysis (Figure 2), in which all granules (from control rats and rats with compensated and decompensated CHF) were analyzed together, the granule equivalent volume demonstrated periodic multimodal distribution, confirming the quantal nature of the granules (Hammel et al. 1988, 1989). The unit granule volume in the present study was calculated to be 0.0048 μm³ for all three groups. Peak locations were at the same values, and therefore all data were aggregated. Data on the mean density of the gold particles attached to the ANP and BNP in the three groups are summarized in Figures 3 and 4. Figure 3 depicts the scattergram analysis of the density of BNP particles plotted in relation to the ANP density in control animals and in rats with CHF. Each point represents mean data from a cluster of granules in a single cell. The data revealed that there was a common trend for both ANP and BNP to increase in density, i.e., cells with high abundance of one peptide also contained high amounts of the other peptide, and vice versa. This might imply that the generation of both peptides in atrial granules shares a common regulatory mechanism both in control animals and in rats with CHF.

Figure 1

Atrial granules of atriocytes double-immunostained with antibodies to ANP (tagged with 10-nm gold particles) and to BNP (tagged with 15-nm gold particles) in control rats (A), compensated rats (B), and decompensated rats (C), atrial granules immunostained with non-immune rabbit serum replacing both primary antibodies (D). Original magnification ×60,000.

Figure 2

Moving bin analysis of all granules (control plus experimental), showing the quantal nature of the granular volume.

An important observation in this study was the significant difference between the compensated group and the decompensated group in the density of gold particles attached to both ANP and BNP (Figure 4B and 4C). On the other hand, there was no significant difference between the control group and the compensated rats (Figure 4A and 4B). However, despite a significant reduction in the absolute values of ANP and BNP in rats with decompensated CHF, the ANP:BNP ratio was not significantly different in this subgroup. Of note is the finding that when the mean densities of gold particles were plotted relative to the mean granule equivalent volumes in the three groups (Figure 4), a similar ratio was obtained irrespective of the granule size.

There was a significant difference between the compensated and control rats compared with the decompensated subgroup. Because peaks were not evident in all three groups, we took data from Figure 2 and labeled the graphs of Figure 4 at the location of the quanta of the calculated granules. In Figures 4A and 4B, in both cases, the modes of the graphs are approximately near the indicated arrows, whereas in the de-compensated subgroup the quantal nature is significantly less obvious. However, it should be noted that because peaks in this graph were not evident in all three cases, the modes might have been due to noise (Hammel et al. 1983).

Discussion

Our study provides important insight into the regulation of both ANP and BNP in the atria of rats with A–V fistula, an experimental model of CHF. Using a quantitative morphometric immunocytochemical method, we confirmed previous observations that both ANP and BNP are stored in the same granules in cardiac atriocytes. Moreover, rats with severe decompensated CHF display a significant decrease in the density of both peptides in the atrial granules, irrespective of granule size, as well as alterations in the quantal nature of atrial granule formation. The results further suggest that the generation of ANP and BNP in the atria shares a common regulatory mechanism, both under normal conditions and in animals with CHF.

The finding that atrial granules contain both ANP and BNP is in agreement with previous reports in the literature. Hasegawa et al. (1991) were the first to report such a co-localization in porcine atria. In their study, using a similar two-face immunogold staining method, they identified two types of granules: a monohormonal granule containing ANP alone (Type 1), and granules containing both ANP and BNP (Type 2). In the present study we did not observe Type 1 granules. Similarly, studies in human cardiac myocytes (Nakamura et al. 1991; Takemura et al. 1998) and in rat atrial cardiocytes (Thibault et al. 1992) co-localized ANP and BNP in the same atrial secretory granules. On the basis of those findings, it has been suggested that the secretion of BNP is regulated by a mechanism similar to that of ANP (Nakamura et al. 1991). Our findings further extend these observations. By combining a morphometric analysis of granule size with a double immunocytochemical labeling of both peptides, we were able to calculate the density of the labeling for each peptide in atrial granules and to follow this parameter during the evolution of heart failure in the rat. Using this approach, we demonstrated that the density of both ANP and BNP (measured as number of gold particles calculated per granule profile) was significantly decreased, by approximately 50–60%, in rats with decompensated CHF compared with control rats, but not in rats with compensated CHF.

Figure 3

Scattergram analysis of the double-labeled granules, demonstrating the relationship between the densities of ANP and BNP, in control rats and in all the experimental animals. Each point represents a cluster of granules in a single cell. Y-bar refers to SEM of BNP density; X-bar refers to ANP density.

Figure 4

Mean density of gold particles attached to ANP and BNP as related to granule profile equivalent volume in control rats (A), compensated rats (B), and decompensated rats (C).

The densities of ANP and BNP cannot be used as a quantitative estimate of the absolute content of ANP and BNP in the granule, because the antibodies may differ in their affinities for the peptides. However, the BNP:ANP density ratio can provide valuable information about the alterations in the granule content of each hormone during the evolution of CHF from a compensated to a decompensated state, thus hinting at either common or separate regulation of the two hormones in this disease state. Indeed, contradictory data are available on the regulation of ANP and BNP in the atria during the development of heart failure. Our findings support previous studies by Pemberton et al. (1997) and Doyama et al. (1998). The former demonstrated that the ANP:BNP mRNA ratio in the atrium was not changed after induction of CHF by rapid ventricular pacing in sheep, compared with control animals. Similar observations were reported by Doyama et al. (1998), who examined the expression of these peptides by Northern blotting analysis in the right atria of patients undergoing cardiac surgery. Their findings suggest that ANP and BNP share a common regulatory mechanism and that their synthesis may be commonly associated with atrial stretch. Indeed, wall stretch is known to be an important factor in the induction of ANP and BNP genes in cardiac tissue in CHF (Yasue et al. 1994; Magga et al. 1997, Lange-nickel et al. 2000).

However, other investigators have shown differences in the ANP:BNP ratio in the atria. Moe et al. (1996) reported that ventricular pacing in dogs caused depletion of ANP in the atrial tissue compared with controls, whereas the atrial content of BNP remained unchanged in this tissue, suggesting a different mode of response of ANP and BNP to hemodynamic changes produced by cardiac pacing. The reason for this discrepancy is not clear but may be related to the mode of induction of cardiac failure as well as to species differences. It is also possible that the regulation of the natriuretic peptide synthesis may differ in atria vs the ventricles in CHF. Thus, Pemberton et al. (1997) showed that, in contrast to the atria, where the ANP:BNP ratio was preserved, this ratio was increased in the ventricles in response to CHF, suggesting different mechanisms in the cardiac atria and ventricles. This notion is also supported indirectly by the measurements of plasma levels of ANP and BNP, which reflect the total production of the hormones by the heart. Such differential regulation is also in line with other studies in clinical CHF that demonstrated a different pattern of release for both peptides during the development of heart failure (Mukoyama et al. 1991; Wei et al. 1993; Yoshimura et al. 1993). Finally, Thibault et al. (1992), using a similar rat model with aortocaval fistula, demonstrated a similar BNP:ANP ratio in atria of CHF and sham-operated rats. In contrast, there was a marked increase in this ratio in the ventricles of CHF rats compared with sham-operated animals, further supporting the concept of independent regulatory mechanisms of the peptides in atria and ventricles.

An additional interesting feature in the present study was the loss of the modes with the deterioration of CHF into a decompensated state (Figure 4). The secretory granule equivalent volume in the decompensated subgroup was multimodal and periodic. However, the scattergram analysis correlating the immunohistochemical data with granule size was not multimodal as in the other two groups. In a previous study from our laboratory, using morphometric analysis of the granules in atriocytes of rats with decompensated CHF, we demonstrated changes compatible with accelerated generation and release of the granules in this severe form of CHF (Avramovitch et al. 1995). We have then further speculated that, because of the high ANP turnover, the granules of CHF rats spent a shorter time in the myocardium and were not provided with the optimal duration required for normal fusion and enlargement (Avramovitch et al. 1995). The disappearance of the histochemical multimodal nature of the granules is consistent with such a concept. Moreover, this alteration was noted only in the decompensated subgroup, which was characterized by a more severe form of CHF, a higher degree of neurohormonal activation, and avid sodium retention by the kidney (Winaver et al. 1988). Taken together with the observation that ANP and BNP densities were significantly decreased in rats with decompensated CHF, it is possible that alterations in the normal maturation of the granules may occur in this subgroup. An alternative explanation for the decrease in the levels of both peptides in the atriocytes of the decompensated rats might be an accelerated fusion of atrial granules with the plasma membrane and premature release of the hormonal content into the circulation (Avramovitch et al. 1995; Pemberton et al. 1997). Both possibilities may result in atrial depletion of the natriuretic peptides relative to the degree of heart failure, which may further contribute to the pathogenesis of this severe form of CHF. Additional studies are required to elucidate the pathophysiological significance of these abnormalities and their importance in heart failure.

In summary, in the present study we used a double-labeling cytomorphometric approach to follow the alterations in the ANP- and BNP-containing granules in atria of rats with experimental CHF. The data demonstrate that the progression of CHF from a compensated to a decompensated state is associated with marked alterations in the density of both peptides in the atrial granules and in the quantal nature of granule formation. The findings further suggest that both ANP and BNP in atrial granules share a common regulatory mechanism in heart failure.