INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

CARDIAC FIBROBLASTS are abundant cellular constituents of the left ventricle and participate in cardiac remodeling and collagen deposition after acute injury, such as after coronary artery occlusion and during the development of left ventricular hypertrophy due to pressure overload. However, regulation of collagen biosynthesis by the cardiac fibroblast has not been fully elucidated. Based on studies from this and other laboratories, a central role of angiotensin II (ANG II) in promoting collagen biosynthesis by cardiac fibroblasts has been advanced. Specifically, the presence of a high-affinity ANG II receptor was demonstrated on cultured rat left ventricular fibroblasts (33) but not on isolated rat cardiac myocytes (16), and ANG II was shown to enhance mRNA levels for collagen (types I and III) and fibronectin (33). These findings, together with reports from other laboratories (7, 34), have led to a widely held hypothesis that ANG II directly targets the cardiac fibroblast to alter collagen synthesis and deposition in ventricular remodeling. In support of this hypothesis, numerous in vivo studies in rats have shown that ANG II AT₁ receptor blockade mimics the ability of angiotensin-converting enzyme (ACE) inhibitors to limit collagen deposition in the myocardium after injury (17, 26, 31). The caveat is that the majority of these studies have utilized the rat myocardium and cultured rat cardiac cells as the model systems for studies of the regulation of left ventricular collagen and extracellular matrix deposition.

The present study was undertaken to examine the potential role of ANG II on cardiac fibroblasts across species and arose from an analysis of the organization and temporal deposition of left ventricular collagen fibers in a rabbit infarct model (10). During the course of this investigation, a preliminary experiment failed to reveal measurable levels of ANG II receptors by direct radioligand binding on cultured rabbit cardiac fibroblasts. A systematic study was undertaken using both direct radioligand binding as well as functional responses to test for ANG II effects on cultured cells and in situ. The importance of this question relates directly both to the putative role of ANG II in the regulation of cardiac collagen biosynthesis and deposition during cardiac remodeling across species and, by extension, to the mechanism of action of ACE inhibitors in limiting collagen deposition after myocardial infarction and in chronic hypertension.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Surgery and euthanasia procedures. All animal experiments conformed to American Association for the Accreditation of Laboratory Animal Care guidelines, and experimental protocols were approved by the University of California San Diego (UCSD) Animal Subjects Committee. For coronary artery ligation in New Zealand White rabbits, a left thoracotomy was performed to expose the anterior surface of the left ventricle. A descending branch of the circumflex artery midway between the apex and base was ligated with 5-0 silk, and the chest was closed in layers. Six rabbits underwent coronary artery ligation with a 100% survival rate. Coronary artery ligation was performed in the rat, as described by Pfeffer et al. (24). The left ventricle was exposed, and a ligature was placed apical to the pulmonary artery outflow to ensure ligation of at least one of the coronary arteries. Six rats underwent coronary artery ligation with a 33% survival rate. For both species, unoperated animals were used to obtain control hearts. This myocardial infarction model was previously shown to produce infarctions of similar size in the rat (40%) and rabbit (48%; see Ref. 10). To isolate the hearts, after anesthesia induction, a sternotomy was performed to expose the heart and permit clearance of adhesions, after which a euthanizing dose was given (100 mg/kg pentobarbital sodium), and the heart was excised. After excision, hearts were rinsed in cold saline, sliced into two sections along the short axis of the left ventricle, snap-frozen in isopentane on dry ice, and stored at 70°C until sectioned. The most basal ring was cut from above the ligature and contained noninfarcted myocardium, whereas the apical ring was cut from below the ligature and contained infarcted tissue. The rings were cryostat sectioned (10 µm) along the short axis for histology and receptor autoradiography.

Cell culture. Normal, noninfarcted hearts were used for the fibroblast cultures with one rabbit or three rat hearts per cell preparation. Atrial tissue was removed, left ventricles were minced, and cells were dispersed with collagenase (Boehringer Mannheim, Indianapolis, IN) and pancreatin (GIBCO-BRL, Gaithersburg, MD). After five separate, sequential digestions, the cell suspensions were combined, centrifuged, and resuspended in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS). Cardiac fibroblast cultures were obtained by differential plating (33). Cells were plated onto 150-mm cell culture dishes for 45 min at 37°C to permit attachment of cells, after which the media were changed to remove unattached cells. Cultured cells were characterized by immunocytochemistry and were immunopositive for -smooth muscle actin, vimentin, and fibronectin but not for desmin or -sarcomeric actin. Smooth muscle cells, used in comparison studies, were immunoreactive for -smooth muscle actin and vimentin but not for -sarcomeric actin with a subset of smooth muscle cells immunoreactive for desmin. Additionally, endothelial cell contamination was checked using uptake of di-I-acetylated low-density lipoprotein (Biomedical Technologies, Stoughton, MA). Bovine aortic endothelial cells (kindly provided by Dr. David Loskutoff, Scripps Research Clinic) were used as positive controls. No endothelial contamination was found. Human cardiac fibroblasts (kindly provided by Dr. Francisco Villarreal) were obtained by explant from human ventricular myocardium excised at the time of cardiac transplant of patients 1-2 yr of age. Ventricular tissue was minced and seeded into 35-mm dishes containing DMEM with penicillin, streptomycin, Fungizone, and 20% FBS. Glass coverslips, adherent at one edge to the dish, secured the tissue to the bottom of the well. Fibroblasts grew out of the explants by 2-4 days of culture and reached confluency after ~2 wk. The cells used here were derived from tissue of six separate individuals. Human cardiac fibroblasts grew exclusively in monolayers rather than multilayered "hills and valleys" characteristic of smooth muscle cells and were immunonegative for the muscle cell structural protein desmin and immunopositive for the cytoplasmic intermediate filament protein vimentin. Cells from all three species were grown to confluency and trypsinized, and stock aliquots were frozen for future use or passaged (1:3). Cells were used between passages 1 and 3.

Preparation of fibroblast membranes. Cell membranes were prepared from confluent cardiac fibroblasts (150-mm cell culture plates) and washed with ice-cold phosphate-buffered saline (PBS) by resuspension and centrifugation. The resulting membrane pellet was resuspended in 50 mM tris(hydroxymethyl)aminomethane (Tris; pH 7.4 at 25°C) containing 5 mM EDTA, 1.0 mg/ml bacitracin, 1.0 mg/ml trypsin inhibitor, and 0.1 mM phenylmethylsulfonyl fluoride, and the membranes were dispersed by brief polytron homogenization in ice. The homogenate was centrifuged two times, and supernatants were pooled and recentrifuged at 100,000 g for 45 min at 2°C. The resulting pellet was resuspended and immediately used in ligand binding assays.

Isolation of ventricular myocytes. Rabbit cardiac myocytes (kindly provided by Dr. Wilbur Lew) were isolated as described previously (14). The excised heart was rinsed in Ca²⁺-free Tyrode solution [136 mM NaCl, 5.4 mM KCl, 1.0 mM MgCl₂, 0.33 mM NaH₂PO₄, 10 mM glucose, and 10 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES)] to remove blood, followed by Langendorff perfusion with Ca²⁺-Tyrode solution (50 µM Ca²⁺) containing 1.0 mg/ml collagenase B and 0.5 mg/ml Dispase (Sigma Chemical, St. Louis, MO). The left ventricle was dissected free and gently dispersed, and the resulting cell suspension was filtered through gauze to remove connective and fatty tissue. The cell suspension was rinsed successively with minimal essential medium containing progressively increasing levels of Ca²⁺ concentration up to 2 mM. To optimize recovery of viable cells and enrich for myocytes, cells were layered onto and centrifuged at 250 g for 4 min through Histopaque 1077 (Sigma Chemical). Cells collected from the interface were 75-90% rod-shaped myocytes.

Radioligand binding. Subconfluent cardiac fibroblasts were incubated with 0.1 nM ¹²⁵I-labeled ANG II or ¹²⁵I-[Sar¹,Ile⁸]ANG II for competition experiments and 0.2 nM ¹²⁵I-ANG II or ¹²⁵I-[Sar¹,Ile⁸]ANG II for equilibrium binding experiments. All radioisotopes were obtained either from New England Nuclear (Boston, MA) or Dr. Robert Speth, Washington State University (Pullman, WA). Competition studies employed either unlabeled ANG II (Bachem, Torrance, CA), [Sar¹,Ile⁸]ANG II, the AT_1a-specific antagonist losartan (kindly supplied by DuPont-Merck, Wilmington, DE), or the AT₂-specific antagonist PD-123177 (kindly supplied by Parke-Davis, Ann Arbor, MI). All binding experiments with cultured cells utilized a binding buffer (TBS) composed of 50 mM Tris, 100 mM NaCl, pH 7.2 at 25°C, with 2.5 g/l bovine serum albumin (BSA; Serva, Heidelberg, Germany). Cultured cells were incubated in TBS with ligands for 120 min at 12-14°C, a temperature designed to minimize internalization of ligand. Unbound ligand was removed with three, ice-cold PBS washes. Cells were lysed with 2 N NaOH, and, after the culture wells were rinsed with distilled water, bound radioligand was determined. To correct for differences in cell number across experiments, during each radioligand binding experiment a minimum of three wells was incubated with binding buffer only, after which the cells were trypsinized, resuspended in DMEM, and counted. For equilibrium binding experiments, the molar concentration of specifically bound ¹²⁵I-ANG II was calculated as described by Mukku (22). The demonstration of specific ANG II binding to subconfluent rabbit aorta vascular smooth muscle cells (kindly supplied by Dr. Daniel Steinberg, UCSD) employed 0.1 nM ¹²⁵I-ANG II (New England Nuclear) in the presence or absence of 1 µM unlabeled ANG II (Bachem). Radioligand binding to suspension cultures of rabbit cardiac myocytes employed 0.1 nM ¹²⁵I-[Sar¹,Ile⁸]ANG II and unlabeled [Sar¹,Ile⁸]ANG II. After an incubation for 120 min at 12-14°C, ice-cold PBS was added, and the cells were immediately centrifuged for 60 s. After three centrifugation washes with ice-cold PBS, bound radioligand was determined along with cell number. Radioligand binding to isolated membranes (70-100 µg of membrane protein) was conducted at 25°C using 0.1 nM ¹²⁵I-ANG II in the absence or presence of 1 µM unlabeled ANG II. All receptor binding studies were analyzed by nonlinear regression analysis using Prism (GraphPad Software, San Diego, CA).

Receptor autoradiography. Cryostat-sectioned tissue (10 µm) mounted on coated glass slides (Fisherbrand Plus; Fisher Scientific, Pittsburgh, PA) was preincubated in modified binding buffer (TBS containing 2 g/l BSA, 0.4 mM bacitracin, and 5 mM Na₂EDTA, pH 7.4) for 15 min at 25°C followed by a 60-min incubation with 0.1 nM ¹²⁵I-[Sar¹,Ile⁸]ANG II in the same binding buffer. Nonspecific binding was measured in the presence of 1 µM unlabeled [Sar¹,Ile⁸]ANG II. After the incubation, unbound ligand was removed by immersion (3×) of the slides in PBS, after which the tissue sections were fixed in 4% glutaraldehyde for 120 min at 25°C. Exposed reactive aldehyde groups were blocked by overnight incubation in 4 mg/ml BSA in PBS at 4°C. The slides were rinsed, dried, and dipped in liquid emulsion (Kodak NTB-2) for high-resolution autoradiography. After exposure at 4°C for 4 wk, slides were developed with Kodak D19, fixed, and counterstained using Gill's hematoxylin. Silver grains were quantified by image analysis (MCID Image; Imaging Research, Ontario, Canada). Silver grains corresponding to total and nonspecific ANG II binding were counted in 15 randomly selected sites, each of constant area (14,000 µm²), in both the noninfarcted and infarcted areas of each heart. The average number of grains per unit area representing nonspecific binding was subtracted from the average number of grains per unit area representing total binding to obtain specific binding for both the infarcted and noninfarcted areas of each heart. Grain counts corresponding to specific binding were used for comparisons and statistical analyses.

Histology. Cryostat-sectioned tissues mounted on coated slides were fixed in 10% Formalin for 15 min, rinsed with distilled water, and stained with 0.01% sirius red in saturated picric acid for 60 min. After staining, the sections were dehydrated or counterstained with Gill's hematoxylin before dehydration and were covered with coverslips.

Fura 2 ratio imaging. Fura 2 ratio imaging was performed as previously described (12, 29). Cells were excited at 340 and 380 nm, and emissions were monitored at 510 nm after loading with 2 µM fura 2-acetoxymethyl ester for 60-120 min (Molecular Probes). Pluronic F-127 (10% wt/vol in dimethyl sulfoxide) was added to 1 mM stock fura 2 with a 1:4 dilution before loading. The resulting fluorescence signals and the ratio of emission at both excitatory wavelengths were stored on-line every 5 s for analysis and plotting. Within the microscopic field, 6-12 areas, each corresponding to an individual cell, were defined, and separate fluorescence data were simultaneously acquired from each area, permitting real-time data acquisition from multiple cells. Standard shading images (zero Ca²⁺) were saved before each experiment using a 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid-buffered solution with fura 2, and background shading images of cell-free fields were saved to control for subtle variations in the optical path between runs (29). Fura 2 emission ratios corresponding to high and low Ca²⁺ concentrations were acquired to standardize the ratio data (30). The standard Ca²⁺-containing bath solution contained (in mM) 135 NaCl, 5 KCl, 2 MgCl₂, 3 CaCl₂ (2 mM free Ca²⁺), 1 ethylene glycol-bis(-aminoethyl ether)-N,N,N',N'-tetraacetic acid, 15 HEPES (pH 7.35 with NaOH), and 10 glucose. The Ca²⁺-free bathing solution was identical, except for equiosmolar substitution of glucose for CaCl₂. ANG II and bradykinin were added to the recording chamber to give the indicated final concentrations.

Arachidonic acid metabolite release. Confluent cardiac fibroblasts were labeled by incubation in DMEM containing 0.5% FBS and 3 µCi/plate [³H]arachidonic acid (New England Nuclear) for 18 h. After removal of the labeling medium, the cells were rinsed two times in Ca²⁺/Mg²⁺-free PBS (pH 7.4), equilibrated in HEPES-buffered medium (20 mM HEPES) at 37°C for 30 min, after which 10% of the volume was removed and counted as an estimate of basal metabolite release under unstimulated conditions. After 30 min of stimulation with 1 µM ANG II, released radioactivity in a second 1/10th volume of medium was counted to measure total metabolites released. Cells from parallel treated plates were scraped, lysed with 0.1 N NaOH, and assessed by Lowry assay to normalize for protein content.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Cultured fibroblasts and myocytes. Cultured rat cardiac fibroblasts exhibited specific and saturable binding of the radioiodinated agonist ¹²⁵I-ANG II, confirming prior observations (33). Equilibrium binding identified a single class of binding sites with a maximal binding of 43,000 ± 15,000 receptors/cell and with a dissociation constant (K_d) of 0.92 ± 0.34 nM (n = 4 in triplicate). ¹²⁵I-ANG II binding to rat fibroblasts was competed fully by ANG II [half-maximal inhibitory concentration (IC₅₀) = 0.89 ± 0.5 nM] and the AT₁- specific antagonist losartan (IC₅₀ = 7.98 ± 3.5 nM) but not by the AT₂-specific antagonist PD-123319 (n = 3 in triplicate).

In contrast to cultures of rat cardiac fibroblasts, cultured rabbit cardiac fibroblasts did not bind either ¹²⁵I-ANG II or the aminopeptidase-stabilized antagonist ¹²⁵I-[Sar¹,Ile⁸]ANG II (Fig. 1). Increasing competing ligand [Sar¹, Ile⁸]ANG II from 10¹² to 10⁶ M displaced only 0.4 fmol of 0.2 nM ¹²⁵I-[Sar¹,Ile⁸]ANG II/10⁶ rabbit cardiac fibroblasts, which computes to fewer than 300 receptors/cell (assuming a K_d of 1.0 nM). To examine whether low radioiodinated ANG II binding observed for rabbit cardiac fibroblasts was attributable to receptor masking or internalized receptors, binding studies were conducted with partially purified membranes from cultured cells. However, ¹²⁵I-ANG II binding was not detected in the membrane preparations using 70-100 µg membrane protein/sample. In contrast to the fibroblast cultures, competitive binding assays identified significant ¹²⁵I-ANG II binding to purified rabbit left ventricular myocytes, yielding 25,400 ± 8,000 ¹²⁵I-ANG II sites/myocyte with an IC₅₀ of 0.69 ± 0.10 nM (Fig. 2). Additionally, cultured rabbit aortic vascular smooth muscle cells exhibited specific ¹²⁵I-ANG II binding (data not shown). Various experimental manipulations, such as variation in incubation temperature or time, ionic concentration, and addition of various protease inhibitors, all failed to detect significant specific angiotensin binding sites by cultured rabbit cardiac fibroblasts.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Representative example of a competitive binding experiment of 125I-[Sar1,Ile8]ANG II from rat () and rabbit () cardiac fibroblasts using unlabeled [Sar1,Ile8]ANG II. Half-maximal inhibitory concentration (IC50) for the rat cells is 0.68 nM (0.60-0.77 nM). Each data point represents the mean of triplicate measurements. cpm, Counts/min.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Competitive binding of 125I-[Sar1,Ile8]ANG II with unlabeled ligand revealed that rabbit cardiac myocytes have 25,400 ± 8,000 ANG II receptors/cell with an IC50 of 0.69 ± 0.10 nM. Each data point represents the mean ± SE of triplicate measurements from 3 separate experiments.

The availability of cultured human cardiac fibroblasts was limited, since growth necessary to provide large numbers of cells would have resulted in high passage numbers. Binding studies were therefore restricted to time course analyses designed to ensure attainment of equilibrium binding and to measure specific radioligand bound. Only 0.5 fmol ¹²⁵I-ANG II bound per 10⁶ human cardiac fibroblasts (equilibrated with 0.1 nM radioligand), which was less than 12% of the binding predicted for rat cardiac fibroblasts (4.1 fmol/10⁶ cells).

Functional responses to angiotensin. Because direct evidence for specific ANG II receptor expression by intact rabbit or human cardiac fibroblasts or on membranes from rabbit cells was not obtained by radioligand binding, studies were undertaken to determine if a low density of receptors could be detected through functional cellular responses. The AT₁-receptor subtype is known to elevate intracellular Ca²⁺ concentration ([Ca²⁺]_i; see Refs. 4 and 7); therefore, intracellular fluorescence fura 2 studies were undertaken. Similar to the kinetics observed for other cell types containing angiotensin receptors, addition of ANG II to cultured rat cardiac fibroblasts resulted in a rapid release of [Ca²⁺]_i (Fig. 3). Measurements at the single cell level indicated that, upon addition of 1 µM ANG II, 87% (69/79) of the rat cardiac fibroblasts responded with increased [Ca²⁺]_i, whereas only 7% (5/69) of rabbit cells and 2% (1/60) of human cells responded to this saturating dose. Responders were defined as cells with a >100 nM rise in [Ca²⁺]_i after treatment. It should be noted that the rabbit cells exhibiting positive responses to angiotensin were all observed in a single experiment (1 of 14 experiments) from a single cell preparation. In contrast, as shown (Fig. 3), rat, rabbit, and human cells all responded strongly to 1 µM bradykinin, indicating that the rabbit and human cells were capable of intracellular Ca²⁺ mobilization but were unresponsive to ANG II.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   ANG II-stimulated intracellular Ca2+ signals in representative cardiac fibroblasts from rats (A), rabbits (B), and humans (C). All cells were bathed in solution containing 2 mM free Ca2+ unless otherwise indicated. Fura 2 ratio imaging was used to determine intracellular Ca2+ concentration ([Ca2+]i). Addition of ANG II, bradykinin (BK), or bath exchanges occurred at the indicated times and with the indicated final concentrations. Note that the concentration and time scales differ in A-C. A: both ANG II and BK stimulate intracellular Ca2+ signals in 4 individual rat cardiac fibroblasts. B: ANG II does not stimulate intracellular Ca2+ signals in 5 individual rabbit cardiac fibroblasts; however, BK produces significant [Ca2+]i in all of these cells. C: ANG II does not stimulate intracellular Ca2+ signals in 3 individual cardiac fibroblasts isolated from a human heart; however, BK produces significant [Ca2+]i in all of these cells.

Because it is well established that angiotensin elicits a cellular release of arachidonic acid (18), this was also employed as a functional response to detect ANG II receptors. Cultured rat and rabbit cardiac fibroblasts, prelabeled with [³H]arachidonic acid, were stimulated with 1 µM ANG II, since the formation of arachidonic acid metabolites (enzymatically or nonenzymatically) would provide a measure of phospholipase activation and release of free fatty acid. As predicted, ANG II stimulated arachidonic acid release from rat cardiac fibroblasts but failed to do so from rabbit fibroblasts (Fig. 4).

View larger version (8K): [in this window] [in a new window]   Fig. 4.   Arachidonic acid metabolites released from rat [Wistar-Kyoto (WKY)] and rabbit cardiac fibroblasts after 30 min stimulation with vehicle and 1 µM ANG II. Released counts were corrected for protein content in each plate of cells by Lowry assay. * P < 0.05 vs. vehicle by unpaired 2-tailed t-test.

ANG II receptor expression in experimental infarction models. To examine if rabbit cardiac fibroblast cultures failed to express an angiotensin binding site due to the cell culture process, studies were conducted to examine in situ expression of receptors in left ventricular tissue sections after coronary artery occlusion. One week after coronary occlusion in both rat and rabbit hearts, the amount of collagen within the area of infarction, as measured by picrosirius red staining, was increased dramatically relative to the noninfarcted tissue in both species (Fig. 5, A and C). Additionally, there was evidence of greater cellular infiltrate within the area of infarction relative to the noninfarcted tissue (Fig. 5, B and D). In situ receptor autoradiography was performed to determine whether the local changes in collagen content were paralleled by local changes in ANG II receptor density. At 6 days postocclusion in the rat, specific ANG II receptor density was increased 34.8 ± 13.4-fold in and around the area of infarction relative to the noninfarcted tissue (Fig. 6). In contrast, at 6 days postocclusion, rabbit hearts (Fig. 7) exhibited only a 3.2 ± 1.6-fold increase in ¹²⁵I-ANG II binding within the infarcted tissue relative to the noninfarcted areas (Fig. 8). More pronounced specific ¹²⁵I-ANG II binding was evident around blood vessels in the rabbit heart (Fig. 9).

View larger version (136K): [in this window] [in a new window]   Fig. 5.   Infarcted rat (A and B) and rabbit (C and D) left ventricle stained with 0.01% sirius red in saturated picric acid (A and C) and the same hearts counterstained with hematoxylin (B and D). Within the area of infarction (I), collagen content (denoted by darker-stained areas) and cell density were increased by 6 days after coronary artery ligation. Scale bar = 100 µm.

View larger version (125K): [in this window] [in a new window]   View larger version (147K): [in this window] [in a new window]   View larger version (131K): [in this window] [in a new window]   Fig. 6.   Border between infarcted and noninfarcted myocardium within the left ventricular free wall of a rat heart 6 days postmyocardial infarction. A: brightfield view of the edge of the infarcted area showing an increase in cell density at the infarction border. Total (B) and nonspecific (C) 125I-[Sar1,Ile8]ANG II binding within the same area showing enhanced ANG II binding in the infarction zone (I) relative to the noninfarcted tissue (N). Scale bar = 250 µm.

View larger version (121K): [in this window] [in a new window]   View larger version (123K): [in this window] [in a new window]   View larger version (113K): [in this window] [in a new window]   Fig. 7.   Border between infarcted and noninfarcted myocardium within the left ventricular free wall of a rabbit heart 6 days after myocardial infarction. A: brightfield view at the border between infarcted and noninfarcted tissue. Total (B) and nonspecific (C) 125I-[Sar1,Ile8]ANG II binding within the same area showing homogeneous ANG II binding within the left ventricular infarcted and noninfarcted areas. Scale bar = 250 µm.

View larger version (14K): [in this window] [in a new window]   Fig. 8.   ANG II binding within noninfarcted and infarcted rat and rabbit hearts as quantified by autoradiographic development of silver grains. Infarcted rat tissue shows a 34.8 ± 13.4-fold increase in number of silver grains (per 14,000 µm2 tissue area) relative to noninfarcted area, whereas rabbit hearts show a 3.2 ± 1.6-fold increase in silver grains in the infarcted tissue relative to the noninfarcted tissue.

View larger version (119K): [in this window] [in a new window]   View larger version (122K): [in this window] [in a new window]   View larger version (124K): [in this window] [in a new window]   Fig. 9.   Brightfield view (A) and corresponding total (B) and nonspecific (C) 125I-[Sar1,Ile8]ANG II binding showing specific ANG II binding around a blood vessel within the noninfarcted septum of the rabbit left ventricle shown in Fig. 7. Scale bar = 100 µm.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The present results document for the first time that the expression of ANG II receptors by left ventricular fibroblasts exhibits a marked species dependence. These studies also confirm and extend our prior findings on ANG II receptor expression by rat left ventricular cardiac fibroblasts and confirm the observations of Sun and Weber (28) that ANG II receptor expression is upregulated within the infarction area of the rat left ventricle days after experimental coronary artery occlusion. In contrast, the present study found few, if any, functional ANG II receptors on cultured rabbit or human fibroblasts, with a small but significant increase in ANG II receptor binding in the rabbit left ventricle after experimental infarction. The lack of receptor binding by cultured rabbit cells was supported by the inability of saturating ANG II to stimulate Ca²⁺ mobilization or arachidonic acid release. The similarity in Ca²⁺ response between rabbit and human cells and our failure to find significant receptor expression on human cardiac fibroblasts indicate significant species differences in cardiac ANG II expression. This finding raises questions about the role of ANG II as a direct regulator of cardiac collagen biosynthesis and deposition across species.

ANG II receptors on cardiac fibroblasts. The classical approach in defining a receptor is the demonstration of specific and saturable high-affinity ligand binding with quantification of the density and affinity of the receptor. Alternatively, the combination of initial functional measurements followed by radioligand binding or measurement of specific mRNA encoding the receptor would suffice as proof of the existence of the receptor. Our initial approach with the rabbit cardiac fibroblast was the former, based on demonstration of a high-affinity ANG II binding site on rat cardiac fibroblasts by our laboratory and others. However, ANG II receptor binding experiments indicated <1% as many ANG II receptors on rabbit cardiac fibroblasts as were present on fibroblasts isolated from rat hearts. Furthermore, ANG II binding sites were not evident on isolated membranes, arguing against receptor translocation to an intracellular compartment. Yet, specific and saturable ANG II binding was demonstrable on rabbit aortic vascular smooth muscle cells and on cultured rabbit left ventricular myocytes, indicating that the assay methods employed should have detected ANG II receptors, if present, on rabbit cardiac fibroblasts.

In theory, ANG II receptors could have been expressed on rabbit cardiac fibroblasts in insufficient quantity to detect using radioligand binding but could still be functionally active. However, our inability to demonstrate ANG II-stimulated intracellular Ca²⁺ transients by the vast majority of rabbit cardiac fibroblasts assayed argued that such was not the case. The small number of positive cellular responses (observed in only 1 of 14 experiments using rabbit cardiac fibroblasts for a total of 6 cells and a single human cardiac fibroblast) seems most consistent with experimental artifact or trace contamination evident when examining single cells. Furthermore, ANG II failed to stimulate arachidonic acid release from rabbit cardiac fibroblasts, as was shown for rat cardiac fibroblasts. Neonatal human cardiac fibroblasts were capable of minimal ANG II binding relative to rat cardiac fibroblasts, and only 1 of 60 cells mobilized Ca²⁺ in response to ANG II. Therefore, the conclusion from the receptor binding and functional studies, and in contrast to rat cardiac fibroblasts, is that cultures of rabbit and human cardiac fibroblasts express few, if any, functional ANG II receptors.

The finding of ANG II receptors on suspended rabbit myocytes is not unexpected, since previous studies reported ANG II receptors in the membrane fractions from rabbit ventricular tissue (3, 5, 25, 27). Additionally, it has been shown that ANG II has a positive inotropic effect in the rabbit (3, 9, 27) and human (21) but not on the rat myocardium (8, 15). This functional assay is in close accordance with our demonstration of ANG II receptors on cultured rabbit myocytes but not on rat myocytes (16).

ANG II receptors within infarcted tissue. The absence of significant ANG II receptor expression on cultures of rabbit cardiac fibroblasts could be due to an influence of culture preparation, passage, or in vitro experimental conditions. Although this appeared unlikely, since rat and rabbit fibroblasts were isolated, cultured, and assayed under identical conditions, it could not be excluded. Fibroblasts from both rat and rabbit exhibited similar immunocytochemical identifiers; however, available immunocytochemical probes are not able to distinguish between fibroblasts and myofibroblasts. In fact, if immunoreactivity for -smooth muscle actin is the criteria, then fibroblasts from both species would be considered myofibroblasts. To address the issue of culture conditions, studies were conducted to analyze receptor expression in vivo after acute myocardial injury. Six days after coronary artery ligation, both rat and rabbit infarcted tissues were similar with regard to increased collagen deposition and cellular infiltration relative to the noninfarcted tissue; however, expression of ANG II receptors was markedly different between species. As reported previously (28), ANG II binding was increased in the fibrotic infarct of the rat. In the present study, we found a 35-fold increase in ANG II binding site density in the infarcted rat heart relative to the noninfarcted tissue. In contrast, ANG II binding within the infarcted rabbit hearts was only increased threefold relative to the noninfarcted tissue. The discrepancy in ANG II binding likely is not due to a difference in extent of injury as previously we showed that the infarction models used in this study produced myocardial infarctions of similar size in the two species (10). We cannot exclude the possibility that ANG II receptors are upregulated on the cardiac fibroblast within the infarcted rabbit heart, but the threefold increase in binding sites in the rabbit tissue may reflect changes on nonfibroblast cells, including invasive inflammatory cells. Nonetheless, the discrepancy in ANG II binding after myocardial infarction between rat and rabbit is further evidence for a difference in ANG II receptor expression between species.

Species differences in expression and function of receptors in cells and tissue have been shown previously. For example, ANG II was shown to subserve an important role in intimal thickening in the rat carotid artery but not in porcine coronary arteries (13). In the central nervous system, ANG II receptor localization was shown to differ among rats, rabbits, and humans (1). In addition, we have data indicating that bradykinin produces intracellular Ca²⁺ signals in cardiac fibroblasts with species-specific patterns, suggesting species variability in the handling of Ca²⁺ (T. D. Bahnson, M. P. Printz, N. N. Kim, and A. M. Gallagher, unpublished observations). On the basis of the ANG II receptor profile, the current study would argue that rabbit cardiac fibroblasts may be a more representative model of the human cardiac fibroblast than cells isolated from rat hearts.

ANG II and the use of ACE inhibitors. The findings of this study also have direct relevance to the ongoing controversy as to the mechanism(s) which underlie the ability of ACE inhibitors to limit left ventricular remodeling after infarction or in the setting of heart failure or chronic hypertension. Long-term postinfarction pharmacotherapy with ACE inhibitors improves survival, decreases the risk of developing heart failure, and is associated with improved hemodynamics and ventricular function (2, 11, 23). The beneficial effects of ACE inhibitors may be due to improved cardiac hemodynamics and/or improved collateral blood flow and/or by reducing and reversing ventricular remodeling. Accordingly, ACE inhibitors may exert their beneficial effects via the cardiac myocytes, vascular smooth muscle cells, endothelial cells, or cardiac fibroblasts. Numerous in vivo studies using rat models of myocardial injury and pressure overload (20, 28, 32) as well as in vitro studies on cardiac cells derived from rats (33, 34) suggest that ANG II plays an important role in regulating collagen synthesis by cardiac fibroblasts. These studies led to the widely held hypothesis that ACE inhibitors limit myocardial fibrosis primarily via effects on tissue ANG II to directly regulate cardiac fibroblast synthetic function. However, the studies that indicate that angiotensin receptor blockade mimics the ability of ACE inhibitors to limit cardiac hypertrophy and fibrosis after myocardial injury have relied almost exclusively on the rat cardiac model (26). In a chronic dog model, specific AT₁-receptor blockade did not mimic ACE inhibitor-mediated decreases in left ventricular mass after myocardial injury (19). Taken together with the present data, these studies raise questions as to the mechanisms by which ACE inhibitors limit fibrosis. The current results would suggest that ANG II does not act directly on cardiac fibroblasts across species to regulate collagen deposition after myocardial injury and that a major, direct or indirect, target of ACE inhibitors to regulate cardiac fibroblast function are signaling molecules other than ANG II.