OUP user menu

Angiotensin IV has mixed effects on left ventricle systolic function and speeds relaxation

Bryan K. Slinker, Yiming Wu, Adam J. Brennan, Kenneth B. Campbell, Joseph W. Harding
DOI: http://dx.doi.org/10.1016/S0008-6363(98)00344-7 660-669 First published online: 1 June 1999

Abstract

Objective: A novel angiotensin receptor has been described and named AT4. Ligands for this receptor include the angiotensin II (Ang II) metabolite Ang II (3–8), known as angiotensin IV (Ang IV). There is 10-fold more AT4 receptor than AT1 receptor in rabbit myocardium. The AT4 receptor has a high affinity for Ang IV (Ki in rabbit myocardium <2×10−9) and similar ligands, but very low affinity for Ang II (Ki in rabbit myocardium >10−6). Although several functions have been attributed to the novel Ang IV peptide/AT4 receptor system, the effect of this system on left ventricular (LV) function has not been studied. We hypothesized (1) that Ang IV would affect LV function and (2) that any effects would be opposite to those of Ang II. Methods: Using the buffer-perfused (30°C) isolated rabbit heart, we studied the effect of the AT4 agonist Nle1-Ang IV on LV systolic function, quantified using both Frank–Starling and end-systolic pressure–volume relationships, and relaxation. We also studied the effect of the AT1/AT2 agonist, Sar1-Ang II on LV function. Finally, because the profile of effect of Nle1-Ang IV was similar to the reported effect of nitric oxide (NO), we also studied the effect of Nle1-Ang IV in the presence of the NO synthase inhibitor NG-monomethyl-l-arginine. Results: Nle1-Ang IV reduced LV pressure-generating capability at any volume but increased the sensitivity of pressure development to volume change. Nle1-Ang IV reduced LV ejection capability. Sar1-Ang II had the opposite effect – increasing both pressure generation and ejection capability. Finally, both Sar1-Ang II and Nle1-Ang IV speeded LV relaxation. Inhibition of NO synthase did not alter the effect of Nle1-Ang IV on LV systolic function or relaxation. Conclusions: AT4 receptor agonism has mixed effects on LV systolic function, depressing pressure-generation and ejection capabilities, but enhancing the sensitivity of pressure development to volume change. It also speeds relaxation. The effect of Ang IV on systolic function is generally opposite to the effect of Ang II, whereas the Ang IV influence on relaxation is similar to the effect of Ang II.

Keywords
  • Angiotensin
  • Contractile function
  • Nitric oxide
  • Renin angiotensin system
  • Ventricular function

Time for primary review 29 days.

1 Introduction

A novel angiotensin receptor has been described recently and named AT4 [1–4]. Unlike other known AT receptors, AT4 is probably not G-protein linked [1, 4–7]. Ligands for this receptor include the angiotensin II (Ang II) hexapeptide metabolite Ang II (3–8), commonly known as angiotensin IV (Ang IV), and the angiotensin I (Ang I) octapeptide metabolite Ang I (3–10) [1, 2]. [125I]-Ang IV binding to the AT4 receptor has been characterized in several cardiovascular tissues and cells, including membrane preparations of bovine vascular smooth muscle cells [8, 9], bovine aortic endothelial cells [5], coronary microvascular endothelial cells [9], and in both rabbit and guinea pig myocardium [1, 2]. The quantity of AT4 receptor in rabbit myocardium (731 fmol mg−1 protein) [1] is an order of magnitude larger than the estimated quantity of AT1 receptor (∼50–70 fmol mg−1 [10, 11]). Thus, there is abundant AT4 receptor in the myocardium and this receptor has a high affinity for Ang IV (Ki in rabbit myocardium <2×10−9) [1, 12] and similar ligands, but very low affinity for Ang II (Ki in rabbit myocardium >10−6) [1].

Several physiological functions have been attributed to the novel Ang IV peptide/AT4 receptor system. Ang IV has been associated with the control of blood flow in the microvasculature of both the cerebral cortex of rats and rabbits [13–15] and the renal cortex of rats [3, 16], where it acts as a vasodilator, perhaps via an endothelium-dependent mechanism, and thus acts opposite to the effect of Ang II. Ang IV has also been shown to antagonize Ang II-induced hypertrophy in cultured chick myocytes [17], and to blunt mechanically induced immediate-early gene expression in the rabbit heart [12]. Other reports have shown that Ang IV increases DNA and RNA synthesis in rabbit cardiac fibroblasts [7] and has a synergistic effect with basic fibroblast growth factor to increase DNA synthesis in coronary microvascular endothelial cells in vitro [9]. Finally, Ang II induction of plasminogen activator inhibitor-1 (PAI-1) in bovine aortic endothelial cells is mediated by Ang IV acting at the AT4 receptor following Ang II→Ang IV conversion [18].

These observations of (1) a large amount of high affinity AT4 receptor in rabbit myocardium, (2) important physiological effects of Ang IV in several organ systems or cell types, including several in the cardiovascular system, and (3) Ang IV having effects opposite to those of Ang II in the heart and vasculature led us to investigate two hypotheses: One, we hypothesized that Ang IV would affect left ventricle (LV) function in the rabbit heart. Two, we further hypothesized that if Ang IV did have an effect it would be opposite to the effects of Ang II. To test these hypotheses we used the isolated rabbit heart preparation to study the effect of the AT4 agonist Nle1-Ang IV both on LV systolic function, as evaluated by both Frank–Starling relations in isovolumic beats and end-systolic pressure–volume relations (ESPVRs) in ejecting beats, and on cardiac relaxation in isovolumic beats. For comparison, we also studied the effect of the Ang II analog Sar1-Ang II on these same aspects of LV function. Finally, because the profile of effect of Nle1-Ang IV was similar to that reported for the effect of nitric oxide (NO), we also studied the effect of Nle1-Ang IV in the presence of the NO synthase inhibitor NG-monomethyl-l-arginine (l-NMMA).

2 Methods

2.1 Isolated heart preparation

This study was approved by the Washington State University Institutional Animal Care and Use Committee and conformed with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Experiments were performed using hearts isolated from 27 adult male New Zealand White rabbits (median weight 3.3 kg; range 2.6–3.9 kg). Details of the procedure for removing the heart from anesthetized rabbits and mounting it on a volume-servo system for pressure- and volume-control experiments have been described previously [19, 20]. Briefly, rabbits were anesthetized by intramuscular injection of ketamine, xylazine, and atropine (35, 7.5, and 0.02 mg kg−1, respectively). After a tracheostomy, surgical anesthesia was maintained with 1–2% isoflurane via positive-pressure ventilation. A midsternal thoracotomy was performed, the brachiocephalic artery was cannulated, and 1000 units of heparin were administered via the cannula. A modified Tyrode solution which contained 35 Mm K+ was vigorously bubbled with 95%O2–5%CO2 and delivered via the brachiocephalic cannula to arrest the heart. This relaxing solution was composed of (in mM); Na+ 121, K+ 35, Cl 138, Ca2+ 0.1, Mg2+ 1.08, HCO3 21, PO43− 0.36, and glucose 11.1, with 2.5 units l−1 regular insulin. The aorta was then ligated, and the heart was removed and perfused at a constant pressure of 90 mm Hg.

The heart was transferred to a perfusion support system, which was maintained at a constant temperature of 30°C. A thin latex balloon (constructed from an adult esophageal balloon) was secured to the flared ostium of the piston cylinder of the volume-servo system. A suture attached to the end of the balloon was passed through the mitral orifice and out through a puncture in the LV apex. The purse-string suture was then tightened around the flared ostium to secure the balloon within the LV chamber. A 5-Fr Millar Micro-Tip catheter pressure transducer was advanced through a side port into the center of the balloon. The perfusate was then changed to a modified Tyrode solution, which contained a lower K+ concentration, to allow contractions. This solution was composed of (in mM); Na+ 147, K+ 7.44, Cl 138, Ca2+ 1.24, Mg2+ 1.08, HCO3 21, PO43− 0.36, and glucose 11.1, with 2.5 units l−1 insulin. The heart beat isovolumically and was paced at 1 s−1 using field stimulation via copper plate electrodes inserted into the wall of the chamber in which the heart was immersed in physiological buffer. Field stimulation activates both ventricles simultaneously, but does not significantly alter LV function other than to increase LV systolic pressure generation or ejection in the baseline state. This effect, which is present uniformly throughout all experimental conditions, can be blocked by 10−6 M propranolol, and thus is probably due to intra-cardiac catecholamine release caused by the pacing pulse.

A single-beat, variably preloaded Frank–Starling protocol [19] was conducted to identify the LV volume (Vmax) at which the peak developed pressure of the Frank–Starling curve was obtained. LV volume was then adjusted so that it was 80% of Vmax, and this was the baseline volume, VBL, used throughout the experiment.

2.2 Preparation of solutions

All drugs were prepared as concentrated stock solutions in distilled water, and stored at −20°C. Stock solution was diluted to deliver drug at the desired concentration. Perfusate buffer was not re-circulated during an experiment. Sar1-Ang II and propranolol were obtained from Sigma, l-NMMA was obtained from CalBiochem, and Nle1-Ang IV was synthesized in our laboratories (J.W.H.).

2.3 Experimental protocols

Hearts were randomly assigned to one of four treatment groups, each consisting of 6–8 hearts. The first two groups comprised the core contrast of the study: the Ang IV Group received 10−10 M of the AT4 receptor agonist Nle1-Ang IV, and the Ang II Group received 10−8 M of the AT1/AT2 agonist Sar1-Ang II. Two other groups provided ancillary information to aid the interpretation of the results from these two core groups: The third group, the Ang II+Propranolol Group, received 10−8 M Sar1-Ang II in the presence of 10−6 M of the β-blocker propranolol in order to show that the effects of Ang II were not mediated by Ang II-induced catecholamine release [21]. The fourth group, the Ang IV+L-NMMA Group, received 10−10 M Nle1-Ang IV, followed by the combination of 10−10 M Nle1-Ang IV and 3×10−5 M of the nitric oxide synthase inhibitor l-NMMA. This group was studied because the effects of Nle1-Ang IV observed in pilot studies were similar to those previously reported for NO [22], and because there is precedent for an endothelium-dependent action of Ang IV in the microvasculature [13, 23]. Angiotensin peptide doses were based on preliminary dose–response studies (data not shown). The dose of l-NMMA was chosen based on the results of Smith et al. [24], who showed that 3×10−5 M l-NMMA produced only about a 10% increase in coronary perfusion pressure at a fixed coronary flow of 25 ml min−1 in Langendorff-perfused isolated rabbit hearts.

After establishing VBL, but prior to treatment with drug, two protocols were conducted while in the control state: (i) a single-beat, isovolumic, variably preloaded Frank–Starling protocol, FS, and (ii) a single-beat, ejecting, variably afterloaded protocol, AL. Records taken during these protocols yielded a family of functional indices to serve as baseline control values. After completing these two protocols, perfusion was switched from control perfusate to one containing a drug, followed by a 15-min period for a stable baseline to be established, as indicated by stable peak pressure. Then, the FS and AL protocols were once again conducted and records taken during the influence of the peptide and/or drug.

The FS and AL protocols, which have been described in detail elsewhere [19, 20, 25], allow individual functional capabilities of the LV to be characterized, including the single-beat Frank–Starling relationship (developed LV pressure, Pdpk, vs. volume, V), ESPVR (end-systolic pressure, Pes, vs. end-systolic volume, Ves) and the isovolumic relaxation-wall stress relationship (relaxation time, T75–25, vs. peak wall stress, σpk, for isovolumic beats from the FS protocol).

From the isovolumic beats recorded in the FS protocol, we measured, as previously described [20, 25, 26], fully relaxed, passive pressure, Ppass, as the lowest pressure during the prolonged pause following each volume-perturbed beat and peak LV developed pressure, Pdpk, as the difference between peak pressure, Ppk, and Ppass. The characteristic time of relaxation, T75–25, was defined as the time of pressure decay from 75% Pdpk to 25% Pdpk.

Similar to the isovolumic beats from the FS protocol, Ppass for the ejecting beats recorded in the AL protocol was determined as the minimum pressure measured before volume was returned to VBL following ejection. The time of the maximal pressure/volume ratio of each pressure-clamped beat was defined as the time of end ejection, Tej, and the pressure at this time was defined as the end-ejection pressure, Pej. Pes was calculated as PejPpass. The volume at Tej was defined as Ves. Stroke volume, SV, was then computed as VBLVes.

2.4 Data analysis

The Frank–Starling relationships were fit, as described previously [20], using a multiple linear regression in which Pdpk was related to volume, V, according to Embedded Image(1) where D is a dummy variable to encode the drug treatment (D=1) condition and the baseline condition to which drug treatment was to be compared (D=0), and the bi are regression coefficients. Similarly, the T75−25 vs. σpk relationships were fit, as described previously [20], using the multiple linear regression equation Embedded Image(2) where the ci are regression coefficients, and D is a dummy variable as defined for Eq. (1).

ESPVRs, obtained by plotting Pes vs. Ves, were constructed from the ejecting beats recorded in the AL protocol. Similar to the approach used for fitting the Frank–Starling curves obtained from the FS protocol, the non-linear ESPVR obtained from these isolated hearts using the AL protocol were fit using the multiple linear regression equation Embedded Image(3) where the di are regression coefficients, and D is a dummy variable as defined for Eq. (1).

Because each heart served as its own control, dummy variables, defined using effects coding [27], were included in each regression equation to account for between-subjects variation in each parameter (except for bv2, cσ2, and dv2). For simplicity, these dummy variables are not shown in Eqs. (1)(3). Unless otherwise stated, all summary statistics are given as mean±SD. All statistical analyses were performed using Minitab 12.1. Regression parameter estimates that were considered significant (p<0.05) were retained in equations that were then used to generate the curve fits shown in Figs. 1–9.

3 Results

All results presented involve multiple linear regression fits using Eqs. (1)(3). In all cases, the data were well fit by these regression equations, with all R2≥0.95. For simplicity of presentation of results, regression parameters are not presented. All interpretations with respect to the drug-induced shifts in the relationships are based on either the set of parameters bd, cd, and dd – which, if significant (p<0.05) for their respective equation, indicate that the drug caused a shift in the intercept in the relationship, compared to baseline – or the set of parameters bvd, cσd, and dvd – which, if significant (p<0.05) for their respective equation, indicate that the drug caused a slope shift in the relationship, compared to baseline.

3.1 Nle1-Ang IV has mixed effects on LV systolic function and speeds relaxation

3.1.1 Effects on systolic function

10−10 M Nle1-Ang IV caused a non-parallel downward shift in the Frank–Starling relationship (Fig. 1). Although the effect of Nle1-Ang IV is minimal at the largest volumes, where the two Frank–Starling relationships converge, the Frank–Starling relationship obtained under the influence of Nle1-Ang IV is increasingly depressed below baseline as volume decreases. This non-parallel shift increases the steepness (“slope”) of the Frank–Starling relationship, which indicates an increased sensitivity of pressure development to volume change.

Fig. 1

Frank–Starling relationship [Pdpk (mm Hg) vs. volume (ml)] for baseline (no drug; ○) and for treatment with 10−10 M Nle1-Ang IV (●). The lines are curve fits (n=6) using multiple linear regression and Eq. (1)(- - -=baseline and ———=Nle1-Ang IV treatment): Pdpk=23.4+90.9V−18V2−24.4D+2.9V·D. Ang IV depresses LV pressure generation at low volumes, but has little effect at the highest volumes. On the other hand, Ang IV increases the sensitivity of pressure development to changes in volume, as can be seen by the steeper “slope” of the Ang IV curve.

Nle1-Ang IV (10−10 M) also caused a parallel down- and right-ward shift in the ESPVR (Fig. 2). The depression of pressure-generating capability and the effect on the ESPVR indicate a modest effect of Nle1-Ang IV to depress LV systolic function. On the other hand, the increased slope of the Frank–Starling relationship indicates an increased sensitivity of pressure development to volume change, which indicates an effect of Nle1-Ang IV to enhance LV systolic function.

Fig. 2

ESPVR [Pes (mm Hg) vs. Ves (ml)] for baseline (no drug) and for treatment with 10−10 M Nle1-Ang IV. For simplicity of presentation, individual data points are not shown (the between heart variability is similar to that shown in Fig. 1). The lines are curve fits (n=6) using multiple linear regression and Eq. (3)(- - -=baseline and ———=Nle1-Ang IV treatment): Pes=−26.3+144.1Ves−33.3Ves2−8.8D. Ang IV depresses LV systolic function at all Ves (i.e., there is a parallel shift of the ESPVR down and to the right).

3.1.2 Effect on relaxation

Nle1-Ang IV (10−10 M) caused a downward shift in the isovolumic stress-dependent relaxation relationship (Fig. 3), indicating that its effects on LV systolic function are accompanied by a positive lusitropic effect, i.e., relaxation is speeded at any wall stress.

Fig. 3

Stress-dependent relaxation relationship [T75–25 (ms) vs. σpk (mm Hg)] for baseline (no drug; ○) and for treatment with 10−10 M Nle1-Ang IV (●). The lines are curve fits (n=6) using multiple linear regression and Eq. (2)(- - -=baseline and ———=Nle1-Ang IV treatment): T75–25=0.1+0.00015σpk+0.0000039σpk2−0.008D−0.0008σpk·D. Ang IV speeds LV relaxation – i.e., causes a downward shift in the stress-dependent relaxation relationship.

3.2 The effects of Nle1-Ang IV are not mediated via NO

3.2.1 Effect on systolic function

The Frank–Starling relationships obtained under the influence of 10−10 M Nle-1-Ang IV alone and under the influence of 10−10 M Nle-1-Ang IV in the presence of 3×10−5 M l-NMMA superimpose (Fig. 4). Similarly, there is no difference in ESPVR (data not shown). Both results indicate that inhibition of NO synthase does not alter the influence of Nle1-Ang IV on LV systolic function.

Fig. 4

Frank–Starling relationship [Pdpk (mm Hg) vs. volume (ml)] for treatment with 10−10 M Nle1-Ang IV alone (○) and for treatment with 10−10 M Nle1-Ang IV in the presence of 3×10−5 M l-NMMA to inhibit NO synthase (●). The lines are curve fits (n=8) using multiple linear regression and Eq. (1)(- - -=Nle1-Ang IV alone and ———=Nle1-Ang IV+l-NMMA): Pdpk=−20.9+114.6V−22.1V2−5.3D+2.2V·D. Although both bd (=−5.3) and bvd (=2.2) are statistically significant, these values relate to effects extrapolated to zero volume. However, there is no appreciable effect over the observed range of volume. Thus, l-NMMA does not alter the effect of Nle1-Ang IV.

3.2.2 Effect on relaxation

The isovolumic stress-dependent relaxation relationships obtained under the influence of 10−10 M Nle-1-Ang IV alone and under the influence of 10−10 M Nle-1-Ang IV in the presence of 3×10−5 M l-NMMA superimpose (Fig. 5), indicating that inhibition of NO synthase does not alter the influence of Nle1-Ang IV on LV relaxation.

Fig. 5

Stress-dependent relaxation relationship [T75–25 (ms) vs. σpk (mm Hg)] for treatment with 10−10 M Nle1-Ang IV alone (○) and for treatment with 10−10 M Nle1-Ang IV in the presence of 3×10−5 M l-NMMA to inhibit NO synthase (●). The lines are curve fits (n=8) using multiple linear regression and Eq. (2)(- - -=Nle1-Ang IV alone and ———=Nle1-Ang IV+l-NMMA): T75–25=0.12+0.00041σpk+0.0000092σpk2. l-NMMA does not alter the effect of Nle1-Ang IV.

3.3 Sar1-Ang II enhances LV systolic function and speeds relaxation

3.3.1 Effect on systolic function

Sar1-Ang II (10−8 M) caused a parallel upward shift in the Frank–Starling relationship (Fig. 6). Sar1-Ang II (10−8 M) also caused a non-parallel up- and left-ward shift in the ESPVR (Fig. 7). Although, the effect of Sar1-Ang II on the ESPVR is minimal at the smallest Ves, where the two ESPVRs converge, the ESPVR obtained under the influence of Sar1-Ang II is increasingly above, and to the left of, the baseline ESPVR as Ves increases (or, conversely, as SV decreases). Both the effect on the Frank–Starling relationship and the effect on the ESPVR indicate that Sar1-Ang II enhances LV systolic function.

Fig. 7

ESPVR [Pes (mm Hg) vs. Ves (ml)] for baseline (no drug) and for treatment with 10−8 M Sar1-Ang II. For simplicity of presentation, individual data points are not shown (the between heart variability is similar to that shown in Figs. 1 and 6). The lines are curve fits (n=6) using multiple linear regression and Eq. (3)(- - -=baseline and ———=Ang II treatment): Pes=−20.9+106.4Ves−20.7Ves2+4.1Ves·D. Ang II enhances LV systolic function, particularly at larger Ves (i.e., smaller SV).

Fig. 6

Frank–Starling relationship [Pdpk (mm Hg) vs. volume (ml)] for baseline (no drug; ○) and for treatment with 10−8 M Sar1-Ang II (●). The lines are curve fits (n=6) using multiple linear regression and Eq. (1)(- - -=baseline and ———= Ang II treatment): Pdpk=86.7V−15.4V2−3.2D+6.7V·D. Ang II enhances LV systolic function – i.e., the Frank–Starling relation shifts up and to the right.

3.3.2 Effect on relaxation

Sar1-Ang II (10−8 M) caused a non-parallel downward shift in the isovolumic stress-dependent relaxation relationship (Fig. 8), indicating that its effect to enhance LV systolic function is accompanied by a positive lusitropic effect, i.e., relaxation is speeded at any wall stress, with the speeding of relaxation being relatively larger at higher values of stress.

Fig. 8

Stress-dependent relaxation relationship [T75–25 (ms) vs. σpk (mm Hg)] for baseline (no drug; ○) and for treatment with 10−8 M Sar1-Ang II (●). The lines are curve fits (n=6) using multiple linear regression and Eq. (2)(- - -=baseline and ———=Ang II treatment): T75–25=0.1+0.00028σpk+0.0000033σpk2−0.0002σpk·D. Similar to the effect of Ang IV, Ang II speeds LV relaxation.

3.3.3 The effects of Sar1-Ang II are qualitatively unchanged by β-receptor blockade

Sar1-Ang II (10−8 M) given in the presence of 10−6 M propranolol to block β receptors shifted both the Frank–Starling relationship and ESPVR in a manner qualitatively similar to the effect observed in the absence of propranolol (data not shown). The shift in the Frank–Starling relationship was slightly non-parallel in the presence of β blockade, with the relationships converging at lower volumes. Similarly, 10−8 M Sar1-Ang II given in the presence of 10−6 M propranolol shifted the isovolumic stress-dependent relaxation relationship downward in a manner qualitatively similar to the effect observed in the absence of propranolol (Fig. 9). For all these effects, the magnitude of shift caused by Sar1-Ang II in the presence of β blockade was about 1/2 the magnitude observed without β blockade, indicating that some of the effects of Sar1-Ang II shown in Figs. 6–8 may have been due to Ang II-induced catecholamine release.

Fig. 9

Stress-dependent relaxation relationship [T75–25 (msec) vs. σpk (mm Hg)] for baseline (no drug) and for treatment with 10−8 M Sar1-Ang II, both in the presence of 10−6 M propranolol to block β receptors. For simplicity of presentation, individual data points are not shown (the between-heart variability is similar to that shown in Figs. 3 and 5). The lines are curve fits (n=7) using multiple linear regression and Eq. (2)(- - -=baseline and ———=Ang II treatment): T75–25=0.11+0.0002σpk+0.0000036σpk2−0.005D−0.0005σpk·D. In the presence of β blockade, Ang II has the same qualitative effect on relaxation as in the absence of β blockade. Quantitatively, however, the effect of Ang II is reduced after β blockade, particularly at higher values of σpk (e.g., compare to Fig. 8).

3.4 Effect of treatments on coronary flow

Nle1-Ang IV (10−10 M) was associated with a decrease in coronary flow from 41 to 38 ml min−1 (7% decrease; p=0.005). Similarly, Sar1-Ang II (10−8 M; data pooled from experiments with and without propranolol) was associated with a decrease in coronary flow from 39 to 37 ml min−1 (5% decrease; p=0.043). Although both these decreases are statistically significant, neither is large and both are consistent with the small decreases in coronary flow that are observed over time in this experimental preparation. l-NMMA (3×10−5 M) was expected to have a small vasoconstrictor effect, and was, in fact, associated with an 18% decrease in coronary flow (from 33 to 27 ml min−1; p<0.01).

4 Discussion

4.1 The AT4 receptor influences LV function

The major findings of this study are that an angiotensin peptide ligand of the AT4 receptor, Nle1-Ang IV, affects LV systolic function and relaxation in the isolated rabbit heart and that it does so, in part, opposite to the effects of Ang II. This is the first report of AT4-mediated effects on LV function.

4.1.1 Effects on systolic function and relaxation

Nle1-Ang IV had mixed effects on LV systolic function: The analysis of pressure-generating capability using the Frank–Starling relationship shows features that can be interpreted as both depressing and enhancing LV systolic function. The downward shift of the Frank–Starling relationship (Fig. 1) means that pressure-generating capability is depressed for any given volume, which can be interpreted as depressed LV systolic function. On the other hand, this downward shift is not parallel such that the slope of the Frank–Starling relationship is increased, which can be interpreted as enhanced LV systolic function. Analysis of ejecting capability using the ESPVR shows that LV systolic function is depressed. These mixed effects emphasize the multi-factorial nature of LV function and the need to consider multiple features of LV function when evaluating the effects of an intervention on LV “contractility” [28].

These effects on LV systolic function were accompanied by a speeding of relaxation, as shown by the downward shift of the T75–25 vs. σpk relationship (Fig. 3). Thus, at any wall stress, relaxation is faster in the presence of 10−10 M Nle1-Ang IV. Although the two curve fits are slightly non-parallel, the effect on relaxation is similar at all σpk.

4.1.2 The response to Ang IV does not depend on NO

Because this profile of effect on LV function – slight depression in systolic function and speeding of relaxation – is similar to the reported effect of NO on LV function [22], and because the effect of Ang IV to decrease vascular resistance has been reported to be endothelial dependent [13, 23], we repeated the study of the effect of Nle1-Ang IV in the presence of 3×10−5 M l-NMMA to inhibit the production of NO. When the effect of Nle1-Ang IV alone is compared to the effect of Nle1-Ang IV in the presence of l-NMMA (Figs. 4 and 5), the systolic function and relaxation curves superimpose, indicating that the effect of Nle1-Ang IV is not changed by inhibiting NO production. Thus, the effects of Nle1-Ang IV on LV function appear to be independent of NO. It is unlikely that the effect of l-NMMA to decrease coronary flow by 18% significantly affects this conclusion. l-NMMA did not change the passive pressure–volume relationship (data not shown) and so unlikely affected LV function via a so-called turgor effect. Further, the average coronary flow of 27 ml min−1 observed with the combination of l-NMMA and Nle1-Ang IV is within the range we observe at a late time point in this experimental preparation (the sequence of events is such that these measurements occurred approximately 45–60 min after the heart had resumed beating after isolation). Thus, although l-NMMA did exert a vasoconstrictor effect, it was mild within the context of our experimental preparation, and so it is unlikely that reduced coronary flow confounded our interpretations.

4.1.3 Evidence that Nle1-Ang IV is acting at the AT4 receptor

AT4 receptor binding has been characterized in both rabbit and guinea pig myocardium [1, 2]. The quantity of AT4 receptor in rabbit myocardium has been estimated at 731 fmol mg−1 protein) [1], which is an order of magnitude larger than the estimated quantity of AT1 receptor (∼50–70 fmol mg−1 [10, 11]). The AT4 receptor has a high affinity for Ang IV (Ki in rabbit myocardium=<2×10−9) [1, 12] and similar ligands, such as Nle1-Ang IV, which shows a two-site competition curve in rabbit myocardium (Ki for the high affinity site=6.18×10−12 and Ki for the low affinity site=1.87×10−9) [12]. Conversely the AT4 receptor has a very low affinity for Ang II (Ki in rabbit myocardium >10−6) [1]. Likewise, Ang IV and related ligands have very poor affinity for AT1 (Ki>10−6) and AT2 (Ki>10−4) receptors [3]. Thus, the response we observed to 10−10 M Nle1-Ang IV was due to action at the high affinity AT4 site in rabbit myocardium. Because these quantitative binding studies have been done on whole heart homogenate, there is no specific published information available regarding the cellular localization of these receptors. Highly specific Ang IV binding sites have, however, been observed in bovine coronary vascular endothelial cells [9], avian myocytes [17], and both rabbit cardiac fibroblasts and myoctyes [[4, 5] Hanesworth et al., unpublished data].

4.2 Protocol design considerations

This study was done in the isolated heart, which was buffer perfused at 30°C. Thus, the results must be viewed in the context of a heart isolated from circulating components of the renin angiotensin system. For the purpose of testing a hypothesis about the influence of angiotensin peptides on LV function, the isolated heart was chosen because, in contrast to an intact circulatory system preparation in the whole animal, the isolated heart minimizes the number of confounding variables and, thus, improves our ability to interpret the results. It would, however, be unwarranted to make extensive extrapolation from the results of this study in order to speculate about the possible physiological role(s) of Ang IV in a heart in an intact circulatory system.

These data were obtained at sub-physiologic temperature and heart rate. Both of these factors are consistent with all previous reports from our laboratory and are concessions we have made routinely in order to obtain better stability from this preparation [19, 20, 25, 26, 29, 30]. Based on our extensive previous experience with the study of LV function using this preparation, the lower temperature and heart rate are unlikely to cause us to misjudge the effects of these angiotensin peptides.

4.3 Effects mediated by the AT4 receptor are, in part, opposite the effects of Ang II

In contrast to the features that show depression of LV systolic function by Nle1-Ang IV (reduced pressure development and decreased ejection capability), all features of LV systolic function were enhanced by 10−8 M Sar1-Ang II, a non-specific AT1/AT2 receptor agonist. This positive effect of Ang II on systolic function is consistent with previous reports in most species, including the rabbit [31, 32].

Unexpectedly, LV relaxation was speeded by both 10−10 M Nle1-Ang IV and 10−8 M Sar1-Ang II. Ang II is generally considered to have either no effect, or a negative effect, on relaxation [31, 33, 34]. The negative effect of Ang II on relaxation has been suggested to be exacerbated in LV hypertrophy and Ang II has been shown to shift the diastolic passive pressure–volume relationship upward in animal models of LV hypertrophy [35, 36]. Similarly, intra-coronary (i.c.) angiotensin converting enzyme (ACE) inhibitors have been shown to improve, at least by a small amount, LV relaxation in human subjects [37, 38]. However, most of the studies of the effects of Ang II on relaxation have been limited in that they did not consider the important load dependence of relaxation. The study of freely shortening adult rat cardiomyocytes by Neyses and Vetter [34] is a good example of this limitation (even though the myocytes were subjected to no external load). Ang II increased the extent of shortening in these cells (a positive inotropic effect). The twitch duration also increased, which was interpreted to mean that relaxation was slowed. This interpretation, however, fails to account for the increased extent of shortening and total twitch duration. Likewise, the studies by Friedrich et al. [37] and Haber et al. [38] are difficult to interpret with respect to the reported changes, or lack thereof, in relaxation in response to i.c. ACE inhibitor: Were the observed effects of ACE inhibition on LV relaxation a direct effect of the intervention on relaxation, or were they secondary to the effects of ACE inhibition to decrease LV systolic function or preload?

Because we took the load dependence of relaxation into account by examining the relationship between relaxation and wall stress (Fig. 8), we are able to demonstrate an unequivocal effect of Sar1-Ang II to speed relaxation (see Slinker et al. [20] for a fuller discussion of the advantages of using the stress-dependent relaxation relationship to better appreciate the effects of an intervention on LV relaxation).

In summary, we showed that Ang IV has mixed effects on LV systolic function, with depressed ejection capability and pressure development, but enhanced sensitivity of pressure development to changes in volume. In contrast, Ang II enhanced all of the LV systolic functional capabilities that we examine. Thus, Ang II and Ang IV have generally opposite effects on LV systolic function, which is consistent with a pattern of effects of Ang IV being opposite to those of Ang II [39]. In contrast, and unexpected in relation to the emerging general pattern of opposing effects of Ang IV and Ang II, we showed that both Ang IV and Ang II speed LV relaxation.

4.4 Ang IV/AT4 system influences on cardiovascular function

Although this is the first report of an AT4-mediated effect on LV function, there have been other reports of the effect of Ang IV, or related ligands, in myocardium. These include the demonstration that Ang IV could independently antagonize Ang II-induced hypertrophic changes in cultured chick myocytes [17] and that Nle1-Ang IV could blunt mechanically induced immediate-early gene expression in the isolated rabbit heart [12]. In a related study [Yang and Slinker, unpublished observations], it was shown that AT1 receptor blockade with 10−6 M Losartan produced a quantitatively similar blunting of mechanically induced immediate-early gene expression.

Hence, there is evidence that Ang IV opposes the effects of Ang II in the heart. The results of this study also provide evidence of opposite effects of Ang IV and Ang II – the effects of Nle1-Ang IV on LV systolic function were generally opposite to the effects of Sar1-Ang II. That Ang IV and Ang II exert opposite effects on cardiovascular function is also supported by studies of the effect of Ang IV on vascular resistance where, opposite to the effect of Ang II, Ang IV acts as a vasodilator and decreases vascular resistance [3, 13, 14].

In addition to their generally opposite effects on LV systolic function, however, we also observed that Ang IV and Ang II had a similar effect on LV relaxation. At present there is no clear explanation for this similar effect of both Ang II and Ang IV on relaxation, when they have different effects on pressure generation and ejecting capability. So little is known about the cell signaling pathways involved in AT4 signaling that speculation about the cellular mechanisms of these effects is difficult. The only information known at present is that the AT4 receptor is not G-protein linked [1, 4–7]. Beyond that, there are no studies of intracellular mechanisms relating to intracellular Ca2+ and pH, as has been done for the effects of Ang II on cardiac function [40].

Although the broad significance of the Ang IV/AT4 system for the heart remains to be determined, we have shown that Nle1-Ang IV, acting at the AT4 receptor can influence LV function in a manner that is, in part, opposite to the effect of Sar1-Ang II acting at AT1 and/or AT2 receptors. Given the importance of the cardiac angiotensin system to cardiac pathophysiology [41–43], the possible role of the Ang IV/AT4 system to interact with, or modulate, the effects of the better understood Ang II/AT1 system merits further investigation.

Acknowledgements

Supported by grant-in-aid WA-96-77290 from the Washington Affiliate of the American Heart Association (to B.K.S.). A.J.B. was supported by a Howard Hughes Undergraduate Biological Sciences Education Initiative Grant.

References

View Abstract