INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IN CONGESTIVE HEART failure (CHF), sympathetic activity increases in parallel with the impairment of cardiac performance (6, 18, 27) and may contribute to the progression of the heart failure (23). In rats with CHF postmyocardial infarction (MI), plasma catecholamines, sympathoexcitatory responses to air-jet stress, and sympathoinhibitory responses to central injection of the ₂-adrenergic-agonist guanabenz (15) increase, whereas arterial and cardiopulmonary baroreflex control of sympathetic activity decreases (13). Brain ouabain-like activity ("ouabain") increases as well (15), and blockade of brain "ouabain" normalizes the sympathetic hyperactivity (15) and the impairment of baroreflex function in rats with CHF (13). In normotensive rats and spontaneously hypertensive rats, the sympathetic hyperactivity and impairment of baroreflex function in response to high cerebrospinal fluid sodium and high-sodium intake, respectively, appear to depend on brain "ouabain" followed by activation of the brain renin-angiotensin system (RAS; see Refs. 11 and 12). Studies on the brain RAS in CHF are so far very limited. Acute intracerebroventricular injection of losartan decreased resting renal sympathetic nerve activity (RSNA) and improved the abnormal arterial baroreflex control of RSNA in rats with CHF, suggesting that the brain RAS contributes to the impairment of baroreflex regulation of RSNA in CHF (5). Effects of chronic blockade of the brain RAS on specific parameters of sympathetic hyperactivity and impairment of baroreflex function in CHF have so far not been evaluated. Chronic blockade evaluates whether compensatory mechanisms develop or whether the brain RAS is essential for the chronic sympathetic hyperactivity in CHF.

To elucidate the role of the central RAS in long-term cardiovascular regulation in CHF, in the present study we examined whether in rats with CHF post-MI the brain RAS chronically contributes to the sympathetic hyperactivity and impairment of arterial baroreflex control of RSNA and heart rate (HR).

For this, rats received from 2 to 4 wk or from 6 to 8 wk post-MI an intracerebroventricular infusion of the AT₁-receptor blocker losartan at a dose of 1 mg · kg¹ · day¹ followed by assessment of mean arterial pressure (MAP), HR, and RSNA responses to air-jet stress and guanabenz and of arterial baroreflex function in the conscious state.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Wistar rats (male, 200-250 g) were obtained from Charles River Breeding Laboratories (Montreal, Quebec, Canada). The rats were fed a standard commercial rat chow and water ad libitum. After 5-7 days of acclimatization, acute left coronary artery ligation was performed to induce MI and heart failure as described previously (15). Under methohexital sodium (50-60 mg/kg ip) anesthesia, an endotracheal tube was inserted and connected with a respirator (model 683; Harvard Rodent Ventilator) with room air. The thorax was opened at the left fourth or fifth intercostal space, and the left coronary artery was ligated 2-3 mm from its origin with a 6-0 suture. Positive end-expiratory pressure was then applied to inflate the lung before the chest was closed in layers. The mortality during surgery and subsequent 48-h period was ~50%. Control (sham-operated) rats underwent the same surgical procedure without coronary artery ligation. The study was carried out in accordance with the guidelines of the University of Ottawa Animal Care Committee.

After the coronary artery ligation or sham ligation (2 or 6 wk), under pentobarbital sodium (60 mg/kg ip) anesthesia, an intracerebroventricular guide stainless steel tube was implanted 0.5 mm above the right lateral cerebroventricle, and an L-shaped stainless steel tube implanted in the opposite cerebroventricle was connected, via a PE-50/60 tube, to an osmotic minipump (model 2002; Alzet) filled with losartan (1 mg · kg¹ · day¹, CHF-Los) or vehicle (physiological saline; CHF-Veh). The pump was implanted subcutaneously on the back of the rat. The infusion rate was 12 µl/day. Chronic intracerebroventricular infusion of vehicle at this rate has no measurable effects on resting hemodynamics (10). Sham-operated rats underwent identical surgical procedures without pump implantation and treatment (Sham). According to previous studies, this dose of losartan does not result in detectable changes in systemic hemodynamics when administered peripherally (11, 14). To further substantiate that peripheral actions of centrally administered losartan did not contribute, in the 6-8 wk post-MI experiment, one group of rats received losartan subcutaneously at 1 mg · kg¹ · day¹ (CHF-per.Los), using the same type of osmotic minipumps. To assess the possible effects of losartan at the above rates in normotensive control rats, in a separate control experiment, rats were randomized to losartan intracerebroventricularly or subcutaneously at a rate of 1 mg · kg¹ · day¹ or vehicle intracerebroventricularly for 2 wk, using the same procedures outlined above. Table 1 provides an outline of the different experimental groups in the three different experiments.

After minipump implantation (2 wk), under halothane and supplemental intravenous methohexital anesthesia, the left femoral vein and artery were cannulated, and a catheter was placed in the right jugular vein with its tip at the level of the right atrium. The left kidney was exposed via a left flank incision. One of the nerves to the left kidney was dissected free from surrounding tissue and hooked by a pair of silver electrodes to record RSNA. When an optimal signal was recorded, the electrodes and nerve were glued together with SilGel 604 (Wacker, Munich, Germany), and the electrodes were fixed to back muscle and were tunnelled to the back of the neck. The rat was then placed in a small cage that allowed it to move back and forth. The experiment was started 4 h after the rats woke up from the anesthesia. The intra-arterial catheter was connected to a pressure transducer to record blood pressure (BP). The electrodes were connected to a Grass P511 bandpass amplifier (100-1,000 Hz), and RSNA (spikes/s) was counted by a nerve traffic analyzer (model 706C; University of Iowa Bioengineering). Intracerebroventricular injections were done via L-shaped stainless steel tubing (26 gauge) placed via the guide cannula so that its tip protruded 1.0 mm from the tip of the guide cannula in the right cerebral ventricle. The outer side of the tubing was connected to a Hamilton micro-syringe via a polyethylene tube (PE-10 fused to PE-50).

After 30 min of stabilization, basal MAP, HR, central venous pressure (CVP), and RSNA were recorded. Subsequently, the responses of these parameters to the following procedures were evaluated: 1) a jet of air (1.5-2 psi) was blown on the rat face two times, with a 3-min interval; 2) after 25 min of rest, phenylephrine (5-50 µg · kg¹ · min¹) was infused intravenously to obtain a ramp increase in MAP with a maximum increase of 50 mmHg over 2 min, and, after 10 min of rest, nitroprusside (10-100 µg · kg¹ · min¹) was infused intravenously to obtain a ramp decrease in MAP with a maximum decrease of 50 mmHg over 2 min; and 3) after 25 min of rest, the ₂-adrenoceptor-agonist guanabenz (75 µg in 5 µl) was injected intracerebroventricularly over 30 s. In the control losartan experiment, in addition, 0.6 µg of ouabain intracerebroventricularly and 30 ng of ANG II intracerebroventricularly {5 min after the vasopressin antagonist [d(CH₂)₅Tyr(Me)]AVP, 30 µg/kg iv} were administered after 25-min rest periods. Figure 1 shows a typical tracing of BP, HR, and RSNA in response to intravenous infusion of phenylephrine and sodium nitroprusside in a sham-operated rat.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Analog recording of blood pressure (BP), heart rate (HR), and filtered renal sympathetic nerve activity (RSNA) in response to intravenous infusion of phenylephrine (PE, left) and sodium nitroprusside (NP, right) in a sham-operated rat.

MAP, HR, and RSNA were recorded throughout the experiment by an on-line computer and a Grass polygraph. At the end of the experiment, the rat was killed by intravenous pentobarbital sodium, and the RSNA was recorded for 30 min, which was considered as the noise level from the recording system, and was subtracted from the total activity. For the present study, the minimal acceptable signal-to-noise ratio was set at 25. Methylene blue was injected intracerebroventricularly to check the accuracy of the intracerebroventricular cannulation. The heart was removed, and the infarct size was measured as previously described (15). Briefly, the left ventricular (LV) tissue was pressed flat, and the circumferences of the entire LV versus visualized infarcted area were outlined on a transparent plastic sheet. The difference in weight between the two marked areas on the sheet was used to determine the infarct size. The right ventricular (RV) free wall wet weight and the LV wet weight were obtained, and their ratio to body weight was calculated. In the 2- to 4-wk experiment, two rats in the CHF-Los group with <30% infarct area of the LV were excluded (24).

Data analysis. Responses of MAP and HR were expressed as net changes from resting levels. The responses of RSNA were expressed as percent changes from resting levels. For establishing best-fit relations between MAP and HR or MAP and RSNA, a nonlinear regression program (Sigmastat and Sigmaplot; Jandel Scientific) was used to analyze the baroreflex curve: Y = a/{1 + exp[b(X  c)]} + d, where X is MAP; Y is HR or RSNA; a is the range of HR or RSNA; b is the slope coefficient; c is MAP at the midpoint range of RSNA or HR; and d is the minimum of RSNA or HR. In each rat, raw data of MAP, HR, and RSNA were fit to the logistic function to generate parameters a, b, c, and d. The maximum gain of the baroreflex curve is defined as ba/4 (5, 28).

All other data were analyzed by using Sigmastat software and are expressed as means ± SE. One-way ANOVA was performed for comparison among groups. When F ratios were significant, Newman-Keuls test was applied to identify which groups were significantly different. Statistical significance was defined as P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

General characteristics. In the CHF groups, the average infarct size was in the moderate range (23), MAP was lower, CVP was increased, and RV weight clearly increased; the LV weight was significantly increased at 8 wk but not yet at 4 wk post-MI compared with those in the Sham groups. There were no differences in infarct size, basal MAP, HR, CVP, RV, or LV weight between the CHF groups (Table 2). In control rats, losartan intracerebroventricularly or subcutaneously for 2 wk had no effects on resting MAP and HR (Table 3).

Responses to stress and guanabenz. Air-jet stress induced modest increases in MAP, HR, and RSNA (Fig. 2). These responses to air-jet stress were significantly larger in both CHF-Veh groups compared with those in the Sham groups (P < 0.05). In both CHF-Los groups, responses of MAP and RSNA were not enhanced and were similar to those in the Sham group. The HR response was only normalized by central losartan treatment from 6 to 8 wk post-MI. Peripheral treatment with the same dose of losartan did not affect the enhanced responses to air-jet stress in CHF rats (Fig. 2). In control rats, losartan intracerebroventricularly or subcutaneously did not affect responses to air-jet stress (Table 3).

View larger version (35K): [in this window] [in a new window]   Fig. 2.   Peak increases in mean arterial pressure (MAP), HR, and RSNA in response to air-jet stress in rats with congestive heart failure (CHF) postmyocardial infarction (MI), treated with losartan icv or sc from 2 to 4 wk (A) or 6 to 8 wk (B) post-MI. Sham, sham-operated rats; CHF-Veh, heart failure rats that received vehicle infusion; CHF-Los and CHF-per.Los, heart failure rats that received losartan infusion (1 mg · kg1 · day1 icv or sc, respectively, for 2 wk). Values are means ± SE (for no. of rats/group, see Table 1). * P < 0.05 vs. Sham group; + P < 0.05 vs. CHF-Veh group.

Intracerebroventricular injection of guanabenz caused significant decreases in MAP, HR, and RSNA (Fig. 3). These responses to guanabenz were markedly larger in both CHF-Veh groups. Central losartan treatment completely normalized these responses (P < 0.05 for CHF-Los vs. CHF-Veh) in both groups from 2 to 4 wk and 6 to 8 wk post-MI. In contrast, peripheral treatment with the same dose of losartan did not affect the enhanced responses to guanabenz in the CHF rats (Fig. 3). In control rats, losartan intracerebroventricularly or subcutaneously did not affect responses to guanabenz intracerebroventricularly (Table 3).

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Peak decreases in MAP, HR, and RSNA in response to guanabenz (75 µg injected icv) in rats with CHF post-MI, treated with losartan icv or sc from 2 to 4 wk (A) or 6 to 8 wk (B) post-MI. * P < 0.05 vs. Sham group; + P < 0.05 vs. CHF-Veh group.

Because the BP was lower in CHF versus Sham rats, the above responses were also evaluated as percent changes. The pattern of responses described above persisted (data not shown).

Arterial baroreflex control of HR and RSNA. Increases or decreases in MAP induced the expected inhibitory or excitatory responses of RSNA (Fig. 4). Compared with the Sham group, both the maximal increase and decrease in RSNA, reflected by the range and minimum value of the RSNA-MAP reflex curve, were significantly attenuated in both CHF-Veh groups (Table 4 and Fig. 4). Moreover, the maximal gain of the curve was significantly less in the CHF-Veh group versus the Sham group (Table 4 and Fig. 4), indicating an impairment of arterial baroreflex control of RSNA in rats with CHF. In rats with CHF receiving central losartan treatment from either 2 to 4 wk or 6 to 8 wk post-MI, the minimum value of the curve was significantly lower, associated with a significant increase in the range. In both CHF-Los groups the maximum gain was normalized and was not different from that in the Sham groups (Table 4). In contrast, peripheral administration of losartan in the same dose did not significantly affect arterial baroreflex control of RSNA (Table 4 and Fig. 4).

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Changes of RSNA (%) against changes of MAP at increments of 5 mmHg elicited by iv infusions of nitroprusside and phenylephrine in rats with CHF post-MI, treated with losartan icv or sc from 2 to 4 wk (A) or from 6 to 8 wk (B) post-MI. Sigmoid curves were generated from averaged parameters as described in METHODS. Values are means ± SE (for no. of rats/group, see Table 1).

Increases or decreases in MAP induced the anticipated responses of HR (Fig. 5). Compared with the Sham groups, the range of the HR-MAP reflex curve was significantly attenuated in both CHF-Veh groups (Table 5). Moreover, the maximal gain of the curve was significantly decreased in the CHF-Veh groups versus the Sham groups (Table 5 and Fig. 5), indicating an impairment of arterial baroreflex control of HR in rats with CHF. When rats with CHF received central losartan treatment from 2 to 4 wk post-MI, the range and maximal gain of the curve were similar to those observed in the Sham group. Similarly, central losartan from 6 to 8 wk post-MI reverted these parameters of the HR-MAP reflex curve back to normal, whereas peripheral administration of losartan at the same rate did not cause significant improvements.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Changes of HR [beats/min (bpm)] against changes of MAP at increments of 5 mmHg elicited by iv infusions of nitroprusside and phenylephrine in rats with CHF post-MI, treated with losartan icv or sc from 2 to 4 wk (A) or 6 to 8 wk (B) post-MI. Sigmoid curves were generated from averaged parameters as described in METHODS. Values are means ± SE (for no. of rats/group, see Table 1).

In rats with CHF post-MI, the MAP midpoints were significantly lower than those of corresponding sham-operated rats for both the RSNA-MAP and the HR-MAP reflex curves. On losartan, the MAP midpoint remained decreased (Tables 4 and 5).

In control rats, losartan intracerebroventricularly or subcutaneously did not affect arterial baroreflex control of RSNA and HR (for maximum slope, see Table 3).

Responses to ouabain and ANG II. Intracerebroventricular injection of ouabain or ANG II caused the anticipated increases in MAP, HR, and RSNA. Intracerebroventricular losartan at the rate of 1 mg · kg¹ · day¹ for 2 wk markedly inhibited these responses, and remaining responses were no longer statistically significant (Table 3). In contrast, losartan subcutaneously at the same rate did not affect responses to centrally administered ouabain or ANG II (Table 3).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As a major new finding, the present study demonstrates that, in rats with CHF post-MI, "chronic" blockade of the brain RAS by intracerebroventricular losartan normalizes the decreased sympathoinhibition, increased sympathoexcitation, and impairment of arterial baroreflex control of RSNA and HR. These results indicate that the brain RAS is essential for the sympathetic hyperactivity and impairment of arterial baroreflex function during the development of CHF in rats post-MI. For the arterial baroreflex, these results with chronic blockade extend the ones with acute blockade by DiBona et al. (5).

CHF post-MI, sympathetic hyperactivity, and arterial baroreflex function. In rats with infarct size >30% after left coronary artery ligation, the modest decrease in MAP, increase in CVP, and the clear increase in RV weight are consistent with the development of moderate CHF by 8 wk post-MI and to a lesser extent at 4 wk (13, 15, 24). In the present study, rats with CHF post-MI showed enhanced sympathoexcitatory responses to air-jet stress and enhanced sympathoinhibitory responses to the ₂-adrenergic-receptor-agonist guanabenz. The latter is consistent with decreased sympathoinhibition resulting in upregulation and/or decreased receptor occupancy of ₂-adrenergic receptors in the anterior hypothalamus and thereby enhanced responses to an ₂-receptor agonist (30). The arterial baroreflex control of both RSNA and HR was clearly impaired in the CHF-Veh groups compared with those in the Sham group. These results are consistent with previous findings by us and others showing that the development of CHF post-MI is early on already characterized by decreased sympathoinhibition and increased sympathoexcitation as well as impairment of arterial baroreflex function (6, 7, 13, 15).

Brain RAS, sympathetic hyperactivity, and impairment of arterial baroreflex function in CHF. ANG II is involved in the regulation of sympathetic activity and baroreflex function (26). In the central nervous system, ANG II can cause dose-related increases of MAP, HR, and RSNA (11) and can desensitize the arterial baroreflex (28). Peripheral administration of ANG II can also increase MAP and sympathetic activity and decrease baroreflex function (1, 2, 16). The possible sites where ANG II may act include the area postrema (17), nucleus tractus solitarii (3), and caudal and rostral regions of the ventrolateral medulla in the brain (29).

The peripheral RAS may be activated in humans and animals with CHF (5, 22, 26). Several studies have shown that blockade of the RAS by an angiotensin-converting enzyme inhibitor or an AT₁-receptor blocker can lower sympathetic activity and improve baroreflex control of HR and RSNA in CHF (5, 9, 19-21). However, except for one report (5) in which losartan was given by acute intracerebroventricular injection, in all other studies the blockers were given orally or intravenously. Because peripheral (hemodynamic) effects may also contribute to an improvement of baroreflex function and lowering of sympathetic activity (9, 19, 21), these studies do not separate central from peripheral actions. In the present study, losartan was infused centrally at a dose that does not induce changes in systemic hemodynamics when administered peripherally (Table 2). Infarct size was similar in the CHF groups, and central or peripheral treatment with losartan did not cause significant changes in basal cardiovascular parameters (Table 2). Central infusion of losartan at 1 mg · kg¹ · day¹ largely normalized the sympathoexcitatory responses to air-jet stress, sympathoinhibitory responses to the ₂-adrenoceptor-agonist guanabenz, and arterial baroreflex function in rats with CHF post-MI. Whether higher doses of losartan would more completely block the brain RAS and induce additional changes cannot be excluded. Peripheral administration of losartan in the same dose of 1 mg · kg¹ · day¹ was ineffective in improving these parameters of sympathetic hyperactivity. Moreover, in control rats, central losartan did not affect responses to air-jet stress, guanabenz, or arterial baroreflex function, indicating that, under physiological circumstances, the brain RAS does not play a detectable role. Thus the effects of losartan in rats with CHF post-MI appear to be related to central AT₁-receptor blockade, and the central RAS appears to play a major role in the chronic sympathetic hyperactivity and impairment of baroreflex function in CHF post-MI. Previously, DiBona et al. (5) reported that acute intracerebroventricular injection of losartan improved baroreflex control of RSNA and lowered basal RSNA. The present results show that these effects persist during chronic treatment, indicating first that ANG II is essential not only for the acute regulation but also the chronic regulation and second that no compensatory mechanisms exist/develop that can lead to decreased baroreflex function in CHF. MAP midpoints were lower in rats with CHF versus sham-operated rats, and intracerebroventricular losartan did not prevent the decrease in midpoint. The brain RAS appears therefore to contribute to the decrease in gain and range of baroreflex function in rats with CHF but not the resetting of the baroreflex curves to lower BP.

Peripheral administration of losartan at the dose of 1 mg · kg¹ · day¹ subcutaneously did not affect the enhanced sympathoexcitatory responses to air-jet stress and enhanced sympathoinhibitory responses to guanabenz but tended (not significant) to cause some improvement in arterial baroreflex function. The latter may reflect a contribution of peripheral ANG II unmasked by this low dose of losartan or some central effects of this peripheral dose of losartan. Nonetheless, it is obvious from these findings that the pattern of changes in sympathetic control caused by the central infusion of losartan cannot be explained by leakage of the centrally administered losartan in the peripheral circulation.

Brain "ouabain" and brain RAS in CHF post-MI. As described in the introduction, the sympathetic hyperactivity and impairment of baroreflex function in CHF are associated with increases in brain "ouabain," and blockade of brain "ouabain" prevents sympathetic hyperactivity and impairment of baroreflex function (13, 15). Blockade of brain AT₁ receptors either acutely (12) or chronically (Table 3) prevents sympathetic and pressor responses to acute intracerebroventricular injection of both ouabain and ANG II, whereas the ouabain antibody blocks the responses to intracerebroventricular ouabain but does not block the responses to intracerebroventricular ANG II (12). In normotensive and hypertensive rats, the responses to brain "ouabain" appear to be mediated via the brain RAS (11, 12). In rats with CHF, blockade of both brain "ouabain" and of brain RAS prevents the sympathetic hyperactivity and impairment of baroreflex function. We therefore propose that, in CHF, activation of brain "ouabain" also leads to activation of the brain RAS and, subsequently, sympathetic hyperactivity and impairment of arterial baroreflex function.

Clinical relevance. In addition to improving hemodynamics, losartan also inhibits neurohumoral activation associated with CHF, e.g., plasma catecholamines (4, 8). The results from the present study suggest that such effects of losartan by oral administration can result from not only peripheral but also central actions and that its central effects may not depend on peripheral effects. Further studies in both animals and humans need to assess the dose-response relationship for chronic oral treatment with losartan regarding peripheral versus central actions. This relationship will likely differ for different AT₁-receptor blockers. In addition, it remains to be established whether blockade of both the peripheral and the brain RAS results in better outcome compared with only peripheral blockade.

In summary, in the present study, rats with CHF post-MI showed sympathetic hyperactivity, as assessed by responses to air-jet stress and intracerebroventricular injection of guanabenz, and attenuated arterial baroreflex function. The development of the decrease in sympathetic inhibition, the increase in sympathetic excitation, and the impairment of arterial baroreflex function in CHF appear to depend on the brain RAS, since this whole pattern of changes can be normalized by chronic central AT₁-receptor blockade with losartan.