The present study tested the hypothesis that nitric oxide (NO) contributes to impaired baroreflex gain of pregnancy and that this action is enhanced by angiotensin II. To test these hypotheses, we quantified baroreflex control of heart rate in nonpregnant and pregnant conscious rabbits before and after: 1) blockade of NO synthase (NOS) with Nω-nitro-l-arginine (20 mg/kg iv); 2) blockade of the angiotensin II AT1 receptor with L-158,809 (5 μg·kg−1·min−1 iv); 3) infusion of angiotensin II (1 ng·kg−1·min−1 nonpregnant, 1.6–4 ng·kg−1·min−1 pregnant iv); 4) combined blockade of angiotensin II AT1 receptors and NOS; and 5) combined infusion of angiotensin II and blockade of NOS. To determine the potential role of brain neuronal NOS (nNOS), mRNA and protein levels were measured in the paraventricular nucleus, nucleus of the solitary tract, caudal ventrolateral medulla, and rostral ventrolateral medulla in pregnant and nonpregnant rabbits. The decrease in baroreflex gain observed in pregnant rabbits (from 23.3 ± 3.6 to 7.1 ± 0.9 beats·min−1·mmHg−1, P < 0.05) was not reversed by NOS blockade (to 8.3 ± 2.5 beats·min−1·mmHg−1), angiotensin II blockade (to 5.0 ± 1.1 beats·min−1·mmHg−1), or combined blockade (to 12.3 ± 4.8 beats·min−1·mmHg−1). Angiotensin II infusion with (to 5.7 ± 1.0 beats·min−1·mmHg−1) or without (to 8.4 ± 2.4 beats·min−1·mmHg−1) NOS blockade also failed to improve baroreflex gain in pregnant or nonpregnant rabbits. In addition, nNOS mRNA and protein levels in cardiovascular brain regions were not different between nonpregnant and pregnant rabbits. Therefore, we conclude that NO, either alone or via an interaction with angiotensin II, is not responsible for decrease in baroreflex gain during pregnancy.
- conscious rabbits
- neuronal nitric oxide synthase
- AT1-receptor blockade
pregnancy decreases baroreflex gain, which results in a reduced ability to maintain blood pressure in response to hypotensive challenges (for reviews, see Refs. 10 and 26). In particular, reduced baroreflex function appears to underlie the increased incidence of orthostatic hypotension in pregnant women (20) and the decreased tolerance of pregnant animals to hemorrhage (11). Indeed, 17% of human maternal mortality in the United States between 1990 and 1999 was due to uncontrolled hemorrhage (13). Despite this clear clinical significance, the mechanism for the suppression of baroreflex gain remains unclear.
One potential contributor is nitric oxide (NO). Indirect evidence to support this possibility is that NO is increased in pregnancy (for reviews, see Refs. 37, 54, and 58). Moreover, in conscious animals, NO can decrease baroreflex gain, as systemic (32) and central (35) NO synthase (NOS) blockade has been shown to improve baroreflex gain. In addition, mRNA levels of the neuronal isoform of the NOS enzyme (nNOS) are normally found in multiple brain regions important in cardiovascular control, including the paraventricular nucleus of the hypothalamus (PVN), nucleus of the solitary tract (NTS) (27), caudal ventrolateral medulla (CVLM), and rostral ventrolateral medulla (RVLM) (44), and NO can influence baroreflex function by acting in these regions (18, 25, 34, 45, 49, 55, 62). Therefore, we tested the hypothesis that during pregnancy NO contributes to impaired baroreflex gain by determining, first, whether acute systemic blockade of the NOS enzyme increases the gain of the baroreflex control of heart rate in conscious pregnant rabbits. We also determined whether the mRNA and protein levels of nNOS in the NTS, RVLM, CVLM, and PVN were altered in pregnant compared with nonpregnant rabbits.
Recent studies suggest that the actions of NO in setting basal sympathetic tone and baroreflex control depend on an interaction with angiotensin II (28, 30, 33, 36, 45). For example, Kumagai et al. (28) found that, in spontaneously hypertensive rats, blockade of either angiotensin II AT1 receptors or the NOS enzyme alone increased the gain of baroreflex control of heart rate and renal sympathetic nerve activity. However, the effects of combined AT1 receptor and NOS blockade were not additive, and, in the case of the baroreflex control of renal sympathetic nerve activity, they occluded each other, suggesting interdependence.
Pregnancy is also associated with elevated angiotensin II levels (9, 11, 53), but whether angiotensin II interacts with NO in the control of baroreflex gain during pregnancy is unknown. Therefore, a third aim of this study was to test the hypothesis that the ability of NO to decrease baroreflex gain during pregnancy is enhanced by angiotensin II. To test this hypothesis, we determined the effect of NOS blockade following AT1-receptor blockade. If angiotensin II enhances the actions of NO, then, following AT1-receptor blockade, NOS blockade should not increase the gain of the baroreflex control of heart rate to the same extent as blockade of NOS alone.
All studies were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of Oregon Health & Science University.
Female New Zealand White rabbits (Western Oregon Rabbit; Philomath, OR) weighing 4.0 ± 0.1 kg (nonpregnant, n = 16) were used for these experiments. The rabbits were received when they were 14 wk old and were allowed a minimum of 6 days to acclimate before surgery.
Surgery was performed to implant nonoccluding abdominal aortic and vena cava catheters as previously described (24) for the measurement of mean arterial pressure and heart rate and the infusion of drugs, respectively. Briefly, the animals were initially anesthetized with a cocktail containing ketamine (58.8 mg/kg), xylazine (5.9 mg/kg), and acepromazine (1.2 mg/kg) administered subcutaneously. The rabbits were then intubated, placed on a respirator, and ventilated with 100% oxygen throughout the surgery. A surgical plane of anesthesia was maintained by administration of ketamine intravenously as needed. A midline abdominal incision was made in all rabbits, and polyethylene catheters with Silastic tips were implanted in the abdominal aorta (one catheter) and vena cava (two catheters). The catheters were tunneled subcutaneously from the abdominal cavity and exited at the nape of the neck.
As this study sought to determine the role of NO on the baroreflex, use of a NO donor, such as nitroprusside, to decrease mean arterial pressure could pose a confound. Therefore, after a minimum 2-wk recovery period, a second surgery was performed to implant a vena cava occluder to decrease blood pressure in tests of baroreflex function. Rabbits were initially anesthetized with ketamine (250 mg sc), intubated, placed on a respirator, and ventilated with 100% oxygen; a surgical plane of anesthesia was maintained with isoflurane (2%). An occluder was implanted around the thoracic inferior vena cava via a right thoracotomy through the fifth intercostal space. The end of the occluder also exited at the nape of the neck, and the incision was closed in layers. The end of the occluder and catheters were protected in a plastic pillbox sutured to the rabbits' skin.
The rabbits were given an intramuscular injection of enrofloxacin (22.7 mg) just before each surgery and an intravenous injection of this antibiotic for 4 days following surgery (22.7 mg/day). The animals were also injected with buprenorphine hydrochloride (0.09 mg sc) 2 to 3 h after surgery, and again the next day to relieve pain. The catheters were flushed immediately after surgery and then flushed three times weekly with sterile 0.9% saline and filled with heparin (1,000 U/ml) to maintain patency. Animals were allowed at least 2 wk for recovery from the thoracotomy before any experiments were performed. During this time, the rabbits were conditioned to the 37 × 17 × 21-cm black Plexiglas box used during experiments.
Baroreflex curve generation.
On the experimental day, the rabbits were placed in the box and allowed ∼30 min to acclimate. Mean arterial pressure and heart rate were measured continuously via the aortic catheter by using a Statham pressure transducer, a Grass tachometer, Grass polygraph, and a Biopac (Santa Barbara, CA) MP100 data acquisition and analysis system. To determine the baroreflex relationship between mean arterial pressure and heart rate, and to ensure adequate time for the heart to respond to changes in the activity of both the sympathetic and parasympathetic nervous systems (16), a slow ramp method of altering mean arterial pressure was performed. A slow inflation of the occluder around the vena cava was used to lower arterial pressure until heart rate reached maximum values. Mean arterial pressure was lowered at an overall rate of 0.47 ± 0.02 mmHg/s, and this part of the baroreflex curve was generated in 64 ± 2 s. After recovery, phenylephrine was infused in increasing doses (0.5, 1, 2, 4, and 8 μg·kg−1·min−1 in 5% dextrose in water vehicle iv); the dose was doubled every 15 s until the increment in mean arterial pressure was ∼20 mmHg over basal (5). Beyond this, arrhythmias develop and a steep decrease in heart rate is observed that deviates from the lower plateau (59). This took 59 ± 2 s so that the entire curve was generated in 123 ± 3 s. More than one occluder inflation was performed in each experiment, and at least 10 min were allowed between inflations to allow mean arterial pressure and heart rate to return to basal conditions. A total of 12 rabbits was used for baroreflex studies; many of the rabbits were studied in more than one of the protocols described below.
Effects of NOS blockade.
To test the hypothesis that pregnancy impairs baroreflex gain via actions of increased NO, we determined whether NOS blockade normalizes baroreflex gain in pregnant rabbits. In six rabbits, baroreflex curves were generated before and 105–134 min (depending on the occluder inflation chosen) after a bolus intravenous injection of the nonspecific NOS inhibitor Nω-nitro-l-arginine (l-NNA; 20 mg/kg dissolved in 10 ml of 0.5% sodium carbonate in water and then adjusted to pH 7.4) This dose produces maximal increases in mean arterial pressure and heart rate in rabbits (6). Following this experiment, rabbits were mated and the effect of l-NNA on baroreflex function was studied again at the end of gestation (day 30; term = 31 days).
Effects of NOS and AT1-angiotensin II-receptor blockade.
The purpose of this protocol was to determine whether the actions of NO require angiotensin II. Animals were studied in either the nonpregnant (n = 8) or pregnant (n = 6) states. After a control baroreflex curve was generated, a continuous intravenous infusion of the AT1-angiotensin II-receptor antagonist, L-158,809 (5 μg·kg−1·min−1 in 5% dextrose in water vehicle) was begun. This dose was sufficient to prevent increases in arterial pressure (20 ± 3 mmHg) following intravenous injection of 100 ng of angiotensin II in three of the nonpregnant and three of the pregnant rabbits in this study. Thirty to sixty minutes after the start of the L-158,809 infusion a second baroreflex curve was generated and then l-NNA (20 mg/kg iv) was administered. Finally, after 105–135 min, a final baroreflex curve was generated following combined angiotensin II/NOS blockade. All pregnant animals in this protocol were studied at day 30 of their second pregnancy, and one of the nonpregnant animals was studied 2 wk following delivery of her second litter. In a previous study (5), baroreflex curves were similar in rabbits during a second pregnancy (baroreflex gain = 12 ± 3 beats·min−1· mmHg−1, n = 5) as during the first pregnancy (baroreflex gain = 13 ± 1 beats·min−1· mmHg−1, n = 3) relative to nonpregnant values (24 ± 3 and 19 ± 3, respectively).
Effects of NOS blockade during angiotensin II infusion.
Since NOS blockade suppresses renin release (22), angiotensin II levels may be reduced in rabbits receiving l-NNA. If the actions of NO require angiotensin II, this suppression could confound the interpretation of the effects of blockade of NOS alone. Therefore, it was determined whether NOS blockade increases baroreflex gain when low doses of angiotensin II were infused to maintain basal levels. Rabbits (n = 4) were studied in the nonpregnant state and during their first pregnancy (at day 30 of gestation). After a control baroreflex curve was generated, a continuous intravenous infusion of angiotensin II was begun. For nonpregnant rabbits, angiotensin II was infused at a rate of 1 ng·kg−1·min−1 (in 5% dextose in water vehicle), which induced a small but measurable increase in mean arterial pressure (∼5 mmHg). During pregnancy, angiotensin II levels are higher (11) than in nonpregnant animals, and more angiotensin II is required to increase blood pressure. To ensure similar effectiveness, angiotensin II was infused at a dose sufficient to again increase pressure by ∼5 mmHg (1.6 to 4 ng·kg−1·min−1).
Levels of nNOS protein and mRNA in brain regions involved in cardiovascular control.
This protocol was employed to determine whether alterations in the balance of nNOS expression in different cardiovascular brain nuclei might be involved in the decreased baroreflex gain of pregnancy. For these experiments, rabbits (n = 4) at day 30 of gestation and age-matched nonpregnant rabbits were euthanized with 1 ml of Euthasol (390 mg pentobarbital, 50 mg phenytoin iv), and their brains were removed, quickly frozen in −80°C 2-methylbutane, and stored at −80°C until shipment to the University of Nebraska Medical Center where the assays were performed.
The PVN was punched from 900-μm slices. The NTS, CVLM, and RVLM sections were punched out of 1200-μm thick slices with a 15-gauge needle stub using the Palkovits technique (43). The punched tissue was put in 0.5 ml of Tri-Reagent and homogenized. Total RNA and proteins in the homogenate were each extracted according to the Tri-Reagent manufacturer's instructions.
Quantification of mRNA levels.
Semiquantitative RT-PCR assays were performed to assess relative mRNA levels. RNA was isolated, followed by a reverse transcription reaction for 40 min at 38°C in the presence of 1.5 μM random hexamers and 200 units of Moloney murine leukemia virus reverse transcriptase. Each 1.5-μl aliquot of the reverse transcriptase product was used for nNOS cDNA amplification. The following polymerase chain reaction primers were used: nNOS, 5′-GATCGCTGACCGTATGCAG-3′ (sense); 5′-GTCGTACTCCTGCTTGGTG-3′ (antisense); and β-actin, 5′-GGGAAATCGTGCGTGACATT-3′ (sense); 5′-CGGATGTCAACGTCACACTT-3′ (antisense). The polymerase chain reaction mixture contained 0.7 μM primers, deoxynucleotide triphosphates (dNTP), bovine serum albumin and 1 unit of Taq DNA polymerase. β-Actin was coamplified with each sample as an internal control. After 4 min of denaturing at 94°C, the amplification was performed at 94°C for 1 min, at 46°C for 1 min, and at 72°C for 1 min for 30 cycles. The products of this reaction (7 μl) were fractionated in a 1% agarose gel and transferred to a Nytran nylon membrane. A Southern blot analysis was then performed using [5′-32P]-antisense deoxyoligonucleotides (5′-GTCGTACTCCTGCTTGGTG-3′ for nNOS, 5′-CCGCCGATCCACAC-3′ for β-actin). The radioactive signal emitted from the cDNA was quantified by phosphor imaging, and the data are expressed as digital light units per unit time (DLU) for nNOS mRNA relative to DLU for β-actin mRNA.
Quantification of protein levels.
The Tri-Reagent protein extract was used for Western blot analysis of nNOS. The protein concentration was measured by using a protein assay kit (BCA kit; Pierce). The protein sample was mixed with an equal volume of 2 × 4% sodium dodecyl lauryl sulfate sample buffer. The sample was boiled for 5 min, and then loaded onto a 7.5% sodium dodecyl lauryl sulfate-polyacrylamide gel for electrophoresis at 40 mA/each gel for 50 min. The fractionated proteins on the gel were electrophoretically transferred onto a polyvinylidene difluoride membrane at 300 mA for 90 min. The membrane was incubated with 5% milk-Tris-buffered saline-Tween 20 solution for 30 min at room temperature and then with primary antibody (1:1,000; BD Transduction Laboratories) at 4°C overnight. After being washed three times, the membrane was incubated with secondary antibody (peroxidase conjugated, 1:5,000; Pierce) for 30 min at room temperature. After being washed three times, the membrane was treated with enhanced chemiluminescence reagent for 5 min and detected by exposing a film. The bands on the developed film were visualized and analyzed using UVP BioImaging Systems. The light signal emitted was quantified, and the values were normalized to the β-tubulin protein band to determine specific changes in nNOS levels.
Baroreflex curve analysis.
The data for the baroreflex curves were collected at 200 Hz and processed by using a Biopac (Santa Barbara, CA) MP100 data acquisition and analysis system. Raw data were grouped into 1-s bins from which mean values were obtained. Since more than one curve was generated using the vena cava occluder, curves were selected that were free from movement artifact and exhibited the best sigmoidal fit. Only one occlusion was used for each curve. Occasionally, basal mean arterial pressure before the pressor (phenylephrine) part of the reflex curve differed slightly from the depressor (occluder) part. When this occurred, to avoid erroneous measurements of baroreflex gain, half of the pressure difference was added to all pressure values in the segment with the lower basal pressure and half the difference was subtracted from the pressures in the segment with the higher pressure so that the two segments were aligned, as previously described (5, 19). Importantly, the adjustments made in nonpregnant (2.1 ± 0.4 mmHg; range, 0–5 mmHg) and pregnant (2.1 ± 0.5 mmHg; range, 0–6 mmHg) rabbits were not different. Figure 1A illustrates representative curves for one rabbit and also shows that pregnancy decreases baroreflex gain.
The sigmoidal baroreflex relationships between mean arterial pressure and heart rate generated in each experiment were fitted and compared using the Boltzmann sigmoidal equation [heart rate = A + B/1 + e(C−mean arterial pressure)/D, where A equals the minimum heart rate, B equals the heart rate range, C equals the mean arterial pressure at the midpoint between the minimum and maximum heart rate (or BP50), and D is the slope coefficient or width]. Maximum baroreflex gain was calculated by dividing the heart rate range by four times the slope coefficient (5, 8, 12, 51). Due to the exponential nature of changes in baroreflex gain, and therefore the high variability in this parameter, the log of the maximum baroreflex gain was used for statistical analysis.
Basal mean arterial pressures, heart rates, and curve-fitting parameters were compared between groups using a two-way ANOVA for repeated measures and the post hoc Bonferroni test. Student's t-test was used to detect differences in mRNA and protein levels of nNOS in different brain nuclei, differences in basal mean arterial pressures, heart rates, and curve-fitting parameters between nonpregnant and pregnant rabbits, as well as between treatment with l-NNA alone vs. l-NNA and L-158,809. In the figures, the sigmoidal curves derived from the averaged parameters are shown along with basal points (means ± SE).
Effects of pregnancy.
As previously reported (10), pregnancy decreased heart rate range, basal mean arterial pressure, the operating pressure for the baroreflex (decreased BP50), and baroreflex gain while it increased baroreflex-induced minimum and basal heart rates and the slope coefficient (Fig. 1, Tables 1–3). Also as previously observed (5), sigmoidal parameters calculated by fitting baroreflex relationships generated during the first pregnancy were not different from those during the second pregnancy. For example, baroreflex gain averaged 7.5 ± 1.2 beats·min−1·mmHg−1 during a first pregnancy and 6.5 ± 1.5 beats·min−1·mmHg−1 during a second pregnancy.
Effects of NOS blockade.
NOS blockade did not increase baroreflex gain in either pregnant or nonpregnant rabbits (Table 1, Fig. 2). However, in both states l-NNA increased basal mean arterial pressure and the operating pressure for the baroreflex, and decreased baroreflex-induced minimum and basal heart rate (Table 1, Fig. 2). In the pregnant animals, l-NNA also decreased baroreflex-induced maximum heart rate (Table 1, Fig. 2B).
Effects of AT1-angiotensin II-receptor and NOS blockade.
Treatment with L-158,809, an inhibitor of AT1 receptors, alone or in combination with NOS blockade did not improve baroreflex gain in either pregnant or nonpregnant rabbits (Table 2, Fig. 3). Blockade of AT1 receptors decreased basal blood pressure and the operating pressures of the baroreflex in both nonpregnant and pregnant rabbits (Table 2, Fig. 3). In addition, with the combined blockade of both AT1 receptors and NOS, baroreflex-induced minimum and basal heart rates decreased in nonpregnant and pregnant rabbits (Table 2), similar to the effect of NOS blockade alone. As with NOS blockade alone, the combined blockade also reduced the baroreflex-induced maximum heart rate in pregnant but not nonpregnant animals (Table 2, Fig. 3). Thus, in nonpregnant rabbits, baroreflex curve parameters were not different whether rabbits were treated with l-NNA alone or were treated with both l-NNA and L-158,809. The same was true for pregnant rabbits, except for changes in the BP50 (see below).
In pregnant animals angiotensin II AT1 blockade reduced the pressor response following l-NNA treatment. For example, basal blood pressure and baroreflex operating pressures were lower (P < 0.05) with combined l-NNA and L-158,809 treatment compared with treatment with l-NNA alone (Table 2, Fig. 3B). Moreover, while in both groups, blockade of NOS following L-158,809 administration increased basal mean arterial pressure above levels observed following treatment with the AT1-receptor antagonist alone, the increase was reduced in pregnant animals (11 ± 1 mmHg pregnant, compared with 16 ± 2 mmHg nonpregnant; P < 0.05). As a result, mean arterial pressure after combined blockade did not differ from the control state in pregnant rabbits, whereas in nonpregnant rabbits, combined blockade increased basal mean arterial pressure above control values.
Effects of NOS blockade during angiotensin II infusion.
Angiotensin II slightly increased basal mean arterial pressure and the operating pressures of the baroreflex in both nonpregnant and pregnant rabbits, but this reached significance only when the animals were pregnant (Table 3, Fig. 4). With the combined administration of angiotensin II and l-NNA, basal mean arterial pressure rose, baroreflex-induced minimum and basal heart rates fell, and the baroreflex operated at higher pressures in both nonpregnant and pregnant rabbits. These changes were similar to the effect of l-NNA alone (Table 3, Fig. 4). However, with the angiotensin II infusion, l-NNA no longer decreased the maximum heart rate in pregnant animals (Table 3, Fig. 4). Importantly, baroreflex gain was again unaffected (Table 3, Fig. 4).
Levels of nNOS protein and mRNA in cardiovascular centers of the brain.
Compared with values in paired (n = 4) nonpregnant rabbits (Table 4) neuronal NOS mRNA was not altered by pregnancy in the NTS (129 ± 25% of nonpregnant), the CVLM (74 ± 6% of nonpregnant), the RVLM (98 ± 17% of nonpregnant), nor the PVN (148 ± 46% of nonpregnant). The protein levels of nNOS were also unaltered by pregnancy in the NTS (154 ± 69% of nonpregnant), the CVLM (101 ± 10% of nonpregnant), the RVLM (122 ± 54% of nonpregnant), or the PVN (112 ± 47% of nonpregnant).
This study tested the hypothesis that NO contributes to the decreased baroreflex gain of pregnancy and that angiotensin II modulates this action. The novel findings are that 1) blockade of NO production with or without AT1-receptor blockade or angiotensin II infusion does not alter the gain of the baroreflex control of heart rate in pregnant or nonpregnant rabbits; 2) NOS blockade decreases baroreflex-induced maximum heart rate only in pregnant rabbits, and this effect is prevented by infusion of angiotensin II; 3) in pregnant rabbits after blockade of AT1 receptors, the increase in mean arterial pressure following NOS blockade is reduced; and 4) pregnancy does not alter nNOS levels in the PVN, NTS, CVLM, or RVLM of rabbits. On the basis of these findings, we conclude that 1) neither NO nor angiotensin II, alone or in combination, contributes to the impaired baroreflex gain of pregnant rabbits; 2) alterations in nNOS levels in the PVN, NTS, CVLM, or RVLM are not responsible for the decrease in baroreflex gain during pregnancy in rabbits; 3) in pregnant rabbits only, when angiotensin II actions are low, NO supports baroreflex-induced maximum heart rate; and 4) in the pregnant rabbit, NO exerts part of its vasodilatory actions through inhibition of the renin-angiotensin system.
As in previous reports, near the end of pregnancy in rabbits, basal mean arterial pressure was decreased (due to reduced vascular resistance), and basal heart rate was increased (for reviews, see Refs. 10 and 26). In response to the fall in blood pressure, the baroreflex curves shifted to the left (BP50 is reduced), presumably due to baroreflex resetting (1, 14). This resetting allows the animal to better defend the new basal blood pressure, and it is likely due to adaptation of the peripheral baroreceptors, as well as to alterations in central neurons involved in baroreflex control (1, 14). In addition, we confirmed that pregnancy decreases the gain of the baroreflex control of heart rate. Reflex gain was assessed by comparing the maximal slopes of sigmoidal relationships between arterial pressure and heart rate. Because, in the present experiments, the resting basal values of arterial pressure and heart rate were situated at the bottom of the steep segment of the sigmoidal baroreflex curve, the gain measurements were derived almost exclusively by the tachycardia responses to inflation of the vena cava balloon. Importantly, sympathoexcitatory responses to vena cava balloon inflation in rabbits are not reduced by elimination of cardiac afferents (via administration of pericardial procaine) and are abolished by sinoaortic denervation alone (19). Thus, the gain of our baroreflex function curves likely reflects responses secondary to unloading of arterial, but not cardiac, baroreceptors. Therefore, these data suggest that pregnancy impairs arterial baroreflex control of heart rate.
Although intensely studied, the cause for the fall in baroreflex gain during pregnancy has remained elusive. Because NO increases during pregnancy (37, 54, 58) and can inhibit baroreflex gain (32, 35), we hypothesized that this factor is involved. To test this hypothesis, we determined whether NOS blockade normalized baroreflex gain. Similar to previous reports in male rabbits (32) and nonpregnant and pregnant female rabbits (6), NOS blockade increased basal mean arterial pressure and decreased basal heart rate. In addition, as in male rabbits (32), nonspecific NOS inhibition increased the baroreflex operating pressure. These data indicate that endogenous NO is vasodilatory and supports basal heart rate. However, contrary to our hypothesis and the work of Liu et al. (32) in conscious male rabbits, we found no effect of systemic NOS blockade on baroreflex gain in either nonpregnant or pregnant rabbits. This lack of effect is similar to that observed by Chiu and Reid (15) also in conscious rabbits. Conflicting results have also been observed in rats in which systemic NOS blockade can either increase baroreflex gain (4, 28, 38) or cause no change in this parameter (21, 29).
Discrepancies between studies are likely explained by the stress state or the tonic levels of NO. At low levels, NO in the NTS either increases or does not affect baroreflex gain (18, 25, 34, 45, 49, 55, 62). However, when NO levels are elevated, for example, by acute psychological stress, NO decreases gain (17). Moreover, increases in NO secondary to substantial increments in angiotensin II also decrease baroreflex gain (45). Indeed, in spontaneously hypertensive rats in which the central actions of angiotensin II are clearly enhanced, NOS blockade improves gain of baroreflex control of heart rate and renal sympathetic nerve activity (28).
In contrast, our data suggest that angiotensin II-induced NO levels are not sufficiently elevated during rabbit pregnancy to significantly impair the gain of the baroreflex control of heart rate. Pretreatment with an AT1-receptor blocker did not alter the subsequent effect of NOS blockade on baroreflex gain in either pregnant or nonpregnant rabbits. Moreover, infusion of angiotensin II to maintain levels also did not influence the effects of NOS blockade. Finally, as shown previously, neither AT1-receptor blockade alone (12, 41), nor low-dose angiotensin II infusion alone affected baroreflex gain in nonpregnant or pregnant rabbits. This latter result is in contrast to the gain-lowering effect of higher angiotensin II doses (7). Collectively, these data importantly suggest that NO and/or angiotensin II must reach high levels before a decrease in gain is produced. On the other hand, in unstressed conscious pregnant and nonpregnant animals, relatively low basal NO and angiotensin II levels, either alone or via interaction, do not decrease the gain of baroreflex control of heart rate.
Our conclusion that angiotensin II and NO do not contribute significantly to the impaired baroreflex gain of pregnancy depends on documentation that adequate blockade of AT1 receptors or NOS activity was achieved. Previous work suggests that both systemic and central NOS were blocked. In a previous dose-response study, we found that the present dose was greater than that required to produce maximal pressor and bradycardia responses, indicating that NOS was maximally inactivated systemically (6). Moreover, this inhibitor is known to cross the blood-brain barrier (56), and Liu et al. (32) demonstrated that the effects of a lower dose of intravenous l-NNA on baroreflex control of heart rate in conscious rabbits is mediated by a central action. Finally, the present dose of l-NNA reversed the effect of psychological stress to decrease baroreflex gain, most likely due to an action in the brain (17). Peripheral and central AT1 receptors were also likely blocked. We confirmed systemic blockade by demonstrating that the AT1-receptor antagonist l-158,809 completely prevented the pressor response to intravenous angiotensin II injection. The presence of complete systemic AT1 blockade is pertinent, since angiotensin II may decrease gain primarily via an action from the systemic circulation to increase NO in the NTS (45) or by acting at the area postrema (31), a brain region that lacks a blood-brain barrier. Nevertheless, it is also well-established that nonpeptide AT1-receptor antagonists cross the blood-brain barrier to act centrally (2, 23, 52). Moreover, O'Hagan et al. (41) showed that losartan failed to increase baroreflex gain in pregnant rabbits, even when it was administered both intracerebroventricularly and intravenously. Thus, the inability of NOS and AT1-receptor blockade to restore baroreflex gain during pregnancy does not appear to be due to suboptimal antagonist dosages.
One novel finding was that in pregnant but not in nonpregnant rabbits, NOS blockade decreased the baroreflex-induced maximum heart rate, suggesting that NO acts during pregnancy to prevent decreases in baroreflex-induced maximum heart rate. Therefore, one role for the elevated NO levels during pregnancy may be to avert a further reduced ability to respond to hypotensive challenges. However, this effect appears to be dependent on low-circulating angiotensin II levels/actions. It was evident following blockade of NOS alone, which decreases renin release (22) and also when NOS blockade followed blockade of AT1-angiotensin II receptors. However, when systemic angiotensin II levels were held constant by infusion, NOS inhibition no longer decreased maximum heart rate. These data suggest a complex interaction between these two vasoactive substances, such that low levels of angiotensin II in plasma do not significantly alter baroreflex-induced maximum heart rate but prevent the decrease in maximum heart rate induced by loss of NO. Because infused exogenous angiotensin II cannot cross the blood-brain barrier, we speculate that angiotensin II likely acts at a circumventricular organ.
As we have observed previously (5, 8), the baroreflex-mediated minimum heart rate induced by increases in arterial pressure was elevated when the rabbits were pregnant. This reduced ability to suppress heart rate is due to an attenuation of parasympathetic control of the heart (8); however, the mechanism of impaired vagal control is unknown. NO (generated via nNOS) restrains vagal activity via a central action (32, 40, 48). Therefore, one hypothesis is that the increases in basal and minimum heart rates are mediated by increased brain actions of NO. However, while systemic NOS blockade suppressed basal and minimum heart rate in both nonpregnant and pregnant rabbits, as previously shown in male rabbits (32, 40), these values remained elevated in the pregnant animals after NOS blockade. Therefore, we conclude that the pregnancy-induced increases in basal and minimum heart rate are independent of increased actions of NO.
After AT1-receptor blockade, the ability of NOS blockade to increase mean arterial pressure was reduced in pregnant rabbits compared with nonpregnant rabbits. This effect is consistent with previous work showing that NO feeds back during pregnancy to attenuate the pressor effects of angiotensin II (39, 57). Angiotensin II, which is increased during pregnancy (9, 11, 53), stimulates NO production in vascular endothelial cells (46, 47, 50). Subsequently, NO counteracts the effects of angiotensin II; it limits the vasoconstrictive actions of angiotensin II by causing vasodilation (3, 46). Therefore, one way in which l-NNA increases mean arterial pressure during pregnancy is by increasing the pressor effects of angiotensin II.
Consistent with our finding that NOS blockade did not normalize baroreflex gain or minimum heart rate in pregnant rabbits, we found no significant differences in either gene expression or protein levels of nNOS in tissue punches of the PVN, NTS, CVLM, or RVLM between rabbits in these two states. Our data are in conflict with previous studies demonstrating elevations in nNOS levels in the whole hypothalamus of pregnant rats (61) and both decreases (42) and increases (60) in NO activity in the magnocellular region of the PVN, which is responsible for oxytocin and vasopressin production and release. In addition, Woodside and Amir (60) found higher levels of NOS activity in the parvocellular region of the PVN of the pregnant rat. Thus, our failure to detect significant differences may be due to opposing changes within the same nucleus (42, 60). Similarly, other isoforms of the NOS enzyme (eNOS or inducible NOS) may be differentially regulated during pregnancy, which could explain the failure of systemic nonspecific NOS inhibition to increase baroreflex gain. Therefore, while it remains possible that opposing changes in NOS and NO are induced by pregnancy, we conclude that net increases in NO production, and regional changes in nNOS levels in the PVN, NTS, CVLM, or RVLM, are not responsible for the decreased the gain of baroreflex control of heart rate.
If neither angiotensin II nor NO is required, then what mechanism mediates the baroreflex impairment induced by pregnancy? In our companion paper (17a), we test the novel hypothesis that the insulin resistance that appears late in pregnancy is involved. Our data support this hypothesis and demonstrate that the degree of baroreflex impairment is highly correlated both temporally and quantitatively to the development of insulin resistance. In addition, pregnancy-induced decreases in baroreflex gain are reversed when pregnant rabbits are treated with the insulin-sensitizing drug, rosiglitazone. While the exact mechanism by which insulin resistance attenuates baroreflex function is unknown, the results of the present paper suggest that neither NO nor angiotensin II is involved.
In summary, near-term pregnant rabbits exhibit decreases in basal mean arterial pressure and the gain of the baroreflex control of heart rate, increases in basal heart rate and baroreflex-induced minimum heart rate, and lower baroreflex operating pressures. Blockade of NO synthesis is without effect on baroreflex gain in either the nonpregnant and pregnant rabbit. Neither blockade of angiotensin II AT1 receptors, nor angiotensin II infusion significantly altered this lack of effect of NOS blockade. Consistent with these results, we found that levels of nNOS mRNA and protein are also unaltered by pregnancy in the PVN, NTS, CVLM, and RVLM. Taken together, these results suggest that substantial elevations in NO and angiotensin II are required for a baroreflex gain-reducing action, and that NO, alone or via interaction with angiotensin II, does not contribute to the decreased gain of the baroreflex control of heart rate during pregnancy.
This work was supported in part by National Heart, Lung, and Blood Institute Grants HL-35872 (to V. L. Brooks), HL-70962 (to V. L. Brooks), and PO-1-HL-062222 (to I. H. Zucker) and by American Heart Association, Pacific Mountain Affiliate Fellowship Grant 0315254Z (to D. L. Daubert).
We thank George D. Giraud and Korrina L. Freeman for technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society