Ghrelin, a neuropeptide originally known for its growth hormone-releasing and orexigenic properties, exerts important pleiotropic effects on the cardiovascular system. Growing evidence suggests that these effects are mediated by the sympathetic nervous system. The present study aimed at elucidating the acute effect of ghrelin on sympathetic outflow to the muscle vascular bed (muscle sympathetic nerve activity, MSNA) and on baroreflex-mediated arterial blood pressure (BP) regulation in healthy humans. In a randomized double-blind cross-over design, 12 lean young men were treated with a single dose of either ghrelin 2 μg/kg iv or placebo (isotonic saline). MSNA, heart rate (HR), and BP were recorded continuously from 30 min before until 90 min after substance administration. Sensitivity of arterial baroreflex was repeatedly tested by injection of vasoactive substances based on the modified Oxford protocol. Early, i.e., during the initial 30 min after ghrelin injection, BP significantly decreased together with a transient increase of MSNA and HR. In the course of the experiment (>30 min), BP approached placebo level, while MSNA and HR were significantly lower compared with placebo. The sensitivity of vascular arterial baroreflex significantly increased at 30–60 min after intravenous ghrelin compared with placebo, while HR response to vasoactive drugs was unaltered. Our findings suggest two distinct phases of ghrelin action: In the immediate phase, BP is decreased presumably due to its vasodilating effects, which trigger baroreflex-mediated counter-regulation with increases of HR and MSNA. In the delayed phase, central nervous sympathetic activity is suppressed, accompanied by an increase of baroreflex sensitivity.
- muscle sympathetic nerve activity
ghrelin is a peptide hormone mainly secreted from the stomach and small intestine with rapidly expanding experimental and clinical relevance. Although its growth hormone (GH) secretagogue and energy-homeostatic properties are rather well characterized (14, 15), the effects of ghrelin on sympathetic nervous regulation of cardiovascular function are largely unexplored. The potential therapeutic use of ghrelin (3), its analogs, and antagonists highlights the importance of further understanding its physiology. Thus, ghrelin proved beneficial for the treatment of cachexia due to malignancy (6), chronic heart failure (22), or chronic obstructive pulmonary disease (19). Moreover, cardioprotective and anti-inflammatory properties were proclaimed in experimental models of myocardial infarction (26, 27), inflammatory bowel disease (7), and sepsis (35). Previous results from human (20, 21) and animal studies (18, 26, 27) strongly suggest that ghrelin acts at central nervous baroreflex centers to reduce sympathetic activity, thereby decreasing blood pressure (BP). On the other side, ghrelin has specific binding sites within the peripheral vasculature (31) and was suggested to additionally exert direct vasodilating effects.
Recently Lambert et al. (16) observed a blood pressure (BP) decreasing effect of continuous intravenous infusion of ghrelin for 1 h in human volunteers. This BP drop was combined with a significant increase of vasoconstrictive sympathetic outflow to the muscle vascular bed to counteract hypotension. Heart rate, however, remained unchanged compared with placebo, and sympathetic and hemodynamic responses to mental stress were blunted, suggesting a combination of peripheral vasodilating and central sympathoneural effects of ghrelin.
Our present placebo-controlled study aims to elucidate the effects of ghrelin on sympathetic nervous BP control in healthy humans. We hypothesize that 1) ghrelin centrally decreases muscle sympathetic nerve activity (MSNA) and BP and that 2) ghrelin modulates the sensitivity of baroreflex loops. Given the pulsatile nature of physiological ghrelin secretion (4), we chose a placebo-controlled experimental design in which the effects of a single intravenous dose of ghrelin on MSNA, as well as hemodynamic and endocrine parameters were evaluated.
Twelve healthy lean male volunteers (body mass index: 20–25 kg/m2), aged between 22 and 30 years participated in the experiments. Participants were nonsmokers and free of any medication. They were asked to abstain from alcohol and caffeinated beverages for 24 h and to have their last meal in the evening prior to the experiments. In a randomized double-blind placebo-controlled crossover design, subjects either received a bolus of human ghrelin or isotonic saline. The study was conducted on two experimental days, i.e., ghrelin or placebo, separated by at least 1 wk. Each subject was assigned to a sequence of experiments randomly. The study was approved by the local ethics committee, and all participants gave their written informed consent.
Experiments were performed in our neurophysiological laboratory starting at 7:30 A.M. and lasting until 12:30 P.M. Participants were investigated in a comfortable supine position and equipped for continuous ECG and BP monitoring (Finometer, Finapres Medical Systems, Amsterdam, Netherlands). In addition, BP was measured oscillometrically (Vital Signs Monitor 300, Welch Allyn, Skaneateles Falls, NY), and Finometer BP was calibrated according to the oscillometric readings. An intravenous cannula was inserted into an antecubital vein for repeated blood sampling. A second cannula was placed into the antecubital vein of the opposite arm for bolus administration of either ghrelin (2 μg/kg body wt, Clinalfa basic, Bachem, Bubendorf, Switzerland) (19, 22) or isotonic saline (placebo), as well as application of vasoactive substances for baroreflex testing. The respective substance was prepared by a team member not involved in the current session in a lab room apart and was provided in a neutral syringe.
MSNA was recorded from the peroneal nerve using tungsten microelectrodes as reported previously (8). In brief, the recording electrode was percutaneously inserted into a sympathetic fascicle of the peroneal muscle nerve. A reference electrode was positioned subcutaneously at a distance of 2 to 3 cm. Signals were amplified, filtered, and passed through an amplitude discriminator to obtain a mean voltage display of the multiunit nerve activity. Technical details and evidence that the recorded activity is of sympathetic origin have been published before (33
Recordings of hemodynamic parameters and MSNA were started about 30 min (t −30 min) prior to the bolus injection of either ghrelin or isotonic saline (t +0 min) and were continued throughout the experiment. To determine the baroreflex setpoint, MSNA, BP, and heart rate (HR) were repeatedly measured during 5–10 min periods of unaffected rest (baseline), according to the protocol presented in Fig. 1. In brief, the first baseline period was timed before injection of the test substance (10-min period starting at t −25 min, preinjection baseline). Similar recording periods of baseline MSNA, BP, and HR were placed immediately (t +2 and t +7 min, early postinjection baseline 1 and 2) and later after administration of the respective test substance (t +85 min, delayed postinjection baseline). To document constant MSNA recording quality, microelectrodes remained in an intraneural position during the entire experiment. Subjective sensations were repeatedly assessed with a simple questionnaire.
Additionally, baroreflex sensitivity was assessed using a novel simplified approach combining the steady-state vasoactive-drug protocol used by our group before (24) with the modified Oxford protocol (MOP) (23). In brief, each cycle of baroreflex assessment consisted of three parts: 1) a brief resting period, followed by 2) an intravenous bolus of 150 μg sodium-nitroprusside, a short-acting direct vasodilator, and 3) an intravenous bolus of 150 μg phenylephrine, a short-acting vasoconstrictor, 60 s later. Cycles for pharmacologic baroreflex assessment were performed before (t −5 min) and after administration of the test substance (t +15 min, t +35 min, and t +65 min; see Fig. 1, MOP 1–4).
Blood samples were taken at t −10 min, t +10 min, t +30 min, t +60 min, and at the end of the experiment (t +110 min). They were centrifuged and stored at −80°C for analysis of ghrelin (total ghrelin RIA kit; Millipore, Billerica, MA), renin (renin III generation, Cisbio, Codolet, France), catecholamines (HPLC; Chromsystems, Munich, Germany), GH, ACTH, cortisol (all IMMULITE, Siemens, Llanberis, United Kingdom) and copeptin (Thermo Fisher Scientific Brahms, Henningsdorf, Germany).
Because of physiological considerations discussed in detail below, we distinguished the preinjection phase from an immediate (<30 min) and a delayed (>30 min) response phase after injection of ghrelin or placebo, respectively. Previous studies have shown distinct hemodynamic and endocrine changes within this time scale (20). BP, HR, and MSNA were analyzed during preinjection baseline (t −25 min), immediate (t +2 min and t +7 min), and delayed postinjection baseline (t +85 min).
For assessment of baroreflex sensitivity, the MOP cycles were analyzed. MSNA, HR, and mean arterial BP were measured for 60-s intervals: 1) before injection of vasoactive substances, 2) from 30 to 90 s after bolus injection of 150 μg sodium nitroprusside and 3) from 90 to 150 s after injection of 150 μg of phenylephrine. For each cycle of measurement, BP was plotted against MSNA or HR (vascular or cardiac baroreflex, respectively), and a linear regression analysis was performed using the ordinary least-squares method. A cycle of measurement was accepted when the coefficient of determination R2 was equal to or greater than 0.7. Resulting slopes of model linear functions were used to characterize baroreflex sensitivity and subjected to further statistical analysis. An example of MSNA recording during baseline, as well as during baroreflex challenge with nitroprusside and phenylephrine, is shown in Fig. 2.
Data are expressed as means ± SE. Statistical analysis of data was based on ANOVA with group factor treatment (ghrelin vs. placebo) and the repeated-measures factor time. A Greenhouse-Geisser corrected P value < 0.05 was considered significant. Post hoc testing was performed using paired Student's t-test as appropriate (PASW Statistics 18). In addition to comparison between both conditions, i.e., ghrelin vs. placebo, within-condition comparison was calculated for the recording periods defined above, prior to vs. after respective bolus injection. Sample size calculations (power and effect size) were performed with G·Power 3.1.3 (Franz Faul, University of Kiel, Germany).
Intravenous administration of ghrelin was generally well tolerated. Under the ghrelin condition, subjects reported sensations of hunger (n = 6), flush (n = 3), dizziness (n = 2), sleepiness (n = 3) and borborygmi (n = 4). Sweating occurred in five subjects. Subjects' ratings of “physical discomfort” were transiently increased during the early phase after ghrelin injection. All ghrelin-related sensations resolved within the first 30 min after administration, except hunger, which persisted for the whole experiment. Under the placebo condition, one subject reported sleepiness.
Hormonal response to ghrelin.
In the placebo condition, plasma levels of total ghrelin remained at preinjection level throughout the experiment. Following ghrelin injection, however, plasma levels increased about 18-fold after 10 min, followed by an exponential decline. As expected, GH significantly increased with short latency (peak at t +30 min, 15-fold compared with preinjection), while a slight decrease of GH was observed under placebo. In accordance with previous studies (20), administration of ghrelin was answered by a highly significant increase in ACTH, cortisol, and epinephrine, but not norepinephrine. Moreover, levels of copeptin, an unspecific marker of circulatory stress, were increased, and plasma renin tended to be higher early after intravenous ghrelin (peak at t +30 min). (See Table 2.)
Blood pressure, heart rate, and MSNA.
BP and HR were analyzed for all 12 participants, while MSNA data were complete for 10, but incomplete for two subjects due to loss of electrode position from the sympathetic fascicle at delayed ghrelin phase. Results are reported in detail in Table 1 and Fig. 3, A–C).
For the whole experiment, ANOVA for interaction between group factor ghrelin/placebo and time of measurement was significant concerning MSNA (P = 0.037) and HR (P < 0.001). It indicated a strong trend regarding BP (P = 0.057).
In detail, resting BP, HR, and MSNA did not differ between both conditions during the preinjection period (t −25 min, t −5 min). After ghrelin injection (t +15 min and subsequent), mean arterial BP was persistently decreased compared with the preinjection period. Compared with the corresponding placebo data, BP was significantly lower early after ghrelin (t +7 to +35 min; P = 0.014, ANOVA for factor ghrelin/placebo) but approached placebo level later (t +65 min and subsequent). HR transiently increased immediately after ghrelin administration (t +2 min: P = 0.018 compared with preinjection (t −25 min); P = 0.117 compared with placebo, t-test). In the delayed postinjection phase (t +35 min and subsequent), HR significantly decreased after ghrelin but increased in the placebo condition compared with preinjection data. Therefore, HR was significantly lower after ghrelin compared with placebo in the delayed phase (P = 0.043, ANOVA for factor ghrelin/placebo; Fig. 3B). MSNA significantly increased immediately after the ghrelin bolus, but not after placebo compared with preinjection levels (t +2 min vs. t −25 min: P = 0.005, t-test). Though this immediate increase persisted for more than 15 min, it did not reach significance compared with the corresponding placebo data (t +2 min: P = 0.088, t-test) and was less than expected with regard to the prevailing BP decrease. After ghrelin, during the delayed phase, MSNA had returned to preinjection levels until the end of the experiment (t +85 min). After placebo, in contrast, MSNA remained constant in the immediate postinjection phase, but increased during the later periods to become significantly higher at t +85 min compared with ghrelin (28.4 vs. 32.7 bpm, P = 0.047, t-test) or to placebo preinjection levels (P = 0.016, t-test).
To identify differential effects on baroreflex performance, vasoactive substances were administered (MOP cycles), and slopes of baroreflex were derived from linear regression, as illustrated in Fig. 4. Preinjection (t −5 min) slopes were comparable under both conditions. Because of poor correlation, immediate postinjection phase data (t +15 min) did not meet the statistical requirements mentioned above. In the delayed postinjection phase (t +35 min and t +65 min, pooled data) sensitivity of MSNA-related vascular baroreflex showed a significant increase following intravenous ghrelin (mean slope −2.25 vs. −1.66 after intravenous ghrelin or placebo, respectively; ANOVA: P = 0.036 for interaction between factors ghrelin/placebo and time of measurement, n = 10). Statistical analysis of sensitivity of HR-related cardiac baroreflex did not show any significant difference between ghrelin and placebo condition (ANOVA: n.s.; n = 8). However, in contrast to MSNA-related baroreflex data, this aspect was underpowered, according to post hoc power analysis.
Immediate and delayed phase of ghrelin action.
In the present study, two distinct phases of ghrelin action could be distinguished over the experimental time course: immediately after ghrelin injection, a decrease of BP combined with an increase of MSNA and a transient increase of HR was observed (immediate ghrelin phase). After 30 min, MSNA had returned to placebo level, while BP and HR were decreased. Finally (at t +85 min), blood pressure reached the placebo level, while HR remained decreased and MSNA was significantly reduced compared with placebo (delayed ghrelin phase).
According to the broadly accepted baroreflex paradigm, a reduced BP should trigger the arterial baroreflex to increase sympathetic outflow toward the heart and vasculature to restore the preset BP level. Vice versa, a BP above the setpoint level should be followed by a decrease of MSNA and HR, unless the setpoint itself was altered (5, 11). Therefore, our findings in the immediate ghrelin phase can be explained by a direct vasodilation leading to an immediate blood pressure decrease, which results in a compensatory increase of MSNA and a short transient increase of HR mediated by the arterial baroreflex. Ghrelin is known to decrease systemic vascular resistance (21), which might result from NO-dependent (36) or -independent (34) vasodilatory mechanisms. However, the reflex-increase of resting MSNA at this phase was modest with regard to the significant BP decrease (e.g., t +15 min; Table 1). In previous studies, NO-related BP decreases of similar degree yielded much stronger sympathoexcitation (25, 30). This observation might indicate a rapid central sympatholytic ghrelin effect already during the immediate phase.
During the delayed ghrelin phase, at first, MSNA remained at the placebo level and HR was decreased during a state of reduced BP (t +35 min). Subsequently, MSNA and HR were lower during a state of unchanged BP (t +65 min, t +85 min) compared with placebo. Therefore, one potential explanation might involve a shift of the setpoint of sympathetic BP control toward lower BP levels during the delayed ghrelin phase. This interpretation is in accordance with previous findings concerning interactions of ghrelin with the sympathetic nervous system (SNS). Thus, in healthy men, ghrelin was found to decrease mean arterial BP without any significant change of HR but with an increase of cardiac index (20). In patients with chronic heart failure, intravenous ghrelin twice a day for 3 wk resulted in a significant decrease of plasma norepinephrine, a global marker of overall SNS activity (22). Moreover, the intracerebroventricular administration of ghrelin decreased BP, HR and renal SNS activity and increased the sensitivity of baroreflex in conscious rabbits (18). Intravenous ghrelin suppressed cardiac SNS activation following myocardial infarction in rats (26, 27) and suppressed sympathetically mediated cytokine release in experimental sepsis (35).
Alternatively, ghrelin might have blunted stress-induced sympathetic feed-forward commands in contrast to placebo. In the present study, MSNA increased in the course of time under placebo condition. Such increases of sympathetic activity are a well-known phenomenon during recording protocols of long duration and are commonly explained as a reaction to urinary bladder filling or discomfort due to restricted body position—though subjective ratings for these categories were not increased in the present study at the end of the placebo condition (25). Ghrelin, in contrast, finally (t +85 min) resulted in an unchanged BP combined with a significantly lower MSNA and HR compared with placebo. MSNA is a direct measure of central nervous sympathetic outflow. Therefore, our findings might indicate that ghrelin prevented stress-related sympathoactivation. This is in accordance with recent findings in humans showing mitigated response to mental stress under ghrelin (16)—an aspect that may have distinct clinical implication on individual hypertensive pressor responses. Microneurographic approaches to assess baroreflex gain represent a simplified input-output relationship of blood pressure and MSNA, which may potentially be modulated by ghrelin independently from its mere baroreflex effects. In accordance with previous findings, we observed that following ghrelin plasma levels of various hemodynamically active hormones like cortisol, copeptin, and epinephrine were strongly increased, transiently (20). These increases can be attributed to the presence of specific GH-secretagogue binding sites within the hypothalamus, pituitary gland, and adrenal medulla (10, 16). In contrast to epinephrine, norepinephrine was not elevated—a pattern similarly observed even after injection of much higher ghrelin doses of 10 μg/kg (20). When injected repeatedly over 3 wk at doses comparable to our study, ghrelin even decreased plasma norepinephrine levels (19). Plasma norepinephrine originates from multiple sympathetic branches by spillover. Therefore, this rather crude marker indicates that overall sympathetic activity was not increased in our study, corroborating the above-mentioned MSNA-findings.
Blood pressure resetting and sensitivity of baroreflex.
In the present study, baroreflex assessment yielded a steepening of the MSNA-BP slope at 35 and 65 min after ghrelin administration. Thus, intravenous ghrelin increased the sensitivity of arterial baroreflex. In consistency, a similar effect was found for renal sympathetic nerve activity in an animal model (18). This finding provides additional support for the hypothesis that peripheral ghrelin modulates circulation control at central nervous system levels, including hypothalamic and brain stem centers (17). There is evidence that ghrelin can pass the blood-brain barrier (2). These mechanisms, however, warrant further investigation.
Ghrelin is well known for its GH-secretagogue function. Since in states of GH-deficiency an increased sympathetic outflow was described (29), whereas GH replacement therapy led to a decrease of MSNA (28), one might hypothesize that GH should inhibit sympathetic outflow. Moreover, intravenous ghrelin increases cortisol plasma levels. Glucocorticoids have been shown to acutely decrease MSNA (8). Hence, as an alternative to the mechanism mentioned above, the delayed decrease of sympathetic outflow observed in our study following intravenous ghrelin could be mediated by GH or cortisol.
Recently, Lambert et al. (16) found that the continuous intravenous infusion of ghrelin for 1 h significantly decreased BP compared with preceding or subsequent placebo infusion (16). This decrease was accompanied by a modest but significant increase of MSNA, indicating a baroreflex-mediated sympathetic counterregulation. The experimental bolus design of the present study differs substantially from the continuous infusion model chosen by Lambert et al. (16). Moreover, conducting the ghrelin and placebo experiment on separate days reduced the effects of substance interaction in our study. Under physiological conditions, ghrelin is secreted following a pulsatile pattern (4) and inactivated by deacylation with a half-life of only about 9–13 min (1). In the present study, the bolus application of ghrelin produced two distinct phases. Presumably, in Lambert's experiment, the physiological mechanisms are comparable to those prevailing during the immediate phase of our study. Absence of a significant increase of HR as well as the blunted sympathoexcitation in response to stressful stimuli described by Lambert et al. suggested an additional central nervous effect of ghrelin as proposed for the delayed phase of the present study.
We found new evidence in human subjects that high circulating levels of the gut-hormone ghrelin can decrease both BP and SNS activity, while increasing baroreflex sensitivity. Low ghrelin levels, in contrast, as found in obese subjects (32), might contribute to obesity-related SNS activation (9), a hallmark of obesity hypertension. On the other hand, physiologically high ghrelin levels during slow wave sleep might play a role in nocturnal BP dipping and resetting of the baroreflex during nighttime sleep (24). Moreover, ghrelin therapy was beneficial in humans suffering from chronic heart failure (22) or pulmonary cachexia (19), i.e., conditions strongly associated with chronically increased SNS activity. In these studies, ghrelin was injected intravenously twice per day for 3 wk at doses equivalent to our study (2 μg/kg). The therapeutic ghrelin effect might at least partially be mediated by central nervous sympathoinhibition.
Our experimental model focused on the acute effects of ghrelin on baroreflex-mediated BP control. Recordings of MSNA, being the most important vasoconstrictive stimulus, provide a direct measure of sympathetic outflow to resistance arteries. This allows us to draw very specific conclusions concerning this branch of the baroreflex arch, whereas other aspects, particularly concerning cardiac baroreflex function, cannot be answered with similar specificity. Though the number of subjects was small in our study, the solid placebo-controlled design permitting within subject comparison is an important strength of our study. Statistical tests were sufficiently powered to detect an alteration of sympathetic baroreflex function, whereas analysis of the cardiac baroreflex slope was underpowered. The baroreflex test chosen in our study represents a simplified method providing basic insight. Other more sophisticated dynamic techniques might provide more detailed information but will not alter our key findings (12, 13).
Perspectives and Significance
Our study suggests a two-phase model of ghrelin effect on sympathetic cardiovascular control: an immediate phase of direct peripheral vasodilation with baroreflex-mediated sympathoactivation, followed by a delayed neurohumoral phase. The latter was characterized by decreased vasoconstrictive sympathetic activity combined with increased baroreflex sensitivity compared with placebo. Our results could offer a new understanding of seemingly conflicting findings in previous studies. To further elucidate long-lasting ghrelin-baroreflex interactions, experimental models of chronic ghrelin administration are desirable. We would expect that under this regimen, an even more distinct central nervous inhibition of vasoconstrictive SNS activity might occur. Subsequently, it would be most interesting to investigate the potentially beneficial effects of ghrelin on MSNA under pathological conditions associated with permanently increased sympathetic activity, like chronic heart failure, liver cirrhosis, morbid obesity, or sleep-disordered breathing.
This study was financially supported by a Grant of the Deutsche Forschungsgemeinschaft (SFB 654/B4) without influencing the study design, data collection, analysis, or interpretation.
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: A.F.K., C.D., and F.S. conception and design of research; A.F.K., J.R., F.M., and F.S. performed experiments; A.F.K. and J.R. analyzed data; A.F.K., J.R., F.M., K.A.I., and F.S. interpreted results of experiments; A.F.K. prepared figures; A.F.K. and F.S. drafted manuscript; A.F.K., J.R., K.A.I., C.D., H.L., and F.S. edited and revised manuscript; A.F.K., J.R., F.M., K.A.I., C.D., H.L., and F.S. approved final version of manuscript.
We thank Martina Grohs and Heidi Ruf for their excellent technical assistance.
- Copyright © 2012 the American Physiological Society