OUP user menu

The endogenous peptide apelin potently improves cardiac contractility and reduces cardiac loading in vivo

Euan A. Ashley, Jennifer Powers, Mary Chen, Ramendra Kundu, Tom Finsterbach, Anthony Caffarelli, Alicia Deng, Jens Eichhorn, Raina Mahajan, Rani Agrawal, Joan Greve, Robert Robbins, Andrew J. Patterson, Daniel Bernstein, Thomas Quertermous
DOI: http://dx.doi.org/10.1016/j.cardiores.2004.08.018 73-82 First published online: 1 January 2005


Objective: The endogenous peptide apelin is differentially regulated in cardiovascular disease but the nature of its role in cardiac function remains unclear.

Methods: We investigated the functional relevance of this peptide using ECG and respiration gated magnetic resonance imaging, conductance catheter pressure–volume hemodynamic measurements, and echocardiography in vivo. In addition, we carried out histology and immunohistochemistry to assess cardiac hypertrophy and to localize apelin and APJ in the adult and embryonic mouse heart.

Results: Intraperitoneal injection of apelin (300 μg/kg) resulted in a decrease in left ventricular end diastolic area (pre: 0.122 ± 0.007; post: 0.104 ± 0.005 cm2, p=0.006) and an increase in heart rate (pre: 537 ± 20; post: 559 ± 19 beats per minute, p=0.03). Hemodynamic measurements revealed a marked increase in ventricular elastance (pre: 3.7 ± 0.9; post: 6.5 ± 1.4 mm Hg/RVU, p=0.018) and preload recruitable stroke work (pre: 27.4 ± 8.0; post: 51.8 ± 3.1, p=0.059) with little change in diastolic parameters following acute infusion of apelin. Chronic infusion (2 mg/kg/day) resulted in significant increases in the velocity of circumferential shortening (baseline: 5.36 ± 0.401; 14 days: 6.85 ± 0.358 circ/s, p=0.049) and cardiac output (baseline: 0.142 ± 0.019; 14 days: 0.25 ± 0.019 l/min, p=0.001) as determined by 15 MHz echocardiography. Post-mortem corrected heart weights were not different between apelin and saline groups (p=0.5) and histology revealed no evidence of cellular hypertrophy in the apelin group (nuclei per unit area, p=0.9). Immunohistochemistry studies revealed APJ staining of myocardial cells in all regions of the adult mouse heart. Antibody staining, as well as quantitative real time polymerase chain reaction identified expression of both APJ and apelin in embryonic myocardium as early as embryonic day 13.5.

Conclusions: Apelin reduces left ventricular preload and afterload and increases contractile reserve without evidence of hypertrophy. These results associate apelin with a positive hemodynamic profile and suggest it as an attractive target for pharmacotherapy in the setting of heart failure.

  • Apelin
  • APJ
  • Angiotensin
  • Heart failure
  • Pressure–volume hemodynamics

This article is referred to in the Editorial by G.A. Losano (pages 8–9) in this issue.

1. Introduction

G protein-coupled receptor systems are important regulators of cardiovascular physiology and represent the greatest proportion of known therapeutic targets. As such, the association of novel G protein coupled signaling pathways with human disease is of great importance. Using expression profiling of offloaded failing left ventricle, we recently postulated a role for the apelin-APJ system in human heart failure [1].

APJ is a 377 amino acid, 7 transmembrane domain, Gi coupled receptor whose gene is localized on the long arm of chromosome 11. It was first cloned in 1993 from genomic human DNA using degenerate oligonucleotide primers [2] and shares significant homology with angiotensin II receptor type 1. Despite this homology, however, angiotensin II does not bind APJ. The natural ligand for the APJ receptor, apelin, has been isolated from bovine stomach by Tatemoto et al. [3]. Spanning 1726 base pairs of genomic DNA with 3 exons, the apelin locus is highly conserved between species. Synthesized as a 77 amino acid preprotein and cleaved to short peptides of different sizes in different tissues [4,5], apelin has been implicated in cardiovascular function, central autonomic control, and fluid homeostasis.

Several authors have localized APJ transcripts to the paraventricular nuclei and supraoptic nucleus of the hypothalamus, the anterior and intermediate lobes of the pituitary, and the pineal gland [6–9]. Reaux et al. [8] have also demonstrated expression of apelin peptide in these hypothalamic nuclei, including colocalization with vasopressinergic neurons in the supraoptic nucleus. Functional studies reveal effects of intra-cerebroventricular injection of apelin 13 on vasopressin levels in conscious mice with free access to water, suggestive of a direct inhibitory action on vasopressin release via APJ in vasopressinergic neurons. This hypothesis was recently confirmed by two sets of authors [10,11].

The direct role of apelin in cardiovascular physiology has however been investigated by few investigators to date. Early studies showed a clear decrease in mean arterial pressure following an intravenous bolus injection of apelin in rats [9,12,13]. In addition, APJ knockout mice show an increased vasopressor response to angiotensin II, suggesting a counter-regulatory role in relation to the renin–angiotensin system. However, another group reported that apelin potently contracts isolated human saphenous vein [14] suggesting the effect of apelin on vascular reactivity is not absolutely clear. In relation to myocardial function, Szokodi et al. [15] showed a positive inotropic effect of apelin on the contractility of the isolated rat heart that was both potent (EC50 in the low picomolar range) and efficacious (maximum developed tension was 70% that of isoproterenol). However, despite these significant effects, and the demonstration that apelin circulates in plasma [1,5], the role of apelin in vivo remains unknown. Specifically, the balance of effects on cardiac loading and intrinsic contractility (ventriculo-vascular coupling) has yet to be described.

2. Methods

The study protocol was approved by the Stanford University Administrative Panel on Laboratory Animal Care and conforms 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). The mouse model was chosen for its relevance to experiments using transgenic models such as those of Ishida et al. [16].

2.1. Peptide reagents

Apelin-12 was purchased from Bachem (Bachem Bioscience, King of Prussia, PA). Pyroglutamylated apelin-13 (PYRapelin13) was purchased from American Peptide (Sunnyvale, CA). The pyroglutamylated form of apelin13 is of equivalent potency to apelin12 but more resistant to degradation by aminopeptidases [4,17]. For this reason, we judged it as a more suitable reagent for chronic infusion. Apelin was dissolved in distilled, autoclaved, degassed water, frozen at −20° C at high concentration, and aliquoted on the morning of use.

2.2. Magnetic resonance imaging

Male C57Bl/6 mice aged 16 weeks (n=9) were scanned twice on subsequent days. The animals underwent general anesthesia while breathing spontaneously via a nose cone fitted carefully to minimize escape of anesthetic into the environment. Two percent isoflurane was administered with an oxygen flow rate of 1–2 l/min. Platinum needle ECG leads were inserted subcutaneously. Respiration was monitored by means of a pneumatic pillow sensor positioned against the abdomen. Mouse body temperature was maintained during scanning at 37 °C by a flow of heated air thermostatically controlled by a rectal temperature probe. Magnetic resonance images were acquired on a 4.7 T Oxford magnet controlled by a Varian Inova console (Varian, Palo Alto, CA) using a transmit–receive, quadrature, volume coil with an inner diameter of 3.5 cm. Image acquisition was gated to respiration and to the ECG R wave (SA Instruments, Stony Brook, NY). Coronal and sagittal scout images led to the acquisition of multiple contiguous 1-mm-thick, short axis slices orthogonal to the interventricular septum. Nine cine frames were taken at each slice level with the following sequence parameters: TE=2.8 ms, NEX=12, FOV=3 × 3 cm, matrix=128 × 128, flip angle=60°. Cine frames were spaced 16 ms apart and acquired through slightly more than one cardiac cycle guaranteeing acquisition of systole and diastole. On the second day of scanning, mice received 300 μg/kg body weight of apelin12 as an intraperitoneal injection 1 h prior to scanning. A pilot study had previously identified 1 h as an appropriate time within which to identify apelin effects resulting from peritoneal absorption. Planimetry measurements of end diastolic and end systolic dimension were derived offline from short axis views of the left ventricle at the level of the papillary muscles using ImageJ software (National Institutes of Health, Bethesda, MD). Ejection fraction was calculated as [LVEDA−LVESA]/LVEDA.

2.3. Pressure–volume hemodynamics

Pressure–volume hemodynamics were assessed using the Aria System (Millar Instruments, Houston, TX). This measurement platform, specifically designed for hemodynamic cardiovascular measurements in small rodents, comprises an ultra-miniature 1.4 F (0.47 mm outer diameter) catheter which incorporates pressure and conductance sensors, processing hardware including analog-digital conversion and analysis software. Pressure is measured directly in mm Hg while the conductivity of blood is used to estimate volume and allow construction of pressure–volume relationships in real time. Derived measurements of conductivity are expressed as relative volume units.

Male C57Bl/6 mice aged 8–12 weeks (n=10) were anesthetized with 1–2% isoflurane in oxygen. The internal jugular vein was cannulated with PE tubing and a 10% albumin solution infused at 5 μl/min following a bolus of 150 μl over 5 min. After tracheotomy, a 19-gauge cannula was inserted into the trachea and the animal was ventilated at a tidal volume of 200 μl at 100 breaths per minute (Harvard Apparatus, Holliston, MA). Mice were warmed throughout the procedure and constantly monitored for depth of anesthesia. Following an incision just dorsal to the xyphoid cartilage, the diaphragm was visualized from below, and after diaphragmatic incision, the left ventricular apex was visualized. The pressure–volume catheter was inserted along the long axis of the left ventricle, from where it was adjusted to obtain rectangular shaped pressure–volume loops. Appropriate position was verified post-mortem (Fig. 2, Panel D). Baseline loops were recorded following volume replacement, at which point, the inferior vena cava was visualized within the chest and occlusion parameters were recorded during and after a 5-s manual occlusion of this vessel. Next, the albumin solution was replaced by one containing 100 nM Apelin12 which was infused at 5 μl/min for 20 min, following which, baseline and occlusion loops were recorded once again. Because the half-life of apelin is unknown, and control of prolonged anesthesia in rodents is challenging, a control series of animals (n=4) was infused with albumin only for an equivalent period (30 min). No significant differences in any measured variables were observed during this time period in these animals (data not shown).

Fig. 2

Pressure–volume hemodynamics in response to acute apelin infusion. Anesthetized, ventilated C57Bl6 mice underwent placement of a specialized conductance catheter along the long axis of the left ventricle via a sub-diaphragmatic approach (Panel D). The pressure–volume relationship at different levels of preload, facilitated by a 5-s manual occlusion of the inferior vena cava, is illustrated. Loops were recorded at baseline (example series of loops from one animal, Panel A) and following 20 min of apelin infusion (example loops from the same animal following apelin, Panel B). Volume is expressed as Relative Volume Units (RVU). Mean data demonstrated that, after apelin infusion, ventricular elastance, the slope of the end systolic pressure–volume relationship, was increased (Panel C, a summary graph illustrating the mean increase in slope and intercept across all animals, p=0.018) along with preload recruitable stroke work (Panel E, mean ± S.E.M., p=0.056). Steady state end systolic pressure was lower following apelin indicating a reduction in afterload (Panel F, mean ± S.E.M., p=0.02).

Signals from the catheter were digitized using the Powerlab system (ADInstruments, Colorado Springs, CO) and stored for offline analysis using the PVAN software (Pressure–Volume ANalysis, Millar Instruments). This software allows analyses of pressure (e.g., end systolic pressure, end diastolic pressure) and volume (e.g., end diastolic volume, end systolic volume) and derivation of pressure–time and volume–time parameters at steady state. In addition, it facilitates standardized derivation of load independent parameters which rely on regression (the end systolic pressure volume relation–Ees; pre-load recruitable stroke work–a regression line fitted to the relationship between end diastolic volume, and stroke work–stroke work representing the area of the pressure–volume loop; the dPdtmax to end diastolic volume relation; and time varying elastance–the maximum slope of a series of lines drawn through each point in the cardiac cycle).

2.4. Chronic apelin infusion

To test longer term effects of apelin, we infused 2 mg/kg/day PYRapelin13 into 8–12 week male C57/Bl/6 mice. A short anesthetic (isoflurane 1% in oxygen 1 l/min) facilitated implantation of a 2-ml osmotic minipump under the scruff with staple closure (Alzet Osmotic pumps, Cupertino, CA, model 1002). Minipumps contained either PYRapelin13 (n=10) or sterile normal saline (n=5). Cardiovascular parameters were recorded at 7 days and 14 days by tail cuff sphygmomanometry (Visitech Systems, Apex, NC) and echocardiography (probe frequency 15 MHz, Acuson Sequoia, Siemens, Malvern, PA). For echocardiography, mice were anesthetized using isoflurane (0.75–1.25% in oxygen 1 l/min) then placed supine and warmed with a heat lamp. Using a gel buffer, parasternal long and short axis views were recorded in each animal to allow estimation of indices of contractility such as fractional shortening (LVEDD−LVESD/LVEDD), cardiac output ([Pi × (Aod)2 × VTI × HR]/4), velocity of circumferential shortening ([LVEDD−LVESD]/[ET × LVEDD]), and LV mass (1.05*[(IVSD+LVEDD+PWTD)3 − LVEDD3), where LVEDD is left ventricular end diastolic diameter, LVESD is left ventricular end systolic diameter, IVSD is inter ventricular septum in diastole, PWTD is posterior wall thickness in diastole, Aod is aortic diameter, VTI is the velocity time integral, ET is ejection time. The last two parameters are derived from Doppler sampling of the outflow tract. All measurements were made by one operator blinded to group. At 7 and 14 days of infusion, these measurements were repeated.

2.5. Assessment of ventricular hypertrophy

Following chronic infusion, mice were sacrificed, weighed and their organs removed for measurement of wet weight. Wet weight was expressed relative to body weight. The hearts were then fixed in formalin, cut in short axis at the papillary muscle level and embedded in paraffin. The paraffin blocks were sectioned and the sections stained with both Masson's trichome and hematoxylin and eosin according to standard protocols. Fields from the mid-portion of the left ventricular free wall were captured under high power, converted to TIFF format, and viewed with ImageJ software (National Institutes of Health). A region of interest was defined of a size approximating 50 nuclei taking care to avoid epicardial and endocardial areas. The number of nuclei within this cross-sectional area was counted for each heart and the mean compared between groups.

2.6. Expression of the APJ receptor

Adult mice were perfusion fixed with 4% paraformaldehyde, and adult heart and whole embryos further fixed over night, embedded in paraffin and sectioned. Five-micron-thick sections were cut and stored at 4 °C. Blocking was achieved using 1.5% goat serum (Vector labs, Burlingame, CA). Sections were stained with APJ polyclonal antibody (Lifespan Biosciences, Seattle, WA, LSA64, 1/100 dilution). Secondary antibody was biotinylated anti-rabbit raised in goat (Vector labs, 1/200 dilution). For embryo sections, the chromagen substrate was BCIP/NBT for alkaline phosphatase (Vector labs), and sections were counterstained with nuclear fast red. For the heart sections, the chromagen substrate was vector red for alkaline phosphatase (Vector labs), and sections were counterstained with hematoxylin.

2.7. Data analysis

Data were analyzed using Student's t statistic (paired) or the repeated measures analysis of variance using the post hoc comparison of Fisher (NCSS 2002). Values reported are mean plus or minus the standard error of the mean. Exact p values are reported for all comparisons.

3. Results

3.1. Magnetic resonance imaging

Contractility of C57Bl/6 mice was first assessed at baseline. Short axis views of the left ventricle at the level of the papillary muscles allowed estimation of end systolic and end diastolic areas by planimetry (Fig. 1, Panel D). Apelin had no effect on the spontaneous rate of respiration (pre: 68 ± 11; post: 66 ± 7 breaths per minute, p=0.7), while heart rate calculated as the inverse of the R–R interval of the electrocardiogram was significantly greater (pre: 537 ± 20; post: 559 ± 19 beats per minute, p=0.03, Fig. 1, Panel A). Ejection fraction tended to increase following apelin injection but this did not reach significance (pre: 63.6 ± 2.7; post: 67.7 ± 1.6%, p=0.16, Fig. 1, Panel B). However, the end diastolic area was very significantly reduced following apelin injection (pre: 0.122 ± 0.007; post: 0.104 ± 0.005 cm2, p=0.006, Fig. 1, Panel C).

Fig. 1

Changes in cardiovascular function following intraperitoneal injection of apelin12 in C57Bl6 mice. Mice (n=9) were anesthetized with isoflurance and warmed to 36–37° before magnetic resonance imaging. Electrocardiography revealed a significant increase in HR following apelin injection (Panel A). ECG and respiration gated cine magnetic resonance images of the left ventricle taken in short axis reveal a significant reduction in left ventricular end diastolic area (Panel C) with an upward trend in ejection fraction (Panel B). Panel D shows example images of end diastole (left) and end systole (right) from pre (above) and post (below) apelin injection.

3.2. Pressure–volume hemodynamics

Separating effects of load and function are not possible with non-invasive imaging and since apelin has effects on both vascular reactivity and intrinsic contractility, we elected to assess ventriculo-vascular coupling via pressure–volume hemodynamics (Table 1, Fig. 2). Here, intraventricular pressure is measured directly while intraventricular volume is estimated by conductance. Thus, effects of load and intrinsic contractility can be simultaneously and independently assessed through the construction of pressure–volume loops, both at baseline and during preload reduction, achieved by manual compression of the inferior vena cava. Baseline measurements were consistent with those previously reported in the literature for the C57Bl/6 mouse (Table 1). Thoracotomy requires greater depth of anesthesia than non-invasive imaging and this contributes to a lower heart rate in invasive studies. This mild but obligatory cardiovascular depression may also explain the lack of increase in heart rate (pre: 376 ± 21; post: 353 ± 33 bpm, p=0.3) which was seen in the MRI studies.

View this table:
Table 1

Hemodynamic indices following acute infusion of apelin

HR (bpm)EDV (RVU)Pmax (mm Hg)EF (%)Ees (mm Hg/RVU)Intercept
Pre apelin3762129.−14.25.9
Post apelin3533326.25.980.77.959.
p value0.
dP/dtmax (mm Hg/s)PRSW (mm Hg × RVU)InterceptdP/dt-EDV (mm Hg/s/RVU)InterceptEmax (mm Hg/RVU)
SV (RVU)CO (RVU/min)
Pre apelin18.62.96724677
Post apelin15.43.75033799
p value0.30.1
  • Invasive hemodynamic indices measured by conductance catheter in the anesthetized, ventilated mouse before and after apelin infusion. HR–heart rate; EDV–end diastolic volume; Pmax–maximum generated pressure; EF–ejection fraction, Ees–end systolic pressure volume relationship (ventricular elastance); dP/dtmax–maximum rate of rise of left ventricular pressure; PRSW–preload recruitable stroke work; dP/dt-EDV–relationship between maximum rate of rise of left ventricular pressure and end diastolic volume; Emax–time varying elastance; RVU–relative volume unit; SV–stroke volume; CO–cardiac output). See text for details.

3.2.1. Systolic function

Consistent with the MR data, LV end diastolic volume was lower, but this difference was not significant (pre: 29.3 ± 6.1; post: 26.2 ± 5.9 RVU, p=0.4). Similarly, steady state, load-dependent measures of contractility such as ejection fraction and dP/dtmax did not change following apelin infusion (EF: pre–61.2 ± 6.8 vs. post–59.1 ± 9.0%, p=0.8; dP/dtmax: pre–6710 ± 1583 vs. post–5564 ± 2914, p=0.2). Consistent with a reduction in mean circulatory filling pressure suggested by a reduction in maximum developed pressure (data not shown) and end systolic pressure (Fig. 2, Panel F), mean stroke volume and cardiac output were reduced following apelin infusion, however this change was not significant (SV: pre–18.6 ± 2.9 vs. post–15.4 ± 3.7 RVU, p=0.3; CO: pre–6724 ± 677 vs. post–5033 ± 799 RVU/min, p=0.1).

Inspection of occlusion parameters revealed highly significant changes in load independent measures of contractility. Both the slope and intercept of the end systolic pressure–volume relationship were significantly different following apelin infusion (Fig. 2, Panels A–C). Corresponding to this finding, the time varying elastance was significantly greater (pre: 6.0 ± 1.5; post: 12.7 ± 3.1, p=0.017). There was also a trend towards an increase in pre-load recruitable stroke work (Panel E). Although, dP/dtmax itself was not different following apelin infusion (Table 1), the relationship between dP/dtmax and end diastolic volume was steeper and its intercept greater (Table 1). Arterial elastance, a steady-state parameter that incorporates peripheral resistance, impedance, compliance, and systolic/diastolic time intervals (approximated by the steady state LV end systolic pressure to stroke volume ratio) did not change significantly, however the maximum developed pressure and end systolic pressure were reduced (Panel F).

3.2.2. Diastolic function

Time constants of relaxation were not different after apelin infusion (data not shown). In addition, the slope and intercept derived from a linear model fit of the end diastolic pressure–volume relationship were not different. However, pressure decay is known to be load-dependent and model dependency of diastolic parameters is well recognized [18]. When the end diastolic pressure–volume relationship was fit by a monoexponential of the form [LVEDP=k1 × exp(k2 × LVEDV)], the constant k1 trended towards an increase (pre: 0.064 ± 0.008; post: 0.111 ± 0.002, p=0.09), while the exponential k2 was clearly unchanged (pre: 0.83 ± 0.22; post: 0.69 ± 0.41, p=0.66).

3.3. Chronic apelin infusion

We infused PYRapelin13 over the course of two weeks at a level previously shown to exert acute hemodynamic effects. No significant changes were seen in saline-infused controls from baseline to 14 days. Significant changes in heart rate and blood pressure, known to occur acutely, were not detected over the period of the chronic apelin infusion: neither conscious heart rates (tail cuff method) nor isoflurane heart rates (echo Doppler) were different at any time point (data not shown) and while systolic blood pressure trended lower during apelin infusion this change was not significant (Fig. 3). End diastolic diameter increased but this was not significant (3.76 vs. 3.85 vs. 4.0 mm at time 0, 7, and 14 days, respectively, p=0.22). However, left ventricular contractility measurements derived from Doppler ultrasound of the left ventricular aortic outflow tract were significantly increased during apelin infusion. The velocity of circumferential shortening was increased at 14 days (p=0.049, Panel C). Similarly, cardiac output was increased at 7 days and this increase was maintained at 14 days (p=0.001, Panel D).

Fig. 3

Effect of chronic apelin infusion in C57Bl6 mice. Long axis and short axis views of the left ventricle with Doppler sampling of the outflow tract were used to estimate left ventricular contractility in vivo (Panel A). The velocity of circumferential shortening (Panel C) and cardiac output (Panel D) were significantly increased from baseline following two weeks of PYRapelin13 infusion. Systolic blood pressure as determined by tail cuff was also lower but this did not reach significance (Panel B).

3.4. Assessment of cardiac hypertrophy

Despite the increases in contractility described above, post-mortem organ weights were not different between the saline and apelin groups (absolute: apelin 0.1256 ± 0.0038 vs. saline 0.1172 ± 0.0053 g; corrected for body weight: apelin 0.0043 ± 3.47E-07 vs. saline 0.004147 ± 7.36E-08). In addition, in histological section, the number of nuclei per unit area was not significantly different between groups (apelin 42.6 ± 3.72 vs. saline 42.8 ± 8.2) suggesting absence of myocyte hypertrophy.

3.5. Expression of APJ

Immunohistochemistry for apelin and APJ revealed specific immunolocalization of both proteins in the developing myocardium, with very similar patterns of expression as early as embryonic day 13.5 (Fig. 4, Panels A, B). Real-time quantitative RT-PCR of isolated heart mRNA validated this finding, and suggested that the relative contribution to the total myocardial RNA by these transcripts remains relatively constant through late gestation and adulthood (Fig. 4, Panel C). In the adult mouse heart, immunolocalization of APJ was identified in association with both atrial and ventricular myocardial cells (Fig. 4, Panels D–F).

Fig. 4

Immunohistochemistry of apelin and APJ in the developing and adult mouse heart. Specific immunolocalization of both proteins in the developing myocardium revealed very similar patterns of expression as early as embryonic day 13.5 (Panels A, B: A–atrium, V–ventricle, li–liver). Real time quantitative RT-PCR of isolated heart mRNA suggested that the relative contribution to the total myocardial RNA by these transcripts remains relatively constant through late gestation and adulthood (Panel C). In the adult mouse heart, immunolocalization of APJ expression was identified in association with both atrial and ventricular myocardial cells (Panels D, E). A control slide without the addition of the secondary antibody is shown in Panel F.

4. Discussion

The principal findings of this study are that acute administration of apelin in vivo causes a reduction in left ventricular end diastolic area and an increase in left ventricular elastance, whereas chronic apelin infusion increases cardiac output without causing hypertrophy. These findings extend and complement previously described physiological effects of apelin and suggest an important role for the apelin-APJ system in cardiovascular control.

Several authors have shown that intravenous bolus delivery of apelin reduces mean arterial pressure with a rapid onset and peak at approximately 5 min [9,12]. In another study, Cheng et al. [13] administered apelin infusions to conscious, unrestrained rats, both with and without ganglionic blockade. They reported that apelin decreased mean arterial pressure and mean circulatory filling pressure in both groups in a dose-dependent manner. The authors concluded that apelin has a venodilator action more efficacious than nitroglycerin or hydralazine, both of which decrease mean arterial pressure but cause a sympathetically mediated reflex increase in mean circulatory filling pressure. In light of this, we hypothesized that apelin would reduce both left ventricular preload and afterload through venous and arterial dilation. Our finding that end diastolic area is significantly decreased after intraperitoneal injection of apelin12 provides the first demonstration that the vascular reactivity of apelin couples to the left ventricle. We also detected decreases in maximum and end systolic pressure in invasive studies, in line with previously observed changes in mean arterial pressure. The significant increase in heart rate seen in our MRI study and by some [9] but not all [12] other investigators is likely explained by a baroreceptor mediated response to decreased mean arterial pressure.

Thus, well-characterized apelin-mediated changes in vascular reactivity translate to the in vivo setting and to changes in loading conditions of the heart. However, an important question remains: what are the related and independent changes in myocardial function? In the isolated rat heart, Szokodi et al. [15] described a positive, efficacious, and potent effect of apelin on externally developed tension and pre-load recruitable maximum rate of developed pressure, however, these effects had not been observed in vivo. Our findings that apelin increases the slope of the end systolic pressure–volume relationship (ventricular elastance), the slope of the end diastolic volume to stroke work relationship, and in the chronic infusion model, the velocity of circumferential shortening, provide the first demonstration of an in vivo effect of apelin on myocardial contractility. Indeed, the change in slope of the end-systolic pressure–volume relationship (Fig. 2, Panel C) is consistent with the change in the end diastolic pressure to dP/dtmax relationship observed by Szokodi et al. Further, this observation reinforces the importance of assessing biological signaling systems active at the level of the myocardium and vasculature in an integrated manner: standard, load-dependent measures of contractility such as dP/dtmax and ejection fraction were not significantly different after apelin infusion in vivo. Such observations can be clearly understood in the context of previously described changes in loading conditions. While apelin increases intrinsic contractility, an effect which, in the absence of changes in loading conditions, would lead to an increase in ejection fraction and dP/dtmax, apelin-mediated reductions in preload will move the Frank Starling curve to the left and thus, alter ejection phase indices (and particularly those such as dP/dtmax which are exquisitely sensitive to preload) downwards. Similarly, although apelin-mediated reductions in systemic blood pressure reduce afterload, something which in itself would tend to augment afterload-dependent measures of contractility such as ejection fraction, the leftward shift in the Frank–Starling relationship will mitigate the increase and result in a net effect equivalent to little change or little increase. This is, in fact, what we observe. Effects of agents which change loading conditions concurrent with the visco-elastic properties of the ventricle (and there is significant conservation in signaling pathways between the vasculature and the myocardium) can be masked by traditional measures of contractility.

An intriguing finding from this study lies in the contrast between acute and chronic effects of apelin-APJ augmentation. In the acute setting at steady state, mean circulatory filling pressure is reduced and although inotropic potential is increased, cardiac output trends downwards. In contrast, following chronic infusion, we detect highly significant increases in cardiac output. Although it is possible, this finding could reflect the different techniques used to assess function invasively and non-invasively (both techniques require assumptions, which may be violated in these models), biological explanations are also attractive. For example, it is possible the acute change seen in filling pressures is curtailed over time (a possibility supported by the lack of significant change in tail cuff blood pressure in the chronic infusion animals, in contrast with the clear reduction in end systolic pressure in the acute invasive studies). This would leave a small baseline inotropic effect which might easily be sufficient to increase cardiac output (in a manner similar to dobutamine). Another exciting possibility is that chronic apelin therapy is associated with a shift to a more ‘efficient’ cardiac stroke, resulting in greater power for energy consumption. Such an effect is exemplified by the calcium sensitization of agents such as EMD-57033 [19], although it seems more likely that an effect of apelin on myocardial efficiency would be mediated via PKC. This is because cardiac apelin-APJ signaling is abrogated by PKC inhibitors and PKC phosphorylation of cardiac TnI has been shown to reduce the requirements of the contractile apparatus for both Ca2+ and ATP (promoting efficient ATP utilization during contraction) [20]. Further experiments will be needed to define such mechanisms, which remain speculative for now.

Effecting a reduction in cardiac loading while increasing contractile reserve makes the apelin-APJ system an attractive target for therapy in heart failure. Indeed, we have shown increases in the myocardial expression of both apelin and its receptor APJ following LVAD offloading in human heart failure [1]. Further, we demonstrated increases in circulating apelin in patients with moderate LV dysfunction [1]. Together, these observations suggest that apelin may act as a ‘good peptide’ in heart failure (akin to the natriuretic peptides Ref. [21]) serving to ameliorate rather than antagonize the abnormal hemodynamic state of that disease. Although no natriuretic quality of apelin has yet been described, one group has shown a decrease in release of vasopressin and in water intake following intracerebroventricular injection of PYRapelin13 in rats [8].

A caveat to chronic pharmacologic augmentation of a positive inotropic is the potential for deleterious effects demonstrated in clinical trials with agents such as milrinone [22,23] and dobutamine [24,25], and in transgenic models overexpressing components of the beta-adrenergic signaling systems [26,27]. However, apelin increases contractile reserve through an increase in elastance and concomitantly decreases loading, so does not overdrive the heart. It has been known for over a decade that the hemodynamic profile of well-characterized inotropic agents is improved by the addition of pre-load reducing agents [28]. In this study, we further demonstrated no evidence of cardiac hypertrophy manifest either at the organ or cellular level, after two weeks of apelin infusion compared to saline control. This is despite increases in cardiac output and in the velocity of circumferential shortening [15], despite significant homology between apelin and angiotensin II, and despite similarity in downstream signaling networks to endothelin, angiotensin II and alpha-adrenoceptor agonists. All of these things might suggest apelin would be involved in remodeling and might cause cardiac hypertrophy via calcium-dependent processes such as calmodulin kinase activation [29]. However, whereas endothelin and angiotensin increase peripheral resistance, apelin is one of the most potent arterial and venous dilators known, and it seems likely that this more global effect of apelin on vascular tone outweighs any local tissue effects leading to a net absence of hypertrophy.

Although apelin expression has been well described at both the RNA and protein level [1,30–32], antibodies for APJ have only recently become available. Here, we show for the first time protein expression of APJ by myocardial cells of the atrium and ventricle. In concert with earlier studies from this laboratory identifying apelin expression in the coronary endothelium, these data suggest a paracrine signaling pathway that links the endothelial cells and myocardial cells for the purpose of regulating cardiac contractility [1]. Interestingly, we have also documented at both the protein and mRNA level expression of apelin and APJ by myocardial cells in the embryonic heart. This finding suggests an autocrine pathway that is important for heart development, and which is functional before establishment of the coronary circulation. Since we did not detect apelin expression by adult myocardial cells, there must be a shift of apelin expression from myocardial to endothelial cells after establishment of the coronary circulation in late gestation. Quantitative evaluation of mRNA levels through late gestation and adulthood indicated that apelin expression is relatively constant, and is consistent with a need to maintain apelin levels for cardiac homeostasis. It is worth noting that in decompensated failing human heart tissues, apelin immunoreactivity was noted in association with myocardial cells, suggesting that this embryonic pathway is reactivated in the setting of congestive heart failure, in parallel with other embryonic programs [1].

In summary, we demonstrate here in vivo effects of the endogenous peptide apelin. Apelin reduces left ventricular preload and afterload, and increases contractile reserve without causing hypertrophy. Together with a propensity to decrease central vasopressin release [8], these findings suggest a prominent role for the apelin-APJ system in cardiovascular homeostasis, and warrant investigation of a potential therapeutic role for exogenous apelin in heart failure.


This work was supported by the Donald W. Reynolds Cardiovascular Clinical Research Center at Stanford University.


  • Time for primary review 15 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
View Abstract