Endocrinology and Metabolism


Atrial natriuretic peptide (ANP) controls lipolysis in human adipocytes. Lipid mobilization is increased during repeated bouts of exercise, but the underlying mechanisms involved in this process have not yet been delineated. The relative involvement of catecholamine- and ANP-dependent pathways in the control of lipid mobilization during repeated bouts of exercise was thus investigated in subcutaneous adipose tissue (SCAT) by microdialysis. The study was performed in healthy males. Subjects performed two 45-min exercise bouts (E1 and E2) at 50% of their maximal oxygen uptake separated by a 60-min rest period. Extracellular glycerol concentration (EGC), reflecting SCAT lipolysis, was measured in a control probe perfused with Ringer solution and in two other probes perfused with either Ringer plus phentolamine (α1/2-AR antagonist) or Ringer plus both phentolamine and propranolol (β-AR antagonist). Plasma epinephrine, plasma glycerol, and EGC were 1.7-, 1.6-, and 1.2-fold higher in E2 than in E1, respectively. Phentolamine potentiated exercise-induced EGC increase during E2 only. Propranolol reduced the lipolytic rate during both E1 and E2 compared with the probe with phentolamine. Plasma ANP concentration increased more during E2 than during E1 and was correlated with the increase in EGC in the probe containing phentolamine plus propranolol. The results suggest that ANP is involved in the control of lipolysis during exercise and that it contributes to stimulation of lipolysis during repeated bouts of exercise.

  • microdialysis
  • insulin
  • epinephrine
  • guanosine 5′,5′-cyclic monophosphate

until recently, the regulation of human adipose tissue lipolysis during exercise was mainly attributed to both the increase in catecholamine levels and the concomitant decrease in plasma insulin concentration. Catecholamines activate lipolysis through β1- and β2-adrenergic receptors (AR) and inhibit it through α2-AR stimulation. The coordinated activation of both receptors modulates the intracellular cAMP concentration, which activates cAMP-dependent protein kinase, leading to the phosphorylation and activation of hormone-sensitive lipase (HSL) (2, 22). The recent discovery that natriuretic peptides (NP) are potent activators of lipolysis in human fat cells raises new questions about the control of lipid mobilization in humans (27, 28). NP stimulate fat cell plasma membrane receptors (NPR-A subtype) bearing an intrinsic guanylyl cyclase activity and elevate intracellular levels of cGMP, which activate a cGMP-dependent protein kinase (PKG). PKG-dependent phosphorylation of perilipin and HSL stimulate lipolysis (27, 28). NP are potent lipolytic agents when they act on isolated fat cells from subcutaneous adipose tissue (SCAT), specifically in primates (29). Intravenous infusion of pharmacological doses of hANP (human atrial natriuretic peptide) in humans promotes a strong lipid-mobilizing effect independently of reflex activation of the sympathetic nervous system (SNS) (12). A similar response occurred when ANP was perfused through a microdialysis probe inserted in SCAT (27). Moreover, within a physiological range, ANP stimulates adipose tissue lipolysis and fat oxidation and increases circulating levels of nonesterified fatty acids (NEFA) (7). Circulating ANP concentrations rise during short-term exercise of increasing intensity; the maximal concentrations of ANP increase two- to threefold (25).

The microdialysis method is useful to monitor local lipid mobilization and changes in cGMP or cAMP release in SCAT. It is the most suitable method available to perform mechanistic explorations of adipose tissue function in vivo (1, 3). Exercise is a prerequisite for both SNS activation (11, 15) and NP release (5, 26). We (25) demonstrated that physiological ANP release during acute endurance exercise contributes to the stimulation of lipid mobilization and NEFA supply for the working muscle.

A previous study showed that hormonal and metabolic responses differed during two repeated exercise bouts of similar intensities (31). A higher secretion of epinephrine and increased lipid mobilization were found in SCAT during the second exercise bout. The hormonal partners responsible for the enhanced lipolysis have not been delineated so far.

The aim of the present study was to examine the relative contributions of catecholamines and ANP in the control of lipolysis during sequential periods of exercise. To do so, healthy male subjects fasted overnight and then performed two exercise bouts (E1 and E2) for 45 min each at 50% of their maximal oxygen uptake (V̇o2 max) separated by a 60-min rest period. To assess the specific role of ANP in the control of SCAT lipolysis, glycerol release was measured under local α- and β-blockade during two repeated bouts of exercise.



Eight healthy young men aged 23 ± 0.6 yr (range 22–27 yr) with a mean body mass index of 24 ± 0.7 kg/m2 (range 21.6–24.9 kg/m2) who had not enrolled in any pharmacological or nutritional protocol before the study were recruited. All had a stable weight during the previous 3 mo. Selection of the subjects was based on detailed medical history, a physical examination, and several blood chemistry analyses. One week before the investigation period, the V̇o2 max was determined during a maximal-effort test performed on an electrically braked bicycle ergometer (Ergometrics 800s, Ergoline) by the use of an incremental procedure (work rate increasing by 25 W/min). Heart rate (Baumann BHL 6000 cardiometer) was monitored throughout. V̇o2 max was measured using a Vmax apparatus (Sensor Medics), and the highest V̇o2 value was considered as V̇o2 max. The mean V̇o2 max was 43.9 ± 1.6 ml·kg−1·min−1 (range 39.5–46.2 ml·kg−1·min−1). The Ethics Committee of the Third Faculty of Medicine, Charles University, approved the study. All of the subjects gave their informed consent for the experimental conditions after detailed explanation.

Experimental protocol.

The subjects entered the hospital at 8:00 AM after an overnight fast and were maintained in a semirecumbent position. An indwelling polyethylene catheter was inserted into the antecubital vein for blood sampling. At 8:30 AM, three microdialysis probes (Carnegie Medicine, Stockholm, Sweden) of 20 × 0.5 mm and 20,000-MW cut-off were inserted percutaneously after epidermal anesthesia (200 μl of 1% lidocaine; Roger-Bellon, Neuilly-sur-Seine, France) into the SCAT at a distance of 10 cm from the umbilicus. The probes were connected to a microperfusion pump (Harvard apparatus, S.A.R.L., Les Ulis, France) and perfused with Ringer solution (in mmol/l: 139 sodium, 2.7 potassium, 0.9 calcium, 140.5 chloride, 2.4 bicarbonate). Ethanol (1.7 g/l) was added to the perfusate to estimate changes in the blood flow, as previously described (10, 32). One probe was perfused with Ringer solution. The second probe was perfused with 100 μmol/l phentolamine (an α1,2-AR antagonist). The third probe was perfused with 100 μmol/l phentolamine plus 100 μmol/l propranolol (a nonselective β-AR antagonist).

After a 60-min equilibration period, a 30-min fraction of dialysate was collected at a flow rate of 0.5 μl/min. Then, the perfusion flow rate was set at 2.5 μl/min for the remaining experimental period. The estimated extracellular glycerol concentration (EGC) was calculated by plotting (after log transformation) the concentration of glycerol in the dialysate measured at 0.5 and 2.5 μl/min against the perfusion rates. The average recovery of probes was calculated as the ratio of the concentration of glycerol in the dialysate at the perfusion rate of 2.5 μl/min and the calculated EGC. The mean recovery was 25 ± 3%. The EGC found in the present study fits with previously reported values (19, 32).

The subjects then performed exercise (E1) for 45 min with a load corresponding to 50% of their V̇o2 max on the cycle ergometer. Heart rate was continuously monitored with a Baumann BHL 6000 cardiometer. The subjects were then allowed to rest in the semirecumbent position for 60 min, after which they performed a second 45-min exercise bout (E2) with the load adapted such that their V̇o2 was kept constant and remained similar to that recorded during the first exercise. Water intake was allowed ad libitum during the exercise and recovery periods.

Before and during the exercise bouts, 5 ml of blood were collected from the catheter. The catheter was kept patent by slow infusion of saline. Blood was collected in 50 μl of an anticoagulant and antioxidant cocktail (Immunotech, Marseilles, France), to prevent the oxidation of catecholamines, and was immediately centrifuged in a refrigerated centrifuge. The plasma was stored at −80°C until analysis.

Drugs and analytic methods.

Propranolol chlorhydrate (Avlocardyl) and phentolamine methanesulfonic acid (Regitine) were obtained from Zeneca Pharma (Cergy, France) and Novartis (Reuil-Malmaison, France), respectively. Glycerol in dialysate and in plasma was analyzed with an ultrasensitive radiometric method (9). Ethanol in dialysate and perfusate (5 μl) was determined with an enzymatic method (6). Plasma glucose and NEFA were determined with a glucose oxidase technique (Biotrol, Paris, France) and an enzymatic procedure (Wako, Unipath, Dardilly, France), respectively. Plasma insulin concentrations were measured using RIA kits from ICN Pharmaceuticals (Orsay, France). Plasma epinephrine and norepinephrine were assayed in 1-ml aliquots of plasma by high-pressure liquid chromatography using electrochemical (amperometric) detection; the detection limit was 10 pg/sample. Extracellular cGMP and cAMP concentrations were determined using an RIA kit (Cayman Chemical, Ann Arbor, MI). Plasma ANP and growth hormone (GH), collected on EDTA (1 mg/ml) and aprotinin, were determined using RIA kits from Peninsula Laboratories (San Carlos, CA).

Statistical analysis.

All values are means ± SE. The responses to exercise were analyzed by a repeated-measurement analysis of variance (ANOVA). Plasma and extracellular concentration response curves were calculated as the total integrated changes over baseline values. Significance values are quoted in the figures and the table. P < 0.05 was considered statistically significant.


General observations.

The heart rate rose rapidly over the first 5 min during exercise, and the power developed was regularly adjusted to maintain the V̇o2 constant over each exercise bout. The heart rates were similar during the two exercise bouts (138 ± 2 and 142 ± 3 beats/min during E1 and E2, respectively). The power developed during E1 (121.5 ± 8.1 W) was significantly higher than that developed during E2 (113.8 ± 1.6 W).

Plasma catecholamines and GH concentrations.

Plasma norepinephrine and epinephrine concentrations are depicted in Table 1. At rest, plasma norepinephrine and epinephrine concentrations were 294 ± 28 and 62 ± 4 pg/ml, respectively. At the end of period E1, the plasma norepinephrine concentration rose significantly (901 ± 75 pg/ml). At the end of the recovery period (duration 135 min), the norepinephrine concentration returned to a value that did not differ from that measured in basal conditions (time 0 min). No significant difference was found in the plasma norepinephrine concentration kinetics during the second exercise bout. At the end of period E2, the plasma norepinephrine concentration rose significantly to 992 ± 45 pg/ml. Plasma epinephrine concentration rose significantly from 62 ± 4 to 109 ± 9 pg/ml at the end of E1. During recovery, epinephrine concentration decreased toward a value that did not differ from that measured in basal conditions (time 0 min). During E2, the plasma epinephrine concentrations rose dramatically to values (185 ± 38 pg/ml) significantly higher than those measured during E1. Plasma GH concentrations increased during E1 and E2 (Table 1). Exercise-induced GH release was about twofold lower during E2 (P < 0.05).

View this table:
Table 1.

Plasma norepinephrine, epinephrine, glucose, insulin, GH, and ANP concentrations during E1 and E2

Plasma glucose, insulin, and ANP concentrations.

No significant changes in plasma glucose level were observed during E1. A significant decrease in plasma glucose concentration was observed at the end of E2 and during the second recovery period. Plasma insulin concentration decreased during E1 (Table 1). During recovery, insulin concentration returned to values that did not differ from those measured in basal conditions (time 30 and 135 min). At the end of period E2, plasma insulin concentrations decreased to values significantly lower than those measured during E1 (P < 0.001). At the end of E1, the plasma ANP concentration rose significantly. At the end of the recovery period, ANP concentrations decreased to a value that did not differ from that measured in basal conditions. At the end of E2, the plasma ANP concentration rose significantly towards values higher than during E1 (Table 1).

Plasma NEFA and glycerol levels.

Plasma NEFA concentration was unchanged throughout E1 (Fig. 1); then, after the first 15 min of the recovery period, it increased (P < 0.01) and finally decreased toward values similar to those found in basal conditions at time 0 min. During E2, the plasma NEFA concentration increased significantly compared with values measured during E1. Plasma glycerol increased progressively during E1 and then, during the recovery period, returned to values that were not significantly different from those determined before exercise (time 0 min). During E2, a significant increase in plasma glycerol concentration was observed compared with E1.

Fig. 1.

Plasma nonesterified fatty acid (NEFA) and glycerol concentrations during repeated bouts of 45-min cycle ergometer exercise (E1 and E2). Data are expressed as means ± SE. +P < 0.05 compared with values measured before E1 or E2. P < 0.05 compared with corresponding values measured during E1.

Changes in EGC during the two exercise bouts.

At rest, before E1, the mean basal EGC did not differ between the control probe (160.6 ± 16.9 μmol/l) and the probe with phentolamine (163.3 ± 20.2 μmol/l). The EGC in the probe with phentolamine plus propranolol (132.6 ± 15.7 μmol/l) was significantly reduced compared with the two other probes. In the probe containing phentolamine, the EGC were not different from those of the control probe during E1 but were higher than the control during E2 (Fig. 2B). In the probe containing phentolamine plus propranolol, the EGC were lower than in the control and also lower than in the probe with phentolamine alone during E1. During E2, the EGC were lower compared with the phentolamine probe but not with the control probe. Figure 2A represents the differences in exercise-induced responses during E1 compared with E2. Regardless of the presence of α-blockade or combined α- and β-blockade, the increase in EGC was more pronounced during E2 than during E1. EGC was significantly increased during E2 in the probe perfused with phentolamine but not during E1. EGC was ∼35% lower in the phentolamine-plus-propranolol probe during E1 than in the control probe. This effect did not reach significance during E2.

Fig. 2.

A: changes in extracellular glycerol concentrations (EGC) during E1 and E2 in the control probe (□), the probe perfused with phentolamine (▪), and the probe perfused with phentolamine + propranolol (gray square). Changes were calculated by the difference between the mean values obtained during exercise (E1 or E2) and the value measured before exercise E1 or E2. Data are expressed as means ± SE. *P < 0.05 compared with E1. NS, not significant. B: EGC in subcutaneous adipose tissue (SCAT) from the control probe (□), the probe perfused with phentolamine (▪), and the probe perfused with phentolamine + propranolol (gray square) during repeated bouts of 45-min cycle ergometer exercise (E1 and E2). Data are expressed as means ± SE. P < 0.05, probe with phentolamine compared with control probe; ¥P < 0.05, probe with phentolamine + propranolol compared with control probe; #P < 0.05, probe with phentolamine + propranolol compared with probe with phentolamine.

Finally, a correlation was found between the increase in plasma ANP and the increase in EGC evaluated from the probe containing phentolamine plus propranolol (Fig. 3).

Fig. 3.

Correlation between plasma atrial natriuretic peptide (ANP) and EGC changes from baseline in the probe with phentolamine plus propranolol during E1 (○) and E2 (•). Data are means ± SE of 6 subjects.

Changes in extracellular cAMP and cGMP levels during the two exercise bouts.

We determined that extracellular cAMP and cGMP levels in SCAT do discriminate between the relative contribution of catecholamines and ANP during exercise-induced lipolysis. These levels were determined from the pooled dialysates obtained at rest and during periods E1 and E2 in the probes with phentolamine alone or associated with propranolol. The determinations were performed in six subjects. The results, depicted in Fig. 4, show that extracellular cGMP concentration increased significantly (P < 0.04) and similarly in each probe during E1. Extracellular cGMP tended to increase more during E2 than during E1, but the difference did not reach significance (P < 0.09). Extracellular cAMP concentration increased significantly during E1 (P < 0.04) regardless of the presence of α-blockade or combined α- and β-blockade, and cAMP increase tended to be higher during E2 than during E1. In addition, it was observed that the level of cAMP was reduced in the probe containing phentolamine plus propranolol at rest and during E1 and E2 compared with the probe with phentolamine alone.

Fig. 4.

Extracellular cGMP and cAMP concentrations from SCAT, the probe with phentolamine (α1/2-blocker), and the probe with phentolamine + propranolol (β1/2-blocker) during repeated bouts of 45-min cycle ergometer exercise (E1 and E2). Data are expressed as means ± SE. *P < 0.05 compared with probe with phentolamine + propranolol.

Modifications of adipose tissue blood flow in SCAT during exercise bouts.

Local changes in adipose tissue blood flow (ATBF) were evaluated using the ethanol washout method. The ethanol ratio was calculated in percent ratio of the ethanol concentration measured in the dialysate divided by the ethanol concentration measured in the perfusate × 100 for the dialysate from each probe and taken as an index of ethanol washout. A high ethanol ratio corresponds to a lower ethanol washout and a lower regional ATBF. During the two resting periods, the ethanol outflow/inflow ratio was not different among the three probes. During the exercise bouts, a significant decrease in ethanol ratio was observed in the probe with phentolamine, the decrease being higher during E2 than during E1. No exercise-induced changes in ethanol ratio were observed in the control probe or in the probe combining the two AR antagonists.


This study reveals the major contribution of ANP in the stimulation of lipid mobilization from SCAT during repeated bouts of endurance exercise. Numerous works have mentioned the role of epinephrine and insulin in this process, but the role of ANP has never been evoked. Sequential 45-min exercise bouts separated by a 60-min rest were used to promote SNS activation and ANP release. On the basis of the use of the microdialysis technique and local selective blockade of α2- and β1/2-ARs within SCAT, we showed that more than 50% of the nonadrenergic lipolysis observed during a repeated bout of exercise is ANP dependent.

Catecholamines released under SNS activation during exercise were long considered to be the major agents controlling lipid mobilization from adipose tissue in humans (16, 34). We recently demonstrated (27) that NP act in a new pathway to control human fat cell lipolysis in vitro or when administered intravenously. Using the microdialysis method, we previously demonstrated that repeated bouts of exercise at moderate intensity (50% V̇o2 max) elicited a greater increase in lipid mobilization in SCAT of healthy men than exercise without repetition. Because when these experiments were performed the lipolytic effect of NP was unknown, the enhanced lipolytic response during the second exercise was attributed to a higher epinephrine response and reduced plasma levels of insulin (31).

We (25) previously verified that local infusion of propranolol was able to efficiently counteract the lipolytic effect of isoproterenol, a β-AR agonist, in SCAT. However, investigations have shown that, during exercise, addition of propranolol to the perfusion probe partly reduced the exercise-induced increase in glycerol levels in human SCAT (2, 15). During exercise, local α2-AR blockade enhances the glycerol output from SCAT (calculated from the probe perfused with phentolamine). This potentiating effect of the antagonist did not reach a significant level during E1 but became significant during the second exercise, E2. Plasma epinephrine concentrations were 1.7-fold higher during E2 than during E1, whereas the exercise-induced increase in plasma norepinephrine level was similar (Table 1). Thus the increased EGC in the presence of phentolamine had previously been attributed to the local suppression of the α2-AR-dependent antilipolytic effect of epinephrine on fat cells (32). In this study, in the presence of both propranolol and phentolamine, exercise-induced lipolysis in SCAT was only partly reduced (∼35% during both E1 and E2) compared with the probe containing phentolamine alone. The residual increment of EGC remaining after β1/2- and α2-AR blockade suggests that nonadrenergic factors (possibly including the exercise-induced inhibition of insulin release and ANP release) contribute to the stimulation of lipolysis during exercise (Fig. 1).

The major evidence supporting the role of ANP in glycerol production is that the lipolytic rate (assessed by EGC) around the probe perfused with both phentolamine and propranolol was significantly higher during E2 than during E1 (during E1 and E2, the increase in EGC representing 65 and 74% of the increase found in the control probe, respectively; Fig. 1). In a previous study (25), we found that, during exercise, extracellular cGMP concentrations also increase, reflecting local ANP action within SCAT. In the present study, it was observed that the amount of extracellular cGMP was similar in the probe perfused with phentolamine alone or associated with propranolol (Fig. 2). In contrast, the concentration of cAMP was reduced in the probe combining the two pharmacological agents, so the lipid mobilization remaining after the local blockade of fat cell β1/2- and α2-ARs has another origin. Nonadrenergic-dependent lipolysis is correlated with both plasma ANP (Fig. 3) and extracellular cGMP measured in SCAT. Thus the ANP-dependent lipolytic pathway appears to be involved in the enhancement of lipolysis in SCAT and responsible for the residual mobilization under adrenergic receptor blockade.

The fall in plasma insulin during exercise could also be involved in the enhancement of lipolysis due to the reduction of its antilipolytic effect. This decrease was more pronounced during E2. Insulin is an important factor controlling lipolysis in fat cells (14, 17, 24). The partial enhancement of extracellular cAMP (via a reduction of phosphodiesterase-3 activity) in the probe containing phentolamine plus propranolol suggests that insulin could be another candidate for the observed increase in lipolysis during exercise. To assess the contribution of insulin to exercise-induced lipolysis, it could be of interest to study the effect of two exercise bouts (E1 and E2) on lipid mobilization under conditions of stable insulinemia and glycemia using euglycemic insulinemic clamp. Another hormonal stimulus possibly contributing to the exercise-induced lipolysis is GH, as its circulating levels increase during exercise (21, 31). Its mechanism of action involves adenylyl cyclase/Gs-Gi protein complex and cAMP production (35). However, GH cannot act in the probe perfused with both phentolamine and propranolol, because these two pharmacological drugs inhibit Gi- and Gs-dependent pathways, respectively. Several other factors are known to modulate adipose tissue lipolysis. Among them, cortisol, parathyroid hormone (PTH), interleukin-6, and leptin are relevant candidates. Cortisol has a permissive effect on catecholamine-induced lipolysis, but an effect during short-term exercise is unlikely. PTH exerts a weak lipolytic effect in vitro in human fat cells (8) and stimulates lipid mobilization in humans; however, it acts at rather high concentrations that are clearly in an extraphysiological range. Interleukin-6 has been shown to stimulate lipid mobilization and oxidation in humans (33). Its effect appears at concentrations that are not encountered during endurance exercise, and a possible role of GH is not excluded (20). Leptin stimulates lipolysis in vitro in rodent fat cells but not in humans (30), and its plasma levels do not change during acute exercise (13). No data are available on a possible role of adiponectin in the regulation of lipolysis. Plasma adiponectin levels are not acutely regulated by exercise either (18). To sum up, a physiological lipolytic role of all the aforementioned factors seems unlikely during the acute bouts of endurance exercise used in this study.

For many years, exercise-induced lipolysis was attributed to catecholamines via the stimulation of fat cell β-AR. After the discovery of an antilipolytic action of catecholamines via the activation of α2-AR (4, 23), it was proposed that epinephrine, which exhibits a higher affinity for α2-AR than for β-AR, could exert a negative control on lipolysis. In vivo studies using microdialysis during exercise have revealed the physiological relevance of this pathway (15, 32). In this study, the lipolytic β-adrenergic responsiveness of catecholamines is totally counterbalanced by activation of α2-adrenergic antilipolytic receptors during E2. This effect could be related to the sustained rise in plasma epinephrine level during E2 in response to the decrease in plasma blood glucose concentration. Finally, after the discovery of the lipolytic effect and lipid-mobilizing action of NP (25, 27), it is proposed that the nonadrenergic lipolysis observed during the second exercise bout depends on ANP release. However, the lack of an ANP receptor antagonist limits reaching a final decision.

To conclude, this study raises the fundamental question of the role of ANP in the control of lipid mobilization during repeated bouts of endurance exercise. Besides epinephrine and insulin, ANP may act as a relevant physiological agent controlling exercise-induced lipolysis. The role played by unsuspected lipid-mobilizing factors like ANP in the control of lipolysis must be studied in greater detail under various physiological and pathological conditions. Obese patients with or without insulin resistance should be included in these studies, as should highly trained subjects, in whom plasma ANP response to exercise or adipose tissue sensitivity to ANP could change.


The study was supported by the Research Project of Czech Republic No. MSM0021620814.


We express our gratitude to Marie-Thérèse Canal and Zuzana Pariskova for contributing to the study. We are also indebted to Marie-Antoinette Tran for laboratory support in catecholamine measurements, and to Prof. Alexander Sorisky and Dr. Peter Winterton for revising the English version of the manuscript.


  • 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.


View Abstract