Author + information
- Received January 17, 2011
- Revision received May 3, 2011
- Accepted May 31, 2011
- Published online September 6, 2011.
- Jan Polak, MD, PhD⁎,‡,§∥,
- Martin Kotrc, MD⁎,
- Zuzana Wedellova, MD§∥,
- Antonin Jabor, MD, PhD⁎,
- Ivan Malek, MD, PhD⁎,
- Josef Kautzner, MD, PhD⁎,
- Ludmila Kazdova, PhD† and
- Vojtech Melenovsky, MD, PhD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Vojtech Melenovsky, Department of Cardiology, Institute of Clinical and Experimental Medicine, Videnska 1958/9, Prague 140 21, Czech Republic
Objectives Our goal was to examine the role of B-type natriuretic peptide (BNP) in lipolysis regulation in heart failure (HF) patients.
Background Enhanced adipose tissue lipolysis can contribute to myocardial lipid overload, insulin resistance, and cachexia in advanced HF. Natriuretic peptides were recently recognized to stimulate lipolysis in healthy subjects.
Methods Ten nondiabetic HF patients (New York Heart Association functional class III, 50% nonischemic etiology) and 13 healthy subjects (control subjects) of similar age, sex, and body composition underwent a microdialysis study of subcutaneous abdominal adipose tissue. Four microdialysis probes were simultaneously perfused with 0.1 μM BNP1–32, 10 μM BNP1–32, 10 μM norepinephrine (NE) or Ringer's solution. Outgoing dialysate glycerol concentration (DGC) was measured as an index of lipolysis.
Results Spontaneous lipolysis was higher in HF patients compared with control subjects (DGC: 189 ± 37 μmol/l vs. 152 ± 35 μmol/l, p < 0.01). Response to NE was similar (p = 0.35) in HF patients and control subjects (DGC increase of 1.7 ± 0.2-fold vs. 1.7 ± 0.4-fold). BNP1–32 10 μM markedly increased lipolysis in both HF patients and control subjects (DGC increase of 2.8 ± 0.5-fold vs. 3.2 ± 0.3-fold), whereas the response to 0.1 μM BNP1–32 was more pronounced in HF patients (p = 0.02). In HF patients, spontaneous lipolysis positively correlated with insulin resistance and the response to BNP1–32 negatively correlated with adiposity.
Conclusions BNP1–32 exerts strong lipolytic effects in humans. Despite marked elevation of plasma immunoreactive BNP, the responsiveness of adipose tissue to BNP1–32 is not attenuated in HF, possibly reflecting a deficiency of endogenous bioactive BNP. Lipolytic effects of BNP can contribute to excessive fatty acid mobilization in advanced HF.
A chronic heart failure (HF) state triggers a neuroendocrine response that stimulates the release of free fatty acids (FFAs) from adipose tissue (AT) (1). Although FFAs are the major substrate for cardiac metabolism, persistently increased FFA efflux from AT might have adverse effects on cardiac function (2) by impairing insulin sensitivity (3), diminishing cardiac efficiency (4), and contributing to myocardial lipotoxicity (5–7). Enhanced lipolysis also contributes to body weight loss and cardiac cachexia, which are linked with a particularly poor outcome in HF (2,8–10). Increased adrenergic-fatty acid load in HF has been suggested as a potential therapeutic target (11–14).
Mechanisms of increased AT lipolysis in HF patients are incompletely understood. Lipolysis in HF patients is stimulated by increased catecholamines (15) and tumor necrosis factor-α (16). Recently, it was recognized that natriuretic peptides (NPs) are also capable of stimulating lipolysis in adipocytes (17,18). Lipolytic role of atrial natriuretic peptide (ANP) was first suggested by in vitro experiments (19,20) and subsequently confirmed by in vivo studies by our and other groups (20,21). ANP appears to be a major regulator of lipolysis during prolonged exercise (22) in healthy subjects, surpassing the contribution of catecholamines (14,23). ANP is also responsible for the regulation of residual lipolysis in situations of pharmacological beta-adrenergic blockade (2,23). Much less is known about the role of B-type natriuretic peptide (BNP) in lipolysis regulation. Importantly, no studies have examined the lipolytic role of BNP in HF patients, despite important links between NP system activation, enhanced lipolysis, and possible progression of HF.
BNP is released from the heart as inactive pro-BNP1–108 in response to increased wall stress (6,9,24) and is then cleft into the bioactive form BNP1–32 by specific proteases (9,10,14,17). Despite the fact that circulating immunoreactive BNP (iBNP) is typically increased in HF and serves as a hallmark of the disease (25,26), the functional effects of BNP on target tissues are influenced by abnormal processing/decreased generation of bioactive BNP species (active-BNP deficiency) or by attenuated tissue response (BNP resistance). Which of these 2 mechanisms prevail in AT of HF patients is currently unknown; however, this question can be addressed by examining in vivo lipolytic responses by the administration of exogenous bioactive BNP1–32. BNP-mediated lipolysis would be attenuated if the latter mechanism prevails or preserved if the former mechanism is operational.
The aim of this study was to examine the effects of BNP1–32 on regulation of lipolysis in vivo and to compare healthy subjects with advanced HF patients. We used the unique technique of microdialysis, which allows tissue delivery of pharmacological substances into subcutaneous AT with simultaneous sampling of extracellular fluid for metabolite analyses.
Men with symptomatic HF (n = 10) were recruited at the Cardiology Department of IKEM, Prague, Czech Republic. HF was defined by Framingham criteria as a history of HF hospitalization, documented systolic dysfunction (left ventricular ejection fraction <35%), and persistent symptoms (New York Heart Association functional class of II or higher for >3 months). Importantly, we excluded patients with diabetes mellitus and hemodynamic instability, infection, severe renal failure or anemia (creatinine >300 μmol/l, hemoglobin <100 g/l), endocrine disease, restrictive or hypertrophic cardiomyopathy. HF etiology was coronary artery disease in 50% and nonischemic in the rest. All patients were treated with furosemide (54 ± 8 mg/day), 9 with beta-blockers and 7 with angiotensin-converting enzyme inhibitors or angiotensin receptor antagonists. Healthy male controls (n = 13), free of any disease or medication, were recruited by advertisement. All subjects were nonsmokers. The study was approved by the local ethics committee, and all subjects signed informed consent.
Symptoms were quantified using a Minnesota Living With Heart Failure Questionnaire, body composition was assessed using dual-energy X-ray absorptiometry (Lunar Prodigy, GE Healthcare, Waukesha, Wisconsin), and skinfold thickness was measured with Best's calipers (nondominant triceps, biceps, subscapular, and supra-iliac region). Patients were weighed (Omron HBF510W, Tokyo, Japan), underwent cardiac ultrasonography (Vivid7, GE Healthcare), and their height was measured with a stadiometer. For biochemical tests, venous blood samples were obtained after an overnight fast and plasma was stored at −80°C until analysis. An oral glucose tolerance test was performed on a separate day (75 g of glucose/300 ml of water), with sampling of venous blood at 0, 30, 60, and 120 min. All variables were collected within ±5 days from the microdialysis experiment.
Microdialysis is a mini-invasive method evaluating AT metabolism in vivo (27). Semipermeable membrane probes are perfused (with or without pharmacological substances) at low flow rates until equilibrium is established between interstitial tissue fluid and dialysate solution in the probe. Glycerol released from adipocytes after triglyceride hydrolysis is used as a lipolysis marker. Microdialysis perfusions with 2 different concentrations of BNP1–32 (0.1 μM and 10 μM) were based on previous in vitro data (28) showing maximum lipolytic response with 10 μM and ∼50% of maximum effect with a 0.1-μM concentration (2,8,29,30). The subjects entered the research unit at 8:00 am after an overnight fast. All heparin was withheld for more than 24 h before tests. At 8:30 am, 4 microdialysis probes (20,000-kDa cutoff, CMA Microdialysis, Stockholm, Sweden) were inserted in the subcutaneous abdominal AT. The probes were connected to a microperfusion pump (Harvard Apparatus, Les Ulis, France) and perfused with Ringer's solution (Baxter, Prague, Czech Republic). Ethanol (1.7 g/l) was added to the perfusate to estimate changes in the blood flow, as described previously (9,10,14). The first probe was perfused with Ringer's solution (control probe), the second probe with 10 μM norepinephrine (Leciva, Prague, Czech Republic), the third probe with 10 μM of human recombinant BNP1–32 (nesiritide-Noratak, Janssen-Cilag, Baar, Switzerland) and the fourth probe with 0.1 μM BNP1–32. The study consisted of 80-min baseline sampling and 60 min of pharmacological stimulation. Dialysate samples were collected at 20-min intervals at a flow rate of 1 μl/min throughout the experiments. The dialysate glycerol concentration (DGC) was measured as a marker of lipolysis.
To evaluate the relative recovery of microdialysis probes (a parameter describing permeability characteristics of the membrane), 4 probes were submerged in a vial containing 0.1 μM BNP1–32 dissolved in saline solution. After 3-h equilibration, probes were perfused with saline solution and 0.1% albumin at flow rates 0.5, 1, 2, 3.5, and 5 μl/min. BNP1–32 was measured in dialysate and in surrounding BNP1–32 solution and log-transformed concentrations were plotted against perfusion rates to calculate the relative recovery at each perfusion rate. This experiment demonstrated a relative recovery of 20 ± 2% for BNP1–32 at perfusion rate 1 μl/min.
The iBNP concentrations were measured by chemiluminescent microparticle immunoassay (Architect BNP, Abbott Laboratories, Abbott Park, Illinois), insulin by IRMA assay (Immunotech, Prague, Czech Republic), FFA by immunoturbidimetry (Wako, Richmond, Virginia), and glycerol by Randox Glycerol Kit (Randox Laboratories Ltd., Antrim, United Kingdom).
Statistical analysis was performed using SPSS version 13.0 for Windows (SPSS Inc., Chicago, Illinois). The response of AT to pharmacological perfusions was evaluated in the HF and control groups separately using Student's paired t test. Differential responses to BNP1–32, norepinephrine, or Ringer's solution perfusion between groups were evaluated using an ordinary t test. Between-group differences in anthropometric and biochemical variables (Tables 1 and 2)⇓⇓ were evaluated using the Mann-Whitney test. Associations between variables were investigated using the Spearman method. Data are presented as mean ± SEM unless stated otherwise. A level of p ≤ 0.05 was considered statistically significant.
Spontaneous lipolysis at baseline was evaluated by analyzing DGC in all samples obtained throughout the last hour of a baseline period. HF patients had higher spontaneous lipolytic rate by 19% than control subjects (189.0 ± 37.3 μmol/l vs. 152.1 ± 35.1 μmol/l, p < 0.01) (Fig. 1A). Spontaneous lipolysis (DGC in a control probe) remained unchanged in healthy volunteers, but decreased by 22% in the HF group (p = 0.01) during the experiment period.
Response to BNP1–32 and norepinephrine
Compared with baseline DGC, the perfusion with 10 μM BNP1–32 increased lipolysis by 2.8 ± 0.5-fold and 3.2 ± 0.3-fold (both p < 0.05) in control and HF subjects, respectively. The increment in lipolysis induced by 10 μM BNP1–32 was similar between groups (p = 0.61). The lipolytic response to the lower BNP1–32 concentration (0.1 μM) differed between HF and control subjects (p = 0.02); lipolysis increased by 17% (p = 0.056) in HF subjects, but remained unchanged in the control group (p = 0.39). No differences were observed between the HF and control groups (p = 0.35) in norepinephrine-induced lipolysis (1.7 ± 0.4-fold and 1.7 ± 0.2-fold increase in controls and HF, p = 0.04 and p = 0.06, respectively). Analysis of absolute differences in DGC (in μmol/l) after BNP1–32 perfusions revealed a trend toward a higher response in HF than in control subjects both at low and high BNP1–32 concentrations (by 23% and 41%, respectively, p = 0.1 and p = 0.07, respectively). Data are summarized in Figures 2 and 3⇓⇓.
Determinants of spontaneous lipolysis and response to BNP1–32
Differential regulation of lipolysis was observed between HF and control groups. Spontaneous DGC was associated with adiposity and insulin resistance evaluated as the area under the plasma insulin curve (AUCinsulin) during the oral glucose tolerance test (r = 0.67, p = 0.01 and r = 0.88, p < 0.01) in the control group, whereas these associations were absent in the HF subjects (r = 0.31, p = 0.5 and r = −0.52, p = 0.18). No associations were observed between basal lipolytic rate and plasma iBNP in the HF or control group (r = 0.41, p = 0.24 and r = 0.5, p = 0.1, respectively) (Fig. 4A). However, lipolytic response to 10-μM BNP1–32 perfusion was inversely related to fat mass (%) in the HF group (r = −0.90, p < 0.01), but no relationship with fat mass or other investigated variables (fasting plasma insulin, homeostasis model assessment of insulin resistance, INS120) was observed in the control group (Fig. 4B).
Changes in AT blood flow and AT cyclic guanosine monophosphate production during BNP1–32 administration
Blood flow in AT was evaluated using the ethanol concentration outflow/inflow ratio, as described previously (14,31). No differences were observed in the HF and control groups at baseline or after 10-μM BNP1–32 perfusion, which induced vasodilation in both groups. In contrast, norepinephrine administration led to vasodilation in control subjects only (Fig. 1B). Additionally, after the 10-μM BNP1–32 perfusion in a separate group of 4 healthy volunteers, we observed a 43.2 ± 12.8-fold increase (p < 0.05) in AT cyclic guanosine monophosphate concentration that was associated with a concomitant increase in DGC (r = 1, p < 0.001). These results indicate that cyclic guanosine monophosphate–dependent lipolytic pathway is activated in adipocytes during BNP administration.
We present several new findings concerning the regulation of lipolysis in AT in HF. In nondiabetic advanced HF patients, spontaneous lipolysis is enhanced compared with body composition–matched healthy controls. In both groups, administration of BNP1–32 leads to a strong lipolytic response. Moreover, the responsiveness of AT to BNP1–32 is not attenuated, but rather increased in HF, and it is inversely proportional to body fat mass.
Increased spontaneous lipolysis in HF
Humoral factors including NPs (23), catecholamines (14,32), insulin (33), and drugs (e.g., beta-blockers) contribute to lipolysis regulation in HF. The finding of increased spontaneous lipolysis in HF is consistent with a report of increased plasmatic FFA flux in the HF state (1). Lipolysis is physiologically inhibited by insulin, and altered insulin action thus regulates lipolytic rate. Indeed, insulin resistance is frequently observed in nondiabetic HF patients (30). Furthermore, higher plasma insulin levels during the oral glucose tolerance test (representing a dynamic test of glucose metabolism) were associated with greater spontaneous lipolysis in the HF group but not in the control group in our study, indicating that insulin resistance and the attenuated antilipolytic effect of insulin might contribute to greater lipolysis in HF.
BNP1–32 induced strong lipolytic effects in both control and HF subjects
The effect of bioactive BNP1–32 on AT lipolysis of both groups is in line with previous experiments using local, systemic, or in vitro administration of ANP (14,34,35). The magnitude of lipolysis induced by 10 μM BNP1–32 is comparable to that observed after administration of 10 μM ANP (14); however, because the subjects in this study differ from the previous in several cofounders (e.g., body composition, age) further BNP1–32 studies are needed to directly compare in vivo lipolytic potency of ANP and BNP. Importantly, because the effects of catecholamines are attenuated in HF by beta-blocker therapy, NPs might be even more important regulators of lipolysis in HF subjects than in control subjects. Despite the fact that HF is often associated with resistance of vascular and renal tissues to the effects of NPs due to multiple mechanisms (36), we did not observe diminished lipolysis induced by BNP1–32 in the HF group, suggesting that NP resistance in HF is not ubiquitous in all tissues.
Heightened AT responsiveness to BNP1–32
A perfusion with 2 concentrations of BNP1–32 allowed us to partly evaluate the BNP sensitivity of AT. Although the response was similar in both groups with 10-μM perfusion, lipolysis activation with 0.1 μM was higher in HF subjects, indicating increased sensitivity of AT to BNP. How then can most of the HF patients maintain their stable fat mass, regardless of a large increase in iBNP? This contradiction could be explained by the notion that bioactive BNP species are actually deficient in HF subjects (36). Mass spectrometry analysis showed that circulating iBNP in HF patients contains mostly biologically inactive BNP pro-forms (20,37). Here we show higher AT sensitivity to BNP1–32 despite elevated circulating iBNP and thus provide indirect in vivo support for the hypothesis of active BNP deficiency in HF. Additionally, the absence of an association between iBNP and spontaneous lipolysis in our study further supports this hypothesis.
Interestingly, we observed an inverse association between BNP1–32-induced lipolysis and fat mass in HF. In obese subjects, lower lipolytic responsiveness to NPs and therefore lower FFA flux may translate into fewer detrimental effects on cardiac function (5–7) and could underlie a “paradoxical” association between increased fat mass and better HF prognosis (38). Increased NP clearance receptors in AT of obese subjects (39) or the effects of a high-fat diet (40) can explain lower lipolytic BNP1–32 responses in obese subjects (40). Furthermore, the ability of BNP to regulate whole-body energy homeostasis (41) via enhanced mitochondrial biogenesis and lipid oxidation in skeletal muscle (40) might also modify plasma FFA levels; however, it remains unclear how these mechanisms are operational in HF. Subsequently, AT might become more sensitive to BNP1–32 as body fat is decreasing (e.g., during the weight loss or in cardiac cachexia), further perpetuating the vicious cycle of weight loss.
Effects of BNP1–32 on AT blood flow
Because changes in AT blood flow affect DGC, we evaluated AT blood flow during perfusions. Neither AT blood flow at baseline nor the vasodilatory response to 10 μM BNP1–32 differed between groups. No effect of norepinephrine on AT blood flow was observed in the HF group, whereas vasodilation occurred in control subjects.
Our work has identifiable limitations. First, it was limited to advanced HF in middle-aged men without diabetes and the number of examined subjects is rather small. Second, concentrations of BNP1–32 used in microdialysis perfusions were substantially higher than physiological levels in AT, and it is thus possible that using lower doses would provide additional information in a more physiological range. Third, the HF group was somewhat older than control subjects. Because aging is typically associated with diminished lipolysis (33,42), our results might underestimate the lipolysis of HF compared with younger controls. Future studies using systemic BNP1–32 infusion at more physiological concentrations are needed to extend our findings, help to delineate the real cause-and-effect relationship, and describe other endocrine and metabolic changes induced by BNP (e.g., changes in adipokine levels and substrate use). Additionally, studies comparing patients with different HF severity across a range of adiposity and insulin resistance will provide important and clinically relevant information.
Our observations suggest that the lipolytic rate is increased in patients with HF. Furthermore, induction of lipolysis by BNP1–32 is increased in HF patients compared with control subjects, which suggests that BNP might play an important role in excessive FFA mobilization and related metabolic abnormalities including cardiac cachexia.
The authors thank Zuzana Parizkova and Hana Gregorova for their excellent technical assistance, nurse coordinators Hana Podlesakova and Josef Halaxa, and Dr. Petr Tuma, PhD, for his help with biochemical analyses.
This work was supported by Ministry of Health (MZO-00023001 to Dr. Melenovsky, NS10497-3/2009 to Drs. Melenovsky and Pola); Ministry of Education (MSMT-1MO510 to Dr. Melenovsky and MSMT-0021620814 to Drs. Polak and Wedellova) and the Grant Agency of the Academy of Sciences of the Czech Republic (305/09/1390 to Dr. Melenovsky). Dr. Kautzner is on the advisory boards of Biosense Webster, Boston Scientific, GE Healthcare, Medtronic, and St. Jude Medical.
All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial natriuretic peptide
- adipose tissue
- B-type natriuretic peptide
- dialysate glycerol concentration
- free fatty acid
- heart failure
- immunoreactive B-type natriuretic peptide
- natriuretic peptide
- Received January 17, 2011.
- Revision received May 3, 2011.
- Accepted May 31, 2011.
- American College of Cardiology Foundation
- Opie L.H.,
- Knuuti J.
- Sharma S.,
- Adrogue J.V.,
- Golfman L.,
- et al.
- Ashrafian H.,
- Frenneaux M.P.,
- Opie L.H.
- Moro C.,
- Crampes F.,
- Sengenès C.,
- et al.
- Sengenès C.,
- Berlan M.,
- De Glisezinski I.,
- Lafontan M.,
- Galitzky J.
- Polak J.,
- Moro C.,
- Klimcakova E.,
- et al.
- Moro C.,
- Pillard F.,
- De Glisezinski I.,
- et al.
- Moro C.,
- Polak J.,
- Hejnova J.,
- et al.
- Dickstein K.,
- Cohen-Solal A.,
- Filippatos G.,
- et al.
- Xu-Cai Y.O.,
- Wu Q.
- Floras J.S.
- Doehner W.,
- Rauchhaus M.,
- Ponikowski P.,
- et al.
- Galitzky J.,
- Sengenès C.,
- Thalamas C.,
- et al.
- Sengenès C.,
- Zakaroff-Girard A.,
- Moulin A.,
- et al.
- Chen H.H.
- Hawkridge A.M.,
- Heublein D.M.,
- Bergen H.R. III.,
- Cataliotti A.,
- Burnett J.C. Jr..,
- Muddiman D.C.
- Lavie C.J.,
- Milani R.V.,
- Ventura H.O.
- Wang T.J.,
- Larson M.G.,
- Levy D.,
- et al.
- Miyashita K.,
- Itoh H.,
- Tsujimoto H.,
- et al.
- Moro C.,
- Smith S.R.