These authors contributed equally to this study.
Distinct roles of free leptin, bound leptin and soluble leptin receptor during the metabolic-inflammatory response in patients with liver cirrhosis
Version of Record online: 2 APR 2007
Alimentary Pharmacology & Therapeutics
Volume 25, Issue 11, pages 1301–1309, June 2007
How to Cite
OCKENGA, J., TIETGE, U. J. F., BÖKER, K. H. W., MANNS, M. P., BRABANT, G. and BAHR, M. J. (2007), Distinct roles of free leptin, bound leptin and soluble leptin receptor during the metabolic-inflammatory response in patients with liver cirrhosis. Alimentary Pharmacology & Therapeutics, 25: 1301–1309. doi: 10.1111/j.1365-2036.2007.03327.x
- Issue online: 2 APR 2007
- Version of Record online: 2 APR 2007
- Publication data Submitted 29 January 2007 First Decision 20 February 2007 Resubmitted 23 March 2007 Accepted 25 March 2007
Alteration of the leptin system appears to play a role in the inflammatory–metabolic response in catabolic diseases such as chronic liver diseases.
To investigate the association between leptin components, inflammatory markers and hepatic energy and substrate metabolism.
We investigated in vivo hepatic substrate and leptin metabolism in 40 patients employing a combination of arterial and hepatic vein catheterization techniques and hepatic blood flow measurements. In addition to metabolic, inflammatory and neuroendocrine parameters, circulating levels of free leptin, bound leptin and soluble leptin receptor were determined.
Compared with controls, bound leptin and soluble leptin receptor levels were significantly elevated in cirrhosis, while free leptin did not increase. In cirrhosis bound leptin was correlated with soluble leptin receptor (r = 0.70, P < 0.001). Free leptin was positively correlated with metabolic parameters such as energy storage (body fat mass; r = 0.36, P < 0.05), insulin and insulin resistance (r = 0.48; r = 0.46, P < 0.01) as well as with hepatic glucose and energy release (r = 0.35 and r = 0.40, P < 0.05). In contrast, bound leptin and soluble leptin receptor were linked to proinflammatory cytokines and sympathetic activity (r = 0.61 and r = 0.56, P < 0.01).
The components of the leptin system (free leptin, bound leptin and soluble leptin receptor) have distinct roles in metabolic and inflammatory processes in patients with liver cirrhosis. The better understanding of this metabolic and inflammatory tissue-repair response may lead to innovative new therapeutic strategies in liver disease as well as in various other catabolic diseases.
Growing evidence suggests that neurohumoral mechanisms link the inflammatory response and metabolic alterations in patients with chronic catabolic diseases like liver cirrhosis.1 The discovery of the adipose-derived hormone leptin in 1994 represents a landmark in the understanding of the neurohumoral regulation of energy metabolism and homeostasis including both obesity and catabolic diseases.2 Leptin exerts multiple biological effects on a variety of different processes such as energy metabolism, immune responses, angiogenesis and wound healing.3 Experimental work in the genetically leptin-deficient ob/ob mouse model as well as clinical studies suggested that leptin is involved in liver diseases by modulating hepatic fibrogenesis, liver regeneration and steatosis, as well as distinct metabolic pathways.4–8 The data also implied a role for leptin in the pathogenesis of non-alcoholic fatty liver disease (NAFLD), and alcoholic as well as hepatitis C-related liver disease.9–11
Leptin effects are mediated via specific leptin receptors (LR) expressed in a variety of tissues including muscle and liver.12–14 Several isoforms of the receptor have been characterized.12 The shortest isoform lacks the hydrophobic transmembrane domain, and is a soluble, secreted variant (sLR) capable of binding free leptin (FL).15, 16 In serum, leptin circulates both as free hormone and attached to a variety of serum proteins,17 but the majority is bound in high-molecular weight complexes by the soluble leptin receptor (sLR).16
Our previous data indicated that FL and bound leptin (BL) are differentially regulated in a variety of physiological and pathophysiological conditions including chronic liver disease.8, 18–20 Although FL, BL and sLR are detected in supernatants of subcutaneous adipose tissue in vitro,13 only FL consistently and independently predicts body fat mass. In contrast, BL appears to be involved in various regulatory pathways such as sympathetic outflow, energy expenditure and substrate utilization.8, 21
Using in vivo measurements, we characterized the differential regulation of the leptin components in patients with liver cirrhosis and examined the potential metabolic consequences of alterations of the leptin system in this patient population suffering from a catabolic disease.
Patients and methods
Clinical data were obtained from 40 patients (age 48 ± 9 years, BMI 23 ± 3 kg/m2) with biopsy-proven liver cirrhosis while hospitalized to be evaluated for liver transplantation (Table 1). Patients with cirrhosis were graded according to the Child-Pugh score. Eleven patients had alcoholic liver disease, 14 had viral hepatitis (HBV = 5, HCV = 7, HBV + HDV = 1; HBV + HCV + HDV = 1), 10 patients suffered from biliary liver disease (primary biliary cirrhosis = 6, primary sclerosing cholangitis = 2, secondary biliary cirrhosis = 2) and in five patients liver disease was cryptogenic. All patients were in a stable clinical condition before entering the study. The patients were on a standard diet containing 30 kcal/kg/day and 1.2 g/kg/day protein for at least 1 week. Following an overnight fast, measurements were performed between 7:00 am and 9:00 am while the patient was still lying in bed. The last meal was served before 7:00 pm, the preceding evening. Patients with renal failure, proteinuria, significant ascites, acute infections, clinically overt diabetes, thyroid dysfunction or evidence for any other endocrine disorder were excluded from the study. No hormone or thyroid regulatory medication was administered. Patency and orthograde blood flow of portal vein and hepatic artery were documented in patients by Doppler ultrasound before admittance to the study.
|Alcohol||Viral hepatitis||Biliary liver disease||Cryptogenic liver disease|
|Age (years)||48 ± 5||47 ± 11||51 ± 1||47 ± 14|
|BMI (kg/m2)||23 ± 4||24 ± 3||22 ± 3||22 ± 3|
|Body fat mass (kg)||21 ± 7||17 ± 7||17 ± 8||14 ± 4|
|ALT (U/L)||16 ± 12||37 ± 21||43 ± 24||21 ± 13|
|γ-GT (U/L)||65 ± 55||59 ± 44*||157 ± 71**||61 ± 45|
|Bilirubin (μm)||49 ± 34||51 ± 46||92 ± 112***||25 ± 9|
|Child-Pugh score||11 ± 2||10 ± 2||9 ± 2||7 ± 2|
|IL-1ß (pg/mL)||4.9 ± 1.2||6.4 ± 1.1||15.5 ± 10.4||5.9 ± 2.4|
|IL-6 (pg/mL)||12.7 ± 3.4||13.2 ± 2.6||9.7 ± 2.1||4.6 ± 0.3|
|FL (pm)||93 ± 98||85 ± 143||63 ± 36||59 ± 71|
|BL (nm)||0.72 ± 0.46||0.61 ± 0.34||0.94 ± 0.79||0.64 ± 0.42|
|sLR (nm)||16.5 ± 14.9||13.2 ± 4.7||26.9 ± 37.0||14.5 ± 19.2|
In addition, 31 healthy controls (age 51 ± 7 years, BMI 25 ± 4 kg/m2, normal liver function tests, no liver or metabolic diseases) were investigated for blood parameters. Hepatic vein catheterization with assessment of hepatic hemodynamics was not performed in controls.
All subjects were thoroughly informed about the rationale and the possible risks of all investigational procedures and gave written consent before entering the study. The study protocol was approved by the local Ethics Committee.
Blood sampling and laboratory analyses
All blood samples were immediately placed on crushed ice and further processed without delay. ICG was measured spectrophotometrically at 800 nm (DU 6 Beckmann Photometer, Beckmann Instruments GmbH, München, Germany). Glucose, lactate, pyruvate, ß-hydroxybutyrate and free fatty acids (FFA) were measured enzymatically using commercially available assays (Boehringer Mannheim GmbH, Mannheim, Germany).
Plasma aliquots for hormone analysis were stored at −80 °C. Commercially available radioimmunosassays were used to determine plasma concentrations of insulin (Pharmacia Insulin RIA 100, Pharmacia Diagnostics AB, Uppsala, Sweden), and glucagon (Glukagon-RIA DAk/PEG, Hermann Biermann GmbH, Bad Nauheim, Germany). Insulin sensitivity was estimated by calculating the quantitative insulin sensitivity check index (QUICKI), which is derived from fasting blood glucose and plasma insulin levels: QUICKI = 1/[lg (I0) + log (G0)], where I0 is fasting insulin (μU/mL), and G0 is fasting glucose (mg/dL). QUICKI has been validated for use in patients with liver cirrhosis.22 Plasma concentrations of adrenaline were measured by high-performance liquid chromatography as described.23 Plasma levels of IL-1β and IL-6 were determined using commercially available ELISAs (Medgenix Diagnostics, Brussels, Belgium).
Assessment of hepatic hemodynamics
Hepatic vein catheterization was performed to obtain hepatic venous blood samples. Using a balloon catheter, free hepatic venous pressure (FHVP) and wedged hepatic venous pressure (WHVP) were measured as described.24 The hepatic venous pressure gradient (HVPG), reflecting portal pressure, was calculated as HVPG = WHVP − FHVP (in mmHg). Arterial blood was drawn simultaneously from a line placed in the right femoral artery. Quantitative hepatic blood flow (HBF) was determined by indocyanine green (ICG) steady-state infusion technique essentially as described.24 Percentage hepatic extraction was calculated by dividing the arterio-hepatic venous concentration difference by the respective arterial concentration (conca − conclv/conca × 100). Hepatic production/extraction rates were calculated by multiplying the respective arterio-hepatic venous concentration differences with HBF.
Based on the hepatic production/extraction rates the hepatic caloric equivalents were calculated by the following formula: Hepatic caloric equivalents (HCE) = Glucose [mg/min] − (lactate + pyruvate [mg/min/2]) × 3.9 + (lactate + pyruvate [mg/min/2]) × 0.21 + acetoacetate [mg/min] × 4.15 + ß-hydoxybutyrate [mg/min] × 4.69. In this formula, the caloric equivalents of glucose are corrected for recycled lactate and pyruvate and for glycolysis 25.
Immunoassays of leptin components
The assays for serum FL, BL and sLR have been described in detail.13, 21 In brief, polyclonal antibodies to carboxy-terminal (amino acids 126–140) and amino-terminal leptin fragments (amino acids 25–39; 25-Tyr) were generated in rabbits. Both were coupled to hemocyanin by the carbodiimide method. Labelled carboxy- and amino-terminal fragments served as tracer.
Body composition analysis
Patients were weighed and body composition analysis was carried out by bioelectrical impedance analysis (BIA) with a radiofrequency current of 800 μA at 50 kHz between a set of electrodes attached to the dorsum of the hand and foot as described (BIA/S RJL Systems, Detroit, MI, USA). It has been validated and demonstrated that small amounts of ascites in cirrhosis do not significantly impact the BIA results.26, 27
Data are presented as mean ± s.d. unless indicated otherwise. The chi-squared or the Fisher’s exact test was used for discrete variables. Either the Student’s t-test or the Mann–Whitney’s rank sum test was used for unpaired data as indicated. The Kruskal–Wallis test was used for comparison of multiple groups of non-parametric unpaired data. The distribution of FL, sLR and insulin showed significant skewness, therefore these parameters were used for correlation analysis after logarithmic transformation. Stepwise regression analysis was performed to determine the association of different parameters with a dependent variable as mentioned in the Results section. In patients with FL values below the detection level (15 pm), the value 15 pm was used. P-values below 0.05 were considered statistically significant. The data were analysed using spss/PC+ V12.0 software (SPSS, Chicago, IL, USA).
Plasma levels of BL and the sLR are elevated in patients with liver cirrhosis
Plasma levels of FL were not different between patients with liver cirrhosis and control subjects (78 ± 97 vs. 106 ± 120 pm, respectively). In addition, FL levels were independent of the aetiology of liver disease (Table 1) and of the clinical severity of cirrhosis as assessed by the Child classification (Table 2).
|BMI (kg/m2)||22.6 ± 1||22.6 ± 2.9||23.5 ± 3.9|
|Body fat mass (kg)||14 ± 2.1||16 ± 1.8||20 ± 1.7|
|FL (pm)||75 ± 46||88 ± 32||69 ± 17|
|BL (nm)||0.52 ± 0.2||0.67 ± 0.09||0.83 ± 0.15|
|sLR (nm)||9.2 ± 2.6||12.5 ± 3.1||24.8 ± 6.8*|
|FL/kg body fat (pmol/kg)||4.9 ± 2.4||6.9 ± 3||3.2 ± 0.6|
In contrast, BL and sLR were significantly elevated in patients with cirrhosis compared with controls (0.73 ± 0.52 vs. 0.56 ± 0.29 nm, and 17.8 ± 23.0 vs. 5.1 ± 2.0 nm, P < 0.01, respectively). As the clinical severity of cirrhosis increased, the levels of both BL and sLR increased (r = 0.38, P < 0.05; Table 2), regardless of the aetiology of liver disease (Table 1). Neither of these factors were correlated with bilirubin as a measure of cholestasis (r = 0.18, N.S.).
In controls, FL was higher in females compared with males (157 ± 103 vs. 93 ± 122 pm, P < 0.01); however, in patients with liver cirrhosis no significant difference in FL levels could be attributed to sex (84 ± 71 vs. 74 ± 110 pm, N.S.). BL was independent of gender in controls and in cirrhosis.
Interestingly, BL and sLR concentrations were significantly correlated in patients with cirrhosis (r = 0.70, P < 0.001, Figure 1a), but not in the control group (r = 0.18, N.S., Figure 1b). On the other hand, FL was neither associated with BL (r = −0.43, N.S.) nor with sLR (r = −0.13, N.S.).
No significant differences between the arterial and the hepatic venous concentrations of FL (90 ± 108 vs. 80 ± 110 pm, N.S.) were observed, with only a general trend towards increased hepatic venous concentrations of BL (0.65 ± 0.44 vs. 0.73 ± 0.44 nm, N.S.) and sLR (12.1 ± 17.0 vs. 19.3 ± 24.0 nm, N.S.). Consequently, no significant mean hepatic extraction or production rate was detectable.
Circulating FL levels correlate with parameters of hepatic substrate metabolism in cirrhosis
Free leptin correlated with body fat mass (r = 0.36, P < 0.05), and also with circulating insulin levels (r = 0.48, P < 0.01) and displayed an inverse relationship with the degree of insulin resistance as determined by the QUICKI index (r = −0.46, P < 0.01).
Free leptin levels increased with hepatic glucose production (r = 0.35, P = 0.05) but was decreased with respect to hepatic FFA extraction (r = −0.40, P < 0.01). Interestingly, FL was correlated with the HCE (r = 0.40, P < 0.05), thus linking FL to the role of the liver in providing energy substrates in the fasting state. Including FL, insulin and FFA serum concentrations in a multiple regression analysis (dependent variable: hepatic glucose production, independent variables: FL, insulin, FFA serum concentrations) revealed that only FL was independently associated with hepatic glucose production [hepatic glucose production (μmol/min) = 0.438 log FL (pm) – 258; P < 0.02].
With regard to the in vivo assessments of hepatic function, FL was not related to HVPG reflecting portal pressure and did not correlate with HBF or ICG half-life, which is a measure of effective HBF and integrates extrahepatic and intrahepatic shunt flow and the degree of capillarization (data not shown). Circulating levels of the proinflammatory cytokines IL-1β and IL-6 had no impact on plasma FL. Additionally, FL was not associated with sympathetic activity reflected by plasma adrenaline concentrations (Figure 2a).
Circulating levels of BL and sLR correlate with parameters of sympathetic activity and systemic inflammation
The circulating levels of BL and sLR were not correlated significantly with either HVPG, with HBF, or with ICG half-life (data not shown).
However, sLR showed a statistically significant association with IL-1β and IL-6 (r = 0.53, P < 0.01 and r = 0.61, P < 0.01), while the association between IL-1β and IL-6 and BL did not achieve significance. In addition, BL and sLR levels were positively correlated with adrenaline plasma concentrations, reflecting sympathetic activity (r = 0.42 and r = 0.56, each P < 0.01; Figure 2b,c). IL-1ß and IL-6 were positively correlated with adrenaline (r = 0.36 and r = 0.51, each P < 0.05).
To delineate the interactions between the different parameters significantly associated with sLR we performed a stepwise multiple regression analysis. Including adrenaline, IL-6 and IL-1β, this demonstrate that adrenaline and IL-6 were independently associated with sLR (log sLR = 0.15 + 0.017 adrenaline (ng/L) + 0.0023 IL-6 (pg/mL), r = 0.77, P < 0.0001).
In contrast to FL levels, BL and sLR serum concentrations exhibited no significant correlations with body fat mass, with circulating insulin levels or with the degree of insulin resistance as determined by the QUICKI index (data not shown).
In addition, BL and sLR showed no significant associations with hepatic glucose production, hepatic FFA extraction or with the HCE (data not shown).
Herein, we show evidence for a distinct role of the leptin components FL, BL and sLR in liver cirrhosis. In chronic liver disease, increased leptin levels have been reported.9, 28–31 Employing a combination of arterial and hepatic vein catherization techniques and HBF measurements we investigated for the first time in vivo hepatic substrate and leptin metabolism in 40 patients. Therefore, this study confirms not only our previous data that increased leptin levels in cirrhosis are mainly related to increased BL levels;8, 32 we provide new inside in hepatic metabolism and the leptin system in patients with a pronounced catabolic state.
A crucial role of leptin in inflammation and hepatic tissue repair mechanisms has been confirmed by several studies.7, 31, 33–35 However, none of these studies discriminated between FL and BL, nor was sLR determined. Our data indicate that FL, BL and sLR have distinct roles in the interaction between the leptin system and metabolic/inflammatory/regenerative responses. The close relationship between BL and sLR in our study supports the role of sLR as the major binding protein for FL, capable of protecting leptin and modulating leptin actions, as we and others have proposed before.13, 17, 21, 36
The consistent finding of increased BL and sLR with regard to the aetiology of liver disease supports the assumption that differential regulation of the leptin system is a general feature of liver cirrhosis. The LR is expressed on hepatocytes and stellate cells.4 Experimental in vitro work by us37 and also a recent study in wild-type mice38 showed an increase of LR RNA in the liver suggesting that the liver is a source of plasma sLR levels. However, our in vivo results in the present investigation for the difference between hepatic venous and arterial concentrations of BL and sLR did not reach statistical significance. The differences between arterial and hepatic venous concentrations may be too small to detect with our methods during a cross-sectional observational investigation. Future mechanistic studies will address whether the human liver is involved in generation of BL and SLR in vivo.
The observed association between sLR with IL-1 and IL-6 links the leptin system to the proinflammatory neuroimmune response in several clinical situations, including anorexia and cachexia.35 Inflammatory cytokines like IL-6 are crucial for liver regeneration after partial hepatectomy or in toxic liver injury. The close link between leptin and IL-6 is demonstrated by the observation that the effect of IL-6 on liver regeneration is impaired in leptin-deficient ob/ob mice and can be restored by leptin administration.39 In addition, hepatoma cell lines and hepatic stellate cells respond to leptin with an IL-6 receptor-like downstream signalling including activation of STAT proteins, which are involved in hepatic glucose production.40, 41
Some effects of leptin are mediated via the sympathetic nervous system (SNS).42 We found a positive association between BL and sLR levels and sympathetic activity. Increased SNS activity is frequently observed in cirrhosis and has been linked to metabolic complications as well as to hepatic fibrosis.6, 43, 44 Data from an analysis with organ-specific disruption of LRs suggest that liver abnormalities of deficient ob/ob mice are secondary to defective leptin signalling in the brain.45 In this context, it is of special interest that the central nervous effect of BL is more pronounced than the effect of FL.18, 21 Recent work by Oben et al. showed that the fibrogenic effect of leptin depends on neurotransmitter and SNS activity.44 Leptin directly stimulates central sympathetic outflow in rodents and primates. Next to the role in fibrogenesis, SNS activation results in a hypermetabolic state. This may explain the positive association between BL and resting energy expenditure in cirrhotics as well as in healthy subjects.8, 21
During starvation or stress the mobilization of endogenous substrates is crucial to provide the liver with substrates. With respect to metabolism, the liver is the most versatile organ, and the crosstalk between the energy stores (adipocytes) and the liver is of special interest. In cultured hepatocytes, leptin (FL) regulates gene expression of phosphoenolpyruvate carboxykinase, a key gluconeogenic enzyme.46 Thus, leptin may be involved in the intrahepatic distribution of substrate fluxes,47 but to date, human in vivo data with respect to the leptin system are limited. The strength of the present study is our in vivo approach. For simultaneous investigation of the hepatic flux rates of the major energy yielding substrates (i.e. glucose, FFA, ketone bodies, amino acids), we used a combination of arterial and hepatic vein catheterization technique and HBF measurements.30, 48 Although the concentration difference across the hepatic artery and the liver veins represents the exchange rate of the splanchnic bed instead of the true hepatic exchange rates, this methods reflects mainly the hepatic exchange rates.25 The obtained substrate exchange rates for glucose (403 ± 300 μmol/min), alanine (59 ± 40 μmol/min) and FFA (43 ± 76 μmol/min) as well as the positive association between FFA uptake and glucose release are in line with the literature (see review 49). Therefore, our method represents a reliable experimental approach to obtain human in vivo data on hepatic metabolism under physiological conditions.
In contrast to experimental work on cultured hepatocytes or ex vivo perfused mouse liver,50, 51 we observed a negative association between FL and splanchnic FFA uptake. This may be explained by different factors: (i) we performed a cross-sectional observational study in vivo in the fasted state, but the experimental data were generated by a intervention during administration of leptin and exogenous substrates in vitro over a longer period; (ii) high levels of FL may reflect a high abdominal fat mass, which influences splanchnic FFA exchange; (iii) we differentiate between FL, BL and sFL, whereas in the other studies, only total leptin was determined and (iv) we investigated patients with liver cirrhosis as opposed to patients with ‘healthy livers’ in the experimental settings.
Increased FFA delivery to the liver increases hepatic gluconeogenesis and glucose output as confirmed by the positive association between splanchnic FFA uptake and glucose release in this study. In patients with liver cirrhosis, this adaptive mechanism is consistent with maintaining an adequate blood glucose level, because hepatic glucose storage is impaired and glucose mobilization in the fasted state is reduced.49 Although FL levels in cirrhosis were not elevated when compared with controls, FL was associated with insulin and insulin resistance, as well as with hepatic glucose and energy release. This is in accordance with previous studies showing an increase in FL, but not in BL or sLR after administration of insulin.52
In conclusion, our data suggest that the components of the leptin system have distinct roles in metabolic and inflammatory processes in patients with liver cirrhosis. While FL is involved in the crosstalk between the adipose tissue (energy storage) and hepatic energy and substrate flow, the formation of BL and sLR is associated with proinflammatory cytokines and may extend their effects by sympathetic activity. During the metabolic – inflammatory response of the organism to injury, the leptin system provides neuroendocrine signals from body fat for modulation of metabolic and immunological activities to save the physiological and anatomic integrity of the individual. A better understanding of this metabolic and inflammatory tissue-repair response may lead to innovative new therapy strategies in liver disease and probably in various other diseases.
We are indebted to S. Ohlendorf and R. Horn for expert technical assistance and Prof. Hecker from the Department of Bio-Informatics for statistical advice.
Declaration of personal interests: J. Ockenga has served as a speaker, a consultant and an advisory board member for Fresenius Kabi Gmbh, Solvay, Roche, and has received research funding from Fresenius Kabi GmbH.
Declaration of funding interests: The study was supported in part by the Humbold Forschungsförderung (HU/2002-105 to J.O.), European Society for Parenteral and Enteral Nutrition (ESPEN Research Grant to J.O.) and by the Netherlands Organization for Scientific Research (VIDI Grant 917-56-358 to U.J.F.T).
- 11Serum leptin levels correlate with hepatic steatosis in chronic hepatitis C. Am J Gastroenterol 2003; 98: 1135–41., , , et al.Direct Link: