Quantification of the Soluble Leptin Receptor in Human Blood by ...

0013-7227/02/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 87(6):2931–2939 Copyright © 2002 by The Endocrine Society

Quantification of the Soluble Leptin Receptor in Human Blood by Ligand-Mediated Immunofunctional Assay ¨ P, ZIDA WU, MARTIN BIDLINGMAIER, CHANGLU LIU, ERROL B. DE SOUZA, MATTHIAS TSCHO KATHERINE M. MORRISON, AND CHRISTIAN J. STRASBURGER Medical Clinic (Z.W., M.B., M.T., K.M.M., C.J.S.), University Hospital Innenstadt, Ludwig Maximilians University, Munich, Germany; Neurocrine Biosciences, Inc. (C.L., E.B.D.S.), San Diego, California 92121 Panels of monoclonal antibodies (mAbs) were raised against recombinant human leptin and the recombinant human soluble leptin receptor. Using these mAbs, we established a ligand-mediated immunofunctional assay (LIFA) to quantify concentrations of the soluble leptin receptor, which has been shown to be a major binding protein for leptin in human serum. In performing the assay, a monoclonal antibody (mAb 2H6) against the soluble leptin receptor, which binds an epitope outside the leptin-binding site and equally recognizes both, free and leptin-occupied soluble leptin receptor, is used to capture the soluble leptin receptor on a microtiter plate. Recombinant human leptin is added to saturate all binding sites, and a biotinylated anti-leptin mAb (4D3) detects the amount of leptin (endogenous and exogenous) bound to the soluble leptin receptor. The same procedure, but without adding exogenous leptin, allows for measurement of the circulating endogenous leptin/soluble leptin receptor complexes. The LIFA assay has a linear working range of 0.5–200 ␮g/liter, intra- and interassay coefficients of variation ranged from 3.2– 6.3% and from 5.2–7.9%, respectively. The assay has a linearity of 102.2 ⴞ 5.2% (mean ⴞ SD) and a recovery of 100.7 ⴞ 6.9%. Size-exclusion chromatography revealed that the assay

L

EPTIN IS A 16-kDa peptide hormone produced in adipocytes, which is involved in the regulation of energy expenditure, food intake, and adiposity. It also has a role in other physiological processes including pubertal development, reproduction, immunoresponses, hematopoiesis, and angiogenesis. In humans, circulating leptin levels are increased in obesity and are regulated by fasting, feeding, and body weight changes (for review see Refs. 1 and 2). As demonstrated by traditional methods using 125I-labeled recombinant leptin and size-exclusion chromatography (3– 5), circulating leptin in humans is bound to high-molecularweight (HMW) components. Furthermore, a spun-column assay was used to analyze leptin-binding activity in human serum (3, 6). More recently, free leptin, bound leptin, and the soluble leptin receptor were measured in normal and diabetic pregnancies by RIA methods using polyclonal antisera to anti-N- and anti-C-terminal fragments of leptin and of the leptin receptor (7). In this study, females with type I diabetes Abbreviations: BMI, Body mass index; CV, coefficient(s) of variability; FPLC, fast performance liquid chromatography; hGH, human GH; hGH-R, full-length GH receptor extracellular domain; HMW, high molecular weight; IFMA, immunofluorometric assay; Ka, affinity constant; Lep-R, recombinant soluble leptin receptor; LIFA, ligand-mediated immunofunctional assay; mAb, monoclonal antibody; TNF-RI/II, TNF receptor types I and II.

measures a protein with a main peak eluted at 340 kDa. The soluble leptin receptor concentration (63.3 ⴞ 22.8 ␮g/liter (mean ⴞ SD), range 17.9 –129.2 ␮g/liter, n ⴝ 43) in normal subjects (body mass index ⴝ 22.3 ⴞ 2.3 kg/m2) was not different from the concentration (54.4 ⴞ 19.8 ␮g/liter, range 23.7–104.8 ␮g/liter, n ⴝ 34, P > 0.05) found in obese subjects (body mass index ⴝ 40.9 ⴞ 15.7 kg/m2). However, the percentage of the total soluble leptin receptor complexed with endogenous leptin was significantly higher in obese subjects, compared with normal subjects (74.9% ⴞ 23.5% vs. 33.1% ⴞ 19.5%, P < 0.001). Higher serum leptin levels in obese subjects (38.4 ⴞ 23.7 ␮g/ liter vs. 7.8 ⴞ 5.5 ␮g/liter in normal subjects, P < 0.001) together with comparable soluble leptin receptor levels result in a lower proportion of leptin bound to the soluble leptin receptor in obese subjects (19.3% ⴞ 19.4%, range 4.9 –97.2%) than in normal subjects (39.0% ⴞ 22.5%, range 15.3–96.5%, P < 0.001). The development of this LIFA for the rapid and accurate quantification of total soluble leptin receptor and circulating leptin/soluble leptin receptor complexes provides a valuable tool for the further understanding of the role of leptin and its soluble receptor in health and disease. (J Clin Endocrinol Metab 87: 2931–2939, 2002)

were shown to have significantly higher soluble leptin receptor levels. Using a charcoal stripping method and HPLC analysis, the leptin-binding activity was found to change with age. Low at birth and high in the prepubertal years, leptin-binding activity declines through puberty and remains stable during adult life (8). Free and bound forms of leptin have also been quantified by HPLC separation of serum samples followed by RIA leptin determination (9). However, although leptin-binding activity has been studied for several years, the leptin-binding components in human circulation have not been thoroughly characterized. Leptinbinding activity was found in a wide molecular weight range when analyzed by electrophoresis and size exclusion chromatography. In addition, it has been demonstrated that transformed ␣-2-macroglobulin forms a complex with leptin (10). Soluble forms of several hormone or cytokine receptors have been shown to circulate as binding proteins in serum (11). The leptin receptor belongs to the class I cytokine receptor superfamily. The extracellular domain of another member of the same receptor superfamily, the GH receptor, functions as a high affinity-binding protein in human circulation. In mice, it has been shown that a soluble form of the leptin receptor appears as a high affinity leptin-binding protein during pregnancy (12). Evidence has accumulated that,

2931

2932

J Clin Endocrinol Metab, June 2002, 87(6):2931–2939

at least in part, the leptin-binding activity in human serum is related to a soluble form of the leptin receptor as well. Leptin-binding components were shown to have immunoreactivity with antibodies raised against a synthetic peptide of the leptin receptor extracellular domain (5, 7), and Scatchard analysis revealed that leptin-binding components have an affinity comparable with that of the leptin receptor (6, 8). To investigate the presence of the soluble leptin receptor in human serum, we produced and characterized a panel of mAbs against the recombinant soluble leptin receptor (13) and against recombinant leptin. Using these monoclonal antibodies (mAbs), it was possible to detect a soluble form of the leptin receptor in human serum, which is able to bind leptin specifically. Herein we describe the development of a rapid and specific method for quantification of the soluble leptin receptor and of circulating leptin/soluble leptin receptor complexes by a sensitive ligand-mediated immunofunctional assay (LIFA) suitable for large-scale analysis of human serum samples. Materials and Methods Reagents and equipment Tween 20 and 40, BSA, NaN3, polyethylene glycol, Tris-(hydroxymethyl)-aminomethane, diethylenetriaminepenta-acetic acid, bovine ␥-globulin, Hunter’s Titer Max, and biotin-amidocaproate-N-hydroxysuccinimide ester were purchased from Sigma (St. Louis, MO). Na2HPO4, NaOH, and NaCl were obtained from Merck & Co., Inc. (Darmstadt, Germany). Fast performance liquid chromatography (FPLC) equipment, HMW and low-molecular-weight standard kit, Sepharose-r-protein A and Superdex 200 Hiload columns were purchased from Pharmacia (Uppsala, Sweden). FCS, DMEM medium, horse serum, and Protein Free II medium were purchased from Life Technologies, Inc. (Eggenstein, Germany). Mini-PERM devices were purchased from Heraeus (Hanau, Germany) and the IgG-subclass kit from Pierce Chemical Co. (Rockford, IL). Recombinant human leptin, a mutant, which has a single amino acid substitution (Trp to Glu at position 100) but with comparable biological activity and dramatically improved solubility (14), mouse leptin and rat leptin were kindly provided by Eli Lilly & Co. (Indianapolis, IN). Recombinant human GH (hGH, IRP 88/624) was obtained from the National Institute for Biological Standards (Hertfordshire, UK). Recombinant IL-6 and human soluble TNF receptor types I and II (TNF-RI/II) were purchased from PromoCell (Heidelberg, Germany). Full-length GH receptor extracellular domain (hGH-R, amino acids 1–246) was a gift from Prof. A. Gertler (Hebrew University, Rehovot, Israel). The extracellular domain of the leptin receptor (recombinant soluble leptin receptor, Lep-R) was produced in COS7 cells and characterized as described previously (13). Streptavidin-Europium, enhancement solution and the DELFIA 1232 time-resolved fluorometer were obtained from Wallac, Inc. (Turku, Finland). Ninety-six-well flatbottom microtiter plates (Maxisorp plates no. 442404) were obtained from Nunc (Rosklide, Denmark).

Wu et al. • LIFA for Soluble Leptin Receptor

For all assays, the end point detection was the same. After incubation with a biotinylated tracer at room temperature, the microtiter plates were washed three times with 0.3 ml wash buffer, and 10 ng europiumlabeled streptavidin was added into each well and incubated (30 min.). After a 6-fold washing step, the addition of 0.2 ml enhancement solution to each well, and a final incubation (15 min) on a horizontal plate shaker, the signal was read using the DELFIA time-resolved fluorometer.

Production and purification of mAbs against leptin and Lep-R Two-month-old female Balb/c mice were immunized with either recombinant human leptin or recombinant human leptin receptor extracellular domain dissolved in Hunter’s TiterMax adjuvant and injected intradermally (10 ␮g antigen/mouse). After 4 – 6 months of repeated immunization, the mice with the highest serum titers were killed and spleen cells were fused with NSO cells in the presence of polyethylene glycol using the hybridoma technique (15). Cells were grown in medium containing 20% horse serum. Hybridoma cell supernatants were screened for antileptin- or antisoluble leptin receptor activity after 10 –12 d of culture using biotinylated soluble leptin receptor or biotinylated leptin, respectively. Hybridoma cells corresponding to the supernatants giving the highest signals were cloned at least three times by limiting dilution. The IgG subclass of the mAbs was determined and large-scale production was carried out in mini-PERM devices in protein-free medium. The IgG concentration of the supernatant was 1–5 g/liter. The mAbs were affinity purified using an r-Protein A FPLC column, pooled IgG-containing fractions were extensively dialyzed against PBS, divided into aliquots, and stored at ⫺20 C until use.

Biotinylation of mAbs, leptin, and leptin receptor The purified mAbs were biotinylated as described previously (16) using a 75-fold molecular excess of the labeling reagent (biotin) in the reaction. Soluble leptin receptor and leptin were biotinylated using the same method but with 25-fold and 10-fold molar excess of biotinamidocaproate-N-hydroxysuccinimide ester, respectively.

Characterization and selection of the mAbs A LIFA as described here requires two mAbs, both not affected by the binding of their respective antigen to its ligand (Fig. 1). Accordingly, the capture mAb against the soluble leptin receptor was selected on the basis of a high-affinity binding with the Lep-R that was not displaceable by recombinant leptin. The other way around, an antileptin antibody was selected, which binds leptin outside the leptin/leptin receptor interaction site and therefore is not displaced by the soluble leptin receptor. The mAb 2H6 has been raised against the Lep-R. To ensure that the mAb also recognizes specifically the soluble leptin receptor molecule as it naturally occurs in human serum, the following assay was developed: Microtiter plates coated with mAb 2H6 (200 ng/well) were incubated with 0.05 ml human serum and 0.150 ml assay buffer overnight at 4 C. After washing, 1 ng biotinylated human leptin was added to each well together with serially diluted nonlabeled leptin (human, mouse, rat), nonlabeled hGH or nonlabeled IL-6 and incubated overnight at 4 C. After a subsequent washing step, plates were processed as described above.

Time-resolved fluorescence immunoassays

Assay standards and serum control samples

Microtiter plates were coated with antibodies diluted in phosphate buffer (50 mmol/liter, pH 9.6) by over night incubation at 4 C. Phosphate buffer was used because previous studies have shown that the presence of phosphate rather than carbonate ions leads to a higher absorption capacity of the microtiter plates for monoclonal antibodies. Washing buffer was prepared freshly for all experiments [50 mmol/liter Na2HPO4, 0.05% Tween 20, and 0.0025% NaN3 (pH 7.5)]. Detection antibodies were diluted in assay buffer [50 mmol/liter Tris-(hydroxymethyl)-aminomethane, 154 mmol/liter NaCl, 20 ␮mol/liter diethylenetriaminepenta-acetic acid, 0.01% Tween 40, and 0.05% NaN3 (pH 7.75)]. BSA (0.5%) and bovine ␥-globulin (0.05%) were added to combat nonspecific binding.

Lep-R diluted in assay buffer is used as the calibrator for the LIFA. To minimize matrix differences between serum samples and calibrators, additional BSA was added to the calibrators (final protein concentration 7%). Recombinant leptin diluted in FCS served as the calibrator for the immunofluorometric assay (IFMA) for leptin. Serum samples were obtained from normal healthy [n ⫽ 43, body mass index (BMI) 22.3 ⫾ 2.3 kg/m2] and obese (n ⫽ 34, BMI 40.9 ⫾ 15.7 kg/m2) adults not receiving medication. Obesity was defined by a BMI greater than 27.3 for males and greater than 27.8 for females, according to the NIH Consensus Development Panel. Serum samples were centrifuged and stored at ⫺20 C until assayed. All serum samples were obtained from studies approved by the Ethics Committee of the medical


J Clin Endocrinol Metab, June 2002, 87(6):2931–2939 2933

biotinylated antileptin mAb 6D9 in 0.175 ml assay buffer. After an overnight incubation at 4 C, the plates were processed as described above for the LIFA. A possible interference of soluble leptin receptor on leptin measurement in this assay was investigated by spiking leptin standards (5, 10, 20, 50, and 100 ␮g/liter) with Lep-R (10, 20, 50, 100, and 200 ␮g/liter). After an overnight preincubation, samples were assayed by IFMA. Less than 5% change in the measured leptin levels even at the highest concentration of soluble leptin receptor indicate that this IFMA for leptin measures total leptin independently from the formation of leptin/leptin receptor complexes.

Size-exclusion chromatography An FPLC system, composed of the control unit (LCC-500) and two P-500 pumps connected to a Superdex 200 Hiload column (Pharmacia), was precalibrated with the gel filtration low-molecular-weight and HMW calibration kits. Eluted proteins were detected by UV absorption at 280 nm, and their molecular masses were calculated based on elution time and UV absorption using the Cricket graph software [version 1.2.1. (D1), Computer Associates, Inc., Islandia, NY]. Serum from a healthy adult (0.5 ml) was filtered through a 0.22-␮m filter and loaded over the preequilibrated column. The column was eluted with PBS [10 mm Na2HPO4, 150 mm NaCl (pH 7.4)] under neutral conditions (pH 7.4, room temperature) at a flow rate of 1 ml/min, and 1-ml fractions were collected in tubes containing 0.1 ml 1% BSA in PBS. Fractions were assayed by LIFA to measure soluble leptin receptor concentrations and using the IFMA for total leptin concentrations.

Mathematics and statistics FIG. 1. Measurement of circulating endogenous leptin/soluble leptin receptor complexes and total soluble leptin receptor by LIFA. Soluble leptin receptor is captured by mAb 2H6, which binds an epitope outside the leptin/leptin receptor interaction site and equally recognizes both free leptin receptor and leptin/leptin receptor complexes. Addition of antileptin mAb 4D3 alone detects preexisting endogenous leptin/soluble leptin receptor complexes only (A), whereas addition of mAb 4D3 together with an excess of recombinant leptin allows for the measurement of total soluble leptin receptor concentration (B). faculty of the Ludwig-Maximilians University after informed consent was given.

LIFA assay procedure Microtiter plates were coated with the antileptin receptor mAb 2H6 (500 ng/well) as described above. Sealed plates were stored at 4 C for up to 1 month. After removal of the coating solution, standards or samples (0.05 ml) were pipetted into each well. Then 50 ng biotinylated antileptin mAb 4D3, either alone (for measurement of endogenous leptin/soluble leptin receptor complexes, Fig. 1A) or together with 5 ng recombinant leptin (for measurement of the total soluble leptin receptor concentration, Fig. 1B) in 0.15 ml assay buffer, were added. After an overnight incubation (4 C) and a 3-fold washing step, plates were processed as described above. The calibration curve was produced by plotting the soluble leptin receptor concentration of standards and the signal from the time-resolved fluorometric end point in a double logarithmic system. A spline-fit (Multicalc software package, version 2.6., Wallac, Inc.) was used, and the concentrations of unknown samples were read by interpolation of the signal obtained on the standard curve. Specificity of the LIFA was investigated using recombinant soluble hGH-R or recombinant human soluble TNF-RI/II (up to 200 ␮g/liter) instead of the soluble leptin receptor. Furthermore, recombinant leptin was substituted by recombinant hGH or IL-6 at concentrations up to 1000 ␮g/liter.

Leptin IFMA assay procedure The previously described IFMA for leptin in serum (17) was modified by extending incubation time and reducing sample volume. Briefly, microtiter plates coated with mAb 4D3 (500 ng/well) were washed, and standards or samples were pipetted into each well together with 50 ng

Nonparametric statistics (Mann-Whitney U test, Spearman correlation) were carried out for comparison between groups using the StatView software program (version 5.0, SAS Institute, Inc., Cary, NC). Unless otherwise noted, all values are given as mean ⫾ 1 sd. An equimolar binding stoichiometry has been shown between leptin and its receptor (13, 18). Therefore, to get an estimate of the affinity constant (Ka) for the interaction between leptin and the soluble leptin receptor in human serum, we used an adaptation of the mass equation:

Ka ⫽

关ligand/receptor complex兴关free ligand兴 ⫻ 关free receptor兴 ⫽

关circulating leptin/soluble leptin receptor complex兴关free leptin兴 ⫻ 关free soluble leptin receptor兴

On a molar basis, free leptin concentration was calculated as total leptin concentration minus circulating leptin/leptin receptor complex concentration. Free soluble leptin receptor concentration corresponds to total soluble leptin receptor concentration minus circulating leptin/leptin receptor complex concentration. In accordance with our previous findings on the Lep-R monomer (13) and with the molecular weight predicted from the amino acid sequence and a 36% N-glycosylation (19), the molecular weight of 130 kDa for the serum soluble leptin receptor was used for the calculation.

Results Production and selection of mAbs against the soluble leptin receptor

Twenty clones secreting mAbs with high affinity binding to the Lep-R were produced through a single fusion with the conventional hybridoma technique. Twelve mAbs were purified and biotinylated. The relationship between the epitopes of these mAbs and the leptin binding site of the soluble leptin receptor was identified by displacement experiments using recombinant leptin. For example, binding of mAb 9F8 to the biotinylated Lep-R is inhibited by leptin, but binding of mAb 2H6 is not (Fig. 2A). This indicates that mAb 9F8 binds the leptin receptor extracellular domain within the

2934



FIG. 2. Epitope and specificity of mAb 2H6. The mAb 2H6 or mAb 9F8, respectively, were immobilized on microtiter plates. Binding of the biotinylated soluble leptin receptor to mAb 9F can be displaced by leptin, whereas binding to mAb 2H6 is independent from leptin. Antisoluble leptin receptor mAb 2H6 binds to a protein in human serum, which is able to bind biotinylated leptin. The binding of this soluble receptor to biotinylated human leptin is displaceable by nonlabeled human, mouse, and rat leptin but not by nonlabeled hGH or IL-6.

leptin-binding site, and mAb 2H6 targets an epitope outside the leptin binding. The specificity of binding of mAb 2H6 to the soluble leptin receptor in human serum was verified by demonstrating its binding to a protein in human serum samples, which was able to bind biotinylated leptin specifically. The binding of biotinylated human leptin is displaceable by nonlabeled human, mouse, and rat leptin but not by hGH or IL-6 (Fig. 2B). Therefore, mAb 2H6 was selected as the capture antibody for the LIFA. Choice of leptin concentration to saturate all soluble leptin receptors

For the measurement of the total soluble leptin receptor concentration using the LIFA, the molar ratios within the wells of the microtiter plate have to be taken into account. On the one hand, all leptin-binding sites of the receptors have to be saturated by recombinant leptin. Therefore, a sufficient molar excess of leptin over the highest Lep-R standard (200 ␮g/liter corresponding to a final concentration of 0.39 nmol/ liter in the well) must be ensured. On the other hand, to allow for a simultaneous incubation of recombinant leptin and the antileptin detection mAb, the leptin concentration must not exceed the concentration of the leptin-binding sites of biotinylated mAb 4D3 (two binding sites per molecule of the antibody, final concentration of binding sites in the well 3.33 nmol/liter). Thus, we tested the effect of the addition of different amounts of recombinant leptin leading to final concentrations in the well of 0.78, 1.57, and 3.13 nmol/liter. Recombinant leptin in the well at 1.57 nmol/liter gave a higher signal than at 0.78 nmol/liter, but a further increase in concentration had no effect. Therefore, a concentration of 1.57 nmol/liter was chosen, corresponding to a final concentration of 5 ng recombinant leptin per well.

To exclude the interference of endogenous leptin present in serum samples with the measurement of soluble leptin receptor, two sets of standards were measured by LIFA. One set was preincubated with recombinant leptin (100 ␮g/liter) overnight at 4 C and the other preincubated with assay buffer alone. A comparable signal was generated for both sets of standards (data not shown). This indicated that the measurement of the soluble leptin receptor by this assay is independent from the concentration of endogenous leptin. LIFA range and sensitivity

The curves obtained from Lep-R and diluted serum samples are parallel, indicating that the assay equally recognizes both the recombinant and the serum-derived soluble leptin receptor (Fig. 3A). The linear working range is 0.5–200 ␮g/ liter (0.004 –1.54 nmol/liter). The lower limit of detection, defined as the concentration corresponding to the mean signal of a 20-fold determination of the blank (zero standard) plus 2 sd, is 0.2 ␮g/liter (1.54 pmol/liter). Assay specificity and assay precision

The specificity of the LIFA was tested by substituting the soluble leptin receptor by another soluble receptor of the same class I cytokine receptor superfamily, recombinant soluble hGH-R. These experiments revealed that the LIFA has less than 0.01% cross-reaction with the recombinant soluble GH receptor. Furthermore, it was proven that recombinant human soluble TNF-RI/II did not cross-react in this assay. Substitution of recombinant leptin by recombinant hGH or IL-6 resulted in values indistinguishable from the blank when tested together with either Lep-R, human soluble TNFRI/II or hGH-R (Fig. 4). Intraassay coefficient of variability (CV) at concentrations


FIG. 3. Standard curve of LIFA and parallelism between recombinant and serum-derived soluble leptin receptor. Lep-R standard and two human serum samples were diluted serially in zero standard and analyzed by LIFA. Signal intensity (counts) detected by the fluorometer are plotted against the concentration of the soluble leptin receptor.

of 21.7, 53.7, and 105.7 ␮g/liter were 6.3%, 5.0%, and 3.2%, respectively (n ⫽ 18). Interassay precision was determined at concentrations of 38.5, 57.4, and 74.9 ␮g/liter. The CV was 7.9%, 5.3%, and 5.9%, respectively (n ⫽ 7). Linearity and accuracy of the assay

Two human serum samples were measured after serial dilution in zero standard. Mean linearity was shown to be 102.2% ⫾ 5.9% in sample 1 (soluble leptin receptor concentration 41.0 ␮g/liter) and 102.2% ⫾ 5.3% in sample 2 (soluble leptin receptor concentration 56.0 ␮g/liter). Linear regression analysis of observed vs. expected values revealed coefficients of correlation of 0.99 or greater, indicating a good linearity. To determine the accuracy of the soluble leptin receptor determinations by the LIFA, five different concentrations of Lep-R diluted in assay buffer were spiked to four serum samples, and all samples were assayed by LIFA. The theoretical soluble leptin receptor concentration for each sample was used to calculate the recovery. The mean recovery was 100.4% (range 90.4 –107.5%). Stability of soluble leptin receptor

The effects of freezing and thawing of serum samples on the quantification of the soluble leptin receptor concentrations were examined in three serum samples, which were measured eight times during an 8-wk period. After each assay, the samples were frozen again and stored at ⫺20 C until the next assay. The CV for these samples with soluble leptin receptor concentrations of 31.5 ␮g/liter, 53.2 g/liter, and 71.8 ␮g/liter were 6.7%, 6.7%, and 5.2%, respectively,


FIG. 4. Lack of cross-reaction of LIFA with other soluble receptors. Signals are indistinguishable from the blank when recombinant soluble hGH-R or TNF-RI together with leptin are used. The same was true for Lep-R together with hGH or IL-6. Signal intensity (counts) detected by the fluorometer are plotted against the concentration of the respective soluble receptor.

indicating that the measurement of the soluble leptin receptor concentration is unaffected by freezing and thawing. Concentrations of total soluble leptin receptor, leptin, and leptin/soluble leptin receptor complexes in normal and obese subjects

The soluble leptin receptor levels (63.3 ⫾ 22.8 ␮g/liter, range 17.9 –129.2 ␮g/liter, n ⫽ 43) in normal subjects (BMI 22.3 ⫾ 2.3 kg/m2) were not different from those (54.4 ⫾ 19.8 ␮g/liter, range 23.7–104.8 ␮g/liter, n ⫽ 34, P ⬎ 0.05) observed in obese subjects (BMI 40.9 ⫾ 15.7 kg/m2). However, the percentage of soluble leptin receptors circulating in a complex with leptin is significantly higher in obese, compared with normal, subjects (74.9% ⫾ 23.5% vs. 33.1% ⫾ 19.5%, P ⬍ 0.001). This indicates a higher degree of saturation of the soluble leptin receptor by leptin in obese subjects (Fig. 5A). No significant difference in total soluble leptin receptor concentrations between males and females (57.6 ⫾ 19.0 vs. 60.7 ⫾ 23.2, P ⬎ 0.05) was observed. Furthermore, no significant correlation with age could be detected in the adult population studied (R2 ⫽ 0.004, P ⬎ 0.05). However, because of the much lower leptin concentration in male, compared with female, subjects (9.9 ⫾ 17.8 ␮g/liter vs. 27.3 ⫾ 22.2 ␮g/liter, P ⬍ 0.0001), the concentration of circulating soluble leptin receptor/leptin complexes is significantly lower in males (37.3 ⫾ 19.6 ␮g/liter vs. 16.5 ⫾ 15.1 ␮g/liter, P ⬍ 0.0001). A higher proportion of circulating leptin (39.0% ⫾ 22.5%, range 15.3–96.5%) was bound to the soluble leptin receptor in normal subjects, compared with only 19.3% ⫾ 19.4% (range 4.9 –97.2%) in obese subjects (P ⬍ 0.001). To a major extent, this can be explained by the significantly higher circulating leptin levels in obese (38.4 ⫾ 23.7 ␮g/liter), compared with normal, subjects (7.8 ⫾ 5.5 ␮g/liter, P ⬍ 0.001, Fig. 5B). The amount of leptin bound to the soluble leptin receptor

2936



FIG. 5. Total leptin, total soluble leptin receptor, and circulating preexisting leptin/soluble leptin receptor complexes in normal and obese subjects (*, P ⬍ 0.001 vs. normal). Comparable concentrations of soluble leptin receptor in normal and obese subjects. A higher proportion of the soluble leptin receptor is bound with leptin in obese subjects. Total leptin concentrations and leptin/ soluble leptin receptor complex concentrations are higher in obese subjects. However, the percentage of leptin bound to the soluble receptor is higher in normal subjects.

FIG. 6. Saturation of the soluble leptin receptor in circulation by endogenous leptin. Total soluble leptin receptor and circulating endogenous leptin/leptin receptor complexes in serum samples obtained from normal and obese subjects were assayed by LIFA. Total leptin was assayed by IFMA. Plotting of leptin concentrations against the ratios of leptin/ leptin receptor complexes to total soluble leptin receptor reveals a typical saturation curve.

was calculated from the quantity of preexisting leptin/soluble leptin receptor complexes. Because of the considerable difference in the molecular mass between leptin and the soluble leptin receptor, the calculated concentration of leptin bound to the soluble leptin receptor shown in Fig. 5 seems to be much lower than the concentration of the soluble leptin receptor bound to leptin (both expressed in micrograms per liter), but they are equal if calculated as molar concentrations. The ratio of leptin/soluble leptin receptor complexes to total soluble leptin receptors plotted against the leptin levels represent a typical saturation curve (Fig. 6). Size-exclusion chromatography profile

Analysis of the fractions obtained after size-exclusion chromatography of a serum sample from a healthy adult by LIFA revealed that the main peak of the soluble leptin receptor elutes at an apparent molecular weight of approximately 340 kDa (Fig. 7). Some small additional elution peaks were observed at lower and higher molecular weight, possibly representing glycosylation variants or multimeric molecules. Under the conditions used in this experiment, the leptin/soluble leptin receptor complex is dissociated (5). Ac-

FIG. 7. Superdex 200 elution profile of the soluble leptin receptor and leptin in human serum. A, Superdex 200 Hiload column (Pharmacia) calibrated with molecular weight standards was loaded with 0.5 ml human serum and run at room temperature. Elution fractions of 1 ml were collected and assayed by LIFA for soluble leptin receptor and by leptin IFMA. The observed molecular mass for the main elution peak of the soluble leptin receptor was about 340 kDa.


cordingly, a single peak was detected for leptin at the expected molecular weight (16 kDa). Calculation of the Ka for the interaction between leptin and the soluble leptin receptor

We analyzed the concentrations of total leptin, total soluble leptin receptor, and circulating leptin/soluble leptin receptor complexes in a series of serum samples obtained from 39 normal and 31 obese individuals. Using these concentrations, we calculated the Ka for the interaction between leptin and its soluble receptor using an adaptation of the mass equation (see “Mathematics and statistics” in Materials and Methods). The mean Ka calculated for all samples is 2.4 ⫾ 1.4 ⫻ 109 m⫺1 but significantly higher in obese (3.0 ⫾ 1.7 ⫻ 109 m⫺1, n ⫽ 31) when compared with normal subjects (1.9 ⫾ 1.0 ⫻ 109 m⫺1, n ⫽ 39, P ⬍ 0.002). Discussion

Since the first reports on leptin-binding activity in human serum, the nature of possible leptin binding proteins has been widely discussed. In mice, the leptin-binding protein has been identified as the soluble leptin receptor (12, 20). It is absent in dbPas/dbPas mice, which carry a duplication of exons 4 and 5 of leptin receptor DNA leading to a premature stop codon at amino acid 281 (21). Several groups provided data supporting the hypothesis that a soluble form of the leptin receptor also could be present in human serum and serve as a high affinity-binding protein (3–5, 7, 8, 22). However, the molecular masses of the leptin-binding components in human serum reported by different authors vary from about 85 kDa (8), 100 and 200 kDa (7), 176 and 240 kDa (4), 280 kDa (5) up to 450 kDa (3) or between 200 and 670 kDa (22). We found that the protein measured by the LIFA presented here is eluted at a molecular mass of 340 kDa. This is in accordance with the data published previously using 125I labeled leptin together with the Lep-R (13). However, if the soluble leptin receptor represents the major leptin-binding protein in human circulation, how can the heterogeneity in the molecular weight of the leptin-binding activity be explained? Complex formation with ␣-2-macroglobulin, which has been shown to bind to leptin under certain experimental conditions, might explain the presence of leptin in very HMW complexes in vitro by some authors. However, it is difficult to estimate the impact of ␣-2-macroglobulin in vivo, in which most of the ␣-2-macroglobulin is not transformed and thus is not able to bind leptin (10). Another source of the heterogeneity may be inherent to the technical nature of liquid chromatography: resolution of the columns, number, and range of the protein used for standardization, the number and size of the elution fractions, and the detection method used may influence the results. Furthermore, as mentioned by Wallace (23), leptin is hydrophobic; thus, aggregation of the radiolabeled leptin used in binding studies might have affected the results. It also can be speculated that the soluble leptin receptor in circulation is dimerized because it has been demonstrated for the membrane bound leptin receptor (18). This could explain the lower molecular masses seen by SDS-PAGE when the molecule is monomeric under


denaturalizing conditions, compared with higher molecular masses estimated by size-exclusion chromatography. In addition, as suggested recently (13, 18, 19, 22), the soluble leptin receptor in human serum could consist of different isoforms, possibly because of alternative splicing or glycosylation. All these aspects could contribute to the observed heterogeneous molecular weights of leptin-binding components in human serum. Here, we provide further evidence that a major leptinbinding component in human serum is a soluble form of the leptin receptor. First, the use of specific mAbs against an epitope outside the leptin-binding domain of the soluble leptin receptor immobilized a protein from human serum, which is able to bind leptin specifically. Second, because the method described here allows for the direct measurement of the circulating leptin/soluble leptin receptor complexes in serum samples, it was possible to estimate the Ka of the soluble leptin receptor according to the mass reaction equation: Using the measured concentrations for total leptin, total soluble leptin receptor, and circulating leptin/soluble leptin receptor complexes, we obtained Ka values comparable with those previously reported by others using different techniques (6, 13, 24). Of course, the determination of the Ka from measured serum concentrations of the ligands and their complexes is not the classical approach. Therefore, the calculated Ka should be interpreted with caution, especially because of the possible influence of binding proteins other than the soluble receptor. However, the similarity to the Ka shown by classical in vitro methods suggests that this influence is low and the soluble leptin receptor is a major binding protein in human serum. In normal subjects, the Ka as calculated by our approach is 1.9 ⫾ 1.0 ⫻ 109 m⫺1 (n ⫽ 39), compared with 3.0 ⫾ 1.7 ⫻ 109 m⫺1 in obese subjects (n ⫽ 31). Although this comparably small difference between normal and obese subjects is statistically significant (P ⬍ 0.002), its possible relevance remains to be further studied. Most current methodologies for analysis of leptin-binding activity in human serum rely on the use of 125I-labeled recombinant leptin and size-exclusion chromatography. These methods are sensitive and, theoretically, can detect all complexes formed by leptin with either low or high affinitybinding proteins. Recently, a method based on HPLC separation and RIA quantification of leptin facilitating the measurement of free and bound leptin has been reported (9). However, chromatography methods are labor intensive and time consuming and therefore of limited suitability for studies with large numbers of samples. Leptin-binding activity has also been determined by a charcoal-stripping method and the spun-column assay, the latter being suitable for multiple determinations, but these two methods still require pretreatment of the samples. In terms of feasibility, the determination of free and bound leptin by RIA using polyclonal antibodies against N- and C-terminal fragments of leptin provides an advantage (7). However, it remains hard to explain why the anti-N-terminal antibody used for this method recognizes leptin only when bound to the soluble receptor. This would imply that the polyclonal antiserum recognizes an epitope present only after a conformational change induced by the formation of leptin/leptin receptor complexes.

2938


The RIA using antisera against fragments of the soluble leptin receptor developed by the same authors directly quantifies the soluble leptin receptor. However, antipeptide antibodies usually have a higher affinity to the peptides than to the intact proteins. Furthermore, polyclonal antisera most likely bind both intact and fragmented antigens. Thus, the combination of a specific, characterized monoclonal antibody against the intact Lep-R together with the natural ligand leptin in the LIFA assay as presented here bears the advantage of measuring a well-defined molecule only if it is functionally active and able to bind leptin. In addition, without adding exogenous leptin, the assay allows the direct determination of preexisting, circulating leptin/leptin receptor complexes. The same approach has already been used for the measurement of the soluble hGH receptor ectodomain or hGHBP and has proven to be highly specific (25). Performance of the LIFA is similar to standard immunoassays, rapid, simple, and easily carried out in most laboratories even for larger series of samples. Despite the differences in the methodology used to measure bound leptin in human serum, the results from other groups are comparable with ours. Sinha et al. (5) have reported that a higher proportion of total leptin circulates in the bound form in lean (46.5% ⫾ 5.5%) than in obese (21.4% ⫾ 3.4%) subjects. More recently, these results were confirmed by Landt et al. (26) using a different method. In accordance with that, we found 39.0% ⫾ 22.5% (range 15.3–96.5%) of circulating leptin bound to the soluble leptin receptor in normal subjects, compared with only 19.3% ⫾ 19.4% (range 4.9 –97.2%, P ⬍ 0.001) in obese subjects. The wide range of values can be explained by the great between-subject variability in leptin levels, compared with the more uniform levels of the soluble leptin receptor. Soluble leptin receptor concentrations did not vary with BMI, age, or gender in the adult population we studied. However, the absolute concentrations of both, total leptin and leptin bound to the soluble receptor, were higher in obese than in normal subjects (Fig. 5B). In addition, they were higher in females than in males. This is in accordance to the recent findings by McConway et al. (27). Using a gel filtration method, they showed higher absolute concentrations of both free and bound leptin in obese, compared with normal, subjects and in women, compared with men. One can speculate that in normal subjects, the soluble leptin receptor may regulate leptin concentration dynamically as a buffer, but the higher degree of saturation of the soluble leptin receptor by leptin in obese individuals may negatively influence this system (Fig. 6). The clinical significance of these findings remains to be determined. However, our data suggest that the leptin resistance in obesity is not a consequence of elevated soluble leptin receptor levels. The quantification of biologically intact soluble leptin receptor molecules by the LIFA also provides an opportunity to study the leptin receptor in vivo. A mutation of the human leptin receptor gene that results in a truncated leptin receptor lacking both the transmembrane and the intracellular domains has recently been described (28). Patients homozygous for the mutation have early-onset morbid obesity, absence of pubertal development, and abnormal secretion of GH and TSH. Elevated ratios of bound/free leptin have been dem-


onstrated by size-exclusion chromatography using radioactive leptin. Although this mutation is an uncommon cause of obesity, the measurement of the soluble leptin receptor by the LIFA could provide a screening method to identify patients with defects in the leptin receptor. These known defects in the circulating leptin-binding protein, the information on this system in diabetic and normal pregnancies (7), the recent reports on the impact of the soluble leptin receptor levels on RIA leptin determination (29) and the bioavailability of leptin in rats (30) illustrate the importance and potential clinical usefulness of the measurement of not only leptin but also its soluble receptor. The specific and sensitive assays described herein therefore provide important tools in the study of leptin and its receptor in health and disease. Acknowledgments We thank Dr. Mark Heiman (Eli Lilly & Co., Indianapolis, IN) and Prof. Arieh Gertler (Hebrew University, Rehovot, Israel) for the generous gift of recombinant leptin and hGHBP, respectively. Received June 28, 2001. Accepted March 10, 2002. Address all correspondence and requests for reprints to: Christian J. Strasburger, M.D., Medical Clinic, University Hospital Innenstadt, Ludwig Maximilians University, Ziemssenstrasse 1, 80336 Munich, Germany. E-mail: [email protected]. This work was presented in part at the 81st Annual Meeting of The Endocrine Society, June 12–15, 1999, San Diego, California (Abstracts P1-426 and P1-429).

References 1. Ahima RS, Flier JS 2000 Leptin. Annu Rev Physiol 62:413– 437 2. Wauters M, Considine RV, Van Gaal LF 2000 Human leptin: from an adipocyte hormone to an endocrine mediator. Eur J Endocrinol 143:293–311 3. Diamond FB, Eichler DC, Duckett G, Jorgensen EV, Shulman D, Root AW 1997 Demonstration of a leptin binding factor in human serum. Biochem Biophys Res Commun 233:818 – 822 4. Houseknecht KL, Mantzoros CS, Kuliawat R, Hadro E, Flier JS, Kahn BB 1996 Evidence for leptin binding to proteins in serum of rodents and humans: modulation with obesity. Diabetes 45:1638 –1643 5. Sinha MK, Opentanova I, Ohannesian JP, Kolaczynski JW, Heiman ML, Hale J, Becker GW, Bowsher RR, Stephens TW, Caro JF 1996 Evidence of free and bound leptin in human circulation. Studies in lean and obese subjects and during short-term fasting. J Clin Invest 98:1277–1282 6. Eichler DC, Root AW, Duckett G, Moore KL, Diamond FB 1999 A spuncolumn assay for determination of leptin binding in serum. Anal Biochem 267:100 –103 7. Lewandowski K, Horn R, O’Callaghan CJ, Dunlop D, Medley GF, O’Hare P, Brabant G 1999 Free leptin, bound leptin, and soluble leptin receptor in normal and diabetic pregnancies. J Clin Endocrinol Metab 84:300 –306 8. Quinton ND, Smith RF, Clayton PE, Gill MS, Shalet S, Justice SK, Walters S, Postel-Vinay MC, Blakemore AI, Ross RJ 1999 Leptin binding activity changes with age: the link between leptin and puberty. J Clin Endocrinol Metab 84:2336 –2341 9. Landt M 2000 Leptin binding and binding capacity in serum. Clin Chem 46:379 –384 10. Birkenmeier G, Kampfer I, Kratzsch J, Schellenberger W 1998 Human leptin forms complexes with alpha 2-macroglobulin which are recognized by the alpha 2-macroglobulin receptor/low density lipoprotein receptor-related protein. Eur J Endocrinol 139:224 –230 11. Muller-Newen G, Kohne C, Heinrich PC 1996 Soluble receptors for cytokines and growth factors. Int Arch Allergy Immunol 111:99 –106 12. Gavrilova O, Barr V, Marcus-Samuels B, Reitman M 1997 Hyperleptinemia of pregnancy associated with the appearance of a circulating form of the leptin receptor. J Biol Chem 272:30546 –30551 13. Liu C, Liu XJ, Barry G, Ling N, Maki RA, De Souza EB 1997 Expression and characterization of a putative high affinity human soluble leptin receptor. Endocrinology 138:3548 –3554 14. Zhang F, Basinski MB, Beals JM, Briggs SL, Churgay LM, Clawson DK, DiMarchi RD, Furman TC, Hale JE, Hsiung HM, Schoner BE, Smith DP, Zhang XY, Wery JP, Schevitz RW 1997 Crystal structure of the obese protein leptin-E100. Nature 387:206 –209


15. Kohler G, Milstein C 1975 Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256:495– 497 16. Strasburger CJ, Kostyo J, Vogel T, Barnard GJ, Kohen F 1989 The antigenic epitopes of human grown hormone as mapped by monoclonal antibodies. Endocrinology 124:1548 –1557 17. Tschoep M, Wu Z, Bidlingmaier M, Heiman M, Strasburger CJ A sandwich immunoassay for human leptin. Program of the 80th Annual Meeting of The Endocrine Society, New Orleans, LA, 1998 (Abstract P2-524) 18. Devos R, Guisez Y, Van der Heyden J, White DW, Kalai M, Fountoulakis M, Plaetinck G 1997 Ligand-independent dimerization of the extracellular domain of the leptin receptor and determination of the stoichiometry of leptin binding. J Biol Chem 272:18304 –18310 19. Haniu M, Arakawa T, Bures EJ, Young Y, Hui JO, Rohde MF, Welcher AA, Horan T 1998 Human leptin receptor. Determination of disulfide structure and N-glycosylation sites of the extracellular domain. J Biol Chem 273:28691–28699 20. Lee GH, Proenca R, Montez JM, Carroll KM, Darvishzadeh JG, Lee JI, Friedman JM 1996 Abnormal splicing of the leptin receptor in diabetic mice. Nature 379:632– 635 21. Li C, Ioffe E, Fidahusein N, Connolly E, Friedman JM 1998 Absence of soluble leptin receptor in plasma from dbPas/dbPas and other db/db mice. J Biol Chem 273:10078 –10082 22. Lammert A, Kiess W, Bottner A, Glasow A, Kratzsch J 2001 Soluble leptin receptor represents the main leptin binding activity in human blood. Biochem Biophys Res Commun 283:982–988


23. Wallace AM 2000 Measurement of leptin and leptin binding in the human circulation. Ann Clin Biochem 37:244 –252 24. Tartaglia LA, Dembski M, Weng X, Deng N, Culpepper J, Devos R, Richards GJ, Campfield LA, Clark FT, Deeds J, et al. 1995 Identification and expression cloning of a leptin receptor, OB-R. Cell 83:1263–1271 25. Carlsson LM, Rowland AM, Clark RG, Gesundheit N, Wong WL 1991 Ligand-mediated immunofunctional assay for quantitation of growth hormonebinding protein in human blood. J Clin Endocrinol Metab 73:1216 –1223 26. Landt M, Horowitz JF, Coppack SW, Klein S 2001 Effect of short-term fasting on free and bound leptin concentrations in lean and obese women. J Clin Endocrinol Metab 86:3768 –3771 27. McConway MG, Johnson D, Kelly A, Griffin D, Smith J, Wallace AM 2000 Differences in circulating concentrations of total, free and bound leptin relate to gender and body composition in adult humans. Ann Clin Biochem 37:717– 723 28. Clement K, Vaisse C, Lahlou N, Cabrol S, Pelloux V, Cassuto D, Gourmelen M, Dina C, Chambaz J, Lacorte JM, Basdevant A, Bougneres P, Lebouc Y, Froguel P, Guy-Grand B 1998 A mutation in the human leptin receptor gene causes obesity and pituitary dysfunction. Nature 392:398 – 401 29. Murakami T, Otani S, Honjoh T, Doi T, Shima K 2001 Influence of the presence of OB-Re on leptin radioimmunoassay. J Endocrinol 168:79 – 86 30. Huang L, Wang Z, Li C 2001 Modulation of circulating leptin levels by its soluble receptor. J Biol Chem 276:6343– 6349