Modulation of Direct Leptin Signaling by Soluble Leptin Receptor

Abstract

Soluble leptin receptor (SLR) represents the major leptin binding activity in plasma. It is generated by alternative splicing of OB-R mRNA (OB-Re) and/or ectodomain shedding of membrane-spanning receptors. To determine the role of SLR in leptin activation of its long-form receptor, OB-Rb, we established in vitro assays using a cell line stably expressing OB-Rb. Human embryonic kidney 293 cell lines stably overexpressing OB-Rb show a dose-dependent leptin-induced signal transducer and activator of transcription 3 (STAT3) tyrosine phosphorylation and STAT3-responsive luciferase (STAT3-luc) activity. We demonstrate that when SLR is incubated with free leptin, binding of leptin to OB-Rb is reduced, with corresponding decrease of leptin-induced STAT3 tyrosine phosphorylation and STAT3-luc activity. However, by preparing leptin/SLR mixtures containing the same amount of free leptin but increasing amounts of leptin-SLR complex, we show that leptin-SLR complex does not inhibit OB-Rb activation by free leptin. These results suggest that in settings in which leptin and SLR coexist, SLR may sequester leptin from productive interactions with OB-Rb. Because plasma SLR levels are independently regulated by many different physiological and pathophysiological conditions, SLR may modulate the actions of leptin in tissues in which direct action of leptin has been demonstrated.

LEPTIN IS AN adipose tissue-derived hormone with potent effects on food intake, metabolism, and many other important physiological processes, ranging from fertility and angiogenesis to bone formation and immune function (1, 2). Leptin acts by binding to its receptor, OB-R, which is alternatively spliced into several isoforms. One of the splice variants, OB-Re, does not encode a transmembrane domain and is predicted to be secreted. Although OB-Re mRNA has been identified in rodents, this splice variant has not been found in any of the many human tissues examined (3, 4). However, soluble leptin receptor (SLR) does circulate in human plasma (5–10). In vitro and in vivo studies indicate that circulating leptin receptor may be generated by ectodomain shedding of membrane-spanning receptors, mediated by a metalloprotease (11, 12). The relative contribution of OB-Re mRNA-derived and ectodomain shedding-derived SLR in plasma is not known, because the sizes of circulating SLR generated by either mechanism are indistinguishable by Western blotting analysis (11). One exception was reported, however. It appears that pregnancy-induced production of SLR from OB-Re mRNA exhibits a different glycosylation pattern, although the mechanisms involved and consequences on leptin action are not known (13).

The generation of OB-Re mRNA and protein is regulated under both physiological and pathophysiological conditions. In mice, expression of OB-Re mRNA in the placenta is strongly induced at late stages of pregnancy, causing an up to 40-fold increase in circulating leptin receptor (14). In humans, levels of SLR are inversely related to adiposity (15–19), increased in patients with advanced chronic heart failure (20), and higher in female insulin-dependent diabetic subjects (21). However, the effect of SLR on leptin action in target tissues has not been determined, especially at peripheral sites where both leptin and SLR are present. The role of SLR in leptin transport across the blood brain barrier also remains unresolved. Because the affinity of leptin toward membrane-spanning receptors and full-length SLR is indistinguishable (22, 23), SLR may not transport leptin to the hypothalamus, where it is required for leptin’s weight reducing and other effects (24). The blood-cerebrospinal fluid barrier at the choroid plexus and the blood-brain barrier at the cerebral endothelium are two major controlling sites for entry of circulating proteins into the brain, locations at which membrane-spanning leptin receptors are enriched and across which leptin is presumably transported (25–27). Transcellular transport of intact leptin in vitro has been reported (28). No SLR has been detected in cerebrospinal fluid (29).

Previously, we and others reported that SLR is the major determinant of plasma leptin levels and acts to stabilize circulating leptin without increasing leptin mRNA transcription (30, 31). We also found that, similar to human SLR, mouse SLR may be generated by ectodomain shedding of membrane-spanning receptors both in vitro and in vivo (11). In this study, we report that when leptin is bound to its soluble receptor, the leptin-SLR complex is not capable of activating OB-Rb, nor does it inhibit the action of free leptin. When the ability of a fixed amount of leptin to activate OB-Rb with or without the presence of SLR is compared, leptin-stimulated OB-Rb signal transduction is reduced in the presence of SLR. These results suggest that SLR serves to sequester leptin from productive interactions with its signaling receptor, OB-Rb, and may play a role in leptin action in tissues in which both leptin and SLR coexist.

RESULTS

Determination of Leptin Signaling Using Stable Cell Lines

Leptin binding to its long-form receptor, OB-Rb, causes activation of the Jak2 kinase, which is constitutively associated with OB-Rb. Activation of Jak2 leads to the phosphorylation of tyrosine residues within the cytoplasmic domain of OB-Rb. Two tyrosines are phosphorylated after leptin stimulation and become docking sites for Src homology 2 (SH2) domain containing signaling proteins (32). In the mouse receptor, phosphorylation of Tyr 985 leads to the recruitment of SHP-2, a protein tyrosine phosphatase that serves to decrease Jak2 phosphorylation and as an adaptor protein to activate the ERK/MAPK pathway (32, 33); phosphorylation of Tyr 1138 leads to the recruitment of signal transducer and activator of transcription 3 (STAT3) in vivo (34) and other STATs in vitro. Activation of STAT3 can be determined by following STAT3 tyrosine phosphorylation or its ability to increase transcription of reporter genes containing multiple copies of STAT3 binding sites in the promoter region (STAT3-luc).

Efforts from many laboratories have failed to identify cell lines with high-level expression of endogenous OB-Rb. To overcome this difficulty in studying leptin signal transduction pathway, we and others have been performing transfection of OB-R expression vectors transiently or permanently to detect leptin signaling (32, 35). Using a Flp-In human embryonic kidney 293 (HEK293) cell line that allows the integration of one copy transfected cDNA per cell, based on Cre-Lox recombination strategy, we were able to generate stable cell lines that express recombinant proteins without the influence of insert copy number, therefore allowing expression of mRNA to similar levels among the cell lines generated (11).

We measured leptin response in stable HEK293 cells overexpressing either OB-Ra or OB-Rb. OB-Ra, which cannot signal by itself, is used as a negative control; OB-Rb, which encodes a long cytoplasmic domain containing both Tyr 985 and Tyr 1138, is a target of Jak2 phosphorylation. Cell surface expression of OB-R protein was quantified by incubation of cells with radiolabeled leptin with or without unlabeled, cold leptin as competitor (Fig. 1B). There is less leptin binding on the surface of OB-Rb-expressing cells, consistent with earlier studies that OB-Rb translocation to plasma membrane is less efficient compared with OB-Ra (36, 37). However, when cells are treated with leptin after incubation under serum-free conditions, leptin caused STAT3 activation only in OB-Rb-expressing cells, as determined by STAT3 tyrosine phosphorylation (Fig. 1C) and STAT3-luc activity (Fig. 1D), confirming the specificity of these cell lines as useful reagents in determining leptin receptor signal transduction in the absence or presence of SLR.

Dose-Dependent Activation of STAT3 in Stable Cell Lines

One purpose of generating stable cell lines overexpressing OB-Rb is to determine the effect of SLR on leptin signal transduction. We reasoned that SLR might either enhance or inhibit leptin action in this setting. To distinguish between these possibilities, our system must be able to respond to different levels of leptin stimulation with graded responses. To determine that our stable cell lines can indeed function in such a manner, we first generated a leptin dosage-responsive curve in the absence of exogenously added SLR. Figure 2 shows that when leptin concentrations range from 0.0016 nm to 1 nm, both STAT3-luc activity (Fig. 2A) and STAT3 tyrosine phosphorylation (Fig. 2B) in treated cells showed parallel increases. In the experiments described below, we have chosen free leptin concentrations of between 0.008 nm and 0.2 nm, a range that gives rise to STAT3 responses close to being linear.

Preparation of Leptin-SLR Complexes with Defined Concentrations of Free and Bound Leptin

To determine the effect of SLR on leptin receptor signal transduction, we also needed to prepare leptin-SLR complexes with defined absolute and relative concentrations of each component. To achieve this goal, we took advantage of an efficient adenoviral vector system encoding OB-Re (Ad-OB-Re), which expresses recombinant SLR, as reported earlier (31). We used commercial preparations of bacterially derived recombinant leptin and added to it conditioned serum-free media from cells infected with Ad-OB-Re or a control virus that encodes β-galactosidase, Ad-β-Gal. Figure 3A shows that SLR accumulation in media of Ad-OB-Re-infected cells is readily detected with a polyclonal antibody specifically recognizing the extracellular domain of all OB-R isoforms (11, 31).

To prepare SLR-leptin complexes with defined free leptin concentration, we took advantage of the differential migration properties of leptin in a gel filtration column in free form or bound to SLR (7). The distribution of peaks corresponding to bound and free leptin is determined by incubating SLR-containing media with [¹²⁵I]leptin tracer, which can be quantitated by scintillation counting. The percentage of free and bound leptin is determined by integration of areas under the curve for each peak. Finally, the absolute concentrations of free and bound leptin are obtained by multiplying that of total leptin added to the incubation mixture with the percentage of each fraction. To validate this procedure, we used either a fixed amount of free leptin and varying amounts of SLR or a fixed amount of SLR but varying amounts of free leptin and performed gel filtration chromatography (Fig. 3B and data not shown). Other leptin binding activity is not present in supernatant of virus-infected cells as incubation of [¹²⁵I]-leptin tracer with an excess volume of conditioned media of Ad-β-Gal-infected cells did not reveal any additional peaks that elute ahead of free leptin (Fig. 3C).

Addition of SLR to a fixed amount of leptin inhibits leptin binding to OB-Rb HEK293 cells as well as OB-Rb signal transduction.

With the establishment of stable cell lines that respond to the concentration of free leptin linearly at a defined range (Fig. 2) and our ability to prepare SLR-leptin complexes with accurate determination of the percentage of free and bound leptin (Fig. 3), we proceeded to determine whether SLR enhances or inhibits leptin action.

Based on the gel filtration data, we mixed 1.6 ng leptin in 1 ml of buffer, corresponding to 0.1 nm leptin, which by itself elicits an intermediate STAT3 tyrosine phosphorylation and luciferase response (Fig. 2), with varying volumes of SLR-containing serum-free media to obtain a final leptin preparation that is 20%, 43%, and 86% bound to SLR (Fig. 3B). OB-Rb HEK293 cells were then treated with these leptin samples, containing either free leptin alone or leptin bound to varying amounts of SLR. Because binding of leptin to SLR is maintained at equilibrium, it is impossible to obtain leptin samples that are 100% bound. As an added control, leptin was mixed with serum-free conditioned media from Ad-β-Gal-infected HEK293 cells.

We first assessed how leptin binding to cell surface OB-Rb is affected in the absence or presence of increasing amounts of SLR. [¹²⁵I]leptin alone or added different volumes of SLR-containing supernatant were incubated with OB-Rb HEK293 cells, and specific leptin binding to cell surface OB-Rb was determined (Fig. 4A). Addition of SLR caused a dose-dependent decrease of cell surface [¹²⁵I]leptin binding, suggesting that SLR sequesters leptin from interaction with membrane-bound receptors.

The effects of SLR on leptin activation of OB-Rb in HEK293 cells are shown in Fig. 4B. As opposed to when all leptin is in the free form, addition of SLR to a fixed amount of leptin caused a dose-dependent decrease in leptin-activated luciferase activity. In agreement with the luciferase response, STAT3 tyrosine phosphorylation is also similarly reduced by increasing the amounts of SLR. This decrease is not due to the nonspecific effects caused by the presence of conditioned media, because addition of supernatant from Ad-β-Gal-infected cells did not significantly decrease leptin-induced STAT3 tyrosine phosphorylation or STAT3-luc activity (Fig. 4B), although a trend toward a slight decrease in STAT3-luc activation was observed when leptin was incubated with increasing volumes of conditioned media from Ad-β-Gal-infected cells (Fig. 4B). In aggregate, these results demonstrate that addition of SLR to a fixed amount of leptin decreases the ability of leptin ability to activate OB-Rb, as measured by both STAT3 tyrosine phosphorylation and STAT3-luc reporter activity.

SLR-Leptin Complex Does Not Affect Activation of OB-Rb by Free Leptin

We then asked whether binding of leptin to its soluble receptor merely sequesters it from binding to OB-Rb, as shown above, or whether the leptin-SLR complex also inhibits OB-Rb activation by leptin in a dominant negative fashion. To address this question, leptin was mixed with conditioned media from Ad-β-Gal-infected cells to a final concentration of 0.08 nm, or mixed with different volumes of serum-free conditioned media from Ad-OB-Re-infected cells containing 0.08 nm free leptin but different amount of leptin-SLR complex. In these samples, 34% or 88% of leptin is complexed with SLR, verified by fast protein liquid chromatography (FPLC) (Fig. 5A). To take into consideration the potential influence of media components on leptin activation of OB-Rb, the difference in volume of SLR-containing media was corrected by adding that from Ad-β-Gal-infected HEK293 cells, so that total media volume in each sample is identical. OB-Rb HEK293 cells were transfected with STAT3-luc and treated with the leptin samples prepared above. Figure 5B shows that at identical concentrations of free leptin, the presence of leptin-SLR complex did not cause appreciable decrease in STAT3-luc activity, suggesting that SLRs simply sequester leptin away from being available for OB-Rb interaction without a dominant negative role. This result is consistent with earlier studies, which demonstrated that leptin receptors tend to form homodimers but do not form heterodimers (38). This result is also reminiscent of the role of the complex of GH/GH binding protein, which has been demonstrated to inhibit or enhance GH signaling under different conditions (39). The role of SLR on leptin action in vivo is also likely to be similarly complex.

DISCUSSION

The ability of leptin to activate its receptor and mediate signal transduction depends on the presence of the signaling form of its receptor, OB-Rb. Leptin activation of OB-Rb centrally causes reduced food intake and increased energy expenditure. These central effects are mediated by a complex neural circuit comprised of neuropeptide Y/agouti-related protein and proopiomelanocortin/cocaine- and amphetamine-regulated transcript neurons, which are modulated by leptin (40–42). In peripheral tissues, direct leptin action has also been shown to mediate a wide range of physiologically important, tissue-specific processes, such as wound healing, angiogenesis, T helper cell function, and AMP-activated protein kinase activation (1, 43). We hypothesized that these central and peripheral actions of leptin might be enhanced or inhibited by leptin binding to SLR. To test this prediction, we used an in vitro luciferase reporter assay to determine how free vs. bound leptin affects leptin signal transduction.

Previously, we generated HEK293 cells stably expressing the signaling form of the leptin receptor, OB-Rb (11). Transfection of a luciferase reporter driven by STAT3 responsive elements into these cells allows leptin activation to be assayed by both STAT3 tyrosine phosphorylation and luciferase activity in the lysate (Fig. 1). We sought to determine how STAT3 phosphorylation and STAT3-luc reporter activity are differentially regulated when leptin is added to these cells either alone or in combination with different amounts of SLR. In this report, we demonstrate that using STAT3 tyrosine phosphorylation and STAT3-luc reporter assays, leptin bound to SLR is not capable of activating its long-form receptor, OB-Rb, nor does it inhibit activation of OB-Rb by free leptin. However, when leptin-free SLR is added to a fixed amount of leptin, OB-Rb activation is reduced, presumably by sequestering leptin from productive interactions with OB-Rb. Soluble receptors for a number of cytokines exist and, in many cases, also inhibit the action of these cytokines at their respective target sites (44, 45). These results, coupled with earlier studies showing that SLR is the main leptin-binding activity in plasma (30), have important implications for our understanding of the regulation of the biological activity of leptin in vivo.

Many earlier studies have demonstrated that leptin signals in a variety of peripheral tissues to mediate a variety of physiological functions (1). These tissues are also exposed to SLR. Acute SLR induction is thus likely to inhibit leptin signaling temporally at these sites. The neutralization of leptin action at these sites by SLR could be protective in times of famine. However, in a leptin resistance state, increased amounts of SLR might dampen leptin action when it is needed, potentially contributing to local leptin resistance and the pathogenesis of obesity.

The SLR may also affect leptin’s activity in the brain, in combination with other receptor isoforms (27, 46, 47). To gain access to the central nervous system, leptin needs to cross the blood brain barrier, a process that is saturable, yet not fully defined (25, 46–50). SLR might play a role in this process. Further studies to determine the effect of SLR on leptin transport into the central nervous system should shed more light on this very important question.

Prior studies have demonstrated that levels of circulating leptin vary, such as between males and females (51). In humans, serum leptin per unit fat mass is significantly higher at 36 wk of gestation than at 3 and 6 months postpartum (52). In rodents, SLR levels are similarly increased by up to more than 40-fold during late stages of pregnancy (14). Mutation of leptin receptor in Zucker diabetic fatty rat also caused an increase of plasma SLR (31). Our present data suggest that SLR may play a significant role in determining the levels of total leptin and free leptin, as well as leptin’s biological activity. Determination of the levels of free and bound leptin in individuals with hyperleptinemia should be useful in the diagnosis and treatment of certain forms of obesity. Given that leptin acts both centrally, where SLR has not been detected, and peripherally, where both free and bound leptin circulate, the biology of SLR on leptin action is likely to be complex. Analysis of transgenic mice overexpressing SLR, currently underway, may provide additional interesting insights on the role of SLR on leptin action in vivo.

MATERIALS AND METHODS

Reagents

Recombinant leptin was from Sigma Chemical Co. (St. Louis, MO) or from Dr. A. F. Parlow, National Hormone & Peptide Program; murine recombinant ¹²⁵I-leptin was from Perkin-Elmer Life Sciences (Boston, MA); DMEM, BSA, penicillin, and streptomycin were from Sigma. FUGENE 6 transfection reagent was from Roche (Indianapolis, IN); luciferase assay system was from Promega Corp. (Madison, WI). Enhanced chemiluminescence Western blotting detection reagents and gel filtration columns were from Amersham Biosciences (Piscataway, NJ). Antiphospho-STAT3 antibody was from Cell Signaling Technology, Inc. (Beverly, MA). Antibody against total STAT3 was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Bradford assay kit was from Bio-Rad Laboratories, Inc. (Hercules, CA).

Cell Culture

Flp-In HEK293 cell lines stably over-expressing OB-R were generated as described (11). Cells were maintained in DMEM containing 10% fetal bovine serum and 100 U/ml penicillin and 100 μg/ml streptomycin in a 5% CO₂, humidified environment at 37 C. The two cell lines used in the current study, encoding OB-Ra and OB-Rb, are schematically diagrammed in Fig. 1A.

[¹²⁵I]Leptin Binding to Stably Transfected Cells

Leptin binding to Flip-In OB-Rb HEK293 cell lines was performed in six-well plates as described (11). Briefly, cells stably transfected with OB-Ra or OB-Rb were grown to about 90% confluence and washed with cold PBS. Cells were incubated with approximately 60,000 cpm of [¹²⁵I]leptin alone, or [¹²⁵I]leptin supplemented with conditioned media containing SLR, in the presence or absence of an excess amount of cold leptin (2 μg/well) for 6 h at 4 C in a final volume of 1 ml binding buffer [PBS containing 1% (wt/vol) BSA (fraction V)]. At the end of incubation, unbound label was removed by two PBS washes; 1 ml of 1 n NaOH was then added, and radioactivity in lysate was measured using a COBRA II AUTO-GAMMA counter from Packard Instruments (Meriden, CT).

Western Blotting Analyses

Cells were lysed in RIPA buffer, containing 150 mm NaCl, 50 mm Tris-HCl, 1% Triton X-100, 2% Igepal CA-630, 1 mm phenylmethylsulfonyl fluoride, 1 mm Na₃VO₄, 1 mm NaF, 10 μg/ml leupeptin, and 10 μg/ml aprotinin. Lysate was incubated on ice for 20 min with gentle rocking and centrifuged at 13,000 × g for 15 min at 4 C. Protein concentration was determined by Bradford assay. Protein samples were boiled in 2× reducing sample buffer and subjected to SDS-PAGE, followed by electrophoretic transfer to nitrocellulose membrane. After incubation with antibodies, target protein was detected using enhanced chemiluminescence, following the manufacturer’s instructions.

STAT3-Driven Luciferase Assay

When cells were approximately 80% confluent, transfection was performed by using FUGENE 6 Transfection Reagent with 0.5 μg/well STAT3-luc DNA, p4Xm67, containing four copies of the STAT3 binding sequences fused to luciferase cDNA (11). Two days after transfection, media were aspirated from cells and replaced with DMEM containing 0.5% BSA. After overnight incubation under serum-free conditions, cells were stimulated with different concentrations of leptin. Cells were harvested 15 min after leptin stimulation to determine the level of STAT3 tyrosine phosphorylation or after 3 h to determine STAT3-luc activity.

Preparation of Conditioned Media Containing SLR

HEK293 cells were grown in 15-cm plates until 90% confluent. Cells were washed with PBS, and then 20 ml of DMEM without serum were added. Purified recombinant adenoviruses encoding OB-Re (Ad-OB-Re) or β-galactosidase (Ad-β-Gal) (20 μl), each with a physical titer of 4 × 10¹² particles/ml, were added to each plate. Two days after virus infection, medium was collected and filtered through a 0.22-μm syringe filter. Expression of SLR was verified by Western blotting.

Titration of Leptin Binding to SLR and Determination of Free and Bound Leptin by FPLC

HEK293 cells were cultured in serum-free DMEM at 37 C, 5% CO₂, and infected with Ad-OB-Re or Ad-β-Gal. After incubation for 3 d, conditioned medium was harvested and mixed with cold recombinant mouse leptin and tracer [¹²⁵I]leptin. Murine recombinant [¹²⁵I]leptin with a specific activity of 2200 Ci/mmol (81.4 TBq/mmol) was used as a tracer. In separate 1.5-ml Eppendorf tubes, 0, 8, 16, and 80 μl of conditioned medium from infected HEK293 cells were mixed with 1.6 ng cold leptin and 2 μl [¹²⁵I]leptin and incubated at 4 C overnight in a total volume of 1 ml. Subsequently, the mixture was analyzed by FPLC, using an AKTA design chromatography system from Amersham Biosciences. Gel filtration was performed on a Superdex 200 HR (10/30) column, equilibrated and eluted with the same buffer (0.05 m NaPO₄ and 0.15 m NaCl, pH 7.4) at 4 C. The 1-ml mixture (200 μl) was loaded onto the column, and 80 fractions were collected, each having a volume of 300 μl. The amount of radioactivity in each fraction was determined by the use of a COBRA II AUTO-GAMMA counter. The areas under the curve of each of the two peaks, corresponding to the elution position of bound and free leptin, were calculated with the software Prism (GraphPad Software, Inc., San Diego, CA). Based on the assumption that tracer distribution reflects that of cold leptin, we obtained the absolute amount of bound and free leptin in the mixture by multiplying the percentage of free and bound leptin with the amount of total leptin added in each incubation mixture. All data were analyzed with Prism 3 software.

Acknowledgments

We thank Drs. Roger Unger and Joyce Repa for many thoughtful comments and a critical review of this manuscript; and Drs. Jeff Friedman and Paul Cohen (Rockefeller University) for many discussions and sharing of unpublished data.

This work was supported by Grant 0255737Y from the American Heart Association, Texas Affiliate, and National Institutes of Health Grant DK 60137. C.L. is the recipient of a Career Development Award from the American Diabetes Association.

Abbreviations

 
• FPLC

  Fast protein liquid chromatography;

 
• HEK

  human embryonic kidney;

 
• SLR

  soluble leptin receptor;

 
• STAT

  signal transducer and activator of transcription.