| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Institute of Laboratory Medicine, Clinical Chemistry, and Molecular Diagnostics (A.G., J.K.), and Department of Dermatology, University of Leipzig (U.A.), 04103 Leipzig, Germany; and Children’s Hospital (A.G., W.K., A.Be., A.Bo.), 04317 Leipzig, Germany
Address all correspondence and requests for reprints to: J. Kratzsch, Ph.D., Institute of Laboratory Medicine, Clinical Chemistry, and Molecular Diagnostics, University of Leipzig, Liebigstrasse 27, D-04103 Leipzig, Germany. E-mail: kraj{at}medizin.uni-leipzig.de
Abstract
Leptin, a hormone of the cytokine family, is mainly synthesized by white adipocytes. As fibroblasts and adipocytes share a common stem cell origin, we hypothesized that connective tissue may be another candidate for leptin synthesis. We demonstrated leptin receptors, inclusive of all isoforms, on cultured fibroblasts (n = 13) by RT-PCR and immunohistochemistry. In contrast to its receptor, basal leptin mRNA expression and protein secretion were found in 8 of 13 cultures, reaching 1.4 ng/350,000 cells·24 h. Incubation with physiological insulin concentrations (1 nmol/liter) increased leptin secretion in fibroblast culture supernatants to 152% of basal levels. A maximal stimulation of the basal level up to 192% was found with 10 nmol/liter insulin after 24 h. Actinomycin D and cycloheximide abolished this effect, providing evidence that active RNA and protein synthesis are involved in insulin’s action. Completing these in vitro results, we could show protein expression for leptin and leptin receptors in fibroblasts by immunostaining of human skin biopsies in situ. In conclusion, we provide evidence of leptin synthesis and secretion by human fibroblasts that are regulated by insulin. Leptin produced by fibroblasts may thus exert important local autocrine and paracrine actions and contribute to the total plasma pool. Hence it might in part account for variations in body mass index-dependent reference ranges of leptin as well as disruptions in the relationship between fat content and leptin.
LEPTIN, A 16-kDa protein, is produced primarily by adipocytes. Its tertiary structure resembles that of the cytokine family, as demonstrated by structural similarities with, for instance, ciliary neurotropic factor and leukemia inhibitor factor (1). From the adipose tissue leptin is released into the peripheral blood and transported across the blood-brain barrier, and acts via central receptors to control body weight homeostasis and energy expenditure (2). The leptin receptor, a single membrane-spanning glycoprotein, belongs to the class I cytokine receptor superfamily, sharing sequence homologies for interaction with Janus kinase as well as signal transducers and activators of transcription (3). To date, the human leptin receptor (Ob-R) is identified as a full-length form B219 (4) and three membrane-spanning isoforms (B219.1–3) with shorter intracellular domains generated by alternative splicing (5, 6). Furthermore, a soluble extracellular domain of the Ob-R seems to act as a leptin-binding protein in the peripheral blood (7, 8, 9). The hypothalamus is supposed to be the only tissue with predominantly expressed full-length Ob-R. However, an increasing number of peripheral tissues are described to express the long form as well as several isoforms (4, 6, 10, 11). Therefore, direct effects on peripheral tissues could underlie some of leptin’s biological actions, as identified by recent studies: leptin induces oxidative stress in endothelial cells (12), leptin acts as an angiogenic factor in the vascular endothelium (13), leptin promotes aggregation of human platelets via the long form of its receptor (14), and leptin can induce proliferation, differentiation, as well as functional activation of hemopoietic cells (15). These actions may be caused by leptin originating from adipose tissue or at least partially from leptin originating from other, as yet unknown, local sources. The latter idea is supported by recent observations demonstrating leptin production by nonadipose tissues and cells, including placenta, human gastric mucosa, hepatic stellate cells, mammary epithelial cells, pituitary, brain, and skeletal muscle (16, 17, 18, 19).
Interestingly, connective tissue, which is one of the most important sources for cytokine synthesis and action, has not been intensively investigated with regard to autocrine or paracrine leptin effects and leptin receptor expression. As fibroblasts and adipocytes are closely related due to their common stem cell origin, we hypothesized that fibroblasts could be a further candidate for leptin production. In this study we provide evidence that Ob mRNA is expressed and leptin protein is secreted by human skin fibroblasts. Similar to adipocytes, in which leptin synthesis is adapted to the blood glucose level through its dependence on insulin levels (20), we demonstrate that leptin secretion by fibroblasts is up-regulated by insulin in vitro. In addition, we showed the expression of different leptin receptor isoforms in fibroblasts, supporting the idea of an ubiquitous presence of autocrine and/or paracrine actions of leptin in peripheral tissues.
Materials and Methods
Materials
Normal skin biopsy material and fibroblast cell cultures were provided by Prof. Dr. H. Klein and Dr. M. Schuett, Medical University of Luebeck (Luebeck, Germany) and Prof. F. Beguinot (University of Naples, Naples, Italy) and were obtained from patients at Department of Dermatology, University of Leipzig (Leipzig, Germany). The study was approved by the local ethical committees of University of Leipzig, University of Luebeck, and University of Naples, Biomed 2 program (BMH-CT96-0751, European Union, Brussels, Belgium). Chemicals were obtained from Sigma-Aldrich Corp. (Steinheim, Germany) or Merck & Co., Inc. (Darmstadt, Germany), and molecular biology reagents were purchased from Roche (Mannheim, Germany) if not otherwise indicated.
Cell culture
Fibroblast cultures were derived after outgrowth from human normal skin biopsies [n = 13, six males and seven females, aged 23–70 yr; body mass index (BMI), 21.1–29.9 kg/m2]. They were cultured in 30 ml DMEM/nutrient mix F-12, (DMEM/F-12 containing 15 mM HEPES and L-glutamine, Life Technologies, Inc., Eggenstein, Germany) supplemented with 10% FBS (Biochrom, Berlin, Germany), 2 mM L-glutamine (Life Technologies, Inc.), 100 U/liter penicillin, 0.1 mg/liter streptomycin, and 0.25 µg/ml amphotericin (Sigma-Aldrich Corp.) in 175-ml tissue culture flasks (Becton Dickinson and Co., Meylan Cedey, France). Medium was replaced every 72 h. For passaging, cells were detached by 0.05% (wt/wt) trypsin and 0.02% EDTA. UV light tests after staining with DNA staining reagent (Bisbenzamid-Fluorochrom, Hoechst, Frankfurt, Germany) were performed, confirming that cell cultures were not contaminated by mycoplasms.
Incubation experiments
Confluent cells (passages 5–12) from 175-ml culture flasks were plated on 25-ml culture flasks (Becton Dickinson and Co.). After reaching confluence, cells were washed with serum-free medium (DMEM/F-12 with 15 mM HEPES, 2 mM L-glutamine, antibiotics, and 0.1% BSA, Sigma-Aldrich Corp.) and grown in 1.25 ml serum-free medium/flask. After resting for 4 h, 100 µl mediators (0.1–10 nmol/liter insulin and/or 10 µg/ml cycloheximide diluted in serum-free medium) were added, and supernatants were collected after defined incubation times (3–48 h) for leptin measurement. Human insulin (Actrapid, Novo Nordisk, Mainz, Germany) was purified from stabilizers using Sepharose columns (PD-10 Sephadex G-25M, Pharmacia Biotech, Uppsala, Sweden), and its concentration was measured by RIA. Degradation of insulin within 24 h was measured by RIA in cell supernatants and was between 70–85%. Cylcoheximide (dissolved in ethanol, diluted in serum-free medium) and actinomycin D (dissolved and diluted in serum-free medium) were obtained from Sigma-Aldrich Corp. Control incubations with corresponding ethanol concentration and medium controls without cells were carried out to check ethanol effects and basal medium leptin levels.
2-(4-Iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt (WST-1) assay
The effects of insulin on cell viability and proliferation were studied using the WST-1 assay (Roche), which correlates the enzyme activity of the mitochondrial succinate-dehydrogenase system with cell number and activity.
After incubation with insulin as described above and removing the supernatants of the cells for RIA measurements, fibroblast cell cultures were incubated with serum-free medium containing 5% WST reagent for 30 min at 37 C in 5% CO2. Afterward, 100 µl supernatant were transferred to a microtiter plate, and extinction was measured by an ELISA reader (SLT Rainbow, SLT Labinstruments, Crailsheim, Germany) at 450 nm using serum-free medium with 5% WST as background control. A possible direct effect of insulin on the cell viability was excluded by a short (3 h) incubation with insulin. A linear standard curve (10,000–400,000 cells) was determined to correlate the extinction to cell numbers.
RNA isolation and RT-PCR
Total RNA was isolated from cells grown to confluence on 75-ml flasks using QIAamp columns (QIAamp, QIAGEN, Hilden, Germany). DNA was digested with ribonuclease-free deoxyribonuclease (QIAGEN) following te original protocol, and complete DNA removal was confirmed by PCR for glyceraldehyde-3'-phosphate dehydrogenase (21) with RNA template. For first strand synthesis, 10 µl reaction mix containing 1 µg RNA, 1 x RT buffer (Life Technologies, Inc.), 200 µM deoxy-NTPs, 200 ng random primers [p(dN)6], and 0.01 M dithiothreitol were denatured at 65 C for 5 min and incubated on ice for 2 min. After adding 200 U Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.) and 40 U ribonuclease inhibitor, reaction was performed at 39 C for 60 min and was stopped by denaturation at 94 C for 2 min. Contamination controls omitting enzyme in the reaction were performed. The integrity of each template was confirmed by control PCR for glyceraldehyde-3'-phosphate dehydrogenase.
PCR was performed in a thermal cycler (GeneAmpPCR system 2400, Perkin-Elmer Corp., Uberlingen, Germany) using the Expand high fidelity system (Roche). PCR primers and conditions were previously described for Ob-R (11) and Ob amplification (10). For both PCRs, 38 cycles were applied, and the denaturation time was shortened to 15 sec. In all PCR experiments amplification in the absence of cDNA was performed as a negative control. For Ob and Ob-R PCR, cDNA from sc adipose tissue served as positive control. PCR products were analyzed in a 1.5% agarose gel containing ethidium bromide, and length was estimated by a 100-bp ladder (Life Technologies, Inc.).
Leptin measurements by RIA
A competitive in-house immunoassay was applied to measure leptin in cell culture supernatants. Leptin standards ranged between 0.2–16 ng/ml. To increase the detection limit for leptin, cell culture supernatants (1.2 ml) were lyophilized using a vacuum concentrator (Bachhofer, Reutlingen, Germany) and subsequently dissolved in 120 µl PBS. Standards, supernatants, or medium controls were mixed with 0.050 ml 125I-labeled leptin and incubated with a 1:10,000 dilution of polyclonal rabbit antibodies raised against human recombinant leptin (PeproTech, Rocky Hill, NJ) for 16–20 h at 4 C. A mixture of antirabbit IgG and polyethylene glycol 6000 was added for double antibody precipitation. The sensitivity of the RIA (2 SD of the 0 ng/ml level; n = 12) was 0.2 ng/ml. Intra- and interassay coefficients of variation were less than 12.5% in the range between 1–8 ng/ml leptin. Standard curves were prepared in media. Dilution experiments with pooled samples as well as samples spiked with leptin gave linear curves for the concentration range of 1–5 ng/ml. Leptin results are presented as the original concentrations of the supernatants and are given as nanograms per cells.
Immunohistochemistry
For immunohistochemical studies four different fibroblast secondary cell cultures (passages 7–12) and skin biopsies from three patients were analyzed. Cells were cultured on glass chamber slides (Nalge Nunc International, Naperville, IL) to confluence. Normal skin biopsies were fixed in 4% paraformaldehyde in PBS at 4 C for 30 min and incubated in 20% saccharose in PBS for 12 h. Afterward they were frozen in isopentane at -70 C, sectioned (5 µm), dried for 30 min, and stored at -80 C.
Leptin, leptin receptor, Thy-1, and prolyl-4-hydroxylase staining.Tissue sections and cells were washed in PBS for 5 min and fixed in 3% paraformaldehyde (for leptin and leptin receptor) or methanol (for Thy-1) for 15 min. Staining was performed using the LSABplus system (DAKO Corp., Carpenteria, CA) following the manufacturer’s protocol. Specimens were additionally treated with Triton X-100 (0.5% in Tris-buffered saline) for 3 min before blocking the endogenous peroxidase for 20 min. Primary antibodies were diluted in TBS containing 2% normal porcine serum. The following antibodies against leptin and leptin receptor were used: rabbit polyclonal antihuman leptin antibody (A-20, IgG, 8 µg/ml), goat polyclonal antihuman leptin receptor antibody (C-20, IgG, 8 µg/ml, corresponding to amino acids 1146–1165 detecting the long form of the leptin receptor), and corresponding blocking peptides (sc-842P and sc-1832P; Santa Cruz Biotechnology, Inc., Heidelberg, Germany) and in-house antibodies: rabbit polyclonal antibody against human leptin receptor, recognizing the extracellular domain (Ob-R1984, IgG, 7 µg/ml) and a rabbit polyclonal antihuman leptin antibody (Ob1514, IgG, 8 µg/ml).
For the characterization of fibroblasts, the mouse antihuman Thy-1 monoclonal antibody (clone AS02, IgG, 4 µg/ml; Dianova, Hamburg, Germany) (22) and mouse antihuman fibroblast monoclonal antibody (clone 5B5, IgG, 1.6 µg/ml, detecting prolyl-4-hydroxylase; DAKO Corp., Hamburg, Germany) were used.
For negative control, antibodies were replaced by corresponding nonimmune IgG preparations in the appropriate protein concentration (Santa Cruz Biotechnology, Inc.). Subcutaneous fat tissue served as a positive control for leptin and leptin receptor staining. Visualization of antibody reaction was achieved by incubation with 3-amino-9-ethyl-carbazole solution (Coulter-Immunotech, Marseilles, France) for 15 min. Specimens were counterstained with hematoxylin for 1 min, rinsed with water for 10 min, and mounted with glycerin gelatin.
Sudan III staining
Staining was performed on frozen slices of skin biopsies with adjacent sc white adipose tissue for positive controls and on cultivated leptin-producing fibroblast cells (passages 7–12). Tissue sections were washed in PBS, fixed in formaldehyde (4% in PBS) for 15 min, and washed in aqua distilled and 50% ethanol each time for 3 min. Cultured fibroblast cells were not fixed but washed in PBS for 5 min. All specimens were incubated in Sudan III staining solution (Dr. K. Hollborn and Soehne GmbH, Leipzig, Germany) for 30 min, followed by washing in 50% ethanol for 3 min, counterstained, and mounted as described above.
FACS analysis: detection of Thy-1 expression
Cells (2 x 105) were detached by 0.025% trypsin/0.01% EDTA (Life Technologies, Inc.), washed twice with PBS containing 10% Gelafusal (Serumwerk, Bernburg, Germany), and labeled with 10 µg/ml mouse antihuman Thy-1 monoclonal antibody (clone AS02, Dianova) at 4 C for 45 min. For negative controls, primary antibody was replaced by nonimmune mouse IgG. After three washes with PBS/10% Gelafusal, cells were incubated with a 1:200 dilution of the secondary fluorescein isothiocyanate-conjugated F(ab)2 goat antimouse antibody (Dianova) at 4 C for 45 min. Flow cytometric analysis was performed after three final washes with PBS/10% Gelafusal on a FACSCalibur (Becton Dickinson and Co.). Duplicate samples with about 6000 cells each were counted.
Statistics
All experiments for hormone measurements were repeated for a minimum of three different human secondary fibroblast cell cultures using quadruplicate determinations. Data are expressed as the mean ± SEM, calculated as percentages of basal levels if not otherwise indicated. All nonstimulated values are defined as basal levels. Statistical analysis was performed on the original (not on percentage) data using SPSS 9.0.1 for Windows, UNIANOVA (SPSS, Inc., Chicago, IL). By Tukey’s test it was determined which treatment pairs had significantly different rates for leptin release. Differences between stimulated and basal values are marked with asterisks, and differences between stimulated values are indicated with crosses. Differences were considered statistically significant for P < 0.05 (*/+), P < 0.01 (**/++), and P < 0.001 (***/+++).
Results
Characterization of fibroblast cultures by staining for Thy-1, prolyl-4-hydroxylase, and Sudan III
Quantification of Thy-1-positive cells by FACS analysis. Human fibroblast cultures were analyzed by flow cytometry for their expression of Thy-1, a fibroblast-specific surface protein. Analysis of three different human skin fibroblast secondary cell cultures (passages 7–12) revealed 99.83 ± 0.085% Thy-1-positive cells, confirming cell type homogeneity of the investigated cultures. The specificity of the analysis was demonstrated by a lack of staining in negative controls (0.37 ± 0.01%).
Immunohistochemistry for prolyl-4-hydroxylase (EC 1.14.11.2).
The mouse antihuman fibroblast monoclonal antibody, 5B5, recognizes the
ß-subunit of prolyl-4-hydroxylase (23).
Prolyl-4-hydroxylase catalyzes the hydroxylation of proline residues in
collagen, which is the major protein synthesized by active fibroblasts.
By immunostaining we could detect this enzyme in the cytoplasm in
nearly all cells of four different fibroblast cell cultures, suggesting
active collagen synthesis. Negative controls with mouse nonspecific IgG
showed a negligible staining (Fig. 1
, A
and B).
|
Ob and Ob-R mRNA expression by RT-PCR in fibroblasts
Ob mRNA expression was observed in 8 of 13 cultures (Fig. 2A). In contrast, the full-length Ob-R
(219) and the three known membrane-spanning isoforms (219.1–3) were
expressed in all fibroblast cell cultures investigated (n = 13;
Fig. 2, B–E). Human sc fat tissue served as a positive control
(24, 25) for the expression of Ob-R as well as for Ob
mRNAs.
|
Three of the human secondary fibroblast cell cultures shown to
express Ob mRNA were examined for leptin protein expression by
immunostaining. Corresponding to the RT-PCR results, all three of them
expressed positive immunoreactivity for leptin using antibody A-20.
Staining was observed within the cytoplasm of fibroblast cells with
varying staining intensities (Fig. 3A).
These cell cultures were also positive for leptin secretion as detected
by RIA measurements of cell-conditioned medium. In addition,
full-length leptin receptors were detected by immunostaining in the
cytoplasma and on the cytoplasmic membrane of fibroblasts (n = 3)
using antibody C-20 (Fig. 3B). Both staining reactions were completely
blocked by a 10-fold excess of the corresponding blocking peptide,
confirming the specificity of the antibody reaction (Fig. 3, C and
D).
|
). Serial sections
stained with an antibody against leptin (A20, Fig. 3F) and leptin
receptor (C-20, Fig. 3G) confirmed leptin and leptin receptor
expression to Thy-1-positive cells, thus providing evidence that
fibroblasts express both proteins in vivo as well.
Interestingly, leptin was also found in epithelial cells of hair
follicles, as described in mice by others (26) (data not
shown). Similar staining results on fibroblast cultures and skin biopsies were obtained with two in-house antibodies against leptin receptor (Ob-R1984) and leptin (Ob1514). Control staining was performed on human sc fat tissue with positive reaction. No staining was observed in controls with primary antibodies replaced by nonspecific IgG (data not shown).
Basal leptin secretion and effect of insulin
Basal leptin secretion. All cell cultures (n = 13) analyzed by RT-PCR for Ob synthesis were also analyzed for leptin protein levels by RIA. RIA measurements in the cell-conditioned medium from nonstimulated cells were in good agreement with the PCR results. No leptin was detected in the five cultures without Ob mRNA expression. In all cultures with detectable Ob mRNA expression we also found leptin protein in the cell-conditioned medium. However, individual hormone levels varied from a minimum value of 0.07 ng/350,000 cells to a maximum of 1.4 ng/350,000 cells after 24 h. The basal secretion of leptin increased asymptotically over a 72-h period up to 258% of the 24-h values (n = 3; P < 0.001). Leptin levels did not demonstrate any correlation with age, sex, or BMI of the donors.
Effect of insulin on leptin secretion. Incubation of
fibroblast cell cultures with insulin (0.1–10 nmol/liter) for 24
h (n = 3) stimulated leptin synthesis in a dose-dependent manner.
Significant effects were achieved with insulin levels as low as 1
nmol/liter that increased the leptin concentration by 52 ± 28%
(P < 0.01) of nonstimulated basal levels. Maximal
stimulation of 92 ± 44.5% (P < 0.001) was found
after incubation with insulin (10 nmol/liter; Fig. 4A).
|
).
Cycloheximide, an inhibitor of ribosomal protein synthesis, as well as
actinomycin D, an inhibitor of the transcription, diminished the effect
of insulin on leptin secretion. Cycloheximide (10 µg/ml) decreased
basal leptin secretion over 24 h by 66 ± 21.6%
(P < 0.01; n = 3), whereas actinomycin D had no
significant effect on basal secretion. The stimulation of leptin
secretion with 10 nmol/liter insulin, however, was completely abolished
by 10 µg/ml cycloheximide as well as by 500 nmol/liter actinomycin D,
providing evidence that active RNA and protein synthesis are involved
in mediation of the insulin-induced increase in leptin concentration
(Fig. 4C). Control incubations with an ethanol concentration comparable
to that of cycloheximide-treated supernatants (containing 0.01%
ethanol) were performed and showed no effect on leptin secretion or
viability of the cells.
Cell proliferation
The incubation of fibroblast cell cultures (n = 3) with 0.1–10 nmol/liter insulin, 10 µg/ml cycloheximide, or 500 nmol/liter actinomycin D did not significantly change the basal extinction after 24 h (0.489 ± 0.034) and 48 h (0.614 ± 0.113) as determined by WST-1 assay. This provides evidence that these mediators did not affect cell proliferation in the investigated time interval. Furthermore, no acute effects on cell viability were found by incubation with the maximal insulin concentration (10 nmol/liter) for 3 h.
Discussion
Fibroblasts are well known to produce and secrete a variety of cytokines. Considering the close ontogenic relationship between fibroblasts and adipocytes, we hypothesized that leptin is produced by human skin fibroblasts, which had not been specifically investigated to date. A potential role for leptin in fibroblast function is supported by recent studies that detected short and soluble leptin receptors in the mouse embryo 3T3-L1 fibroblast-like cell line in which leptin was observed to affect the Na,K-pump activity via a wortmannin-sensitive signaling pathway (27). Ring et al. (28) reported that systemic or topic administration of leptin induced a highly significant acceleration of wound repair in the ob/ob mouse, but not in the db/db mouse, indicating that leptin’s effect was mediated by the leptin receptor. Finally, Bornstein et al. showed that the Ob-R was localized to fibroblast-like cells of immature human adipocyte cell cultures by immunohistochemistry (29).
In this study we demonstrate the expression of long and short leptin receptor isoforms and of leptin itself in mature human fibroblasts that is regulated by insulin in a time- and dose-dependent manner.
Fibroblasts and adipocytes have common progenitor cells. In adults, these seem to be microvascularly localized mesenchymal cells from which both cell lines differentiate (30). Also, differentiated fibroblasts are likely to maintain the ability to differentiate into preadipocytes and mature adipocytes (31, 32). Therefore, we characterized our fibroblast secondary cell cultures to confirm cell type homogeneity and activity to be specific to fibroblasts and that the effects observed were not confounded by adipocytes. Light microscopic evaluation and staining with Sudan III, a lipid-soluble color reagent that stains lipid droplets, excluded the presence of adipocytes or preadipocytes that otherwise may have accounted for the leptin synthesis observed. Furthermore, flow cytometry using a fluorescence-labeled antibody against Thy-1, a highly specific identification marker of human fibroblasts that does not recognize fat cells (22, 33, 34), revealed 99.8% Thy-1-positive cells in these cultures. Finally, immunohistochemical detection of prolyl-4-hydroxylase, a key enzyme in collagen formation of active fibroblasts (35), confirmed the maintenance of specific cell functions in the fibroblast cultures. Nearly all cells showed an intensive granular staining pattern, indicating the endoplasmic localization of this enzyme as described by others (36).
The full length Ob-R (mRNA and protein) and the three shorter membrane-spanning isoforms of the Ob-R (mRNA) were shown to be uniformly expressed by all human secondary fibroblast cell cultures investigated. This finding also applies to the situation in vivo, as demonstrated by immunohistochemical detection of the full-length leptin receptor in human skin biopsies. The expression of the functional Ob-R is thus likely to constitute the basis for local effects exerted by leptin on fibroblast function.
In addition, we provide evidence that fibroblasts produce leptin under cell culture conditions and also express leptin in vivo. This supports very recent findings in orbital fibroblast cultures from patients with Graves’ ophthalmopathy, which were shown to express Ob mRNA (31); however, in this study cultures in the second passage were used, not excluding the coexistence of various kinds of other cell types.
In our cell culture system, nonstimulated fibroblasts secreted leptin protein into supernatants in amounts reaching those observed in cultured human adipocytes (37). In contrast to white adipose tissue, however, leptin production by fibroblasts appears to be independent of sex, age, and BMI. However, the number of patients investigated was too small to allow definite conclusions about an association between these parameters. It is also not related to the passage of the culture. As we observed that a high cell density dramatically increased leptin synthesis of the cultures (data not shown), we speculated that leptin synthesis is increased in cells not engaged in propagation, as shown for other proteins, e.g. IGF-binding protein-3 (38) and prolyl-4-hydroxylase (39).
Plasma leptin and insulin concentrations are closely associated in humans (40, 41), in that leptin inhibits pancreatic insulin secretion (42), and insulin, in turn, stimulates adipocyte leptin synthesis (20). In accordance with the conditions observed in adipocytes (43), leptin synthesis in fibroblasts is regulated by insulin. This effect occurs at physiological to supraphysiological concentrations of insulin, with the first significant increases of leptin secretion after 14 h, indicating chronic, but not acute, effects, which may suggest an induction of Ob mRNA transcription and/or translation by insulin. Indeed, we showed that addition of the transcriptional and translational inhibitors actinomycin D and cycloheximide blocked the insulin-induced stimulation of leptin secretion.
Our findings correspond with studies on insulin-dependent leptin synthesis in rat (44) and porcine (45) adipocytes, which are also described to be cycloheximide sensitive. However, although insulin was shown to increase leptin mRNA in human adipocytes (46), contradictory effects of actinomycin D on insulin-stimulated leptin synthesis were reported in rat adipocytes, in which neither abolished the stimulatory effects of insulin (44, 47). The existence of preformed insulin-regulated pools as suggested by this study (44) is not supported by our findings in fibroblasts, as coincubation of insulin and cycloheximide did not change the leptin concentration compared with incubation with cycloheximide alone. It should be considered, however, that leptin was localized in vesicle-like structures in human adipocytes (29). In contrast to rodents, most of the in vivo studies in humans have also described increased leptin expression and secretion after chronic, but not after acute, hyperinsulinemia (48). In addition, in human adipocyte cultures, a long-term effect of insulin on leptin production was demonstrated, whereas insulin did not stimulate leptin production acutely (49). Hence, these findings indicate that the insulin-stimulated leptin response affects transcriptional and translational processes and involves de novo protein synthesis.
The specificity of our findings of insulin’s effect on leptin secretion as opposed to effects on fibroblast proliferation were confirmed by a WST test that did not reveal significant alterations in the cell number and viability of human fibroblast secondary cell cultures (up to 48 h). Even though insulin is known to exert mitogenic effects on fibroblast cultures (50, 51), this does not contrast with our results. Because the concentrations of insulin used in this study were lower, the medium was not supplemented with additional stimulatory factors, such as epidermal growth factor, dexamethasone, or FCS, and incubation times were relatively short. We also did not find any signs of hypertrophy (e.g. excess of synthesis of extracellular matrix proteins) or changes in cell size by light microscopic evaluation after incubation of the fibroblasts with insulin.
The physiological consequences of leptin synthesis in human fibroblasts are unclear. Fibroblast-derived leptin might be 1) of local importance, in that it exerts autocrine/paracrine mechanisms, and/or 2) contribute to the effects of the endocrine active leptin pool in the circulation.
This latter assumption might explain in part the wide interindividual variation in leptin levels at any given amount of body fat (40, 52). Furthermore, a disruption of the association between fat content or BMI and leptin levels has been observed in some physiological and pathophysiological conditions, such as aging in humans (53, 54) or during a long-term hypocaloric diet, in which leptin derived from sources other than white adipose tissue may be responsible for the dissociation of serum leptin concentration and body fat content (55).
The degree to which leptin produced by fibroblasts contributes to its level in the peripheral blood remains unclear. In patients with liver cirrhosis, a condition in which hepatic fibrosylation occurs, the rate of increase in circulating leptin levels in relation to increased BMI was higher than that in healthy control subjects of both genders (56).
In addition to possible endocrine effects, fibroblast-derived leptin may act locally in an autocrine or paracrine manner, as observed for other cytokines. The Ob-R is almost ubiquitously expressed in peripheral tissues, and fibroblasts express the functional form of the leptin receptor themselves, as shown in this study. Hence, local actions of leptin may well affect collagen synthesis and cytokine release, which, in turn, might contribute to improved wound repair through leptin (57, 58). This hypothesis is further supported by the finding that leptin influenced specific murine T lymphocyte responses (59), leading to an enhancement of the production of cytokines and the phagocytotic activity of murine peritoneal macrophages, and by the fact that the leptin receptor is expressed on granulocytes, monocytes, and mature peritoneal macrophages (15). Vice versa, inflammatory conditions with increased cytokine release were shown to be associated with increased plasma leptin levels in rats (60) and therefore might also influence leptin synthesis in fibroblasts.
In conclusion, we demonstrate that human fibroblasts express leptin in vitro and in situ. Fibroblast leptin synthesis is stimulated by insulin in vitro and might contribute to the total leptin plasma pool. In addition, leptin might exert local paracrine or autocrine actions mediated via its own receptor expressed in connective tissues.
Acknowledgments
We thank Elke Jeschke, Christina Liebig, and Ines Schindler for excellent technical assistance.
Footnotes
This work was supported by the Bundesministerium für Bildung und Forschung, Interdisciplinary Center for Clinical Research, University of Leipzig (01KS9504; Project B15) as well as a grant from the Eli Lilly & Co. International Foundation.
Abbreviations: BMI, Body mass index; Ob-R, leptin receptor; WST-1, 2-(4-Iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt.
Received November 15, 2000.
Accepted May 16, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  M Bokarewa, D Bokarew, O Hultgren, and A Tarkowski Leptin consumption in the inflamed joints of patients with rheumatoid arthritis Ann Rheum Dis, October 1, 2003; 62(10): 952 - 956. [Abstract] [Full Text]
|
|
|
|
 |
|
 T. Reimer, D. Koczan, B. Gerber, D. Richter, H.J. Thiesen, and K. Friese Microarray analysis of differentially expressed genes in placental tissue of pre-eclampsia: up-regulation of obesity-related genes Mol. Hum. Reprod., July 1, 2002; 8(7): 674 - 680. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Shintani, H. Ikegami, T. Fujisawa, Y. Kawaguchi, M. Ohishi, T. Katsuya, J. Higaki, K. Shimamoto, and T. Ogihara Leptin Gene Polymorphism Is Associated with Hypertension Independent of Obesity J. Clin. Endocrinol. Metab., June 1, 2002; 87(6): 2909 - 2912. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
