Evidence for a key role of leptin in osteoarthritis




To evaluate the contribution of leptin (an adipose tissue–derived hormone) to the pathophysiology of osteoarthritis (OA), by determining the level of leptin in both synovial fluid (SF) and cartilage specimens obtained from human joints. We also investigated the effect of leptin on cartilage, using intraarticular injections of leptin in rats.


Leptin levels in SF samples obtained from OA patients undergoing either knee replacement surgery or knee arthroscopy were measured by enzyme-linked immunosorbent assay. In addition, histologic sections of articular cartilage and osteophytes obtained during surgery for total knee replacement were graded using the Mankin score, and were immunostained using antibodies to leptin, transforming growth factor β (TGFβ), and insulin-like growth factor 1 (IGF-1). For experimental studies, various doses of leptin (10, 30, 100, and 300 μg) were injected into the knee joints of rats. Tibial plateaus were collected and processed for proteoglycan synthesis by radiolabeled sulfate incorporation, and for expression of leptin, its receptor (Ob-Rb), and growth factors by reverse transcriptase–polymerase chain reaction and immunohistochemical analysis.


Leptin was observed in SF obtained from human OA–affected joints, and leptin concentrations correlated with the body mass index. Marked expression of the protein was observed in OA cartilage and in osteophytes, while in normal cartilage, few chondrocytes produced leptin. Furthermore, the pattern and level of leptin expression were related to the grade of cartilage destruction and paralleled those of growth factors (IGF-1 and TGFβ1). Animal studies showed that leptin strongly stimulated anabolic functions of chondrocytes and induced the synthesis of IGF-1 and TGFβ1 in cartilage at both the messenger RNA and the protein levels.


These findings suggest a new peripheral function of leptin as a key regulator of chondrocyte metabolism, and indicate that leptin may play an important role in the pathophysiology of OA.

Leptin, a small (16 kd) polypeptide encoded by the obese (ob) gene, is produced predominantly in white adipose tissue and regulates food intake and energy expenditure at the hypothalamic level (1–5). Leptin-deficient mice (ob/ob) fail to produce the functional protein and exhibit severe obesity (6). This hormone achieves its biologic effects by interacting with specific receptors (Ob-R) that result from the alternative splicing of the db gene transcript (7). Only Ob-Rb (the long isoform of the leptin receptor) contains intracellular motifs required for the signal transduction pathway through activation of signal transducers and activators of transcription (8, 9). Because both leptin and its receptor share common structural and functional properties with the family of interleukin-6 (IL-6) cytokines (10, 11), this adipose-derived protein has been classified as an adipocytokine.

Because of the wide pattern of Ob-R expression in peripheral tissues (12), leptin may be considered a pleiotropic hormone involved in the control of various physiologic processes, such as lipid homeostasis (13), insulin secretion (3), reproductive functions (14), thermogenesis (3, 15), or angiogenesis (16). In addition, this adipocytokine plays an essential role in immune functions (17). The dysregulation of cytokine production in leptin-deficient mice leads to an increased susceptibility to infection and inflammatory stimuli (18, 19). Recently, leptin was described as a new regulator of bone growth, acting either through a neuronal network to release an unidentified antiosteogenic factor (20, 21) or directly by inducing osteoblast proliferation, collagen synthesis, and bone mineralization (22, 23). Leptin is also a skeletal growth factor that stimulates endochondral ossification (24). Because functional leptin receptor has been shown to be expressed in normal human cartilage (25), it can be speculated that leptin could also act on this articular connective tissue, especially during osteoarthritis (OA), a disease in which great metabolic changes of chondrocytes are observed.

OA is a degenerative joint disorder and represents one of the most frequent and disabling diseases encountered in elderly individuals. In the early stages of the disease, the cartilage surface undergoes fibrillations before full-thickness matrix loss occurs. OA changes are not limited to cartilage, because remodeling of the underlying bone and development of osteophytes are also observed in osteoarthritic joints. Several inflammatory components such as tumor necrosis factor α and IL-1 have been detected in the synovial fluid (SF) of patients with OA and have been implicated in the degenerative process by inhibiting extracellular matrix synthesis and increasing catabolic activities of metalloproteinases (26–29). In response to cartilage damage, various growth factors, including insulin-like growth factor 1 (IGF-1) or transforming growth factor β (TGFβ), are activated and stimulate chondrocytes to repair the damaged extracellular matrix by forming cell clusters and increasing their anabolic activity (30–33). In fact, TGFβ has a dual role in OA, having a beneficial effect on cartilage repair but inducing osteophyte formation (34, 35). Despite the increased synthetic response induced by growth factors, the enhanced synthesis of degrading enzymes ultimately results in cartilage erosion. Although the role of these factors in OA has been clearly demonstrated, the pathologic processes involved in this degenerative articular disease remain unclear.

The potential role of leptin in OA is supported by the relationship between high body mass index (BMI) and an increased risk of OA (36). This positive association is observed not only for knee joints but also for non–weight-bearing joints such as the hands (37, 38), suggesting that OA may be initiated by metabolic disorders, with progression being worsened by high mechanical stress on abnormal cartilage. In order to better understand the molecular mechanisms involved in OA, the present study was undertaken to evaluate the contribution of leptin to cartilage changes associated with this joint disease.

This study is the first to show that leptin was detected in SF obtained from patients with OA, and that levels of leptin correlated with the BMI. Moreover, leptin was strongly overexpressed in human OA cartilage and in osteophytes. Human OA chondrocytes also produced growth factors, the topographic localization and staining intensity of which varied according to the histologic grade of cartilage destruction and paralleled those of leptin. Results of animal studies indicated that this adipocytokine stimulated anabolic functions of chondrocytes and induced the synthesis of growth factors in cartilage. Taken together, results of the current study provide evidence for a key role of leptin in the pathophysiology of OA.


Patients and samples.

Specimens of hyaline cartilage and osteophytes were obtained from 11 patients (8 women and 3 men, ages 56–80 years [mean age 69.3 years]) who had OA according to the American College of Rheumatology criteria (39). Tibial plateaus and osteophytes from the condyles were collected during total knee replacement surgery. For comparison with diseased tissues, normal knee joint cartilage was obtained from 2 transplant donors (a 56-year-old man and a 41-year-old woman) through an agreement with the Etablissement Français des Greffes.

SF samples were obtained from OA patients who either were undergoing knee replacement surgery (n = 11; 4 women and 7 men, ages 56–80 years [mean age 70.5 years]) or required knee arthroscopy (n = 9; 2 women and 7 men, ages 45–72 years [mean age 62.1 years]). This human study was conducted in conformity with the declaration of Helsinki principles, and written informed consent was obtained from all participants.

Leptin measurement in SF.

To determine whether leptin is present within the joint, SF samples were analyzed for leptin, using specific enzyme-linked immunosorbent assays (R&D Systems, Abingdon, UK) according to the manufacturer's instructions. Leptin concentrations were determined in duplicate, after 10-fold dilution with a solution supplied by the manufacturer.

Histologic assessment.

Tibial plateaus and osteophytes were fixed for 24 hours in 4% paraformaldehyde immediately after removal, decalcified in rapid bone decalcifier (RDO; Apex Engineering Products, Plainfield, NJ) for 8 hours, and further fixed in 4% paraformaldehyde. For each lateral and medial plateau, 5 full-depth cartilage biopsy specimens were sampled using a biopsy punch (6 mm diameter) according to a standardized diagram. Cartilage biopsy specimens and osteophytes were then dehydrated in a graded series of alcohol and embedded in paraffin.

For all cartilage samples, hematoxylin–eosin–saffron and toluidine blue stainings were performed on 5-μm serial sections to determine histologic grading. For each biopsy, the severity of OA cartilage lesions was evaluated by 2 independent observers and was graded using the Mankin score (40). Forty-nine specimens representative of all grades were selected for the immunohistochemical study.

Animal studies.

Male Wistar rats (150–175 gm) were obtained from Iffa Credo (L'Arbresle, France). The animals were housed under controlled temperature and lighting conditions, and received food and water ad libitum. Animals were acclimatized to the laboratory environment for 1 week before the experiment. The local Animal Care and Use Committee approved the experimental protocols, and guidelines for laboratory procedures were followed at all times.

In order to investigate the in vivo effects of leptin on articular cartilage, rat recombinant leptin (endotoxin <0.01%) (R&D Systems), in a volume of 50 μl sterile saline, was injected into the joint cavity of the right knee of rats. Rats injected with saline served as controls. The tibial plateaus from rats killed 24 hours after intraarticular administration of leptin (30 and 100 μg) were collected for reverse transcriptase–polymerase chain reaction (RT-PCR) analysis. The tissue samples were decalcified overnight in 150 mM EDTA at 4°C, and cartilage was isolated from the underlying bone before total RNA extraction. Tibial plateaus, which were removed 48 hours following leptin injection (10, 30, 100, and 300 μg), were processed for proteoglycan synthesis determination. For immunohistochemical analysis, total knee joints were dissected 48 hours after leptin administration (100 μg), fixed in 4% paraformaldehyde for 24 hours, decalcified in RDO for 4 hours, further fixed in paraformaldehyde for 24 hours, and then embedded in paraffin.

Immunohistochemical analysis.

Paraffin sections (5 μm) from human and rat cartilage specimens were deparaffinized in Tissue Clear (Bayer Diagnostics, Puteaux, France) and rehydrated in a graded series of ethanol. After blocking nonspecific sites for at least 1.5 hours with phosphate buffered saline (PBS) containing bovine serum albumin (BSA, 4% weight/volume), sections were incubated overnight at 4°C with the different primary polyclonal antibodies diluted in BSA. For human cartilage tissues, antibodies for leptin and TGFβ1 (sc-842 and sc-146, respectively; Santa Cruz Biotechnology, Le Perray en Yvelines, France) were both used at a final concentration of 4 μg/ml, and antibody for IGF-1 (sc-7144; Santa Cruz Biotechnology) was used at a final concentration of 7 μg/ml. For rat specimens, antibodies for leptin, TGFβ1, and IGF-1 were used at final concentrations of 1 μg/ml, 0.7 μg/ml, and 2 μg/ml, respectively. Tissues obtained from normal rats were also analyzed for the expression of the long isoform of the leptin receptor (Ob-Rb) in articular cartilage, by using a specific antibody (sc-1834; Santa Cruz Biotechnology) diluted at 7 μg/ml.

After washing twice in PBS, biotinylated secondary antibodies (1 μg/ml) (goat anti-rabbit IgG [Vectastain ABC kit; Novocastra, Le Perray en Yvelines, France] and bovine anti-goat IgG [sc-2347; Santa Cruz Biotechnology]) were applied for 45 minutes at room temperature. Slices were then treated with hydrogen peroxide (0.3%) to quench endogenous peroxidase. The signal was amplified with preformed avidin–biotinylated horseradish peroxidase complexes for 45 minutes at room temperature (Vectastain ABC kit; Novocastra), and staining was developed with 3,3′-diaminobenzidine (0.05% in hydrogen peroxide) (Liquid DAB Substrate Kit; Novocastra). Counterstaining of nuclei was performed with hematoxylin, and slices were mounted in Eukitt (CML, Nemours, France). Sections incubated without primary antibody served as controls.

The entire surface of human cartilage specimens was examined by 2 independent observers. The percentage of positively stained cells was rated on a scale of 0 (negative) to +++, with + representing 5–20%, ++ representing 21–50%, and +++ representing >50% staining.

Polymerase chain reaction.

For the semiquantitative RT-PCR, total RNA was isolated from cartilage using a commercially available phenol–chloroform solution (TRIzol). Following precipitation with isopropanol, one-fifth of the RNA samples were reverse transcribed to complementary DNA (cDNA) using oligo(dT) primers (100 pmoles) and Moloney murine leukemia virus reverse transcriptase (200 units) (Gibco BRL, Cergy-Pontoise, France). PCR amplification was then performed on one-tenth of the RT products with Taq polymerase (2.5 units) (Gibco BRL). PCR conditions were 94°C for 45 seconds, 55°C for 45 seconds, and 72°C for 45 seconds, unless otherwise indicated. As an internal control, the same cDNA was subjected to PCR analysis for the L27 gene (melting temperature = 61°C), encoding for a ribosomal protein.

PCR primers were selected to amplify a 373-bp fragment for leptin (forward 5′-CCT-CAT-CAA-GAC-CAT-TGT-CAC-C-3′; reverse 5′-AGA-ATG-TCC-TGC-AGA-GAG-CCC-3′), a 453-bp fragment for TGFβ (forward 5′-CTA-CGC-CAA-AGA-AGT-CAC-CC-3′; reverse 5′-GTT-CAT-GTC-ATG-GAT-GGT-GC-3′), a 356-bp fragment for IGF-1 (forward 5′-TCA-CAT-CTC-TTC-TAC-CTG-GC-3′; reverse 5′-TCC-TGC-ACT-TCC-TCT-ACT-TG-3′), and a 225-bp fragment for L27 (forward 5′-TCC-TGG-CTG-GAC-GCT-ACT-C-3′; reverse 5′-CCA-CAG-AGT-ACC-TTG-TGG-GC-3′) (MWG Biotech, Ebersberg, Germany). The PCR products were separated on 1.2% agarose gels, and bands were semiquantitatively analyzed with a GelDoc digital imaging system (Bio-Rad, Marne la Coquette, France).

Assessment of proteoglycan synthesis in rat cartilage.

Tibial plateaus were dissected with a minimal amount of adjacent tissue and incubated for 3 hours at 37°C in RPMI-HEPES 1640 medium supplemented with 2 mML-glutamine, 100 μg/ml streptomycin, 100 IU/ml penicillin, and 0.6 μCi/ml of Na235SO4. After 5 washing steps in saline, tibial plateaus were fixed overnight at room temperature in 0.5% cetylpyridinium chloride dissolved in 10% formalin buffer. Plateaus were then decalcified for 6 hours in 5% formic acid, and 4 specimens were obtained from each lateral and medial plateau. Each specimen was dissolved in 0.5 ml Soluene overnight at 42°C. The 35S content was measured by liquid scintillation counting.

Statistical analysis.

For human studies, the correlation between SF leptin levels and BMI was determined by linear regression analysis, using StatView for Windows, version 5.0 (SAS, Cary, NC). For experimental studies, each value is the mean ± SEM of 4 samples for RT-PCR analysis, and of 6 samples for the determination of proteoglycan synthesis analyzed from 3 independent experiments. The different groups were compared using a 2-way analysis of variance test. P values less than 0.05 were considered significant.


Correlation between SF leptin levels and BMI.

Synovial fluid samples were obtained from 11 patients undergoing knee replacement surgery and from 9 pa- tients requiring knee arthroscopy. In the 14 male patients, the BMI ranged from 22 kg/m2 to 34.2 kg/m2 (mean 28.15 ± 4.26 kg/m2), and in the 6 female patients, the range was 22 kg/m2 to 31.5 kg/m2 (mean 26.04 ± 4.03 kg/m2). This study is the first to show that leptin was detected in SF obtained from OA patients, at concentrations of 0.60–17.40 μg/liter (mean 8.16 ± 5.50 μg/liter) in men and 5.38–28.50 μg/liter (mean 12.95 ± 8.92 μg/liter) in women. Leptin levels in SF were significantly correlated with BMI (r = 0.572, P < 0.01) (Figure 1).

Figure 1.

Correlation between body mass index (BMI) and the concentration of leptin in synovial fluid, in 20 patients with osteoarthritis.

All OA grades exhibited in cartilage biopsy specimens.

Normal cartilage slices showed very few surface irregularities and intense toluidine blue staining in all areas (score = 0). For each lateral and medial part of the tibial plateaus derived from OA patients, 5 cartilage biopsy specimens were removed according to a standardized scheme, except for regions with subchondral bone exposure. Cartilage samples exhibited variability in their histologic appearance, ranging from surface irregularities to marked fibrillations, cluster formation, and clefts. Proteoglycan loss was evident histologically in samples of moderately damaged OA cartilage, but the overall thickness of the cartilage was preserved. In contrast, specimens showing advanced cartilage destruction exhibited severe proteoglycan loss and cartilage erosion. Overall, 49 biopsy specimens obtained from 11 knees were selected and divided into 3 groups for further immunohistochemical analysis, as follows: Mankin scores of 1–4, 5–9, and 10–14 were considered to represent mild, moderate, and severe OA, respectively (Table 1).

Table 1. Frequency of cells positive for leptin, TGFβ1, and IGF-1 in human OA cartilage specimens of various histologic grades*
OA groupLeptinTGFβ1IGF-1
  • *

    Values are the number (percentage) of cartilage biopsy specimens in each group stained for leptin, transforming growth factor β1 (TGFβ1), or insulin-like growth factor 1 (IGF-1). Histologic grades of cartilage destruction were determined using the Mankin score after examining cartilage sections stained with hematoxylin–eosin–saffron and toluidine blue. Scores of 1–4, 5–9, and 10–14 were considered mild, moderate, and severe osteoarthritis (OA), respectively. The percentage of positive chondrocytes throughout the entire biopsy specimen is rated as follows: + = 5–20%, ++ = 20–50%, and +++ = >50%. In normal samples, <5% of total cells exhibited positive staining for leptin, and no expression of growth factors was detected.

Mild (n = 13)10 (77)03 (23)11 (85)2 (15)012 (92)01 (8)
Moderate (n = 26)15 (58)5 (19)6 (23)17 (65)4 (16)5 (19)20 (77)4 (15)2 (8)
Severe (n = 10)3 (30)2 (20)5 (50)3 (30)3 (30)4 (40)3 (30)2 (20)5 (50)

Relationship between leptin immunostaining in human articular cartilage and OA grade.

In biopsy specimens of normal cartilage, few positive chondrocytes were detected in the middle cartilage layer (Figure 2). Our immunohistochemical experiments indicated that leptin expression was up-regulated in human OA chondrocytes. No background staining was observed in negative control studies when primary antibody was omitted (Figure 2). Leptin immunostaining was scored according to the percentage of positive chondrocytes over the entire biopsy specimen (+, ++, +++). The frequency of cells positive for leptin in each specimen was related to the histologic OA score (Table 1). In fact, 77% of cartilage biopsy specimens showing mild OA had low leptin staining, and 50% of cartilage specimens showing severe OA exhibited strong staining.

Figure 2.

Expression of leptin in representative serial sections of human cartilage. Control staining was performed without primary antibody. Positive cells are brown; nuclei are counterstained with hematoxylin. OA = osteoarthritis (original magnification × 40; × 160 in insets).

In addition, the distribution of leptin was associated with the severity of the lesions. Leptin immunostaining was detected in the middle zone of cartilage samples showing mild OA. In cartilage specimens showing moderate OA, superficial cells and more cells of the middle layer were positive (Figure 2). Leptin expression was most prominent in the superficial and middle layers of severely damaged cartilage and was especially associated with areas of matrix depletion, fibrillations, and chondrocyte clusters. In contrast, leptin immunostaining was rarely detected in the deep layer of cartilage, except in a few chondrocyte clusters. Finally, osteophytes obtained from condyles exhibited strong leptin expression throughout the entire thickness of the slice (Figure 3). Increased immunoreactivity was observed in the transition area between the fibrous tissue and the underlying cartilage.

Figure 3.

Expression of leptin, transforming growth factor β1 (TGFβ1), and insulin-like growth factor 1 (IGF-1) in representative serial sections of osteophytes. Control staining was performed without primary antibody. Positive cells are brown; nuclei are counterstained with hematoxylin. Arrowheads indicate strong staining in the transitional layer between cartilage and fibrous tissues (original magnification × 40).

Distribution pattern of leptin, TGFβ1, and IGF-1 in human tibial cartilage.

Biopsy specimens of normal cartilage showed no staining for TGFβ1 or IGF-1 in any cartilage layer. In contrast, growth factors were expressed in OA cartilage (Figure 4). Interestingly, the topographic distribution and staining intensity for these growth factors varied according to the grade of cartilage destruction and paralleled those of leptin. As shown in Table 1, 85% and 92% of cartilage biopsy specimens showing mild OA exhibited weak expression of TGFβ1 and IGF-1, respectively. Conversely, 40% and 50% of specimens showing severely damaged cartilage had high levels of TGFβ1 and IGF-1, respectively (Table 1). Furthermore, as was observed for leptin, strong TGFβ1 expression was demonstrated in osteophytes obtained from patients with OA (Figure 3). In contrast, expression of IGF-1 was barely detectable in these marginal osteophytes.

Figure 4.

Expression of TGFβ1 and IGF-1 in representative serial sections of human cartilage. Control staining was performed without primary antibody. Positive cells are brown; nuclei are counterstained with hematoxylin. OA = osteoarthritis (see Figure 3 for other definitions). (Original magnification × 40; × 160 in insets.)

Leptin stimulation of proteoglycan synthesis in rat cartilage.

We investigated the effects of leptin on proteoglycan anabolism by measuring 35S incorporation in 4 areas of the tibial plateaus (medial and lateral) after intraarticular injection of leptin (10, 30, 100, and 300 μg/50 μl) into the right knee joint (Figure 5A). After a 30-μg injection, a marked increase in proteoglycan synthesis was observed in all cartilage areas of the plateau (Figure 5B). This stimulatory effect of leptin decreased at the highest dose (300 μg). Variation in this anabolic response was observed in different sites of the plateaus, as follows: the increase in proteoglycan synthesis following a 30-μg dose of leptin was significant in the medial plateau only (+65% in area 3 [P = 0.043] and +94% in area 4 [P = 0.015]), whereas the lateral plateau responded to higher doses of leptin (+76% in area 1 [P = 0.015] and +131% in area 2 [P = 0.003] after a 100-μg dose; and +65% in area 1 [P = 0.036] and +87% in area 2 [P = 0.034] after a 300-μg dose).

Figure 5.

Effect of leptin on proteoglycan synthesis in rat cartilage. A, Localization of biopsy specimen punched out from the tibial plateau. B, Proteoglycan synthesis in cartilage specimens was measured by Na235SO4 incorporation, 48 hours following intraarticular injection of 10, 30, 100, or 300 μg of rat recombinant leptin. Results (mean and SEM) are expressed as the percentage of 35S incorporation in leptin-injected rats compared with that in saline-injected rats, and represent 1 of 3 experiments (n = 6 per group). ∗ = P < 0.05; ∗∗ = P < 0.01.

Leptin-induced expression of TGFβ1, IGF-1, and leptin messenger RNA (mRNA) in rat cartilage.

RT-PCR analysis was performed to evaluate the effects of intraarticular injections of leptin (30 μg and 100 μg) on the expression of IGF-1, TGFβ1, and leptin mRNA in rat cartilage plateaus. The results showed that the low dose of leptin induced leptin mRNA expression in rat tibial plateaus (Figure 6). Regardless of which dose was administered, leptin up-regulated the expression of TGFβ1 and IGF-1 transcripts in tibial cartilage (Figure 6). The stimulating effect of leptin on growth factor expression was dose-related, with a significant effect at the highest dose used in this experiment (100 μg) (P < 0.05).

Figure 6.

Effect of leptin on mRNA expression for leptin, IGF-1 and TGFβ in rat cartilage. A, Total RNA was isolated from cartilage of tibial plateaus collected 24 hours following intraarticular injection of 30 μg or 100 μg of leptin into the knee joint, and was subjected to reverse transcriptase–polymerase chain reaction (PCR) analysis using specific primers for each gene. PCR amplification of L27 was performed on the same RNA preparation, as an internal control for the relative amounts of template. B, Semiquantitative analysis of expression levels. Results (mean and SEM) are expressed as the percentage of gene expression in leptin-injected rats versus that in saline-injected rats. Assays were performed in triplicate, using at least 4 samples at each time point. ∗ = P < 0.05. See Figure 3 for other definitions.

Leptin-induced stimulation of leptin, IGF-1, and TGFβ1 protein expression in rat cartilage.

This is the first study to demonstrate expression of leptin and its receptor (Ob-Rb) in rat cartilage (Figure 7). Injection of 100 μg of leptin into the knee joint induced overexpression of leptin in each cartilage layer (Figure 7). Growth factor expression in tibial-plate cartilage was also up-regulated after administration of 100 μg of leptin. Negative control studies without primary antibody demonstrated no staining for any of the proteins studied.

Figure 7.

Immunohistochemical analysis of the expression of leptin and its receptor (Ob-Rb) in rat cartilage (top), and effect of leptin on leptin, TGFβ1, and IGF-1 synthesis in representative sections of tibial plateaus obtained from rats killed 48 hours following intraarticular injection of 100 μg of rat recombinant leptin (bottom). Sections from untreated rats and from leptin-injected rats were stained for leptin, TGFβ1, and IGF-1. Control staining was performed without primary antibody. Positive cells are brown; nuclei are counterstained with hematoxylin. See Figure 3 for definitions (original magnification × 40).


The present study was designed to investigate the contribution of leptin to cartilage changes in human OA. This is the first study to demonstrate the presence of leptin in SF obtained from patients with OA. In addition, this adipose-derived cytokine was expressed in human chondrocytes, and its synthesis was up-regulated in OA, in both cartilage and osteophytes. Interestingly, the pattern of leptin expression was related to the grade of cartilage destruction and was shown to be similar to that of growth factors. The stimulatory effect of leptin on both proteoglycan synthesis and growth factor expression in rat cartilage suggested a new peripheral function of this adipocytokine as a key regulator of chondrocyte metabolism. Taken together, these findings indicate that leptin may play an important role in the pathogenesis of OA by modulating chondrocyte functions and by contributing to osteophyte formation.

The mean levels of leptin in SF obtained from OA patients were in the same range as those previously reported in serum, and the leptin level in women was higher than that in men (41). The significant correlation between BMI and the concentration of leptin in SF suggests that this circulating adipocytokine with low molecular weight reaches the joint by diffusion through the synovial membrane.

The expression of leptin in cartilage obtained from OA patients was examined and compared with that observed in normal tissue. In order to take into account the focal character of the degenerative disease, cartilage specimens were removed from various areas of the tibial plateaus and further analyzed for histologic grading as well as immunostaining for leptin. In normal cartilage, few chondrocytes in the middle zone exhibited positive staining for leptin. This immunoreactive staining strongly increased in OA cartilage, especially in specimens showing severe damage, in which leptin was also observed in chondrocyte clusters and in cells located in the superficial zone. Data obtained from animal studies showed that rat chondrocytes also produced leptin and expressed the long form of the leptin receptor. Leptin synthesis was up-regulated at both the mRNA and protein levels when rats received injections of leptin into the joint. Because leptin was detected in SF from patients with OA, it can be speculated that the protein may penetrate into cartilage and may target the recently identified leptin receptors on human chondrocytes (25).

Taken together, these findings suggest a direct stimulatory effect of this adipocytokine on its own synthesis by chondrocytes. However, adjacent cartilage specimens with various levels of OA severity exhibited different staining patterns for leptin, in terms of both topographic distribution and intensity. This focal response may be explained either by different levels of leptin receptor expression depending on the disease state of cartilage samples or by better access to chondrocytes of leptin derived from the SF due to fissuring or damage in the extracellular matrix. The strong staining observed around clefts in the damaged cartilage further supports the latter hypothesis. Finally, chondrocytes themselves may act through a paracrine and/or autocrine pathway to induce leptin synthesis.

Results of the present study indicate that in specimens of human cartilage showing moderate OA, the positive immunoreactive staining for leptin was located mainly in chondrocytes in the middle layer. Interestingly, OA chondrocytes in this zone were shown to be hyperactive toward matrix synthesis, while cells in the upper zone down-regulated their expression of matrix components (42). The stimulatory effect of leptin on proteoglycan synthesis observed in rats suggests that this protein may be involved in the up-regulated anabolic activity observed in OA chondrocytes. This contribution may also occur in the early stages of the disease, which are characterized by enhanced chondrocyte synthetic activity (43). This observed stimulatory effect of leptin on cartilage anabolism is consistent with recent studies in which leptin was shown to promote bone growth by targeting osteoblasts directly and therefore has been considered to be a new hormonal regulator of bone growth (23, 44). The similar anabolic response of osteoblasts and chondrocytes to leptin is not very surprising, because both cell types are derived from mesenchymal precursors and may share common regulatory mechanisms to produce their respective extracellular matrix.

Human OA chondrocytes were shown to produce both leptin and growth factors, in the same topographic distribution and to the same extent, depending on lesion severity. To determine whether an increase in leptin synthesis in pathologic tissues may be responsible for high levels of growth factors, rats received injections of the recombinant protein into the knee joint. Intraarticular administration of leptin induced expression of IGF-1 and TGFβ in rat chondrocytes at both the mRNA and protein levels. This stimulating effect may explain the enhanced proteoglycan synthesis observed in leptin-injected rats, which may be further increased by the up-regulation of growth factor receptors, as was previously shown for IGF-1 receptor in leptin-stimulated mandibular condyles (45).

All of these findings support the concept that leptin may have a role in OA. High levels of IGF-1 and TGFβ were indeed measured in the SF from patients with OA (46, 47), and synthesis of these growth factors was increased in human OA cartilage (48–50) as well as in animal models of OA (33, 51). These growth factors are believed to have a beneficial function in the cartilage repair process that may occur during OA (51–53). However, besides their protective role against cartilage damage, they may also trigger degeneration of this connective tissue. Excessive and/or prolonged exposure to TGFβ leads to development of lesions similar to those observed in mice with spontaneous OA (43, 54). Because of the dual effect of growth factors (especially TFGβ) on cartilage, it would be interesting to identify any agent that is able to modulate their production. Although several compounds, including IL-1 or prostaglandin E2, have been shown to promote the release of growth factors (55, 56), no locally produced factors that might stimulate their synthesis in chondrocytes have been characterized. Because leptin induced IGF-1 and TGFβ expression at both the mRNA and protein levels, this study demonstrated, for the first time, that this adipocytokine is a key regulator of growth factors synthesis, and thus implied that leptin may contribute to the pathologic process of OA.

In addition to their effects on chondrocyte functions, growth factors are also involved in osteophyte formation, which is a prominent feature of OA-affected joints (51). TGFβ and IGF-1 have been detected in osteophytes (57–59), and repeated injections or overexpression of TGFβ in mice knee joints led to osteophyte generation (43, 60). In the current study, leptin (as well as TGFβ) was strongly produced in osteophytes from OA patients, which is similar to what was observed in cartilage. This adipose-derived protein may contribute to osteophyte formation either indirectly by stimulating TGFβ expression or directly by inducing endochondral ossification (24). In addition, leptin was strongly expressed within the fibrous mesenchymal tissue in the upper zones of osteophytes, where a sequential process of differentiation of pluripotent cells may result in the formation of new cartilaginous outgrowths that ultimately may ossify.

After examining various osteophytes, Aigner et al suggested that different cell and tissue types in osteophytes were similar to those observed in normally developing fetal epiphyses (61). Interestingly, several studies ascribed a role for leptin in the early stages of fetal bone and/or cartilage development (62, 63). The high levels of leptin detected in osteophytes indicated that this adipocytokine may contribute to the formation of these new bone structures by promoting fetal-like skeletal development processes, with synthesis of new extracellular matrix.

In conclusion, our data demonstrate that leptin is overexpressed in the human OA knee joint and provide evidence that leptin contributes to the pathogenesis of OA through stimulation of growth factor synthesis. Therefore, this adipocytokine may have dual effects on the joint: variations in the levels of systemic and/or locally produced leptin might regulate chondrocyte proliferation and anabolic function, and these variations might induce osteophyte formation during OA. Results of the current study further support the hypothesis that OA is a systemic disorder in which disturbances of lipid homeostasis might be involved (64). Moreover, the effects of leptin on articular cartilage may explain the relationship between BMI and the increased risk of OA. Further investigation, throughout the course of OA, is required to characterize the mechanisms implicated in the regulation of leptin expression in articular cartilage. In addition, the use of leptin-deficient mice would be helpful in defining the overall contribution of leptin to the pathophysiology of OA.


The authors thank Michel Thiery for taking good care of the animals, and the Etablissement Français des Greffes for providing normal tissue from transplant donors.