Article Text

Extended report
Leptin produced by joint white adipose tissue induces cartilage degradation via upregulation and activation of matrix metalloproteinases
  1. Wang Hui1,
  2. Gary J Litherland1,
  3. Martina S Elias2,
  4. Gareth I Kitson1,
  5. Tim E Cawston1,
  6. Andrew D Rowan1,
  7. David A Young1
  1. 1Musculoskeletal Research Group, Institute of Cellular Medicine, Medical School, Newcastle University, Newcastle-upon-Tyne, UK
  2. 2Department of Dermatology, Institute of Cellular Medicine, Newcastle University, Newcastle-upon-Tyne, UK
  1. Correspondence to David A Young, Musculoskeletal Research Group, Institute of Cellular Medicine, 4th Floor Cookson Building, Medical School, Newcastle University, Newcastle upon Tyne NE2 4HH, UK; d.a.young{at}


Objectives To investigate the effect of leptin on cartilage destruction.

Methods Collagen release was assessed in bovine cartilage explant cultures, while collagenolytic and gelatinolytic activities in culture supernatants were determined by bioassay and gelatin zymography. The expression of matrix metalloproteinases (MMP) was analysed by real-time RT–PCR. Signalling pathway activation was studied by immunoblotting. Leptin levels in cultured osteoarthritic joint infrapatellar fat pad or peri-enthesal deposit supernatants were measured by immunoassay.

Results Leptin, either alone or in synergy with IL-1, significantly induced collagen release from bovine cartilage by upregulating collagenolytic and gelatinolytic activity. In chondrocytes, leptin induced MMP1 and MMP13 expression with a concomitant activation of STAT1, STAT3, STAT5, MAPK (JNK, Erk, p38), Akt and NF-κB signalling pathways. Selective inhibitor blockade of PI3K, p38, Erk and Akt pathways significantly reduced MMP1 and MMP13 expression in chondrocytes, and reduced cartilage collagen release induced by leptin or leptin plus IL-1. JNK inhibition had no effect on leptin-induced MMP13 expression or leptin plus IL-1-induced cartilage collagen release. Conditioned media from cultured white adipose tissue (WAT) from osteoarthritis knee joint fat pads contained leptin, induced cartilage collagen release and increased MMP1 and MMP13 expression in chondrocytes; the latter being partly blocked with an anti-leptin antibody.

Conclusions Leptin acts as a pro-inflammatory adipokine with a catabolic role on cartilage metabolism via the upregulation of proteolytic enzymes and acts synergistically with other pro-inflammatory stimuli. This suggests that the infrapatellar fat pad and other WAT in arthritic joints are local producers of leptin, which may contribute to the inflammatory and degenerative processes in cartilage catabolism, providing a mechanistic link between obesity and osteoarthritis.

Statistics from

Osteoarthritis is a complex disease with genetic, mechanical and environmental components leading to a number of changes within the joint but typified by the irreversible destruction of articular cartilage.1 Articular cartilage is composed predominantly of an extracellular matrix (ECM) containing proteoglycan (aggrecan) and collagen (mainly type II collagen). Within this ECM chondrocytes control the turnover and remodelling of the cartilage matrix, through regulation of the expression of matrix components and matrix-degrading enzymes.2 Type II collagen is cleaved by matrix metalloproteinases (MMP), including MMP1, MMP8, MMP13 and MMP14, while aggrecan is cleaved by ‘a disintegrin and metalloproteinase with thrombospondin motifs’ (ADAMTS) enzymes, most probably ADAMTS5.3 Together, these proteases release specific type II collagen or aggrecan fragments that can be measured in vitro and in vivo.4 Whereas the loss of aggrecan from cartilage is reversible, collagen breakdown is not and therefore represents an irreversible step in cartilage degradation.5 6

A variety of cytokines and growth factors, such as the pro-inflammatory cytokines tumour necrosis factor alpha (TNFα) and interleukin (IL) 1 can induce the expression of many of the metalloproteinases described. Furthermore, when in combination with oncostatin-M, an IL-6-family cytokine, these pro-inflammatory cytokines can synergistically enhance the production of catabolic MMP.7,,9 These cytokines are key mediators of inflammation in rheumatoid arthritis but are increasingly being recognised as important in osteoarthritis pathogenesis.10

One of the most significant risk factors for osteoarthritis development is obesity.11 12 The effects of obesity on osteoarthritis can only partly be explained by an increase in biomechanical loading, because body mass index (BMI), or more specifically adiposity, is also a risk factor for osteoarthritis in non-weightbearing joints, such as hands.11 13 These data suggest that an adipose-derived inflammatory factor acts as an obesity-related risk factor for osteoarthritis. Adipose tissue secretes numerous cytokines and growth factors as well as the adipose-derived hormones, the adipokines (including adiponectin, resistin, visfatin and leptin),14 to which chondrocytes are known to respond and in some cases produce. For example, adiponectin induces IL-8 expression by chondrocytes,15 osteoarthritis chondrocytes are also responsive to both visfatin, which stimulates the expression of ADAMTS4, ADAMTS5, MMP3 and MMP13 and induces prostaglandin E2 release, and resistin, which induces MMP1, MMP13 and ADAMTS4, both of which can thus be considered catabolic.16 17

The best characterised role of leptin is to act as a signal for the central nervous system to inhibit food intake and to stimulate energy expenditure;18 however, it also plays a role in various physiological processes such as lipid metabolism, haematopoiesis, immune function, angiogenesis, reproduction, bone formation and inflammation.19,,25 It is mainly secreted by adipocytes, and the levels of leptin within the circulation correlate with the amount of white adipose tissue (WAT).18

Articular cartilage produces leptin26 27 and expresses the functional leptin receptor Ob-R;28 the expression of these two proteins is further increased in advanced osteoarthritis and correlates with the BMI of osteoarthritis patients.29 Leptin has also been shown to have a pro-inflammatory effect on chondrocytes by inducing nitric oxide synthase.30 31

The aim of the present study was to investigate the effects of leptin alone or in combination with pro-inflammatory cytokines on cartilage metabolism by measuring its effects on the release of cartilage collagen, expression and production of collagenases and gelatinases, and by evaluating the mechanisms involved in these effects. Importantly, we show that the infrapatellar fat pad, and other WAT, produce significant amounts of leptin, which can contribute to the pro-inflammatory milieu of the joint, the induction of collagenase gene expression by chondrocytes, and potentially cartilage catabolism, providing a biological link between obesity and osteoarthritis.

Materials and methods

IL-1α was a generous gift from Dr K Ray (Glaxo-SmithKline, Stevenage, UK), and recombinant human oncostatin-M was prepared in-house using expression vectors kindly provided by Professor J Heath (University of Birmingham, UK) and methods described.32 TNFα, recombinant human leptin and adiponectin, anti-human leptin monoclonal antibody and human leptin ELISA kit were from R&D Systems (Abingdon, UK). Resistin and visfatin were from First Link (UK) Ltd (Birmingham, UK). The mouse lgG control was from Becton-Dickinson (UK). Primary antibodies against phospho-Erk1/2 (phospho-p44/42) (#9101), phospho-p38 (#9211), phospho-JNK1/2 (#9251), phospho-p65 (Ser 536; #3031) and phospho-STAT were all from Cell Signaling Technology (New England Biolabs, Hitchin, UK). The phospho-Akt antibodies used have been described previously.33 A rabbit anti-β tubulin antibody (ab6046) was from Abcam (Cambridge, UK). Secondary immunoglobulins/horseradish peroxidase were from Cytomation (Dako, Glostrup, Denmark). Chemical pathway inhibitors LY294002, U0126, SB203580, SP600125 and Akt inhibitors IV and VIII were from Merck Chemicals (Nottingham, UK). Tissue culture reagents were obtained from Lifetechnologies (Paisley, UK).

Cartilage degradation and enzyme activity assay

Bovine nasal cartilage has previously been validated as a model relevant to human disease34 and was cultured as previously described.7 Briefly, three discs (∼2 mm3 discs) of bovine nasal cartilage/well were incubated with or without test reagents in serum-free Dulbecco's modified eagle medium (DMEM) for 14 days, with a medium change at day 7. Selective pathway inhibitors were added 30 min before test reagents at the concentrations indicated. Day 7 and day 14 supernatants were stored at −20°C. Cartilage collagen degradation was measured using a hydroxyproline assay7 and collagenase activities in culture supernatants were determined by the 3H-acetylated collagen diffuse fibril assay.35 Aminophenylmercuric acetate (0.67 mM) was used to activate pro-collagenases. Gelatin zymography was used to measure gelatinase activity in the culture supernatants.36

Chondrocyte isolation and culture

Primary human articular chondrocytes (HAC) were derived from articular cartilage obtained from joint replacement patients diagnosed with osteoarthritis. Normal cartilage was obtained from neck of femur fracture patients (with no history of osteoarthritis) undergoing joint replacement surgery. All tissue was obtained with informed consent and ethics committee approval from the Newcastle and North Tyneside Health Authority. Enzymatic digestion of tissue and maintenance and culture of cells were as previously described.9 When cells reached 80–90% confluence without passage they were serum-starved for 16 h before pretreatment with selective pathway inhibitors (for 30 min at the indicated concentrations) before the addition of leptin, cytokines or WAT conditioned media at the concentrations and duration indicated or as previously described.37 Bovine nasal chondrocytes were isolated and cultured as previously described.7 9

RNA extraction and real-time reverse transcription PCR

Total RNA was isolated from cultured HAC or bovine chondrocytes and reverse transcribed using the Ambion Cells-to-cDNA II Kit (Lifetechnologies). Oligonucleotides were purchased from Sigma-Genosys (Poole, UK) and were as previously described.36 38 Relative quantification of genes was performed using the ABI Prism 7900HT sequence detection system (Lifetechnologies). Bovine metalloproteinase expression profiles were determined using SYBR Green (Takara, Cambrex, Wokingham, UK) in accordance with the manufacturer's protocol and as described.38 Throughout, the 18s ribosomal RNA gene was used as an endogenous housekeeping control. TaqMan probe-based assays were performed using TaqMan Gene Expression mastermix (Lifetechnologies) according to the manufacturer's protocol as previously reported.36 Adipokine receptor expression in neck of femur fracture and osteoarthritis hip cartilage patient samples was determined as part of a TaqMan low-density array screen using assays ADIPOR1-Hs00360422_m1, ADIPOR2-Hs00226105_m1, ADIPOQ-Hs00605917_m1, LEPR-Hs00174497_m1 and LEP-Hs00174877_m1 (Lifetechnologies), again normalised to the housekeeping gene 18s rRNA.


Cells were lysed after the indicated time points with ice-cold buffer (50 mM Tris-HCl, pH 7.5; 1.2 M glycerol; 1 mM ethylene glycol tetraacetic acid; 1 mM EDTA; 1 mM Na3VO4; 10 mM β-glycerophosphate; 50 mM NaF; 5 mM sodium pyrophosphate; 1% (v/v) Triton X-100; 1 μM microcystin-LR; 0.1% (v/v) β-mercaptoethanol; Roche protease inhibitor complex (Roche, UK)), particulate matter removed by centrifugation at 13 000g, 5 min at 4°C, and lysates stored at −80°C until required. Lysates were resolved by sodium dodecyl sulphate polyacrylamide gel electrophoresis, transferred to polyvinylidene fluoride membranes (Millipore, Watford, UK) and subsequently probed using the antibodies described. The optimum time points for signalling pathway activation were based upon our previous work with HAC.33 36 37 39

WAT-conditioned media

The infrapatellar fat pad or peri-enthesal deposits from human osteoarthritis knee joints were the source of WAT. This was separated from the rest of the joint and any fibrous material, such as connective tissue or blood vessels, before dissection into small pieces of approximately 50 mg. The tissue was then washed twice with phosphate-buffered saline and once with serum-free DMEM. After washing, the WAT was incubated for 24 h at 37°C in DMEM culture medium containing 5% fetal calf serum and antibiotics (streptomycin (100 μg/ml) and penicillin (100 IU/ml)), which was replaced with serum-free DMEM (1 ml/0.3 g WAT) and incubated for 72 h at 37°C. After harvesting, the media were filter sterilised and stored at −80°C before use. To quantify the amount of leptin present in the WAT-conditioned media, the Human Leptin Quantikine ELISA Kit (R&D Systems, UK) was performed according to the manufacturers' instructions.

Statistical analysis

For cartilage degradation assays and collagenase bioassays one-way analysis of variance with Bonferroni post hoc statistical tests were performed. Statistical differences between sample groups in the real-time reverse transcription (RT)–PCR experiments were assessed using the two-tailed Student's t test. SPSS 15.0 software was used for all statistical analyses. Significance levels were indicated as *p<0.05, **p<0.01 and ***p<0.001.


Leptin induces collagenase gene expression and cartilage resorption

To study the effects of leptin on cartilage catabolism, bovine cartilage was stimulated with leptin in the absence or presence of several pro-inflammatory cytokines; IL-1, TNFα or oncostatin-M. Leptin alone induced low but significant (p<0.05) collagen release (approximately 5%) compared with control (figure 1A). IL-1 or TNFα alone induced approximately 30% collagen release but when combined with leptin this synergistically increased to over 60% (figure 1A). This increase in collagen release was accompanied by an upregulation of total and active collagenolytic and gelatinolytic (MMP2 and MMP9) activities (figure 1B, C), demonstrating a role of leptin in the activation of a number of proMMP. Oncostatin-M alone or in combination with leptin did not effect cartilage resorption.

Figure 1

Leptin in combination with pro-inflammatory cytokines induces cartilage collagen release from bovine cartilage by enhancing collagenolytic and gelatinolytic activities. Bovine nasal cartilage was treated with leptin (25 µg/ml unless indicated otherwise) with oncostatin-M (10 ng/ml), IL-1α (0.25 ng/ml) or tumour necrosis factor α (10 ng/ml) in serum-free medium for 14 days. The release of collagen and enzyme activities were assayed as described in the Materials and methods section. (A and D) Cumulative collagen release by day 14 is expressed as a percentage of the total (mean±SD, n=4). Inset shows an expansion of the leptin alone data for clarity. (B) Active (open bars) and total (solid bars) collagenolytic activity values (mean±SD, n=4). (C) Gelatin zymography of day 14 media. Pro and active MMP2 and MMP9 are indicated. *p<0.05; **p<0.01; ***p<0.001 versus each respective control. Data are representative of three separate cartilage experiments. (E and F) Primary human articular chondrocytes were stimulated with leptin as indicated with or without IL-1 (0.25 ng/ml) for 24 h. The expression of MMP1 and MMP13 and 18s were assayed by real-time reverse transcription PCR. The levels of MMP1 (solid bars; E) and MMP13 (grey bars; F) were expressed as fold against 18s rRNA. *p<0.05; **p<0.01 versus control or cytokine alone. Experiments were performed as n=4 and are representative of three independent experiments. IL, interleukin; MMP, matrix metalloproteinase; OSM, oncostatin-M; TNFα, tumour necrosis factor alpha.

Our previous work has focused on synergy between various inflammatory mediators, especially with IL-1 and the potent regulation of the collagenases.7,,9 33 36 40 We therefore assessed the synergy between leptin and IL-1. A concentration-dependent increase in cartilage collagen degradation was observed with leptin concentrations exceeding 10 μg/ml, although when in combination with IL-1 (0.25 ng/ml) even the lowest concentration of leptin tested (1 μg/ml) elicited a synergistic increase in cartilage collagen release (figure 1D). We also investigated whether other adipokines (visfatin, resistin and adiponectin) could induce or potentiate cartilage resorption but, surprisingly, only leptin showed such efficacy (see supplementary figure 1A, available online only) despite previous literature demonstrating their ability to induce a catabolic phenotype in chondrocytes15,,17 and significant expression by chondrocytes of adiponectin receptors (see supplementary figure 2, available online only).

When added to primary HAC, leptin induced the gene expression of the major collagenases MMP1 and MMP13. Again, when in combination with IL-1 (0.25 ng/ml) higher concentrations of leptin (>1 μg/ml) synergistically induced these collagenases (figure 1E, F). A similar induction of MMP1 and MMP13 by leptin and leptin plus IL-1 was also observed in isolated bovine chondrocytes (see supplementary figure 1B,C, available online only), emphasising the suitability of the bovine cartilage model.

Leptin activates STAT, MAPK and Akt signalling in primary HAC

Stimulation of primary HAC with leptin with or without IL-1 activated multiple signalling pathways. Leptin alone induced STAT1, STAT3 and STAT5 tyrosine phosphorylation (at positions Y701, Y705 and Y694, respectively) but failed to induce robust STAT serine phosphorylation, with the exception of a modest increase in STAT3 S727 phosphorylation. This was in contrast to IL-1, which only induced STAT serine phosphorylation. Some synergy between leptin and IL-1 occurred with the phosphorylation of STAT1 and STAT3 at S727. These activation events declined by 60 min poststimulation (figure 2A). Leptin alone activated all three mitogen-activated protein kinase (MAPK) pathways, evidenced by phosphorylation of stress-activated protein kinase (SAPK)/cJun-amino-terminal kinase JNK (T183/Y185), Erk (T202/Y204 of Erk1) and p38 (T180/Y182). When combined with IL-1, an increase in phosphorylated p38 and JNK1/2 were detected (figure 2B), the latter only becoming obvious by 60 min. Leptin also led to the phosphorylation of Akt (at S473 and Y308), although no additional phosphorylation was observed when combined with IL-1 (figure 2C). Finally, leptin led to a modest phosphorylation of the nuclear factor kappa B (NF-κB) subunit p65 at 20 min but, as with the serine phosphorylation of STAT-1/3, this was not as pronounced as that elicited by IL-1 alone (figure 2C).

Figure 2

Signalling pathways activated after leptin and leptin plus IL-1 stimulation of chondrocytes and their role in collagenase expression. Primary human articular chondrocytes were serum starved overnight and then stimulated for 20 and 60 min with either leptin (25 μg/ml) or IL-1 (0.25 ng/ml) alone or in combination. Cells were lysed, and extracts immunoblotted with antibodies against (A) phospho-STAT1 (Y701 or S727), phospho-STAT3 (Y705 and S727), phospho-STAT5 (Y694) and β-tubulin as loading control; (B) phospho-Erk (p44/42) (T202/Y204 of Erk1), phospho-JNK (SAPK)(T183/Y185), phospho-p38 (T180/Y182); or (C) phospho-Akt (S473 and Y308) and phospho-p65 (S536). (D and E) Primary human chondrocytes were stimulated with leptin (25 μg/ml) with or without the inhibitors, LY (LY249002, 5 μM), U0126 (5 μM), SB (SB203580, 10 μM), SP (SP600125, 10 μM), Akt IV (3 μM) or Akt VIII (3 μM) for 20 h. Cells were pretreated for 30 min with inhibitors or vehicle control before the addition of leptin. The expression of MMP1 (D) and MMP13 (E) was assayed by real-time reverse transcription PCR and expressed as fold relative to the control after normalisation to 18s rRNA. *p<0.05; **p<0.01; ***p<0.001 versus control. Data are representative of three separate experiments. IL, interleukin; MMP, matrix metalloproteinase.

Different signalling pathways control leptin-induced collagenase expression and cartilage collagen release

Next, to test whether leptin regulated the expression of MMP1 and MMP13 via the identified signalling pathways, HAC were cultured in the presence of selective pathway inhibitors. All the inhibitors, with the exception of Akt VIII, significantly suppressed the leptin-induced expression of MMP1, while only the JNK inhibitor, SP600125, failed to inhibit leptin-induced MMP13 (figure 2D, E).

In the bovine cartilage resorption assay, all the inhibitors significantly reduced the modest collagen release elicited by leptin (figure 3A). When the combination of IL-1 and leptin was used as the stimulus, only the JNK and Akt VIII inhibitors were ineffective at blocking resorption (figure 3B). This inhibitor profile was reminiscent of that seen for the MMP13 gene expression experiment (compare figures 2E and 3B).

Figure 3

Effects of signalling pathway inhibitors on leptin-induced cartilage collagen release and WAT-conditioned media cartilage degradation. (A) Bovine cartilage was stimulated with leptin (25 μg/ml) with or without inhibitors, LY (LY249002, 5 μM), U0126 (5 μM), SB (SB203580, 10 μM), SP (SP600125, 10 μM), Akt IV (3 μM) or Akt VIII (3 μM) or (B) leptin plus IL-1 (0.25 ng/ml) with or without the pathway inhibitors described above, in serum-free medium for 14 days. Cartilage was pretreated for 30 min with inhibitors or vehicle control before the addition of leptin or leptin plus IL-1. (C and D) WAT samples were collected from the infrapatellar fat pad of osteoarthritis patients as described and incubated in 5% fetal calf serum/DMEM culture medium (1 ml medium/0.3 g WAT) at 37°C for 24 h, then incubated in serum-free DMEM for 3 days. 20% (v/v) WAT-conditioned media from four patients were then added to bovine cartilage alone (C) or plus IL-1 (0.25 ng/ml) (D) in serum-free medium for 14 days. In all experiments on day 7, media were collected and replaced with identical test reagents. The release of collagen was assayed as described in the Materials and methods section. Cumulative collagen release by day 14 is expressed as the percentage of total (n=4). *p<0.05; **p<0.01; ***p<0.001 versus control. Data are representative of three separate experiments. DMEM, Dulbecco's modified eagle medium; IL, interleukin; WAT, white adipose tissue.

WAT-conditioned culture media induce MMP and cartilage resorption

In order to assess whether WAT-derived leptin WAT might have an impact on cartilage degradation we initially confirmed that WAT-conditioned, not control, conditioned media were able to induce MMP1 and MMP13 by primary bovine chondrocytes (see supplementary figure 3A, available online only). Next, WAT-conditioned culture media (using fat pads from four different osteoarthritis patients) were incubated with bovine nasal cartilage with or without IL-1. Alone, the WAT-conditioned media induced modest, but significant, collagen release compared with control-conditioned medium (figure 3C) and when combined with IL-1, a further synergistic release of proteoglycan (not shown) and collagen was observed (figure 3D).

Leptin in WAT-conditioned media contributes to the induction of MMP1 and MMP13 expression

Next, HAC were cultured with WAT-conditioned media taken from fat pads of 14 osteoarthritis patients. Ten WAT media samples significantly upregulated MMP13, while the same 10 samples, plus N1688, also significantly upregulated MMP1 expression. The levels of induction of MMP1 and MMP13 obtained were comparable to those induced by leptin (25 μg/ml) alone (figure 4A). To determine if the induction of the collagenases induced by the WAT media was partly mediated by secreted leptin we first quantified the amount of leptin present in the WAT-conditioned media using a specific leptin immunoassay (figure 4B). After the removal of outliers, 12 WAT-conditioned media contained a wide range of leptin levels from 99 to 1777 pg/ml (figure 4B). Using regression analysis, the amount of leptin showed a trend towards a positive correlation with the MMP1 and MMP13 fold induction by these media (R2=0.1798 and 0.4897, respectively), which approached but importantly did not reach significance, suggestive of a role for leptin in combination with other WAT-produced pro-inflammatory factors in collagenase induction. To attempt to test conclusively if the leptin produced by WAT was involved in the upregulation of the collagenases induced by the conditioned media, HAC were incubated with WAT-conditioned media with or without an anti-leptin antibody. When analysing the WAT-conditioned media that induced each collagenase more than five-fold, the leptin blocking antibody led to a significant reduction in the induced collagenase levels (figure 4C, D). Finally, a similar experiment was repeated in which HAC were again incubated with WAT-conditioned media with or without an anti-leptin or control (IgG) antibody. As expected, the anti-leptin but not the control antibody significantly reduced the level of induction of both collagenases by leptin, and moreover did so in four out of five WAT-conditioned media examined (see supplementary figure 3B,C, available online only).

Figure 4

Leptin in WAT-conditioned media and its effect on MMP1 and MMP13 expression. (A) Primary HAC were cultured in serum-free medium for 24 h and then incubated with 20% (v/v) WAT-conditioned media in DMEM for 24 h. The expression of MMP1 (black) and MMP13 (grey), normalised to 18s rRNA, were assayed by real-time RT–PCR (n=4/WAT-conditioned medium). (B) The levels of leptin in WAT-conditioned media collected from osteoarthritis fat pad were assayed by ELISA then expressed relative to the fold induction of MMP1 (○) or MMP13 (■) for each WAT-medium sample. Best fit lines are leptin versus MMP1, broken; leptin versus MMP13, solid. (C and D) Primary HAC were co-incubated with an anti-leptin antibody (20 µg/ml) and 20% (v/v) WAT-conditioned media in DMEM for 24 h, then the expression of MMP1 (C) and MMP13 (D) assayed by real time RT–PCR and normalised to 18s rRNA levels (n=3). Only WAT media that induced a greater than five-fold increase in collagenase expression are included. *p<0.05; **p<0.01 and ***p<0.001 versus basal control or WAT-conditioned medium induced levels. HAC, human articular chondrocytes; MMP, matrix metalloproteinase; RT–PCR, reverse transcription PCR; WAT, white adipose tissue.


Several studies have linked leptin, obesity and osteoarthritis.27 41,,47 For example, serum leptin concentrations correlate with BMI and are closely related to the radiographic severity, joint inflammation and disease severity of osteoarthritis. Others have reported a marked increase in leptin protein expression in osteoarthritis cartilage compared with normal cartilage.26 However, the most compelling evidence that leptin plays a direct, rather than correlative, role in osteoarthritis pathogenesis comes from studies using leptin signalling-deficient mice. These mice are severely obese yet exhibited the same (increased) level of osteoarthritis incidence as that of obese wild-type mice.48

In the current study, we demonstrated for the first time that leptin, but surprisingly not the other adipokines visfatin, resistin or adiponectin, can directly induce cartilage collagen degradation, alone and when combined with pro-inflammatory cytokines (IL-1 or TNFα). Adiponectin has previously been shown to induce the expression of IL-8 by chondrocytes15 but also to have a chondroprotective role.49 We found little response to this adipokine in our assays even though one of its receptors (ADIPOR2) was highly, and moreover, differentially expressed in diseased cartilage. We also found bovine cartilage to be unresponsive to both resistin and visfatin (see supplementary figure 1, available online only), both of which have previously been reported to induce a number of metalloproteinases when added to chondrocytes.16 17 These discrepancies could be due to a number of factors, including the tissue examined (cartilage vs cells), species, source of adipokine and culture conditions. The effects of leptin upon cartilage collagen release were accompanied with an upregulation of collagenase and gelatinase activities in the culture medium. A recent study has similarly shown an induction of the collagenases from human osteoarthritis cartilage following leptin or leptin plus IL-1 stimulation.50 The same group also showed that leptin alone or in combination with IL-1 enhanced the expression of the pro-inflammatory mediators inducible nitric oxide synthase and cyclooxygenase 2 and the production of nitric oxide, prostaglandin E2, IL-6 and IL-8.43

Leptin signalling through leptin receptors (Ob-R/LEPR) is mediated via the janus kinase (JAK)/STAT pathway, but additionally the phosphatidylinositol 3 kinase (PI3K)/Akt, MAPK and NF-κB pathways, all of which have been reported to be important in leptin signalling by chondrocytes,30 43 in accordance with the data presented here. Interestingly, our data show that leptin alone is a better inducer of STAT phosphorylation at the tyrosine position rather than serine, the latter only becoming prominent when leptin acts synergistically with IL-1. This is consistent with the role of the tyrosine kinase family JAK, probably JAK2, in Ob-R (LEPR) signalling.51 We also showed for the first time that the MAPK and PI3K pathways are important mediators of leptin signalling in chondrocytes and cartilage as chemical blockade of any of the these pathways prevented cartilage collagen release. Interestingly, the JNK inhibitor, SP600125, did not inhibit either leptin plus IL-1-induced cartilage collagen release or leptin-induced MMP13 by chondrocytes but did block MMP1 levels, supporting the theory that MMP-13 is the major type II collagen-degrading enzyme52 and validating the bovine cartilage model.

Gradual loss of articular cartilage is a major characteristic of osteoarthritis, with initial changes including the loss of proteoglycan. With this in mind, the in-vivo injection of leptin into the joints of rats induces cartilage proteoglycan loss and increases the expression of MMP2 and MMP9 in articular cartilage.46 The destructive process in osteoarthritis is determined by an imbalance between anabolic and catabolic mediators (and their regulators) in the joint, and the local distribution of these mediators in the cartilage.10 52 Leptin appears to be an important local and systemic factor influencing cartilage/bone homeostasis. It is likely that locally produced, rather than circulatory, leptin may be more important in regulating cartilage homeostasis because the concentration of leptin in synovial fluid is higher than in the corresponding sera of osteoarthritis patients,29 and these levels are associated with disease activity.26 53 Moreover, expression of leptin and LEPR have previously been detected in human cartilage, with levels significantly increased in end-stage osteoarthritis cartilage.26 41 The endogenous leptin produced by chondrocytes has also been shown to induce MMP13 expression.54 However, although we could detect leptin and LEPR expression in cartilage, neither were differentially expressed in osteoarthritis in our experiments and the level of leptin expression detected was deemed negligible (see supplementary figure 2, available online only). Therefore, we focused on leptin production by the infrapatellar fat pad, a joint tissue reported to be a site of leptin production.41 For the first time to our knowledge we have shown that WAT-conditioned media, collected from cultured infrapatellar fat pad and other WAT of osteoarthritis knee joints, contains significant quantities of leptin, the amount of which correlated with the ability of the WAT-conditioned media to stimulate MMP1 and MMP13 expression in chondrocytes. These WAT-conditioned media were also able to induce collagen release from cartilage cultures. Furthermore, the ability of these WAT-conditioned media to stimulate MMP1 and MMP13 expression was partly abolished by an antileptin antibody. Together, these data suggest that leptin produced by the WAT-conditioned media was partly responsible for the cartilage degradation observed, but importantly that these media contain other pro-inflammatory cytokines, such as IL-1, whose expression by WAT tissue we could detect (data not shown), which synergise with leptin to induce collagenase expression.

Taken together our findings suggest that the infrapatellar fat pad of arthritic joints can be a significant local producer of leptin and may contribute to inflammatory and degenerative processes in osteoarthritic cartilage catabolism via the upregulation of the collagenolytic metalloproteinases MMP1 and MMP13, thus providing a direct mechanistic link between obesity and osteoarthritis.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Funding This work was supported by the Action Medical Research, Arthritis Research UK grants 18726 and 19485, the JGWP Foundation, the Newcastle University Hospitals special trustees, UK and the UK NIHR Biomedical Research Centre for ageing and age-related disease award to the Newcastle upon Tyne Hospitals NHS Foundation Trust. Clinical and translational research in the Musculoskeletal Research Group is supported by the Northumberland, Tyne and Wear comprehensive local research network.

  • Competing interests None.

  • Ethics approval All tissue was obtained with informed consent and ethics committee approval from the Newcastle and North Tyneside Health Authority.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.