Upregulation of Bone Morphogenetic Protein-2 Synthesis and Consequent Collagen II Expression in Leptin-stimulated Human Chondrocytes

Bone morphogenetic proteins (BMPs) play positive roles in cartilage development, but they can barely be detected in healthy articular cartilage. However, recent evidence has indicated that BMPs could be detected in osteoarthritic and damaged cartilage and their precise roles have not been well defined. Extremely high amounts of leptin have been reported in obese individuals, which can be associated with osteoarthritis (OA) development. The aim of this study was to investigate whether BMPs could be induced in human primary chondrocytes during leptin-stimulated OA development and the underlying mechanism. We found that expression of BMP-2 mRNA, but not BMP-4, BMP-6, or BMP-7 mRNA, could be increased in human primary chondrocytes under leptin stimulation. Moreover, this BMP-2 induction was mediated through transcription factor-signal transducer and activator of transcription (STAT) 3 activation via JAK2-ERK1/2-induced Ser727-phosphorylation. Of note, histone deacetylases (HDACs) 3 and 4 were both involved in modulating leptin-induced BMP-2 mRNA expression through different pathways: HDAC3, but not HDAC4, associated with STAT3 to form a complex. Our results further demonstrated that the role of BMP-2 induction under leptin stimulation is to increase collagen II expression. The findings in this study provide new insights into the regulatory mechanism of BMP-2 induction in leptin-stimulated chondrocytes and suggest that BMP-2 may play a reparative role in regulating leptin-induced OA development.


Introduction
Our findings elucidate the detailed mechanism of BMP-2 expression and indicate a possible reparative role for BMP-2 in leptin-stimulated chondrocytes.

Cell culture
Human primary OA chondrocytes were isolated from OA patients who underwent knee replacement surgery (n = 70; ages 65-82 years). The diagnosis in all patients were based on the American College of Rheumatology criteria for knee OA [25]. The Ethics Committees of Taipei Veterans General Hospital approved the study protocol, and written informed consent was obtained from all patient subjects. Primary chondrocytes were obtained from the non-lesional areas of OA cartilage. The preparation of first passage chondrocytes from human cartilage was performed according to the published report [26]. In brief, the articular cartilage was washed with phosphate buffered saline (PBS) and removed from the underlying bone and then cut into pieces of around 0.5 cm 2 . After sequential enzyme digestion with Pronase (2 mg/mL) (Calbiochem, La Jolla, CA) and then with collagenase I (0.25 mg/mL), the cell suspension was filtered through 100 μm nylon meshes. The isolated chondrocytes were washed repeatedly with PBS and then cultured in DMEM medium containing 10% FBS and 1% antibiotics for 5-7 days before use (~70% confluence). To ensure that the cell phenotype was maintained, experiments were performed using only first-passage cultured chondrocytes. In the experiment, cells were seeded in 6-well plates (~70% confluence) in complete DMEM medium and serum starved for 12 h and then treated with leptin for various time intervals.

Western blot analysis
Cells were collected and lysed with a RIPA buffer containing 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, and a protease inhibitor mixture (PMSF, aprotinin, and sodium orthovanadate). The total cell lysate was separated by SDS-PAGE and transferred onto a nitrocellulose membrane. The membrane was then incubated with the designated antibodies. Immunodetection was performed by using the Western-Light chemiluminescent detection system (Applied Biosystems, Foster City, CA).

siRNA transfection
For siRNA transfection, cells at 70% confluence were transfected with designed siRNA (40 nM), which caused~50-80% reductions in expressions of the corresponding proteins (S1 Fig

Immunoprecipitation
The cells were scraped and lysed with a RIPA buffer and the same amount of protein from each sample was incubated with a designated antibody. The immune complex was then incubated with protein A/G agarose and collected by centrifugation. This agarose-bound immunoprecipitates were washed and incubated with boiling sample buffer and then were subjected to SDS-PAGE and Western blotting.

Chromatin immunoprecipitation (ChIP) assay
ChIP assays were performed using EZ ChIP Assay Kit (Upstate Biologicals) in accordance with recommended protocols. In brief, after stimulation with leptin for indicated time, the cells were fixed with 1% formaldehyde. Immunoprecipitation analysis was carried out using anti-STAT3 antibody. Real-time PCR was performed with primers (sense: 5 0 -CACTCAATTTCCAGC CTGCT-3 0 ; antisense: 5 0 -AAACACCTCAATTACAGCG-3 0 ) that amplified the part of the human BMP-2 promoter that contains a putative STAT3 binding site. Then 10% of the chromatin DNA used for immunoprecipitation was similarly subjected to real-time PCR analysis and indicated as an internal control.

Statistical analysis
Results are expressed as mean ± standard error of the mean (SEM). Statistical analysis was performed by using an independent Student t-test for two groups of data and analysis of variance (ANOVA) followed by Scheffe's test for multiple comparisons. A P value less than 0.05 was considered significant. Results were calculated from at least 3 repetitions generated from individual patients.

Leptin induces BMP-2 expression in human primary chondrocytes
We examined whether leptin induces BMP expression in human primary chondrocytes. Cells were kept as controls or treated with leptin (5 ng/ml) for 1, 4, 8, and 24 h, and their expression of BMP-2, BMP-4, BMP-6 and BMP-7 mRNA was examined. Treatment of cells with leptin resulted in significant increases (within 4 h) in BMP-2 mRNA expression, which reached a maximal level of approximately 6.5 times that of untreated controls within 8 h and then slightly declined but remained elevated after 24 h of treatment ( Fig 1A). The expression of BMP-4, BMP-6 and BMP-7 mRNA did not increase in leptin-stimulated cells. Moreover, leptin also induced BMP-2 protein secretion into the medium within 4 h, which persisted for 24 h, compared with untreated control cells ( Fig 1B).

Leptin-induced BMP-2 mRNA expression is mediated by JAK2-ERK1/2 signaling in human primary chondrocytes
The JAK and ERK pathways are known as leptin-directed downstream signaling pathways [15]. We investigated the roles of JAK1/2 and ERK1/2 in leptin-induced BMP-2 mRNA expression in human primary chondrocytes. Cells stimulated with leptin (5 ng/ml) induced a rapid and transient increase in JAK2 (not JAK1) and ERK1/2 phosphorylation within 1 h compared with untreated control cells (Fig 2A). The increased levels of JAK2 and ERK1/2 phosphorylation declined after 8 h of stimulation. Transfection of cells with JAK2-specific (not JAK1-specific) siRNA significantly inhibited the leptin-induced ERK1/2 phosphorylation ( Fig 2B) and BMP-2 mRNA expression ( Fig 2C). Transfection of cells with ERK1-or ERK2-specific siRNA also significantly inhibited the leptin-induced BMP-2 mRNA expression ( Fig 2D).

STAT3 regulates leptin-induced BMP-2 mRNA in human primary chondrocytes
STATs play a critical role in mediating leptin-regulated gene transcription [15]. Treatment of human primary chondrocytes with leptin (5 ng/ml) induced increases in STAT3 phosphorylation at Ser727, but not Tyr705, within 1 h that persisted until 24 h. Leptin induced STAT5 phosphorylation at Tyr694 within 4 h, which then persisted until 8 h and declined thereafter. Moreover, leptin also induced STAT5 protein expression within 4 h that persisted until 24 h ( Fig 3A). Transfection of cells with ERK1-or ERK2-specific siRNA significantly inhibited leptin-induced STAT3 Ser727-phosphorylation, but not that of Tyr705 ( Fig 3B). Moreover, transfecting cells with STAT3-or STAT5-specific siRNA showed that only STAT3-specific siRNA transfection abolished leptin-induced BMP-2 mRNA expression ( Fig 3C). Chromatin immunoprecipitation (ChIP) assays using an antibody against STAT3 and primers against the putative STAT3-binding element of the BMP-2 promoter demonstrated that exposure of cells to leptin (5 ng/ml) for 1, 4, or 8 h induced in vivo STAT3 binding to the STAT3-binding element of the BMP-2 promoter in human primary chondrocytes (Fig 3D).

HDAC3/4 mediate leptin-induced BMP-2 mRNA expression in human primary chondrocytes
Pretreating cells with DMSO or TSA (1 μM), an HDAC activity inhibitor, and then keeping them as controls or stimulating them with leptin (5 ng/ml) for 4 or 8 h, showed that HDAC activity inhibition blocked leptin-induced BMP-2 mRNA expression in human primary chondrocytes ( Fig 4A). Next, we determined which HDAC members were involved in mediating leptin's effect on BMP-2 mRNA expression. Human primary chondrocytes were kept as controls or treated with leptin (5 ng/ml) for 1, 4, 8, or 24 h, and the expression of class I (HDAC1/2/3) and class II (HDAC4/6/7) HDACs was examined. Treatment of cells with leptin induced HDAC3 and 4 expression within 4 h and that persisted until 24 h compared with untreated control cells ( Fig  4B). Cellular fractionation assays, involving separation of cytoplasmic and nuclear fractions of cell lysates derived from leptin-treated cells, demonstrated that leptin induced sustained increases in HDAC3 expression in both the cytoplasm and nucleus and HDAC4 expression in the nucleus (Fig 4C). Moreover, transfecting cells with HDAC3-or 4-specific siRNA also inhibited leptininduced BMP-2 mRNA expression (Fig 4D) in human primary chondrocytes.
STAT3-HDAC3 and HDAC4 are divergent pathways mediating leptininduced BMP-2 mRNA expression in human primary chondrocytes Next, we determined whether HDAC3/4 and STAT3 cooperatively regulate BMP-2 mRNA expression under leptin stimulation. Human primary chondrocytes were kept as controls or treated with leptin (5 ng/ml) for 1, 4, or 8 h. Cell extracts were immunoprecipitated with an antibody against STAT3 followed by Western blot analysis with an antibody against HDAC3 or 4. Leptin induced rapid increases in the association of HDAC3 with STAT3 within 1 h, as shown by their coimmunoprecipitation, in comparison with untreated control cells (Fig 5A). In contrast, leptin did not induce an association between HDAC4 and STAT3 in human primary chondrocytes. Further, a STAT3 deacetylation assay using Western blotting showed that deacetylation of STAT3 was not increased by treating cells with leptin for 1, 4, 8, or 24 h, even under the formation of the HDAC3 and STAT3 complex (Fig 5B).

BMP-2 induction in leptin-stimulated human primary chondrocytes increases collagen II expression
Finally, we determined whether the role of leptin-induced BMP-2 expression in human primary chondrocytes is to drive the cells into the osteoblast lineage and hence promote osteophyte formation or to alter the metabolic homeostasis of chondrocytes. Cells were kept as controls or treated with leptin (5 ng/ml) for 10', 30', 1, 4, 8, or 24 h, and the expression of col1a1, runx2 and osteocalcin (osteoblast markers) mRNA and collagen II (anabolic marker) and COX2 (catabolic marker) protein were examined. Stimulation of cells with leptin did not affect the mRNA expression of col1a1, runx2 or osteocalcin (Fig 6A). However, cells stimulated with leptin simultaneously induced both collagen II and COX2 protein expression after 4 h that persisted until 24 h compared with untreated control cells (Fig 6B). Treating cells with Noggin (100 ng/mL), a specific antagonist that binds to BMPs to block their binding to BMP receptors, blocked leptin-induced collagen II, but not COX2, expression (Fig 6C). Moreover, transfecting cells with BMP-2-, Smad1-, or Smad5-specific siRNA further showed that the gene knockdown of BMP-2 and Smad1/5 abolished only leptin-induced collagen II expression (Fig 6D).

Discussion
This study has elucidated the mechanism whereby leptin upregulates BMP-2 synthesis and consequently increases collagen II expression in human primary chondrocytes (summarized in Fig 7). The systematic experiments demonstrated that (i) leptin induces expression of BMP-2, but not that of BMP-4, BMP-6, or BMP-7, in human primary chondrocytes. (ii). Leptininduced BMP-2 expression is regulated by JAK2-ERK1/2-induced STAT3 Ser727-phosphorylation. (iii) Both HDAC3 and 4 are involved in regulating leptin-enhanced BMP-2 mRNA expression but through different pathways: HDAC3, but not HDAC4, associates with STAT3 to cooperatively regulate BMP-2 mRNA expression. (iv) The role of leptin-enhanced BMP-2 synthesis in human primary chondrocytes is to increase collagen II expression. Our findings provide new insights into the regulatory mechanism and potential reparative effect of BMP-2 synthesis in human chondrocytes under leptin-initiated OA development.
The regulation of various stages of chondrogenesis by BMPs has been well defined [27][28][29]. However, recently, accumulating evidence has indicated that some BMPs are detected in osteoarthritic adult cartilage. These studies have further implicated that the expression of BMPs in OA lesional areas might play important roles in the control of cartilage structure and composition, for example by improving cartilage integrity and promoting of osteophyte formation [2][3][4][5][6][7][30][31]. Our data revealed the upregulation of BMP-2 synthesis in human primary chondrocytes when leptin induced OA marker expression. Further experiments showed that leptin did not affect the mRNA expression of col1a1, runx2, or osteocalcin (osteoblast markers), but it did increase the protein expression of collagen II (an anabolic marker) through BMP-2. Therefore, although we cannot completely rule out a possible role for BMP-2 in promoting osteophyte formation under leptin stimulation, the results from our current study provide evidence that BMP-2 might contribute to the reparative mechanism to neutralize the proinflammatory cytokines' catabolic effects, including leptin. However, we suggest that BMP-2 activity alone is not sufficient to adequately protect against the destruction of cartilage. This might also be the reason why the rescue response arises only in the early phase of OA under proinflammatory cytokine stimulation and why OA is a deleterious and irreversible disease [17]. Hence, although BMP-2 has been suggested to be a potential rescue factor that eases OA progression by previous studies and by our present study, it remains to be further determined how BMP-2 could be applied clinically in the future.
STATs are the major transcription factors of leptin signaling that regulate target gene transcription [15]. In chondrocytes, both STAT3 and STAT5 have been reported to increase the expression of matrix metalloproteinases under leptin stimulation [32]. In addition, STAT3 has also been demonstrated to regulate the expression of multiple genes in leptin-induced chondrocytes, e.g., type X collagen and Wnt-specific receptors (Frizzled receptors) [33][34]. In the present study, we found that leptin induced both STAT3 and STAT5 phosphorylation and increased STAT5 expression in human primary chondrocytes. However, our further examination elucidated that leptin-induced BMP-2 expression was regulated only by STAT3, and not STAT5. We hence suggest that the role of STAT5 induction in leptin-induced chondrocytes may be in regulating the catabolic effects of leptin in these cells. Our results showed that STAT3 was activated in leptin-induced BMP-2 expression by JAK2-ERK1/2-induced STAT3 Ser727-phosphorylation, but not Tyr705-phosphorylation. The phosphorylation of residues Tyr705 and Ser727 on STAT3 through the JAK and ERK pathways has been regarded as characteristic of STAT3 activation, increasing its transcription activity. It is generally thought that STAT3 activation requires dimerization through Tyr705-phosphorylation, and its transcriptional activity can be further modulated by Ser727-phosphorylation in a manner that is dependent on the cell type and the stimulus used [35][36]. However, accumulating evidence has indicated that the functional regulation of Ser727-phopshorylation on STAT3 is more complex than traditionally expected [36][37][38]. For example, Sakaguchi M et al. elucidated that STAT3 Ser727-phosphorylation is not necessarily a secondary event after Tyr705-phosphorylation in melanoma cells and that this Ser727-only phosphorylation on STAT3 regulates STAT3 nuclear translocation and cell survival [36]. Zhou Z et al. also indicated that the leptinenhanced MCP-1 mRNA levels are due to the increased phosphorylation of STAT3 on Ser727 but not on Tyr705 in adipose stem cells [39]. These reports could explain our results indicating that leptin-induced STAT3 activation occurs only through Ser727 phosphorylation. Hence, these findings indicate that Ser727-phosphorylation only on STAT3 is one of the important mechanisms of leptin's effect and also one of the general mechanisms of STAT3 transcription activity induction.
HDACs play a critical role in removing the Lys acetylation of histone and non-histone proteins and consequently regulating their activity [19][20]. Mounting evidence indicates that HDACs activity is involved in the control of chondrocyte gene expression and function [21][22][23][24]; however, which HDACs modulate these processes and their precise role in modulating chondrocyte responses to leptin remain poorly understood. Our present data shows that both HDAC3 and 4 are involved in leptin-induced BMP-2 mRNA expression but through distinct pathways: HDAC3, but not HDAC4, associates with STAT3. These results suggest that HDAC4 might interact with other transcription factors and/or regulate histone proteins to increase BMP-2 mRNA expression, warranting further exploration. Previous studies have shown that STAT3 is a substrate protein of HDAC3 and hence increases STAT3 phosphorylation and activation in multiple cancers [40][41][42]. We therefore examined the impact of HDAC3 activity on STAT3 acetylation. However, inconsistently with previous studies in cancer cells, we did not observe the downregulation of STAT3 acetylation after STAT3-HDAC3 complex formation in human primary chondrocytes in response to leptin. We propose the possibility that their interaction might serve to bring HDAC3 to the promoter of the BMP-2 gene, where it could modify chromatin. Recently, some evidence has also indicated that HDAC inhibitor suppression of JAK2/STAT3 signaling in cancer cells occurs through inducing the acetylation of histones connected to the promoter of SOCS3, a potent feedback inhibitor of JAK2/STAT signaling [43][44]. Thus, the question of whether HDAC4 and the HDAC3/STAT3 complex regulate the deacetylation of other transcriptional cofactors and/or chromatin factors and consequently mediate leptin-induced BMP-2 expression in human chondrocytes will require ongoing study.
The present study has reported that leptin increases BMP-2 expression to initiate a potent reparative effect when triggering catabolic damage in human chondrocytes. These results provide new insights into the understanding of the complexities of leptin-initiated OA development. Accumulating studies indicate that leptin and its associated signaling may serve as promising novel drug targets and appear to be potential strategies in the therapy of OA, especially in obese patients [17,45]. Our data concerning the anabolic activity of leptin may contribute new ideas to drug design and development for treating OA progression. Moreover, our findings also suggest the possible clinical application of BMP-2 and should be taken into account in OA therapy in the future.