CXCL8 and CCL20 Enhance Osteoclastogenesis via Modulation of Cytokine Production by Human Primary Osteoblasts

Generalized osteoporosis is common in patients with inflammatory diseases, possibly because of circulating inflammatory factors that affect osteoblast and osteoclast formation and activity. Serum levels of the inflammatory factors CXCL8 and CCL20 are elevated in rheumatoid arthritis, but whether these factors affect bone metabolism is unknown. We hypothesized that CXCL8 and CCL20 decrease osteoblast proliferation and differentiation, and enhance osteoblast-mediated osteoclast formation and activity. Human primary osteoblasts were cultured with or without CXCL8 (2–200 pg/ml) or CCL20 (5–500 pg/ml) for 14 days. Osteoblast proliferation and gene expression of matrix proteins and cytokines were analyzed. Osteoclast precursors were cultured with CXCL8 (200 pg/ml) and CCL20 (500 pg/ml), or with conditioned medium (CM) from CXCL8 and CCL20-treated osteoblasts with or without IL-6 inhibitor. After 3 weeks osteoclast formation and activity were determined. CXCL8 (200 pg/ml) and CCL20 (500 pg/ml) enhanced mRNA expression of KI67 (2.5–2.7-fold), ALP (1.6–1.7-fold), and IL-6 protein production (1.3–1.6-fold) by osteoblasts. CXCL8-CM enhanced the number of osteoclasts with 3–5 nuclei (1.7-fold), and with >5 nuclei (3-fold). CCL20-CM enhanced the number of osteoclasts with 3–5 nuclei (1.3-fold), and with >5 nuclei (2.8-fold). IL-6 inhibition reduced the stimulatory effect of CXCL8-CM and CCL20-CM on formation of osteoclasts. In conclusion, CXCL8 and CCL20 did not decrease osteoblast proliferation or gene expression of matrix proteins. CXCL8 and CCL20 did not directly affect osteoclastogenesis. However, CXCL8 and CCL20 enhanced osteoblast-mediated osteoclastogenesis, partly via IL-6 production, suggesting that CXCL8 and CCL20 may contribute to osteoporosis in rheumatoid arthritis by affecting bone cell communication.


Osteoblast Culture
Trabecular bone samples (surgical waste) from 4 female and 3 male donors (63.3±7.8 yrs; range 53-75 yrs) were obtained from the iliac crest during sinus floor elevation surgery using autologous bone graft. Serum C-reactive protein (CRP) levels of all donors was <2.5 mg/l, indicating no inflammatory disease. The protocol was approved by the Ethical Review Board of the VU University Medical Center and all subjects gave written informed consent.

Osteoblast Proliferation
Proliferation of osteoblasts was tested in cells seeded at 2.5x10 3 cells/well of 96-well culture plates (Greiner Bio-One). The following day, medium was replaced by fresh medium with or without CXCL8, CCL20, CXCL8+CCL20, or TNF-α. Cell proliferation was determined after 3, 5, and 7 days of culture using a Cell Proliferation Kit II (XTT; Roche, Mannhelm, Germany).
The total DNA content of the cell layer was quantified using a Cyquant Cell Proliferation Assay (Molecular Probes, Eugene, OR).
Procollagen Type 1 Amino-terminal Propeptide (P1NP) and Alkaline Phosphatase (ALP) Activity P1NP in CM and ALP in cell lysate of osteoblast cultures were measured as described earlier [25].

RNA Isolation and Real-time RT-PCR
Total RNA of primary osteoblasts was isolated using an RNeasy Micro kit with an on-column DNase I digestion (Qiagen, Basel, Switzerland). Total RNA concentrations were measured with a Nanodrop spectrophotometer (Nanodrop Technologies, Wilmington, DE). cDNA synthesis was performed in a thermocycler GeneAmp PCR System 9700 PE (Applied Biosystems, Foster City, CA), using a SuperScript VILO cDNA Synthesis Kit (LifeTechnologies, Inchinnan, UK), with 0.1 μg of total RNA in 20 μl reaction mixture consisting of VILO Reaction Mix and Super-Script Enzyme Mix.
Real-time PCR reactions were performed using 2.5 μl cDNA and SYBR Green Supermix (Roche Laboratories, Indianapolis, IN) in a LightCycler (Roche Diagnostics). In each PCR run, the reaction mixture without cDNA was used as a negative control. For quantitative real-time PCR, the values of relative target gene expression were normalized for relative mean of YWHAZ and HPRT housekeeping gene expression. Real-time PCR was used to assess expression of the following genes: KI67, collagen 1 (COL1), ALP, osteopontin (OPN), osteocalcin (OCN), macrophage colony stimulating factor (MCSF), RANKL, OPG, IL-1β, IL-6, IL-17, CXCL8, CCL20, and cysteine rich protein 61 (CYR61). All primers used were from Life-Technologies. The primer sequences are listed in S1 Table. IL-6 Protein Quantification CM was collected from osteoblasts after 14 days of culture in the presence or absence of chemokines, and IL-6 protein was quantified using a PeliKine human IL-6 ELISA kit (Sanquin Blood Supply, Amsterdam, Netherlands).

Osteoclastic Bone Resorption
To quantify osteoclast activity, resorption pits in bone slices were visualized and counted as described earlier [27]. The resorbed area was measured using Image Pro-Plus Software (Media Cybernetics, Silver Spring, MD) and expressed as percentage of total bone surface area. The percentage of bone resorption was expressed per number of TRACP-positive multinucleated cells.

Statistical Analysis
Data on KI67 gene expression, TRACP-positive multinucleated osteoclast number, and osteoclastic bone resorption was expressed as mean±SEM. Data on expression of other genes analyzed and IL-6 protein production were expressed as median with 5-95 percentile range of the treatment-over-control ratio. Differences in gene expression and IL-6 protein production between chemokine-treated and untreated control cultures were tested using the Wilcoxon Signed Rank test. The effect of control-CM or chemokine-CM on osteoclastic bone resorption, and the effect of control-CM or chemokine-CM and CM+IL-6 inhibitor on osteoclast formation was tested using ANOVA followed by Bonferroni's Multiple Comparison test. Differences were considered significant if p<0.05. Statistical analysis was performed using GraphPad Prism 5.01 (GraphPad Software, Inc., La Jolla, CA, USA).

Discussion
In this study we analyzed whether chemokines potentially play a role in the emergence of osteoporosis in inflammatory diseases by studying their effect on osteoblast function and osteoblast-mediated osteoclast formation and activity. We found that CXCL8 and CCL20 did not inhibit osteoblast proliferation nor gene expression of the main matrix proteins COL1, OPN, and OCN. CXCL8 and CCL20 enhanced osteoblast-mediated osteoclastogenesis, partly via IL-6 production, suggesting that CXCL8 and CCL20 may contribute to generalized osteoporosis in RA by affecting bone cell communication.
The osteoblastic nature of human primary osteoblasts, as well as their osteocyte-like behavior has been shown earlier [28][29][30][31]. We found that these cells express functional chemokine receptors CXCR1 and CCR6, but not CXCR3, which is the receptor for CXCL9 and CXCL10. Lisignoli and colleagues reported that osteoblasts express functional CXCR3 [18]. This discrepancy in CXCR3 expression might be related to differences in anatomical location of the bone biopsies; we used biopsies from the iliac crest, and Lisignoli and colleagues from the tibial plateau [18].
Both CXCL8 and CCL20 enhanced gene expression of the cell proliferation marker KI67 in osteoblasts. However we could not detect an effect of chemokines on osteoblast proliferation using the XTT assay or by measuring the total DNA content, which is in contrast with data published by others showing that CCL20 enhances proliferation of osteoblasts from RA pg/ml), CCL20 (500 pg/ml), CXCL8 +CCL20 (20 pg/ml+50 pg/ml), and TNF-α (100 ng/ml) enhanced IL-6 production by osteoblasts. Values are median with 5-95 percentile range of treatmentover-control ratios from 3 experiments, n = 9. Significant effect of chemokines and TNF-α, *p<0.05, **p<0.01. doi:10.1371/journal.pone.0131041.g003 CXCL8 and CCL20 Enhance Osteoclastogenesis patients [19]. This difference might be due to the fact that we obtained cells from patients without any signs of systemic inflammation, while Lisignoli and colleagues obtained cells from RA patients [19]. Our findings indicate that CXCL8 and CCL20 do not affect osteoblast proliferation.
Since proinflammatory cytokines inhibit osteoblast function [14,[32][33][34] we expected that CXCL8 or CCL20 would also inhibit osteoblast function. However we found that CXCL8 and CCL20 enhanced gene expression of some early osteogenic differentiation-related markers but did not affect P1NP production, ALP activity, nor gene expression of late osteogenic differentiation-related markers. CXCL8 and CCL20 increased ALP and COL1 gene expression, but not P1NP and ALP activity, which might be due to altered post-transcriptional processing. In any case, our data suggests that CXCL8 and CCL20 do not negatively affect osteoblast function in vitro.
CXCL8 and CCL20 enhanced IL-6 gene expression and protein production by osteoblasts. CXCL8 and CCL20 did not directly affect osteoclastogenesis, which is in accordance with findings by others [19,20]. Here we show that CM from osteoblasts cultured for 14 days (control-CM; without chemokines) enhanced osteoclastogenesis in vitro, indicating that osteoblasts produce factors essential for osteoclast formation. Furthermore, our study reveals that CXCL8-CM, CCL20-CM, and TNF-α-CM enhanced osteoclastogenesis. CCL20-CM, but not CXCL8-CM, enhanced osteoclast activity. Osteoclast activity is not only regulated by the number of osteoclasts and tartrate-resistant acid phosphatase present in osteoclasts, but also by a number of other regulators such as cathepsin K, lysophosphatidic acid receptor type 1 (LPA1), lysosome associated membrane protein-2, and chloride channels CIC3 and CIC7 [35][36][37][38]. Osteoclast activity regulators might present differently in osteoclasts formed in the presence of CXCL8-CM compared with CCL20-CM, which might cause the difference in osteoclastic bone resorbing activity. We also found that CXCL8 and CCL20 enhanced IL-6 gene expression and protein production by osteoblasts. Moreover IL-6 production was strongly enhanced by TNFα treatment; a similar finding has been reported by Chaudhary et al. [9]. Inhibition of IL-6 robustly reduced the effect of CXCL8-CM, CCL20-CM, and TNF-α-CM on osteoclastogenesis. CCL20 has been suggested to play a role in the pathogenesis of rheumatoid arthritis [39]. IL-6 is one of the most potent stimulators of osteoclastic bone resorption and central to the pathogenesis of generalized osteoporosis in RA [40][41][42]. This corroborates with our data showing a stimulatory effect of CXCL8 and CCL20 on IL-6 production by osteoblasts, which enhanced osteoclastogenesis. CXCL8 and CCL20 might not only affect IL-6 production, but also the production of a whole cocktail of factors by osteoblasts. Therefore it is unlikely that IL-6 inhibition alone will completely block the effect of CXCL8 and/or CCL20 on osteoblast-mediated osteoclastogenesis. RANK, RANKL, and OPG are known key molecules involved in osteoclast formation and function [43]. In this study, RANKL gene expression by osteoblasts was below the detection limit, and therefore we did not test the effect of RANKL and OPG signaling molecules produced by osteoblasts on osteoclast formation. Our model uses exogenous recombinant RANKL to allow osteoclast formation, since osteoclast formation does not occur in the absence of RANKL, which created further limitation to analyze the role of RANKL and OPG in  CXCL8 and CCL20 Enhance Osteoclastogenesis osteoclastogenesis. Detailed information on our model is provided in S1 Fig. Based on our findings, we created a pathophysiological model illustrating how CXCL8 and CCL20 might influence bone remodeling in inflammatory conditions and contribute to osteoporosis (Fig 6).
CXCL8 and CCL20 are abundantly present in synovial tissue, synovial fluid, and serum in RA, and their levels correlate with disease activity [8,20,44,45]. Blockade of CXCL8 reduces inflammation in a murine RA model [46]. A CXCR1 antagonist has been shown to decrease clinical disease scores in a murine collagen-induced arthritis model [47]. A polymorphism of the CCR6 gene has been associated with RA susceptibility [48]. Moreover CCL20 has been suggested as an emerging player in the pathogenesis of rheumatoid arthritis [39]. Our study provides insight in the mechanism of action of CXCL8 and CCL20 with regard to the regulation of bone metabolism. Future studies in which CXCL8 and CCL20 are blocked could confirm whether CXCL8 and/or CCL20 might be new targets to prevent bone loss in inflammatory diseases such as RA. The combination of CXCL8 and CCL20 had a similar effect on osteoblastmediated osteoclastogenesis as the individual CXCL8 and CCL20 at a 10-fold higher concentration. This indicates that the effect of CXCL8 and CCL20 on bone loss in vivo where combinations of cytokines and chemokines are present in the circulation, might be more pronounced than in vitro. A limitation of this study might be the relatively low number of patients included. Statistical significance between groups was not easily obtained due to the fairly large data variation, probably due to donor variation. Another limitation of this study is that the addition of exogenous recombinant RANKL (4 ng/ml) is essential for osteoclastogenesis in our experimental set up. This prevents analysis of the role of chemokine-induced osteoblast-mediated RANKL/OPG signaling on osteoclastogenesis. However, osteoblastic RANKL gene expression was below the detection level, even after treatment with CXCL8 or CCL20. Furthermore osteoblastic OPG expression was not affected by CXCL8 or CCL20. In this study IL-6 was a key molecule produced by osteoblasts as a result of CXCL8 and/or CCL20 treatment, which enhanced osteoclastogenesis and this effect of IL-6 was reduced by blocking IL-6.
In conclusion, our results indicate that CXCL8 and CCL20 did not significantly inhibit osteoblast proliferation and function, nor directly enhanced osteoclastogenesis. However, CXCL8 and CCL20 strongly enhanced osteoblast-mediated osteoclastogenesis, which seems partially mediated by CXCL8 and CCL20-induced IL-6 production by osteoblasts. Based on these findings, we speculate that CXCL8 and CCL20 may play a role in generalized osteoporosis during systemic inflammation in which serum levels of CXCL8 and CCL20 are elevated.