Cross-talk between primary osteocytes and bone marrow macrophages for osteoclastogenesis upon collagen treatment

Homeostasis of osteoclast formation from bone marrow macrophages (BMM) is regulated by paracrine signals of the neighbourhood bone cells particularly mesenchymal stem cells (MSC), osteoblasts and osteocytes (OC). Besides paracrine cues, collagen and glycosaminoglycan are involved in controlling bone homeostasis. Towards this approach, different molecular weight collagens were reacted with MSC, OC and BMM to understand the bone homeostasis activity of collagen. The up-regulating effect of collagens on osteogenic cell growth was confirmed by the presence of mineralized nodules in the osteoblastogenic lineage cells and increased osteogenic stimulatory gene expression. The decreased BMM-derived TRAP+ osteoclasts number and osteoclastogenic regulatory gene expression of OC could demonstrate the exploitive osteoclastogenic activity of collagens. Osteoclastogenesis from BMM was triggered by paracrine cues of OC in some extend, but it was down-regulated by collagen. Overall, the effect of collagen on osteoclastogenesis and osteoblastogenesis may depend on the molecular weight of collagens, and collagen suppresses osteoclastogenesis, at least in part by downregulating the secretion of cytokines in OC.


Results and Discussion
Effect of collagen on osteocytes. The result of proliferation assay is given in Fig. 1. On day 4, type II collagen (CII) and 57 kDa type II collagen polypeptide (57 kDa) treated groups (50 μg/ml) had a greater number of osteocytes than control, however, there were no significant differences observed between treated and control group on day 7. Similarly, rat tail type I collagen (150 μg/ml) coated osteocyte-like cell lines (MLO-Y4 cells) found to be an effective way to maintain the growth of osteocyte-like phenotype 9 . In another study, primary osteocytes culture condition was optimized using type I collagen matrix with 0.2% FBS 10 .
In general, osteocytes would not produce a high level of ALP, therefore, as an exploratory study, we estimated ALP activity to understand the regulatory mechanism of collagen and polypeptides in osteocytes. The results showed that the ALP level was elevated more in CII (50 μg/ml), 40 kDa type II collagen polypeptide (40 kDa) (1 μg/ml) and 57 kDa polypeptide treated groups than control on day 7 (p < 0.05) (Fig. 1). On day 4, there were no changes on ALP level in the treated groups. ALP activity is considered as an important factor in determining bone cell differentiation and mineralization and is used as a biochemical marker for determining mature bone cell phenotype. Therefore, an increased level of ALP in osteogenic cells would indicate increased osteocytes activity. As shown in Fig. 1, ALP activity was elevated with increasing concentrations of collagen and its polypeptides from 1 to 50 μg/ml compared to control. Similar to the present findings, ALP activity of murine osteoblasts was significantly increased in the presence of 1 mg/mL porcine collagen hydrolysates 7 . Collectively, the above findings support an osteogenic capability of type-II collagens and also osteocytes might secrete ALP in some extent. mRNA expression. mRNA levels of genes of interest in osteocytes treated with collagen and polypeptides are shown in Fig. 2. Col2a1 mRNA levels were not significantly altered by either collagen or polypeptides treatment, except high dose (50 μg/ml) of polypeptides at 1 h treatment (p < 0.05) and, low dose of CII (1 μg/ml) and 57 kDa (1 and 10 μg/ml) had increased the level of Col2a1 mRNA at 6 h. Il6ra mRNA levels were increased between 1 h and 6 h in untreated wells (p < 0.05). At 1 h and 6 h, Il6ra mRNA levels decreased with increasing CII and 57 kDa (50 μg/ml) (p < 0.05). It was proved that local stimulatory factors IL-6 and IL-6ra had predictive for osteoclast maturation, activation, and recruitment. Several studies demonstrated the higher level of IL-6ra responsible for bone loss in early post-menopause 11,12 .
Tumor necrosis factor receptor superfamily member 11B (Tnfrsf11b) mRNA levels were not significantly altered by either collagen or polypeptide treatment in 1 h. At 6 h, Tnfrsf11b mRNA levels were increased with increasing collagen (10 and 50 μg/ml) and 57 kDa (50 μg/ml) (p < 0.05). There was no significant difference in Tnfrsf11b mRNA levels between controls and collagen polypeptides (1 μg/ml and 10 μg/ml)-treated cells.
Tnfrsf11b gene provides instruction for making a protein osteoprotegerin, which plays a vital role in bone remodeling and regulation of osteoclast cells. Tnfrsf11b acts as a decoy receptor for the receptor activator of nuclear factor kappa B ligand (RANKL), a major cytokine for osteoclastogenesis 13 . Therefore, higher expression of Tnfrsf11b mRNA level by collagen and 57 kDa should reduce the osteoclast formation through osteocytes signaling mechanism.
Tumor necrosis factor ligand superfamily member 11 (Tnfsf11), also known as RANKL is a type-II membrane protein and has been identified to affect the immune system and control bone regeneration and remodeling. At 1 h and 6 h, Tnfsf11 mRNA levels were decreased with increasing collagen and polypeptides (57 kDa and 40 kDa) concentration (50 μg/ml) (p < 0.05). There is strong evidence that osteocytes produce paracrine signals and regulate osteoclast formation through RANKL production 3 . The present study clearly depicted that treatment of collagen had obstructed the mRNA expression of RANKL in osteocytes.
Sost mRNA provides an instruction for producing a protein called sclerostin. It is produced in osteocytes and acts on the bone to stop bone formation. At 1 h, Sost mRNA levels were decreased with increasing collagen and polypeptides concentration (50 μg/ml) when compared to control cells (p < 0.05). There was no significant difference in Sost mRNA levels by collagen and polypeptides in 6 h. From the data, collagen had down-regulate bone formation inhibitor gene, i.e., sost mRNA expression, which further supported the proliferation effect of collagen and polypeptides on osteocytes (Fig. 1). The recent study reported the decreased formation of the mineralized matrix and an increased mRNA expression of Sost in osteocytes, indicating a reduction in bone formation via the NF-κB pathway 14 .
Alpl is the gene for the alkaline phosphatase (ALP) enzyme, which helps in the growth and development of bones through bone mineralization. This enzyme helps for the deposition of calcium and phosphorus in developing bones 15 . Alpl mRNA levels were increased significantly between 1 h and 6 h in untreated cells (p < 0.05). Except for CII, 57 kDa and 40 kDa (10 μg/ml)-treated cells, there was no significant difference in Alpl mRNA levels between controls and collagen/polypeptides-treated cells in 1 h. At 6 h, Alpl mRNA levels decreased in collagen/polypeptides-treated cells (p < 0.05). Though, the proliferation rate and cellular alkaline phosphatase level of collagen treated osteocytes increased, the level of Alpl mRNA expression was not triggered at 6 h treatment. Several hypotheses could be possible for the above result because in general, osteocytes are not the major sources for ALP productions for bone formation. The earlier study could demonstrate that an induction of Alpl mRNA during the in vitro osteogenic differentiation process expected for bone forming capacity of the bone marrow stromal cells in vivo 16 .
Taken together, these findings suggest that the expression of osteogenic regulatory genes such as Col2a1 and Alpl, was triggered initially at 1 h and declined later at 6 h in collagens treated osteocytes; and osteoclasts regulatory genes, such as Il6ra, and Tnfsf11 were suppressed in collagen and polypeptides treated osteocytes that could predict to down-regulate osteocytes cues for osteoclast formation. It was earlier confirmed by few authors that certain specific amino acid residues (asparagine, glutamine, glycine, and alanine) of collagen interact with integrin α 2 β 1 on the mesenchymal stem cell membrane and can lead to up-regulate bone matrix synthesis through inhibition of TGF-beta signaling and activation of RUNX2 through FAK-JNK signaling 7,17,18 . Effect of collagen on MMSC-bm. MMSC-bm activity contains three important phases including proliferation, matrix protein synthesis and mineralization of the bone matrix during new bone formation. The result of proliferation assay of MMSC-bm cells treated with different concentration of collagen and polypeptides is given in Fig. 3. In general, the proliferation rate of MMSC-bm cells was accelerated with increasing concentration of collagen and polypeptides. Notably, the rate was significantly higher on day 7 at 50 μg/ml of CII than other treatments. Similarly, Hennessy et al. and Gao et al. disclosed the osteogenic stimulatory activities of type I collagen Collagen treated MMSC-bm cells cultured with osteoblast medium containing osteoblast supplements for 21 days and stained with Alizarin red. Control and Negative control cells were cultured in osteoblastic medium with and without osteoblast supplements, respectively. CII-type II collagen, 57 K and 40 K-57 and 40 kDa molecular weight collagen polypeptide, respectively. Bars with different alphabets (a and b) are significantly different (P < 0.05) among CII and its polypeptides in their respective concentration. *P < 0.05, vs. control (MMSC-bm cells without collagen).
peptides and rat tail type I collagen on mesenchymal stem cell 19,20 . In contrary, Song et al. observed decreasing trend of human bone marrow mesenchymal stem cells with coating and/or direct treatment of human type I collagen (5 μg/ml) 21 . Vleggeert-Lankamp et al. reported that bovine dermis type I collagen did not promote/ inhibit the proliferation of human Schwann cells 22 . In some studies, type I collagen was shown having the effect to increase cell proliferation, whereas, in some others, it did not have such effect, even decreased the proliferation of MSC cells. We speculated that this could be due to different molecular composition of collagens from different sources. Recently, Chiu et al. 17 reported that the level of integrin α 2 β 1 complex (VLA-2) expression on MSC-bm surface increased by type-II collagen treatment on day 4 and then gradually reduced from day 6 to a lower level than that of the control on day 10.
Collagens and polypeptides treated MMSC-bm cells cultured in osteogenic media were able to differentiate into the osteoblastogenic lineage cells. This was confirmed by the presence of mineralized nodules in collagens treated MMSC-bm cells after Alizarin red staining on day 21 (Fig. 3). In collagens untreated cultures smaller aggregates were observed. There was no nodular aggregate deposited in negative control cultures (cells cultured without osteoblast supplements). This was clearly demonstrating that collagens and polypeptides accelerated osteoblast lineage cells differentiation from MMSC-bm and nodular aggregates present in collagens treated cultures were calcium deposits 23 . This is consistent with previous work showing that bone marrow-derived mesenchymal stem cells (BM-MSC) cultured on mammalian type-II collagen coated plates exhibited significant higher calcium deposition on day 12 and day 16, which suggests osteogenic induction properties of type-II collagen 17 . Collagen treated cultures had larger alizarin red positive aggregates and stained more intensively, indicating that a more extensive calcium deposition had occurred. The presence of alizarin red stain with nodular cell aggregates observed in collagen and polypeptides treated cultures establishes that these amorphous deposits contain calcium and suggests that calcium deposits are made up of MMSC-bm cells differentiated to the osteoblastic lineage. Numerous possible mechanisms have been proposed for stimulating MMSC-bm osteogenic differentiation 24,25 . Actual signaling mechanism of type II collagen during early osteogenic differentiation of MMSC-bm occurred by facilitating RUNX2 phosphorylation activation through integrin α 2 β 1 -FAK-JNK signaling axis, thus enhanced bone defect repair through an endochondral ossification-like process 17 .
Effect of collagen on osteoclast differentiation. Macrophages harvested from mouse bone marrow were treated with RANKL and mCSF in the presence and absence of different concentration of collagen and polypeptides to assess the effects of the collagens on osteoclastogenesis. The result shows that increasing concentration of collagen and its polypeptides reduced osteoclast numbers considerably (p < 0.01) (Fig. 4). In contrast, the murine oncostatin M (m-OSM) treated group increased osteoclast cell numbers (3+ nuclei) compared to control. But, the number of double nuclei osteoclast cells was decreased in m-OSM treated group than control (p < 0.01). The higher concentration of collagen and polypeptides treated groups had considerably decreased both double nuclei and 3+ nuclei TRAP+ cells compared to control (p < 0.01). Similarly, Guillermin et al. 7 reported that murine osteoclasts differentiation reduced by the treatment of porcine collagen hydrolysates (2 or 5 kDa) with 1 mg/ml concentration, which might be due to inhibition of Transforming growth factor beta (TGF-β) through interaction of collagen-derived peptides (asparagine, glycine, glutamine and alanine).
The micrograph clearly depicted the lower number of osteoclast cells in the presence of higher concentration (50 μg/ml) of collagen and its polypeptides (Fig. 5). Among the treatment, 1 μg of 57 kDa collagen polypeptide-treated group had a high number of osteoclast cells than other treatment groups. In higher magnification (10×), "ghost" outlines remaining after osteoclast apoptosis cells were found in wells treated with the higher doses of CII but not in lower dose treatment. This observation indicated that increasing concentration of collagen increased apoptosis in osteoclasts. Co-culture model. It was reported that differentiated primary osteoblasts express cytokines, RANKL and mCSF in presence of stimulators (IL-6, soluble IL-6 Receptor, prostaglandin E2 (PGE2), 1,25-Dihydroxy vitamin D3 (1,25 (OH) 2 D 3 ) and oncostatin M), which bind with mouse BMM and support osteoclast differentiation 24 . In contrast, Nakashima et al. 3 identified that purified osteocytes also express a much higher amount of RANKL and have a greater capacity to support osteoclastogenesis. Therefore, in the present study, we used a co-culture model with different stimulators to understand the effect of inducers such as m-OSM, IL-6, sIL-6R, 1,25 (OH) 2 D 3 and PGE2 on osteoclast formation. The primary co-culture of BMM with pOC was grown in presence of inducers for 7 days. m-OSM significantly increased the number of single and double nuclei TRAP-positive cells, but few cells with more than 2 nuclei were observed. Compared to m-OSM, less number of TRAP-positive cells found in 1,25 (OH) 2 D 3 and PGE2 group. In contrast, IL-6 in the presence of sIL6R supported only the production of TRAP-positive single and double nuclei cells to some extent (Fig. 6). No TRAP-positive cells were observed in wells treated with IL-6 or sIL-6R alone; this can also be seen in the micrographs (Fig. 6). Our above study confirmed that osteocytes stimulated with m-OSM might support osteoclast differentiation to the point of a multinucleated cell, as previously observed with other stimuli in freshly isolated osteocytes 26 . Further, we aimed to understand whether collagen and its polypeptide might play a similar role in the presence and absence of inducer, OSM in the co-culture of BMM and pOC. TRAP-positive single and double nuclei cells were formed in the collagens-OSM treated group (Fig. 7), but it was lower than m-OSM alone treated group (Fig. 6), which confirmed that collagen and collagen polypeptides might suppress m-OSM-stimulation of osteocytes towards osteoclast formation to some extent. In contrast, others reported no effect of hydrolyzed collagen on primary co-culture of osteoblast and osteoclasts growth 7 . The down-regulation of osteoclastogenesis by collagen might be justified by the higher level of Tnfrsf11b mRNA (a decoy receptor for RANKL) (Fig. 2) that coupled with m-OSM-triggered-RANKL and leads to inhibition of RANK-RANKL ligand binding towards osteoclast formation.
The level of inhibition was higher (p < 0.05) in a collagen-OSM group than polypeptides treated groups, which might be due to the suppressive mechanism of osteoclastogenic regulatory genes such as Il6ra and Tnfsf11 expression of pOC by collagen (Fig. 2). Thus, pOC could not secrete necessary cytokines in order to stimulate osteoclast formation from collagen treated BMM-pOC co-culture, though the presence of positive inducer, m-OSM.

Effect of collagen and collagen polypeptides on MMSC-bm and osteocytes.
Cell culture. Primary osteocytes (pOC) cells were harvested from mice 27 (Sino-British Sippr/BK Lab Animal Co., Ltd, Shanghai, China) long bones (femora, and tibia) using 300 U/mL collagenase (Sigma-Aldrich) dissolved in α-minimal essential medium (α-MEM) (Gibco, Shanghai, China) and were cultured in α-MEM medium supplemented with 10% fetal bovine serum (Gibco) at 37 °C in a CO 2 incubator (Shanghai Hengyue Medical Instruments Co., Ltd, Shanghai, China), respectively. Animal study protocols and procedures were approved by the Shanghai Ocean University institutional animal care and use committee (Permit Number: 13-0012). All methods were employed in accordance with the relevant guidelines and regulations of Scientific and Ethical Care and Use of Laboratory Animals of Shanghai Ocean University. Proliferation assay. After confluence, cells pOC and MMSC-bm at passage 5 were seeded (5 × 10 5 cells/well/48 well plate) in microtiter plates (Costar, Shanghai) along with the different concentration of collagen and polypeptides (1, 10 and 50 μg/ml). Controls consisted of uncoated (without collagen) wells. The total number of viable cells was counted using an Invitrogen cell counter (Countess II Automated Cell Counter, ThermoFisher Scientific, Shanghai, China) at 0, 4 and 7 days after seeding.
Alkaline phosphatase (ALP) assay. pOC cells were generated as described above. At each time point, cells were washed twice with 1 × phosphate buffered saline (PBS), harvested with lysis buffer (10 mM Tris buffer, pH 7.4) and sonicated (40 w, Shanghai Kedao Ultrasonic Instrument Co, Shanghai, China) with an ice water-bath for 30 sec to release cellular components. After a brief centrifugation (13,000 × g for 1 min at 4 °C, Universal 320 R, Andreas-Hettich, Buckinghamshire, Germany), the supernatant was used for the ALP assay. In brief, 25 µl per sample was added to 100 µl substrate (10 mM p-nitrophenyl phosphate (Sigma-Aldrich) in assay buffer (0.1 M Na 2 CO 4 buffer, pH 10.0)). The reaction proceeded for 30 min at 37 °C; the reaction was terminated by addition of 50 µl NaOH (1 M) and absorbance measured at 410 nm using a plate reader (Bio-Rad Model 550, Shanghai). ALP activity was determined from a standard curve using 10 mM p-nitrophenol from 4000 to 62.5 nmoles and without substrate maintained as a control. The same volume of sample was used to determine protein content using bicinchoninic acid (BCA) as per the manufacturer's instructions (Pierce, IL, USA). ALP activity was expressed as nmoles/min/mg protein determined as follows:  Real-time polymerase chain reaction (RT-PCR) was done using a 96-well plate using an ABI 7500 Fast Real-Time PCR System (Applied Biosystems, Shanghai, China) using SYBR Green Fast qPCR RT Master Mix (Invitrogen, Shanghai, China). pOC regulatory primers designed using Primer-BLAST (Shanghai Biotechnology Co., Ltd, Shanghai, China) were listed in Table 1. The total volume of each PCR reaction was 10 μl, containing 5 μl SYBR Green Fast qPCR RT Master mix, 1.5 μl cDNA template sample, 0.5 μl of forward and reverse primers and 2.5 μl water. The PCR reaction was carried out at 95 °C for 30 min, 40 cycles at 95 °C for 5 min, 60 °C for 30 min, one cycle of 95 °C for 1 h, 55 °C for 30 min and 95 °C for 30 min.
Osteogenic differentiation and quantification. MMSC-bm cells at passage 5 were initially seeded at a density of 5 × 10 4 cells/well in microtiter 6 well plates. After 2 h, 50 μg/ml collagen or collagen polypeptides were added into each well. To induce osteogenesis, the cells were grown in osteoblast medium (Shanghai Zhong Qiao Xin Zhou Biotechnology Co., Ltd, Cat. No. 4601) with the addition of osteoblast growth supplement (ObGS) (Shanghai Zhong Qiao Xin Zhou Biotechnology Co., Ltd, Cat. No. 4652) composed of 100 nM dexamethasone, 10 mM b-glycerolphosphate, and 0.05 mM 2-phosphate-ascorbic acid for 21 days with media changes every 3 days. Control and negative control cells were grown in the culture medium without sample and ObGS, respectively. The effect of collagen on osteogenesis was confirmed by observing Ca deposition in cultured cells using the Alizarin red staining method 28 . Briefly, cells were washed with PBS and fixed in 4% paraformaldehyde for 30 min and stained with 1% alizarin red (Sinopharm Chemical Reagent Co., Ltd, Shanghai, China) containing 1% NH 4 OH.

Conclusion
In summary, the in-vitro results obtained with pOC, MMSC-bm and osteoclast cells demonstrated that type-II collagen and its polypeptides were able to stimulate osteogenesis and supress osteoclastogenesis. The effect of collagen and its polypeptides in osteocytes and MMSC-bm cell growth is dose-dependent, which ultimately maintained the bone formation. Osteoclast numbers were reduced by the concentration of collagen and its polypeptides. Intriguingly, a higher concentration of type-II collagen and polypeptides treatment had up-regulated the osteoblast lineage cell differentiation from MMSC-bm and had down-regulated the BMM-derived osteoclast cell formation, suggesting that collagen or collagen hydrolysate could be helpful in the management of bone diseases like osteoporosis. There are numerous reports disclosing the chondrogenic properties of type II collagen for cartilage repair, but here we are reporting that not only chondrogenesis, type II collagen might also play some role in ossification through modulating osteocytes paracrine signals for osteoclastogenesis. Therefore, the present work concluded that type II collagen and collagen polypeptides isolated from shark cartilages might be conceivable novel biomaterials to modulate bone formation and resorption activity; and could be of potential curiosity as a nutritional supplement in the preclusion of bone loss in osteoporosis. This work also highlighted that in some parameters native collagen was more efficient than that observed in collagen polypeptides. However, one important issue that needs to be addressed by further studies is the concern on the molecular interaction of collagen with osteocytes in order to maintain bone homeostasis.
Data Availability. The datasets generated and analyzed during the present study are available from the corresponding author upon reasonable request.