Collagen Extract Derived from Yeonsan Ogye Chicken Increases Bone Microarchitecture by Suppressing the RANKL/OPG Ratio via the JNK Signaling Pathway

Yeonsan Ogye is a traditional Korean chicken breed (Gallus domesticus, GD), with a dominant gene for fibromelanosis, showing entirely black fluffy head feathers, ear lobes, and pupils. GD collagen extract (78.6 g per 100 g total protein) was derived from the flesh of Yeonsan Ogye. The effects of GD collagen on bone mass, microarchitecture, osteogenic, osteoclastogenic differentiations, and function factor expression were investigated in ovariectomized (OVX) rats. GD collagen stimulated osteogenesis in OVX rats and increased tibial bone strength and calcium content. Micro-computed tomography analysis of tibia cross-sections revealed that GD collagen attenuated the OVX-induced changes in trabecular thickness, spacing, and number. GD collagen stimulated alkaline phosphatase activity, bone-specific matrix proteins (alkaline phosphatase (ALP), osteocalcin, collagen type I (COL-I)) and mineralization by activating bone morphogenetic protein 2 (BMP-2)/mothers against decapentaplegic homolog 5 (SMAD5)/runt-related transcription factor 2 (Runx2). GD collagen inhibited osteoclast differentiation and function gene markers (TRAP, cathepsin K) by interfering with the Wnt signaling, increasing OPG production, and reducing the expression of RANKL, TRAP, and cathepsin K. GD collagen promoted osteogenesis by activating the p38 signal pathway and prevented osteoclastogenesis by lowering the RANKL/OPG ratio and blocking the JNK signaling pathway. Dietary supplementation with GD collagen might inhibit osteoclastogenesis, stimulate osteoblastogenesis, and regulate bone metabolism.


Introduction
Bone is a highly mineralized connective tissue that continually undergoes resorption of the bone matrix by osteoclasts and deposition of the osteoid matrix (containing~90% of collagen type I) by osteoblasts to maintain the skeleton's structural integrity [1,2]. Imbalance triggered by increasing bone matrix resorption, without corresponding new bone tissue formation, can lead to bone loss, quality deterioration, microarchitecture, and osteoporosis [3]. Traditional first-line prevention and treatment of osteoporotic fractures, such as hormone replacement therapy, have several side effects in chronic use [4,5]. Therapeutic agents in osteoporosis, such as bisphosphonates and modulators of estrogen and estrogen receptors, delay bone loss by inhibiting osteoclast activity or differentiation; however, they showed

Analysis of ALP Activity
ALP activity was measured as the rate of p-nitrophenyl phosphate (pNPP) hydrolysis. MG-63 cells (5 × 10 4 ) were seeded in 96-well plates, incubated in RPMI medium for 24 h, and then treated with the indicated concentrations of GD collagen for 96 h. The cells were lysed in 0.1% Triton X-100 and incubated with pNPP at 37 • C for 60 min. ALP activity was determined based on the optical density measured at 405 nm, using an Asys UVM 340 microplate reader (Biochrom Ltd., Cambridge, UK).

Analysis of Mineralization by Adherent Cells in Culture Using Alizarin Red S Staining
The degree of mineralization was analyzed after a 14-day treatment with GD collagen. The MG-63 cells were washed with phosphate-buffered saline (PBS) and fixed with 70% ethanol for 1 h, prior to staining with 40 mM alizarin red S in deionized water (pH 4.2) for 15 min. After aspiration of the alizarin red S solution, the cells were incubated in PBS for 15 min in an orbital rotator, rinsed once with fresh PBS, and de-stained for 15 min with 10% (w/v) cetylpyridinium chloride in 10 mM sodium phosphate (pH 7). Finally, the extracted stain was transferred to a 96-well plate, and the absorbance values were measured at 550 nm, using an Asys UVM 340 microplate reader.

TRAP Activity and Staining
RAW 264.7 cells (5 × 10 4 ) were seeded in 96-well plates containing DMEM with 10% fetal bovine serum. Four days after stimulation of the cells with 50 ng/mL RANKL, the cells were washed with PBS. To measure the TRAP activity, the cells were fixed in 3.5% formaldehyde for 10 min, washed with distilled water, and incubated in 50 mM citrate buffer (pH 4.5) containing 10 mM sodium tartrate and 6 mM pNPP. After a 1-h incubation, the mixtures were moved into new 96-well plates containing an equal volume of 0.1 N NaOH, and the absorbance values were measured at 405 nm using an Asys UVM 340 microplate reader. For TRAP staining, the cells were stained using a leukocyte acid phosphatase assay kit (Sigma Chemical Co., St. Louis, MO, USA), according to the manufacturer's instructions.

Animals and Diet
All animal experiments were approved by the Institutional Animal Care and Use Committee of the Konkuk University (IACUC approval number, KU 15133). Four-week-old Wistar rats were purchased from Doo Yeol Biotech (Seoul, Korea). The animals were housed in a room maintained at 22 • C, under 12-h light-dark cycles. After a one-week acclimatization period, the rats were split into 2 groups, either a sham-operated group (normal, one group, n = 12) or an OVX group (four groups, total n = 48). Ovariectomy was performed via ligation and excision of the ovaries, whereas the sham surgery involved exposing the ovaries without excision. After a one-week period of adaptation, the initial mean rat body weight was 117 ± 7 g. The sham and positive control groups were fed a normal diet (TD.97191), whereas the negative control group was fed an OVX diet (TD.95027) containing a lower calcium (0.01%) than the normal diet (0.6% of calcium) ( Table 1). OVX rats fed GD collagen also received an OVX diet (TD.95027) containing a lower calcium (0.01%) content, and the added GD collagen quantity was deducted from the casein content in the OVX diet. The animals were randomly assigned to five groups (12 rats per group) as follows-(1) Sham (normal diet); (2) OVX diet (OVX + low calcium); (3) positive control (OVX + normal diet); (4) GDC2 (OVX + casein-deducted normal diet + 2 g/Kg GD collagen); and (5) GDC3 (OVX + casein-deducted normal diet + 3 g/Kg GD collagen). GD collagen was orally administered (gavage). Food intake was monitored daily, and the rats were weighed once a week. At the end of the 8-week feeding period, the rats received a lethal intraperitoneal ether overdose. Tibial bones were dissected and stored at -20 • C. Blood samples were collected from the heart under light anesthesia, and serum was then obtained via centrifugation (848× g for 30 min) and stored at −80 • C, prior to the biochemical assays.

Determination of Biochemical Variables and Estrogen Level in Serum
Serum lipid, including total cholesterol (TC), high-density lipoprotein (HDL) and triglyceride (TG) contents were measured using the T-CHO ELISA kit (3I2020; Asan Pharmaceutical, Hwaseong-si, Korea) for TC, HDL-CHO ELISA kit (3I2030; Asan Pharmaceutical, Hwaseong-si, Korea) for HDL, and the TG-S ELISA kit (3I1570; Asan Pharmaceutical, Hwaseong-si, Korea) for TG. Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels were determined using an automatic biochemistry analysis system (Hitachi 7060; Hitachi, Japan). The estrogen levels in the rat serum were measured using a rat estradiol ELISA kit (Biovision Inc., Milpitas, CA, USA), according to the manufacturer's protocol.

Calcium Content Measurement
Calcium content in the rat tibia was quantified using a microwave digestion system (Multiwave 3000; Anton Paar, Graz, Austria) and inductively coupled plasma mass spectrometry (HP-4500; Hewlett-Packard, Avondale, PA, USA). All experiments were performed according to official methods, using the Association of Analytical Chemists (AOAC) procedures.

Bone Strength Test
Tibial firmness was measured using an A/WEG wedge fracture probe (Stable Micro Systems, Godalming, UK). The tibial bone was fractured with the downward motion (3 mm/s) of a steel blade (30 mm wide). The maximum force (N) applied to break the bone was used to quantify bone firmness.

Micro-CT
Tibial morphometric parameters were determined in the distal tibia using a high-resolution, cone-beam micro-CT system (Inveon PET; Siemens Medical Solutions, Knoxville, TN, USA). The trabecular and cortical BMD and bone structure parameters, such as bone surface area/bone volume (BSA/BV), bone volume/total volume (BV/TV), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), and trabecular number (Tb.N), were determined by measuring trabecular bone mass and its distribution. Other parameters, such as cortical thickness (Ct.Th), were three-dimensionally determined via the measurement of cortical bone mass and its distribution, according to the standard procedures. Scans were performed using an applied voltage (80 kV) with a 1 mm aluminum filter. The image resolution of all cross-sections contained 2048 × 2048 pixels, with an isotropic voxel size of 9.31 µm. Data analysis was performed using the Inveon acquisition workplace software (Siemens). Z-scores were calculated using the following formula: Z-score = [measured BMD -age-matched BMD]/age-matched population SD [24].

RT-PCR Assay
Total RNA from the MG-63 cells and RAW264.7 cells that had were differentiated after 96 h, as well as tibial RNA, were isolated using TRIzol reagent (Thermo Fisher Scientific, Inc., Waltham, MA, USA). A 1-µg aliquot of total RNA was reverse-transcribed using SuperScript III Reverse Transcriptase (Invitrogen). The resultant cDNA was used for the determination of BMP-2, RUNX2, Wnt3a, osteocalcin, COL-1, OPG, RANKL, TRAP, and cathepsin K mRNA levels in the tibia, through PCR amplification, using Taq DNA polymerase (KAPA Biosystems, London, UK). As an internal control, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used. Different primer sequences were presented as follows:

Western Blotting Analysis
Differentiated cells and tibia were homogenized in a lysis buffer containing a protease inhibitor (Roche, Mannheim, Germany) and centrifuged at refrigerated temperature (10,000 × g for 10 min at). Total protein content in the samples was quantified using a Bio-Rad DC protein assay kit. Afterward, the proteins were separated in an SDS-PAGE experiment using a 12% polyacrylamide gel. The separated proteins were transferred onto immobilon-P transfer membranes, which were blocked with 5% bovine serum albumin before incubation with specific primary antibodies against osteocalcin, COL-1, BMP-2, or RUNX2 (Abcam), SMAD5 (Santa Cruz, Texas, USA), phosphorylated serine/threonine kinase (p-AKT), p-ERK, p-p38, p-JNK, or β-actin (Cell Signaling Technology, MA, USA). The membranes were incubated with the appropriate secondary antibody, either a goat anti-rabbit IgG (H + L)-HRP conjugate or goat anti-mouse IgG (H + L)-HRP conjugate (Zymax). The immune complexes antigen-antibody were visualized using enhanced chemiluminescence. Finally, densitometric analysis of the signals was performed using a C-DiGit Blot Scanner (Li-COR, NE, USA).

Histology, Immunohistochemical Analysis, and TRAP Staining of Tibia
Cleaned tibias were fixed for 2 days at 40 • C in 10% neutral-buffered formalin. Decalcification was performed by immersion-stirring in 10% ethylenediaminetetraacetic acid (EDTA, pH 7.4), which was replaced daily for 20 days at room temperature (25 • C). The bones were then washed with tap water for 4 h. Afterward, the tibias were paraffin-embedded, and longitudinal 4 µm sections were cut using a microtome. Sections were mounted, stained using hematoxylin and eosin, and observed using a light microscope at a magnification of 100×.
For the immunohistochemical analysis, the decalcified and paraffin-embedded tibias were deparaffinized in xylene, and rehydrated using an ethanol series, and then rinsed in PBS. A 30 min incubation in 0.3% hydrogen peroxide was performed to quench the endogenous peroxidase activity. Next, the 4 µm sections of tibia were incubated with 10% goat serum for 30 min and incubated overnight at 4 • C, with the specific primary antibodies against RUNX2, BMP-2, Wnt3a, osteocalcin, or COL-1 (Abcam, Cambridge, MA, UK). The sections were then incubated with either a biotinylated goat anti-rabbit IgG (H + L)-horseradish peroxidase (HRP) conjugate or goat anti-mouse IgG (H + L)-HRP conjugate (Zymax, San Francisco, CA, USA). The sections were then stained with 3,3 -diaminobenzidine (DAB substrate kit; Vector Lab., CA, USA) and counterstained with hematoxylin. Negative controls were incubated with normal goat IgG instead of the primary antibody. Finally, the specimens were observed, and the images were recorded using a Nikon Eclipse TS100 microscope (Nikon, Tokyo, Japan) at 200× magnification. Data were analyzed using the Optiview 3.7 image analysis software (Korea Lab Tech, Sungnam, Korea).
Additionally, TRAP and nuclear staining of the tibia were performed using the leukocyte acid phosphatase assay kit (Sigma Chemical Co., St. Louis, MO, USA), following the manufacturer's instructions. The specimens were examined at 200× magnification, using the Nikon Eclipse TS100 microscope.

Statistical Analysis
Data are presented as the mean ± standard deviation of triplicate experiments. Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 18.0 (SPSS Inc., Chicago, IL, USA). Comparisons between different groups were carried out using one-way analysis of variance, followed by Duncan's multiple range post-hoc test. p-values < 0.05 were considered statistically significant.

Collagen Contents of GD
The collagen contents of GD are given in Figure 1. The proximate chemical composition per 100 g of GD collagen showed 56.6% of total protein, 15.6% of moisture, 13.2% of fat, 11.6 of carbohydrate, and 2.9% of ash. Interestingly, the collagen fraction represented 78.6% of the total protein, indicating a collagen-rich product.

GD Collagen Regulated the Expression of Osteoblastogenesis and Mineralization in MG-63 Cells
To evaluate the cytotoxicity, osteoblastogenesis and mineralization induced by GD collagen (500 and 1000 µg/mL), we investigated the effects on cell viability, ALP activity, bone matrix mineralization, and the expression of proteins in osteoblast differentiation in MG-63 cells ( Figure 2). There was no significant change in MG-63 cell viability in the presence of GD collagen aqueous extracts, at all tested concentrations ( Figure 2A). As shown in Figure 2B, MG-63 cells exposed to GD collagen exhibited a dose-dependent increase in ALP activity, compared to the control cells (p < 0.05). GD collagen significantly increased the mineralization level in the cells ( Figure 2C), inducing a stronger red color in the cultured cells than in the control cells, indicating the promotion of bone mineralization ( Figure 2D). Figure 2E shows that the respective concentrations (500 µg/mL and 1000 µg/mL) of GD collagen significantly increased the expression of genes involved in bone anabolism, including BMP-2 (11.9-and 9.5-fold), SMAD5 (3.7-and 4.2-fold), Runx2 (5.2-and 5.1-fold), osteocalcin (1.5-and 1.5-fold), and COL-1 (7-and 7.2-fold) in GD collagen-treated MG-63 cells, compared with the control cells (p < 0.05).
Next, we examined the effect of GD collagen on p-ERK, p-p38, and p-JNK expression levels using osteoblast-like MG-63 cells. As shown in Figure 2F, all tested concentrations (500 and 1000 µg/mL) significantly upregulated the levels of p-ERK (2.8-and 3.0-fold, respectively) and p-p38 (4.8-and 2.7-fold, respectively), compared with the control cells, whereas p-JNK levels were significantly inhibited (1.5-and 1.9-fold, respectively) (p < 0.05). Thus, GD collagen might be safe for osteoblast-like MG63 cells, as it stimulates osteoblast differentiation and bone nodule mineralization, and activates the ERK and p38 pathways.

GD Collagen Suppressed Osteoclast Formation and the Expression of Osteoclastogenesis-Related Genes in RAW 264.7 Cells
To evaluate the effect of GD collagen (500 and 1000 µg/mL) on the proliferation, differentiation, and function of osteoclast precursor RAW 264.7 cells, we evaluated cell viability and identified markers for osteoclast differentiation and activity ( Figure 3A). No cytotoxic effects were observed in these osteoclast precursor cells at all tested concentrations of GD collagen. Since the TRAP plays an important role in osteoclast-mediated bone resorption [25], we measured TRAP activity in RAW 264.7 cells exposed to GD collagen ( Figure 3B). The results showed that GD collagen (500 and 1000 µg/mL) significantly inhibited the TRAP activity (1.5-and 1.8-fold, respectively) compared to the positive control cells (p < 0.05). The RAW 264.7 cells also exhibited significantly fewer TRAP-positive multinucleated cells when cultured in GD collagen-treated media, as opposed to the positive control cells ( Figure 3C).  Figure 3D,E shows that the mRNA levels of the soluble decoy Opg receptor were significantly more upregulated in 500 µg/mL and 1000 µg/mL of GD collagen-treated cells (2.6-and 2.1-fold, respectively) than that observed in the control cells (p < 0.05). The mRNA levels of Rankl were significantly reduced to~1.6-fold, regardless of the GD collagen dose. The Rankl/Opg ratio was markedly reduced at 500 µg/mL (4.2-fold) and 1000 µg/mL (3.4-fold-fold) of GD collagen, compared to the control cells (p < 0.05). When compared to the control cells, the GD collagen (500 µg/mL and 1000 µg/mL)-stimulated RAW 264.7 cells also exhibited a significant decrease in the mRNA levels of Trap (1.8-and 2.5-fold, respectively) and cathepsin K (~1.4-fold at all tested concentrations) (p < 0.05). These results suggested that GD collagen was less toxic to osteoclast precursors in short-time exposure, but might strongly prevent osteoclast differentiation and function.

GD Collagen Feed Enhanced Body Weight Gain
As shown in Table 2, all types of diets improved body weight gain in rats. However, there was no significant difference among experimental groups with regard to feed intake, body-weight gain, and food efficiency ratio (FER) ( Table 2). ; GD collagen groups = GD collagen proportion deduced from casein content in OVX diet, administration by oral gavage, GDC2, GD collagen 2 g/Kg, GDC3, GD collagen 3 g/Kg; 1 FER: Food efficient ratio = body weight gain (g/week)/feed intake (g/week).

GD Collagen Regulated Serum Biochemical Parameters in OVX Rats
Lipid contents (TC, HDL, and TG), liver function parameters (AST, ALT), and estradiol levels are presented in Table 3. When compared with the sham (normal) group, TC levels significantly increased in the negative control (OVX diet group) (1.3-fold, p < 0.05). Interestingly, the GD collagen dosages, especially GDC2 supplementation markedly decreased the TC levels (1.3-fold), compared to the negative control group. In addition, the GDC2 effect on TC levels was slightly better than that in GDC3 and the positive control groups, and comparable to the sham (normal) group (p < 0.05). In the negative control group, the serum HDL levels were significantly reduced (1.4-fold), compared to the sham group (p < 0.05). However, the GD collagen dosages, particularly GDC2, substantially enhanced the HDL levels, compared to the negative control group (p < 0.05). GDC2 induced slightly better improvement in the HDL levels, compared to the GDC3 and positive control groups, and showed similar results as in the sham (normal) group. Finally, no significant difference was observed among the TG levels in the sham (normal) and the negative control groups. At the same time, the GD collagen dosages, especially the GDC3 markedly decreased the TG levels (~1.3-fold), compared to the sham (normal) and the negative control groups (p < 0.05). We evaluated the effects of GD collagen treatments on the serum indices of liver function, such as AST and ALT (Table 3). There was a slight increase in the serum AST levels in the negative control group, compared to the sham (normal) group. Additionally, no significant difference was noticed, despite a minor increase in ALT. Intriguingly, the GDC2 and GDC3 diets significantly decreased the AST levels (1.3-and, 1.4-fold, respectively) compared to the negative control group (p < 0.05). The GDC3 inhibitory effect on AST surpassed that in the sham (normal) and the positive control groups (p < 0.05). Despite the small decrease in ALT levels in all GD collagen groups, compared to the negative control, no significant difference was reported.
The OVX rats showed a significant decrease in estrogen (2.5-fold), compared to the sham (normal) group (p < 0.05) ( Table 3). However, OVX rats fed the GDC2-and GDC3-supplemented diets, showed a significant increase in estrogen levels (1.8-and 1.5-fold), respectively, compared to the negative control group (p < 0.05). In addition, among GD collagen dosages, GDC2 induced a greater improvement in estrogen levels, compared to the positive control (p < 0.05).

GD Collagen Increased the Calcium Content and Bone Strength of the Tibia in OVX Rats
Rat tibial calcium content was measured to determine the influence of GD on bone mineralization ( Figure 4A). Results showed that the negative control group had a significantly lower calcium content (1.6-fold), compared with the normal (sham) group (p < 0.05). Interestingly, the GDC2 and GDC3 groups exhibited negligible higher calcium levels (approximately 1.04-and 1.1-fold, respectively) than the negative control group (p < 0.05).
No significant difference was observed among the GD collagen groups. Despite slightly lower calcium levels in the tibia of the GD collagen-fed rats, there was no significant difference compared to the positive control group.
We further investigated the effects of the GD collagen diet on bone strength. Breaking force measurements were performed using a texture analyzer ( Figure 4B). We found that the force required to break the tibia was significantly lower in the negative control group (299% lower) than in the sham (normal) group. However, the breaking forces required in the GDC2 and GDC3 groups was 218% and 190% higher than in the negative control group, respectively ( Figure 4C). Intriguingly, GDC2 and GDC3 exhibited significantly higher strength than the positive control group (p < 0.05). There was no significant difference between the GDC2 and GDC3 breaking forces. Thus, the GD collagen diet increased tibial bone strength compared to the negative control group.

GD Collagen Improved Bone Microarchitecture
To evaluate the bone structure improvements induced by GD collagen diets, we performed ex vivo micro-computed tomography (micro-CT) on all animals after 8 weeks, and investigated tibial architecture using the maximum intensity projection images ( Figure 5). Figure 5A illustrates the process and measurements used to analyze trabecular and cortical bones. After 8 weeks, the OVX diet rats (negative control group) showed an overall loss of tibial bone, with increased trabecular spacing, reduced bone volume, reduced trabecular number, and decreased trabecular thickness, compared to the sham group. Longitudinal sections of tibias from the negative control group showed the central spaces in the trabecular bone ( Figure 5B). The upper and lower cross-sections of the tibia of the negative control group presented large spaces within the marrow, as opposed to the sham group. In contrast, longitudinal and cross-sectional analyses showed the bone marrow space was more occupied in the GDC2 and GDC3 groups than that of the negative control group. The bone structures of the GDC2 and GDC3 groups were distributed in a relatively uniform fashion to form a well-connected network, which was greater than that observed in the negative control group, and quite similar to that in the positive control ( Figure 5B).
Furthermore, micro-CT analysis of the tibia metaphysis was used to determine trabecular and cortical BMD and bone structural parameters (BSA/BV, BV/TV, Tb.Th, Tb.Sp, Tb.N, and Ct.Th) ( Figure 5C). Following OVX surgery, the rats exhibited significantly reduced trabecular and cortical BMD values, and BV/TV, Tb.Th, and Ct.Th, and increased BSA/BV, Tb.N, and Tb.Sp, compared to the sham group (p < 0.05). Intriguingly, our data showed that the GDC2 and GDC3 rats presented higher trabecular and cortical BMD values and increased BV/TV, Tb.Th, Tb.N, and Ct.Th, compared to the negative control group. Other parameters, such as BSA/TV and Tb.Sp, were significantly lower in the GDC2 and GDC3 groups than in the negative control group (p < 0.05). Although there was no significant difference between the GDC2 and GDC3 overall structural parameters, GDC2 induced better BV/TV and Tb.Th than did GDC3 (p < 0.05). Finally, GDC2 also exhibited better bone structural change results than did the positive control group, while the GDC3 results were quite similar to those of the positive control group. These data indicated that dietary supplementation with GD collagen ameliorated the loss of trabecular bone integrity observed in OVX rats.
Finally, we calculated the Z-score of trabecular and cortical bones ( Figure 5D) in which the BMD was compared to age-matched rats. Results showed significant decreases in both trabecular and cortical bones of the negative control group, compared to the sham group (normal) (p < 0.05). In contrast, the OVX rats fed with GDC2 and GDC3 showed significantly increased Z-scores compared to the negative control group (p < 0.05), whereas no significant difference was observed among GDC2 and GDC3. In addition, no significant difference was observed between the GD collagen and the positive control group for the trabecular Z-score; whereas all GD collagen dosages displayed the highest cortical Z-score, compared to the positive control (p < 0.05). These findings were confirmed by a histological examination using hematoxylin and eosin staining ( Figure 6).

GD Collagen Increased Trabecular Connections
The rat tibias were stained with hematoxylin and eosin for histopathological examination ( Figure 6). As opposed to the sham group, the trabeculae of rats in the negative control group were disorderly arranged and were thinner, or completely disappeared; the remaining connections were incomplete. A non-optimal discontinuity in the trabeculae and signs of matrix resorption were apparent. In contrast, the GDC2 and GDC3 groups showed a greater number of well-connected trabeculae, arranged in a more orderly fashion than those observed in the negative control group. These findings indicated that the GD collagen diet improved the trabecular structure and continuity.
In addition, immunohistochemical staining was performed to measure the immunoreactivity of the osteogenesis-related proteins (BMP-2, Wnt3a, RUNX2, osteocalcin, and COL-1), as well as a specific marker for osteoclast function (TRAP) (Figure 6). For all osteogenesis-related proteins, the negative control group displayed mild to moderate staining, compared to the sham group. The staining for the GDC2, GDC3, and positive control groups was slightly more intense, compared to that of the negative control group. Multinucleated TRAP-positive osteoclasts were found at 8 weeks after surgery in the subchondral bone marrow space of the tibia (Figure 6). In contrast, in the negative control group, TRAP-positive osteoclasts were mainly detected in the tibial subchondral bone marrow space, as compared to the other groups. These results showed that the GD collagen diet increased the bone levels of osteogenesis-related proteins, BMP-2, RUNX2, Wnt3a, osteocalcin, and COL-1, and suppressed the mRNA expression of genes involved in osteoclastogenesis in vivo.

GD Collagen Inhibited Osteoclastogenesis Associated Genes
As shown in Table 4, the negative control group expressed a significantly lower Opg (3.7-fold), and increased Rankl (5.3-fold), Rankl/Opg ratio (20-fold), Trap (4.0-fold), and cathepsin K (9.5-fold), compared to the sham group (p < 0.05). However, compared to the negative control group, the GDC2 and GDC3 groups showed markedly enhanced OPG (3.3-and 2.6-fold, respectively) and significantly lowered levels of Rankl (1.6-and and 2.8-fold, respectively), Rankl/Opg ratio (5.4-and 7.3, respectively), Trap (1.7-and 1.5-fold, respectively), and cathepsin K (7.5-and 15-fold, respectively) (p < 0.05). Among the GD collagen dosages, GDC2 induced a higher increase in Opg and a better decrease in Trap, while GDC3 was more effective at alleviating the Rankl/Opg ratio, Rankl, and cathepsin K. The results from all GD collagen dosages were more efficient than those of the positive control group. These findings suggested that GD collagen might drive the Rankl/Opg ratio to inhibit the mRNA expression of osteoclastogenesis-associated genes in vivo.

GD Collagen Regulated MAPK Signaling
To examine the effects of GDC2 and GDC3 diets on MAPK signaling, we evaluated the activation of phosphorylated ERK, p38, and JNK in the tibia. As shown in Figure 7, the negative control group significantly inhibited p-ERK (3.2-fold) and p-p38 (1.6-fold), but promoted p-JNK (2.0-fold), compared to the sham group (p < 0.05). GDC2 and GDC3 significantly activated levels of p-ERK (1.9-and 1.3-fold, respectively) and p-p38 (1.5-and 1.6-fold, respectively), while deactivating p-JNK levels (1.2-and 1.8-fold, respectively) compared to the negative control group (p < 0.05). Both GD collagen dosages induced similar p-p38 level activation, but GDC2 induced higher p-ERK activation. GDC3 was more effective in inhibiting p-JNK. Overall, the positive control group showed higher p-ERK and p-p38 activation, compared to both GD collagen dosages; however, GDC3 was the most effective in p-ERK inhibition. Thus, the GD collagen diet modulated the MAPK signaling pathways, leading to osteoblast differentiation and inhibition of osteoclast differentiation. Figure 7. Effect of GD collagen on the MAPK signaling pathway. p-ERK, p-p38, and p-JNK protein expression in rats' tibias were determined by Western blot analysis. GDC2, GD collagen 2 g/kg and GDC3, GD collagen 3 g/kg. Expression was quantified using ImageJ, relative to that of β-actin. The values represent the mean ± standard deviation. Values with different letters differed significantly using Duncan's multiple range test (p < 0.05).

Discussion
Yeonsan Ogye is a traditional Korean chicken breed, Gallus gallus domesticus, with the dominant gene responsible for dermal hyperpigmentation (or fibromelanosis) [26] Abnormal accumulation of eumelanin (photoprotective and antioxidant pigment) in tissues is responsible for the hyperpigmentation of Yeonsan Ogye, causing its unique dark, gray-black coloration, including black fluffy head feathers, earlobes, and pupils [27,28]. This was the first study to simultaneously examine the anti-osteoclastogenic and osteogenic effects of GD collagen (collagen extract obtained from Yeonsan Ogye flesh), via the impairment of the RANKL/OPG ratio, activation of p38 and inhibition of the JNK signaling pathways, in an OVX rat model.
Estrogen plays a key role in bone metabolism regulation by inhibiting osteoclast-mediated bone resorption and upregulating apoptosis in osteoclasts [29]. The estrogen deficiency caused by bilateral ovariectomy in rats dramatically reduced bone mineral content, density, and biomechanical strength [30,31]. The combination of ovariectomy with a low-calcium diet in rats was used to establish a rapid animal model of osteoporosis [32]. We measured the serum level of estradiol, which is the main product resulting from the biosynthesis of estrogen. Moreover, estradiol is the most potent estrogen during the premenopause period [33]. Our results showed that the additive effects of OVX and a low-calcium diet (0.01%) significantly decreased the serum estrogen levels in the negative control group. In contrast, the GD collagen supplemented the significantly enhanced estrogen levels, and these results were better than that of the OVX rats receiving a normal diet (high calcium content, 0.6%). Moreover, our results showed a significant decrease in tibial calcium levels in the negative control group, while the GD collagen diet alleviated calcium loss in rat tibia ( Figure 4A). A low calcium feed might induce a decrease in calcium content in rat tibia. Kim et al. reported that when the calcium level falls due to low calcium intake and increased excretion in urine, parathyroid hormone release rises, provoking bone resorption stronger than building-up [34]. Mizuno and Kuboki showed via an in vitro study that a type I collagen supplement enhances calcium precipitation in the medium of bone marrow stromal fibroblastic cultured cells [35]. Furthermore, we demonstrated that the break force of tibias from the GDC2 and GDC3 groups was 218% and 190% higher than that required for the negative control group, respectively ( Figure 4C). Our results revealed that feeding GD collagen diets to OVX rats prevented hypocalcemia and increased bone strength. Therefore, the GD collagen diet might be useful for the maintenance of bone health.
Using Micro-CT, which offers nondestructive, high-resolution 3D imaging of the internal structure of objects [36], we found that both the upper and lower cross-sections of the tibias from the negative control group exhibited fewer contents of trabecular bone, higher void spaces in the marrow area, and a rod-like structure in the bone marrow ( Figure 5B). In contrast, the cross-sections of the sham (normal), the GDC2, and the GDC3 groups displayed relatively uniform patterns, well-connected networks, and better spatial distributions of a plate-like trabecular bone structure in the bone marrow. Compared with the negative control ( Figure 5C), the GD collagen diets increased trabecular and cortical BMD scores, enhanced bone structure parameters (BV/TV, Tb.Th, Tb.N, and Ct.Th), and decreased other BSA/BV and Tb.Sp parameters. Additionally, trabecular and cortical bone Z-scores significantly decreased below −2.5, whereas GDC2 and GDC3 significantly increased these parameters above −1.5.
In particular, all GD collagen dosages resulted in better improvement in cortical bone Z-score than the positive control. A study reported that lower Z-scores were associated with a severe osteoporotic status [37] Furthermore, a lower Z-score associated with low BMD was indicative of factors other than natural menopause and aging, which were known to adversely affect the health of the skeletal system [38]. These findings revealed an improvement in the overall bone strength and tibial trabecular bone density. These results go beyond those of previous studies [39], showing a decrease in BV/TV, Tb.Th, Tb.N, and Ct.Th, and an increase in BSA/BV and Tb.Sp, succeeding the OVX operation [39,40]. Collagen supplementation was reported to enhance BMD and increase bone strength and the external diameter of the cortical areas in the thighbone of OVX mice [17,18]. Accordingly, the improvements in bone mass and the microarchitectural network induced by the GD-collagen diet, might explain the increased biochemical strength of the bones.
Our in vitro results demonstrated that the osteoblast precursor MG-63 cells treated with GD collagen exhibited increased ALP activity and deeply positive staining in the calcified nodules, as compared to the control cells ( Figure 2B-D). Mechanistic studies revealed that the GD-collagen diet significantly upregulated the expression of osteocalcin, COL-I, Runx2, Wnt3a, BMP-2, and SMAD5, as opposed to the control cell (in vitro) and the negative control group (in vivo) ( Figure 2E, Figure 6) ( Table 4). Previous studies reported that enzymatically hydrolyzed collagen, gelatin, and collagen derived from tilapia, bovine, and porcine, improved various osteoblastic differentiation markers (BMP-2, Runx2, SMAD5, osteocalcin, and COL-1), increased ALP activity, and accelerated bone matrix mineralization [17,18,[41][42][43]. Our findings confirmed that GDC2-and GDC3-collagen supplements promote bone formation by stimulating bone-specific matrix protein expression, bone nodule mineralization, and Wnt3a/BMP-2/SMAD5/RUNX2 anabolic signaling pathways (Figure 8). The present study showed that the GD-collagen diet activated Wnt3a signaling and increased the synthesis of OPG in OVX rats, while decreasing the levels of RANKL, TRAP, and cathepsin K, when compared to the negative control group (Figure 8). The impairment that occurred in the RANKL/OPG ratio suggests that GD collagen supports bone formation and reduces bone resorption by inhibiting the osteoclastic differentiation [44]. Upregulation of Wnt3a expression might activate the β-catenin signals, inducing OPG synthesis by osteoblasts and bone marrow stromal cells, which in turn inhibit osteoclastic differentiation [45,46]. Synthesized soluble OPG inhibits the binding of RANKL to its RANK receptor in osteoclast precursors, thus, downregulating osteoclastic differentiation [47][48][49]. In contrast, OPG deficiency promotes RANKL-RANK binding, activating signaling pathways, and enhancing the expression of osteoclast-specific genes, such as TRAP and cathepsin K, in osteoclasts [50]. Our results showed that GDC2 and GDC3 suppressed the expression of cathepsin K, which is upregulated by RANKL and degrades bone type I collagen. These findings agree with those of Elango et al., showing that collagen treatment decreases the number of bone marrow macrophage-derived TRAP-positive osteoclasts and suppresses the expression of osteoclastogenic regulatory genes [12]. Our results suggest that GDC2 and GDC3 inhibit osteoclast formation by activating the canonical Wnt pathway, promoting OPG synthesis, and blocking the expression of RANKL, TRAP, and cathepsin K.
Studies have shown that the p38 and JNK signaling pathways are involved in the regulation of several cellular physiological functions, including cell proliferation, osteoblast differentiation, and skeletal development [8,[51][52][53]. The present study found that the GD-collagen diet upregulated osteoblast-specific genes and osteoblast differentiation by activating p-38 signaling in MG-63 osteoblast cells and OVX rats (Figures 2F and 7). The p38 signaling pathway is involved in osteogenesis [52], and p38 is probably involved in the regulation of COL-1 and osteocalcin expression in differentiating cells [54]. A study on the ugonin K flavonoid obtained from the roots of Helminthostachys zeylanica revealed that the p38 signaling pathway helps promote osteoblastic differentiation and mineralization [55]. However, OVX rats fed with GDC2 and GDC3, exhibited significantly decreased JNK expression, compared to the negative control group (Figure 7). Similarly, proanthocyanidins used as a therapeutic in an in vitro osteoporosis treatment model inhibited osteoclast formation and function by blocking the JNK signaling pathway in bone marrow macrophages and RAW264.7 cells [56]. Another report demonstrated that SP600125 suppressed PMMA particle-induced osteolysis by inhibiting the JNK signaling pathway in human osteoclast precursor cells and mouse calvaria model [57]. Thus, dietary supplementation with GD collagen might simultaneously promote osteoblast differentiation by activating the p38 signaling pathway and suppressing osteoclastogenesis by blocking the JNK signaling ( Figure 8).

Conclusions
This was the first study to show that collagen extract derived from the flesh of the Yeonsan Ogye chicken breed (Gallus domesticus) simultaneously promoted osteoblastogenesis and prevented osteoclastogenesis both in vitro and in vivo. These findings revealed the biosafety of GD collagen. The orally administered GD collagen alleviated the severe bone calcium loss induced by a low-calcium diet intake. GD collagen enhanced BMD and bone microarchitecture, as well as improved bone structure parameters in the tibia of OVX rats fed low calcium diets. GD collagen intensified the osteoblast number and blocked osteoclastic differentiation, thereby reducing the osteoclast number. We have shown both in vitro and in vivo that GD collagen improved the expression of bone-specific matrix proteins (ALP, osteocalcin, and COL-1), Runx2 transcription factors, and osteoinductive BMP-2, and SMAD5; increased the expression of OPG; and decreased RANKL. GD collagen downregulated the expression of osteoclast-specific genes, including TRAP and cathepsin K, activated the ERK and p38 signaling pathways, and blocked the JNK pathway; it exhibited osteogenic and anti-osteoclastogenic effects. The current study provided evidence of the positive effects of GD collagen, which might help prevent and treat dysfunction in the bone remodeling process, such as osteoporosis. Future research should investigate the clinical effects of GD collagen supplements in combined estrogen deficiencyand calcium deficiency-induced bone disorders.