Mechanism of CK2.3, a Novel Mimetic Peptide of Bone Morphogenetic Protein Receptor Type IA, Mediated Osteogenesis

Background: Osteoporosis is a degenerative skeletal disease with a limited number of treatment options. CK2.3, a novel peptide, may be a potential therapeutic. It induces osteogenesis and bone formation in vitro and in vivo by acting downstream of BMPRIA through releasing CK2 from the receptor. However, the detailed signaling pathways, the time frame of signaling, and genes activated remain largely unknown. Methods: Using a newly developed fluorescent CK2.3 analog, specific inhibitors for the BMP signaling pathways, Western blot, and RT-qPCR, we determined the mechanism of CK2.3 in C2C12 cells. We then confirmed the results in primary BMSCs. Results: Using these methods, we showed that CK2.3 stimulation activated OSX, ALP, and OCN. CK2.3 stimulation induced time dependent release of CK2β from BMPRIA and concurrently CK2.3 colocalized with CK2α. Furthermore, CK2.3 induced BMP signaling depends on ERK1/2 and Smad1/5/8 signaling pathways. Conclusion: CK2.3 is a novel peptide that drives osteogenesis, and we detailed the molecular sequence of events that are triggered from the stimulation of CK2.3 until the induction of mineralization. This knowledge can be applied in the development of future therapeutics for osteoporosis.


Introduction
Osteoporosis is a skeletal disorder and is the leading cause of fractures in the USA. 10 million Americans are diagnosed with osteoporosis and another 44 million Americans are at an increased risk of a fracture due to their low bone density. In 2005, there were 2 million osteoporosis-related fractures that resulted in a total cost of 18 billion dollars, and this number is estimated to increase to 3 million fractures and cost 25.3 billion dollars by 2025 [1,2]. This is due to the rapid increase in the world's aging populations, thus placing a substantial burden of disability and costs on the individuals, as well as on the society. The underlying cellular pathogenesis causing osteoporosis is due to: (a) excessive bone resorption by osteoclasts and/or (b) failure of osteoblasts to produce and replace the resorbed bone, leading to weak and porous bones [3][4][5][6]. The majority of the existing treatments for osteoporosis [4,[7][8][9] such as bisphosphonates (e.g., alendronate, risedronate, and ibandronate) [10], selective estrogen receptor modulators (SERMs) (e.g., tamoxifen and raloxifene) [11,12], calcitonin, and denosumab, target inhibition of osteoclast activity. However, prolonged use of such treatments can prevent bone remodeling, thereby micro-damages and fractures in the bone may go unresolved. Teriparatide, a recombinant human-parathyroid hormone (rh-PTH), composed of 1-34 amino acid fragments of the parathyroid hormone (PTH) [13,14], is an anabolic treatment that addresses the issue of increasing bone mineral density (BMD) in osteoporotic patients, but excess concentration of serum PTH Stimulation of C2C12 cells with 100 nM of CK2.3 induced a time-dependent release of CK2β from BMPRIA. Additionally, we confirmed a gradual increase in co-localization between CK2α and CK2.3 starting at 6 h, 12 h, and 18 h post-stimulation using CK2.3-Qdot ® s. We determined that CK2.3 activated the BMPRIA downstream signaling pathways such as Smad1/5/8, ERK1/2, and Akt1/2/3. However, using pharmacological inhibitors such as U0126-EtOH, SB202190, and MK-2206 2HCl against MEK1/2, p38 MAPK, and Akt1/2/3, respectively, and Smad4 siRNA to silence Smad1/5/8, we were able to identify that the ERK1/2 and Smad1/5/8 signaling pathways were critical for CK2.3mediated osteogenesis. Current research focuses on Smad1/5/8-mediated osteogenesis and neglects the role of Smad-independent signaling pathways. Here, we report the ERK1/2 signaling pathway to be essential for osteogenesis by CK2.3; an alternate proposal to a commonly held dogma. The significance of ERK1/2 and Smad1/5/8 signaling pathways on CK2.3 mediated osteogenesis was also verified in primary BMSCs. Now that we are able to characterize the activity of CK2.3, it can be regarded as a potential candidate for the treatment of osteoporosis, subject to further research in advanced animal models, or at the very least, the information that we have gained from this research can be implemented in the development of future treatments for osteoporosis.

CK2.3 Stimulation Led to the Upregulation of Osterix, Alkaline Phosphatase, and Osteocalcin Genes in C2C12 Cells
Differentiation of osteoblasts from pluripotent MSCs is a highly sequenced process that involves the expression of particular genes at specific stages of differentiation [40][41][42][43][44]. C2C12 cells were treated with 100 nM of CK2.3 from day 1-5. Total RNA was isolated at each respective time point and a twostep RT-PCR was performed. Expression of osteoblastic genes RUNX2, OSX, ALP, and OCN were analyzed using RT-qPCR. C2C12 cells stimulated with 100 nM of CK2.3 elevated the expression of RUNX2 mRNA on day 4, however it was not statistically significant ( Figure 1A). However, CK2.3 treatment resulted in a gradual increase in the expression of OSX starting from day 1 and was significantly elevated on day 2 and day 3 by 3.95 ± 0.66 and 4.1 ± 0.9 fold, respectively ( Figure 1B). The expression pattern of OSX by CK2.3 is similar to the BMP2-mediated induction of OSX in C2C12 cells [45][46][47]. ALP expression was significantly upregulated on day 5 by 3.98 ± 0.8 fold, following treatment with 100 nM of CK2.3 ( Figure 1C), the upregulation of ALP mRNA expression is similar to C2C12 cells stimulated with BMP2 [48,49]. It is reported that, as the osteoblast differentiation progresses, the expression of OCN gene is upregulated [34,49,50] and a similar result was observed in our experiment with CK2.3 in C2C12 cells, i.e., on day 5, OCN expression was significantly increased by 3.63 ± 1.36 fold ( Figure 1D). GAPDH was used as the house-keeping gene.  OCN were analyzed using RT-qPCR over the course of 5 days. GAPDH was used as the house-keeping gene. Data (n = 4) was normalized to unstimulated cells and analyzed using one-way anova and Tukey-Kramer post hoc statistical test (p < 0.05). a = statistically significant difference to unstimulated, b = statistically significant difference to Day 1, c = statistically significant difference to Day 2, d = statistically significant difference to Day 3, e = statistically significant difference to Day 4, and f = statistically significant difference to Day 5.

Time-Dependent Release of CK2 from BMPRIA at 12 h and 18 h Post CK2.3 Stimulation
The interaction between CK2.3 and BMPRIA must be determined in order to investigate the signaling pathways upregulated by CK2.3. We previously identified an association between CK2 and BMPRIA [19], however the time frame of interaction and release of CK2 were unknown. C2C12 cells were cultured and stimulated with 100 nM of CK2.3 for 6 h, 12 h, and 18 h. Cells were fixed, endogenous BMPRIA and CK2α were immuno-fluorescently labeled, and the co-localization of BMPRIA and CK2α were visualized using a confocal laser scanning microscope (Figure 2A). Colocalization of BMPRIA (green) and CK2α (red) was observed in unstimulated cells (Figure 2A(A)) and at 6 h post CK2.3 stimulation (Figure 2A(B)). However, with prolonged CK2.3 stimulation, decreased co-localization between CK2α and BMPRIA was observed at 12 h (Figure 2A(C)) and 18 h (Figure 2A(D)), respectively. This indicates that CK2.3 was mediating the release of CK2 from BMPRIA after 6 h. To confirm this finding, C2C12 cells were either left unstimulated or stimulated with 100 nM of CK2.3, lysed, and immuno-precipitated for BMPRIA. Detection of the association with CK2β was completed by SDS-PAGE followed by Western blot. There was a decreased amount of CK2β detected in stimulated cells when compared to unstimulated cells, with the maximum disassociation observed at 18 h ( Figure 2B). This indicates and confirms that CK2 was fully released from BMPRIA at 18 h post CK2.3 stimulation. OCN were analyzed using RT-qPCR over the course of 5 days. GAPDH was used as the house-keeping gene. Data (n = 4) was normalized to unstimulated cells and analyzed using one-way anova and Tukey-Kramer post hoc statistical test (p < 0.05). a = statistically significant difference to unstimulated, b = statistically significant difference to Day 1, c = statistically significant difference to Day 2, d = statistically significant difference to Day 3, e = statistically significant difference to Day 4, and f = statistically significant difference to Day 5.

Time-Dependent Release of CK2 from BMPRIA at 12 h and 18 h Post CK2.3 Stimulation
The interaction between CK2.3 and BMPRIA must be determined in order to investigate the signaling pathways upregulated by CK2.3. We previously identified an association between CK2 and BMPRIA [19], however the time frame of interaction and release of CK2 were unknown. C2C12 cells were cultured and stimulated with 100 nM of CK2.3 for 6 h, 12 h, and 18 h. Cells were fixed, endogenous BMPRIA and CK2α were immuno-fluorescently labeled, and the co-localization of BMPRIA and CK2α were visualized using a confocal laser scanning microscope ( Figure 2A). Co-localization of BMPRIA (green) and CK2α (red) was observed in unstimulated cells (Figure 2A(A)) and at 6 h post CK2.3 stimulation (Figure 2A(B)). However, with prolonged CK2.3 stimulation, decreased co-localization between CK2α and BMPRIA was observed at 12 h (Figure 2A(C)) and 18 h (Figure 2A(D)), respectively. This indicates that CK2.3 was mediating the release of CK2 from BMPRIA after 6 h. To confirm this finding, C2C12 cells were either left unstimulated or stimulated with 100 nM of CK2.3, lysed, and immuno-precipitated for BMPRIA. Detection of the association with CK2β was completed by SDS-PAGE followed by Western blot. There was a decreased amount of CK2β detected in stimulated cells when compared to unstimulated cells, with the maximum disassociation observed at 18 h ( Figure 2B). This indicates and confirms that CK2 was fully released from BMPRIA at 18 h post CK2.3 stimulation.

Time-Dependent Co-Localization of CK2 with CK2.3-Qdot ® s at 6 h, 12 h, and 18 h Post-Stimulation
To determine the activity of CK2.3 in vitro, CK2.3 is conjugated to fluorescent nanoparticles called quantum dot ® s (Qdot ® s) to track CK2.3 within C2C12 cells. Qdot ® s are semiconductor nanocrystals and their unique photophysical properties make them ideal for bio-imaging applications. The development and characterization of CK2.3-Qdot ® s has been published [32].
C2C12 cells were either left unstimulated or stimulated with 100 nM of CK2.3-Qdot ® s for 6 h, 12 h, and 18 h. After each respective time point, cells were fixed and stained for CK2α and nucleus of the cell. Images of the cell were taken using confocal laser scanning microscopy. At 6 h post-stimulation, slight co-localization between endogenous CK2α (red) and CK2.3-Qdot ® s (green) ( Figure 3B) was observed. However, the co-localization peaked at 12 h post-stimulation, as seen in the merged images at zoom-4 and zoom-10 ( Figure 3C). At 18 h, we mostly observed the co-localization between CK2α and CK2.3-Qdot ® s around the plasma membrane of the cell ( Figure 3D). These data indicate that CK2.3 co-localized with protein kinase 2 intracellularly.  In the secondary control, unstimulated cells were stained using only the fluorescent secondary antibody to determine their specificity, and lack of staining in the cells shows that the antibody was specific against the antigen. (B) Immuno-precipitation of BMPRIA in C2C12 cells, not stimulated (-) or stimulated (+) with 100 nM of CK2.3 for 6 h, 12 h, and 18 h; followed by a co-immuno-precipitation for CK2β that showed reduced interaction of CK2β with BMPRIA at 12 h and 18 h post-stimulation. Con represents the negative control of the immuno-precipitation, where lysis buffer was used instead of cell lysate. the cell. Images of the cell were taken using confocal laser scanning microscopy. At 6 h poststimulation, slight co-localization between endogenous CK2α (red) and CK2.3-Qdot ® s (green) ( Figure  3B) was observed. However, the co-localization peaked at 12 h post-stimulation, as seen in the merged images at zoom-4 and zoom-10 ( Figure 3C). At 18 h, we mostly observed the co-localization between CK2α and CK2.3-Qdot ® s around the plasma membrane of the cell ( Figure 3D). These data indicate that CK2.3 co-localized with protein kinase 2 intracellularly.

B
A D C Figure 4. Smad1/5/8, ERK1/2, and Akt1/2/3 signaling pathways are upregulated following stimulation with 100 nM of CK2.3 in C2C12 cells. C2C12 cells were either left unstimulated or stimulated with 100 nM of CK2.3 for 6 h, 12 h, and 18 h over days 1-5 and, after each respective time points, protein was extracted, normalized, and analyzed using SDS-PAGE, followed by Western blot. (A) p-Smad1/5/8 and total Smad1/5/8 expression gradually increased with CK2.3 treatment. (B) p-ERK expression was elevated and total ERK1/2 expression gradually increased through the 5-day experiment. Similarly, (C) p-Akt1/2/3 expression increased with CK2.3 treatment, however, (D) p38 MAPK expression stayed relatively constant. β-actin was used as the loading control. An absolute timing of the CK2.3 response was variable over the 5-day experiment. Thus, we did not include a quantitative representation of the Western blots.

Inhibition of ERK1/2 and Smad1/5/8 Signaling Pathways Led To Decreased Mineralization by CK2.3 in C2C12 Cells
To determine the pathway responsible for CK2.3 mediated osteogenesis, we used inhibitors such as U0126-EtOH, SB202190, and MKK-2206 2HCl against ERK1/2, p38 MAPK, and Akt1/2/3 signaling pathways, respectively. An optimal concentration of specific signaling inhibitors that resulted in significant reduction in mineralization was determined by performing von Kossa assay. Furthermore, to confirm that the reduction in mineralization was solely due to the inhibition of the signaling pathway and not due to reduced number of viable cells, viability of the cells were determined by MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoilum bromide) tetrazolium reduction assay and total number of cells were quantified using ImageJ analysis. Smad1/5/8 signaling pathway was silenced using Smad4 siRNA.

500 nM of ERK1/2 Inhibitor (U0126-EtOH) Induces Significant Reduction in Mineralization Without Affecting the Viability of C2C12 Cells
Von Kossa assay is a commonly used technique to detect calcium deposits [63]. C2C12 cells were either left unstimulated or stimulated with 500 nM, 1 µM, and 5 µM of ERK1/2 Inhibitor (U0126-EtOH), p38 MAPK inhibitor (SB202190), and Akt1/2/3 (MKK-2206 2HCl). Also, only in the case of p38 MAPK inhibitor, we used an additional concentration of 10 µM, as neither of the concentrations of SB202190 yielded a significant reduction in mineralization. After 24 h, the cells were then treated with 100 nM of CK2.3. Stimulation of cells with varying concentration of signaling inhibitors, followed by CK2.3 stimulation, was repeated for a total of two times over the course of the experiment. On the sixth day, cells were fixed, and calcium deposits were stained using von Kossa assay. The analysis of images of cells using ImageJ revealed that 500 nM of ERK1/2 inhibitor (U0126-EtOH) ( Figure 5A), 5 µM of Akt inhibitor (MKK-2206 2HCl) ( Figure 5B), and 10 µM of p38 MAPK inhibitor (SB202190) ( Figure 5C) had a significant reduction in mineralization compared to cells only treated with 100 nM of CK2.3.
Viability of cells following stimulation with varying concentrations of pharmacological inhibitors and CK2.3 was determined by MTT reduction assay and quantification of total number of cells. C2C12 cells were either left unstimulated or stimulated with 500 nM of U0126-EtOH, 5 µM of MKK-2206 2HCl, and 10 µM of SB202190 to inhibit ERK1/2, Akt1/2/3, and p38 MAPK signaling pathways, respectively. The next day, the cells were then treated with 100 nM of CK2.3. Halfway through the experiment, the cells were once again treated with the respective concentrations of signaling inhibitors and 100 nM of CK2.3. In the end, viability of the cells was determined by conducting MTT assay. Our data show that 500 nM of ERK1/2 Inhibitor (U0126-EtOH) did not induce significant reduction in the number of viable cells compared to unstimulated cells or cells only treated with 100 nM of CK2.3, however, cells treated with 5 µM of Akt1/2/3 inhibitor (MKK-2206 2HCl) and 10 µM of p38 MAPK inhibitor (SB202190) resulted in a significant decrease in the number of viable cells ( Figure 5D). This data was further verified by performing a cell counting experiment ( Figure 5E). C2C12 cells that were either left unstimulated or stimulated with signaling inhibitors and CK2.3 were then fixed after 6 days. Images of cells were taken, and total number of cells were quantified using ImageJ analysis. Our data once again revealed that there was no significant difference amongst cells that were left untreated, cells that were only treated with 100 nM of CK2.3, or cells that were treated with 500 nM of ERK1/2 inhibitor (U0126-EtOH) and 100 nM of CK2.3; however, cells that were treated with 5 µM of Akt inhibitor (MKK-2206 2HCl) and 10 µM of p38 MAPK inhibitor (SB202190) and additionally stimulated with 100 nM of CK2.3 had a significant reduction in the number of cells. In conclusion, we report that ERK1/2 signaling pathway plays a key role in CK2.3 mediated osteoblast differentiation in C2C12 cells.

200 pM of Smad4 siRNA Induces Significant Reduction in Mineralization Without
Affecting the Viability of C2C12 Cells Smad1/5/8 is one of the most well studied signaling pathway in mediating osteogenesis, however there are not any commercially available pharmacological inhibitors to prevent the activation of Smad1/5/8 molecule. Therefore, we utilized Smad4 siRNA to silence the Smad1/5/8 signaling pathway. Smad4 is a common mediator of Smad1/5/8, which aides in the translocation of Smad1/5/8 into the nucleus, where Smad1/5/8 acts as a transcription factor by directly binding to the DNA and regulating the expression of osteoblast specific genes [51,52]. Furthermore, Smad4 is reported to be essential in regulating osteoblast activity, as its deletion in mature osteoblasts resulted in reduced osteoblast function [64]. Therefore, by employing Smad4 siRNA we effectively prevent the mRNA synthesis of Smad4, thereby Smad1/5/8 is unable to translocate to the nucleus. In our study, C2C12 cells were either left untreated or treated with 200 pM of Smad4 siRNA. The siRNA and the reagent used as the vehicle to transfect the cells are identical to a previously published paper [31] and it is reported that this particular Smad4 siRNA is successful in knocking down 50% of Smad4 [65]. Primary BMSCs were isolated from the tibia and femur of 4-month-old female C57BL/6J mice [18,66,67]. C57BL/6J mice are an ideal mice model to study age-related forms of osteoporosis due to their inherent low bone mineral density [68]. BMSCs were either left unstimulated or stimulated with 500 nM of U0126-EtOH (ERK1/2 inhibitor), 200 pM of Smad4 siRNA, and 200 pM of control siRNA-A. The next day, the cells were again either left unstimulated or stimulated with 100 nM of CK2.3. Treatment of cells with signaling inhibitor/siRNAs, followed by CK2.3 stimulation, was repeated for a total of two times over the course of the experiment. Later, the cells were fixed and stained for calcium deposits using von Kossa assay. Our data in primary BMSCs revealed that cells treated with only 100 nM of CK2.

CK2.3 Mediates Osteogenesis via the Smad1/5/8 and ERK1/2 Signaling Pathways in Primary BMSCs Isolated From 4-Month-Old Female C57BL/6J Mice
Primary BMSCs were isolated from the tibia and femur of 4-month-old female C57BL/6J mice [18,66,67]. C57BL/6J mice are an ideal mice model to study age-related forms of osteoporosis due to their inherent low bone mineral density [68]. BMSCs were either left unstimulated or stimulated with 500 nM of U0126-EtOH (ERK1/2 inhibitor), 200 pM of Smad4 siRNA, and 200 pM of control siRNA-A. The next day, the cells were again either left unstimulated or stimulated with 100 nM of CK2.3. Treatment of cells with signaling inhibitor/siRNAs, followed by CK2.3 stimulation, was repeated for a total of two times over the course of the experiment. Later, the cells were fixed and stained for calcium deposits using von Kossa assay. Our data in primary BMSCs revealed that cells treated with only 100 nM of CK2. 3   3. Treatment of cells with signaling inhibitor/siRNAs, followed by CK2.3 stimulation, was repeated for a total of two times over the course of a 6-day experiment. Later, the cells were fixed and stained for calcium deposits using von Kossa assay. Data (n = 4) was normalized to unstimulated cells and analyzed using one-way anova and Tukey-Kramer post hoc statistical test (p < 0.05). a = statistically significant difference to unstimulated, b = statistically significant difference to 100 nM CK2.3, c = statistically significant difference to 200 pM control siRNA + 100 nM CK2.3, d = statistically significant difference to 200 pM Smad4 siRNA + 100 nM CK2.3, and e = statistically significant difference to 500 nM ERK1/2 Inh + 100 nM CK2.3.

Discussion
Osteoporosis is a skeletal disorder and it is associated with loss of bone mass, structure, and integrity over a period of years, thereby increasing the incidence of fractures either at the hip, spine, or wrist. Mortality and morbidity of patients following hip and vertebral fractures is a common consequence of osteoporosis and has a far-reaching impact on the individuals' quality of life [69]. It is observed that even though women are disproportionately affected by osteoporosis, men are at a higher risk of mortality post hip surgery, at a rate of 36% compared to women at 21%. Mortality rate is seen to be high in the months following hip surgery and peaks at 1 year [70]. We have developed a BMPRIA mimetic peptide called CK2.3 and we have reported about its unique capability to induce calcification and bone growth in in vitro and in vivo studies, respectively [16][17][18][19]33]. However, the mechanism of CK2.3-mediated osteogenesis had been elusive. Here, we undertook the task of characterizing the molecular sequence of events initiated by CK2.3 inducing the differentiation of C2C12 cells and primary BMSCs towards osteoblasts.
Our RT-qPCR data show that OSX is significantly upregulated on day 2 and day 3 post-CK2.3 stimulation ( Figure 1B) but did not induce a statistically significant increase in RUNX2 mRNA expression ( Figure 1A). It is well known that RUNX2 is indispensable in the process of osteogenesis, homozygous deletion of RUNX2 in mice resulted in complete loss of osteoblasts [71][72][73] and haploinsufficiency of RUNX2 in humans leads to a disease known as cleidocranial dysplasia, which is characterized by defects in bone formation [42,71,74,75]. It also functions by inducing expression of other genes necessary for osteoblast differentiation, such as Osterix [72]. OSX is required for the 3. Treatment of cells with signaling inhibitor/siRNAs, followed by CK2.3 stimulation, was repeated for a total of two times over the course of a 6-day experiment. Later, the cells were fixed and stained for calcium deposits using von Kossa assay. Data (n = 4) was normalized to unstimulated cells and analyzed using one-way anova and Tukey-Kramer post hoc statistical test (p < 0.05). a = statistically significant difference to unstimulated, b = statistically significant difference to 100 nM CK2.3, c = statistically significant difference to 200 pM control siRNA + 100 nM CK2.3, d = statistically significant difference to 200 pM Smad4 siRNA + 100 nM CK2.3, and e = statistically significant difference to 500 nM ERK1/2 Inh + 100 nM CK2.3.

Discussion
Osteoporosis is a skeletal disorder and it is associated with loss of bone mass, structure, and integrity over a period of years, thereby increasing the incidence of fractures either at the hip, spine, or wrist. Mortality and morbidity of patients following hip and vertebral fractures is a common consequence of osteoporosis and has a far-reaching impact on the individuals' quality of life [69]. It is observed that even though women are disproportionately affected by osteoporosis, men are at a higher risk of mortality post hip surgery, at a rate of 36% compared to women at 21%. Mortality rate is seen to be high in the months following hip surgery and peaks at 1 year [70]. We have developed a BMPRIA mimetic peptide called CK2.3 and we have reported about its unique capability to induce calcification and bone growth in in vitro and in vivo studies, respectively [16][17][18][19]33]. However, the mechanism of CK2.3-mediated osteogenesis had been elusive. Here, we undertook the task of characterizing the molecular sequence of events initiated by CK2.3 inducing the differentiation of C2C12 cells and primary BMSCs towards osteoblasts.
Our RT-qPCR data show that OSX is significantly upregulated on day 2 and day 3 post-CK2.3 stimulation ( Figure 1B) but did not induce a statistically significant increase in RUNX2 mRNA expression ( Figure 1A). It is well known that RUNX2 is indispensable in the process of osteogenesis, homozygous deletion of RUNX2 in mice resulted in complete loss of osteoblasts [71][72][73] and haploinsufficiency of RUNX2 in humans leads to a disease known as cleidocranial dysplasia, which is characterized by defects in bone formation [42,71,74,75]. It also functions by inducing expression of other genes necessary for osteoblast differentiation, such as Osterix [72]. OSX is required for the differentiation of pre-osteoblasts into fully functioning osteoblasts, by inhibiting the differentiation of MSCs towards chondrogenesis [42,76]. Furthermore, deletion of OSX in mouse embryos results in total absence of osteoblasts and bone formation in the embryo stage [77]. It has been reported that OSX functions downstream of RUNX2 during osteoblast differentiation since in RUNX2-null mice, the expression of OSX was completely abolished, however, deletion of OSX in embryos had no effect on the expression of RUNX2 [77]. However, there are also some conflicting reports that suggest that RUNX2 is not involved in the induction of OSX [78] and BMP2 treatment induced OSX expression in RUNX2-deficient mesenchymal cells [79]. Collectively, these reports imply that OSX expression was induced and activated independent of RUNX2.
We show, through immuno-precipitation of BMPRIA, followed by a Western blot, that CK2.3 induces the release of CK2 from BMPRIA starting at 6 h and reaching complete dissociation at 18 h post-stimulation ( Figure 2). Simultaneously, CK2.3 colocalized with CK2 starting at 6 h post stimulation (Figure 3), indicating that CK2.3 binds to CK2 and causes it to dissociate from the BMPRIA receptor, thereby activating BMPRIA downstream signaling pathways and osteogenic specific genes. We developed a fluorescent bio-imaging CK2.3-Qdot ® s probe to track CK2.3 and visualize its interaction with intracellular proteins. In Figure 3, we show that CK2.3-Qdot ® s interacted with CK2α starting at 6 h and the interaction was maximal at 12 h post stimulation. We further wanted to verify the interaction between CK2.3 and CK2α using an immuno-precipitation method, and also determine whether CK2.3 phosphorylates CK2α protein, but due to the low molecular weight of CK2.3 (3.6 kDa), it has been very difficult to characterize the effect of CK2.3 on protein kinase CK2.
We have previously shown that CK2.3 induces osteogenesis downstream of BMPRIA, independent of BMP2 ligand. However, the function of CK2.3 was not well characterized. Here, we delineated the mechanism of CK2.3. Our Western blot ( Figure 4) and signaling inhibition data in C2C12 cells (Figures 5 and 6) and primary BMSCs (Figure 7) reveal that CK2.3 mediates osteogenesis via ERK1/2 and Smad1/5/8 signaling pathways. Interestingly, based on our Western blot data, it appears that CK2.3 mediates the phosphorylation of either only Smad1 or Smad5 signaling molecule ( Figure 4A). The role of Smad1/5/8 in mediating osteogenesis is well-documented [22,80,91]. Furthermore, Smad1/5/8 molecules are reported to directly bind to DNA sequences [92] and regulate the expression of target genes known to be involved in osteoblast differentiation [93,94] such RUNX2 [95] and Dlx5 [78,96], which then induces the expression of other downstream osteoblastic genes. Even though the role of ERK signaling in mediating osteogenesis is less extensively studied, it is still a critical player [97][98][99][100]. The ERK signaling pathway is reported to mediate OSX expression [101] and MAP kinases can also induce the expression of ALP and OCN [58,59].
Our Western blot data ( Figure 4) revealed that stimulation of C2C12 cells with 100 nM of CK2.3 activated Smad1/5/8, ERK1/2, and Akt1/2/3 signaling pathways; this can be attributed to the fact that most signaling pathways share common mediators and hence there is extensive cross-talk amongst one another [55]. It has been reported that Ras family of small GTPases, a member of the ERK signaling pathway, interacts [102] and regulates PI3K of Akt pathway [103]. Furthermore, the upregulation in expression of ERK1/2 and Smad1/5/8 molecules by CK2.3 could be due to the involvement of Ras/MEK/ERK pathway in the activation of Smad1. Yue et al. (2000) have shown that in untransformed epithelial cells, inactivation of Ras/MEK pathway greatly affected the phosphorylation of endogenous Smad1 by TGF-β and BMP [104]. Additionally, four consensus ERK phosphorylation sites are reported to be located in Smad1 and these ERK residues are reported to be essential in facilitating the interaction between Smad1-Smad4 complex, nuclear translocation of the complex, and activation of the Smad binding element (SBE) [105].
Taking all the data together, we propose the following mechanism of CK2.3 mediated osteogenesis. In native C2C12 cells, we hypothesize that protein kinase CK2 interacts with BMPRIA ( Figure 8A). Stimulation of cells with CK2.3 results in its internalization starting at 6 h through caveolae mediated endocytosis ( Figure 8B). At 12 h post stimulation, CK2.3 co-localizes with endogenous CK2α, which then results in the release of CK2 from the receptor BMPRIA ( Figure 8C). The dissociation thereby activates BMPRIA downstream signaling pathways such as Smad1/5/8 and ERK1/2, followed by upregulation of osteoblast specific genes like OSX, ALP, and OCN ( Figure 8D). This ultimately causes the differentiation of C2C12 cells into osteoblasts. upregulation in expression of ERK1/2 and Smad1/5/8 molecules by CK2.3 could be due to the involvement of Ras/MEK/ERK pathway in the activation of Smad1. Yue et al. (2000) have shown that in untransformed epithelial cells, inactivation of Ras/MEK pathway greatly affected the phosphorylation of endogenous Smad1 by TGF-β and BMP [104]. Additionally, four consensus ERK phosphorylation sites are reported to be located in Smad1 and these ERK residues are reported to be essential in facilitating the interaction between Smad1-Smad4 complex, nuclear translocation of the complex, and activation of the Smad binding element (SBE) [105]. Taking all the data together, we propose the following mechanism of CK2.3 mediated osteogenesis. In native C2C12 cells, we hypothesize that protein kinase CK2 interacts with BMPRIA ( Figure 8A). Stimulation of cells with CK2.3 results in its internalization starting at 6 h through caveolae mediated endocytosis ( Figure 8B). At 12 h post stimulation, CK2.3 co-localizes with endogenous CK2α, which then results in the release of CK2 from the receptor BMPRIA ( Figure 8C). The dissociation thereby activates BMPRIA downstream signaling pathways such as Smad1/5/8 and ERK1/2, followed by upregulation of osteoblast specific genes like OSX, ALP, and OCN ( Figure 8D). This ultimately causes the differentiation of C2C12 cells into osteoblasts.

CK2.3 Peptide
CK2.3 (1.915 μM) is a custom designed osteogenic peptide synthesized at GenScript (Piscataway, NJ, USA). It is a 29 amino acid long peptide and has a molecular weight of 3655.34 Da. It consists of the CK2 phosphorylation site, which is found on BMPRIA, at amino acids 213-217 (SLKD)

CK2.3 Peptide
CK2.3 (1.915 µM) is a custom designed osteogenic peptide synthesized at GenScript (Piscataway, NJ, USA). It is a 29 amino acid long peptide and has a molecular weight of 3655.34 Da. It consists of the CK2 phosphorylation site, which is found on BMPRIA, at amino acids 213-217 (SLKD) incorporated in its sequence, along with the Antennapedia homeodomain signal sequence at the N-terminal of the peptide. Our lab has an active IACUC approved protocol (AUP #1194, protocol approved on 2/11/2017) to isolate BMSCs from C57BL/6J mice. All animals for this study were under the direct care and supervision of the Office of Laboratory Animal Medicine at University of Delaware. All animal users complied with IACUC and OLAM guidelines for the care and use of laboratory animals. Collection and culturing of BMSCs were carried out as per the following publication [66]. Cells were cultured in MEM-Alpha Medium (Corning Cellgro, Manassas, VA, USA) supplemented with 20% (v/v) heat in-activated FBS (Gemini Bio-products, West Sacramento, CA, USA), and 1% (v/v) penicillin/streptomycin (Hy-clone, Pittsburgh, PA, USA). Then, 0.25% Trypsin (Cellgro, Manassas, VA, USA) containing 0.02% EDTA (Fisher Chemical, Fair Lawn, NJ, USA) was used to detach the cells from the flasks.

Quantitative Reverse Transcription Polymerase Chain Reaction
C2C12 cells were grown in 35 mm dishes at a cell density of 4.5 × 10 4 cells/dish for two days in DMEM media containing 20% FBS and 1% penicillin/streptomycin, and at the end of day 2, cells were serum starved with DMEM media supplemented with 1% penicillin/streptomycin, but without FBS for 18 h. Following starvation, cells were either left unstimulated or stimulated with 100 nM of CK2.3 for 6 h, 12 h, 18 h, day 1, day 2, day 3, day 4, and day 5. After each respective time point, total RNA was isolated using Trizol (Ambion life technologies, Carlsbad, CA, USA) method in accordance with Direct-zol™ RNA miniprep manufactures protocol (catalog #R2050, Zymo Research, Irvine, CA, USA). Two-step RT-PCR was performed with 2 µg of RNA obtained, using high-capacity cDNA reverse transcription kit (catalog #4368814, Thermo Fisher Scientific, Waltham, MA, USA). In the second step, cDNA obtained was used for PCR amplification using specific primers. The primers were purchased from Integrated DNA Technologies (Coralville, IA Following blocking, cells were labeled for CK2α using 100 µL (2 µg/mL) of rabbit polyclonal IgG casein kinase IIα antibody (catalog #sc-9030, Santa Cruz Biotechnology, Dallas, TX, USA) diluted in 3% protease-free BSA in the ratio of 1:100, at room temperature for 1 h. The primary antibody was then stained against using 100 µL (1 µg/µL) of fluorescently tagged-secondary antibody, donkey anti-rabbit IgG H&L (Alexa Fluor ® 568, catalog #ab175470, Abcam, Cambridge, MA, USA) diluted in the ratio of 1:500 in 3% protease-free BSA, at room temperature for 1 h. Later, the nucleus was stained using 100 µL   × 3). After lysing the cells, the cell debris was removed by centrifugation at 13,000× g for 20 min and the supernatant containing the cytoplasmic extract was transferred to a new tube [111]. Protein content of the cell lysates were determined using the BCA assay (Thermo Scientific, Rockford, IL, USA) and were normalized to equal protein content. Protein extracts were analyzed using SDS-polyacrylamide gel electrophoresis followed by Western blot. The samples were analyzed for Smad1

Immunoprecipitation
C2C12 cells were grown in 6 well plates at 1.0 × 10 5 cells/mL for two days in DMEM with 10% FBS and 1% penicillin/streptomycin. After the second day, the cells were serum starved for 18 h. They were then stimulated with 100 nM of CK2.3 or left unstimulated for 6 h, 12 h, and 18 h. After the designated time point the cells were lysed following the same protocol as outlined above in the Western blot section. PureProteome Protein G magnetic beads (EMD Millipore, Burlington, MA, USA) were incubated and washed three times with lysis buffer (NP-40), centrifuging at 14,000 rpm for five minutes at 4 • C. The beads were then incubated with BMPRIA (rabbit polyclonal, Santa Cruz, Dallas, TX, USA) for 4 h at 4 • C. After lysates were collected, they were incubated with BMPRIA/magnetic beads overnight at 4 • C. The beads/BMPRIA were then washed three times with lysis buffer (NP-40), centrifuging at 14,000 rpm for five minutes at 4 • C. The IP lysates were analyzed using SDS-polyacrylamide gel electrophoresis, followed by a Western blot. They were detected for BMPRIA (rabbit polyclonal, Santa Cruz, Dallas, TX, USA), and CK2β (rabbit polyclonal, Santa Cruz, Dallas, TX, USA).

siRNA Transfection
C2C12 cells were grown in a 24-well plate at a cell density of 1 × 10 4 cells/well for two days in DMEM media containing 20% FBS and 1% penicillin/streptomycin and at the end of the second day, cells were serum starved for 18 h with DMEM media containing only 1% penicillin/streptomycin. Cells were then either left unstimulated or stimulated with 200 pM of Smad4 siRNA (m) (catalog #sc-29485, Santa Cruz Biotechnology, Dallas, TX, USA) and 200 pM control siRNA-A (catalog #sc-37007, Santa Cruz Biotechnology, Dallas, TX, USA) for 24 h. Cells were transfected with siRNA using Turbofect transfection reagent (catalog #R0531, Thermo Scientific, Waltham, MA, USA) as per the protocol. The next day, cells were either left untreated or treated with 100 nM of CK2.3. Transfection of cells with siRNA, followed by 100 nM of CK2.3 was repeated again on day 4. On the sixth day, cells were either fixed using 4.4% (w/v) paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA) and stained for mineralization using von Kossa assay or viability of the cells was determined using MTT reduction assay.

von Kossa Assay
von Kossa assay is a biochemical test used to quantify the amount of mineralization. In this assay, the silver cations of the silver nitrate solution react with phosphates and carbonates present in the calcium deposits, which are then reduced to black metallic silver stain when exposed to UV light [63,112].

Inhibition of BMPRIA Downstream Signaling Pathways in C2C12 Cells
C2C12 cells were grown in a 24-well plate at a cell density of 1 × 10 4 cells/well for two days in DMEM media containing 20% FBS and 1% penicillin/streptomycin and at the end of the second day, cells were serum starved for 18 h with DMEM media containing 1% penicillin/streptomycin and without FBS. Cells were then either left unstimulated or stimulated with 500 nm, 1 µM, and 5 µM of U0126-EtOH (MEK1/2 inhibitor, catalog #S1102, Selleckchem, Houston, TX, USA). U0126-EtOH is an inhibitor of intracellular Raf/MEK/ERK signaling pathway. One set of C2C12 cells were either left unstimulated or stimulated with 500 nm, 1 µM, and 5 µM of MK-2206 2HCl (Akt1/2/3 inhibitor, catalog #S1078, Selleckchem, Houston, TX, USA). The Akt inhibitor, MK-2206 2HCl is reported to inhibit autophosphorylation of Akt at T308 and S473, as well as prevent Akt-mediated activation of downstream molecules. Also, one more set of cells were either left unstimulated or stimulated with 500 nm, 1 µM, 5 µM, and 10 µM of SB202190 (FHPI) (p38 MAPK inhibitor, catalog #S1077, Selleckchem, Houston, TX, USA). SB202190 (FHPI) is a potent inhibitor of p38 MAPK and it functions by targeting p38α/β. The above-mentioned inhibitors have been previously used to inhibit the respective signaling pathways in C2C12 cells [113][114][115]. The cells were stimulated with the reported concentrations of the inhibitors on day 1 and then again on day 4. Cells were also stimulated with 100 nM of CK2.3 on day 2 and day 5. Further, cells were supplemented with DMEM media containing 10% FBS and 1% penicillin/streptomycin on day 3 and day 6. After 6 complete days of stimulation, cells were fixed using 4.4% (w/v) paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA) for 20 min and washed with ice cold 1X PBS pH 7.4. 5% (w/v) silver nitrate (Chem-Impex international, Wood Dale, IL, USA) dissolved in diH 2 O, which was applied to each well, and the plate was put under the UV light for 1 h. Cells were then washed with diH 2 O and were allowed to dry overnight. Twelve random images of the cells were taken per well using Zeiss Axiovert 10 microscope at 5×/.12 The achrostigmat objective and images were analyzed using ImageJ software (NIH, Bethesda, MD, USA). Images were first converted to 8 bit and threshold was set to the positive control. The same threshold was used for all treatments. Mineralized areas were quantified using the analyzing particles plugin function of ImageJ.

Inhibition of Smad1/5/8 and ERK1/2 Signaling Pathways in Primary BMSCs
Primary BMSCs isolated from the tibia and femur of 4-month-old female C57BL/6J mice were grown in a 24-well plate at a cell density of 1 × 10 4 cells/well in MEM alpha medium containing 20% FBS and 1% penicillin/streptomycin. Once the cells reach 90% confluency, cells were then either left unstimulated or stimulated with 500 nM of U0126-EtOH (ERK inhibitor, catalog #S1102, Selleckchem, Houston, TX, USA), 200 pM of Smad4 siRNA (m) (catalog #sc-29485, Santa Cruz Biotechnology, Dallas, TX, USA), and 200 pM of control siRNA-A (catalog #sc-37007, Santa Cruz Biotechnology, Dallas, TX, USA) on day 1 and then again on day 4. Cells were also stimulated with 100 nM of CK2.3 on day 2 and day 5. After 6 complete days of stimulation, cells were fixed using 4.4% (w/v) paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA) for 20 min and washed with ice cold 1X PBS pH 7.4. 5% (w/v) silver nitrate (Chem-Impex international, Wood Dale, IL, USA) dissolved in diH 2 O, which was applied to each well and the plate was put under the UV light for 1 h. Cells were then washed with diH 2 O and were allowed to dry overnight. Twelve random images of the cells were taken per well using Zeiss Axiovert 10 microscope at 5×/.12 The achrostigmat objective and images were analyzed using ImageJ software (NIH, Bethesda, MD, USA). Images were first converted to 8 bit and threshold was set to the positive control. The same threshold was used for all treatments. Mineralized areas were quantified using the analyzing particles plugin function of ImageJ. 4.9. MTT reduction assay.
MTT reduction assay is a marker for viable cell metabolism. Viable cells with active metabolism reduce MTT into formazan (a purple colored byproduct), however, when cells die, they lose the ability to convert MTT into formazan. The quantity of formazan is linearly proportional to the number of viable cells and measured by changes in absorbance at 570 nm using a spectrophotometer [116]. C2C12 cells were grown in a 24-well plate at a cell density of 1 × 10 4 cells/well for two days in DMEM media containing 20% FBS and 1% penicillin/streptomycin, and at the end of the second day, cells were serum starved for 18 h with DMEM media containing only 1% penicillin/streptomycin. Cells were then either left unstimulated or stimulated with 500 nM of U0126-EtOH (ERK1/2 inhibitor, catalog #S1102), 5 µM of MK-2206 2HCl (Akt1/2/3 inhibitor, catalog #S1078), 10 µM of SB202190 (FHPI) (p38 MAPK inhibitor, catalog #S1077), 200 pM of Smad4 siRNA (m) (catalog #sc-29485), and 200 pM of control siRNA-A (catalog #sc-37007 on day 1 and day 3. Cells were further treated with 100 nM of CK2.3 on day 2 and day 5, cells were also supplemented with DMEM media containing 10% FBS and 1% penicillin/streptomycin on days 4 and 5 of the experiment. On the sixth day, viability of the cells following treatment with inhibitors and CK2.3 was determined using MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide) assay per the manufacturer's protocol (catalog #M6494, Thermo Fisher Scientific, Waltham, MA, USA).

Statistical Data Analaysis
All the results are presented as mean ± standard error of the mean (SEM). All the experiments were repeated at least three times, outliers were removed from the data based on Chauvenet's criterion [117], and data was analyzed using the single factor analysis of variance (ANOVA), followed by Tukey-Kramer post hoc statistical test at a 95% confidence level.

Conflicts of Interest:
The authors declare no conflict of interest.