Magnesium Chloride promotes Osteogenesis through Notch signaling activation and expansion of Mesenchymal Stem Cells

Mesenchymal stem cells (MSC) are osteoblasts progenitors and a variety of studies suggest that they may play an important role for the health in the field of bone regeneration. Magnesium supplementation is gaining importance as adjuvant treatment to improve osteogenesis, although the mechanisms involving this process are not well understood. The objective of this study was to investigate the effects of magnesium on MSC differentiation. Here we show that in rat bone marrow MSC, magnesium chloride increases MSC proliferation in a dose-dependent manner promoting osteogenic differentiation and mineralization. These effects are reduced by 2-APB administration, an inhibitor of magnesium channel TRPM7. Of note, magnesium supplementation did not increase the canonical Wnt/β-catenin pathway, although it promoted the activation of Notch1 signaling, which was also decreased by addition of 2-APB. Electron microscopy showed higher proliferation, organization and maturation of osteoblasts in bone decellularized scaffolds after magnesium addition. In summary, our results demonstrate that magnesium chloride enhances MSC proliferation by Notch1 signaling activation and induces osteogenic differentiation, shedding light on the understanding of the role of magnesium during bone regeneration.

Magnesium (Mg 2+ ) is particularly interesting because of its abundance in the organism, where it participates in numerous biological processes, like osteogenesis of progenitor cells. In addition, low Mg 2+ concentrations have been associated with osteoporosis or osteopenia 6 . In biomaterial engineering, Mg-based implants have been used to enhance bone formation "in vivo" [7][8][9][10] . Due to corrosion resistance of Mg 2+ alloys 11 , new formulas based on Mg 2+ are being investigated. Currently, Mg 2+ is used in combination with other materials as calcium and phosphate, or as Mg-coated structures to enhance osteogenesis. However, little is known about the mechanisms whereby Mg 2+ salts affect osteogenesis and bone formation. It is known that Wnt/β-pathway and Notch signaling are involved in bone marrow mesenchymal stem cells (MSC) osteogenesis 12,13 . Although other works show that Notch signaling activation is also related to the maintenance of stemness 14 even with the inhibition of osteogenesis 15 .
The characterization of the pro-osteogenic effects of Mg 2+ will add knowledge on the biology of bone cells; thus, new or improved strategies can be developed aiming to enhance osteogenesis, and osseointegration of bone prosthesis.
The present study evaluates the effects of magnesium chloride on osteogenesis of bone marrow MSC and its capability to repopulate decellularized bone scaffolds. Moreover, the mechanisms whereby magnesium chloride triggers its pro-osteogenic effect are also investigated.
The enhancement of mineralization was accompanied by a significant increase in the expression of osteogenic master genes such as RUNX2 (Fig. 1c), regulator of the early osteoblast differentiation, OSTERIX (Fig. 1d), transcription factor required for the transition of pre-osteoblasts to osteoblasts, and OSTEOCALCIN ( Fig. 1e), produced by mature osteoblasts. The addition of MgCl 2 increased the expression of these genes in a concentration-dependent manner. As FGF23 is released by mature osteoblasts and osteocytes, the amount of FGF23 in the supernatant demonstrates the presence of mature osteoblasts. FGF23 production during 24 h was measured at 21 days of osteogenic differentiation. Undifferentiated cells did not produce FGF23 while in MSC differentiated into osteoblasts with 1.8 mM of MgCl 2 FGF23 levels were 31-fold increased as compared with 0.8 mM (Fig. 1f).
The inhibition of the Mg 2+ transporter TRPM7 by 2-APB produced a decrease in ALP activity (Fig. 2a) as well as in matrix calcification (Fig. 2b). It also resulted in a significant down-regulation of the osteogenic marker genes RUNX2, OSTERIX and OSTEOCALCIN (Fig. 2c-e, respectively). Moreover, 2-APB treatment reduced FGF23 secretion (Fig. 2f).
Protein levels of Cyclin D1 and PCNA, markers of cellular proliferation, were analyzed by Western blot. The expression of both proteins was increased in osteoblasts as compared with undifferentiated controls. The addition of Mg 2+ to the medium induced a dose dependent increase in both PCNA and Cyclin D1 ( Fig. 2g and h). Notably, the inhibition of the Mg 2+ channel TRPM7 with 2-APB significantly reduced Cyclin D1 and PCNA protein expression.
Magnesium increases osteogenic differentiation through activation of Notch1 signaling in MSC. The Wnt/β-catenin and Notch signaling are pathways closely involved in bone development. The effect of Mg 2+ on both pathways was explored. Immunofluorescence analyses revealed that osteogenic differentiation is associated with nuclear translocation of β-catenin. However, increasing MgCl 2 concentrations in the culture medium did not produce a further increase the nuclear translocation of β-catenin ( Fig. 3a and b).
Confocal microscopy analysis of Notch1 intracellular domain (NICD) showed that Mg 2+ supplementation produced an increase in nuclear translocation of NICD (Fig. 4a). Osteogenic differentiation decreased the nuclear NICD as compared with UC while Mg 2+ supplementation increased the presence of nuclear NICD. In OB cells treated with 2-APB the DAPI-NICD co-localization was almost inexistent. As it is observed in the last column of Fig. 4a as compared to 0.8 mM, after Mg addition there was an increase of green pixels (NICD) matching with blue pixels (DAPI), demonstrating nuclear colocalization of NICD protein. Similarly, Fig. 4b shows that a moderate increase of Mg 2+ in the osteogenic medium upregulated the mRNA expression of HEY2, one of the classic Notch target genes. The opposite effect was observed by the inhibition of the Mg 2+ transporter TRPM7 with 2-APB, which reduced the nuclear translocation of NICD (Fig. 4a) and Notch activation (Fig. 4b). To examine a likely direct effect of MgCl 2 , the nuclear expression of NICD was evaluated after 24 h of Mg supplementation on MSC or MSC plus osteogenic stimulus. The addition of MgCl 2 for 24 h increased the nuclear protein expression of NICD as it was confirmed by western blotting (Fig. 4c). However, when MgCl 2 was administrated for 24 hours in presence of an osteogenic stimulus the levels of nuclear NICD expression were similar to those found with basal Mg 2+ content (Fig. 4d). The administration of MgCl 2 for 24 h to differentiated osteoblasts or 2-APB treated cells for 21 days did not modify the nuclear NICD protein expression (Fig. 4e). A high NICD protein expression was also detected by immunoblotting of nuclear protein extracts from MSC and differentiated osteoblasts with Mg 2+ supplementation after 21 days (Supplementary Figure S1). However, NICD expression was again highly induced in undifferentiated MSC. Taken together, these data suggest that Mg 2+ activates NICD nuclear translocation on undifferentiated MSC but not in differentiated osteoblasts.
Scientific RepORTS | 7: 7839 | DOI:10.1038/s41598-017-08379-y Magnesium promotes maturation and distribution of osteoblast into decellularized scaffolds. Scanning Electron Microscopy (SEM). In decellularized scaffolds (Fig. 5a), osseous matrix was observed as series of uniform and regular depressions corresponding to ducts osteons (Fig. 5b). Whole bone surface and cavities were covered with a thin membrane with scarce cells. In the scaffolds recellularized with MSC and treated with basal levels of Mg 2+ (0.8 mM) (Fig. 5c), bone matrix was covered with an irregular membrane with abundant cells following an unspecific distribution and presenting a stellate morphology. After Mg 2+ supplementation (1.2 mM), MSC formed a uniformly organized layer over the bone surface, of smooth aspect and with abundant cellular proliferation filling the osteon canals (Fig. 5d). In the scaffolds treated with 1.8 mM Mg 2+ , a continuous and regular layer of germinal bone tissue that fully occupied the scaffolds, was observed (Fig. 5e).
Transmission Electron Microscopy (TEM). Cellular layers on decellularized bone scaffolds were analyzed with TEM. During the osteoinduction of MSC with basal Mg 2+ levels (0.8 mM), collagen-producing blastic cells were poorly organized and with immature matrix (Fig. 6a). After Mg 2+ supplementation (1.2 mM), the number of osteoblasts increased and they were organized in layers, similar to appositional growth of bone, and more mature osteoblasts (osteocytes) were observed. These osteocytes showed bone canaliculi prolongation and a more organized collagen matrix (Fig. 6b). The highest concentration of MgCl 2 (1.8 mM) produced a further increase in the number of osteoblast layers (appositional growth), with more mature osteocytes, large bone canaliculi and an organized deposition of collagen (osteoid matrix) (Fig. 6c).

Discussion
In the present study, we have investigated the potential effects of MgCl 2 in bone marrow MSC during osteogenic differentiation including its effect on anchorage, cellular attachment and differentiation when they are cultured on decellularized bone scaffolds. As it is illustrated in Fig. 7, high concentrations of MgCl 2 (1.2 mM or 1.8 mM) enhanced proliferation of rat bone marrow MSC in a dose-dependent manner, and increased the subsequent osteogenesis. These effects were not observed after inhibition of the Mg 2+ channel TRPM7 by 2-APB, which confirmed the key role of intracellular Mg 2+ in the osteogenesis of MSC. We evaluated the intracellular pathways whereby Mg 2+ promotes osteogenesis of MSC, finding that MgCl 2 addition did not increase canonical Wnt/β-catenin pathway activation, although this pathway was activated under osteogenic stimuli. It was interesting to observe that MgCl 2 supplementation induced the nuclear translocation of NICD and HEY2 expression. Our results show a direct effect of Mg 2+ on Notch activation in MSC rather on differentiated osteoblasts from MSC (Fig. 4), suggesting a specific role of Mg 2+ on the maintenance of stemness of MSC rather on osteogenic process. Furthermore, moderate concentrations of MgCl 2 considerably promoted osteocyte maturation and enhanced cell attachment to the decellularized bone surface. Note that MgCl 2 increased the effects of the osteogenic stimuli while conditioned medium with basal levels of Mg 2+ only produced blastic cells, poorly organized, and without osteocytic phenotype.
The beneficial effects of several Mg 2+ alloys on bone formation have been widely reported. The balanced combination of Mg 2+ with different elements such as calcium and phosphate or in Mg-coating prosthesis has demonstrated osteoinductive effects 7,16 . These effects are supported by the formation of a better structure. The structure and surface characteristic of the biomaterials/scaffolds are key for the attachment and function of the cells and, it may affect the absorption and/or integration of proteins; and in turn, the quality of the anchorage, influences the subsequent cellular responses and tissue regeneration. Minardi et al. suggested that Mg 2+ provides the scaffolds with structural characteristics similar to those of bone, allowing anchorage and proliferation of progenitor cells 9 . Few studies have investigated the active effects of Mg 2+ on osteogenesis. Recently, Zhang et al. have demonstrated that cement formed by a combination of calcium, phosphate and Mg 2+ increased osteogenesis through a specific interaction between fibronectin and integrin α5β1 17 . Of interest, the authors observed a significant increase in Mg 2+ concentration (approx. 2.5 mM) after soaking these scaffolds with culture medium; this study support our results demonstrating that Mg 2+ salts may enhance osteogenesis of progenitor cells. Yoshizawa et al. observed that supra-physiological concentrations of Mg 2 SO 4 (10 mM) also promoted the expression of transcription factors related to COL10A1 expression 18 . Therefore, these findings suggest that Mg 2+ salts promote bone formation in  stimulates gene expression of TRPM7 channel and promotes osteoblasts proliferation 21 . Other authors have demonstrated the requirement of TRPM7 for growth and skeletogenesis 22 . The results of these studies are in line with our findings, where TRPM7 inactivation by 2-APB inhibited osteogenesis of bone marrow MSC and significantly reduced Notch1 signaling. The contrary effect was observed with the addition of moderate concentrations of MgCl 2 , which promoted Notch1 signaling activation and increased osteogenesis of the MSC. The central role of this pathway during osteogenesis has been described already 13,23 , although other works highlights too an inhibition of this pathway during osteogenesis 15 . The interrelationship among Notch1 signaling pathway, Mg 2+ and osteogenesis is unknown; taken together, our in vitro and in vivo results suggest that Mg 2+ supplementation promotes Notch activation tissue-specifically in MSC increasing the pool of osteoprogenitor cells susceptible to be differentiated into osteoblasts in the presence of osteogenic stimuli.
In a recent study, it is demonstrated that the inhibition of Notch pathway in MSC leads to a lesser proliferation and osteogenic capability avoiding fracture healing 24 . Zanotti et al. have observed that Notch effects in the skeleton are cell-context-dependent finding different effects on immature osteoblasts or osteocytes 25 . These and other works highlight the potential dimorphic effects of Notch signaling in bone homeostasis 26 .
Finally, and with respect to Mg 2+ supplementation more studies, considering other important parameters for bone turnover such as PTH or vitamin D, should be led to evaluate with precision the in vivo effects of Mg 2+ supplementation on the mineralization or the osteoid production in the bone.
In addition, we studied the effect of Mg 2+ on cell organization and osseointegration of progenitor cells on rat decellularized bone. Our results reveal that the supplementation with Mg 2+ improves cell attachment and increases osteogenic differentiation of MSC cultured on decellularized bone. Transmission electron microscopy analysis showed that Mg 2+ supplementation increased proliferation and maturation of MSC differentiated into osteocytes with the presence of cytoplasmic prolongations between osteocytes (bone canaliculi) and a well-organized osteoid matrix with a large number of layers similar to the appositional growth observed in bones. Similarly, using scanning electron microscopy analysis, it was shown that Mg 2+ induced proliferation, differentiation and a more organized distribution of the osteocytes coating the surface of decellularized bone, allowing mineralization and formation of new bone. These data indicate that the bone tissue formed on decellularized rat bone after Mg 2+ supplementation has a better internal distribution with highly organized structures. Several studies have shown that Mg 2+ ions may enhance cell attachment and promote bone formation. Moreover, previous studies have demonstrated that Mg 2+ ions support the initial cell adhesion of MSC and increase proliferation and activation of integrins 27 . Recently, Galli et al. also showed that the immersion of threaded screws in a solution of MgCl 2 (10 mg/ml) leads to enhanced osteogenesis and improves osseointegration 28 .
The data reported in this study could represent a new, feasible and economic strategy to improve bone formation in different types of scaffolds. In addition, these results also suggest that local administration of moderately high levels of Mg 2+ could be suitable to promote osteogenesis in pathologies in which bone formation is necessary.
This approach would significantly impact the orthopaedic field, and further provides new targets to improve the quality of the materials currently available.
Osteogenic differentiation of MSC and treatments. MSC were cultured for 21 days in αMEM with FBS (10%), ultraglutamine (1%), penicillin (100 U/mL), streptomycin (100 µg/ml) and osteogenic stimuli based on dexamethasone (1 µM; Sigma-Aldrich), β-glycerol phosphate (10 mM; Sigma-Aldrich) and ascorbic acid (0.2 mM; BAYER, Barcelona, Spain). Basal Mg 2+ concentration in the medium was 0.8 mM. In order to increase the Mg 2+ content in the pro-osteogenic medium, MgCl 2 (Carlo ErbaReagentiSpA, Milano, Italy) was added to achieve final Mg 2+ concentrations of 1.2 mM and 1.8 mM during the osteogenic stimulus. Fresh medium alone or osteogenic medium supplemented with MgCl 2 was replaced every 3 days. Furthermore, 2-APB (50 µM; Tocris Bioscience, Bristol, UK) was added to the osteogenic medium containing 0.8 mM of Mg 2+ during differentiation to determine the effects of inhibiting the Mg 2+ channel TRPM7. To examine a direct effect of Mg 2+ supplementation (1.2 and 1.8 m, or 2-APB) for 24 hours on undifferentiated MSC or at the different stages of differentiation (early or mature osteoblasts obtained from MSC) the NICD expression was also analyzed by western blot. All experiments were repeated at least three times.

RNA isolation and quantitative RT-PCR.
Total RNA was extracted with TRI reagent (1 mL; Sigma-Aldrich) and quantified by spectrophotometry (ND-1000, Nanodrop Technologies, Wilmington, DE, USA). cDNA was synthesized from 1 μg of total RNA with a first strand cDNA synthesis kit (Qiagen, Alkaline phosphatase activity quantification. 2 µg of cytoplasmic cell lysates were incubated in p-nitrophenol phosphate (2 mM; Sigma-Aldrich) for 30 min at 37 °C. The reaction was stopped by adding NaOH (3 M), and alkaline phosphatase (ALP) activity was measured by quantifying absorption at 405 nm. ALP activity was expressed as µmol of hydrolyzed p-nitrophenol phosphate per min and per mg of protein versus undifferentiated control cells.
Alizarin red S staining. Matrix mineralization was detected by alizarin red S staining. Cells were washed twice with PBS, fixed with para-formaldehyde (2%) and sucrose (1%) for 15 min and subsequently washed 3 times with PBS. Then, cells were stained with alizarin red S pH 4.1 (40 mM; Sigma-Aldrich) for 20 min, and washed 4 times for 5 min with water at pH 7. Finally, water was removed and samples were dried at room temperature. Plates were scanned in a WIFI OKI Scanner (Madrid, Spain).
FGF23 secretion quantification. Supernatants from cultures were collected and pooled, and intact fibroblast growth factor 23 (FGF23) secretion was determined by using a specific ELISA kit (Kainos Laboratories, Tokyo, Japan).

Rat femurs decellularization and scaffolds preparation.
Bone scaffolds were obtained as it has been previously reported by Shahabipour 29 . Briefly, rat femurs and tibiae were cut longitudinally in 5 mm pieces. Subsequently, bone pieces were boiled 4 times for 5 minutes to remove fat tissues. Pieces were stored overnight at −20 °C before decellularization. Bone specimens were thawed at room temperature, washed with PBS and placed in liquid nitrogen for 2 minutes. Then, pieces were maintained in distilled water at room temperature and washed with PBS. The freeze-thaw process was repeated five times to lysate the cells. Bone scaffolds were decellularized in SDS (2.5%) for 24 hours at 37 °C with gentle shaking. Then, bone specimens were washed with PBS twice for 15 minutes to remove SDS, washed in ethanol (70%) and maintained in PBS for 30 minutes with shaking at room temperature.
Scanning Electron Microscope (SEM). Decellularized bone fragments were cut in small pieces and washed in buffer solution for 15 minutes to remove debris. Then samples were kept in glutaraldehyde (2.5%). Decellularized bones without addition of MSC were also analyzed to ensure the decellularization process. Bone scaffolds were mounted on the SEM specimen stubs with carbon tape and were carbon-coated. Samples were analyzed and photographed with a Hitachi S520 SEM (Tokyo, Japan).
Scientific RepORTS | 7: 7839 | DOI:10.1038/s41598-017-08379-y Transmission Electron Microscope (TEM). Bone marrow MSC were cultured in the presence of different concentrations of MgCl 2 (0.8, 1.2 and 1.8 mM) on rat femur scaffolds. After 21 days of culture, the cellular layer was removed from the surface of the bone scaffolds and they were analyzed by electronic microscopy. For the ultrastructural study, randomly selected samples of decellularized bone scaffolds were primarily fixed in a glutaraldehyde (2%) solution in phosphate buffer (0.1 M) pH 7.4 overnight at 4 °C and then re-fixed in osmium tetroxide (1%) in phosphate buffer (0.1 M) pH 7.4 for 30 minutes. After dehydration in graded ethanol series and embedding in araldite, semi-thin and ultra-thin sections were cut with a LKB ultramicrotome. Ultra-thin sections were viewed and photographed in a Philips CM10 transmission electron microscope.