Following implantation of demineralized bone matrix (DBM), MSCs undergo directed migration in response to matrix chemoattractants [40]. Indeed, the induction of bone formation requires three key components: an osteoinductive soluble signal, an insoluble substratum, and responding host’s cells [41, 42]. We hypothesized that CaSO4 could act as a promigratory signal and would induce the recruitment of such endogenous cells. We checked this hypothesis using a gelatin sponge as a substratum to provide initial attachment to the osteoprogenitor cells and agarose to both provide a sustained Ca2+ release and serve as a binding agent. Our findings strongly suggest that CaSO4 has the ability to recruit osteoprogenitor cells in vitro and in vivo.

Our results showed that there is an optimal range of CaSO4 concentration to promote MSC migration. In our model this range is between 3 and 5 mM in vitro, whereas in vivo a threshold for an osteoinductive effect was determined in those scaffolds soaked in a solution of CaSO4 20 mM. Yamaguchi et al. [5] found that exposure of MC3T3-E1 cells to high CaCl2 (up to 4.8 mM) in vitro resulted in dose-dependent chemotaxis stimulation. In addition, we also observed that the addition of higher CaSO4 concentrations disturbed MSC migration in vitro and bone regeneration in vivo according to the micro-CT and histological analysis. It has been shown that extracellular fluid at sites of injury, infection, or inflammation contains high concentrations of calcium [43]. Interestingly, rapid resorption of CaSO4 results in a Ca2+-rich fluid that could modulate inflammation and apoptosis [44, 45].

Furthermore, we evaluated the effect of CaSO4 on MSC differentiation. Osteoblast differentiation is regulated by the sequential expression of several osteogenic marker genes [46, 47]. In this study, CaSO4 lowered the expression of Osterix at 24 hours and Alpl or Osteocalcin up to 4 days but increased all of them after 10 days in culture. In agreement with our results, Lazary et al. [48] have shown that expression of Osteocalcin (Bglap), Bone Sialoprotein (Ibsp), and Col1a1 was decreased when MC3T3-E1 cells were cultured on CaSO4 discs or with medium supplemented with CaCl2 25 mM. The Smad pathway transduces signals from BMP receptors and leads to transcriptional induction of key osteogenic transcription factors such as Runx2 and Osterix. It has been shown that other growth factors, such as TGF-β, HGF, EGF, FGF, or IGF, also induce a chemotactic response on MSCs. However, these growth factors individually also induce an antagonistic effect on BMP-induced osteoblast differentiation by inhibiting the nuclear accumulation of Smads [4951].

In addition, it has been reported that genes downstream of G-protein coupled receptor (GPCR) signaling pathways may be the earliest response to calcium-based ceramics [52]. Indeed, calcium sensing receptor (CaSR), a receptor belonging to the GPCR family, modulates the chemotactic response of MSCs in response to extracellular calcium [5, 6, 53]. It has been reported that CaSO4 induces a significant increment in Smad3 and Smad6 expression [48]. Smad3 and Smad6 have an inhibitory effect on BMP signaling during osteoblast differentiation by targeting the BMP-responsive Smad1/5/8 complex [49, 54]. It has been demonstrated that high extracellular calcium decreased the levels of phosphorylated SMAD1, ERK 1/2, and p38 and that Ca2+ could bind BMP-2 extracellularly [55, 56]. In contrast, low concentrations of extracellular calcium (0.18 mM) enhanced BMP-2-induced osteogenic differentiation [55]. Interestingly, Sr2+ produces a similar inhibitory effect on BMP-2 osteogenic capacity via the canonical Smad pathway [57]. Altogether, these results suggest that CaSO4 transiently attenuates BMP-2 signaling, antagonizing the canonical Smad1/5/8, with subsequent Osterix downregulation.

In our study we observed a differential activation of AKT levels by CaSO4 and BMP-2. A lower AKT phosphorylation was observed in CaSO4-treated conditions. Moreover, LY294004 and Wortmannin treatment abolished the migration induced by CaSO4, supporting the evidence that Ca2+-induced migration could be mediated by AKT. In agreement with our results, it has been reported that extracellular calcium produces early AKT activation with maximal effect between 5 and 60 minutes [5860]. Class I PI3Ks are divided into class IA (PI3Kα, PI3Kβ, and PI3Kδ) and class IB (PI3Kγ). It has been demonstrated that BMP-2 induces AKT phosphorylation through the specific activation of the PI3Kα isoform [32]. In contrast, PI3Kγ is almost exclusively activated by GPCRs [61, 62]. Several studies have reported that GPCR activation inhibits PI3K signaling [63]. Of note, class IA isoform p110α (activated by BMP-2) might be inhibited by the GTP-bound Gβγ subunit which is downstream of GPCRs [6365]. Altogether, our results suggest a novel mechanism by which CaSO4 modulates MSC migration by attenuating BMP-2 activation of AKT.

Migration of undifferentiated MSCs dramatically decreases during further steps of osteogenic differentiation [66] and also leads to lower response to chemotactic factors [9]. Therefore, an initial modulation of osteoblast differentiation could promote progenitor cell recruitment. There are at least two aspects in our study relevant to understanding some of the biological mechanisms whereby Ca2+ promotes bone regeneration. First, CaSO4 promotes MSC migration in a concentration-dependent fashion and modulates BMP-2-induced migration. Second, CaSO4 exerts a biphasic effect on MSC differentiation. An initial transient attenuation of BMP-2 promoted differentiation which is followed by a progressive increment in the expression of osteoblastic genes, such as Osterix, Alpl, or Osteocalcin. Therefore, Ca2+ may act on undifferentiated MSCs promoting migration by modulating PI3K/AKT activation and simultaneously delaying a mature osteoblast phenotype which is correlated with decreased motility.