Priming Dental Pulp Stem Cells from Human Exfoliated Deciduous Teeth with Fibroblast Growth Factor-2 Enhances Mineralization Within Tissue-Engineered Constructs Implanted in Craniofacial Bone Defects

Abstract

The craniofacial area is prone to trauma or pathologies often resulting in large bone damages. One potential treatment option is the grafting of a tissue-engineered construct seeded with adult mesenchymal stem cells (MSCs). The dental pulp appears as a relevant source of MSCs, as dental pulp stem cells display strong osteogenic properties and are efficient at bone formation and repair. Fibroblast growth factor-2 (FGF-2) and/or hypoxia primings were shown to boost the angiogenesis potential of dental pulp stem cells from human exfoliated deciduous teeth (SHED). Based on these findings, we hypothesized here that these primings would also improve bone formation in the context of craniofacial bone repair. We found that both hypoxic and FGF-2 primings enhanced SHED proliferation and osteogenic differentiation into plastically compressed collagen hydrogels, with a much stronger effect observed with the FGF-2 priming. After implantation in immunodeficient mice, the tissue-engineered constructs seeded with FGF-2 primed SHED mediated faster intramembranous bone formation into critical size calvarial defects than the other groups (no priming and hypoxia priming). The results of this study highlight the interest of FGF-2 priming in tissue engineering for craniofacial bone repair. Stem Cells Translational Medicine 2019;8:844&857.

Keywords: Bone engineering; Calvaria; Hypoxia; Intramembranous ossification; Mesenchymal stem cells.

Figure 1

Effect of hypoxia and FGF‐2 primings on mineralization within tissue‐engineered scaffolds cultured under osteogenic conditions. (A): Micro‐computed tomography three‐dimensional reconstructions of stem cells from human exfoliated deciduous teeth‐seeded scaffolds for the 3 conditions (normoxia, hypoxia priming, and FGF‐2 priming) at day 21. Mineralized areas, which appear in red, were found throughout the collagen scaffolds in gray. Scale bars: 1 mm. Normalized quantitative analysis of the mineralization within each matrix revealing higher mineral formation in hypoxia‐ and FGF‐2‐primed scaffolds than in the normoxia ones. (B): Representative images of von Kossa stained scaffold sections showing mineralized nodules (arrows) in all the conditions at day 14. Scale bar: 50 μm. Inset detail for each condition showing the area of interest at higher magnification. (C): Alizarin red (AR) staining observed under fluorescent microscopy allows detecting mineral formation (white arrows) at days 7 and 14. Scale bar: day 7, 100 μm; day 14, 50 μm. (D): At day 7, quantification of the AR staining shows significantly higher mineral formation in the hypoxia‐primed and control groups when compared with the FGF‐2‐primed group. At day 14, this group shows significantly higher formation when compared with the other conditions. *, p < .05; **, p < .01; ****, p < .0001.

Figure 2

Effect of hypoxia and FGF‐2 primings on stem cells from human exfoliated deciduous teeth in tissue‐engineered scaffolds. (A): Cell proliferation was calculated using the AlamarBlue reaction. All the conditions were metabolically active throughout the experiment. At day 4, the FGF‐2‐primed group showed a significant higher metabolic activity (×3) compared with the other conditions. (B): Hydroxyproline titration measured at days 7 and 14 showing that the concentration of collagen is significantly increased in hypoxia‐primed scaffolds compared with the other conditions. The FGF‐2‐primed group showed a significant increase in collagen at day 14 compared with day 7. (C): Representative sections of Sirius red staining to visualize the collagen fibers (arrows), which appear in red under polarized light (pol) at day 14. Scale bars: 25 μm. (D): Detection of alkaline phosphatase (ALP) activity (arrows) within the scaffolds at day 14 showing magenta staining associated to cells adjacent to the mineralized nodules, especially in FGF‐2 primed constructs. Scale bars: 50 μm. Inset detail for each condition showing the area of interest at higher magnification. Quantification of staining showed significantly higher ALP activity in FGF‐2 primed constructs (×2.5) compared with the control group. *, p < .05; **, p < .01; ***, p < .001; ****, p < .0001.

Figure 3

Osteogenic differentiation of stem cells from human exfoliated deciduous teeth in tissue‐engineered scaffolds. (A, B): Representative images of osteopontin (OPN; A) and dentin matrix protein 1 (DMP1; B) immunohistochemistry showing, for all the conditions, positive immunolabeling (arrows) adjacent to the cells at day 7 and day 14. Scale bars: 10 μm. Inset detail showing the area of interest at higher magnification. (C): Western blotting of DMP1 and OPN with a normalized quantitative analysis. DMP1 expression showed an increased abundance of the 57 kDa fragment in the FGF‐2 condition (×2.5 and ×1.8 at days 7 and 14, respectively) compared with the hypoxia and control groups. At day 14, OPN showed an increased abundance of 40 kDa (arrowhead) and 60 kDa (arrow) forms in the primed groups, especially FGF‐2 group (×1.9) compared with the control group. No significant difference was observed at day 7 for both OPN forms. Each band was normalized using GAPDH band as housekeeping protein (n = 3 per group). *, p < .05; **, p < .01; ****, p < .0001.

Figure 4

Longitudinal follow up of bone formation in calvarial defects by micro‐computed tomography. (A): Representative images of mouse skull three‐dimensional rendering at baseline, day 14, and 2 and 3 months. (B): Newly formed bone volumetric fraction expressed as a percentage of volume (BV/TV) on the total area of the defect. At 2 months, bone formation was slightly increased in the defects filled with hypoxia or FGF‐2 primed constructs when compared with the control constructs, although there was no significant difference. At 3 months, the defects appeared almost fully repaired in the FGF‐2 primed group with a statistically higher bone formation when compared with the other groups. ns, not significant; ****, p < .0001.

Figure 5

Mineral formation in calvarial bone defects. (A): von Kossa staining revealing bone formation (arrows) in all the conditions within the defects (delineated by a dotted line), at day 14, and 2 and 3 months. Inset detail at day 14: normoxia (Aa), hypoxic priming (Ab), and FGF‐2 priming (Ac). Scale bars: 50 μm. (B): Quantitative analysis of von Kossa staining showing that mineral formation was significantly increased in FGF‐2 primed samples at all‐time points when compared with control (stem cells from human exfoliated deciduous teeth with no priming) or hypoxic primed samples. At 2 and 3 months, bone formation was also increased in the defects filled with hypoxia primed constructs when compared with control constructs. ****, p < .0001.

Figure 6

Characterization of the bone healing process at day 14. (A): Representative images of resin embedded calcified sections stained with toluidine blue revealing, for all the conditions, new bone formation (in purple) at both the edge and the center of the defects (Aa–Ac). Scale bar: 200 μm. Inset detail for each condition (Ad–Af) showing, at higher magnification, osteoblast‐like cells (arrows). These cells were well‐aligned and separated from the mineralized tissue by an osteoid‐like tissue stained in pink (asterisk). Scale bar: 25 μm. (B): Representative images of alkaline phosphatase activity showing, for each condition, mineral formation within the healing defects (arrows). Scale bars: 200 μm. (C): Representative images of osteopontin immunohistochemistry showing positive labeling of the mineralized nodules within the healing defects in the three conditions. Scale bars: 25 μm. (D): Representative images of tartrate‐resistant acid phosphatase staining revealing osteoclast remodeling activity (arrows) within newly forming bone. Scale bar: 200 μm.

Figure 7

Ultrastructural analysis of the grafted calvaria at 3 months postimplantation. (A, C, E, G): Representative SEM image for each condition revealing the different areas of the calvaria: grafted and original part (scale bar: 5 μm). Inset detail showing collagen fibers organization within the defect and the calvaria at higher magnification (scale bar: 1 μm). (B, D, F, H): Representative transmission electron microscopy image for each condition revealing the ultrastructural difference for each area and conditions (scale bar: 1 μm). (A, B): Representative image of the native mouse calvarial bone showing well‐organized collagen fibers with similar same orientation. Osteocytes were observed in their lacuna associated with their delicate network of canaliculi (asterisks). (C, D): Heterogeneous matrix and poorly organized collagen fibers (white arrows) were observed in the control (normoxia) constructs at both magnifications. (E, F): In hypoxia primed constructs, the collagen fibers (white arrows) were more packed and better organized. Lacuno‐canalicular system (asterisks) was observed in the constructs. (H, I): In FGF‐2 primed constructs, the fibers (white arrows) were well‐aligned and showed a dense organization very similar to the adjacent bone. Of note, well‐distinguishable osteocytes (inset) with their lacuno‐canalicular system (asterisks) were observed associated with dense regenerated bone matrix in the construct.