Bone marrow mesenchymal stem cells combine with Treated dentin matrix to build biological root

Treated dentin matrix (TDM) as a kind of scaffolding material has been proved odontogenic induction ability on dental-derived stem cells. Given the limited resources of dental stem cells, it is necessary to seek new seed cell which easily obtained. Jaw bone marrow mesenchymal stem cell (JBMMSC) as non-dental-derived stem cell relates to the development of teeth and jaws which suggest us JBMMSCs could act as a new seed cell for tooth tissue engineering. To assess the odontogenic and osteogenic potential of JBMMSCs, cells were induced by TDM extraction in vitro and combined with TDM in vivo. Results were analyzed by PCR, Western Blotting and histology. PCR and Western Blotting showed odontogenic and osteogenic makers were significantly enhanced in varying degrees after induced by TDM extraction in vitro. In vivo, JBMMSCs expressed both odontogenic and osteogenic-related protein, and the latter showed stronger positive expression. Furthermore, histological examination of the harvested grafts was observed the formation of bone-like tissue. Therefore, osteogenic differentiation ability of JBMMSCs were enhanced significantly after being inducted by TDM which illustrates that non-odontogenic derived stem cells are still promising seed cells in tooth root tissue engineering.

Fabrication of human TDM. Human TDM was fabricated using the method described in our previous study 3 . It was produced to the dimensions of 1.0-1.5 cm in length and 1.0 mm in thickness. SEM examination demonstrated that the dentinal tubules were sufficiently exposed and fiber bundles of intertubular and peritubular dentin were loosened after being treated with EDTA ( Fig. 1B,C).

Real-time PCR analysis.
Real-time PCR was performed to assess the effect of the liquid extraction of TDM on the JBMMSCs. Odontogenic-related and osteogenic-related genes like DMP-1, ALP, Periostin, OPN not only showed higher expression than control group at all time points but also expressed stronger and stronger as time went on. Expression of DSPP, Decorin and OCN kept stable with control group at 3 d and began to rise at later time points. However, there was no significant difference between experiment group and control group about the expression of fibronectin and COL-III which were not the mineralization related marker (Fig. 3A). showing no inflammatory response after implant surgery indicating good biocompatibility of human TDM. H&E staining showed planted cells had ideal adherence and proliferation on TDM (Fig. 4A). Besides we observed that there was a lighter stained layer inside TDM. We thought that structure was demineralized dentin due to demineralization in fabricating. JBMMSCs performed connective tissue-like construction with new vessel generation (Fig. 4B). However expected new tissue was not observed in control group (single TDM) (Fig. 4D). Then we used immunohistochemistry to evaluate the expression of DSPP, DMP-1, OCN, OPN, Decorin, TGF-β , COL-I, COL-III, Periostin, Fibronectin and ALP to explored the differentiation status of JBMMSCs. The result showed that JBMMSCs were strongly positive for protein osteogenic marker OPN (Fig. 5D) and ALP (Fig. 5F) not only in the cells but also in dentin and demineralized dentin layer. Positive expression of OCN (Fig. 5E), Decorin (Fig. 5C) and Periostin (Fig. 5H) was also observed in cells and demineralized layer. Odontogenic markers DSPP (Fig. 5A), DMP-1 (Fig. 5B) and normal dental pulp collagen fibrils components COL-I (Fig. 5J), COL-III (Fig. 5K) showed weakly positive expression. However, most cells were stained negatively for TGF-β (Fig. 5G) and Fibronectin (Fig. 5I). It is interesting that we found the bone like tissue which formed on the pulp cavity surface of TDM in some samples (Fig. 6A). Many osseous lacunas spreaded over neoformative bone matrix (Fig. 6B). We could also observe the osteoblast-like cells at the edge of bone matrix (Fig. 6C). In Masson staining, neo-formative bone-like tissue was stained red instead of blue suggested that the tissue contained fewer collagen components (Fig. 6D).

Discussion
Many materials have been adopted as scaffolds for dentin regeneration [24][25][26] . However, these materials can hardly contribute to prefabricated-shaped and complete dentin regeneration. The reason may be that while these materials support cell growth and mineralization, they are not capable of inducing differentiation towards an odontogenic specialization. As the main component of tooth, dentin is less mineralized and more elastic than enamel. It is reported that the organic matrix of dentin contain approximately 30 volume percent of collagen, Second-passage cells were typical spindle or triangle shape morphology (A) and positive for Vimentin but negative for CK14 in immunofluorescence staining (D). TDM (B) as scaffold was fabricated from human healthy teeth. SEM (C) illustrates that dentinal tubules exposed sufficiently and loosened fiber bundles of intertubular and peritubular dentin. (E) JBMMSCs were planted on the surface of TDM. Scale bars = 100 μ m, 10 μ m for (A,D) respectively. noncollagenous proteins (NCPs), and growth factors, and many of these proteins and factors have been shown to be important in dentin development, mineralization, and regeneration 17,27,28 . Besides, it has been reported that a three-dimensional (3D) microenvironment with specific properties that could promote in vitro proliferation of chondrocytes, hepatocytes, endothelial cells, osteoblasts, neuronal cells, and stem cells 29 . As a natural decellularization matrix scaffold, TDM provided a 3D odontogenic inductive microenvironment. The odontogenic potential of dentin matrix on dental stem cells has been confirmed, whether TDM could also induce non-dental stem cells differentiate into an odontogenic specialization is not realized.
The present results showed that BMMSCs stemmed from human jaw were generally typical spindle or triangle shape. Importantly, cells stained positive for vimentin but negative for CK14. As positive staining for vimentin and CK14 is indicative of mesenchymal and epithelial cells, respectively, this result indicated that these cells were mesenchymal but not contaminated by epithelial cells. FCM showed that cell surface molecules were positive for mesenchymal stem cell surface marker like CD29, CD73, CD90, CD105, CD146 but negative for hemopoietic stem cell marker CD31 and CD34 which belongs to platelet endothelial cell adhesion molecules. Furthermore, these cells had potential for differentiating to osteoblast, lipoblast, nerve-like cell and vascular endothelial-like cells which illustrated they were certainly the stem cells. These results proved that the cells from jaw marrow were indeed BMMSCs. TDM as scaffold was fabricated from human healthy teeth with exposed dentinal tubules which. JBMMSCs could adhere to surface of framework due to good biocompatibility of TDM. In our previous study, we found TDM expressed COL-I, TGF-β 1, decorin, biglycan, DMP-1, and DSPP 3 . In this study, we detected not only the odontogenic markers above but also some other markers such as OCN, OPN, Fibronectin.
As the most abundant protein, COL-I forms a visible network around the dentinal tubules in the peritubular region and in the intertubular predentin region, with multiple spherical mineral foci, which are distinctive features of dentin mineralization 30 . COL-1 is secreted by odontoblasts into predentin and then recruited into the mineralization front as bundles of collagen fibrils, thus forming mineralized. However, biglycan and decorin are considered to interact with COL-I fibrils in dentin mineralization 31,32 . By contrast, the content of COL-III is far less than that of COL-I. COL-III could be detected in dental pulp and periodontium and it plays a role in the development of trabecular bone through its effects on osteoblast differentiation 33 . TGF-β 1 appeared in tooth morphogenesis and odontoblast differentiation during embryogenesis and could also be detected immunohistochemically in mature human teeth 34,35 . During tooth development, TGF-β expressed with a high level, which was considered to relate to the differential activity on the cells at distinct stages of development 36,37 . TGF-β 1 could lead to the sequestration of these growth factors within the dentin matrix that could be released as a result of matrix changes associated with caries or trauma 38 . Other studies have also demonstrated that application of TGF-β 1 could stimulate reparative dentin formation 39,40 . These data implicated TGF-β 1 played a key role in the dentin tissue regeneration. As a primary component of noncollagenous proteins, DMP-1 and DSPP were also detected. DMP-1 is a multifunctional protein involved in the mineralization of bones and dentin 41 . DMP-1 can induce organization of apatitic crystalites into bundles with their c-axes co-aligned with the axis of the bundle, which is a hallmark of the collagenous mineralized tissues such as bone and dentin 42 . DSPP is also an important nucleator for dentin mineralization and proper dentin formation 28,43 . DSPP is expressed and secreted by odontoblasts, the cells that make tooth dentin and that also maintain cell processes extending into the mineralized tissue. Some studies suggest that a deficiency in DSPP is a causative factor in dentinogenesis imperfecta (DI) 43 . DSPP and DMP-1 expression was therefore up-regulated perhaps because TDM could lead JBMMSCs to osteogenic differentiation. Besides, biglycan and decorin were also expressed in a relatively high level of concentration, which was widely distributed in mammalian tissues, including mineralized tissues such as bone and teeth 32,44 . The ability of biglycan and decorin to bind to and inhibit the activity of TGF-β in previous studies makes these two proteoglycans likely candidates for the sequestration of TGF-β reservoirs in the dentin matrix 45,46 . In addition to modifying the extracellular environment, interactions of biglycan and decorin with cells, and with growth factors such as TGF-β 1 affect the proliferation of cells 47 . Furthermore, increased levels of biglycan or decorin may be the cause or consequence of hypomineralization and widened predentin in Dspp− /− mice 48 . Both biglycan and decorin single gene knockout mice exhibited altered numbers of collagen fibrils in the molar predentin at a younger age 49 . Cementum attachment protein (CAP) is considered as cementum specific protein because of significantly higher expression in cementum than bone tissue. In addition, CAP is usually regarded as marker protein of cementoblast precursors with ability to promot proliferation and differentiation of periodontal ligament cells 50 . Runx2 as an important transcription factor has profound effects on osteoblast differentiation in the bone microenvironment. Runx2 target genes include regulators of cell growth control, components of the bone extracellular matrix, angiogenesis, and signaling proteins for development of the osteoblast phenotype and bone turnover 51 . Alkaline phosphatase (ALP) as a mineralization-related marker could participate in the differentiation of osteoblasts. Expression of ALP would become stronger following the increase in the degree of osteogenic differentiation which was similar to RUNX2. OCN secreted by osteoblasts, odontoblasts, and cementoblasts is a major noncollagenous protein found in bone and dentin and could restrict the mineralization of bone and tooth tissue, including 52 , and is regarded as a terminal symbol in hard tissue regeneration 53 . OPN is a multifunctional phosphorylated glycoprotein highly expressed in the hard tissue like bone, tooth. And as a kind of effective inhibitor of hydroxyapatite formation 54 , it could regulate the development of dentin. Periostin expressed in ameloblast, subodontoblast, and odontoblast cells. It plays a novel direct role in controlling postnatal tooth formation, which is required for the integrity of both enamel and dentin 55 . Periostin is the cell adhesion molecule of osteoblasts and their precursor cells. Previous research has shown that periostin located between the cytoplasmic processes of periodontal fibroblasts and cementoblasts and the adjacent collagen fibrils, and bear the mechanical stress in the process of tooth eruption and occlusion 56 .
In summary, these proteins and factors not only play important roles in inducing JBMMSCs proliferation and differentiation into odontoblasts but also form a network to form dentin tissues and control mineralization during dentin regeneration and development.
From the results of real-time PCR and Western Blotting, we observed that mineralization-related markers DMP-1, ALP, RUNX2, OCN, CAP, Periostin and COL-I were high-expression after induction by TDM, while the COL-III and Fibronectin were not significantly up-regulated after induction by TDM. This indicated TDM has the ability for odontogenic and osteogenic induction which may depend on the release of active factors due to exposure of dentin tubules after treated. And it is worth noting that osteogenesis-related makers such as OPN, RUNX2 possessed stronger expression compare with odontogenesis-related markers such as DSPP, Decorin. This suggested that JBMMSCs as a kind of non-dental stem cells may be inclined to differentiate in the direction of osteogenesis under the influence of the active factors from TDM. In order to further verify the validity of the odontogenic functions of JBMMSCs and TDM in vivo, TDM scaffolds seeded with JBMMSCs were implanted into mice at subcutaneous pocket with no mineralization capacity. JBMMSCs has an ideal attachment and proliferation on TDM, which was demonstrated by SEM (Fig. 1E). These represented the superior biocompatibility and bioactivity of TDM. With the induction of growth factors secreted by TDM, JBMMSCs have a greater mobility. Cell cytotoxicity test indicated good biocompatibility of TDM in our previous study 9 .
At first, we detected expression of human-mitochondria (Does not react with other species) in harvest tissue with immunohistochemistry to prove that these cells came from implanted human cells. Result showed the cells within TDM stained positive for human-mitochondria ( Supplementary Fig. S1) which illustrated they were human cells while not the mouse cells from the host. From HE staining, we could observe the pulp-like tissue and new blood vessel generated (Fig. 4B), new fibrous tissue were also observed by Masson stain (Fig. 4C), but there was no obvious dentin or predentin formation. Immunohistochemistry showed JBMMSCs gained a greater degree in osteogenic differentiation than odontogenic differentiation (Fig. 5A-L). Moreover, JBMMSCs were capable of promoting mineralization of demineralized dentin layer (Fig. 3A-F,H). Therefore JBMMSC tended to osteogenic differentiating instead of differentiating to dentin or dental pulp tissue with the effect of TDM.
Our results indicated that JBMMSCs were capable of differentiating into multiple cell lineages when grown under defined conditions in vitro ( Fig. 2A,B), which was in accordance with previous studies 57,58 . A previous study showed that human dental pulp stem cells (DPSCs) and bone marrow stem cells had similar gene expression levels of more than 4,000 known human genes 59 . JBMMSCs share numerous characteristics with dental stem cells and are both able to form bone-like or tooth-like structures. In this study, we found that the odontogenic potential of JBMMSCs was enhanced by TDM. The reasons may be that a lot of odontogenic proteins and molecular signals regulating the fate of tooth development, such as Notch signals, TGFs (transforming growth factors), BMPs (bone morphogenetic proteins) and FGFs, are expressed permanently in TDM. However, we found that JBMMSC showed higher level osteogenic potential according to results of PCR and immunochemistry. JBMMSCs display a lower odontogenic potential than dental stem cells, indicating that MSC from different embryonic origins are not equivalent 11 . Furthermore, the formation of bone-like tissue also indicated the osteogenic potential of JBMMSC and TDM could induce the osteogenic differentiation of JBMMSC. On the other side, as we know, JBMMSCs could be induced to osteoblasts due to favorable microenvironments, and also osteoclasts due to unfavorable microenvironments. So it is necessary to avoid internal absorption process of TDM framework in vivo. There exists a kind of coupling relationship between osteoblast and osteoclast so that bone mass is maintained locally by the balance between osteoclastic bone resorption and osteoblastic bone formation 60 . So osteoclasts are inevitable when TDM with JBMMSCs is implanted in alveolar bone. TDM could induce mineralization by releasing bioactive factors 3 . The existence of osteoclasts would be conducive to bone remodeling instead of bone resorption if there is no infection or immunologic rejection after implantation. Furthermore, our previous studies showed that there was no bone resorption but tissue regeneration observed after implanted TDM combined with dental follicle cells into alveolar bone 4,18,61 . We, therefore, believe that without infection or immunologic rejection, severe bone resorption will not occur when implantation in alveolar bone.
As for no dentin, predentin was observed, in our opinion, the formation of complete dentin, predentin, pulp nerve structures is a complex process which requires regulation of many genes and proteins. While JBMMSCs as a non-dental stem cells combine with TDM generate bone-like tissue subcutaneously in nude mice makes synosteosis possible between TDM and alveolar bone when being transplanted into the jaws in the future.

Conclusion
Here we have described the odontogenic potential of JBMMSCs induced by TDM both in vitro and in vivo. Our results confirmed the good biocompatibility and odontogenic microenvironment of TDM on non-dental derived stem cells. We found JBMMSCs could express odontogenic and osteogenic markers which demonstrated that JBMMSCs have odontogenic potential, the ability to differentiate into odontoblast-like cells and contribute to odontogenesis. Although JBMMSC possesses stronger ability in promoting mineralization and osteogenic differentiation, even could form bone-like tissue with TDM inductive effect, it is still prospective seed cell in building biological root for tooth regeneration.

Materials and Methods
All experiments were conducted in accordance with the ethical protocol approved by the local ethical committee of the Sichuan University. All the methods were carried out in accordance with the approved guidelines. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.
JBMMSCs isolation and culture. Human JBMMSCs were isolated from mandibular bone marrow and the cancellous bone fragment of the consenting donor who received orthognathic surgery due to malocclusion. The donor was examined to exclude hematopoietic neoplasms and the samples were found to be histologically normal. We applied the whole bone marrow culture method to culture JBMMSCs and purified the cells through differential adherent time. Depurated cells were cultured in the medium consisting of Dulbecco's modified Eagle's medium (DMEM, Hyclone, USA), 10% fetal bovine serum (Hyclone, USA), 100 U/mL penicillin, and 100 μ g/mL streptomycin (all Sigma, USA). Cells were seeded at a density of 2 × 10 5 cells/cm 2 . After 72 h, non-adherent cells and debris were washed out by the first media exchange. Then, cells were incubated in DMEM supplemented with 10% fetal bovine serum in a humidified atmosphere at 37 °C and 5% CO 2 . Cell culture medium was changed every 2 d and cells from passages 2-4 were used for the experiments. After about 7 d, the JBMMSCs were observed under a phase-contrast inverted microscope (Nikon, Japan). Immunofluorescent microscopy of JBMMSCs. One day prior to staining, JBMMSCs were released and seeded onto 0.8 cm × 0.8 cm coverslips in a 24-well plate for further culture. JBMMSCs were fixed with 4% polyoxymethylene for 30 min. Subsequent steps were performed according to the manufacturer's recommendations. Antibodies used in immunofluorescent staining included: vimentin and CK-14. All antibodies were from Abcam, USA. All samples were examined under a fluorescence microscope (Leica Optical, Germany).

Surface molecule characterization and gene expression pattern of JBMMSCs. Approximately
Scientific RepoRts | 7:44635 | DOI: 10.1038/srep44635 FITC-conjugated, isotype-matching immunoglobulins were used to determine non-specific staining. The secondary reagents included goat anti-mouse and goat anti-rat IgG-FITC (Santa Cruz). Cells were analyzed on an Accuri C6 Flow cytometer (Becton Dickinson, CA, USA), and the data were analyzed using CellQuest software.
Multipotential differentiation of JBMMSCs. Osteogenic differentiation. A total of 1 × 10 5 JBMMSCs were seeded into each well of a six-well plate. At 80% confluence, JBMMSCs were cultured in osteogenic medium containing 10% FBS, 5 mL glycerophosphate (Sigma, USA), 100 nM dexamethasone (Sigma, USA), and 50 mg/ml ascorbic acid (Sigma, USA) for 2 weeks. The control group was cultured in DMEM with 10% FBS. The medium was changed every two days. After 2 weeks, cells were washed twice in PBS after being fixed in 4% paraformaldehyde for 10 min and then incubated in 0.1% alizarin red solution (Sigma, USA) in Tris-HCl (pH 8.3) at 37 °C for 30 min. After being washed twice with PBS, cells were routinely observed and photographed under a light microscope (OLYMPUS, Japan).
Adipogenic differentiation. A total of 1 × 10 5 JBMMSCs were seeded into each well of a six-well plate. At 80% confluence, JBMMSCCs were cultured in complete adipogenic medium consisted of DMEM supplemented with 10% FBS, 2 mM glutamine (Sigma, USA), 100 U/ml penicillin-streptomycin (Sigma, USA), 100 μ M ascorbic acid Fabrication of treated dentin matrix (TDM). Human TDM was fabricated according to our previous method 3 . Briefly, premolar teeth that were removed for clinical reasons at the West China Hospital of Stomatology of Sichuan University were collected.
Periodontal tissues were completely scraped away with a curette. By grinding along the tooth profile, outer cementum and part of the dentin were removed. Dental pulp tissues and pre-dentin were also mechanically removed. The resulting human dentin matrix was soaked in deionized water for 5 h and mechanically cleaned for 20 min every hour using an ultrasonic cleaner. The deionized water was changed every hour. Human dentin matrices were then treated with 17% Ethylene Diamine Tetra-acetic Acid (EDTA, Sigma, USA) for 5 min, 10% EDTA for 5 min, 5% EDTA for 10 min. Human TDM were maintained in sterile phosphate buffered saline (PBS) with 100 units/ml penicillin (Hyclone, USA) and 100 μ g/ml streptomycin (Hyclone, USA) for 72 h, then washed in sterile deionized water for 10 min in an ultrasonic cleaner, and finally stored in DMEM (Hyclone, USA) at 4 °C. Morphological observations of human TDM were observed by scanning electron microscope (SEM) (Inspect F, FEI, Netherlands).

Effect of TDM on biological characteristics of JBMMSCs in vitro.
To evaluate the effect of TDM on JBMMSCs, we first collected the liquid extract of the dentin matrix according to the protocol of the International Standardization Organization (ISO 10993). Briefly, scaffold samples of dentin matrix were pulverized using a cold mortar and pestle after freezing in liquid nitrogen. DMEM was added to the pulverized scaffold with a ratio of 20 g scaffold powder per 100 mL DMEM. Slurries were incubated for 2 d at 37 °C, and the samples were filtered using a 0.22 μ m filter.

Real-time PCR analysis.
Tissue regeneration in vivo using TDM and JBMMSCs. Twelve immunodeficient mice were divided into three groups including one test group (TDM combined with JBMMSCs) and one control group (single TDM). JBMMSCs were seeded onto TMD at a density of 1 × 10 4 cells/scaffold and incubated at 37 °C for three days. Implantation of scaffolds was performed under deep anesthesia. Eight weeks later, all samples were obtained from the immunodeficient mice under deep anesthesia. Implants were fixed with 4% paraformaldehyde overnight at 4 °C, demineralized with 10% EDTA (pH 8.0), and embedded in paraffin. Paraffin sections were stained with hematoxylin and eosin stain (H&E), Mason stain and immunohistochemical stains.
Immunohistochemical All animal experiments were conducted in accordance with the committee guidelines of Sichuan University for animal experiments, which also, meets the NIH guidelines for the care and use of laboratory animals. Immunodeficient mice were obtained from the Laboratory Animal Research Centre of Sichuan University and were maintained on a daily ration of Purina rodent chow in housing quarters with cycled light (12 h on/off), regulated temperature, and sterile water. Statistical analysis. All data were expressed as the mean ± SD. Statistical significance was analyzed using SPSS 16.0 software (SPSS, USA). A value of p < 0.05 was considered statistically significant.