Enhanced bone tissue regeneration of a biomimetic cellular scaffold with co‐cultured MSCs‐derived osteogenic and angiogenic cells

Abstract Objectives The bone tissue engineering primarily focuses on three‐dimensional co‐culture systems, which physical and biological properties resemble the cell matrix of actual tissues. The complex dialogue between bone‐forming and endothelial cells (ECs) in a tissue‐engineered construct will directly regulate angiogenesis and bone regeneration. The purpose of this study was to investigate whether co‐culture between osteogenic and angiogenic cells derived by bone mesenchymal stem cells (MSCs) could affect cell activities and new bone formation. Materials and methods Mesenchymal stem cells were dually induced to differentiate into osteogenic cells (OMSCs) and ECs; both cell types were co‐cultured at different ratios to investigate their effects and underlying mechanisms through ELISA, RT‐qPCR and MTT assays. The selected cell mixture was transplanted onto a nano‐hydroxyapatite/polyurethane (n‐HA/PU) scaffold to form a cell‐scaffold construct that was implanted in the rat femoral condyles. Histology and micro‐CT were examined for further verification. Results ELISA and gene expression studies revealed that co‐cultured OMSCs/ECs (0.5/1.5) significantly elevated the transcription levels of osteogenic genes such as ALP, Col‐I and OCN, as well as transcription factors Msx2, Runx2 and Osterix; it also upregulated angiogenic factors of vascular endothelial growth factor (VEGF) and CD31 when compared with cells cultured alone or in other ratios. The optimized OMSCs/ECs group had more abundant calcium phosphate crystal deposition, further facilitated their bone formation in vivo. Conclusions The OMSCs/ECs‐scaffold constructs at an optimal cell ratio (0.5/1.5) achieved enhanced osteogenic and angiogenic factor expression and biomineralization, which resulted in more effective bone formation.

functional bone matrix-producing cells has been considered as an alternative to bone grafting. 2 Following the recognition of the limits associated with mimicking complex biological environments when introducing single-cell phenotypes, the co-culture of two or more types of cells in vitro and in vivo is now being granted more attention due to their ability to more closely model natural bone regeneration. This provides additional insight into that cell-cell interactions may improve the efficiency of current bone tissue engineering. 3,4 Cell-cell communication between diverse cell types is vital to the tissue healing process. 5,6 Cells co-cultured with other cell types can produce bioactive factors that allow different crosstalk schemes between cells, promoting endocrine, paracrine, autocrine, and electric signalling routes and direct effects that are dependent on cell contact. Several studies have shown synergistic effects in response to the use of co-culture systems, which have the ability to induce stem cell differentiation. 7,8 The previous studies suggested that the synergistic interplay between osteogenesis and angiogenesis plays a pivotal role in the bone regeneration process, 9,10 while rapid revascularization is crucial for transplanted cell survival and new bone formation. Because bone is a calcified and peripherally vascularized tissue consisting of various cell types, including osteogenic cells and endothelial cells, co-culture of cells with osteogenic and angiogenic potential draw much attention in bone tissue engineering. 5 Herzog et al found that the co-culture of primary osteoblasts and the outgrowth of endothelial cells (ECs) positively influenced vessel formation and bone repair, which was associated with rising levels of growth factors and proteins of different origins. 11 Osteoblasts produce angiogenic factors, such as vascular endothelial growth factor (VEGF) and matrix components, which are important in vessel component differentiation; in turn, these factors stimulate ECs to produce osteogenic factors, such as BMP-2. 12,13 The association of these two essential cell types in a biomaterial can provide a live bone graft that can be used to repair bone defects, 14 which may be beneficial for rebuilding the vascular network within tissueengineering constructs and subsequently promoting bone tissue regeneration.
In addition to the selection of co-cultured cell types, the ratio of the different cell types in the co-culture system can also influence cell characteristics, survival and behaviours. Therefore, the proper ratio of co-cultured cells may be important to guarantee an excellent bone tissue-engineering construct. However, in view of the available literature, few systematic studies assessing optimal cell ratios between ECs and tissue-specific cells have been reported. In most studies, researchers selected a 1:1 cell ratio 15,16 ; however, this may be a matter of keeping things simple, rather than utilizing the full potential of co-cultures. 17,18 An early study by Kim et al reported that the optimal ratio (0.5/1) of two different cell types, adipose-derived stromal cells (ASCs) and bone marrow stromal cells, promoted osteogenic differentiation and osteogenesis in a co-culture model. 19 The effect of the co-cultured cells at different ratios was also investigated by Ma et al using human umbilical ECs and human marrow stromal cells 20 ; however, the optimal ratio (1:1) of co-cultured cells remained poorly understood and required more systematic investigations. To optimize the co-cultured cell ratio, the various tests should not only investigate the proliferation or viability of cells, but also assess the impact of this ratio on gene expression, the related signal-transduction pathway and the desired phenotypic expression within the co-culture system.
Mesenchymal stem cells, which are primarily present in bone marrow, are multipotent stem cells that can differentiate into target cells such as osteoblasts, 21 chondrocytes 22 and endothelial cells 23 under specific conditions. MSCs have been extensively investigated and were shown to be the most suitable cell source for bone tissue engineering due to their excellent osteogenic potential; furthermore, researchers also revealed that MSCs promote angiogenesis through proteolytic mechanisms. 24 Considering the lower osteogenic potential of MSCs compared with osteogenic-induced differentiated MSCs (OMSCs), 25 we hypothesize that the co-culture of MSCs-derived osteogenic and angiogenic cells at an optimal ratio may be a promising strategy for vascularized bone tissue regeneration. The purpose of this study was to investigate whether the co-culture of MSCs-derived ECs to OMSCs, as well as cell ratio, affected cell activities and new bone formation. Thus, we first induced the osteogenic and angiogenic differentiation of MSCs into OMSCs and ECs, respectively. Subsequently, the optimal ratio of OMSCs/ECs in the co-culture system was determined by exploring the level of cell crosstalk based on the functional markers of osteogenic and angiogenic expression.
After screening for the optimal ratio of OMSCs/ECs using in vitro experiments, the selected co-cultures were transplanted onto a biocompatible and bioactive n-HA/PU composite scaffold to form a cellular bone graft. 26 A condylar femur defect model in rat was used to demonstrate the effect of new bone formation.

| Cell isolation and cultivation
Sprague Dawley (SD) rats of about 100 g in weight were used as donors of femurs and tibiae for bone marrow harvesting and primary MSCs isolation, according to an established procedure ( Figure 1A).
Briefly, bone marrow was flushed out using α-minimum essential medium (α-MEM) supplemented with 1% antibiotic/antimycotic and 20% foetal bovine serum. Cells were plated on a culture flask, changing α-MEM every 3 days. After 1 week of incubation, MSCs were regularly subcultured and the fourth passage cells were used for experiments.

| In vitro mixed co-culture of OMSCs/ECs
In vitro mixed co-cultures of OMSCs and ECs at the fourth passage were carried out in OM shown in Table 1. As controls, the monoculture of OMSCs and ECs was performed in the same cell numbers in OM and α-MEM separately.
F I G U R E 1 Schematic representation of the in vitro and in vivo experimental procedure. A, MSCs isolation and dual induced differentiation. B, Co-culture model systems used for the analysis of cell-to-cell interactions and their mixtures at optimal ratios co-cultured with the n-HA/PU scaffold. C, In vivo experimental procedure

| ELISA assay
From the co-culture settings, the medium was taken from the culture flask after 3 days of in vitro culture and assayed to measure the level of ALP, osteocalcin (OCN) and VEGF. ALP and OCN assays (Thermo Fisher Scientific, Waltham) were performed to detect early osteogenic cell differentiation. A VEGF ELISA kit (Thermo Fisher Scientific) was used to quantify VEGF, according to the manufacturer's instructions.

| Reverse transcription and quantitative polymerase chain reaction
The osteogenic and angiogenic differentiation of co-cultured cells was  Table S1.

| Morphology and mineralization of MSCs, OMSCs, ECs and co-cultured cells on the n-HA/ PU scaffolds
The n-HA/PU scaffold was prepared according to our previous report 26 and cut into square samples (10 × 10 × 2 mm 3 ). After an ultrasonic rinse in distilled water and sterilization with an autoclave, the samples were seeded with cells statically. The selected ratio of the co-cultures with the scaffold is presented in Table 2; constructs of OMSCs/ECs were cultured in OM, while the same ratio of MSCs/ ECs was in α-MEM medium in 24-well plates as control in a humidified incubator (37°C, 5% CO 2 ).

| Construction of n-HA/PU-seeded cells
Scaffolds were cut into cylinders (diameter: 3 mm; thickness: 3 mm) using a trephine bur. Following ultrasonic rinse in distilled water and sterilization with an autoclave, the scaffolds were incubated overnight in fresh α-MEM and then co-cultured with cells. For the co-cultured series design, as stated in Section 2.3, the cell suspensions were statically seeded on the scaffold (where the scaffold was without cells as control) and cultured for 14 days in a CO 2 incubator at 37°C to obtain the cellular constructs ( Figure 1B).

| Histological evaluation
The samples were decalcified and then dehydrated through gradient ethanol, cleaned in xylene and embedded with paraffin wax. Finally, the samples were cut into sections (5 μm in thickness) along the sagittal plane, stained with haematoxylin and eosin (HE) staining and observed under optical microscopy.

| Statistical analysis
Quantitative data were presented as the mean ± standard deviation (SD). Statistical analysis was carried out using one-way analysis of variance (ANOVA) with a Tukey test. Differences were considered to be statistically significant when P < 0.05.

| Effects of ECs in mixed co-culture with OMSCs
According to the existing and well-known methods used to induce stem cell differentiation, we separately induced MSCs differentiation into OMSCs 25  and VEGF markers ( Figure S2). Thus, we successfully obtained two desired cell phenotypes derived from same MSCs source by different induction conditions, which could be used for the subsequent experiments.
To determine the interaction between OMSCs and ECs, the optimal cell ratio between these two cell types and their effect on osteogenesis and angiogenesis, the co-cultures were utilized in fixed numbers of total cells (2.0 × 10 5 ) with variable ratios of ECs to OMSCs. It is apparent that the group with a ratio of 0.5/1.5 had the highest OCN and VEGF content, as determined by the ELISA data ( Figure 2A). Moreover, VEGF amount in this group is even higher than the ECs monoculture group. The effects of osteogenic-induced OMSCs on the gene expression of co-cultured OMSCs/ECs were also mirrored by RT-qPCR. It was found that the level of osteogenic genes (OCN, Msx2, Runx2, Osterix) expression was relatively higher at the ratio of 0.5/1.5 ( Figure 2B). When the vascular genes (CD31 and VEGF) were detected, the level varied with different ratios; however, the highest level was achieved with an OMSCs/ECs ratio of 0.5/1.5. MTT assays demonstrated that the total cells proliferated with increased ECs numbers and culture time. Overall, the co-cultured cells of OMSCs/ECs at ratio of 0.5/1.5 were advantageous for osteogenic and vascular expression when compared with the other ratio. As controls, four groups of OMSCs, MSCs, ECs and MSCs/ECs that were, respectively, seeded on the scaffolds were also investigated ( Figure 3(a-d)). Although OMSCs seeded scaffold was cultured in OM medium, only few mineral particles were found on or in the scaffold. The pictures also confirmed that the cell densities of osteogenic-induced group (OMSCs/ECs, OMSCs) were lower than that in the non-induced groups (MSCs, ECs, MSCs/ECs). It is known that differentiated cells (OMSCs) proliferate less than undifferentiated mates (MSCs), which may be due to an inhibitory effect that occurred by inducing factors. Furthermore, abundant ECs formed an arrangement on the scaffold.

| In vivo effects of cellular constructs on bone regeneration
To evaluate the different cell mixture types on angiogenesis and osteogenesis in vivo, the cellular constructs were implanted in the bilateral femoral condyles of rats. There was no evident inflammation found, and the new bone tissue formation that occurred during defect healing was grossly evaluated using micro-CT and microscopically assessed using HE staining at different time points. implantation, which became thicker in the newly generated bone region at 8 weeks. The group of OMSCs/ECs presented a mineralized matrix at 4 weeks, followed by the development of a mature trabecular bone meshwork. These results suggested that the OMSCs/ ECs mixtures co-cultured had a greater effect on bone formation and integrity when compared with the MSCs/ECs mixtures, which promoted a rapid bone-healing process. In addition, the co-cultured cells groups showed more active vascularization than the groups featuring either the scaffold alone or the scaffold with monoculture.
Although capillary vessels (red arrow) in the groups of MSCs/ECs occurred at some sites in the trabecular bone, the numbers of capillaries were significantly less than those of the OMSCs/ECs group at 8 weeks ( Figure 5C).

| D ISCUSS I ON
Clinical situations that require cell transplantation for bone regen- can trigger the cells networking and their interaction with the surrounding biomaterials. 37 In this study, we selected biomimetic n-HA/PU composite scaffold as a substrate, which has been proved to provide a support structure, enhance cell engraftment and survival, and further produce strong vitality in bone regeneration and reconstruction. 26 We found that mixtures of OMSCs/ECs were more effective in inducing bone repair, and it facilitated better restoration of osseous structures than mixtures of MSCs/ECs or the application of one cell type alone, while the control (without con- In this study, we confirmed that the construct seeding with OMSCs/ECs mixtures at a certain ratio (0.5/1.5) promoted biomineralization and bone regeneration both in vitro and in vivo due to their synergistic effects. Further systematic studies need to illuminate the mechanism that how the vascularization of tissue-engineering construct stimulates bone regeneration in vivo. Successful results from these studies will be beneficial in the progression of bone tissue engineering.

| CON CLUS ION
This study demonstrated that osteogenesis and angiogenesis could be enhanced by augmenting the paracrine effects between OMSCs and ECs interactions at an optimal ratio (0.5/1.5) in co-culture treatment. Transplantation of an optimal ratio of OMSCs/ECs co-culture in a scaffold, which mimics natural tissue complexities, provides a live tissue-engineering construct that-when co-implanted-can rapidly generate new bone tissue. The mechanism underlying this effect seems to involve the upregulation of angiogenic factors (VEGF and CD31); the key transcription factors involved in osteogenic differentiation (Msx2, Runx2 and Osterix); the subsequently increased osteogenic markers of ALP, OCN and Col-I; and the stimulated mineralization. The findings in this study highlight that this approach holds great promise in regenerative medicine.

CO N FLI C T O F I NTE R E S T
The authors declare no competing financial interest.

DATA ACCE SS I B I LIT Y
The data that support the findings of this study are available from the corresponding author upon reasonable request.