Three different cell types were used in this study namely, HMSCs, DPSCs, and PDLSCs. HMSCs were obtained from the Tulane Cancer Center. The human DPSCs and PDLSCs were a gift from Dr. Songtao Shi (University of Southern California). All the three cell types were cultured in minimum essential medium alpha (αMEM), (GIBCO) containing 20% fetal bovine serum (FBS), (GIBCO), 1% L-glutamine, (GIBCO) and 1% antibiotic and antimycotic solution, (GIBCO). This will be referred to as growth media throughout this manuscript.

The dental pulp stem cell derived ECM scaffolds were generated as per published protocols (Ravindran et al., 2013). Briefly, DPSCs were cultured within collagen/chitosan scaffolds under the influence of differentiation media The prepared scaffolds were stored in Hank's balanced salt solution (HBSS) containing 5% antibiotic and antimycotic solution. The scaffolds were used within a week of being generated. For DPP blocking, the scaffolds were incubated overnight at 4°C in 500 μ l of DPP antibody (1/100) or rabbit IgG (1/100). They were washed 4× in HBSS prior to cell seeding.

DPSCs, PDLSCs, and HMSCs (2 × 10⁶ per scaffold) were cultured within the ECM scaffolds for up to 2 weeks in standard cell culture conditions in the presence of growth media. At 24 h, 1 week, and 2 weeks post seeding, live and dead cells in triplicate samples were analyzed using the live-dead cell assay kit (Molecular Probes) as per the manufacturer's recommended protocol. The calcein AM and the ethidium homodimer concentrations were 10 μ M each to indicate the live (green) and dead (red) cells respectively. The samples were imaged using a Zeiss Axioobserver D1 fluorescent microscope equipped with the required filter sets. Five individual fields of view were imaged for each and the percentage of dead cells to live cells was calculated.

DPSCs, PDLSCs, and HMSCs were seeded within the ECM scaffolds as described previously. Twenty-four hours, 1 week, and 2 weeks post seeding, triplicate samples were fixed in 4% formalin, permeablized with 0.5% triton-x100 for 1 h and stained with phalloidin probe conjugated with TRITC fluorescence probe to visualize the actin fibers. The samples were then z-stack imaged using a Zeiss LSM 710 confocal microscope. 3D renderings were produced using the Zeiss Zen imaging software.

DPSC, PDLSCs, and HMSCs were seeded within the ECM scaffolds as described above. They were fixed in 4% formalin, dehydrated in graded ethanol solutions, dried, coated with gold/palladium and imaged using a Hitachi SU8030 SEM.

HMSCs were cultured within the ECM scaffolds or control collagen/chitosan scaffolds in quadruplicate for 2 weeks using growth media. After 2 weeks, RNA was isolated using the Qiagen RNA isolation kit as per the manufacturer's instructions. First strand synthesis was performed using superscript III and the generated cDNA was subjected to gene specific quantitative PCR amplification. Fold change was calculated using the comparative delta Ct method. Statistical significance between the control and ECM samples were calculated using student's t-test. A greater that 95% confidence interval was deemed significant. Table 1 lists the gene specific primer sequences used in this study.

DPSCs, PDLSCs, and HMSCs were seeded at a density of 2 × 10⁶ cells/scaffold in triplicates within control collagen/chitosan scaffold, the ECM scaffold, the DPP blocked ECM scaffold (DPSCs only) and the IgG blocked ECM scaffolds as described previously. The DPP blocked scaffolds were prepared by incubating the odontogenic ECM scaffolds with a DPP antibody as published previously (Ravindran et al., 2013). The DPP antibody is an in-house made antibody. The antibody concentration was 2 mg/ml and it was used at a dilution of 1/100. Rabbit IgG blocked scaffolds served as controls. The cell seeded scaffolds were cultured in vitro for 48 h and were then implanted subcutaneously on the back of athymic nude mice (Charles River Laboratories). Four weeks post implantation, the animals were sacrificed and the scaffolds were retrieved, fixed in 4% neutral buffered formalin, embedded and sectioned into 5 μm thick sections for histological evaluation. All animal experiments were performed as per protocol approved by the UIC animal care committee (Assurance number A-3460-01).

Hematoxylin and Eosin (H&E) staining was performed as per published protocols (Ravindran et al., 2010, 2013). Alizarin red staining to visualize calcium deposition was performed as per standard procedures. Immunohistochemistry using peroxidase conjugated secondary antibodies and fluorescent probes was performed as per previously published protocols (Ravindran et al., 2010, 2012). The following antibodies were used: Mouse anti-DMP1 antibody, (1/2000, a gift from Dr. Chunlin Qin from Baylor College of Dentistry), rabbit anti-DSP antibody, (1/250, in house generated antibody), rabbit anti-DPP antibody (1/100, in house generated antibody), rabbit anti-VEGF antibody (1/100, Santa Cruz Biotechnology), mouse anti-von Willebrand factor (VWF) antibody (1/100, Santa Cruz Biotechnology). All fluorescently stained sections were imaged using a Zeiss LSM 710 confocal microscope and peroxidase stained sections were imaged using a Zeiss axioobserver D1 microscope. All comparative fluorescence images using the confocal microscope were imaged using the same imaging conditions.

A principle requirement for a scaffold in tissue engineering is that it must promote cell adhesion, survival, and interaction with the ECM. The cell-matrix adhesion and survival of DPSCs, PDLSCs, and HMSCs within the pulp ECM scaffolds were evaluated. Figure 1 shows representative images from the live-dead cell assay for up to 2 weeks post seeding. Green staining in these images represents live cells and red staining represents dead cells. It is evident from the figures that all three cell types survived over a period of 2 weeks. Quantitative analysis of the fluorescent images indicated that the percentage of live cells in the three cell types studied at 24 h, 1 week, and 2 weeks exceeded 98%. Additionally, there was no statistically significant difference between the percentages of live cells between time points for the same cell type or between different cell types at the same time point.

Fluorescent actin staining followed by 3D confocal imaging was performed to evaluate the three dimensional orientation of the three cell types in the pulp ECM scaffold. The images in Figure 2 are color coded for depth. It is evident from the figure that DPSCs, PDLSCs, and HMSCs show a three dimensional orientation with cell processes spanning several microns in all three dimensions. Figure 2 also provides a qualitative overview of the proliferation of the three cell types in the pulp ECM scaffold. An increase in cell density can be observed between A1 and A3, B1 and B3, and C1 and C3.

SEM imaging was performed to study the ultrastructural characteristics of the cells grown within the pulp ECM scaffolds. Results in Figure 3 show pronounced cell-matrix interactions of the three cell types with pulp ECM scaffold. The arrows in Figure 3 point to cell processes interacting with the scaffold.

Collectively, these experiments demonstrated that the DPSCs, PDLSCs, and HMSCs adhered, survived, proliferated, and interacted with their extracellular environment when cultured within the pulp ECM scaffolds.

Real time Quantitative PCR was used to analyze the expression of several differentiation marker genes that can be a predictor of odontogenic differentiation of HMSCs. Figure 4 is a graphical representation of the fold change in the expression profile of several marker genes along with their statistical significance. Notably, when HMSCs were cultured within the pulp ECM scaffold, they showed significantly higher expression of odontogenic differentiation markers such as Runx2, osteocalcin (OC), and type I collagen. Importantly, DSPP, a key marker of odontogenic differentiation was up-regulated ~3-fold. The average Ct values for DSPP expressed by HMSCs when cultured in the control and pulp ECM scaffolds were 32.83 and 30.97 respectively (average GAPDH values were around 25). This suggests that although there was a significant increase in DSPP expression, the gene expression level was low. A longer time point showed increased DSPP gene expression in PDLSCs (Ravindran et al., 2013). Therefore, HMSCs might require a longer time frame for showing increased expression levels of DSPP gene.

Growth factors TGFβ 1, FGF-1, VEGF, and GDF10 were highly expressed. TGFβ 2 was negatively regulated and BMP2 and FGF2 did not show a statistically significant change in expression. Additionally, phosphate regulating X-lined neutral endopeptidase (PHEX) that was not expressed by the HMSCs cultured within the control scaffolds (Ct value undetermined in 40 cycles) was expressed with an average Ct value of 32.66 (n = 3) by the HMSCs cultured within the pulp ECM scaffold indicating that this phosphate regulating gene expression was turned on.

Table 2 shows the expression levels of a few genes that were significantly regulated at the 2-week time point by at least 2 cell types. The data presented for DPSCs and PDLSCs are published data (Ravindran et al., 2013). Comparing the expression levels, it can be noted that the three stem cell types regulate their differentiation via different mechanisms. PHEX gene did not show a statistically significant regulation after 2 weeks in DPSCs and PDLSCs. However, the basal expression level of this gene was higher in DPSCs and PDLSCs cultured within the control scaffolds with an average Ct value of 31.83 and 28.01 respectively (n = 3).

Sections from pulp ECM scaffolds and control scaffolds containing PDLSCs and DPSCs after 4 weeks of implantation were analyzed by immunohistochemistry for odontogenic marker proteins DMP1, DSP, and DPP. The sections were also analyzed for the expression of VEGF, the pro-angiogenic growth factor. Images in Figures 5, 6 show that in the control scaffold sections, the PDLSCs and DPSCs do not show immunohistochemically detectable levels of DSP, DPP, and VEGF expression (Figures 5, 6A2–A4). However, expression of all the three proteins was observed in the sections from the pulp ECM scaffolds containing PDLSCs and DPSCs (Figures 5, 6B2–B4 with arrows pointing to positive staining). No significant change in DMP1 protein expression levels could be observed in DPSCs (comparing Figures 5A1,B1). A small increase in DMP1 protein expression was observed in PDLSCs cultured within the ECM scaffold (Comparing Figures 6A1,B1). No non-specific fluorescence was observed in rabbit and mouse secondary antibody controls (Figure 7C).

Sections from control and pulp ECM scaffolds containing HMSCs were analyzed by immunohistochemistry for the expression of DMP1, DSP, DPP, and VEGF. Figure 7 shows the result of this experiment (performed in triplicates). Comparing images in Figures 7A,B, it is evident that HMSCs expressed DSP and DPP proteins when implanted with the pulp ECM scaffolds (Figures 7B2,B3) and not the control scaffolds (Figures 7A2,A3). The white arrows in the images show the expression of these proteins. The yellow arrows in Figure 7B3 indicate non-specific fluorescence from the red blood corpuscles. Incidentally, this image also serves to show that the pulp ECM scaffolds also promoted vascularization in vivo. Additionally, an increase in the protein expression levels of DMP1 and VEGF was also noted (Comparing Figures 7A1–A4 with Figures 7B1–B4). No non-specific fluorescence was observed in rabbit and mouse secondary antibody controls (Figure 7C).

Calcium deposition within the control and ECM scaffolds implant sections was analyzed qualitatively using alizarin red staining. Figure 8 shows sections of the pulp ECM scaffold sections containing DPSCs, PDLSCs, and HMSCs. A stronger alizarin red staining in the ECM scaffolds containing all the three cell types showed increased amount of calcium deposition (Figures 8A2,B2,C2) when compared with the respective control sections (Figures 8A1,B1,C1) under the same imaging conditions. Additionally, in the DPP blocked pulp ECM sections containing DPSCs, the calcium deposition was increased with respect to the control sections (comparing Figures 8A1,A3). However, the expression was reduced with respect to the untreated pulp ECM scaffold sections containing DPSCs (comparing Figures 8A2,A3).

von Willebrand factor (VWF), a marker for endothelial cells was used to look for the presence of endothelial cells and thereby, capillary networks in the explant sections of the control and pulp ECM scaffolds containing DPSCs, PDLSCs, and HMSCs. Figures 9–11 shows representative images of the results from this experiment. Overall, we observed that in the control scaffolds (plain collagen/chitosan scaffolds) that do not contain embedded ECM, endothelial cells were present in the periphery of the scaffolds for all the three cell types [Figure 9A1 (DPSCs), 10A1 (PDLSCs), and 11A1 (HMSCs)] showing peripheral migration of host endothelial cells. However, endothelial cells were not observed within these scaffolds for all the three cell types (Figures 9A2, 10A2, 11A2).

On the other hand, in the pulp ECM scaffolds containing all the three cell types, more endothelial cells were present in the periphery (Figures 9B1, 10B1, 11B1). Additionally, endothelial cells were also observed within the scaffolds (Figures 9B2, 10B2, 11B2, black arrows). In the explant sections of ECM scaffolds containing HMSCs, in addition to individual endothelial cells (black arrows in Figure 11B2), we were also able observe the presence of capillary-like networks positively stained by VWF antibody within the scaffold. These networks showed a different staining pattern when compared to individual endothelial cells and are indicated by a white arrow in Figure 11B2. Together, these results suggest that the pulp ECM scaffolds can trigger vascularization irrespective of the mesenchymal cell type present within.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.