**3. Discussion**

Recent evidence indicates that dental pulp harbours a population of stem cells, such as DPSCs [43]. Nowadays, specific markers that may allow to uniquely define the immunophenotype of DPSCs are still unknown [48]. However, these cells express antigens that are characteristic of "classic" MSCs derived from mesodermal tissues. Such cells do not express hematopoietic markers [49]. Importantly, DPSCs were also shown to exhibit the expression of stemness-associated markers, such as *Oct-4, Nanog* and *Sox-2* [50] along with vimentin, as well as neural-specific markers such as nestin, N-tubulin and neurogenin-2 [51]. Due to the previously described unique features of such cells, one of the aims of this study was to analyse differentiation potential and other essential functional and morphological features of DPSCs in comparison with UC-MSCs. UC-MSCs may be used for allogenic or autologous applications. Moreover, accumulating evidence indicates that the transcriptomic profile of UC-MSCs indicating the expression of stemness- and bone development-related genes is similar to BM-MSCs. There is a possibility of elimination of restrictions that are characteristic for autologous BM-MSC. This might result in the replacement of these cells that are considered as gold standard for MSC research and clinical applications [16,17,52]. In the second phase of the current study, we focused on the osteogenic differentiation of DPSCs in 3D cultures in the hypoxic microenvironment resembling the stem cell niche.

We isolated DPSCs from dental pulp extracted from permanent healthy teeth of donors, whereas UC-MSCs were isolated from Wharton's Jelly of UC. Isolated DPSCs represent an adherent fraction of dental pulp tissue and are spindle-shaped, fibroblast-like cells similar in morphology to UC-MSCs. By employing flow cytometry platform, we established the antigenic profiles of DPSCs containing the expression of CD45−/CD14−/CD34−/CD29+/CD44+/CD73+/CD90+/CD105+/Stro-1+/HLA-DR<sup>−</sup>. The immunophenotype of DPSCs is similar to the antigenic profile of UC-MSCs. Subsequently, by employing quantitative RT-PCR and histochemical staining, we confirmed that DPSCs, as well as UC-MSCs exhibit the capacity for tri-lineage mesenchymal differentiation. Collectively, DPSCs, as well as UC-MSCs, fulfil the classification criteria of MSCs published by ISCT [3]. It is important to emphasize that some investigators have pointed out that the exact status of DPSCs as MSCs is still not fully defined and requires further investigation [48]. Based on the fact that the typical characteristics and immunophenotype of DPSCs fulfil MSC classification criteria, DPSCs may be considered as a population of MSC with certain unique properties resulting from their developmental origin.

Due to the ectomesenchymal origin [53], the DPSCs may be effectively differentiated into functionally active neurons [51,54] demonstrating an attractive prospect for the treatment of injuries of the nervous system. However, the comparisons between the differentiation potential of DPSCs and other "classic" MSCs toward mature cell types other than neural cells have never been performed. Thus, in the current study, we compared the differentiation capacity of DPSCs with UC-MSCs isolated from human neonatal tissue, focusing on osteogenic, chondrogenic, adipogenic, cardiomyogenic, and angiogenic differentiation in vitro. In the case of osteogenic differentiation, the expression levels of mRNA for *Runx2* and osteocalcin were comparable between the DPSCs and UC-MSCs, whereas the

expression of osteopontin was elevated in the UC-MSCs notably on the latter days following stimulation (14d). Other studies have shown that osteopontin may be expressed in many organs, with or without matrix, suggesting that this molecule may act as a structural molecule or humoral factor or cytokine and may not be a specific marker of osteogenesis [45], which may explain the elevated expression observed in UC-MSCs. Despite the comparable expression level of osteogenesis associated genes between DPSCs and UC-MSCs, histochemical staining demonstrated larger deposits of calcium phosphate following DPSC osteogenic stimulation in comparison with those in the stimulated UC-MSCs, which may indicate higher osteogenic capacity of DPSCs. Deposition of calcium phosphate, a bone mineral, is a characteristic for osteoblasts [55]. Since some investigators indicate a poor correlation between mRNA and protein expression levels, it may be necessary to extend the studies examining mRNA and protein expression in DPSCs in comparison with UC-MSCs to explain the observed higher deposition of calcium phosphate in DPSCs following osteogenic stimulation despite the low expression of mRNA [56]. Osteogenic differentiation of DPSCs was also confirmed by other investigators [57]. A higher alkaline phosphatase activity has been detected in human DPSCs cultured in osteogenic differentiation medium than that in BM-MSCs [11].

Interestingly, we also confirmed the chondrogenic potential of DPSCs and UC-MSCs. We observed higher expression of *Sox9, Acan*, and *Col10A1* mRNA in UC-MSCs, whereas the expression of *Col2A1* was comparable between both groups of SCs. Histochemical staining of DPSCs and UC-MSCs indicated the extracellular secretion of sulphated proteoglycans by both types of cells. Kinetics of differentiation was also similar between the analysed populations of SCs. Recent evidence indicates that *Col10A1* is a marker of hypertrophic chondrocytes, which may be implicated in bone growth. In contrast, it has also been observed in skeletal dysplasia and osteoarthritis disorders [44]. Thus, further investigation may be necessary to confirm the proper chondrogenic differentiation of UC-MSCs as well as DPSCs. Some investigators, however, have confirmed the higher chondrogenic potential of BM-MSCs in comparison to DPSCs [11]. Moreover, we observed a higher level of adipogenesis in UC-MSCs than that in DPSCs following stimulation, as measured by adipogenic gene expression at mRNA level. Both MSC populations secrete oil droplets, a characteristic of the ongoing adipogenic differentiation [58]. A lower adipogenic potential of DPSCs in comparison with BM-MSCs has also been demonstrated by Guo et al. [11].

In the next step, we analysed differentiation potential of DPSCs toward cells with cardiac and endothelial phenotype. After cardiomyogenic stimulation, we observed higher expression of *Gata-4, Nkx2.5*, and *Myl2c* mRNA in UC-MSCs than in DPSCs, which was confirmed by immunocytochemical staining. In the case of endothelial cell formation, elevated expression of angiogenesis-related genes (*Gata-2, Tie2*, VE-cadherin) was observed in UC-MSCs after seven days of stimulation with VEGF, whereas DPSCs were typically unresponsive toward such proangiogenic stimulation. The results suggested q limited switch towards cardiac and endothelial phenotype in UC-MSCs, while the signs of differentiation were hardly detected in DPSCs. However, other investigators have shown that in the co-culture of DPSCs with neonatal rat cardiomyocytes, the expression of cardiac-specific markers (including troponin I and β-myosin heavy chain) was increased in DPSCs shortly after co-incubation, suggesting a possible transient phenotypical switch into cardiac lineage under certain circumstances [59].

Since DPSCs may be effectively differentiated into osteoblasts, we also encapsulated these cells in the hydrogel matrix to create a 3D culture niche mimicking tissue and organ-specific microarchitecture to conduct osteogenic differentiation in vitro. Nevertheless, we observed a greater relative proliferation of DPSCs in 2D culture in comparison with 3D culture. Moreover, contrary to our initial expectations, we did not observe any influence of oxygen concentration on the proliferation ratio of DPSCs in both 2D and 3D culture at predetermined intervals. However, the increased proliferation of DPSCs in 2D culture was correlated with higher metabolic activity, whereas DPSCs in 3D culture exhibited lower metabolic activity correlating with lower proliferation and higher differentiation capacity in such conditions. It has been demonstrated that 3D culture allows the cells to maintain a better cell-to-cell

contact and intercellular signalling network [60]. In the case of a hypoxic environment, we set up oxygen concentration equal to 2%, which was in the range of O2 concentration measured by other investigators in murine BM (1.54–2.0%) [61]. For a normoxic culture environment, Wegner et al. showed that the O2 concentration should be approximately 18% [62]. Hence, we assumed that the O2 concentration in the normoxic environment was approximately 18%. The typical mammalian cell culture conditions inside an incubator also include the temperature of 37 ◦C, 95% relative humidity and CO2 concentration of approximately 5%.

Kwon et al. has shown increased proliferation rate and percentage of DPSCs in S-phase under 5% hypoxic culture environment from day 5 of the culture accompanied with enhanced osteogenic di fferentiation capacity of these cells in comparison with the normoxic environment [63], whereas Iida et al. observed the inhibition of the osteo/odontogenic di fferentiation capacity of such cells, despite the enhanced proliferation rate of DPSCs under hypoxic environment [64]. Results related to the proliferation of DPSCs cultured under hypoxia have been controversial, which may be attributed to the lack of standardization of cell culture protocols [65]. In this study, we cultured DPSCs in the presence of 2% O2, whereas other investigators cultured cells in the presence of 5% or 3% O2 [63,64].

A decrease in cell proliferation might often be accompanied by the induction of cell di fferentiation [66] and 3D culture may facilitate the developmental processes allowing the cells to di fferentiate into more complex structures [67]. Thus, we analysed the osteogenic di fferentiation of DPSCs in 3D cultures under hypoxic or normoxic conditions.

We also analysed the osteogenic di fferentiation of DPSCs in 3D culture under hypoxic and normoxic conditions. The expression of mRNA corresponding to Runx2–osteogenesis regulating transcription factor-was elevated by approximately thrice the initial amount in DPSCs in 2D and 3D culture in the presence of 2% O2, compared to that in the undi fferentiated cells. The elevated expression of *Runx2* in a 3D culture in an environment containing 2% O2 persisted up to day 14, suggesting long-term osteogenic induction. We observed the highest increase in the expression of early and late osteogenic markers such as *Runx2* and *Col1A*, respectively, at every interval in 3D culture in microenvironment under hypoxia than that in the 3D culture in a normoxic microenvironment. In the case of 3D culture, we also observed large rounded deposits of calcium phosphate, primarily in the presence of hypoxia.

Taken together, our results indicate that 3D cell encapsulation as well as low concentrations of oxygen may favour the osteogenic di fferentiation of DPSCs in vitro. The culture of human MSCs under hypoxic conditions enhances the di fferentiation into osteocytes. These cells produce more osteomatrix and the expression of osteogenesis-related genes (e.g., *Runx2, ALP*) is up-regulated [68,69]. Burakova et al. suggested that the e ffect of hypoxia on MSCs may be driven by specific factors explaining the opposite influence of low oxygen concentration observed by many investigators [70]. Recent evidence indicates that the 3D sca ffold may have a major impact on the proliferation and activation of MSCs, whereas the 3D architecture supports the osteogenic di fferentiation of MSCs and enhances the production of mineralized bone matrix [71]. Increased expression of *Runx2* and *Col1A1* characterizing osteogenic di fferentiation of DPSCs in 3D culture under hypoxic environment observed in our study is consistent with recent evidence demonstrated by other investigators, confirming that the di fferentiation of MSCs encapsulated in alginate beads significantly increases the osteogenic di fferentiation of these cells in vitro [35]. Moreover, in the novel bone development or bone regeneration, vasculature delivers nutrients, growth factors, minerals, and oxygen. The sca ffolds for bone regeneration should allow endothelial cell migration and the growth of new vessels, which may induce VEGF expression promoting the migration and proliferation of endothelial cells and positively influence the bone regeneration process [72]. After the co-culture of DPSCs and endothelial cell line EA.hy926 on chitlac-coated thermosets, an increased proliferation with enhanced DPSC osteogenic di fferentiation and EA.hy926 vessel formation was observed [73], confirming positive feedback of both cell fractions.

In our experiments, we used primary human DPSCs isolated from dental pulp derived from the teeth of three donors, whereas UC-MSCs were isolated from two umbilical cords. Moreover, for qRT-PCR analysis, every sample (prepared for each DPSCs line from each donor) was run in duplicate, whereas for the proliferation and metabolic activity assessment, every sample was run in triplicate. In some cases, we observed a high error of the mean (SEM), which may be the result of biological variability related to primary human tissues obtained from various donors. Data published by other investigators indicate that the DPSCs isolated from the patients at different ages exhibit different doubling time, expression of CD29 antigen, and percentage of apoptotic cells during propagation along with differentiation potential [74], which may have an additional impact on the variability of the cell response. Moreover, the DPSCs isolated from different teeth may possess varied differentiation capacity or immunomodulatory potential [75]. The high variation of the response and data presented in our study is not unusual during the analyses of SCs, including DPSCs. Other investigators have also demonstrated highly variable data from analysis of different populations of SCs, including DPSCs (results with high SEM values) [76]. Therefore, further investigations are required to find a correlation between the type of dental pulp derived from different teeth, the age of the donor, and the biological potential of DPSCs. Proper protocols might be developed for the preparation of optimal DPSC fraction as a product for advanced therapy with high pro-regenerative potential.

To summarize, we observed that human DPSCs possess a higher potential for osteogenic differentiation and similar chondrogenic and adipogenic differentiation potential compared to the "classic" MSCs such as UC-MSCs but a limited ability for the cardiac or endothelial phenotype in vitro. Moreover, the hypoxic environment especially in 3D culture conditions may enhance the osteogenic capacity of DPSCs in vitro. DPSCs encapsulated in injectable hydrogel might be a candidate for innovative advanced therapy for the treatment of bone (including alveolar bone) loss.
