Mesenchymal progenitor cells in human dental pulp: purification and characterisation
INTRODUCTION: Characterising the dental pulp perivascular cells will allow understanding whether these blood vessel associated cells have tissue regenerative potential that can be used for craniofacial regeneration, including bone, nerve, and tooth repair. Perivascular cells (pericytes and adventitial cells) can be purified from multiple human organs including placenta, bone marrow, pancreas, fat and muscle, and exhibit in culture and in vivo multilineage mesodermal developmental potential, implying the perivascular origin of the elusive, culture-derived mesenchymal stem cells (MSCs). Perivascular cells are ubiquitous in the organism and their developmental potential is partially imprinted by their tissue of origin. Pericytes in the dental pulp, which are deriving from primary or permanent teeth, originate from migrating cranial neural crest cells, which, theoretically, could exhibit superior potential to regenerate neural tissue. To test this, my aim was to characterise dental pulp cells and to evaluate their regenerative capacity. MATERIAL AND METHODS: To achieve this goal, we identified and isolated perivascular cells present in the human dental pulp. Dental pulp tissue was isolated from healthy third molar teeth from adults younger than 30 years old. Firstly, immunohistochemistry was performed to determine the presence of cell surface markers on dental pulp pericytes and adventitial cells using markers such as αSMA, PDGFRβ, LepR, CD107a, CD146, NG2 and CD34. Blood vessels were next characterised using MSC markers such as CD44, CD90, and CD29 and endothelial cell markers CD54, CD34, UEA-1, and vWF on pulp sections by immunofluorescence microscopy. Secondly, purification of dental pulp perivascular cells was performed using flow cytometry, followed by culturing and differentiation assays towards osteogenic, adipogenic and neurogenic cell lineages. RESULTS: We found that the phenotype of pericytes in the dental pulp in vivo is similar to that of pericytes in other human organs previously documented. They express CD146, ALP and PDGFRβ and lack endothelial cell markers such as CD34, vWF, CD54 or UEA-1. Importantly, we also found that pericytes are heterogeneous and can be further discriminated based on NG2, αSMA and LepR that coincide with a particular blood vessel type or blood vessel diameter. For example, αSMA is mainly expressed in microvascular pericytes and occasionally in capillaries, similar to LepR. NG2 is not expressed in venules but it is expressed in capillaries and arterioles. In contrast, MSC markers CD90, CD44, and CD29 were expressed on all pericytes suggesting that MSCs reside in heterogeneous perivascular niches. Dental pulp mesenchymal stromal cells were next derived from CD146+CD34-CD56-CD45- pericytes or NG2+ and NG2- pericyte subsets and from CD146-CD34+CD56-CD45- adventitial cells, another perivascular cell source of MSCs located around large vessels. Our data show that, similar to MSCs, all perivascular cell subsets we purified were able to adhere to plastic dish, to expand in culture, and to efficiently accumulate calcium deposits upon osteogenic induction, reminiscent of bone development. However, most of cell lines tested were not adipogenic with the exception of rare cells in the NG2+ pericytes subset. Indeed, lipid droplets were consistently found accumulating only in these purified pericytes from 14 days onwards, albeit at a low frequency. Finally, we discovered that, in situ, neural cells, positive for CD56 and vimentin, share markers with pericytes such as NG2 and CD146. We also found that in vitro, our perivascular cell-derived mesenchymal stromal cell lines express tubulin β III upon neuronal induction although these results require further confirmation. CONCLUSIONS: We documented the molecular signature of pericytes in the human dental pulp. We demonstrated that perivascular cell subsets can be prospectively purified to homogeneity and that perivascular cell-derived stromal cell cultures can be established in vitro. Intriguingly, our in vitro MSC functional assays showed that these cells are not multipotent, and thus are not genuine MSCs. They are, in contrast, very potent bone progenitors with a low frequency of arterial pericytes being adipogenic. Importantly, we here showed a marker overlap between neural cells and pericytes. Whether these cell types are developmentally linked remains to be established.