The Microanatomy of ETCs
In sprouting angiogenesis, endothelial cells must orientate in the tissue environment to effectively invade tissues and form vascular patterns according to the local needs. ETCs respond to vascular endothelial growth factor-A (VEGF-A) by guided migration while the proliferative response to VEGF-A occurs in the sprout stalks (Gerhardt et al.,2003). ETCs prolongations, which are usually described as filopodia (Gerhardt and Betsholtz,2005; Cullen et al.,2011; Siemerink et al.,2012), act as environmental sensors (Cullen et al.,2011) and the ETCs undergo chemotaxis toward angiogenic factors (Tung et al.,2012). Sprouting ETCs are followed by proliferating endothelial stalk cells that are rapidly ensheathed by pericytes (Cullen et al.,2011).
The oral mucosa appears as a good model for studying the process of sprouting angiogenesis and the ETCs. The in situ evidence denotes intensive processes of angiogenesis, seemingly with no obvious differences between the normal and the reparatory mucosa. CD34 labeling and TEM evaluation of ETCs demonstrated that these cells have processes of two morphological types: (a) short protrusions, filopodial like, closely related to neighbor StrCs; (b) long protrusions which are moniliform, and not filopodial as usually considered (see the previous paragraph). The moniliform morphology could be determined by attractive and repulsive forces of the local environment, similar to those guiding axons during the formation of the insect tracheal system (Gerhardt et al.,2003; Gerhardt and Betsholtz,2005). Moreover, collaterals arising from the dilated segments of the ETCs long processes mechanically support the guided migration of those processes while the dilations of these processes could simply indicate their stage of elongation. Observed by TEM the presence of such moniliform processes in the stromal compartment could lead to misinterpretations. A novel class of interstitial cells was recently characterized: the telocytes, which are resident StrCs with prolongations termed telopodes (Kostin,2010; Popescu and Faussone-Pellegrini,2010; Popescu et al.,2011; Nicolescu et al.,2012; Rusu et al.,2012a,b). Telopodes are defined as cells presenting moniliform prolongations, with dilated segments termed podoms. Therefore, TEM distinction between telopodes of telocytes and the moniliform processes of ETCs should come only after identifying the cell body.
The Stromal Environment of ETCs in the Oral Mucosa
In both control and reparatory mucosa, StrCs were identified. These cells had a fibroblastic appearance: spindle shaped, bipolar and were immune positive for the same panel of markers: CD34/CD44/CD45/CD105/PDGFR-α/vimentin. As these cells had mixed morphological and molecular characteristics of hematopoietic stem cells, monocytes and fibroblasts, they were identified as CFCs (see also “Introduction” Section). CFCs were closely related to, or were contacting stromal ETCs. The presence of CFCs, which are believed to promote angiogenesis (Hartlapp et al.,2001; Pilling et al.,2003), appeared to support the extensive sprouting angiogenesis observed in the present study. It should be emphasized also that angiogenesis is not the only process that can occur in adult tissues. For example, endothelial progenitor cells may also participate in neovascularization during ischemic conditions (adult vasculogenesis), and these cells are bone marrow derived (Tepper et al.,2005). There are studies which suggest that bone marrow derived mesenchymal stem cells, endothelial progenitor cells and fibrocytes may be involved in wound healing processes (Wu et al.,2007).
During wound healing of the oral mucosa, CFCs around fibrotic lesions acquired a myofibroblastic phenotype being α-SMA positive. Loss of progenitor markers and gain of a myofibroblastic phenotype correlates with the maturation of the CFCs (Chen et al.,2010). As CFCs populate the mucosa in the absence of mucosa wounds, a prompt phenotypic switch from CFCs to myofibroblasts occurred that likely ensured a quick wound contraction response. This would reduce scar formation during the healing process.
When observed in TEM most of the nonimmune StrCs had morphologies that suggested they were more involved with signaling than with collagen secretion, as are the canonical fibroblasts. The scarcity of secreting fibroblasts in the oral mucosa may provide clues to help understand reduced scar formation during oral mucosa healing, when compared to normal skin. This argument is supported by the results of an experiment which demonstrate that even though the population of fibroblasts increased immediately after surgery, at 4 weeks the fibroblasts density decreased (Berglundh et al.,2007).
The ultrastructural phenotype of CFCs resembled that described for interstitial Cajal cells (Faussone-Pellegrini and Thuneberg,1999; Thuneberg,1999; Davidson and McCloskey,2005) and for telocytes (Popescu and Faussone-Pellegrini,2010; Popescu et al.,2011). However, the interstitial Cajal cells populate the gastrointestinal tract, and the telocytes can be identified in TEM firstly based on the presence of telopodes (Rusu et al.,2012b), and secondly by different morphological features (Rusu et al.,2012a). Telopodes were not observed in any of the samples used in this study.
The CD117/c-kit phenotype of CFCs was negative in control samples, but it was positive in the reparatory samples, and this phenotypic switch could be related with the reparatory process. PDGFR-α positive and CD117/c-kit negative mesenchymal cells, juxtaposed to interstitial Cajal cells, were presumed as being involved in gastrointestinal functions (Iino et al.,2009; Iino and Nojyo,2009; Cobine et al.,2011), but their function is still controversial (Kurahashi et al.,2011). PDGFR positive cells were also associated with chronic tissue inflammation (Rubin et al.,1988). The PDGFR phenotype of the CFCs could be related to the extensive processes of angiogenesis that were observed in this study.
Structurally, PDGFs consist of dimeric isoforms of A and B chains in all three combinations possible with the α-receptor binding to all three forms of PDGF (Sundberg et al.,1993). PDGFR-α acts as an inductor of angiogenesis, while PDGFR-ß is essential for vascular stability (Zhang et al.,2009). Taking these into account, the CFCs and pericytes of the mandibular ridge mucosa appear to ensure an adequate microenvironment for processes of angiogenesis. Moreover, the induction of angiogenesis is also related to MCs. Specifically, MCs products stimulate migration and/or proliferation of endothelial cells and degrade the extracellular matrix to provide space for angiogenic sprouts to form, PDGF facilitates the recruitment of MCs to sites of angiogenesis, and the MC tryptase is an angiogenesis-inducing molecule (Meininger and Zetter,1992; Metcalfe et al.,1997; Norrby,2002; Ribatti et al.,2002). FGF acts also as a key factor in sprouting angiogenesis (Tepper et al.,2005; Woad et al.,2012). FGF2 was consistently expressed in non-reparatory, but not in reparatory samples; this suggests that FGF2 acts in physiological reparatory processes, but it may not intervene in wound healing.
The FGF2 positive epithelial phenotype that was found may be associated to a degree with epithelial dysplasia that may be supported, in turn, by the basal and suprabasal Ki67 positive labeling. For example, it was shown that in oral epithelial dysplasia, Ki67 is frequently expressed in the basal and suprabasal epithelial layers (always jointly) (Gonzalez-Moles et al.,2000).
Pericytes are a crucial target for anti-angiogenic therapies (Morikawa and Ezaki,2011). The presence of pericytes appears to protect endothelial cells against inhibition of VEGF signaling. The simultaneous inhibition of PDGF receptors on pericytes improves the effect of VEGF inhibitors on endothelial cells and thus enhances anti-angiogenic therapies (Hellberg et al.,2010). Pericytes are recruited from the periphery and are involved in blood vessel stabilization during ischemia-induced angiogenesis (Kokovay et al.,2006). Additionally, also bone marrow derived pericytes can contribute to vascularization (Cai et al.,2009).
A synergy between the endothelial cells and pericytes is essential to the stabilization and maturation of blood microvessels. Using an in vitro tissue-engineered model of angiogenesis, it was demonstrated that the newly formed endothelial tubes recruit pericytes from the fibroblast population in that model. PDGF appeared to be a major factor involved in the recruitment of pericytes (Berthod et al.,2012).
Some caution must be exercised when defining cells as pericytes to avoid confusion. A periendothelial cell embedded in the endothelial basal lamina is correctly termed as pericyte. Endothelial stalk cells are indeed closely related to pericytes, as also are the sprouting ETCs. However, ETCs emerge from the endothelial basal lamina and penetrate the connective stroma where various cells may influence them. This differs from pericytes. CFCs contact ETCs, and consequently, when a new endothelial tube is established, the CFCs could be recruited as pericytes. The density of MCs is also directly related to angiogenesis and microvessel density (Muramatsu et al.,2000; Takanami et al.,2000; Ribatti et al.,2007).
An experimental study performed in mice on a nerve scar model identified two types of pericytes: type B, desmin– and/or α smooth muscle actin–positive, and type A, expressing PDGFR–α and –ß, and CD13. It was observed that the lesion sites were populated with blood vessel sprouts with an increased density of associated type A pericytes, and many such pericytes “lost” contact with blood vessels at the lesion site and passed into the perivascular stroma (Goritz et al.,2011). Interestingly, in that study (Goritz et al.,2011), TEM revealed a telocyte morphology of the type A pericyte which, in our opinion, should be considered as a StrC after it breaks the basal endothelial lamina. Unfortunately, the respective study did not take into account and discuss these features, and neither did it consider the presence of ETCs within the scar tissue.
The scenario presented above is different from that described in studies of human skeletal muscle study that brought arguments to support the fact that “the cells identified so far as pericytes by their immunophenotype might also include other types of interstitial cells such as telocytes” (Suciu et al.,2012). This study did not distinguish the telocytes from ETCs, nor did it discuss the stromal influences on this peculiar cell type. Furthermore, it did not consider approaching a differential identification between telocytes and CFCs. It is, however, more supportive of our findings in the oral mucosa, in which the stromal compartment is already configured when ETCs invade it. The distinction between CFCs, lacking telopodes, and telocytes, defined mostly by their specific telopodes (Popescu and Faussone-Pellegrini,2010; Popescu et al.,2011; Nicolescu et al.,2012; Rusu et al.,2012a,b,c), is important since telocytes are resident StrCs, but CFCs are bone marrow derived cells.
Changes in wound bed vascularity are significantly less in oral wounds than in skin wounds (Szpaderska et al.,2005). The difference could be attributed to the CFCs which normally populate the oral mucosa. Maintenance and healing of the oral mucosa are supported by extensive processes of angiogenesis, likely guided by ETCs. The later are influenced by both CFCs and resident StrCs. There is some doubt, however, about the presence of telocytes in the oral mucosa stromal compartment.