The equine periodontal ligament
The equine PDL simultaneously meets the opposing requirements of tooth support and conti-nuous tooth eruption at an exceptionally high rate under physiological conditions [25, 31]. These distinct functions depend on dynamic properties which allow continuous periodontal remodeling in terms of renewal of dental cementum, periodontal ligament and alveolar bone [13, 32, 34]. Accordingly, periodontal remodeling is based on the presence of multiple cell types which are able to replenish the different tissues of the periodontium in a well orchestrated process [12, 35, 36].
It is widely considered that such complex processes as periodontal remodeling and functional regeneration depend on the presence of MSC within the PDL [18, 32, 37]. The presence of MSC has already been demonstrated in the PDL of men, e.g. [18–20], rats and sheep [14, 21] but not in the PDL of horses. To the best of our knowledge this is the first study demonstrating the presence of MSC in the equine PDL.
Equine periodontal and subcutaneous MSC were identified according to distinct in vitro characteristics, i.e. plastic adherence, self-renewal capacity, expression of the stemness markers CD90 and CD105 and trilineage differentiation potency. Due to the fact that equine periodontal MSC have never been reported before, no comparative data exist. We therefore compared our results with non-equine periodontal MSC and also with non-periodontal equine MSC.
All harvested tissue samples contained plastic adherent cells which gave rise to primary cul-tures of fibroblast-like cells. These cells matched the typical in vitro appearance of fibroblasts from the PDL of non-equine species [38–40]. The presence of vimentin confirmed an ectomesenchymal origin of the cells and distinguished them from epithelial cells and endothelial cells.
In vitro self-renewal capacity of MSC is routinely demonstrated by CFU-F assays and doubling time experiments. Obtained results provide valuable information when considering prospective utilization of the investigated MSC for therapeutical use. The CFU-F efficiency is correlated with the quantity of MSC within their original in vivo tissue . The in vitro doubling time provides quantitative information about the ability of the cells to expand in culture.
Stro1+/CD146+ selected human periodontal MSC exhibited a CFU-F efficiency of 19.3% , which is in the same range as the values obtained for eP-MSCap (18.45%) and for eP-MSCm (17.45%). However, cells from other parts of the equine periodontium (eP-MSCsg and eG-MSC) possess lower CFU-F efficiencies of 13.43% and 13.50%. Given that the CFU-F efficiency represents an in vitro enumeration of a clonogenic subset of MSC in vivo as demonstrated by Kuznetsov et al. (2009) , the apical part of the equine PDL contains more MSC than other parts of the equine periodontium. The knowledge of site-specific quantities of MSC in the equine PDL might be of practical relevance regarding protocols for MSC isolation for further investigations. For non-equine species site-specific differences in the availability of PDL MSC have not been reported so far; and it is assumed that this issue remains important only for equine periodontal research due to the enormous length of equine teeth compared to other investigated species. However, even the highest CFU-F values of our cells were lower compared to CFU-F values obtained in equine bone marrow MSC, which show mean CFU-F rates of 27% .
Population doubling time
Reported doubling times for equine MSC from different sources (bone marrow, adipose tissue) are in a range of 2.4 to 5 days [44, 45]. These values are almost the same as demonstrated in equine periodontal MSC which double in number in 3.5 to 6 days. Cultured eP-MSCap and eP-MSCm proliferate at a significantly higher rate than eP-MSCsg. These in vitro findings are supplemented by corresponding in situ studies of the equine PDL. Warhonowicz et al. (2006)  demonstrated an elevated proliferation index in the apical level with decreasing proliferation indices towards the subgingival level. Such an asymmetric proliferation index with highest proliferative activity in the apical part of the PDL has been identified as playing a crucial role in continued tooth eruption [46, 47].
During osteogenic differentiation cells altered their shape and assembled in clusters. The cluster formation is considered a typical feature of osteogenic MSC differentiation in vitro . Positive von Kossa staining confirmed the presence of calcium apatite in the extracellular matrix and thus also demonstrated successful osteogenic differentiation . In addition previous studies showed that non-periodontal equine MSC formed visibly mineralized nodules within three weeks when cultured under osteogenic conditions [43, 44, 48]. In contrast our cells (ePDL-MSC, eG-MSC, eSc-MSC) did not show any presence of mineralized areas before day 28. This observation is in line with studies on non-equine periodontal MSC and suggests a suppressed capacity for extracellular matrix mineralization [20, 48, 49].
Equine periodontal and subcutaneous MSC showed early stages of adipogenic differentiation when cultured for 23 days in a conventional differentiation medium. The adipogenic differentiation was confirmed by the detection of small intracellular lipid droplets (oil red O staining). Yet, a final differentiation into mature adipocytes containing large, fused lipid vacuoles was not achieved. This is in line with the findings of other authors who have reported that equine MSC did not differentiate into mature adipocytes . In a second experimental setting adipogenic differentiation was induced by incubation in a differentiation medium containing rabbit serum (5%) but no additional growth factors (i.e. dexamethasone, indomethacine, 3-isobutylmethylxanthine, insulin). This method had been proven to induce adipogenic differentiation in non-periodontal equine MSC at an optimal rate with minimized detachment of cells . By using rabbit serum we observed accelerated adipogenesis (only three days culture time in induction medium) as also reported by recent investigations [44, 51, 52]. However, also in these experiments a terminal adipogenic differentiation with cells containing large, fused lipid vacuoles has not been observed. Specific investigations addressing the adipogenic differentiation capacity of adipose derived MSC assessed the expression of peroxisome proliferator activated receptor γ2 (PPARγ2) which plays an essential role in adipogenesis and has been widely accepted as a marker for terminal adipogenic differentiation [53, 54]. Interestingly, even those equine adipose derived MSC did not display terminal adipogenic differentiation in vitro . Nevertheless, although terminal adipogenic differentiation seems to be hampered in currently used in vitro culture systems for equine MSC, the demonstrated early adipogenic differentiation stage has been generally accepted to test adipogenic differentiation capacity [43, 50, 55].
Collagens and sulphated proteoglycans are characteristic constituents of the extracellular matrix in cartilage. Their deposition in chondrogenic-induced pellet cultures can be easily demonstrated by Masson-Goldner-Trichrome and toluidine blue staining . However, staining intensity is influenced by several factors and obtained data should be confirmed by determining marker mRNAs for chondrogenic differentiation . We therefore conducted RT-PCR experiments and demonstrated the expression of mRNA for collagen I, COMP and aggrecan in all chondrogenic-induced cell cultures. Interestingly, only eP-MSCap and eP-MSCm expressed transcripts for collagen II. According to the current concept of in vitro chondrogenesis of MSC, collagen I and COMP become upregulated in an early stage of chondrogenic differentiation. The expression of mRNA for aggrecan represents an intermediate stage and collagen II mRNA is expressed in a final stage . Thus, only eP-MSCap and eP-MSCm passed all stages of in vitro chondrogenesis, while eP-MSCsg, eG-MSC and eSc-MSC displayed hampered chondrogenesis in vitro. Further investigations are needed to clarify whether these in vitro results reflect in vivo characteristics of the investigated MSC.
Current therapeutic use of equine MSC
Recently, the use of so-called mesenchymal stem cells in equine medicine has gained a lot of scientific and commercial interest. However, applied cellular products have been defined according to different protocols and it is impossible to verify whether true MSC are used in different investigations [44, 57].
Nevertheless, supposed MSC from different tissue sources (sternal bone marrow and adipose tissue) have been therapeutically used for regenerative therapies of typical equine musculoskeletal diseases, i.e. osteoarthritis [58–60] and core lesions in the superficial digital flexor tendon [61, 62]. Cell-based regenerative therapy of equine tendinopathies turned out to improve clinical outcomes compared to conservative therapies [59, 63, 64]. Further, cell injections resulted in significantly improved tendons histologically [65, 66]. At present, reported outcomes are still far from the biomechanical features of a healthy tendon and the currently used regenerative therapies need to be improved. The search for a tendon-like tissue containing available MSC populations has been identified as a promising approach in order to optimize cell-based therapy of equine tendon injuries . This consideration is supported by the finding that MSC form different tissue sources possess different cellular properties due to the regulatory influence of their natural local microenvironment [67, 68]. Significantly, tendon-derived MSC show a higher capacity for tenogenic differentiation when compared with bone marrow-derived MSC [49, 69]. Unfortunately, the prevalence of MSC in tendons is very low and isolating suitable cell numbers appears to be impractical .
Future prospects for the use of equine periodontal MSC
Considering a suggested tendon-like tissue source of MSC for equine tendon therapies, the obtained equine periodontal MSC might be promising candidates for such MSC. These isolated cells definitely possess typical MSC characteristics (plastic adherence, self-renewal capacity, and trilineage differentiation potency) and are obtained from a natural niche which greatly resembles tendon tissue. The particular in vivo function of periodontal MSC is reflected in their high in vitro expression of scleraxis, a tendon-specific transcription factor. Scleraxis expression is significantly higher in human periodontal ligament MSC when compared with human bone marrow derived MSC [20, 70, 71].
Equine periodontal MSC might also be a useful tool in order to develop successful therapies for equine periodontal disorders. Especially in aged horses periodontal diseases are a frequent problem with an incidence of up to 60 percent, often leading to tooth loss . The search for predictable periodontal regeneration utilizing periodontal MSC has also attracted a lot of interest in the field of human periodontology and several promising therapeutical strategies have been proposed in the last years (for review see Huang et al. 2009 ).
Yet, a major problem of the use of equine periodontal MSC arises from their limited accessi-bility. Obtaining these cells for autologous applications can not be taken in to consideration. Hence, allogenic application techniques are required. Fortunately, recent investigations confirmed that MSC avoid or suppress immunological responses, usually causing rejection of allogeneic delivered cells [74, 75]. Such remarkable immunomodulatory properties of MSC have also been explicitly demonstrated for human periodontal MSC .
Moreover, allogeneic application of equine MSC in diseased tendons was already been performed in experimental studies without causing immune response or tumor formation [62, 77, 78]. These clinical results have recently been supplemented by in vitro investigations which demonstrated the absence of MHC class II (a crucial immune activator) on equine MSC derived from bone marrow .
Future prospects for classification and characterization of equine periodontal MSC
To provide an objective and comprehensive classification of the cells investigated, the rec-ommendations of the International Society for Cellular Therapies (ISCT) for the identification of non-human MSC  were applied. In our study the colony-forming cells showed the required characteristics, i.e. adherence to plastic culture dishes, and in vitro differentiation into osteoblastic, adipogenic and chondrogenic cells . As those minimal criteria for the definition of non-human MSC were met by the isolated cells they were termed MSC.
For human MSC a third criterion is required, i.e. a well-defined profile of surface antigens [55, 33]. A human MSC population should contain more than 95% of cells which express the surface makers CD73, CD90 and CD105, and less than 2% of the cells should express CD45, CD34, CD14 or CD11b, CD79α or CD19 and HLA class II [55, 33, 79]. Such a strict definition leads to a standardized and clear denomination of MSC and provides a substantial basis to compare results from experiments with MSC derived from different tissues of the body [55, 79]. Only recently has the surface marker expression of human MSC from different dental tissues been thoroughly investigated and a useful panel of identifying marker molecules been recommended [73, 80].
The difficulties associated with the establishment of uniform parameters for the characteriza-tion of putative MSC in human research are even more complicated in veterinary science. Unfortunately, the surface antigen expression of equine cells, in particular of equine MSC, must still be regarded as largely unknown [50, 79, 81]. Nevertheless, encouraging investigations showed reactivity of available antibodies against CD90 in equine bone marrow derived MSC , against CD 90 in equine adipose tissue derived MSC [44, 50] and against CD105 in equine adipose tissue derived MSC . Our results demonstrate that also MSC derived from other tissues of the equine body express CD90 and CD105 suggesting that these proteins might be used as universal markers for MSC in the horse. However, currently the panel of available equine stemness markers is very limited. The human MSC surface marker CD73 has been detected at an mRNA-level in equine MSC but not at a protein level so far . Conflicting data exists for the expression of another putative MSC marker, CD13. This marker was recognized on equine MSC derived from peripheral blood but it was absent in MSC derived from adipose tissue [82, 83]. Also demanded proof of the non-expression of particular CD antigens is still undetermined. Guest et al. (2008)  confirmed the non-expression of CD14 in equine MSC. However, others identified CD14 as an equine-specific characteristic of MSC . Similar contradictory results have been reported for the expression  or non-expression  of the embryonic stem cell gene Oct4 in equine MSC. These inconsistent results emphasize the urgent need for future studies to identify and establish a useful and reliable panel of specific surface markers for equine MSC. Such surface markers would supplement and alleviate cell characterization and, even more importantly, would enable effective techniques for cell selection (immunomagnetic or fluorescence activated cell sorting). However, as long as an identifying antibody panel for equine MSC is not established, plastic adherence of colony-forming cells and trilineage differentiation capacity, should be regarded as minimal but adequate criteria for the identification of equine MSC [55, 84].