In the equine veterinary field, orthopaedic injuries are a major cause of retirement of athletic horses
. As a result, it is not surprising that equine regenerative medicine is primarily focused on the treatment of musculoskeletal defects. The present cell therapy studies are carried out with MSCs
[12, 13, 18] and non-adult stem cells
[19–21]. To better understand the mechanisms of action of MSCs in vivo, a large number of studies to characterise equine MSCs have been reported over the last five years
[22–26]. However, because the overall objective of regenerative treatments is the use of MSCs in live horses, it is important to determine all of the properties of MSCs in an oxygen environment that closely emulates the original physiological niche from which the cells derive. To our knowledge, the current work constitutes the first study to perform an analysis of the influence of oxygen tension on proliferation, viability, stemness and marker expression in equine MSCs derived from bone marrow and adipose tissue.
The effects of hypoxia on MSC proliferation have been studied specifically in humans and mice. Enhancements in cell growth following exposure to hypoxia have been described
[10, 11, 27]. However, there is no unanimous consent, Feher et al. (2010) reported no difference in the growth of normoxic and hypoxic cells, and Volker et al. (2010) described similar numbers of cells for both oxygen conditions at the conclusion of the culture period. In addition, Holzwarth et al. (2010), Zeng et al. (2011) and Wang et al. (2005) reported that low oxygen tension inhibited the proliferation of MSCs. Similarly, canine MSCs derived from bone marrow and adipose tissue exposed to atmospheric O2 show more proliferative capacity than those expanded from passage 1 to passage 3 under hypoxic conditions (1% or 5% O2)
. In agreement with these findings, our results describing the proliferation of equine cells as a function of oxygen tension showed that the growth of AT-MSCs was significantly higher at atmospheric oxygen tension, while BM-MSCs underwent also more proliferation in 20% O2.
Differences in cell growth between cultures expanded under different oxygen conditions could result from cell cycle changes or alterations of cell viability. Human MSC populations derived from umbilical cord and bone marrow accumulate cells in G0/G1 phase under low oxygen tension
[9, 29]. Similar to these experiments, we found that hypoxic BM-MSCs displayed a higher percentage of cells in G0/G1 phases than normoxic BM-MSCs throughout the entire culture period. Moreover, the significantly higher proportion of normoxic BM-MSCs involved in the active stages of cell division (S or G2/M) during the median days of culture led to a higher number of BM-MSCs at the conclusion of proliferation assay in the normoxic culture. Cellular arrest in G0/G1 phase in hypoxic BM-MSCs might be caused by up-regulation of cyclin-dependent kinase inhibitors that control the cell cycle checkpoint
In contrast to BM-MSCs, differences observed in the proliferation of normoxic and hypoxic AT-MSC cultures were not due to cell cycle variations, but to variations in cell viability. Similarly to rat MSCs, that undergo a reduction in cell viability when permanently exposed to hypoxia
, in our work the proportions of viable AT-MSCs in hypoxic cultures were always lower than those in normoxic cultures. Reduced viability in hypoxic conditions reflects insufficient adaptation of AT-MSCs at 5% O2, as higher percentages of non-viable cells were found in hypoxic conditions relative to populations at 20% O2. No detectable changes in apoptosis have been previously described for hypoxic MSCs
[34, 35]; our results corroborate these reports since the proportion of AnV+PI- did not display statistical differences between normoxic and hypoxic MSCs derived from the same source.
Moreover, AT-MSCs under either oxygen tension adapted more poorly to the culture environment following trypsinisation than BM-MSCs, as shown by a significantly higher proportion of PI+ AT-MSCs at days 1 and 2 for both 5% and 20% O2 atmospheres. This is reflected in the increased lag phase displayed by AT-MSCs. However, AT-MSCs also showed a significantly increased proportion of cells undergoing cell division during the first days of culture. The increase in cell division of viable AT-MSCs might compensate for cell death in the population because the final number of cells obtained at the end of the experiment was higher in AT-MSC cultures than in BM-MSC cultures, which indicates a higher proliferative ability for AT-MSCs than BM-MSCs. This result is in agreement with previous reports in horses
 and other species as canine
 and human
, which demonstrated AT-MSCs proliferated more rapidly than BM-MSC. In our experimental conditions, AT-MSCs in normoxic condition did not display a plateau phase in the proliferation curve and, at the light microscope, AT-MSCs start growing in several layers instead of an only monolayer (data not shown). These observations might point out a lack of contact inhibition of growth in AT-MSCs. In addition, in a previous study, we described the more rapid decrease of apoptosis in AT-MSCs compared with BM-MSCs in cultures at 20% O2 using a limited number of animals (n = 2)
. The current study confirms that finding because normoxic AT-MSCs showed significantly lower proportion of apoptotic cells than normoxic BM-MSCs.
Flow cytometric immunophenotype analysis of horse MSCs revealed that the surface antigen CD90 was detectable in all MSC types
[38–41]. In addition, cross-reactivity with human antibodies has been demonstrated for the CD29 antigen in a previous report from our group and also in other studies
[42, 43]. Because in other species hypoxia does not alter the immunophenotype of MSCs with regard to CD29
 and CD90
[9, 27, 44], we attempted to characterise this phenotype in equine MSCs and to analyse the presence of these molecules in both BM-MSCs and AT-MSCs in hypoxia and normoxia. According to the literature, equine MSCs displayed the same immunophenotype for CD29 and CD90 independently of the cell source and oxygen tension.
The lack of immunoreactivity of commercial antibodies with equine MSC antigens remains a challenge in determining the immunophenotype of these cells by flow cytometry. As a supplement to this technique, RT-qPCR has been used to establish the expression profiles of various cell surface markers in equine MSCs
[39, 42]. Similar gene expression patterns were demonstrated in AT-MSCs when they were compared to BM-MSCs in their respective oxygen conditions. AT-MSCs at both oxygen tensions expressed higher levels of CD29, CD44, CD90, CD146 and CD34 transcripts respect to BM-MSCs; in contrast, only normoxic AT-MSCs expressed lower mRNA levels of CD49d compared to normoxic BM-MSCs. These results are in agreement with our previous report
. The differences in CD105 expression, with respect to our previous work, might be due to individual differences because different animals were used in the present study. Hypoxia seemed to significantly modify mRNA levels of CD49d in BM-MSCs and CD44 in AT-MSCs, which is in agreement with other studies that have described different expression profiles for CD49d and CD44 in hypoxia. The remaining surface markers analysed in this study showed similar gene expression pattern at the different oxygen conditions studied.
HIF-1α is a transcription factor that is expressed constitutively in cells, although is ubiquitinated and degraded under normoxic conditions. In our study, the gene expression of this factor was detected in normoxic and hypoxic cultures, although HIF-1α was up-regulated in cultures exposed to low oxygen tension.
The expression of specific markers characteristic of embryonic stem cells have been described before for equine BM-MSCs
 and AT-MSCs
. However, to our knowledge, this is the first work that compares the gene expression of the pluripotency markers OCT4, NANOG and SOX2 in equine AT-MSCs and BM-MSCs that were exposed to different oxygen concentrations. In our experimental conditions, equine MSCs expressed all three pluripotency markers. In general, higher expression of each marker was detected in AT-MSCs and was statistically significant for OCT4 in hypoxia and for NANOG in normoxia. The consistently higher expression of all genes in hypoxia might reflect the enhanced stemness of hypoxic equine MSCs
. These results agree with other studies that have described up-regulation of pluripotency-associated markers of hypoxic MSCs
[11, 49, 50]. To our knowledge the relationship between pluripotency markers and HIF-1α has never been investigated in MSCs. However, studies in cancer cells have revealed the expression of HIF-1α induces a gene expression increase of genes involved in stemness
; in accordance with this, the higher expression of HIF-1α observed in hypoxic cultures of equine MSCs could enhance the gene expression of the pluripotency markers. Taken together, the results might suggest that low oxygen tension helps maintain the undifferentiated stem cell phenotype.