Human umbilical cords represent a promising source of hMSCs, which can be conveniently isolated and induce low immunogenicity. In addition, as compared to other stem cell sources these present relatively few ethical issues. The hMSCs isolated from umbilical cords were positive for CD44, CD29 and CD105, and negative for CD106, CD40, CD34, CD45 and HLA-DR surface antigens, confirming their identity as MSCs . In addition, the isolated hMSCs were able to differentiate into adipogenic and osteogenic cells under specific culture conditions.
The intracellular cytoskeleton responds to external mechanical stimuli . Tapping mode AFM was used to observe the surface structures. The position of the adipogenic cell nuclei and their granular quality of the membranes is suggestive of early S phase of the cell cycle. Osteogenic-induced hMSCs appeared larger, squarer and more rigid as compared to undifferentiated cells, which is consistent with that reported by Danti et al. . In addition, they were easily distinguishable from each other as the differentiation process prevented their proliferation, which also correlates with Danti et al .
Differentiation of MSCs is also influenced by hypoxia. Qu et al.  reported that DFO increased the osteoblastic differentiation of bone morphogenetic protein-2-treated MSCs. Valorani et al.  and Ren et al.  observed that hypoxia (2 and 8%, respectively) increased the adipogenic differentiation potential of MSCs. These studies suggest that pre-culturing MSCs under hypoxic conditions prior to transplantation may enhance their efficacy.
As evidenced by clinical trials in hypoxic-ischemic diseases, MSCs-based therapy has potential value in tissue replacement and regeneration [26–28]. The effects of in vivo oxygen concentrations, which are much lower than common experimental conditions, on hMSCs were largely unknown. Therefore, hMSCs proliferation and morphology in response to hypoxia induced by DFO and CoCl2 was assessed. After treatment with DFO and CoCl2, hMSCs were more elongated, and gaps appeared between adjacent cells. Thus, DFO and CoCl2 may inhibit hMSCs growth by weakening cell-to-cell signaling. The reduction in cell-cell junctions may also mediate hMSCs migration induced by hypoxia (3% oxygen) . Further studies are necessary to determine the effects of DFO and CoCl2 treatment on the signaling pathways that govern hMSCs migration.
The influence of hypoxia on hMSCs ultrastructure was also explored by TEM. After treatment with DFO, hMSCs contained a large number of unidentifiable vacuoles that are early markers of apoptosis, which is consistent with Ren et al.  using 8% oxygen. In addition, the observed shrinkage, disintegration, and dissolution of the nucleus, along with chromatin condensation indicated early apoptosis. Consistent with the signs of apoptosis, we found that hMSCs proliferation decreased in a dose-dependent manner with increasing concentrations of CoCl2 and DFO. However, discrepant effects of physical hypoxia and hypoxia mimetics on MSCs proliferation indicate differences. For example, Ren et al.  reported that low oxygen levels (8%) promoted MSCs proliferation, whereas DFO (120 μM) and CoCl2 (100 μM) inhibited their growth. Lavrentieva et al.  reported that 1.5-5% oxygen levels increased the proliferative capacity of hMSCs. Whereas Qu et al.  found that DFO, ranging from 0 to 100 μM, inhibited cell growth in a dose-dependent manner. HIF-1α levels may differ under physically-induced hypoxia as opposed to CoCl2-induced hypoxia [12, 16] which may account for these differences; however, further analysis is required.
The effects of CoCl2 and DFO on hMSCs proliferation may be mediated by cell cycle changes. A larger percentage of DFO- and CoCl2-treated cells in the G0/G1 phase was observed, while the ratio of those in the G2/M/S phase decreased. Similar results were reported by Holzwarth et al. , who reported hMSCs accumulation in the G1 phase at 1% oxygen. Further studies are necessary to determine if the cell cycle effects of hypoxia mimetics can be recapitulated under low oxygen conditions.
There are several study limitations that warrant discussion. Firstly, the present study analyzed the effects of hypoxia mimics on hMSCs; however, the effects of physical hypoxia were not assessed. Further studies will be carried out to compare the effects of physical hypoxia and hypoxia mimetics on hMSCs morphology and growth. In addition, the effects of DFO and CoCl2 on hMSC HIF-1α levels were not analyzed. However, previous studies using the same concentrations of DFO and CoCl2 used in the present study have reported upregulation of HIF-1α expression [12, 32]. Furthermore, the present study did not analyze the influence of DFO and CoCl2 on MSCs differentiation. Although the effects of these agents on MSCs morphology is suggestive of greater self-renewal capacity as hMSCs broaden and flatten with differentiation , and differentiation was associated with changes in nuclear morphology [34, 35], determining their effects on MSCs differentiation will be the focus of future studies. Finally, the effects of hypoxia-induced morphological changes on MSCs function (e.g., cell migration, homing or immune regulatory effect) was not explored in the present study, but will be analyzed in future studies. Although the present study has its limitations, determining the ways in which MSCs respond to environments with lower than atmospheric oxygen concentrations, such as the blood, bone marrow, and cartilage, is crucial for their successful use regenerative medicine. The present study advances our understanding of the influences of hypoxia on MSCs morphology and proliferation.