First, we looked at the differentiation potential of BMSC. Cells either differentiated into adipocytes or osteocytes when treated with corresponding differentiation media (Figure 1A,B), or formed spheroids of cartilage-like nature [22] upon high seeding density in chondrocitic differentiation medium (Figure 1C). Differentiation towards these lineages was confirmed by specific staining (Figure 1C–E). Upon treatment with neuronal differentiation medium, BMSC showed neuronal-like morphology (Figure 1F). Analysis of the CD profile showed that cell population at day 7 of culture contained cells expressing the hematopoietic/endothelial markers CD31, CD45, CD11b (Figure 1G). These cells became less frequent in culture with time. We also identified cells expressing CD106, CD34, and Sca-1 that are reported to be present on mMSC (Figure 1G) [7,23]. These results indicate that our BMSC preparations contained a population of MSC that was confirmed by adherence to plastic, multilineage differentiation and a set of CDs previously reported to be expressed by mMSC.

By taking the advantage of the fact that only MSC give rise to colonies when cultured at low seeding density on plastic [24], we estimated the amount of MSC in BMSC cultures by calculating the number of colonies growing in CFU-A. The results from CFU-A revealed the presence of 4.8 (±0.7) MSC per 10⁵ plated nucleated bone marrow cells (0.005%) (Figure 1H).

We checked if hypoxia had any effect on BMSC growth. Upon isolation from bone marrow, only few cells attached to the plastic. There was no visible increase in the amount of cells up to day 4 when rapid clonal expansion took place. At day 7, cultures reached confluence. The total number of cells recovered in 3% O₂ at this point was reduced by almost half compared to 21% O₂ (Figure 2A). The analysis of the cellular growth rate revealed that 3% O₂ prolonged the lag phase of cultured cells. From day 4, the growth rate of BMSC cultured in 3% O₂ continuously increased and on day 7 cells proliferated even faster in 3% O₂ than in 21% O₂ (Figure 2A,B). We confirmed this trend by cell cycle analysis. On day 4, the fraction of cells in S-G2/M phases was lower in 3% O₂ when compared to 21% O₂, but it significantly increased by day 7 and even exceeded the equivalent fraction in 21% O₂ (Figure 2C). Switching the cells after growing four days at 3% O₂ to three more days at 21% O₂ reduced their fraction in S-G2/M. On the contrary, change from 21% to 3% O₂ augmented the amount of cells in S-G2/M (Figure 2D).

It is known that low O₂ induces stabilization of the transcription factor Hif-1α, which regulates cellular response to hypoxia [25]. Even though 3% O₂ was not sufficient to trigger a measurable stabilization of the Hif-1α protein in BMSC cultures in our studies, we found that the expression of the Hif-1α responsive gene Vegfr1 increased upon exposure of cells to 3% O₂ (Figure 2E) as previously described [26].

These data suggest that 3% O₂ effectively triggers hypoxic response and, after a period of adaptation, stimulates the growth of BMSC.

CFU-A was performed on total BMSC to evaluate the effect of O₂ tension on the number of MSC. Three percent of O₂ augmented the amount of colony-forming cells by 1.6 fold in comparison with 21% O₂ (Figure 3A,B). In addition, the size of colonies and the average number of cells per colony increased in 3% O₂ (Figure 3B,C).

Obtained results suggest that 3% O₂ enhances the yield and proliferation of MSC contributing to the increased growth rate of the total BMSC population.

Next, we tested the differentiation potential of BMSC in 3% and 21% O₂. In agreement with previously reported data, few adipocyte-like cells, with accumulated lipid droplets and characteristic round adipocyte morphology, were detected when BMSC were both propagated and differentiated in 21% O₂[8] (Figure 4A). The frequency of adipocytes remarkably increased if both these steps were performed in 3% O₂ (Figure 4A).

Osteocytic differentiation was also more efficient when BMSC were propagated and differentiated in 3% O₂. BMSC cultured in 3% O₂ produced higher amount of mineralized extracellular matrix in comparison with cells cultured in 21% O₂ (Figure 4B).

We observed the similar pattern of differentiation in CFU-A, where 3% O₂ augmented both the amount of colonies undergoing adipocytic and osteocytic differentiation and the efficiency of their differentiation (Figure 4C,D).

Three percent of O₂ may affect the outcome of BMSC differentiation by regulating intrinsic differentiation processes or by selecting cells capable of differentiation (stem cells). To address this issue, BMSC initially amplified at 3% O₂ were induced to differentiate at 21% O₂. This culture scheme further promoted the outcome of BMSC differentiation in comparison with cells cultured at 3% O₂ only (Figure 4E). In contrast, when cells were amplified in 21% O₂ and switched for differentiation to 3% O₂, the differentiation yield was drastically reduced, almost abolished. We confirmed this pattern of BMSC adipocytic differentiation by qPCR, which showed the highest expression of adipocytic differentiation markers Pparγ2 [27], aP2 [28] and adipoQ [29] in cells amplified in 3% O₂ and switched for the differentiation to 21% O₂ (Figure 4F).

Thus 3% O₂ inhibits the differentiation process per se.

Since cells amplified in 3% O₂ subsequently differentiated more efficiently, we hypothesized that 3% O₂ selects for stem cells. Therefore, we checked whether 3% O₂ affected the stemness of BMSC cultures.

Oct-4 and Rex-1 are pluripotent stem cell markers, which were initially described in embryonic stem cells [30,31]. Induction of their expression by low O₂ is observed in various types of cells [32] and is associated with increased stemness in hMSC [10,13].

Our qPCR results revealed that 3% O₂ enhanced expression of Rex-1 and Oct-4 in BMSC (Figure 5A).

Adult stem cells divide occasionally [33] to renew themselves and generate progeny committed to differentiation [34]. Thus, the amount of rarely dividing cells correlates with the amount of stem cells. We estimated the amount of quiescent BMSC under different O₂ conditions by staining the culture with the vital fluorescent dye PKH-26 that stably integrates into the plasma membrane. When labeled cells divide, daughter cells receive half of the dye, hence PKH-26 signal decreases upon proliferation. In our experiments, both the number of PKH-26 positive cells and their fluorescence intensity were increased in BMSC propagated in 3% O₂ in comparison with control cells cultured in 21% O₂ (Figure 5B–E).

To validate the role of ROS in O₂’s effect on the stemness and differentiation, we tested differentiation of BMSC treated with antioxidants such as lipoic acid or N-Acetylcysteine (NAC), or differentiation of BMSC derived from p66^Shc-deficient mice characterized by reduced ROS levels ([35] and Figure S1). Results revealed that reducing ROS levels by chemicals or by genetic means had no effect on the ability of BMSC to differentiate under various O₂ levels (Figure 6A,B).

Next, we checked if low O₂ tension regulated BMSC differentiation in a p53 dependent manner. To this aim, we investigated differentiation of BMSC from p53 deficient mice in 3% and 21% O₂. P53^−/− BMSC showed significantly higher differentiation ability. Nevertheless, the pattern of BMSC differentiation under various O₂ tensions remained unchanged (Figure 6C–E). It appears that the O₂ availability modulates BMSC differentiation independently from ROS and p53.

If not stated otherwise, all reagents were purchased from Sigma (St. Louis, MO, USA).

Wild type, p53^−/− or p66^Shc−/− 8–10 week-old male mice in C57Bl/6 background were sacrificed by cervical dislocation, and tibia and femurs were collected. Bone marrow was flushed out with growth medium (DMEM-HG supplemented with 20% FBS, 2 mM glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 20 mM HEPES pH 7.4) and passed through 27 G needle. The obtained suspension was centrifuged at 300× g for 10 min. Pellet was resuspended in growth medium and cells were seeded at a density of 2 × 10⁵ nucleated cells/cm² (day 0) and left undisturbed for 72 h. Afterwards, the medium was changed every 48 h in growing cultures and twice per week in differentiating cultures. At day 7, cells were split at a density of 10⁶/cm² and cultured in DMEM-HG supplemented with 10% FBS, 2 mM glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin. The next day, differentiation was induced. Cultures were kept at 37 °C in a gas mixture containing 3% or 21% O₂, 5% CO₂ and balanced with N₂.

All the in vivo experiments were performed in accordance with Italian laws and regulations.

For adipocytic differentiation cells were cultured in adipocytic differentiation medium (DMEM-HG and Ham F12 mixed 3:2, supplemented with 10% horse serum, 2 mM glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 100 nM dexamethasone, 5 μg/mL insulin, and 10 mM nicotinamide) for 14 days. For certain experiments the differentiation medium was supplemented with 250 μM lipoic acid or 500 μM N-acetylcystein (NAC). Oil Red O staining was performed as described by others [43].

For osteogenic differentiation, cells were cultured in osteogenic differentiation medium (DMEM-HG supplemented with 10% FBS, 2 mM glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 50 μM ascorbic acid, 10 mM beta-glycerol-3-phosphate, and 100 nM dexamethasone) for 14 days. For Alizarin Red S staining, cells were fixed in formalin, washed with water, incubated for 5 min with 2% Alizarin Red solution, pH 4.1 and washed twice with water.

For chondrogenic differentiation a method of spheroids culture was used [22]. For Alcian Blue staining, obtained spheroids were fixed in 10% formalin and overloaded with 1% Alcian Blue-8GX solution overnight.

For neuronal differentiation, cells were split every other day, starting from passage 1 in growth medium containing 10 ng/mL bFGF. At passage 5, neuronal differentiation medium (alpha-MEM, supplemented with 0.1% of FBS, 1% DMSO, 10 μM forskolin, 1 μM hydrocortisone, 5 μg/mL insulin, 10 ng/mL bFGF) was applied for 12–24 h.

For CFU-A cells isolated from bone marrow (passage 0) were seeded at a density of 50 × 10³ nucleated cells/cm². At day 7, cells were washed with PBS, fixed in 10% formalin, incubated with Giemsa dye for 10 min and washed three times with water.

Differentiation of cultures seeded in CFU-A was induced at day 8, as described above, but without split of the cells.

Staining was performed with PKH-26 Red Fluorescent Cell Linker Mini Kit. At day 6 cells were trypsinized and collected by centrifugation. Pellet was resuspended in Diluent C, mixed with 4 × 10⁻⁶ M PKH-26 solution and incubated for 2 min. The reaction was terminated by the addition of an equal FBS volume for 1 min. Next, an equal volume of complete medium was added and the suspension was centrifuged. Afterwards cells were washed three times with PBS at 400× g for 5 min and seeded. Fluorescence of cultures was observed after 48 h with a fluorescent microscope (Olympus BX 61) or analyzed by FACS Callibur Cytometry system.

Total RNA extraction was performed using the RNeasy isolation kit (Qiagen, Hilden, Germany).

RNAs were reverse transcribed using SuperScript II reverse transcriptase and random primers (Life Technologies, Carlsbad, CA, USA). Obtained cDNA was used for determination of relative levels of specific mRNA with a 5′ nuclease assay (TaqMan) chemistry system. Gapdh expression was used for normalization. QPCR was performed on an ABI 7900HT sequence detection system. Primer sequences are listed in Table S1.

For flow cytometry analysis, labeling of cells was performed with the following antibodies: fluorescein isothiocyanate-conjugated anti-CD31, anti-CD34, anti-CD106 (eBioscience); phycoerythrin-conjugated anti-Sca-1, anti-CD45 (BD Pharmingen), anti-CD44 (eBioscience), and biotin-conjugated anti-CD11b (BD Pharmingen), all according to manufacturer’s instructions.

For cell cycle analysis, cells were resuspended in ice-cold 70% ethanol/PBS, incubated for 30 min on ice, washed in 1% BSA/PBS, and incubated for 1 h in 250 μg/mL RNase/50 μg/mL propidium iodide solution and analyzed by using Cell Quest software (Beckman Ltd., Brea, CA, USA).

Cells were stained with 70 μM DCFDA and analyzed by FACS as described [19].

Experiments were repeated a minimum of three times, with the exception of CD profile analysis, which was performed twice. Data are presented as means ± standard deviation and were analyzed by Student’s t test. Differences between means were assessed by one-way analysis of variance. The minimum level of significance was set at p < 0.05.