Bone marrow samples were harvested from the sternum or the posterior iliac crest of healthy volunteers after administration and signature of an informed consent. In some cases, we obtained cells from the washouts of discharged bags and filters of hematopoietic stem cell transplantation as previously described (41). We defined young and old donors as subjects having less than 18 years or more than 55 years, respectively. The Table 1 displays the age distribution among the two groups. MSC were isolated using the classical adhesion method. Mononuclear cells were isolated by density gradient centrifugation (Pancoll, Palm Biotech, Aidenbach, Deutschland), washed in Hank’s buffered salt solution (HBSS, Lonza Europe, Verviers, Belgium) and seeded in Dulbecco’s modified Eagle’s medium–low glucose (DMEM-LG, Lonza) supplemented with 15% of fetal bovine serum (FBS, Sigma-Aldrich, Saint-Louis, MO, USA), 2 mM L-glutamine and 1% penicillin-streptomycin (both from Lonza). Cells were incubated at 37°C in a 5% CO2 and 95% humidified atmosphere. After 48 h, non-adherent cells were removed by washing and the medium was changed twice a week. Thereafter, cells were detached with TrypleSelect solution (Lonza) and subcultured at 10³ cells/cm² for their expansion by passage. MSC were immunophenotypically characterized according the ISCT criteria and their adipogenic, osteogenic and chondrogenic differentiation potential evaluated (9).

Briefly, MSC immunophenotype was established by flow cytometry using the following monoclonal antibodies: anti-CD73-PE (Miltenyi Biotec, Bergisch Gladbach, Germany), anti-CD90-PE (Miltenyi Biotec), anti-CD105-FITC (Ancell Corporation, Bayport, USA), anti-CD45-PC7 (BD Biosciences), anti-CD34-PE (Miltenyi), anti-CD14-PC7 (BD Biosciences), CD11b-APC (Miltenyi Biotec) and CD19-PC5 (BD Biosciences). After labeling, acquired results were analyzed by using a MACSQuant analyzer (Miltenyi Biotec). Cells were incubated for 30 min with these antibodies and after washing with PBS, the cells were fixed with 8% formaldehyde. BM-MSC were cultured in appropriate induction medium to assess their adipogenic, osteogenic and chondrogenic lineage differentiation capacities (NH media, Miltenyi Biotec). Lipid vacuole formation, mineralization (calcium deposits) and presence of proteoglycans corresponding to each lineage commitment were demonstrated by Oil Red O (Sigma-Aldrich), Alizarin Red S (Sigma-Aldrich) and Alcian Blue (Sigma-Aldrich) staining respectively.

Population doubling time (PDT) was calculated between P1 and P2 as t/n, where t is the duration of culture in days and n is the number of population doublings calculated by using the formula n = (log Nh - log Ni)/log 2, where Nh is the number of cells harvested at the time of counting at P2 and Ni is the number of cells initially plated at P1.

Senescence was evaluated by two methods: β-Galactosidase (β-Gal) staining and flow cytometry using a fluorescent probe. Ten thousand BM-MSC cells were plated in a 48 well-plate for 24 h before staining cells with the Senescence Detection Kit (BioVision, Milpitas, CA, USA). Blue cells were observed using an inverted microscope and the number of blue cells out of 100 total cells was scored. For the second method, cells were incubated with diluted Green probe (1000x) for 2 hours at 37°C without CO2 (Invitrogen CellEvent™Senescence Green detection kit). After incubation, cells were washed with PBS and analyzed by flow cytometry (MACSQuant, AlexaFluor™ 488/FITC filter set). The results were presented as the mean percentage (± SEM) of senescent cells.

Intracellular reactive oxygen species (ROS) were measured by a DCFDA fluorescent probe (Sigma-Aldrich). Briefly, young and old BM-MSC were wash with PBS twice and incubated with serum-free DMEM containing 10 µM DCFDA at 37°C for 20 min. ROS generation was determined by flow cytometry and the percentage of positive cells and the mean fluorescence intensity (MFI) in each group was evaluated. A shift to the right indicates increased ROS levels.

To estimate the number of mesenchymal clonogenic cells, a colony-forming unit fibroblast (CFU-F) assay was performed. Briefly 5000 cells were plated in a Petri dish (100 mm diameter, Greiner) with culture medium for 10 days. After May-Grünwald/Giemsa staining, colonies were defined as more than 50 fibroblastic cells and scored using an inverted microscope.

Cell cycle was evaluated by standard propidium iodide (PI) staining. Cells to be analyzed were collected, washed, permeabilized, and incubated with solution containing PI and RNAse (Coulter DNA-Prep Reagent, Hialeah, FL, USA). Tubes were placed at 4°C in the dark overnight before analysis by flow cytometry to identify the cell cycle phases. Data collection was gated using forward light scatter and side light scatter to exclude cell debris and aggregates. PI fluorescence of individual nuclei was measured using MACSQuant flow cytometer and a least 2 10⁴ cells of each sample were analyzed by FCS Express 4 Flow software.

The impact of an inflammatory environment on BM-MSC was evaluated as previously described (42). Briefly, cells were stimulated for 24h using a cocktail of pro- inflammatory cytokines: 25 ng/mL IL-1β (Miltenyi Biotec), 50 ng/mL interferon (IFN)-γ (RD Systems, Abingdon, United Kingdom), 50 ng/mL tumor necrosis factor alpha (TNF-α) (Miltenyi Biotec), and 10 ng/mL IFN-α (RD Systems).

Human monocytic THP-1 cells were maintained in culture in RPMI 1640 culture medium (Biowest, Nuaillé, France) containing 10% of heat inactivated fetal bovine serum (Sigma). THP-1 monocytes were differentiated into macrophages by 24 h incubation with 200 ng/ml phorbol 12-myristate 13-acetate (PMA, Sigma). Macrophages were polarized in M1 macrophages by incubation with 20 ng/ml of IFN-γ (R&D systems) and 100 ng/ml of LPS (Sigma, #8630). Macrophage M2 polarization was obtained by incubation with 20 ng/ml of interleukin-4 (Tebu-Bio, Boechout, Belgium) and 20 ng/ml of interleukin-13 (Tebu-Bio).

BM-MSC and THP-1-derived macrophages were co-cultured using a cell culture insert (Greiner Bio-One, Vilvoorde, Belgium) with a 0.4-μm porous membrane to separate the upper and lower chambers. The THP-1 monocytes (1 × 10⁵ cells/ml) were seeded into the lower chamber of the Transwell apparatus (12 well plates), stimulated to differentiate into macrophages by the addition of PMA during 24h. BM-MSC were placed in the upper chamber at a density of 1 × 10⁵ cells/ml for 24 h to allow their adherence to the membrane. The chambers with the BM-MSC were then placed directly on top of the 12-well plates containing the THP-1-derived macrophages and the resulting co-culture systems were incubated for 24 h. Co-culture supernatants were then collected to quantitate cytokine levels and THP-1 derived macrophages were harvested for mRNA and flow cytometry assessment of specific cell surface antigens.

Total RNA was isolated from BM-MSC or THP-1 and was extracted in a single step using TriPure Isolation Reagent (Roche Applied Science, Vilvoorde, Belgium). We performed the reverse transcription reaction with 1 mg RNA using qScript cDNA SuperMix (Quanta Biosciences). Transcripts were quantified by qRT-PCR using 20 ng of cDNA, SYBR Green PCR Master Mix (Applied Biosystems, Lennik, Belgium) and 0.32 mM forward and reverse primers. The primers were designed with Primer Express 2.0 software (Applied Biosystems) or ProbeFinder online software (Roche) and are available in Table 2 . To control variations in input RNA amounts, the GAPDH gene was used as a housekeeping gene to quantify and normalize the results. The reactions were carried out using the ABI Prism 7900 HT system (Applied Biosystems). The comparative ΔΔCt method was used for data analysis. To evaluate the fold change, data were normalized with the GAPDH gene to obtain the ΔCt and were then calibrated with the geometric mean of the GAPDH ΔCt to generate the ΔΔCt. Fold changes were then calculated as fold change=2–^ΔΔCt.

MiRNA were quantified by real-time PCR using SYBR green. Briefly, miRNA will be first polyadenylated using a poly(A) polymerase. Reverse Transcription will be performed using an oligo-dT adapter primer. The adapter primer (GCATAGACCTGAATGGCGGTA) has a unique sequence at its 5’ end which allows amplification of cDNA in real-time PCR reactions. Individual miRNA were quantified in real-time SYBR Green PCR with the desired customer miRNA complementary to the target miRNA (miR-193b: GCCCTCAAAGTCCCGCTAA, miR-21: GCAGCTTATCAGACTGATGTTGAAA) and with a specific primer to the unique sequence of the oligo-dT adapter primer. The custom miRNA provides maximum sensitivity and specificity in real-time PCR amplification and quantification of miRNA. Cell miRNA will be normalized using RNU44 (CCTGGATGATGATAGCAAATGC) and RNU48 (CTCTGAGTGTGTCGCTGATGC), largely described as stable endogenous controls (43).

Secreted protein levels of CCL2, TNF-α and IL-6 were quantitated in culture supernatants of THP-1 derived macrophages co-cultured or not with BM-MSC, in the presence or not of M1 activation using specific cytokine ELISA kits (Quantikine, RD Systems) as per manufacturer’s instructions. The sensitivity of CCL2, TNF-α and IL-6 was respectively 10, 6.23 and 2.95 pg/ml. In some experiments, culture supernatants were analyzed for cytokine levels by fluorescent flow cytometry using the Bio-Plex Pro Human Cytokine Panel, 48-plex (Bio-Rad, USA). The assay was performed using a Bio-Plex machine (Bio-Plex 200 System, Bio-Rad) following the manufacturer’s protocol, and the data were analyzed with the Bio-Plex manager software version 6.0.

Data are presented as mean ± standard error of the mean (SEM). Comparison between data was evaluated with the unpaired Mann-Whitney U test. In some cases, the Wilcoxon matched pair test (two-tailed) was used and differences were significant for p< 0.05. All analyses were performed with GraphPad Prism version 5.00 for Windows (GraphPad Software, www.graphpad.com).

A flow cytometry assay conducted after MSC isolation and expansion showed the co-expression of CD73, CD90 and CD105 in more than 98% of analyzed cells, while the expression of hematopoietic markers CD45, CD14, CD34, CD11b, CD19 and HLA-DR was inferior to 2%, compatible with International Society for Cellular Therapy (ISCT) criteria for MSC identification ( Figure 1A ). Both yBM- and oBM-MSC were capable of adipogenic, osteogenic and chondrogenic differentiation ( Figure 1B ). β-Gal staining showed a significantly increased percentage of senescent cells for oBM-MSC compared to yBM-MSC (39.6 ± 9.4 vs 7 ± 2.4, n=5; p=0.016) ( Figures 2A, B ). These results were confirmed by flow cytometry (47% ± 14% vs 10% ± 6%, n=5; p=0.046) ( Figures 2C, D ). DCFDA staining showed that 20 ± 5% of yBM-MSC and 60 ± 3% of oBM-MSC were ROS positive (p<0.0001) ( Figures 2E, F ) with an increased level of intracellular ROS for oBM-MSC (MFI of 38,85 ± 3.17, n= 20) compared with yBM-MSC (MFI of 28 ± 1.7, n= 8, p= 0.027) ( Figure 2G ). Concerning morphology, oBM-MSC showed significantly increased size (81.38 ± 2.51 vs 66.62 ± 2.38; p=0.005) and granularity (46.35 ± 3.9 vs 33 ± 1.7; p=0.007) compared to yBM-MSC ( Figures 3A, B ). Mean population doubling time was significantly longer for oBM-MSC compared to yBM-MSC (4.68 ± 0.5 vs 3.04 ± 0.28 days; p=0.001) ( Figure 3C ). We observed a reduced clonogenic capacity of oBM-MSC (n= 8) compared to yBM-MSC (n= 16) (36 ± 16.5 colonies vs 149 ± 37 colonies for 10⁶ cells; p=0.0001) ( Figure 3D ). The analysis of the expression of cell cycle regulators showed significantly higher levels in oBM-MSC for p16 (41.78 ± 3.34 vs 31.67 ± 2.71; p=0.03) and p21 (287.7 ± 32.03 vs 166.4 ± 16.46; p=0.01), while no difference was shown for p53 (34.64 ± 4.32 vs 32.11 ± 3.4) ( Figure 4A ). The evaluation of cell cycle phase showed a higher number of oBM-MSC in the S phase (18.72 ± 1.71 vs 11.91 ± 1.38; p=0.02) and a lower number in the mitotic phase (3.05 ± 0.28 vs 5.16 ± 0.69; p=0.03) ( Figure 4B ). At mRNA and protein levels, we observed a higher expression and secretion of IL-6 and IL-8 by oBM-MSC compared to yBM-MSC. The relative IL-6 and IL-8 gene expression was respectively 477 ± 34.6 vs 846 ± 119 (n= 22; p=0.0267) and 224 ± 58 vs 369 ± 174 (n= 15; p=0.1). By ELISA, we confirmed higher IL-6 and IL-8 production by oBM-MSC compared to yBM-MSC (respectively 130 ± 101 vs 989 ± 394 pg/ml; p=0.0076 for IL-6 (n=9) and 224 ± 58 vs 369 ± 174; p=0.0016 for IL-8 (n= 8) ( Figure 4C ). These results were consistent with the presence of SASP in oBM-MSC.

We evaluated the variation in mRNA expression for genes implicated in immunomodulation, before and after exposing MSC to a proinflammatory priming, as described in “Material and methods” section. The cocktail was chosen by our group and others because these cytokines are those mostly present at inflammatory sites. Before exposing yBM-MSC and oBM-MSC to stimulating cytokines, we did not report any difference in the two groups in terms of gene expression for the analyzed markers, except for IL-6 and IL-8 (as shown in the previous section). Inflammation modulates several biological features of MSC (secretome, multilineage potential, immunomodulation) and allows to evaluate the MSC efficiency. Indeed, after the priming, we reported a significant up-regulation in oBM-MSC, when compared to yBM-MSC, for galectin-1 (GAL-1; 22915 ± 2558 vs 14511 ± 1838; p=0.03), interleukin-6 (IL-6; 28622 ± 4203 vs 19564 ± 2938; p=0.004), interleukin-8 (IL-8; 29574 ± 5334 vs 17696 ± 2549; p=0.04), transforming growth factor beta (TGF-β; 739 ± 91 vs 424 ± 24; p=0.003) and tumor necrosis factor stimulating gene-6 (TSG-6; 39239 ± 8658 vs 12221 ± 1606; p=0.002). For other genes, such as cyclooxygenase-1 (COX-1; 6 ± 1.2 vs 4.3 ± 0.69; p=0.569) and -2 (COX-2; 1619 ± 249 vs 1881 ± 207; p=0.470), hepatocyte growth factor (HGF; 176 ± 48 vs 110 ± 15; p=0.301), indoleamine 2,3-dioxygenase (IDO1; 2845 ± 306 vs 2006 ± 201; p=0.176) and leukemia inhibitory factor (LIF; 73 ± 11 vs 55 ± 5.7; p=0.233), no significant differences were observed ( Figure 5 ).

Through a bio-plex multiplex analysis we identified, in macrophage culture medium, several cytokines with significantly increased concentration after polarization to M1 status ( Figure 6A ). Co-culture of MSC with M1 polarized THP-1 derived macrophages was associated to a decrease of M1 cytokine production (n=4) ( Figure 6B ). We compared the inhibitory activity exerted by yBM- and oBM-MSC through the evaluation of the concentration of chemokine ligand 2 (CCL2), tumor necrosis factor-α (TNF-α), human interferon-inducible protein 10 (IP10), IL-6, monokine induced by interferon gamma (MIG): we found a significantly higher anti-M1 activity, determining a reduction for these cytokines, after adding in co-culture yBM-MSC ( Figure 6C ).

We compared the inhibitory effect towards macrophage polarization exerted by yBM-MSC and oBM-MSC, evaluating the mRNA expression of three target cytokines involved in M1 response: CCL-2, TNF-α and IL-6. The relative gene expression of these cytokines increased after M1 polarization of THP-1 macrophages: 1941 ± 290 vs 63296 ± 9051 for CCL2 (p<0.0001), 14 ± 2 vs 141 ± 10.5 (p<0.0001) for TNF-α and 0.8 ± 0.25 vs 39.6 ± 3.2 for IL-6 (p<0.0001) ( Figure 7A ) corresponding to an increase of 59.08 ± 29.42 fold, 10 ± 2.7 fold and 48.45 ± 17.26 fold respectively for CCL2, TNF-α and IL-6. When comparing the impact on the mRNA expression, we reported a significantly reduced expression after co-culture with yBM-MSC if compared to oBM-MSC for CCL-2 (26809 ± 7135 vs 44146 ± 20886, p= 0.04) and TNF-α (50 ± 3.6 vs 86 ± 12, p= 0.0019). yBM-MSC inhibited CCL2 expression of 54 ± 6.8% and 68 ± 3% for TNF-α while the inhibitory activity by oBM-MSC only reached 34 ± 6% and 39 ± 7%. For IL-6, we reported a reduction for yBM-MSC (20.09 ± 1.6 in comparison to 39.6 ± 3.2 for M1 condition corresponding to an inhibition of 52 ± 4%) and an augmentation after oBM-MSC co-culture (52.88 ± 17.5 vs 39.6 ± 3.2; p= 0.014).

We also investigated and compared variations in medium concentration for these three cytokines through an ELISA assay. Concentrations of CCL-2 (27974 ± 2435 pg/ml vs 1889 ± 391 pg/ml; p< 0.0001) and TNF-α (1603 ± 142 pg/ml vs 519 ± 101 pg/ml; p< 0.0001) were significantly higher for M1 macrophages if compared to M0, while for IL-6 we did not find a significant difference (2628 ± 728 vs 812 ± 700 pg/ml) ( Figure 7B ). Concerning the impact of MSC co-culture on these M1 markers, yBM-MSC were associated to a higher percentage reduction in CCL-2 (25% ± 9% vs 20% ± 5.7%, p= 0.01) and TNF-α (65% ± 3.35% vs 40.59% ± 6.4%, p=0.01) levels. For IL-6 we reported increased values after co-culture with MSC, with non-significantly higher levels for oBM-MSC compared to yBM-MSC (1082% ± 148% vs 828% ± 89%; p=0.393).

We performed a miRNA expression profile analysis comparing yBM- and oBM-MSC, by using selected miRNA previously identified with a potential role in the immunomodulating abilities exerted by MSC and particularly in macrophage polarization ( Figure 8A ). After this preliminary phase, we performed further investigations on two miRNA, miR21 and miR193b-3p: we found a significantly increased relative expression after inflammatory priming for both miR21 (1310 ± 495.2 vs 554 ± 114; p=0.001) and miR193b-3p (31844 ± 7967 vs 14748 ± 3044; p=0.001) in yBM-MSC, while no significant differences were noted for oBM-MSC (miR21: 723.1 ± 56.56 vs 622.5 ± 57.82; p=0.577; miR193b-3p: 16376 ± 1960 vs 15431 ± 2369; p=0.243) ( Figure 8B ).