Background
More than 40 years ago Alexander Friedenstein described cells that could be isolated from the bone marrow (BM) and cultured as fibroblastoid colony forming cells
ex vivo[
1,
2]. Depicted as mesenchymal stromal cells (MSCs), these cells were shown to support hematopoiesis [
3] and osteogenesis [
4]. Since then, it has become clear that BM-MSCs offer promising therapeutic potential in regenerative medicine and immunotherapy [
5‐
7]. The most commonly applied technology in research and clinic to isolate and culture BM-MSCs from fresh BM is the removal of non-adherent BM cells and subsequent expansion of the adherent BM-MSCs (BM-MSC preparations) [
6,
8‐
10]. Studies reporting on variations in the growth kinetics and gene expression of BM-MSC preparations obtained from different donors [
11], variable properties of BM-MSC clones [
12], and surface markers identifying BM-MSC subpopulations, such as CD173, CD271, GD2, stage-specific embryonic antigen (SSEA)-4, or mesenchymal stem cell antigen (MSCA)-1 [
13‐
15], highlighted the heterogeneity of MSC preparations. Moreover, their mechanisms of action are still not fully elucidated. Recently, MSC therapies have come under criticism [
16] as, despite decades of research, relevant translational questions of MSC biology and function remain unanswered. Many studies reported on the osteogenic, adipogenic and chondrogenic differentiation potential and the immunomodulatory properties of BM-MSCs [
15,
17,
18]; however, controversies prevail about their myogenic or even non-mesodermal differentiation capacity [
10,
19,
20].
In vivo efficacy paired with poor survival and homing rate to the damaged tissue points toward mechanisms that most presumably are mediated by factors secreted by BM-MSCs [
21,
22]. Recently, more straightforward studies reported on gene expression profiling and phenotype of freshly sorted CD271
+ cells from the BM, and some transcripts appeared to be related to the donor age [
23,
24]. However, the presence or absence of one single marker as like as CD271 does not sufficiently define all BM-MSC subpopulations within human BM-MSC preparations. Therefore, clarification of how the phenotypes defining BM-MSC subpopulations, as well as age and gender of donors, might affect functional properties of BM-MSCs would mark a significant step forward in our understanding of BM-MSC biology. However, different isolation/expansion technologies/reagents [
15] and donor-to-donor variations [
11] result in variable fractions of MSC subpopulations per donor/preparation and hamper reliable statistical analyses. Therefore, we obtained BM-MSC preparations from 53 donors of both genders. These multiple BM-MSC preparations were isolated using the most commonly applied technology in research and the clinic [
6,
8‐
10], that is, removal of non-adherent BM cells and expansion of the adherent mononuclear BM cells. The BM-MSCs were cultured under identical conditions and analyzed at early passage for phenotype, proliferation capacity, cell size, clonogenic potential, differentiation potential, immunomodulatory potential, secretion of trophic factors, gene expression profile and telomerase activity. Hereby, donor-to-donor variations and variations within the BM-MSC preparations could be identified; however, the great number of BM-MSC preparations allowed statistically robust correlation analyses of phenotype, donor gender and age to functional properties of BM-MSCs.
Methods
Isolation and culture of human BM-MSCs
Human BM-MSCs were isolated and cultured as described previously [
14]. After written informed consent and approval of the ethical committee of the University Hospital Tübingen, Germany, BM from patients without metabolic or neoplastic diseases was obtained during orthopedic operations. BM mononuclear cells were isolated by density gradient centrifugation on Lymphoflot (Biotest, Dreieich, Germany), washed twice with PBS (Lonza, Walkersville, MD, USA), counted and seeded at a density of 1 × 10
5 cells per cm
2 in standard culture medium (SCM), composed of α-MEM (Lonza), 1% penicillin-streptomycin (Lonza) and 10% pooled human blood group AB serum (PHS) (ZKT Tübingen, Germany), to tissue culture flasks (Nunc, Roskilde, Denmark). The average concentrations of sex hormones in the PHS (obtained only from male donors) were in the normal ranges for male individuals above 15 years (testosterone: 16.18 nmol/l; estradiol (eE2): 106.8 pmol/l; estrone (E1): 169.4 pmol/l) and the average bFGF concentration was 75.12 pg/ml. The resulting passage (P)0 cultures were kept under standard culture conditions at 37°C in humidified atmosphere with 5% CO
2. Non-adherent cells were removed after 24 hours and the adherent cells were cultured in SCM. SCM was changed twice a week until cells reached subconfluency, defined as 90% surface coverage by cells corresponding to 15,000 to 20,000 cells per cm
2. At this point, the BM-MSCs were detached using Trypsin-EDTA (Lonza), counted using a CASY® cell counter (Roche, Basel, Switzerland) and plated to fresh tissue culture flasks for the next passage (P1) at a density of 1,000 cells per cm
2.
Flow cytometry
Flow cytometry analysis of all BM-MSC preparations was performed at the end (subconfluency) of P1 with a FACScan instrument (BD Biosciences, Franklin Lakes, NJ, USA) using BD CellQuest Pro software and the following (secondary) PE-labeled antibodies: anti-CD10, -CD14, -CD19, -CD29, -CD31, -CD34, -CD43, -CD44, -CD45, -CD56, -CD59, -CD71, -CD73, -CD80, -CD86, -CD90, -CD105, -CD106, -CD117, -CD119, -CD130, -CD140a, -CD140b, -CD146, -CD166, -CD273, -CD274, -GD2, SSEA-1, -SSEA-4 and -HLA class I (BD Biosciences); -CD93, -Galectin 1 (R&D Systems, Minneapolis, MN, USA); -CD133 and -CD271 (Miltenyi Biotec, Bergisch Gladbach, Germany); -CD243 (Chemicon (Millipore Corporation), Billerica, MA, USA); -CD173 (AbD Serotec, Puchheim, Germany) and –MSCA-1 (BioLegend, San Diego, CA, USA). PE-conjugated or non-labeled IgG1, -IgG2a, IgG3 and -IgM antibodies (BD Bioscience) were used as isotype matched controls. Secondary antibody was a polyclonal PE-conjugated goat anti-mouse Ig (BD Bioscience). Dead cells were excluded by uptake of 7-Aminoactinomycin D (for gating strategy see Additional file
1: Figure S4, for density plots see Additional file
2: Figure S5). Analysis of percentage of antigen positive cells and fluorescence intensity was performed using FlowJo-7.2.5 software (Tree Star, Ashland, OR, USA). For compensation of unspecific antibody binding, the positivity of the respective matched isotype control was subtracted from all samples.
Analysis of cell size and PDT
Cell count and analysis of cell size was performed at the end of each passage using a CASY cell counter (Roche). Population doubling time (PDT) during P1 was calculated by the equation PDT = (culture time*ln2)/ln(cell numberharvested/cell numberseeded). Seeding density was kept constant at 1,000 cells per cm2.
For assessment of colony formation capacity, subconfluent primary BM-MSCs (P0) were trypsinized, counted and seeded at a density of 100, 200 and 500 cells per well (for two MSC populations additional data points with 1,000 cells per well were acquired) in six-well plates at P1 (Nunc, Wiesbaden, Germany). To address possible effects of seeding density on colony formation each MSC preparation was seeded at the aforementioned densities. Cells were cultured during 10 days in MesenCult™ Proliferation Kit (human) medium (Stem Cell Technologies, Vancouver, BC, Canada), then fixed and stained with crystal violet containing 4% formalin. Colonies containing >50 cells were counted microscopically. The number of colonies per 100 seeded cells (percentage colony formation) was calculated for each seeding density for each MSC preparation. These percentages were averaged for each MSC preparation (two experiments for each BM-MSC preparation) and used as one CFU-F data point for the respective MSC preparation for statistical analysis.
In vitrodifferentiation assays
Functional characterization of BM-MSCs included induction of adipogenic, osteogenic and chondrogenic differentiation
in vitro as described previously [
14]. Briefly, for adipogenic and osteogenic differentiation, BM-MSCs were seeded at a density of 1,000 cells per cm
2 at P1 and kept under standard culture conditions until reaching subconfluency. Subsequently, either adipogenic differentiation was induced using the hMSC Adipogenic BulletKit (PT-3004, Lonza), or osteogenic differentiation was induced using “osteogenic medium” composed of SCM with 10
-8 M dexamethasone, 0.2 mM ascorbic acid and 10 mM β-glycerolphosphate (Sigma). After three weeks under differentiation conditions cells were processed for RNA isolation or for lineage specific staining: lipid vacuoles in adipogenic cultures were stained with Oil Red O and calcium deposits of osteogenic cultures with Alizarin Red S, respectively.
For chondrogenic differentiation, 2.5 × 105 BM-MSCs at P1 were kept in micromass pellet cultures for subsequent staining or in monolayer cultures for RNA isolation respectively. Differentiation was induced using the hMSC Chondrogenic Differentiation BulletKit (PT-3003, Lonza), supplemented with TGF-β 3 (PT-4124, Lonza) as a growth factor. After four weeks of differentiation, cells were processed for RNA isolation or frozen sections of fixed pellets were stained with Safranin O.
Quantitative RT-PCR
For quantitative analysis of lineage specific mRNA indicative for adipogenic, osteogenic and chondrogenic differentiation potential, as well as for expression analysis of octamer-binding transcription factor (Oct4), Nanog, PR domain containing (Prdm)14, sex determining region Y-box (SOX)2, indoleamine 2,3-dioxygenase (IDO)1 and IDO2 mRNA from differentiated and undifferentiated BM-MSCs was isolated and reversely transcribed to cDNA. To compare Oct4, Nanog, Prdm14 and SOX2 mRNA expression in BM-MSCs to pluripotent stem cells, mRNA was isolated from the human pluripotent teratocarcinoma cell line NCCIT (ATCC-CRL-2073, Ch# 5097030, ATCC, Manassas, VA, USA), and from the human pluripotent embryonic stem cell (hESC) line HUES9 (generously provided by the Harvard Stem Cell Institute, Cambridge, MA, USA). Quantitative PCR was performed on resulting cDNA and gene expression was normalized to the expression of the housekeeping gene GAPDH for each sample. Gene induction for differentiation markers was calculated by normalization of the gene expression of differentiated cultures to undifferentiated controls.
RNA isolation
RNA from adipogenic, osteogenic and chondrogenic differentiated MSCs, as well as from undifferentiated BM-MSCs (controls) at P1 was isolated using peqGOLD TriFast according to the manufacturer’s instructions.
RNA for analysis of Oct4, Nanog, Prdm14, SOX2, IDO1 and IDO2 expression was extracted using the RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Remaining genomic DNA was digested using the RNase-Free DNase Set (Qiagen).
RNA concentration was assessed using a NanoDrop photometer (Thermo Scientific, Wilmington, DE, USA. RNA was stored at −80°C for up to three months.
Reverse transcription
For analysis by quantitative PCR, 250 μg RNA from each sample was reversely transcribed using the Transcriptor First Strand cDNA Synthesis Kit (Roche) according to the manufacturer’s instructions. Resulting cDNA was stored at −20°C for up to six months.
Quantitative PCR
For the adipogenic differentiation markers LPL and PPAR-γ, the osteogenic markers AP and OPN, and the chondrogenic markers SOX9 and COLL2 as well as for GAPDH as housekeeping gene, PCR analysis of cDNA obtained from differentiated and undifferentiated BM-MSC cultures was performed using ready-to-use amplification primer mixes for RT-PCR (search-LC, Heidelberg, Germany) in combination with LightCycler FastStart DNA Master SYBR Green I (Roche) and the LightCycler Instrument (Roche), according to the manufacturer’s instructions. Quantitative PCR assays for Oct4, Nanog, Prdm14 and SOX2 expression of undifferentiated BM-MSCs were performed in the same way.
For the PCR targets IDO1 and IDO2, primers were ordered from Sigma Aldrich (St. Louis, MO, USA) and used in combination with the QuantiTect SYBR Green PCR Kit (Qiagen). Primer sequences were CGGGACACTTTGCTAAAGGCGCT and GGGGGTTGCCTTTCCAGCCAG for IDO1 and CCTGCAGAGGTCCTGCCAAGGAA and ATGCAGGCTCTCTCCCCCAGG for IDO2 cDNA. The primer annealing step was performed at 60°C. PCR specificity for IDO1 and IDO2 was confirmed by product sequencing (4base lab, Reutlingen, Germany).
PCR results were analyzed by normalizing the expression of each target gene to the expression of the housekeeping gene GAPDH in each sample. Gene induction for differentiation markers was calculated by normalization of the gene expression of differentiated cultures to undifferentiated controls.
T cell proliferation assays
The immunosuppressive properties of BM-MSCs were analyzed in co-culture assays with activated allogeneic T cells. BM-MSCs at P2 were mitotically inactivated by treatment with 40 μg/ml mitomycin C for 30 minutes at 37°C in serum free medium. Thereafter, BM-MSCs were washed three times in SCM and seeded into 96-well plates (Nunc, Roskilde, Denmark) in triplicates at a density of 2 × 104 cells per well. After 24 hours, 1 × 105 PBMNCs, freshly isolated by density-gradient centrifugation from heparinized blood of healthy donors and normalized to the lymphocyte number, were added to the appropriate wells of BM-MSC cultures. T cells were stimulated by addition of 10 μg/ml Phytohaemagglutinin-L (Sigma) and cultured for 72 hours under standard culture conditions. During the last six hours, 100 μM Bromodeoxyuridine (BrdU) labeling solution from the Cell Proliferation ELISA, BrdU (colorimetric) Kit (Roche) was added to the (co-) cultures and cell proliferation was assessed by the subsequent ELISA, using the anti-BrdU antibody provided in the kit according to the manufacturer’s instructions. Photometric light absorption was measured in a plate reader and absorbance values were averaged and normalized to the PBMNC only control of the respective blood donor. Supernatants of T cell proliferation assays were collected and stored at −80°C for subsequent analysis of cytokine production.
Multiplex analysis of cytokine production
Cytokine production of BM-MSCs alone and of BM-MSC-PBMNC co-cultures was analyzed in supernatants of T cell proliferation assays. Samples of 25 μl cell culture supernatant were analyzed in triplicates using human cytokine/chemokine multiplex kits (Millipore) according to the manufacturer’s instructions. The Kits were composed of beads for detection of IL-2, IL-4, IL-6, IL-8, IL-10, IFN-γ and TNF-α. Analysis was performed using a Luminex® 200 instrument (Luminex Corporation, Austin, TX, USA).
Analysis of trophic factor secretion
For evaluation of trophic factor secretion, BM-MSCs MSCs from 11 unrelated donors (5 female, 6 male; age: 50.9 y ± 13.3 y) (P2) were seeded to six-well plates at a subconfluent density of 20,000/cm2. The cells were kept under standard culture conditions and the medium was renewed after 24 hours to a precise volume of 3.0 ml per well. After 72 additional hours, culture supernatants were transferred to Eppendorf cups and frozen at −80°C for subsequent cytokine analysis.
Samples were sent to Multimetrix (Heidelberg, Germany) for multiplex analysis of HGF, LIF, VEGF-A, bFGF and NGFB expression in a commercial Luminex® system. Expression of BMP4 and Angiopoietin-1 was analyzed with Quantikine™ ELISA systems (R&D Systems) per the manufacturer’s instructions. ELISA standard curves were aligned using a five-parameter logistic regression model for sample quantification. Background concentration of all factors was measured in complete culture medium and subtracted from the detected concentrations of the respective culture supernatants for both Luminex® analysis and ELISAs. All samples were analyzed in technical duplicates.
Analysis of hormone and growth factor receptor expression
For analysis of androgen receptor and FGFR3 expression, BM-MSCs from 14 unrelated donors (7 female, 7 male; age: 53.7 y ± 13.8 y) (P2) were seeded into 12-well plates in a density of 20,000 cells/cm2 and kept under standard culture conditions for 24 hours. Thereafter, the cell monolayers were washed with ice cold PBS, lysed in 300 μl Procarta™ lysis buffer (Affymetrix, Santa Clara, CA, USA) and frozen at −80°C.
Samples were analyzed for FGFR3 and androgen receptor content with ELISA systems (USCN Life Science, Wuhan, China) per the manufacturer’s instructions. ELISA standard curves were aligned using a five-parameter logistic regression model for sample quantification. Background concentration was measured in pure lysis buffer and subtracted from the detected concentrations of the respective cell lysates. All samples were analyzed in technical duplicates.
Telomerase activity assay
To quantify telomerase activity, the Telo TAGGG Telomerase PCR ELISA kit (Roche Applied Science, Mannheim, Germany), based on the Telomeric Repeat Amplification Protocol (TRAP), was used on 2 × 105 BM-MSCs at P1 according to the manufacturer’s instructions. Samples were classified as telomerase negative when the extinction value was <0.2 after subtracting the negative control.
Statistical analysis
The statistical analyses were performed using ANOVA analysis of variance, the Spearman two-tailed correlation test, or the two-tailed Student’s t-test. Differences were considered significant when P <0.05.
Discussion
BM-MSC preparations
in vitro are known to be a heterogenic mix of adherently growing cells sharing similar, that is, spindle shaped, morphology [
25]. Studies have revealed that BM-MSC preparations
in vitro are composed of poorly defined subpopulations, and that a minority of clonally expanded MSCs showed full “tri-lineage” (adipogenic, osteogenic and chondrogenic) differentiation potential [
12,
14,
26,
27], pointing toward functional differences of BM-MSC subpopulations. In addition to intra-individual heterogeneity, donor-related variations of differentiation potential and growth kinetics of MSC preparations point toward significant inter-individual heterogeneity [
11,
12]. Studies that identify
in vivo stromal cell subpopulations are rare and focus on very few markers [
24,
28,
29]. Attempts were, therefore, made to find distinct markers to identify the most “potent” MSC subpopulations for possible clinical use [
13,
15,
23,
30‐
32], resulting in greatly improved characterization of MSC preparations
in vitro. Whether a distinct phenotype correlated with specific functional properties was addressed by sorting and analyzing few subpopulations but likely left many uncharacterized. As to heterogeneity,
in vitro data are not conclusive due to study-to-study variations on BM-MSC performance (for example, differentiation potential) and reproducibility challenges. Therefore, it is difficult to understand the
in vivo nature and composition of the BM stroma as well as the impact of
in vitro selection and culture conditions on BM-MSC subpopulations. Analysis of sorted BM-MSC subpopulations, requiring thorough optimization of culture conditions for each subpopulation, is one option. We analyzed great numbers of human BM-MSC preparations to identify statistically robust correlations of phenotype, donor gender and age to functional properties.
We identified three surface marker patterns on adherent BM-MSCs in vitro: (1) markers uniquely expressed by all cells; (2) markers not expressed by any cells; and (3) markers heterogeneously expressed by subpopulations. Here we saw prominent inter-individual heterogeneity. With respect to distinct phenotypes of BM-MSC subpopulations, these markers appear most interesting and probably allow correlation of MSC preparations to potency. The great heterogeneity of MSC preparations still hampers the development of marker-based potency assays for MSCs, such as the well-established quantification of CD34+ cells in hematopoietic stem cell preparations.
Quantification of cells expressing (or lacking) these markers allowed us to distinguish markers expressed on rare (≤10%) subpopulations versus those on more frequent subpopulations.
To what extent, if any, adult BM-MSCs exhibit stem cell properties (that is, self-renewal and broad differentiation capacity) has been under debate for over a decade [
33]. Despite difficulties in distinguishing self-renewal from proliferation or selection of immortalized clones, studies suggest MSCs might have self-renewal capacity [
34]. Clonal growth is associated with self-renewal, regulated by factors such as LIF and BMPs [
35,
36]. However, determination of a stem cell character cannot be based on clonal growth alone because the differentiation potential of cell clones derived from BM is highly variable and only a minority of clones exhibit mesodermal “tri-lineage” differentiation capacity [
26]. Regarding differentiation capacity, functional data on non-mesodermal or myogenic differentiation of human BM-MSCs are missing or controversial [
10,
19,
37] and no study has reported on differentiation of adult human BM-MSCs, isolated by plastic adherence, into cells or tissue(s) of all three germ layers upon blastocyst transfer or teratoma formation
in vivo. The surface antigen SSEA-4 and the transcription factors Oct4, Nanog and Prdm14 are regarded as stem cell markers [
33,
38,
39]. SSEA-4
+ MSCs [
40,
41], multipotent adult progenitor cells (MAPC), marrow-isolated adult multi-lineage inducible cells (MIAMI) and very small embryonic-like (VSEL) stem cells were shown to express stem cell markers, and could be differentiated into various mesodermal and non-mesodermal cell types [
42‐
47]. We have also shown
in vivo that adult human BM harbors distinct stromal cell entities expressing Oct4, Nanog and SSEA-4 on the protein/antigen level [
48]. To put these observations into context with BM-MSCs
in vitro, we analyzed clonogenic potential, mesodermal differentiation potential, expression of stem cell markers, and telomerase activity of the BM-MSC preparations. A variety of antigens, but not SSEA-4, were correlated to BM-MSC preparations with high clonogenic potential. Notably, the clonogenic potential of BM-MSCs did not correlate to differentiation potential or telomerase activity, and BM-MSCs expressed much lower levels of stem cell marker mRNA compared to pluripotent stem cells. Our data suggest that BM-MSC preparations isolated and cultured under standard conditions do not contain cells with stem cell properties, in possible contrast to MAPC, MIAMI or VSEL stem cells that require specialized isolation and culture conditions.
Our correlation analyses suggest the presence of BM-MSC phenotypes with either higher (CD10, CD119) or lower (GD2) adipogenic differentiation potential
in vitro. Human BM harbors adipogenic stromal cells expressing CD10 but not GD2
in vivo[
48]. We speculate that this cell type was, among others, positively selected by plastic adherence thereby being able to give rise to terminally differentiated adipocytes when exposed to the respective differentiation medium
in vitro. To verify this hypothesis, we suggest future studies on sorted cells investigating the adipogenic differentiation potential of CD10
+CD119
+GD2
- and CD10
-CD119
-GD2
+ BM-MSC subpopulations.
BM-MSC-mediated immunomodulation mainly comes as potent immunosuppression [
18,
49]. Therefore, BM-MSCs are used in the clinic for treatment of graft-
versus-host disease (GvHD) [
6,
50] and multiple sclerosis (MS) [
9,
51,
52], and clinical trials to treat type-1 diabetes (T1D) are underway. Here we investigated the expression of factors in BM-MSC preparations that are involved in BM-MSC immunomodulation. We show for the first time that human BM-MSCs express
IDO1 but not
IDO2, consistent with the fact that IDO1 but not IDO2 is involved in tryptophan degradation [
53]. François
et al.[
54] reported on the great variability of seven BM-MSC preparations to suppress T cell proliferation. Moreover, in analyzing 21 data points, they observed a positive correlation between IDO production by BM-MSCs and their potential to suppress T cell proliferation [
54]. In our study, despite donor-to-donor variability, MSCs from female donors showed significantly increased suppression of T cell proliferation compared to male donors. Although we observed only trends in greater
IDO1 mRNA expression in female BM-MSCs and stronger suppression of T cell proliferation by BM-MSCs expressing more
IDO1 mRNA, we hypothesize that the superior immunosuppressive properties of female BM-MSCs might be mediated by IDO1.
Galectin 1, located either on the surface or in the cytoplasma of BM-MSCs, plays an important role in MSC-mediated suppression of T cell proliferation [
55] and Galectin 1
+ stromal cells contribute to the hematopoietic niche in the BM [
56]. We detected Galectin 1, and for the first time on human BM-MSCs, PDL-1 and PDL-2 on approximately 50% of the cells within BM-MSC preparations, however, at relatively low densities per cell. Activation of Programmed Death 1 (PD-1) by the negative co-stimulatory molecules PDL-1 and PDL-2 leads to inactivation of T cells [
57] and murine BM-MSCs were shown to exert their immunosuppressive function using the PD-1 pathway [
58,
59]. We did not observe significant correlations of function, in particular immunomodulation, to the expression of Galectin 1, PDL-1 or PDL-2 on BM-MSCs. However, our results do not negate potential contributions of these molecules in human BM-MSC function as the functional analyses of the immunomodulatory properties were performed with fewer BM-MSC preparations, possibly hampering the identification of significant differences or correlations.
Many functional parameters of BM-MSCs did not correlate to donor age or gender (Table
1). We, however, found that high-clonogenic BM-MSCs were smaller, divided more rapidly and were more frequent in BM-MSC preparations from younger, female donors. These findings correspond to a study that reported a higher growth rate and clonogenic potential of BM-MSCs from children
versus adults [
60]. Alves
et al.[
61] recently associated several
in vitro characteristics of BM-MSCs to donor age. We found no correlation of donor age to adipogenic, osteogenic and chondrogenic differentiation
in vitro as confirmed by an extended panel of lineage specific markers. We identified several differentially expressed proteins on BM-MSCs from younger donors compared to older donors. Interestingly, none of these targets correlated to donor age at the mRNA level in the Alves
et al. study [
61]. Besides possible age-related post-transcriptional impact on protein expression, the exposure of fetal bovine serum plus addition of bFGF in this study (
versus PHS without any additional growth factors in our study) could have contributed to these discrepant results.
Studies have shown that MSC function
in vivo is mediated by secreted trophic factors [
7,
62‐
66] although few studies, mainly on rodent MSCs, report gender dimorphism of MSC secretion of trophic factors or cytokines [
67,
68]. We propose the pre-transplantation assessment of BM-MSC-secreted trophic factors as a clinically relevant issue. We, therefore, analyzed the secretion of factors shown to mediate tissue regeneration and immunomodulation: VEGF-A (angiogenesis), HGF (angiogenesis, anti-proliferative effects on T-cells), LIF (induction of Foxp3
+ T cells, anti-proliferative effects on T-cells), Angiopoietin-1 (angiogenesis), bFGF (angiogenesis, cell proliferation and migration), BMP4 (cell proliferation, osteogenesis) and NGFB (neuroprotection) [
18,
22]. We detected Angiopoietin-1 in the supernatants of only 4 of 11 tested BM-MSC preparations thereby identifying “Angiopoietin-1-secretor” and “Angiopoietin-1-non-secretor” BM-MSC preparations. Angiopoietin-1 secretor status was not determined by age, gender or marker expression. For further characterization we recommend RNA microarray screening analyses followed by functional assessments
in vitro and
in vivo. Previous studies reported on rat BM-MSC secretion of BMP4, a factor in proliferation and differentiation [
69,
70]. Interestingly, we could not detect BMP4 in the supernatants of human BM-MSC preparations, which suggests that secreted BMP4 might not be an ideal target to assess human BM-MSC preparations
in vitro.
PDGF and bFGF play pivotal roles in MSC proliferation [
71,
72]. We detected CD140b (PDGFRb) on nearly all cells in the BM-MSC preparations. For CD140a (PDGFRa) expression, though possibly defining a subpopulation, we did not find a difference in gender distribution or correlation to BM-MSC proliferation. We, therefore, lack evidence of a crucial role for PDGF in gender dimorphism of BM-MSC proliferation. We then evaluated the possible influence of bFGF on BM-MSC proliferation. FGF receptors play important roles in proliferation, differentiation and possibly self-renewal of MSCs [
73‐
75]. Kim
et al.[
73] showed that human synovium-derived MSCs transcribed
FGFR 1–4 mRNA but expressed only FGFR3 at the protein level. We analyzed FGFR3 protein expression in BM-MSCs without detecting differences between female and male donors, or identifying a correlation to donor age. The PHS used in our study was obtained from male donors only, with consequently more testosterone than expected from female individuals. To evaluate the role of testosterone in BM-MSC gender dimorphism, we analyzed the expression of androgen receptors in BM-MSCs. We found no difference between genders or a correlation to donor age. This corresponds to studies reporting that treatment with sex hormones (dihydrotestosterone, 17β-estradiol) did not increase the proliferative capacity of human MSCs [
76,
77].
Last, we identify three markers associated with several functional properties of BM-MSCs, that is, CD119, CD146 and HLA ABC (Table
2). We suggest future studies on sorted CD119
+, CD146
+ and HLA ABC
+ BM-MSCs to analyze the expression profile of genes that mediate therapeutic potential as well as performance in functional assays.
Table 2
Assignment of functional and phenotypical properties to CD119, CD146 and HLA ABC expression
CD119 (IFNγR1) antigen density per cell (ΔGeo Mean Fluorescence) | CFU-F | Positive | 25 |
diameter | Negative | 17 |
PDT | Negative | 21 |
Induction PPARγ mRNA | Positive | 22 |
Female donors > male donors* | n.a. | 28 |
CD146 (MCAM) antigen density per cell (ΔGeo Mean Fluorescence) | CFU-F | Positive | 31 |
PDT* | Negative | 27 |
Donor age | Negative | 34 |
HLA ABC antigen density per cell (ΔGeo Mean Fluorescence) | CFU-F | Positive | 31 |
Diameter | Negative | 22 |
PDT | Negative | 27 |
Conclusions
It has become evident that the currently most widely used BM-MSC isolation technology, that is, culture of plastic adherent BM-cells and removal of non-adherent BM-cells, is a significant selection process for BM-MSC subpopulations. Hereby, rare MSC subpopulations that possibly feature stem cell-related properties will be lost. On the other hand, a heterogenic mix of MSC subpopulations adapts very well to these conditions and proliferates in vitro as BM-MSC preparations. By analyzing and extensively characterizing a great number of BM-MSC preparations, we overcame a major challenge in research on primary BM-MSCs of donor-to-donor variation. We hereby identify phenotypes featuring functional properties that are partially donor-related, and propose future studies on CD119+, CD146+, HLA ABC+ BM-MSC subpopulations, as well as on Angiopoietin-1 secreting and non-secreting BM-MSC preparations.
For clinical production of BM-MSCs, markers that correlate positively and negatively to functional properties of BM-MSC preparations and markers that define rare subpopulations could be useful for the development of assays to define release or characterization criteria for quality control purposes and clinical applications. Moreover, the gender-related effect on growth kinetics of BM-MSCs could help to plan their scale-up production.
For clinical applications, our data might, on one hand, provide evidence to initiate clinical phase I trials comparing the immunomodulatory potential of allogeneic BM-MSC preparations from female to male donors for diseases such as GvHD, MS or T1D. On the other hand, donor age or gender might not affect BM-MSC performance in clinical applications where BM-MSC-derived trophic factors are considered to contribute substantially to efficacy of the cell therapy (for example, stroke or myocardial infarction).
Authors’ contributions
GS contributed to the conception and design of the study and to manuscript writing, and was responsible for collection of data. TK and KB were involved in the conception and design of the study, provision of the study material and the collection of data.. UHK collected the data. HN provided administrative support and helped with manuscript writing. RS was responsible for the conception and design of the study, assembled, analyzed and interpreted data, and wrote the manuscript. All authors read and approved the final manuscript.