5-Azacytidine Improves the Osteogenic Differentiation Potential of Aged Human Adipose-Derived Mesenchymal Stem Cells by DNA Demethylation

The therapeutic value of adipose-derived mesenchymal stem cells (Ad-MSCs) for bone regeneration is critically discussed. A possible reason for reduced osteogenic potential may be an age-related deterioration of the Ad-MSCs. In long term in vitro culture, epigenomic changes in DNA methylation are known to cause gene silencing, affecting stem cell growth as well as the differentiation potential. In this study, we observed an age-related decline in proliferation of primary human Ad-MSCs. Decreased Nanog, Oct4 and Lin28A and increased Sox2 gene-expression was accompanied by an impaired osteogenic differentiation potential of Ad-MSCs isolated from old donors (>60 a) as compared to Ad-MSCs isolated from younger donors (<45 a). 5-hydroxymethylcytosine (5 hmC) and 5-methylcytonsine (5 mC) distribution as well as TET gene expression were evaluated to assess the evidence of active DNA demethylation. We observed a decrease of 5 hmC in Ad-MSCs from older donors. Incubation of these cells with 5-Azacytidine induced proliferation and improved the osteogenic differentiation potential in these cells. The increase in AP activity and matrix mineralization was associated with an increased presence of 5 hmC as well as with an increased TET2 and TET3 gene expression. Our data show, for the first time, a decrease of DNA hydroxymethylation in Ad-MSCs which correlates with donor-age and that treatment with 5-Azacytidine provides an approach which could be used to rejuvenate Ad-MSCs from aged donors.


Introduction
Extensive bone loss after trauma or diseases often results in delayed or impaired bone healing [1]. Deteriorated bone regeneration and repair is observed with increasing age of patients [1,2]. The use of mesenchymal stem cells (MSCs) in tissue engineering has great potential for a novel approach in acute bone and cartilage repair, e.g. some pioneer clinical study have already entered the phase III stage [3].
MSCs are one kind of adult stem cells. In the mechanism of osteogenesis, MSCs are considered to be one type of progenitor cells, which are able to proliferate and later on differentiate into osteogenic cells [4]. Hence, the process of bone regeneration requires the recruitment, expansion and differentiation of MSCs [5]. In vitro experiments demonstrated an active self-renewal capacity and multi-lineage differentiation potential of MSCs [6,7]. MSCs can be isolated from various tissues; most frequently from adipose tissue and bone marrow. Due to their easy access, low immune rejection as well as a low risk of tumorigenesis [8,9], MSCs derived from adipose tissue (Ad-MSCs) could be an ideal source for patient-specific cell therapy. However, the osteogenic differentiation potential of these Ad-MSCs has been critically discussed. Furthermore, it has been reported many times that adult stem cells, including MSCs, suffer from a decline in stem cell function with increasing age during long term culture of the cells [10,11]. The decline observed in the self-renewal capacity, resulted in an incomplete differentiation into the committed cell lineage [12].
Epigenetic modification of the genome is considered to be one of the most important regulatory pathways affecting stem cell aging. These dynamic changes mainly appear in DNA methylation and/or chromatin remodeling [13]. Although selected histone-deacetylase inhibitors could improve osteogenic function in differentiated MSCs, they are not suitable for use because of their negative effects on cell proliferation due to DNA damage and cell-cycle inhibition [14]. Efforts have been made to investigate a milder epigenetic modification approach. It has been demonstrated that DNA demethylation can be induced by 5-Azacytidine, a DNA methyltransferase (DNMT) inhibitor [15]. In our hands 5-Azacytidine treatment improved the hepatic differentiation capacity of Ad-MSCs, in correlation to the DNA demethylation [16]. Just recently an active DNA-demethylation mechanism was described in embryonic stem cells. In these cells induction of TET gene expression promoted the conversion of nuclear 5-methylcytosine (5 mC) to 5-hydroxymethylcytosine (5 hmC) and thus DNA demethylation, in order to maintain the self-renewal capacity and the pluripotency state [17].
Thus, aim of the present study was to investigate donor-agerelated changes in the self-renewal of Ad-MSCs as well as their osteogenic differentiation potential and which role DNA methylation plays in in this process. Furthermore, we want to investigate if these donor-age-related changes can be reversed by epigenetic modifications of the DNA. The stem cell capacity will be assessed by measuring the expression levels of transcription factors characteristic for embryonic and induced pluripotent stem cells, namely, Sox2, Lin28A, Oct4 and Nanog. Cell proliferation will be determined by the amount of cells stained positive for Ki67. The osteogenic differentiation potential will be will be defined by measuring the AP-Activity and matrix mineralization after differentiation, as well as by measuring the expression levels of the osteogenic marker gene osteocalcin and the osteogenic transcription factors Runx2 (early) and osterix (late) [18].

Isolation, Expansion and Characterization of Ad-MSCs
Ad-MSCs were isolated from adipose tissue obtained from surgery with the written consent of the patients. This study was specifically approved by the ethical committee of the Universitä tsklinikum Tübingen (Reference: 385/2012 B02) who works in accordance with national regulations and the ICH-GCP guidelines. The study was performed according to the declaration of Helsinki in its newest version. For this study adipose tissue from 23 donors was collected and isolated. Adipose tissue was minced and washed extensively with PBS before digestion with collagenase II (0.075% in PBS) at 37uC for 30 to 60 min. The digestion was stopped with culture medium (DMEM, 10% FCS, 100 U/ml penicillin, 100 mg/ml streptomycin). After centrifugation (600 g, 10 min) the cell pellet was collected, re-suspended and filtered through a 70 mM cell strainer. The cells were seeded in flasks which culture medium. To remove non-attached cells the Ad-MSCs were washed once with PBS 24 hours after isolation. Ad-MSCs were expanded in culture medium until passage 3 in order to minimize the number of contaminating cells in the culture. Purity of the Ad-MSCs was determined by flow cytometry as recently published by our group (CD90 + , CD105 + , CD14 2 , CD45 2 ) [16].

Pre-treatment of Ad-MSCs with 5-Azacytidine
Ad-MSCs from young (,45 a) and old ($60 a) donors were seeded at a cell density of 10,000 cells/cm 2 . After cell attachment the medium was changed to freshly made culture medium containing 5-Azacytidine (5 mM and 20 mM). After 24 h the 5-Azacytidine containing culture medium was refreshed for another 24 h (total pre-treatment: 48 h).

Alkaline Phosphatase (AP) Activity
In order to determine the AP activity, the differentiated cells were incubated with substrate solution (2 mg/ml 4-Nitrophenyl phosphate disodium salt hexahydrate, 50 mM glycine, 1 mM MgCl 2 , 100 mM Tris, pH = 10.5) for 120 min. The subsequent 4-Nitrophenol production in the supernatant was measured photometrically at l = 405 nm with a FLUOstar Omega Plate reader (BMG Labtech, Offenburg, Germany), and normalized to relative cell numbers measured by Sulforhodamine B Staining as described before [19,20].

Matrix Mineralization
Matrix mineralization was assessed by Alizarin red and von Kossa staining. For both stainings cells were fixed with 99% ethanol for 1 h at 220uC. For alizarin red staining cells were incubated with 0.5% Alizarin red S solution (PH = 4.0) for 20 min, then washed extensively with tap water to remove unbound stain. For von Kossa staining cells were incubated with 3% AgNO 3 solution for 30 min, then washed extensively with tap water, exposed to 5% sodium carbonate (in 10% formaldehyde) solution for 2 min, followed by an incubation in 5% sodium-thiosulfate solution for 5 min in order to develop color. For both staining methods the plates were first scanned to obtain an overview image, than images with the light microscope were taken. Alizarin red staining was quantified photometrically (l = 562 nm) with a FLUOstar Omega Plate reader, after resolving the stain with 10% cetylpyridinium chloride solution. The von Kossa staining was quantified densitometrically (Pixel intensity) with the Image J. 1.45 s software (National Institutes of Health, USA).

Immunofluorescence Staining
Ad-MSCs were plated onto cover slips and treated according to the experimental setup. Cells were fixed with 4% paraformaldehyde solution for 15 min at RT and then washed with PBS.  ml in PBS) for 20 min at RT. After a final washing step with PBS, the stained cells were mounted with mounting medium (Fluoromount G, Southern Biotech, NJ, USA). Images of the staining were taken with an EVOS fluorescence microscope (AMG, USA) under standardized condition, processed and analyzed with Image J 1.45 s software [21].

Quantification of Gene Expression
Total RNA from Ad-MSCs was extracted with GeneJET RNA Purification Kit. 2-3 mg of the total RNA was digested with DNase I in order to remove remaining genomic DNA. cDNA was synthesized using the First Strand cDNA Synthesis Kit. Gene expression changes were analyzed either by conventional RT-PCR or by qRT-PCR depending on the amount of target mRNA present in the sample. PCR protocols were optimized for each gene separately to be in a quantifiable (linear) range in our samples.
For quantitative real time PCR (qRT-PCR) a standardized amount of template cDNA was tested for the expression level of each target gene (primer sequences are listed in Table 1) using SYBR Green qPCR (Finnzym, Vantaa, Finnland) and the Step One Plus Real-Time PCR System (Applied Biosystems, CA, USA). GAPDH was used as endogenous control. Relative changes in gene expression were calculated using the DDC T Method with either untreated young cells (changes with age) or untreated age matched cells (changes du to 5-Azacytidine treatment) as reference.
For semi-quantitative (conventional) RT-PCR a standardized amount of template cDNA was tested for the expression level of each target gene (primer sequences are listed in Table 1) using KAPA Fast ready mix (Peqlab, Erlangen, Germany). Products, resolved by gel electrophoresis in a 2% (w/v) agarose gel, were visualized with ethidiumbromide. Densitometric analysis of signals was performed using ImageJ software (NIH, Bethesda, USA). Pixel Intensities of the PCR signal were normalized to GAPDH using a normalization factor (multiple of the average GAPDH signal). Relative changes in gene expression were calculated with either untreated young cells (changes with age) or untreated age matched cells (changes du to 5-Azacytidine treatment) as reference.

Statistical Analysis
Results are expressed as box blot (Whisker's: min to max) or bar chart (mean 6 SEM) of at least 3 independent experiments (biological replicates, N$3) measured as triplicates or more (technical replicates, n$3). The precise of replicates for each experiment is given in the figure legends. Individual data sets were compared by Student's t test or one way ANOVA analysis (Prism 5.04, GraphPad Software, CA, USA), with consideration of the sample distribution. p,0.05 was set as the minimum level of significance.

Ad-MSCs from Elderly Donors showed Impaired Proliferation and Osteogenic Differentiation Capacity
In order to assess the proliferation capacity of the cells a Ki67 staining was performed. Expression of Ki67 was significantly lower in Ad-MSCs isolated from aged donors as compared to Ad-MSCs isolated from younger donors ( figure 1A & B), representing an impaired proliferation capacity of Ad-MSCs from elderly donors. In order to evaluate the differentiation potential, Ad-MSCs were differentiated for 14 days in the presence of osteogenic differentiation medium. As osteogenic markers alkaline phosphatase (AP) activity, matrix mineralization and relative gene expression of Runx2, osterix and osteocalcin were determined after the differentiation process. In comparison to Ad-MSCs from young donors, Ad-MSCs from old donors exhibited a significantly lower AP activity ( figure 1C) and a significant weaker matrix mineralization, as determined by von Kossa (figure 1D & F) and Alizarin red staining ( figure 1E & G). Increased expression of the early osteogenic transcription factor Runx2 and reduced expression of the late osteogenic transcription factor osterix suggest a delay in osteogenic differentiation in Ad-MSCs from aged donors. This is supported by reduced expression of osteocalcin, the most abundant noncollagenous protein in bone (figure 1H-J).

Gene Expression of Pluripotency Markers Demonstrated Age-related Differences in Ad-MSCs
The differentiation potential was evaluated by relative gene expression levels of the pluripotency markers Nanog, Oct4, Lin28A and Sox2 by qRT-PCR. GAPDH was used as endogenous control. Expression of Lin28A was significantly lower in Ad-MSCs from elderly donors ( figure 2A). Similarly, expression of Oct4 was reduced in Ad-MSCs from aged donors ( figure 2B). Furthermore, expression of Nanog was significantly lower in Ad-MScs from elderly donors (figure 2C). Conversely, expression levels of Sox2 were significantly elevated in Ad-MSCs from aged donors ( figure 2D). The observed expression changes were confirmed on protein level by Western blot (figure 2E).

TET Gene Expression and 5 mC/5 hmC Distribution in Ad-MSCs from Young and Old Donors
Members of the TET family (TET1-3) are reported to actively regulate DNA methylation. Therefore, we determined relative gene expression levels of TET1, TET2 and TET3 by qRT-PCR. GAPDH was used as endogenous control. Expression of TET1 was near/below the detection limit in Ad-MSCs from young and elderly donors. Therefore, no significant difference could be detected (data not shown). Expression levels of TET2 were

5-Azacytidine Treatment Induced Expression of TET2 and TET3 while Increasing Nuclear 5 hmC Levels
5-Azacytidine is a well characterized DNA-methyltransferase inhibitor. In order to evaluate the effect of 5-Azacytidine on the DNA methylation status, relative gene expression levels of TET1, TET2 and TET3 were determined by qRT-PCR in Ad-MSCs from young and old donors treated for 48 h with 5 mM 5-Azacytidine. GAPDH was used as endogenous control. After 5-Azacytidine treatment TET1 expression was still near the detection limit in Ad-MSCs from both young and old donors, where no significant difference could be detected (data not shown). In both age groups, TET2 expression levels were significantly increased by 5-Azacytidine treatment ( figure 4A). Similarly, TET3 expression levels were increased by 5-Azacytidine treatment in both age groups, with more pronounced changes in Ad-MScs from aged donors (figure 4B). Expression changes of TET2 and TET3 due to 5-Azacytidine treatment were confirmed by Western blot on the protein level (data not shown). Genomic distribution of 5 hmC and 5 mC, in Ad-MSCs from old donors, was strongly affected by 5-Azacytidine treatment, as determined by immunofluorescence staining. The 5 hmC-positive fraction was significantly increased in Ad-MSCs from old donors after treatment with 5-Azacytidine ( figure 4C). This resulted in a significant decrease in the 5 mC-positive fraction in Ad-MSCs from old donors after treatment with 5-Azacytidine (figure 4D). Consequently relative expression levels of the pluripotency markers Nanog, Oct4, Lin28A and Sox2 were changed by 5-Azacytidine treatment, as determined by qRT-PCR. Expression levels of Sox2 were significantly decreased by 5-Azacytidine treatment only in Ad-MSCs from aged donors (figure 4E). Expression of Lin28A was significantly induced by 5-Azacytidine treatment in Ad-MSCs from elderly donors (figure 4F). Expression of Oct4 was significantly induced by 5-Azacytidine treatment in Ad-MSCs from both age groups (figure 4G). In contrary to Lin28A expression of Nanog was induced by 5-Azacytidine treatment only in Ad-MScs from young donors (figure 4H). On the protein level, pluripotency markers showed a comparable regulation due to 5-Azacytidine treatment (figure 4I).

5-Azacytidine Pre-treatment Improved Proliferation and Osteogenic Differentiation of Ad-MSCs from Aged Donors
In the following experiments we wanted to analyze of the reduction in DNA methylation by 5-Azacytidine improves the proliferation and the osteogenic differentiation capacity of aged Ad-MSCs. Following the treatment of Ad-MSCs from old donors with 5 mM 5-Azacytidine for 48 h significantly more cells were Ki67-positive (1.3-fold, p,0.01), representing an improved proliferation. Conversely, no improvement of the proliferation was observed in Ad-MSCs from young donors following the treatment with 5 mM 5-Azacytidine for 48 h ( figure 5A & B). After pretreatment with 5 mM 5-Azacytidine for 48 h, Ad-MSCs from aged donors were differentiated for 14 days in the presence of osteogenic differentiation medium. As osteogenic markers alkaline phosphatase (AP) activity, matrix mineralization and relative gene expression of Runx2, osterix and osteocalcin were determined after the differentiation process. Pre-treatment with 5-Azacytidine improved AP activity (

Discussion
Stem cell functions include self-renewal and multipotency or even pluripotency. In particular, self-renewal of stem cells does not only include proliferation, but also the maintenance of their differentiation potential after cell division [23]. Our results demonstrate a decreased proliferation rate of Ad-MSCs from aged donors as well as an impaired osteogenic differentiation potential, as characterized by a declined AP activity, a weaker mineralization of extracellular matrix and reduced expression of late osteogenic marker genes osterix and osteocalcin in the differentiated cells from elderly donors, which was accompanied by a strong DNA methylation. Increased global DNA methylation status is a well-established phenomenon that has been observed in many aged adult tissue/cell types [24].
Our preliminary research has shown that 5-Azacytidine treatment significantly reduces the global level of DNA methylation [16]. In the past, the demethylation effect of 5-Azacytidine was described as a cause of direct or indirect inhibition of the DNA methyltranferases (DNMTs) [25]. This theory resembles mainly the passive DNA demethylation process. Recently, more active mechanisms of DNA demethylation have been uncovered [26]. TET proteins mediated conversion of 5 mC to 5 hmC was shown to play a crucial role in active demethylation of DNA [27]. Here, we have presented for the first time, that Ad-MSCs display reduced levels of 5 hmC during aging. With the conversion of osteogenic differentiation was induced. In the differentiated cells osteoblasts function was assessed by measuring (C) AP activity (N = 7, n = 3/age group), (D-G) matrix mineralization and (H-J) gene expression changes. Matrix mineralization was determined by (D & F) von Kossa (N = 4, n = 3/age group) and (E & G) Alizarin red staining (N = 6, n = 3/age group). mRNA levels of the osteogenic transcription factors (H) Runx2 and (I) Osterix, as well as the osteogenic marker (J) Osteocalcin were determined by semi-quantitative RT-PCR (N = 5, n = 3/age group). GAPDH was used for normalization. All data are presents as relative changes (log 10 ) of 5-Azacytidine pre-treated differentiated Ad-MSCs compared to the corresponding differentiated Ad-MSCs without pre-treatment.*p,0.05, **p,0.01, ***p,0.001 as compared to differentiated Ad-MSCs without pre-treatment (student's t-test). doi:10.1371/journal.pone.0090846.g005 5 mC to 5 hmC by TET proteins [17], further modifications can occur in order to launch the active DNA demethylation, i.e. oxidation to 5-formylcytosine (5 fC) and 5-carbonylcytosine (5 caC). 5 caC eventually will be removed by a base excision pathway and thus will be replaced with unmethylated cytosine [26]. Thus, the finding of reduced 5 hmC levels in aged Ad-MSCs could account for a DNA hypermethylation pattern in these [16] and other adult cell types [24]. This is further supported by the finding that 5 hmC has been proven to act like a poor substrate for DNMT1 recognition [28]. In this way, it is also able to take part in the passive DNA demethylation by preventing factors that interact with methylcytosine [26].
Most interestingly, treatment of aged Ad-MSCs with the DNMT-inhibitor 5-Azacytidine not only reduced DNA methylation passively [16] but also actively, by inducing TET expression and nuclear 5 hmC. In contrast to embryonic stem cells, where only TET1 plays a key role in DNA hydroxylation [17,29], in aged Ad-MSCs TET2 and TET3 expression was significantly induced by 5-Azacytidine treatment. Controversially to the hypothesis that increased TET levels lead to an accumulation of 5 hmC, aged Ad-MSCs show slightly increased basal expression levels of TET2. However, this hypothesis is based on work with embryonic stem cells [17,29], where TET1 is the master regulator. In our cells from both age groups, however, TET1 expression is negligible as it is close to/below the detection limit. Furthermore, it is still unclear, whether expression and activity of DNA methyl-transferases is altered with aging as suggested by Casillas and co-workers [30].
The work of Koh et al reports that expression of TET1 and TET2 is strongly dependent on the expression of Oct4, while TET3 expression negatively correlates with Sox2 expression levels [31]. In Ad-MSCs from elderly donors expression of the pluripotency-associated genes Nanog, Oct4 and Lin28A was strongly reduced, while Sox2 expression was induced. After 5-Azacytidine Oct4 and Lin28A expression significantly increased and Sox2 expression was significantly decreased in aged Ad-MSCs. Considering the data from Koh et al this might explain the observed increase in expression of TET2 and TET3 after 5-Azacytidine treatment in Ad-MSCs from aged donors [31]. Why TET2 expression was not affected by 5-Azacytidine treatment in Ad-MSCs from young donors might be explained by the interaction of the different pluripotency factors itself [32], as in this age group 5-Azacytidine treatment induced the expression of Oct4 and Nanog but not Lin28A.
Oct4 and Nanog are described as core transcription factors for the regulation and the maintenance of pluripotency in embryonic stem cells [33,34]. It has been reported that Lin28A is able to promote cell proliferation in embryonic stem cells and breast tumor cells [35,36]. Our data imply that the observed age-related loss of capacity for self-renewal and differentiation could be triggered by the reduced expression of these and other genes controlling stem cell renewal. This implication was supported by the work of Tsai et al, showing, that with increasing time in culture, Nanog and Oct4 expression as well as osteogenic differentiation capacity decreases in MSCs [37]. Lin28A and Oct4 may play a special role in this regulatory mechanism as the treatment with 5-Azacytidine, which strongly induced the expression of Lin28A and Oct4, was able to improve proliferation and osteogenic differentiation in Ad-MSCs from elderly donors. Basal expression levels of the pluripotency factor Sox2 was very low in young Ad-MSCs, which is in line with other publications [38]. Interestingly, basal Sox2 expression levels were significantly elevated in Ad-MSCs from elderly donors and decreased with the 5-Azacytidine treatment. The decline in Sox2 expression after 5-Azacytidine treatment may be pivotal for the improved osteogenic differentiation, as there are reports showing that depletion of Sox2 promotes osteogenic differentiation in a great variety of stem cells [39,40,41,42].
In our aged Ad-MSCs 5-Azacytidine not only induced the cells proliferation but also accelerated their osteogenic differentiation. This is shown by a significant increase in AP activity and matrix mineralization. Supporting this hypothesis the early osteogenic transcription factor Runx2 is reduced in favor for a faster induction of the late osteogenic marker genes osterix and osteocalcin [18].

Conclusion
Summarizing, our data suggest that DNA methylation plays a crucial role in adult stem cell aging, as DNA methylation increases with donor age in Ad-MSCs. We could show that 5-Azacytidine, a small molecule inhibitor for DNMTs, not only passively but also actively promotes DNA demethylation in Ad-MSCs from aged donors. This balanced the expression levels of pluripotency genes and improved the osteogenic differentiation potential, thus rejuvenating aged Ad-MSCs.