Resistance for Genotoxic Damage in Mesenchymal Stromal Cells Is Increased by Hypoxia but Not Generally Dependent on p53-Regulated Cell Cycle Arrest

Adult stem cells including multipotent mesenchymal stromal cells (MSC) acquire a high amount of DNA-damage due to their prolonged lifespan. MSC may exert specific mechanisms of resistance to avoid loss of functional activity. We have previously shown that resistance of MSC is associated with an induction of p53 and proliferation arrest upon genotoxic damage. Hypoxia may also contribute to resistance in MSC due to the low oxygen tension in the niche. In this study we characterized the role of p53 and contribution of hypoxia in resistance of MSC to genotoxic damage. MSC exhibited increased resistance to cisplatin induced DNA-damage. This resistance was associated with a temporary G2/M cell cycle arrest, induction of p53- and p21-expression and reduced cyclin B / cdk1-levels upon subapoptotic damage. Resistance of MSC to cisplatin was increased at hypoxic conditions i. e. oxygen <0.5%. However, upon hypoxia the cisplatin-induced cell cycle arrest and expression of p53 and p21 were abrogated. MSC with shRNA-mediated p53 knock-down showed a reduced cell cycle arrest and increased cyclin B / cdk1 expression. However, this functional p53 knock down did not alter the resistance to cisplatin. In contrast to cisplatin, functional p53-knock-down increased the resistance of MSC to etoposide. We conclude that resistance of MSC to genotoxic damage is influenced by oxygen tension but is not generally dependent on p53. Thus, p53-dependent and p53-independent mechanisms of resistance are likely to contribute to the life-long functional activity of MSC in vivo. These findings indicate that hypoxia and different resistance pathways contribute to the phenotype that enables the prolonged lifespan of MSC.


Abstract
Adult stem cells including multipotent mesenchymal stromal cells (MSC) acquire a high amount of DNA-damage due to their prolonged lifespan. MSC may exert specific mechanisms of resistance to avoid loss of functional activity. We have previously shown that resistance of MSC is associated with an induction of p53 and proliferation arrest upon genotoxic damage. Hypoxia may also contribute to resistance in MSC due to the low oxygen tension in the niche. In this study we characterized the role of p53 and contribution of hypoxia in resistance of MSC to genotoxic damage. MSC exhibited increased resistance to cisplatin induced DNA-damage. This resistance was associated with a temporary G2/M cell cycle arrest, induction of p53-and p21-expression and reduced cyclin B / cdk1-levels upon subapoptotic damage. Resistance of MSC to cisplatin was increased at hypoxic conditions i. e. oxygen <0.5%. However, upon hypoxia the cisplatin-induced cell cycle arrest and expression of p53 and p21 were abrogated. MSC with shRNA-mediated p53 knock-down showed a reduced cell cycle arrest and increased cyclin B / cdk1 expression. However, this functional p53 knock down did not alter the resistance to cisplatin. In contrast to cisplatin, functional p53-knock-down increased the resistance of MSC to etoposide. We conclude that resistance of MSC to genotoxic damage is influenced by oxygen tension but is not generally dependent on p53. Thus, p53-dependent and p53-independent mechanisms of resistance are likely to contribute to the life-long functional activity of MSC in vivo. These findings indicate that hypoxia and different resistance pathways contribute to the phenotype that enables the prolonged lifespan of MSC. PLOS

Introduction
During their lifelong presence the various types of adult stem cells in the human body contribute to the functional maintenance of tissues but are also exposed to a high amount of DNAdamage. Therefore, these cells are supposed to exert specific mechanisms of resistance to genotoxic damage and thereby avoid loss of functional activity as well as malignant transformation [1; 2]. Several reports indicate that p53 contributes to such resistance in gastrointestinal, hematopoietic and hair bulge stem cells [3][4][5][6]. Multipotent mesenchymal stromal cells (MSC) comprise a population of pericytic cells with adult stem cell characteristics [7]. They are present in several compartments of the human organism including bone marrow and adipose tissue [8]. Their ability of multipotent differentiation, immunomodulation and differentiation to carcinoma-associated fibroblasts (CAF) enable them to act as niche cells controlling normal tissue turnover like hematopoiesis [9]. But MSC may also contribute to malignant growth [10].
Based on their isolation mainly from bone marrow, MSC are present in the adult and elderly human body [11; 12]. Data on the impact of age on MSC frequency in vivo and their characteristics are conflicting [11; 13]. However, apart from a yet disputed role in sarcoma [14] MSC are not overtly prone to malignant transformation [15]. Therefore, MSC as other adult stem cell types need to harbor mechanisms to cope with genotoxic damage.
We and others have previously shown that MSC are resistant to genotoxic damage [12; 16; 17]. This resistance is characterized by a high threshold for apoptosis induction. Resistance is accompanied by p53 induction and proliferation arrest [12]. These results suggested a specific role for p53-regulated cell cycle arrest in MSC resistance.
Cell cycle progression and thereby resistance of cells and specifically MSC is also impacted by local oxygen tension [18; 19]. In vivo, MSC are exposed to conditions with 2 -8% oxygen concentration [20]. Specifically, in the bone marrow MSC exist in near-hypoxic conditions [21]. We therefore hypothesized that low oxygen contributes to a p53-regulated resistance in MSC.
With the present study we aimed to characterize the role of p53 in resistance of MSC to genotoxic damage and the contribution of hypoxia to this resistance. Our data show, that resistance of MSC to genotoxic damage is increased by hypoxia but is not generally dependent on a p53-regulated cell cycle arrest.
For genotoxic treatment MSC at passage 2 or 3 were grown to 50% confluence. The cells were overlaid with growth medium containing cisplatin (2 μM or 20 μM) or etoposide (0.75 μM; both Sigma-Aldrich). This time point was defined as day (d) 0. After 24 h the drug containing medium was removed and the cells were washed with PBS. MSC were subsequently cultivated in drug-free growth medium for later analyses (d3 or later) or harvested by trypsinization for immediate analyses (d1).
p53-knock down in MSC was introduced via lentiviral mediated shRNA. Lentiviral particles targeting human p53 (sc-29435-V) and control particles encoding a scrambled sequence (sc-108080) were purchased from Santa Cruz Biotechnologie. Staining for senescence-associated beta-galactosidase activity was performed according to manufacturer's instructions using Senescence β-Galactosidase Staining Kit (Cell Signaling Technology).

Cellular platinum uptake
Cellular platinum drug accumulation was analyzed by atomic absorption spectroscopy (AAS) as described previously [22]. Briefly, MSC and TGCT cells were treated for 24 h with 3 μM cisplatin (Sigma-Aldrich). After rinsing with phosphate buffered saline (PBS; PAN-Biotech) adherent cells were harvested by trypsinization and dried. Platinum concentration was expressed as content [ppm] in lyophilized cells.

Cell cycle analysis
Detached cells were prepared with BD CycleTEST™ PLUS DNA Reagent Kit following manufacturer's instructions. Analyses were performed on a FACSCalibur using CellQuest software (all Becton, Dickinson and Company, San Jose, USA).
For quantification of quiescent (G0 phase) cells, DNA and RNA were differentially stained with 7-amino-actinomycin D (7-AAD) and pyronin Y [23]. Detached cells were fixed in 70% ethanol for at least 2 h at 4˚C, rinsed with PBS/1% FCS and incubated with 10 μg/ml 7-AAD (Sigma) for 20 min at room temperature. After rinsing, cells were incubated on ice for 5 min before adding 1 μg/ml pyronin Y. Following incubation for another 10 min on ice samples were acquired on a FACSCalibur.

Cytotoxicity testing
Cytotoxicity was tested using the sulforhodamin-B assay [24] as described previously [12]. Briefly, MSC were seeded into 96-well plates with 1,000 cells per well in growth medium and after 24 h the cells were treated with different concentrations of cisplatin or etoposide. We implemented 2 different treatment schedules: (I) MSC were treated for 24 h, washed with PBS, incubated for 72 h with drug-free medium fixed with 10% trichloroacetic acid and (II) MSC were treated for 96 h and fixed subsequently. After staining with 0.4% sulforhodamine-B the percentage of surviving cells relative to untreated controls are represented in semilogarithmic dose-response plots.
For cytotoxicity testing under different oxygen conditions MSC were seeded in parallel into three 96-well plates with 2,000 cells per well in growth medium and the plates were brought to normoxia, physioxia and hypoxia, respectively. After 24 h the cells were overlaid with different concentrations of cisplatin and kept under the respective oxygen conditions for another 24 h and fixed subsequently.

Comet assay
MSC were treated for 24 h with 20 μM cisplatin or etoposide. DNA strandbreaks upon cytotoxic damage were evaluated using the comet assay (alkaline single-cell gel electrophoresis) according to published protocols [25]. Short, detached cells were embedded in low melt agarose and layered on a microscope slide (10,000 cells per slide). After incubation for 1 h in lyse buffer (2.5 M NaCl, 100 mM EDTA, 10 mM TRIS, 1% triton X-100, 10% DMSO) the slides were pre-equilibrated in alkaline electrophoretic buffer (300 mM NaOH, 1 mM EDTA, pH13) for 1 h, followed by 30 min electrophorese at 25 V. Slides were rinsed in 0.4 M TRIS, pH 7.5 and stained with 10μg/ml propidium iodide.
Samples were evaluated using (I) the method of visual scoring which comprised a classification of comets according to the extent of tail DNA into class 0 (no tail) to 4 (almost all DNA in tail) [26] and (II) the description of introduced damage as% DNA in tail applying CASP Lab Software (http://casplab.com).

Statistical analysis
Statistical analysis was performed using SPSS 16.0 software (SPSS Inc., Chicago, IL, USA). If not otherwise stated depending on the equality of variances as assessed by Levene testing, results from the t-test were used in case of equal variances and results from an updated t-test were used in case of non-equal variances. In any case, a p-value < 0.05 was considered significant.

Results
Resistance of MSC to genotoxic damage is accompanied by a temporary cell cycle arrest and p53 target gene expression MSC showed increased resistance to genotoxic damage in comparison to pluripotent cells. Upon cisplatin exposure MSC showed significantly higher IC50 and IC90 values than embryonic carcinoma cells TGCT H12.1 and 2102EP (Fig 1A). For subsequent mechanistic evaluations we chose a subapoptotic damage (2 to 3 μM cisplatin, 24 h).
Temporary subapoptotic damage resulted in a proliferation stop but was followed by a complete restoration of proliferation (Fig 1B). At the time of restored proliferation (dR; 15.7 ± 3.6 d), cells showed typical MSC-features including phenotype and multipotent differentiation potential (S1C and S1D Fig). Analysis of cell cycle distribution revealed accumulation in S phase immediately after damage (d1) which was followed by a significant G2/Mphase accumulation on d3 and complete return to pre-treatment distribution on dR (Fig 1C,  S1B Fig). Expression analysis revealed an altered expression of p53-related cell cycle-controlling proteins with an increased expression of p53 and p21 and a reduced expression of cdk1 and cyclinB ( Fig 1D). Again, a return to pre-treatment expression levels was seen on dR.
We therefore assumed that resistance to subapoptotic cisplatin-induced damage relied on a temporary, p53-dependent G2/M-cell cycle arrest.
Hypoxia increases the resistance of MSC to genotoxic damage without impacting cell cycle MSC are physiologically exposed to lower oxygen concentrations [20] than used for in vitro culture. Hence, we explored response to genotoxic damage at physiologic (~6%, "physioxia") and hypoxic (< 0.5%) conditions. Under both conditions MSC maintained an unaltered immunophenotype and multipotent differentiation potential for up to 14 days (S2A and S2B We determined sensitivity of MSC to subapoptotic damage under physioxic and hypoxic conditions. Analyses were performed immediately after cisplatin exposure to avoid bias of a subsequent differential proliferation at the different conditions. Upon hypoxia, MSC showed a reduced sensitivity for cisplatin compared to normoxic conditions. In contrast, physioxia did not alter the sensitivity of MSC (Fig 2A). To verify that hypoxia protects MSC, we exposed hypoxic MSC to an apoptosis-inducing damage (20 μM cisplatin, 24h). In line with a hypoxiaconferred resistance, the proportion of SubG1 cells, i. e. apoptotic cells was significantly reduced in hypoxic compared to normoxic MSC (Fig 2B).
We postulated that as for resistance upon normoxia, hypoxia-conferred additional resistance would be related to cell cycle alterations. In untreated hypoxic MSC the proportion of Sphase cells was only slightly increased compared to both, normoxic as well as physioxic MSC (Fig 2C). Upon subapoptotic damage (2μM cisplatin, 72h), the G2/M arrest as described for normoxic MSC (see Fig 1C) was similarly observed in physioxic MSC. However, hypoxic MSC showed no alterations in cell cycle distribution upon subapoptotic damage. In contrast to subapoptotic damage, apoptosis-inducing damage (20μM cisplatin, 72h) resulted in a similar cell cycle distribution in normoxic, hypoxic and physioxic MSC (Fig 2C). In accordance with these cell-cycle alterations, the expression of p21 was increased in physioxic MSC upon subapoptotic as well as the expression of p53 and p21 upon apoptosis-inducing damage similarly as in normoxic MSC. In contrast, damage of hypoxic MSC neither at subapoptotic nor at apoptosisinducing dose resulted in an induction of p53 and p21 expression (Fig 2D).
These findings suggested that resistance to genotoxic damage in MSC at normoxic and physioxic conditions relies on p53-controlled cell cycle arrest. Thus, we hypothesized that inhibition of p53 would abrogate damage-induced arrest and increase the sensitivity of MSC.

Knock down of p53-knock abrogates damage-induced cell cycle arrest but does not sensitize MSC
To examine this hypothesis, we introduced a lentiviral shRNA-mediated p53-knock down in MSC (MSC p53kd ) resulting in a sufficient reduction of protein levels (Fig 3A). In some experiments, MSC p53kd showed enhanced proliferation but finally reached senescence as typical for MSC (S3A and S3B Fig). Transduction with neither sh-p53 nor control-shRNA did alter con- Upon subapoptotic damage, MSC with p53 knock-down showed a reduced induction of p21 compared to controls. Also, expression of cyclin B and cdk1 was more stable in MSC p53kd when compared to treated control-MSC. This expression pattern indicated that knock-down of p53 was indeed functional (Fig 3A). Again, subapoptotic damage did not alter consensus MSC characteristics in MSC p53kd .
Distribution of cell cycle fractions was similar in cisplatin-treated and untreated MSC p53kd (Fig 3B). Therefore, in line with our hypothesis and the altered expression of regulating proteins, p53-knock-down abrogated the cell cycle arrest in response to subapoptotic genotoxic damage in MSC.
Despite this altered cell cycle response, MSC p53kd and control-MSC showed identical sensitivity to genotoxic damage by cisplatin (Fig 3C), i. e. knock-down of p53 resulted in an altered cell-cycle regulation without impacting the resistance of MSC to genotoxic damage. Thus, resistance to genotoxic damage of MSC did not depend on p53-controlled cell cycle arrest.
We next wanted to test, whether the additional resistance conferred by hypoxia was also independent of p53. Sensitivity to cisplatin induced genotoxic damage was decreased in MSC with p53 knock-down upon hypoxia compared to exposure at normoxia and physioxia ( Fig  3D). This paralleled the decreased sensitivity upon hypoxia as seen for wildtype MSC (Fig 2A). As for wildtype MSC no damage-induced G2/M-arrest was seen in MSC p53kd upon hypoxia (S3E Fig). Accordingly, hypoxia did not result in changes of p53 and p21 expression in MSC p53kd upon damage (S3F Fig). Thus, hypoxia induced increased resistance did not depend on p53. Taken together, these data show that p53-dependent cell cycle regulation was not a general prerequisite for resistance to genotoxic damage in MSC.

p53 knock down sensitizes MSC for damage by etoposide
Cisplatin exerts its damage through direct DNA-crosslinking. We hypothesized that p53 may impact other forms of DNA-damage in MSC. We had previously shown that MSC are resistant to etoposide [12] which induces DNA-strand breaks by topoisomerase inhibition.
Based on our previous studies we used a prolonged 96 h exposure for initial sensitivity testing in order to achieve apoptosis-inducing damage. As previously reported and as seen for cisplatin, MSC showed resistance to etoposide-related damage in comparison to sensitive pluripotent cells (S4A Fig).
Similar to cisplatin a subapoptotic damage by etoposide resulted in a G2/M arrest ( Fig 4A) and altered expression of cell cycle-regulating proteins with upregulation of p53 and p21 and down regulation of cyclin B and cdk1 (Fig 4B). However, in contrast to cisplatin-conferred damage, MSC with p53-knock down showed increased sensitivity to etoposide-induced damage (Fig 4C). This resulted in significantly lower IC50 (93.3 ± 11.5 μM vs. 45.0 ± 26.0 μM, n = 3; p = 0.042) for etoposide in MSC p53kd compared to control-MSC.
We hypothesized that knock down of p53 in MSC allows for more DNA-damage by etoposide but does not affect cisplatin-induced DNA-damage. To validate this, we compared DNAdouble strand breaks induced by cisplatin and etoposide using the comet assay. In order to compare damage at a defined response, cells were treated at apoptosis-inducing concentrations respectively. Analysis was performed with software-based direct evaluation of tail-DNA-content (S4B Fig) as well as visual scoring (Fig 4D). Visual scoring was expressed in arbitrary units and corresponded closely to computer image-based analysis [26]. Both evaluations revealed no difference in DNA-damage between control-MSC and MSC p53kd upon cisplatin treatment. In contrast, for etoposide a significant increase in DNA-damage was seen in MSC p53kd compared to control-MSC (visual score: Chi 2 analysis, p < 0.001; tail-DNA-content: t-test, p < 0.001).
Conclusively, resistance of MSC to genotoxic damage by etoposide depended on p53.

Discussion
The main finding of our study is the observation that the resistance of MSC to genotoxic damage is enhanced by hypoxia but does not generally depend on a p53-regulated cell cycle arrest. This conclusion is derived from the following observations: i) MSC showed a relative resistance and cell cycle arrest upon cisplatin-induced damage, ii) hypoxia increased the resistance to cisplatin without inducing a cell cycle arrest, iii) p53-shRNA transduced MSC showed reduced cell cycle arrest and elevated cyclin B / cdk1 expression-thus demonstrating functional p53 knock down-but showed no altered resistance to cisplatin, iv) in contrast to cisplatin, functional p53-knock-down decreased the resistance of MSC to etoposide. We conclude that our data reflect either the parallel activity of differential p53-dependent and p53-independent mechanisms of resistance in MSC or their activity in different subpopulations of cultivated MSC.
A resistance of MSC to different DNA-damaging agents [12; 17] as well as to radiationinduced damage [27] has been described by us and others. In accordance with our previous observation the resistance was evident at subapoptotic (i. e. IC50) as well as apoptotic (i. e. IC90) doses of DNA-damage by either cisplatin or etoposide. Our data confirm that for cisplatin this resistance is not due to reduced accumulation of damage as we observed an even higher degree of DNA-platination in MSC compared to sensitive tumor cells.
Our data show that MSC respond to genotoxic damage with an increased expression of p53 and p21 and an ensuing G2/M cell cycle arrest. A similar response has been observed in hematopoietic stem cells [28]. Thus MSC like other adult stem cells are distinct from pluripotent embryonic cells which show a p53-independent cell cycle progression [29]. The dependence on p53 is supported by the reduced expression of cdk1 and cyclin B upon genotoxic damage as p53 has been shown to inhibit the promoter activation of both genes [30].
A p53-induced cell cycle arrest may be followed by senescence as has been shown for radiation-induced damage in MSC [31]. In contrast to this, our recent and current data show that the cisplatin-or etoposide-induced cell cycle arrest in MSC is only temporary [12]. MSC resumed proliferation without alteration of their main characteristics [12].
The cisplatin-induced cell cycle arrest in MSC is p53-dependent as it was completely abrogated by knock-down of p53. This loss of damage-induced p53-dependent arrest did not affect sensitivity of MSC to cisplatin but increased their sensitivity to etoposide. We conclude, that response of MSC to genotoxic damage is not generally dependent on p53. This is supported by the observation that hypoxia increases resistance to cisplatin-induced damage but abrogates the cisplatin-induced cell cycle arrest similarly as p53 knock down does. An effect of oxygen tension on characteristics of MSC has been well established [18; 19]. In contrast to others, we observed no effect on proliferation and resistance at physioxia, i. e.~6% oxygen, but only at severe hypoxia. This may be explained by different culture methods. In our assays, no media change was performed, i. e. no intermittent exposure to higher oxygen concentration occurred. However, incubation at lower oxygen was much shorter than in other reports [18; 19]. Although hypoxia often is described as an inducer of p53 there are variations among different cell types [32]. Our data showed no influence of hypoxia on p53 protein level which is in accordance with other reports [19; 33]. As it was not the aim of our study no definite conclusions regarding the impact of oxygen concentration on damage resistance in MSC can be drawn. However, our data on damage at hypoxia suggested that cell-cycle arrest is not a prerequisite for resistance of MSC to cisplatin-induced damage.
Using shRNA, we achieved an efficient and functional knock down of p53 as demonstrated by protein expression for p53 and down-stream proteins p21, cyclinB and cdk1 as wells as by abrogated cell cycle arrest upon cisplatin damage. However, residual activity of p53 may have had a differential impact on response to cisplatin and etoposide as a dose dependence of p53-mediated effects has been reported [34].
Our data show that resistance to etoposide in MSC depends on p53 thereby suggesting different, damage-specific mechanisms of resistance in MSC. Which other mechanisms may contribute to this result is purely speculative. Similar to our study, p53-knock down in murine MSC was accompanied by an increased proliferation [18]. Etoposide acts as a topoisomerase II inhibitor and therefore affects only actively dividing cells whereas cisplatin-induced DNA-platination is independent of cell cycle stage. Therefore, increased proliferation upon p53 knock down may facilitate etoposide-induced damage.
A p53-independent resistance to radiation has been shown for adult hematopoietic and mammary gland stem cells and a concept of p21-directed exhaustion of damaged stem cells has been proposed [3]. Our data correspond to this concept. MSC cultivated in vitro represent an inhomogenous population [35]. Subpopulations of MSC with differential p53 activity may respond differently and less damaged MSC may compensate for damaged MSC with exhausted proliferation potential.
Taken together our data provide evidence that resistance of MSC to genotoxic damage does not generally rely on p53-dependent cell cycle regulation. Different p53-dependent and -independent mechanisms may contribute to damage-specific resistance of MSC. These data contribute to the understanding of MSC as life-long persisting niche cells with limited risk of malignant transformation.