Interaction between ROS dependent DNA damage, mitochondria and p38 MAPK underlies senescence of human adult stem cells.

Human endometrium-derived mesenchymal stem cells (hMESCs) enter the premature senescence under sublethal oxidative stress, however underlying mechanism remains unknown. Here, we showed that exogenous H2O2 induces a rapid phosphorylation and co-localization of ATM, H2A.X, 53BP1 leading to DNA damage response (DDR) activation. DDR was accompanied with nuclear translocation of p-p53 followed by up-regulation of p21Waf1 and the permanent hypophosphorylation of pRb. Additionally, the increased p38MAPK/MAPKAPK-2 activation persisted in H2O2-treated cells. We suggest that both p53/p21/pRb and p38MAPK/MAPKAPK-2 pathways are responsible for establishing an irreversible cell cycle arrest that is typical of senescence. The process of further stabilization of senescence required prolonged DDR signaling activation that was provided by the permanent ROS production which in turn was regulated by both p38MAPK and the increased functional mitochondria. To reverse senescence, the pharmacological inhibition of p38MAPK was performed. Cell treatment with SB203580 was sufficient to recover partially senescence phenotype, to block the ROS elevation, to decrease the mitochondrial function, and finally to rescue proliferation. Thus, suppression of the p38MAPK pathway resulted in a partial prevention of H2O2-induced senescence of hMESCs. The current study is the first to reveal the molecular mechanism of the premature senescence of hMESCs in response to oxidative stress.


INTRODUCTION
Cellular senescence defined as an irreversible proliferation arrest promotes age-related decline in mammalian tissue homeostasis [1]. At present, investigation of this phenomenon becomes more and more widespread due to the following reasons: firstly, senescence is thought to contribute to a multiple agerelated pathologies [2,3], and secondly, senescence acts as a tumor suppressor mechanism that is able to block proliferation of incipient cancer cells [1,4]. Similar to other types of normally proliferating cells that are characterized by a finite lifespan (Hayflick limit) human mesenchymal stem cells undergo the replicative senescence after a fixed number of cell divisions [5].

Research Paper
Moreover, the recent findings have revealed that human mesenchymal stem cells may respond to a variety of subcytotoxic stresses (UV-, γ-radiation, H 2 O 2 , histone deacetylase inhibitors, etc.) by induction of premature senescence [6,7,8].
Oxidative stress has been shown to play an important role for the development of aging and age-related diseases [9]. According to the free-radical theory of aging, reactive oxygen species (ROS), including the oxygen singlet, the superoxide anion (O2 .-), the hydroxyl radical (OH . ) and hydrogen peroxide (H 2 O 2 ) might be the candidates, which are responsible for cellular senescence. Being membrane permeable and long-lived molecule, H 2 O 2 can directly affect the cellular DNA, inducing both single-and double-strand breaks (SSBs and DSBs, respectively) [10]. DNA damage, in turn, triggers a specific DNA damage response (DDR), which involves the following events -(1) activation of any sensor kinases (ATM, ATR, DNA-PK), (2) phosphorylation of adaptor protein 53BP1, and (3) formation of the discrete foci, containing phosphorylated histone H2A.X and p53BP1 [11]. Finally, DDR activation leads to cell cycle arrest via activation of p53/p21 [12] and/or p16/pRb pathways [13,14]. Today it is generally accepted that both the replicative and stress-induced senescence are the outcome of DDR [15].
Conversion from proliferative arrest to irreversible senescence, a process named geroconversion, is driven in part by growth-promoting pathways in particular mammalian target of rapamycin (mTOR) which is mostly responsible for loss of RP (replicative/regenerative potential) and hypertrophy [16,17]. Inhibitors of mTOR such as rapamycin [18] and hypoxia [19] can suppress geroconversion, maintaining quiescence instead. Previously, the mTOR was the only pathway known to be involved in acquiring classic markers of a senescent phenotype, including cyclin D1 accumulation. Recent studies have been revealed an additional MEK/ERK pathway that is required for the acquisition of at least one hallmark of senescence: hyperaccumulation of cyclin D1 [20]. Furthermore, it was shown that p70S6K, a crucial substrate of mTOR, and MEK play different roles in geroconversion [21].
According to recent publications, a so-called feedback loop between the permanent DDR activation and the increased ROS production is necessary for development of senescence [22]. The main effectors of DDR, p53 and p21, were shown to be involved in regulation of ROS generation, leading to enhanced intracellular ROS during the establishing of senescence [23,24]. In the senescent cells, elevated ROS can cause a direct DNA damage and the persistent DDR activation, thereby forming a feedback loop.
It is well-established that the senescent cells are characterized by increased ROS levels. Taken into consideration that the overwhelming majority of intracellular ROS are of mitochondrial origin, it is reasonable to posit that the elevated ROS production might be caused by alteration in mitochondrial function during senescence. Mitochondria are the intracellular organelles responsible for ATP synthesis through the coupling of oxidative phosphorylation to respiration in mammalian cells. Currently, there are different points of view regarding age-related changes in mitochondrial physiology. Several authors considered cellular senescence to be accompanied by mitochondrial dysfunction defined by the decline of mitochondrial membrane potential (MMP). MMP decline leads to respiratory-chain defects and thus to enhanced ROS production [22,25]. On the contrary, the others hypothesized the absence of defects either in electron transport system or in oxidative phosphorylation during senescence [26,27]. In this case, ROS levels are elevated due to (1) the growing number of functional mitochondria and, (2) the age-related alterations in mitochondrial coupling that correlates with the increase in mitochondrial membrane potential. Importantly, both hypotheses manifested the involvement of mitochondria in the aging process via a relevant contribution to intracellular ROS generation.
p38, a member of the family of mitogen-activated protein kinases (MAPKs) is activated by cellular stresses including ROS, UV-and γ-radiation, and proinflammatory cytokines [28]. In response to stress factors, p38 MAPK (hereafter p38) can rapidly phosphorylate and activate MAP kinase-activated protein kinases (MAPKAPKs), particularly MAPKAPK-2 (hereafter MK-2) that is a direct target of p38 [29]. The p38 effector kinase lies downstream of the MKK3/6 activator kinases, whose activity can be regulated by stress-sensitive apoptosis signal-regulating kinase-1 (ASK1) [28]. p38 plays an important causative role in cellular senescence induced by oxidative stress, radiation, genotoxic agents, Ras overexpression [30]. p38 has been demonstrated to participate in feedback relationships during senescence as, on the one hand, it can function as mediator of ROS signaling and, on the other hand, can directly phosphorylate p53 [22]. The significance of p38 MAPK pathway in stress-induced senescence was investigated predominantly in the cultures of either human fibroblasts [22,31] or transformed cells [32]. However, there is no information available as to whether the functional p38 is required for premature senescence of human endometrium-derived mesenchymal stem cells (hMESCs).
hMESCs are a relatively new source of adult stem cells intensively studied over the past decade. The fact that hMESCs isolation does not require invasive and traumatic procedures facilitates their use in regenerative medicine. So to date, promising results concerning the experimental and clinical application of these cells for treatment of heart failure, myocardial infarction, diabetes, stroke, Parkinson's disease, multiple sclerosis, Duchenne Muscular Dystrophy and infertility were obtained [33,34,35]. Despite the different nature of these disorders, namely oxidative stress is well known to play an essential role in their progression. The goal of www.impactaging.com the present study was to clarify the underlying mechanisms of both induction and further maintenance of premature senescence in hMESCs subjected to oxidative stress.

RESULTS
Recently, we have provided the reliable evidence that hMESCs undergo the premature senescence in response to the sublethal concentration of H 2 O 2 [36]. H 2 O 2treated hMESCs were permanently arrested, lost Ki67 proliferative marker, and exhibited senescent phenotype, including cell hypertrophy and increased SA-β-Gal activity, indicating that the cells were driven into stress-induced premature senescence. However, the molecular mechanism of senescence induction in hMESCs under oxidative stress is far from being elucidated.

DDR activation in response to a rapid accumulation of exogenous H 2 O 2 in hMESCs
H 2 O 2 by conventional diffusion may easily pass through the membrane into the intracellular space, causing damage to lipids, proteins, and DNA [37,38]. To ascertain the dynamics of H 2 O 2 penetration into the cells in our experimental conditions (200 µM H 2 O 2 , 1 h), in H 2 DCFDA-stained cells, the changes in H 2 O 2 concentration were monitored for 60 min by either FACS or the confocal microscopy. Of note, the confocal microscopy enabled to estimate the individual changes of fluorescence occurring in each of the cells selected to test ( Fig. 1 A, B), whereas FACS analysis evaluated the average fluorescence per cell ( Fig. 1 C). Despite the apparent distinctions in fluorescence levels presented in Figures 1A and 1C, resulting from the behavior of individual cells in the heterogenous population, the dynamics of H 2 O 2 diffusion evaluated by both methods were alike. H 2 O 2 treatment of cells led to a rapid elevation of intracellular ROS levels, peaking at 15 min and returning progressively to the baseline by 30 min in majority of cells, indicating that exogenous H 2 O 2 is almost completely utilized by cells during 1 h treatment. Among intracellular ROS, which are able to induce DNA damage, H 2 O 2 is known to provoke an appearance of both SSBs and DSBs that can trigger DDR [38]. Generally, DDR is characterized by activation of ataxia-telangiectasia mutated kinase (ATM) and formation of DNA-damage foci, containing γH2A.X and p53BP1 in chromatin, surrounding DSBs. Presently, it is generally accepted that ƴH2A.X, as well as p53BP1 are recognized as the reliable markers of DSBs [11]. To examine the possibility of DDR activation in H 2 O 2 -treated hMESCs, the functional status of key proteins involved in DDR was estimated. Immunofluorescent analysis with the use of specific antibodies against pATM, γH2A.X, p53BP1 revealed a rapid ATM phosphorylation (within 5 min after beginning of treatment) (Fig. 2 A) and further simultaneous phosphorylation of 53BP1 and H2A.X in 15 min (Fig. 2 B, C). Moreover, we observed colocalization of pATM with either ƴH2A.X or p53BP1 in 60 min after beginning of treatment ( Fig. 3 A, B). The fast H 2 O 2 -induced ATM phosphorylation peaking at 30 min was confirmed by Western blot analysis (Fig. 4 A).  In previous study, we have shown that an irreversible cell cycle arrest is the main marker of H 2 O 2 -induced premature senescence of hMESCs, however the signaling pathways providing the cell cycle arrest require the detail investigation. Senescence program is thought to be developed as result of DDR, leading to functional activation of the p53/p21 pathway, which can establish and maintain the growth arrest [12]. In order to determine whether cell cycle block may be realized via p53/p21 pathway in H 2 O 2 -treated hMESCs, we first investigated the functional status of p53 protein.
Western blot analysis with using rabbit polyclonal antibodies against phospho-p53 at Ser15 revealed a rapid (within 10 min) p53 phosphorylation, gradually increasing during 60 min of H 2 O 2 treatment (Fig. 4 B). At the same time, translocation of phospho-p53 into the nuclei and its partial colocalization with pATM was observed (Fig. 4 C). In H 2 O 2 -treated cells, enhanced p53 phosphorylation was also detected at 7 h posttreatment, however, over next 8 days it was dramatically diminished (Fig. 8 C). Remarkably, the decrease of the functional activity of p53 had no effect on elevated p21 induction during the entire observation period (Fig. 8 D).
It is known that p53 activated acts as a transcription factor, inducing expression of p21 which may mediate the initiation of the cell cycle arrest by inhibiting various cyclin-dependent kinases (CDK) that contribute  Immunofluorescent analysis with the use of specific antibodies against pATM, γH2AX, p53BP1 revealed co-localization of pATM with either ƴH2AX (A) or p53BP1 (B) in 60 min after beginning of H 2 O 2 treatment. DAPI was used as nuclear stain (blue). Images are taken at magnification X100. Scale bar is 20 µm and valid for all images.
www.impactaging.com cell cycle phase progression. Therefore, we next examined mRNA and protein expression levels of p21. H 2 O 2 promoted a significant elevation in mRNA and protein expression of p21 already at 7 h post-treatment ( Fig. 4 D, E). An inducible expression of p21 was upregulated, at least, during 7 days with following decline to insignificant, but not the control levels, which persisted up to 21 days. The elevated p21 expression was accompanied with the cell cycle arrest at the same time (data not shown). Retinoblastoma protein (pRb) whose activity is regulated by elevated p21 plays a crucial role for establishing the growth arrest. It is known that pRb in active hypophosphorylated state halts cell proliferation by suppressing the activity of E2F transcription factor that regulates cell cycle progression. To examine the functional status of pRb during establishing senescence, we performed monitoring the kinetics of pRb activation in H 2 O 2treated hMESCs. As expected, beginning 7 h post H 2 O 2 treatment, no pRb phosphorylation was observed in the senescent cells, in contrast to the control proliferating cells, which displayed the high levels of pRb phosphorylation ( Fig. 4 F). Collectively, our findings demonstrate that the p53/p21/pRb signaling pathway leading to the growth arrest is required to drive the premature senescence and apparently to maintain the long-term senescent state in hMESCs.

An interplay between enhanced ROS levels and prolonged DDR activation
As mentioned above, the exogenous H 2 O 2 induced a strong increase in intracellular ROS levels within 1 h of cell treatment ( Fig. 1 A, C) and accordingly triggered a premature senescence of hMESCs. To find out whether the intracellular ROS levels can be modulated during the senescence development, DCF fluorescence intensity was measured in H 2 O 2 -treated cells over the next 9 days. Surprisingly, on day 5 post-treatment, the senescent cells were characterized by strongly increased DCF fluorescence, consistent with higher levels of intracellular ROS that remained elevated further over 9 days (Fig 5A,  B). These results were in agreement with the continuous elevated levels of intracellular peroxides measured by DHR123 in the senescent cells (  Previous studies have reported that there is the functional link between enhanced ROS production and DDR activation during the development and stabilization of senescence [22]. Therefore, we further characterized the functional status of DDR in the senescent cells by testing ATM, H2A.X and 53BP1 for their phosphorylation and an intracellular localization using the fluorescent microscopy. Remarkably, on 5 days post-treatment all of proteins tested remained in an active state and mostly co-localized in so-called senescence-associated DNA-damage foci (SDFs) (Fig. 5 E, F). It should be noted that, in the senescent cells, enhanced ROS production and DDR activation has been contemporized. Together, these observations allow us to suspect that enhanced intracellular ROS could be responsible for long-term DDR activation.

An increase in mitochondrial activity in the senescent hMESCs
Although we have shown that permanently enhanced ROS are typical of the senescent state of hMESCs, nevertheless the actual reason of long-term ROS production remained unclear. We hypothesized that this phenomenon might be associated with the significant modulation of the mitochondrial function in time that www.impactaging.com was postulated to be one of the main contributory factors in senescence [39]. In order to examine this suggestion, the cells were assessed for cellular peroxide production, mitochondrial mass and mitochondrial membrane potential (MMP) by DHR123, NAO and Rho123 staining, respectively. Nonfluorescent dye DHR123 selectively accumulates in mitochondria, where it can be oxidized by mitochondria-derived ROS to a fluorescent rhodamine derivative. As seen in Fig. 5 C, at 24 h post H 2 O 2 treatment the cellular peroxides levels were almost 2-fold higher than in the control cells and then gradually enhanced for 7 days, indicating that, in the senescent cells, there are permanently elevated ROS levels derived from mitochondria. Next, to determine whether H 2 O 2 may cause the proliferation of mitochondria, NAO dye to monitor the mitochondrial mass was used. The relative NAO intensity of H 2 O 2treated cells gradually increased in time and, on 7 days post-treatment, it was found to be 2.5-fold higher than that of the control cells ( Fig. 6 A, B). These results indicate that H 2 O 2 can promote an increase in the number of mitochondria in hMESCs in a timedependent manner. To analyze, whether the extra mitochondria in H 2 O 2 -treated cells were functional at the same time, MMP of cells was measured. In treated cells, the relative intensity of Rho123 fluorescence was substantially higher than that of the control cells (Fig. 6 C, D). Importantly, the increase in MMP correlated with corresponding increase in the mitochondrial mass over the entire observation period. In addition, both characteristics also correlated well with elevated cellular peroxide production.
Taken together, these results clearly indicate that oxidative stress induced by the sublethal H 2 O 2 led to an increase in the amount of functional mitochondria. We assume that an increased mitochondrial activity may be responsible, at least in part, for long-term ROS production observed in the senescent hMESCs.

The role of p38 in the regulation of H 2 O 2 -induced premature senescence of hMESCs
First, we examined whether p38 is activated by sublethal doses of H 2 O 2 in hMESCs. As shown in Fig. 7 A, 200 µM H 2 O 2 induced a significant increase in p38 phosphorylation within 1 h of treatment, whereas in untreated cells p38 phosphorylation was undetectable. Interestingly, H 2 O 2 maintained the elevated p38 phosphorylation up to 8 days without affecting steadystate protein levels of p38 (Fig. 7 B). Phosphorylation of MK-2, a natural substrate of p38, was detected at the same time (Fig. 7 C). www.impactaging.com To explore the role of p38 in regulation of premature hMESCs senescence, we employed SB203580 (hereafter SB), a specific inhibitor of the p38 MAPK pathway. SB is a small molecule that displaces ATP from the ATP-binding pocket of p38, thereby preventing from phosphorylation of p38 targets, in particular MK-2, without preventing p38 phosphorylation itself [40]. As expected, in H 2 O 2treated cells, p38 phosphorylation was unaffected by SB (Fig. 7 B), nonetheless, MK-2 phosphorylation was abolished throughout the 8-days period of the experiment (Fig. 7 C).
Selective inhibition of p38 kinase activity with SB prevented the increase of the size of H 2 O 2 -treated cells (Fig. 8 B) but just slightly reduced their SA-β-Gal activity (data not shown), indicating some modulation of the senescence phenotype of H 2 O 2 -treated cells. Because the growth arrest was shown to be the major mechanism of the integral growth-inhibitory effect of H 2 O 2 in hMESCs under our experimental conditions [36], we next checked the role of p38 in the regulation of cell proliferation. As presented in Fig. 8 A, blocking of p38 with SB led to marked increasing the number of proliferating cells compared with H 2 O 2 -stimulated cells. In order to relieve the proliferation block of the senescent cells to a great extent, in the separate experiments, SB was added at 24 or 48 h post H 2 O 2 treatment unlike the routine immediate adding after H 2 O 2 removal. Interestingly, in this case SB had no effect on the proliferative status of senescent cells, albeit was still able to reduce ROS levels (data not shown). Consequently, the p38/MK-2 inactivation could in part prevent the loss of the proliferative potential of pre-senescent hMESCs. In this case the senescence program has already been initiated however the proliferative arrest could yet be reversed. By contrast, suppression of p38/MK-2 activity in the senescent arrested cells was insufficient to resume the proliferation. These findings demonstrate that p38, acting as a negative regulator proliferation of hMESCs in response to oxidative stress, is required for establishing premature senescence, whereas its inhibition may at least in part rescue the cells from senescence induction.
Remarkably, in H 2 O 2 -treated hMESCs, p38 inhibition did not affect phosphorylation status of p53 (Fig. 8 C) and did not prevent p21 protein induction over the  www.impactaging.com entire observation period (Fig. 8 D). These results suggest that p38/MK-2 and p53/p21 signaling pathways can act independently during establishing and maintaining of the premature senescence of hMESCs. On the other hand, suppression of p38/MK-2 in H 2 O 2treated cells noticeably elevated the pRb phosphorylation levels, indicating an inactivation pRb (Fig. 8 E). These findings correlate well with an increase in the proliferative potential of cells after common treatment with H 2 O 2 and SB as compared with H 2 O 2 -treated cells (Fig. 8 A).
To confirm that the effect of SB was specific to p38, we used another p38 inhibitor with an unrelated chemical structure, BIRB796 [41]. According to preliminary results, the observed effect of BIRB796 at concentration of 5µM was similar to the effect of SB, partially preventing both H 2 O 2 -induced growth inhibition and the increase in the size of H 2 O 2 -treated cells, thereby demonstrating that p38 participates in the establishing premature senescence in hMESCs induced by H 2 O 2 .

p38 MAPK implication in feedback loop via ROS
As described above, the senescent hMESCs are characterized by persistently elevated ROS levels. We presumed that long-term activation of p38 might be involved in regulation of ROS production. To test this idea, in the SB-treated and untreated senescent cells, the intracellular ROS production measured by H 2 DCFDA staining was evaluated. SB treatment of H 2 O 2 -stimulated cells led to a dramatic drop in the intracellular ROS levels in any of time points tested (Fig. 9 A). Taking into account the fact that increased mitochondria could mediate the elevation of ROS levels in the senescent hMESCs, we next examined whether p38 is able to affect the mitochondrial function. SB treatment of the senescent cells equally reduced the cellular peroxide levels and MMP compared with H 2 O 2 -treated cells as evaluated by DHR123 and Rho123 staining, respectively ( Fig. 9 B, C). In addition, in SB-treated cells we observed the similar decrease of mitochondrial mass measured by NAO staining ( Fig. 9 D). Overall, the results obtained indicate p38 implication in continued ROS production mediated by increased mitochondrial function in the senescent hMESCs. www.impactaging.com

DISCUSSION
In the current study, we have examined the molecular mechanism of premature senescence of hMESCs in response to oxidative stress. According to the modern concepts, the process of stress-induced senescence comprises two sequential stages -establishing and maintaining (stabilization) that can be regulated by different mechanisms, depending upon the specific stimuli used, the cell context and other factors. Understanding of the interplay between various signaling pathways that provide the control either stage would have significant therapeutic implications and should be highly useful to determinate the strategy for the reversal of cellular senescence. DDR activation induced by ionizing radiation, chemotherapeutic drugs or oxidative stress has been investigated in different stem cell types, including bone marrow-derived hMSCs [8], hematopoietic stem cells (HSC) [14] and embryonic stem cells [42]. In the literature, the similar studies performed on hMESCs have not been described thus far. The results presented herein demonstrate that, in hMESCs subjected to sublethal oxidative stress, H 2 O 2 generated a persistent DDR signaling associated with DNA double-strand breaks (DSBs) which are a signal for activation of ATM and downstream pathways, leading to cell cycle block, as well as the accumulation of DNA foci marked by γH2A.X and phospho-53BP1. Activation of DDR signaling can be a trigger for switching on senescence and it is essential for establishing and maintaining senescent phenotype of cells [15].
We have reported earlier [36] that, in hMESCs, the exogenous H 2 O 2 caused an irreversible arrest of cell cycle with predominant accumulation of cells in G0/G1phase. To find out the molecular mechanism, triggering the cell cycle arrest in H 2 O 2 -treated hMESCs, first of all we focused on p53-mediated signaling pathway which may lead to the cell cycle block. It is well documented that the p53/p21 pathway is critical for establishing the replicative senescence of human cells [12,13], as well as the premature senescence of hMSCs [8] and HSC [14]. However, little is known about which signaling pathways are responsible for the induction of the premature senescence of hMESCs under oxidative stress. In agreement with a canonical p53/p21 model, our results demonstrate that DDR-activated p53 upregulated the CDK inhibitor p21 that, in turn, prevented the phosphorylation and inactivation of pRb. Besides, pRb activity may be controlled by another CDK inhibitor, p16 (INK4a). Previous studies suggested that p16 was crucial for long-term maintaining of senescence of both human fibroblasts [1,13,43] and HSC [14]. Interestingly, in these cases p16 was expressed much later than p21, forming a second barrier to prevent the cells from cell cycle re-entering. In contrast, our preliminary data indicate that upregulation of p16 occurs solely at initial stage of senescence (within 1 h) but not at delayed time period. These findings support the possibility that p16/pRb pathway, in addition to p53/p21, is responsible for establishing the growth arrest, preventing entering the cells into S-phase.
In consideration of strongly decreased p53 activity observed in senescent cells from 3 days, we can speculate that p53 is critical for rather establishing the senescence growth arrest than prolonged maintaining the senescent state of hMESCs. Unlike p53, an inducible p21 expression was persistently upregulated throughout experiment (up to day 21) and was accompanied with the cell cycle arrest [36]. Accordingly, the elevated p21 induction was indispensable to promote the senescence, as well as to maintain this state in hMESCs in response to H 2 O 2 . Previously, it was reported that the activated checkpoint kinase 2 (Chk2) can induce p21 transcription in the absence of functional p53 and that this contributes to Chk2-mediated senescence [44]. It was attractive to suppose that, in our experimental conditions, Chk2 permanently activated by ATM also is able to mediate a long-term p21 induction during senescence. Taken together, our findings definitely demonstrate that, in H 2 O 2 -treated hMESCs, the senescence program is triggered by DDR signaling, activation of which leads to an irreversible cell cycle arrest through p53/p21/pRb pathway.
Further, we tested the role of stress-activated kinase p38 (in complex with MK-2, a direct downstream target of p38) as the most prominent mediator of stress-induced cellular senescence. MK-2 is known to be a negative regulator of cell cycle progression because it is directly responsible for phosphorylation-dependent inactivation of members of the Cdc25 family of phosphatases, which are positive regulators of Cyclin/CDK complexes. Finally, pRb was found to promote stress-induced growth arrest as a downstream molecule of p38 [45]. At present, MK-2 is recognized as a new member of the DNA damage checkpoint kinase family that functions in parallel with Chk1 and Chk2 to integrate DNA damage responses and cell cycle arrest [46]. Previous studies reported that the p38 pathway is implicated in H 2 O 2induced senescence of human fibroblasts however the data presented by various groups were controversial. Several authors observed in H 2 O 2 -treated cells the continuous p38 activation [45] while the others demonstrated that transiently elevated p38 kinase www.impactaging.com activity was reversible down-regulated after H 2 O 2 removal [47]. Moreover, no activation of p38 was detected in H 2 O 2 -treated fibroblasts, regardless of endogenous p38 expression [48]. By contrast, our results indicate that elevated p38/MK-2 activation in response to H 2 O 2 was persisted for a long time that suggests the importance of p38/MK-2 pathway in the control of both induction and a long-term maintaining of hMESCs senescence. The fact that the permanent p38/MK-2 activation was accompanied by pRb inactivation argues in favour of the p38/MK-2/pRb pathway that is likely to be mediated by Cdc25 family members. Interestingly, during treatment of hMESCs with H 2 O 2 , p38 activation was accompanied with increasing in phosphorylation of ASK1 at Thr845 (results not presented) that is correlated with ASK1 activity. Therefore, at this stage, we cannot exclude the possibility that ASK1 is involved in up-regulation of p38.
It is noteworthy that inactivation of p38/MK-2 induced by SB did not affect the functional status of p53 and p21, suggesting that p38/MK-2 and p53/p21 pathways are uncoupled however can cooperate to induce an irreversible proliferative arrest, and further to maintain the premature senescence of hMESCs. Generally, this suggestion is supported by a recent report demonstrating that p38 participates in oxidative stressinduced senescence via an alternative ATMindependent pathway, implicating lamin B1 accumulation [49].
The pharmacological inhibition of p38 activation may be considered as a possible strategy for senescence prevention [45,49]. According to our results, selective inhibition of p38 kinase activity with SB abrogated H 2 O 2 -induced cell enlargement and flattened morphology, but did not produce any significant effect on SA-β-Gal activity. At the same time, SB treatment of hMESCs allowed to avoid an irreversible cell cycle arrest in response to H 2 O 2 however the recovery of proliferation was incomplete. This may point at the probable dissociation of hallmarks of senescencesenescent morphology, RP and SA-β-Gal staining. Likewise, elimination of cyclin D1 (a universal marker of cellular senescence) by specific inhibitors of the MEK/ERK pathway did not affect at least three classical hallmarks of senescence: loss of RP, senescent morphology and SA-β-Gal staining [20]. The findings that inhibition of p38 partially suppressed the H 2 O 2induced senescent phenotype of hMESCs, as well as prevented the proliferation arrest indicate that activation of p38 contributes to H 2 O 2 -induced cellular senescence. A plausible explanation for the partial effect of SB on growth arrest is that SB being a specific inhibitor of p38α and p38β cannot suppress the redundant γ and δ isoforms of p38 [50]. In addition, SB was reported to produce antiproliferative effect related to inhibition of pRb phosphorylation [51]. Thus, suppression of p38 kinase activity can at least in part rescue stressed hMESCs from the cell cycle arrest and entering premature senescence induced by H 2 O 2 . It is of noted, that the permanent growth arrest correlated with ROS accumulation during development of senescence.
Our attempts to find out the possible reasons for maintaining the H 2 O 2 -induced senescence of hMESCs led to the observation that there is the interplay between permanently elevated ROS and the persistence of DDR signaling. In fact, the senescent cells displayed the persistent accumulation of DNA damage foci marked by p53BP1 and ƴH2A.X associated with pATM, as well as continuously increased levels of both intracellular ROS and mitochondrial peroxides. These findings are in line with previous study, presuming that the feedback loop between DDR and ROS production is necessary and sufficient to maintain senescent growth arrest during establishment of irreversible senescence [22]. Consistent with supposed mechanism, the senescent cells persistently accumulate senescence-associated DNA damage foci (SDFs), which contain proteins associated with DNA damage, particularly ƴH2A.X and p53BP1. The prolonged DDR activation results in upregulation of p53 and p21 that may induce the increase in intracellular ROS levels. All together, the intracellular ROS are able directly to damage DNA and thus sustain DDR in an active state.
To clarify the mechanism which may regulate the feedback loop, we utilized an inhibitor analysis of longterm ROS production. In senescent hMESCs, the specific inhibition of p38/MK-2 activity by SB had a pronounced negative effect on the intracellular (cytosolic) ROS production while the mitochondrial ROS production was diminished just in part. Similarly, mitochondrial mass and mitochondrial membrane potential were partially decreased. It is important to note that in senescent hMESCs preserving the metabolic activity we revealed the significant increase in the amount of functional mitochondria which might be responsible, at least in part, for long-term ROS production. These findings collectively allow speculation of p38/MK-2 involvement in modulation of intracellular ROS that are critical for maintaining feedback loop during senescence of hMESCs. Although our results suggest that the activated p38/MK-2 complex may regulate ROS generation via functional mitochondria, it is most likely that such regulation is indirect. Accumulating evidence pointed to an important role for TGFβ/TGFβ receptor and GADD45 www.impactaging.com in mitochondrial ROS production mediated by p38 [22,28].
In conclusion, the present study is the first to elucidate the molecular mechanism of premature senescence of hMESCs under oxidative stress. The induction of senescence includes a prompt activation of response to DNA damage induced by H 2 O 2 and following signal transduction through p53/p21 and p38/MK-2 pathways which are necessary and sufficient to establish the irreversible cell cycle arrest that is typical of senescence. We believe the prolonged induction of p21 as well as elevated activation of p38/MK-2 also might be indispensable to maintain persistent proliferative block in senescent cells. Additionally, p38 which may regulate both intracellular and mitochondrial ROS production is possibly involved in senescence stabilization via the feedback loop that provides sustained activation of DDR signaling (Fig. 10).
The main properties of hMESCs -the capacity for selfrenewal, multilineage differentiation and noninvasive isolation procedures bring them to the cutting edge of regenerative medicine. Although hMESCs transplantation for treatment of heart failure, myocardial infarction, Duchenne Muscular Dystrophy has been successfully applied, it should be taken into account the possibility of premature senescence of these adult stem cells in stress conditions with following loss of the regenerative potential and, consequently, of the capability to regenerate the injured tissues. Understanding the mechanism of premature senescence of hMESCs induced by H 2 O 2 should provide more effective strategies in transplantation of these cells into the recipients with age-related disorders inherently associated with increased levels of oxidative stress.
RT-PCR assay. To analyze gene expression, total cellular RNA was isolated with RNesy Micro Kit (Qiagen) according to manufacturer's instructions. cDNA synthesis was performed with 1 g of total RNA using RevertAid H Minus First Strand cDNA Synthesis Kit (Fermentas) according to manufacturer's instructions. Specific genes were amplified by Taq DNA polymerase (Fermentas) with C1000 TouchThermal Cycler amplifier (Bio-Rad Laboratories). The program was described earlier [30]. Primers p21Waf1/Cip1 and beta-actin were obtained from SYNTOL (Russia). The electrophoresis of amplified products was performed in 2% agarose gel with TAE buffer and ethidiumbromide. 100 kb DNA ladder (Fermentas) was used as molecular weight marker. Amplified products were visualized in UV-light (302 nm) with transilluminator and registered with a digital Canon camera.
FACS analysis of cell viability and cell size. Adherent cells were rinsed twice with PBS and harvested by trypsinization. Detached cells were pelleted by centrifugation. Finally, detached and adherent cells were pooled and resuspended in PBS. 50 µg/ml propidium iodide (PI) was added to each sample just before analysis and mixed gently. Samples were analyzed on a Coulter EPICS XL Flow Cytometer (Backman Coulter). The cell size was evaluated by cytometric light scattering of PI-stained cells with using Win MDI program version 2.8. To discriminate the live and dead cells, two-parameter histogram was used (FL4LOG vs. FSLOG). Analysis of each sample (at least 10,000 cells) was performed for 100 sec with high sample delivery.
Statistics. All data are presented as the mean and standard deviation of the mean from at least three separate experiments performed. Statistical differences were calculated using the Student's t-test and considered significant at *, § p< 0.05; **, § § p< 0.005; *** , § § § p<0.001. www.impactaging.com