Monovalent Ions and Stress-Induced Senescence in Human Mesenchymal Endometrial Stem Cells


 Monovalent ions are involved in growth, proliferation, differentiation of cells as well as in their death. This work concerns the ion homeostasis during senescence induction in human mesenchymal endometrium stem cells (hMESC): hMESCs subjected to oxidative stress (pulse H2O2 treatment) enter the premature senescence accompanied by persistent DNA damage, irreversible cell cycle arrest, cell hypertrophy, lipofuscin accumulation, enhanced β-galactosidase activity. Using flame photometry to estimate K+, Na+ content and Rb+ (K+) fluxes we found that during the senescence development in stress-induced hMESCs, Na+/K+pump-mediated K+ fluxes are enhanced due to the increased Na+ content in senescent cells, while ouabain-resistant K+ fluxes remain unchanged. Senescence progression is accompanied by a peculiar decrease in the K+ content in cells from 800-900 µmol/g to 500-600 µmol/g. Since cardiac glycosides are offered as selective agents for eliminating senescent cells, we investigated the effect of ouabain on ion homeostasis and viability of hMESCs and found that in both proliferating and senescent hMESCs, ouabain (1 nM-1 µM, 24-48 h) inhibited pump-mediated K+ transport (ID50 5x10-8 M), decreased cell K+/Na+ ratio to 0,1-0,2, however did not induce apoptosis. Comparison of the effect of ouabain on hMESCs with the literature data on the selective cytotoxic effect of cardiac glycosides on senescent or cancer cells suggests the ion pump blockade and intracellular K+ depletion should be synergized with target apoptotic signal to induce the cell death.


Introduction
Monovalent ions are involved in the control of cell growth, proliferation, and death. Unlike Ca 2+ , which is an important signaling player in the cell, the role of monovalent ions such as K + , Na + and Cl − , in these cellular processes is not well understood. It is commonly suggested that ion transporters and ion channels are involved in the intracellular signaling network, and K + , Na + , and Cl − are important for setting the membrane potential and the intracellular pH and Ca 2+ concentrations during cell cycle progression.
For example, the concentration of Na + in cells can affect cell cycle progression by pH i changes: it has been shown that Na + /H + exchanger activity regulates G 2 /M progression by increasing pH i which, in turn, affects cyclin B1 expression and cdk2 activity [1][2][3] . It is assumed that cellular Cl − is involved in the hyperpolarization of cell membrane during G 1 /S transition 4 .
In addition to signaling role, the transmembrane movement of monovalent ions may be important in context of cell volume control. Cell division depends on cell volume increase, and monovalent ions (such as K + , Na + , Cl − ) play important role in the regulation of cell volume [5][6][7][8][9] . Cell volume is considered as an important component in the regulation of cell cycle progression. In studies of transformed cells of different origin and human mesenchymal endometrial stem cells we revealed signi cant changes in cell K + content and K + in uxes which were related to cell accumulation in G 1 phase of cell cycle and proliferation slowing [10][11][12] . Analysis of K + content changes in cell cultures with different proliferative status and in human blood lymphocytes, stimulated to growth, suggested that K + as the main intracellular ion might be involved in regulation of cellular water content during cell transit from quiescence to proliferation 13.14 . On the whole, being elements of cellular "housekeeping", main monovalent permeable ions are able to modulate intracellular signaling and provide a speci c intracellular ion context for development of peculiar cellular response.
In recent years, ion channels and transporters, Na + , K + -ATPase in particular, are suggested as anti-aging targets. It has been revealed that cardiac glycosides selectively kill senescent cells via apoptosis 15,16 . It is proposed that cardiac glycosides damage harmful cells by blocking Na + , K + -ATPase, however the mechanism underlying their selective killing effect has not been established.
Cellular senescence is de ned as an irreversible cell cycle arrest that can be triggered in cells in response to various intrinsic and extrinsic stimuli, as well as developmental signals [17][18][19][20] . Senescence plays physiological role during normal cell development, it underlies stem cell aging and is also proposed as a tumor suppression mechanism [21][22][23][24] . Senescence markers such as DNA damage, increased expression of the cell cycle inhibitors (p16, p21, p53) as well as distinct phenotypic alterations, including chromatin remodeling, metabolic reprogramming, characteristic messaging secretome can be used to identify senescent cells, but these markers are neither speci c nor universal. Despite profound metabolic changes and impaired protein synthesis, alterations in mitochondria and lysosomes physiology, senescent cells remain metabolically active for a long time. Whether monovalent ions participate in senescence development as well as in maintaining the viability of senescent cells has not been investigated. To date, there are few studies on changes in ion content during senescence. Higher concentration of intracellular Ca 2+ is observed in senescent cells compared to proliferating cells, and chelation of Ca 2+ with BAPTA partly prevents senescence [25][26][27] . There are data on the activation of the Na + /H + exchanger during stressinduced proliferation arrest 3 . The membrane potential of both cells and mitochondria appears to be lower in senescent cells than in cyclic cells 16 . Using uorescent probes, an increased content of K + and Na + was detected in the senescent human lung broblast IMR90 15 . Direct analysis of K + and Na + in senescent cells has not yet been performed.
In this study, we investigated changes in monovalent ion homeostasis during the development of premature senescence. Using ame photometry to assess K + and Na + content as well as Rb + (K + ) uxes, we found peculiar changes in K + content in cells during oxidative stress-induced senescence of human endometrial mesenchymal stem cells (hMESCs). The signi cance of ouabain-induced disorders of ion homeostasis (in particular, cellular K + ) for the death of senescent hMESCs is also discussed.

Results
H 2 O 2 -stressed hMESCs enter premature senescence. It has recently been shown that hMESCs subjected to sublethal oxidative stress enter irreversible cell cycle arrest and exhibited a senescent phenotype including persistent DNA damage foci, cell cycle arrest mediated by the p53/p21/Rb pathway, cell hypertrophy, enhanced β-galactosidase staining 32 . In the present study, the growth of hMESCs, that were treated with the hydrogen peroxide (200 µM H 2 O 2 for 1 h) and then returned to H 2 O 2 -free serumcontaining DMEM, stopped by the 3d day and displayed typical senescence markers (Fig. 1A). Under these conditions, the cells enlarged as evidenced by increased forward scattering (Fig. 1B). Also, the enhanced auto uorescence of stressed cells indicated accumulation of lipofuscin, which is considered as a marker of cellular senescence 33,34 (Fig. 1C). As can be seen in Fig. 1D, stress-induced arrested hMESCs displayed the decreased mitochondrial membrane potential as evaluated by increased TMRM uorescence signal 31 . Finally, the SA-β-galactosidase activity, a common marker of senescence, was increased in arrested hMESCs (Fig. 1E, F). After H 2 O 2 treatment hMESCs retain high viability. As evaluated by FACS analysis, at the 5th day the percentage of viable cells in stressed population was 91±7 (n=3) instead of 96±5 (n=3) in control proliferating cells. Taking into account all these data we considered hMESCs, which received a short H 2 O 2 pulse, as a model for studying the ion homeostasis during premeture senescence progression.
Cell K + and Na + content during senescence progression of stressed hMESCs. As recently shown and con rmed in this study, short oxidative stress (1 h treatment with 200 µM H 2 O 2 ) leads to a decrease in K + content and an increase in the Na + content in proliferating hMESCs 12 (Fig. 2A). As a result of reciprocal changes in the content of K + and Na + in H 2 O 2 -treated cells, the intracellular K + /Na + ratio decreased from 7-8 to 3-3.5 which indicates disordered ionic gradients during oxidative stress. After replacing the medium with a fresh medium that did not contain H 2 O 2, the ionic gradients were restored within a day; however, the K + content in stressed cells was lower than in control proliferating cells ( Fig. 2A).
We then assessed the cellular monovalent ions during culturing stressed hMESCs. In our experiments, during the rst 2 days, the stressed cultures continued to grow, though at a slower rate than the control cultures, and then stopped growing (Fig. 1A). Under these conditions, the K + content in both proliferating and stressed cells decreases, which is associated with an increase in cell density and a decrease in the proliferative activity of the culture 12 . By the 8th day, the K + content decreases to a constant level, but in stressed cells it is lower (652±41 µmol/g) than in control cyclic cells (795±39 µmol/g) ( Fig. 2A). Thus, stress-induced cessation of cell proliferation is accompanied by a decrease in the K + content calculated for the protein content in the cell.
To assess the content of cations in a cell, in our studies the measured amount of cations was normalized to the mass of cellular protein in the same sample. In cell biology, such evaluation of intracellular ions is widely used. Indeed, there are signi cant di culties in assessing intracellular ion concentrations (ion content per cell water content) because of the di culties in measuring the volume and water in adherent cells. The most adequate method for assessing the water content of cells in suspension culturesmeasuring the buoyant cell density -is not applicable for monolayer cell cultures. It should be noted here that there are attempts to assess the ionic and other physiological parameters of cells in monolayer cultures after their treatment with a trypsin-containing medium. Our experience has shown, that cells shortly treated with trypsin (0.05%) have an increased Na + content (while maintaining a high K + content) and a K + /Na + ratio close 1 (Table 1). It also turned out that a low K + /Na + ratio persists if the cells are washed from trypsin in a fresh medium and then kept in suspension for up to 1-2 hours (longer observations were not carried out). It is noteworthy that only after attachment to the adhesive surface the low Na + content and the high K + /Na + ratio are restored in these cells (Table 1). Based on these data, we believe that the ame emission method used to measure the intracellular ion content is the most appropriate for studying the ionic homeostasis of monolayer cultures. The method allows one to determine both the content of basic cations in cells and the ion in uxes, using analogous cations (for example, Rb + to assess the in ux of K + ). It is also important that normalization of the amount of ions (in our case, K + ) to the amount of protein in each sample allows us to obtain data that contribute to understanding the mechanism of K + participation in cell growth and proliferation.
Senescent hMESCs remain viable in culture for a long time. We asked if the late stressed cells retain a high K + /Na + ratio. As can be seen in Fig. 2A, in long-term cultures (up to 22 days), senescent hMESCs had a K + content (562±39 µmol/g), comparable to that in early senescent cells. After oxidative stress, the content of intracellular Na + decreases and does not change for 8 days, however during senescence progression, the content of Na + in cells increases from 120±10 µmol/g to 160±19 µmol/g (Fig. 2C). Taken together, these data indicate that, during long-term culture, senescent hMESCs maintain the high K + /Na + ratio typical for functionally active animal cells. A peculiar feature of senescent hMESCs in comparison with proliferating cells is a lower K + content per cellular protein.
Na + ,K + -ATPase-mediated Rb + (K + ) transport is increased during senescence progression in stressed hMESCs. Short-term Rb + uptake was used to assess changes in K + transport during senescence. In proliferating hMESCs, ouabain-inhibitable Rb + in ux, mediated by the Na + ,K + -ATPase pump, accounts for more than half of the total Rb + ux. In the rst days after stress, the ouabain-inhibitable Rb + ux decreases, as well as in proliferating cultures, the in ux decreases (Fig. 2B). A decrease in pumpmediated K + transport in growing cell culture is associated with density-dependent inhibition of cell proliferation 12 . In stress-induced hMESCs, a decrease in ouabain-inhibitable Rb + in ux also re ects the transition to cell cycle arrest and the cessation of cell proliferation.
We then investigated the transport activity of the Na + , K + -ATPase pump in hMESCs during senescence development and compared ouabain-inhibitable Rb + in uxes in proliferating and early stressed hMESCs with that in late stressed hMESCs. As can be seen in Fig. 2B, in late stressed cells, the ouabain-inhibitable Rb + uptake was increased accounting 65±4 µmol/g, 30 min (n=4) instead of 40±4 µmol/g, 30 min (n=6) in early stressed cells. These data indicate elevated pump-mediated K + transport in established senescent hMESCs. To determine whether the observed increase in pump activity is proportional to changes in intracellular Na + concentration, we compared pumping rate coe cients calculated as the ratio of ouabain-inhibitable Rb + uptake to intracellular Na + content during senescence development [35][36][37] . It turned out that the rate coe cients do not differ for early and late stressed cells. Thus, the increased ouabain-inhibitable K + transport in senescent cells is not associated with a change in the intrinsic properties of Na + /K + -ATPase pump, but is a consequence of ux-concentration relations in existing ion pumps and is probably associated with an increase in cellular Na + in late senescent hMESCs.
Comparison of ion changes during the growth of hMESC culture with those in senescent cells shows that stress-induced cell cycle arrest and senescence progression are accompanied by a decrease in K + content per gram of cellular protein mass. The senescence progression is also associated with elevated cell Na + content and increased pump-mediated K + in uxes.
Cell K + and Na + content and Rb + uxes in ouabain-treated cycling and senescent hMESCs.
Having data on ion homeostasis of hMSCs, both proliferating and senescent, we analyzed the cytotoxic effect of ouabain on these cells. Increasing evidence suggests that cardiac glycosides are capable of inducing apoptosis and selectively killing senescent, but not cycling cells 15,16 . We asked whether ouabain-induced changes in cellular K + and Na + content could contribute to apoptotic cell death in senescence, and examined the relations between changes in ionic homeostasis in the presence of ouabain and ouabain's ability to kill senescent hMESCs. So far, direct measurements of intracellular K + and Na + content have not been performed with ouabain-treated senescent cells compared with cycling cells.
First, we con rmed that ouabain at high concentrations stops cell proliferation. Starting from concentration 10 −7 M by 24 h ouabain inhibits the growth of hMESCs cultures causing S-G 2 /M delay in cell cycle (Fig. 3A, B). As expected, after incubating with ouabain for 24 h, there occur reciprocal changes in cell K + and Na + contents (Fig. 3C, D). With increasing the ouabain concentration from 10 −9 to 10 −7 M the cell K + decreased from 879±30 µmol/g (n=4) to 110±12 µmol/g (n=3) in control (young) cells and from 674±48 µmol/g (n=3) to 96±17 µmol/g (n=3) in senescent cells with the most signi cant decrease between 5x10 −9 to 10 −8 M concentrations in both cell populations. Simultaneously with a decrease in cellular K + content, the Na + content in ouabain-treated cells increased reaching the highest level with 10 −7 M ouabain (Fig. 3B). Ouabain induced dose-dependent decrease in ouabain-inhibitable Rb + in ux with the same IC 50 (5x10 −7 M) for both young and senescent hMESCs (Fig. 3E, F). Within the wide range of ouabain concentrations, ouabain-resistant Rb + in ux did not change (Fig. 3G). Thus, ouabain already at a concentration of 10 −7 M completely stops ion pumping thus leading to disruption of cell K + /Na + gradients.
Next, we tested how the long-term ouabain affects the viability of young and senescent hMESCs. As evaluated by FASC analysis, 10 −6 M ouabain led to an increase in number of PI-stained proliferating (young) cells (8,66% PI+ and 14,43% PI+ after 1 and 2 days) whereas in the population of ouabain-treated senescent hMESCs the number of PI-stained cells remained as low as in control (5,02% PI+ and 6,07% PI+) (Fig. 3H). Annexin V test carried out after 10 −6 M ouabain treatment did not show apoptosis induction in proliferating (young) and senescent cells (Fig. 3I). Taken together, the above data indicate that in both proliferating and senescent hMESCs, ouabain causes the same disturbances in monovalent ion homeostasis and does not induce apoptosis.

Discussion
In present study, we investigated the ion homeostasis during stress-induced cell cycle arrest and premature senescence development in hMESCs. Using ame photometry to estimate cellular K + and Na + content and transmembrane Rb + (K + ) uxes we found that senescent hMESCs maintain high K + /Na + ratio typical for functionally active animal cells. The senescence progression is accompanied by elevated Na + content in cells and increased pump-mediated K + in uxes. Stress-induced cell cycle arrest does not affect passive, ouabain-resistant K + uxes across plasma membrane. A peculiar feature of senescent in comparison with proliferating hMESCs is lower K + content per cell protein mass.
Premature senescence caused by various stresses in cells is associated with the cessation of cell proliferation. The lower ratio of the K + content to protein mass in senescent cells is in good agreement with our previous nding that a decrease in this index correlates with a decrease in cell proliferation 12,13,38,39 . Recently, when studying the activation of human T lymphocytes, we also found that the transition of cells from quiescence to proliferation is necessarily accompanied by an increase in the content of both K + and water per g of protein so that during the growth of lymphocytes and an increase in their volume, the intracellular concentration of K + remains constant 14 . These data allow us to suggest that K + may be involved in maintaining cell growth and proliferation as an intracellular ion, which participates in the regulation of cell volume by adjusting the water balance of cell.
In the present study, by analogy with our previous experimental data when reliable measurements of K + , water and volume were carried out simultaneously on proliferating and resting cells cultured in suspension 14,40−42 , relying on a theoretical analysis of ion and water balance in animal cells 5,9,43−46 , and also taking into account that K + is the main cation compensating intracellular anions we assume that the lower ratio of cell K + content to cell protein mass can be interpreted as lower water content in senescent hMESCs. To nd out if this is really so, it is necessary to take reliable measurements of the volume of senescent cells. Various experimental approaches to measuring cell volume have been developed; however, this is still not an easy task to measure the volume of adherent cells in growing monolayer cultures 47 . Nevertheless, we can discuss our assumptions on the signi cance of the ratio of K + content to protein in cells, taking into account the recent theoretical studies of cell volume regulation by Model and Petruccelli 45 . Using the fundamental basic equations of the osmotic balance of the cell and the electroneutrality of the system, the authors investigated the relationship between cellular organic matter, cell volume and inorganic ions, which made it possible to estimate how the water content in the cell changes with the change in the content of small intracellular cations. Applying this approach it can be concluded that the higher the intracellular content of K + and Na + , the higher the water content in the cell.
As found in the present study stress-induced senescence is accompanied by a decrease in the sum of K + and Na + , mainly due to a decrease in the K + content in cells, which indicates a lower water content in nonproliferating senescent cells.
There are few experimental data on changes in cell hydration state associated with a change in the proliferative status of cells. Based on some evidence of higher embryonic and cancer cell hydration, increased cell hydration has been proposed as an important factor in cell malignant growth and carcinogenesis 48,49 . The relationship between hydration and cell proliferation may re ect the effect of macromolecular crowding on cellular metabolism: macromolecular crowding reduces metabolic processes and cytoplasm uidity [50][51][52][53][54][55][56][57] . It has also been hypothesized that macromolecular crowding plays a role in signaling volume perturbations 58-62 . Finally, measurements of water content in aging erythrocytes have suggested that cell water loss and macromolecular crowding may be a common mechanism of cellular senescence 63 .
To further investigate the role of monovalent ions in the maintenance of senescence, we decided to test the participation of the Na + /K + pump in the survival of senescent hMESCs. To access further the involvement of monovalent ions in maintaining senescence we wanted to test the role of Na + /K + pump in survival of senescent hMESCs. This problem is discussed in view of proposed anti-aging effects of cardiac glycosides: cardiac glycosides have been reported to selectively kill senescent cells via apoptosis and can be used as senolytics in the treatment of aging-related diseases 15,16 . In our experiments, in proliferating and senescent hMESCs, ouabain inhibits pump-mediated K + transport in a dose-dependent manner, which leads to profound changes in ionic homeostasis, such that Na + substitutes K + inside the cell and after a 24 h incubation with 0.1 µM ouabain, in cells of both types, intracellular K + becomes only 90-100 instead of 800-900 µmol/g protein in untreated cells. Despite such serious disturbances in ion homeostasis, as assessed by PI/Annexin staining, long-term ouabain (1 µM, up to 48 h) did not induce death and apoptosis in senescent hMESCs.
Earlier it was reported about high apoptosis resistance of human mesenchymal stem cells 32,64−66 .
However, the high resistance of senescent hMESCs to ouabain is not consistent with proposed selective cytotoxic (senolytic) activity of cardiac glycosides in relation to senescence 15,16 . Note that there is a lot of data in the literature that cardiac glycosides also kill high proliferating cancer cells, but not cells that normally cycle 67-72 . The question arises as to what may underlie the selective toxic effect of cardiac glycosides on senescent or cancer cells, and whether disturbances in cell ion homeostasis (in particular, K + content) can be involved in the anti-aging or anticancer effects of these drugs.
The Na + /K + -ATPase pump, a cellular target of cardiac glycosides, is a major player in glycoside cytotoxic effects. Depletion of Na + /K + -ATPase by RNA interference inhibited glycoside-induced apoptosis in cells 72 .
Several studies correlate the expression levels of the alpha-1 Na + , K + -ATPase to the susceptibility of cells towards cardiac glycosides. Indeed, the expression level of Na + , K + -ATPase is dependent on cell proliferative status: quiescence is characterized by a lower expression of α1-and β1-Na + , K + -ATPase subunits, while the transition of cells to proliferation is accompanied by increasing their transcription and the synthesis of new α1-Na + , K + -ATPases 29,73,74 . As a result, highly proliferating cancer cells are characterized by a higher Na + , K + -ATPase expression and may be more sensitive to glycosides than quiescent, differentiated or normally cycling cells. As for senescent cells, there are no studies indicating an increased expression of Na + ,K + -ATPase during senescence: the expression level of α1-Na + ,K + -ATPase did not differ between cycling and arrested senescent lung cancer cells A549 cells 16 . In the present study, no differences were found in the relationship between K + in ux and ouabain dose, or in the rate coe cient for the Na + , K + -ATPase pump in proliferating and senescent hMESCs. Thus, it is unlikely that the Na + , K + -ATPase pump in senescent cells is more sensitive to cardiac glycosides. In addition, the glycoside concentrations which are toxic to senescent cells but do not to normal cycling cells fall into a narrow window 16 . These features of drugs indicate a rather low toxic speci city of glycosides in relation to senescent cells.
The Na + , K + -ATPase pump is the primary ion transport system responsible for the maintaining the Na + and K + concentrations and the resting membrane potential in the cell [75][76][77][78][79] . Animal cell stability and survival depend on the operation of the Na + /K + pump through the pump-leak mechanism 5 . Impermeant anionic molecules in cells establish an unstable osmotic condition (the Donnan effect), which is counteracted by the operation of Na + /K + pump, thus creating an asymmetric distribution of Na + and K + and preventing excessive water in ux. In most animal cells, the Na + /K + pump blockade leads to cell swollen, membrane disruption and necrotic death. To understand the mechanism of selective toxic action of glycosides on dangerous cancer or senescent cells, it is important to know what type of death they cause.
Accumulating evidence indicate that in senescent and cancer cells, cardiac glycosides induce apoptosis or "mixed" cell death with signs of apoptosis and necrosis or autophagy, causing mitochondria dysfunction 15,69-73,80−84 . In this context, K + transport plays a central role in mitochondrial physiology. Under normal physiological conditions, the high electric potential difference generated by proton pump across the inner mitochondrial membrane is used to make ATP and determines also the inwardly directed K + ux into matrix; changes in matrix volume due to this K + in ux and concomitant water entry are compensated by activity of the electroneutral K + /H + antiporter (so called K + cycle) 85 . Disruption of K + cycle inevitably leads to disturbances in mitochondria function and contribute to intrinsic apoptosis induction. It is important, that the activities of K + /H + antiporter and K + channels are regulated by the apoptotic Bcl-2 proteins: pro-apoptotic Bax can interacts with K + channels in the inner mitochondrial membrane, thus blocking K + in ux 86-90 .
The question is what might be the role of cellular K + depletion due to ion pump inhibition in mitochondria dysfunction. Induction of apoptosis by valinomycin alone suggests that disruption of K + homeostasis may be su cient to induce apoptosis in cancer cells 81 . Excessive K + e ux and intracellular K + loss are key early step in apoptosis induction 7,91−96 . It is noteworthy that this early loss of K + occurs simultaneously with the loss of water by cells and does not lead to a decrease in the concentration of K + in cells 40 . At later apoptosis, however, before the release of cytochrome c, a signi cant decrease in the content of K + and Cl − assayed by X-ray microanalysis or cryo-correlative microscopy was found in the cytoplasm and mitochondria [97][98][99] . Late stages of apoptosis are also associated with cell shrinkage, a decrease in water content in cells and a simultaneous decrease in the cell K + concentration 100 . Altogether, these data suggest that in cells with the Na + /K + pump turned off, a sever decrease of cytoplasmic K + could promote intrinsic apoptosis by disruption of the mitochondria ion and volume homeostasis.
Apoptosis as a programmed cell death is controlled by a well-orchestrated genetic program so that the cell fate is dependent on the balance between pro-and anti-apoptotic members of Bcl-2 family proteins, and the ability of cardiac glycosides to induce apoptosis depends on whether the pro-apoptotic proteins are expressed in cells treated with glycoside. It has been revealed that in cancer cells, cardiac glycosides induce apoptosis by down-regulating the anti-apoptotic proteins Bcl-XL and Bcl-2 as well as Mcl-1 and increasing the pro-apoptotic proteins Bid and Bax 68-70,83,84,101−104 . In senescent human lung broblast IMR90 cells, cytotoxic ouabain increased the pro-apoptotic NOXA protein 15 . Based on these studies, we can conclude that if a cell is proned to apoptosis (i.e. pro-apoptotic proteins are in proper functional position in cell) cardiac glycosides can contribute to the apoptosis.
Cardiac glycosides do induce apoptotic death in human embryonic stem cells (hESCs) but not in hESCderived mesenchymal stem cells and human bone marrow mesenchymal stem cells (hBMMSCs) 66 . As reported recently and also shown in our study, cardiac glycosides do not induce apoptosis in both hMESCs and their senescent partners 105 . Obviously, the high resistance of hMESCs to cardiac glycosides is provided by a highly expressed anti-apoptotic program in these cells 32,64,65,106,107 . Taken together, the above data suggest that the depletion of cellular K + is necessary event in cardiac glycoside-induced apoptosis; however, a separate apoptotic signal is required to kill a cell, which acts together with low cellular K + .
In summary, oxidative stress-induced senescence in hMESCs is associated with speci c changes in cell ion homeostasis, which primarily concerns intracellular K + . In senescent cells, pump-mediated K + transport is enhanced due to the increased Na + content, while passive ouabain-resistant K + in uxes remain unchanged. In the course of senescence development, cellular K + content is decreases and in established senescent cells, the ratio of K + content to cellular protein becomes lower that of in cycling cells which may indicate a decrease in the hydration of senescent cells. Evaluating the cytotoxic selectivity of cardiac glycosides toward senescence, we conclude that K + being a key ion in mitochondria physiology is involved in intrinsic apoptotic events in cells treated with cardiac glycosides. However, to kill senescent cells, the ion pump blockade and intracellular K + depletion should be synergized with the target apoptotic signal. Given the pleiotropic effects of cardiac glycosides, their use as selective drugs for eliminating the dangerous cells warrants further study: cardiac glycosides as inhibitors of the ion pump that ensures the stability and survival of all animal cells are able to kill not only harmful tumorous or senescent cells but also functionally important differentiated cells.

Methods
Cells and experiment design. The procedures involved human cells were performed in accordance with the standards of the Declaration of Helsinki (1989) and approved by the Institute of Cytology Ethics Committee. Written informed consent was obtained from all patients who provided tissue. and re-cultured in fresh complete culture medium (Burova et al., 2013). Cells were analyzed either immediately after H 2 O 2 shock or at selected time points after prolonged cultivation, depending on the aim of the study. To assess the effect of ouabain on the viability of hMESCs, young proliferating cells (passage 10) and cells whose senescence was induced by peroxide after an additional 5 days of cultivation were used.
Analysis of cell K + and Na + content and K + in ux. Measurements of ions were performed essentially as described previously 13,29 . To estimate K + in ux Rb + was used as the physiological analog of K + . RbCl ( nal concentration 5 mM) was introduced into the culture medium for 20 min. To evaluate the Na + , K + -ATPase K + in ux, prior to RbCl, 10 −4 M ouabain (Sigma-Aldrich, USA) was added to culture medium. Then, cells were rapidly washed 5 times with ice-cold isotonic MgCl 2 and cations were extracted with 1 ml of 1% trichloroacetic acid (TCA). TCA extracts were analyzed for Rb, K and Na by emission ame photometry on a Perkin-Elmer AA 306 spectrophotometer. TCA precipitates were dissolved in 0.1 N NaOH and analyzed for protein by Lowry procedure. Ouabain-sensitive Rb + uptake was calculated as the differences between the mean values measured in samples incubated with and without ouabain. The intracellular ion content was expressed as amount of ions per amount of protein in each sample analyzed.
FACS analysis of cell viability, cell proliferation, mitochondrial health and apoptosis. For FASC analysis, adherent cells were rinsed twice with PBS and harvested by trypsinization. Detached cells were pelleted by centrifugation and suspended in PBS. Samples were analyzed with CytoFLEX ow cytometer (Beckman Coulter, Brea, CA, USA) or CytoFLEX S ow cytometer (Beckman Coulter, Brea, CA USA).
To determine cell viability propidium iodide (PI) staining was used. PI is excluded by viable cells but can penetrate cell membranes of dead cells and intercalates into double-stranded nucleic acids. 50 µg/mL of PI was added to each sample just before analysis. At least 3000 events were usually collected as the main cell population. Triplicate counts were obtained for each procedure. Representative PI versus FSC dot plots allowed us to distinguish between PI-negative "live" cells and PI-positive "dead" cells. Senescence-associated β-galactosidase assay. Cells expressing senescence-associated β-galactosidase (SA-β-gal) were identi ed using a β-galactosidase staining kit (Cell Signaling Technology, Beverly, MA, USA) according to the manufacturer's recommendations. The kit allows determining the activity of βgalactosidase at pH 6.0, which is well detected in senescent cells.
Statistical Analysis. All data are presented as the mean with standard error of the mean from at least three independent experiments. Statistical signi cance was assessed using either ANOVA-Tukey test in case of multiple comparisons or Student's t-test in case of pair comparisons. Control cells (1) were from con uent proliferating culture of hMSECs. Cellular K and Na was measured by emission ame photometry as described in Methods. Figure 1 Oxidative stress induces premature senescence in hMESCs. Cells were seeded in 3-cm dishes (10x104 cells per dish) and at the second day after plating, some cultures were subjected to 200 μM H2O2 for 1 h with the following H2O2 replacement and cell cultivation under normal conditions. (A) Growth curves of control (1) and H2O2-treated (2)    Ouabain treatment lead to severe changes in cell K+, Na+ content and in Na+/K+ pump-mediated transport, but did not induce apoptosis in both young (proliferating) and senescent hMESCs. increases the number of PI-positive "dead" cells in young hMESCs (H) but did not induce apoptosis both in young and senescent hMESCs (I). Flow cytometry assay for Annexin-V was performed on cells treated with 10-6 ouabain for 1 day. The bar-graph indicate the mean percentage ± SD of PI+ cells, n=3, t-test, *p < 0.05 vs the control, ns -insigni cant.