Having Established The Fact That Senolytic Effect Of Ouabain Is Cell Type-dependent
Sep 08, 2022
Please contact oscar.xiao@wecistanche.com for more information
Disturbance of K*/Nat homeostasis caused by ouabain leads to different physiological reactions in control and senescent A549 and END-MSCs
Having established the fact that senolytic effect of ouabain is cell type-dependent, further we tried to elucidate what may underlie the revealed differences in responses of hMSCs and A549.To this end, we analyzed oxidative stress-induced senescence of END-MSCs and etoposide-induced senescence of A549more precisely. cynomorium benefits The main molecular mechanism of action of various cardiac glycosides is inhibition of Nat/K+-ATPase [44].Nat/K+-ATPase is a plasma membrane enzyme that pumps sodium out of the cell while pumping potassium into the cell against concentration gradients and thus participating in the maintenance of K+ and Na+homeostasis. Therefore, we asked whether there might be any differences in basal and ouabain-induced ion homeostasis between control and senescent cells of two types.

Please click here to know more
The disturbance in K+ and Na+ homeostasis caused by ouabain should ultimately lead to the dissipation of the plasma membrane potential. Thus, further, we used specific fluorescent probe DiBAC4(3) to measure the transmembrane potential of non-excitable cells. Increased fluorescence of this dye reflects membrane depolarization. desert hyacinth In each cell type tested ouabain induced predictable membrane depolarization, confirming ionic disbalance (Fig. 10a). Notably, we observed comparable levels of membrane depolarization in control and senescent END-MSCs, while senescent A549 was characterized by significantly more depolarized membrane compared to control A549 cells.
Other important consequences of ionic deregulation are changes in mitochondrial membrane potential (MP).
Thus, using fluorescent probe JC-1 we assessed MMP in control and senescent cells upon ouabain treatment. It should be specifically highlighted that senescent cells are commonly characterized by decreased MMP reflecting malfunctioning of mitochondria, alterations in energy metabolism, and increased intracellular reactive oxygen species levels[45]. As expected, MMP in senescent END-MSCs and A549 cells was lower than in the control ones,

though this decrease did not affect the viability of senescent cells (Figs. Id, 6a,10b). flavonoid extraction method pdf Similar to the above results, ouabain induced pronounced MMP depolarization in senes-cent A549 compared to their proliferating counterparts, while the effect on MMP of control and senescent END-MSCs was minimal (Fig.10b). Summarizing the results described within this part, we can conclude that blocking Na+/K+-ATPase by ouabain leads to the dramatic depolarization of plasma membrane and drop of MMP in case of senescent A549, contrarily to senescent END-MSCS.
Alterations in the mechanisms of K+ import mediate ouabain-resistance of senescent END-MSCs
To clarify mechanisms mediating ouabain-induced changes in membrane polarization and MMP between senescent A549 and END-MSCs, further, we employed an advanced bio-informatic approach. The aim of the analysis was to reveal possible transcriptomic features mediating cell type depend-ent ouabain-resistance or -sensitivity of senescent cells. In other words, we asked the question: how senescence of END-MSCs (ouabain-resistant cells) differs from senescence of A549 (ouabain-sensitive cells)? To answer this question, we utilized three independent RNA sequencing (RNA-Seq) datasets. The first one was RNA-Seq analysis

Fig.6 Validation of etoposide-induced A549 senescence model. Senescent A549 cells display a loss of proliferation, b acquire elevated cell size, c autofluorescence level, and d SA-b-Gal activity compared to the control ones. e Phosphorylation levels of p53 and Rb and expression levels of p21 and HMGBI proteins in control and senescent A549. The values presented are mean±SD. For multiple groups comparisons at a one-way ANOVA were applied, n=3, ns not significant,***p<0.001. For pair comparisons at b,c, and d Welch's test was used,n=3 for b and c,n=50 for d,***p<0.001. flavonoids Scale bars for images are 500 μm. GAPDH was used as loading control of control and senescent END-MSCs performed by us. The second one was the dataset for control and senescent A549 downloaded from GEO [43]. Of note, senescence-inducing conditions coincided with those applied in the present study and in the pilot studies of cardiac glycosides-induced seno-lysis. And the third dataset was for control and senescent IMR-90 obtained directly from the study on cardiac glycosides as senolytic compounds [25]. Importantly, IMR-90 was proved to be prone to ouabain-induced analysis. It should be noted that comparing only two datasets relevant for END-MSCs and A549 would ultimately lead to improper results, as in such analysis the sought-for difference between the senescence process in END-MSCs and A549 would be indistinguishable from the variations mediated by diverse technical batch effects (i.e.type of sequencing machine, the conditions of the sample ran). To minimize any batch effects, we compared one dataset for ouabain-resistant cells (END-MSCs)and two for ouabain-sensitive cells(A549, IMR-90). To validate the relevance of the further comparative analysis, we performed a Principal Component Analysis for the subset of genes related to the GO term "cellular senescence"(GO 0090398) based on the summary dataset containing samples from all three described experiments. Indeed, for each cell type samples clustered into two separate groups—the control and the senescent ones (Fig.10d).

cistanche can anti-aging
To perform an analysis of differentially expressed genes (DEGs) during senescence of ouabain-resistant and -sensitive cells, we generated the statistical model that summed up two predictors and their interaction. The schematic presentation of the analysis is displayed in Fig. 10c.In brief, the first predictor in the model divided all samples by control and senescence subgroups, the second—by ouabain resistance or ouabain sensitivity, and the last component in the model reflected the interaction of both predictors. The result of differential expression testing for the last variable is presented in Supplementary Table 1.
We then conducted gene set enrichment analysis (GSEA) in Gene Ontology (GO) terms for Biological Processes (BP) (Supplementary Table 2). Notably, among the significantly enriched processes, we found 'Potassium ion import' and 'Positive regulation of cation transmembrane transport to be up-regulated during senescence of ouabain-resistant END-MSCs as compared to senescence of ouabain-sensitive cells (Fig. 10e). The core enrichment gene lists related to these processes included KCNJ2, KCNJ14, KCNJ8, SLC12A7, SLC12A8, WNK4 which were significantly upregulated in senescent END-MSCs (Fig. 10f). The proteins encoded by KCNJ genes are inward-rectifier type potassium channels, which have a greater tendency to allow potassium to flow into a cell rather than out of a cell. hesperidin uses SLC12 a family of cation-you-pled chloride transporters. WNK4 gene encodes for serine/threonine kinase that acts as an activator of sodium-coupled chloride cotransporters and inhibitor of potassium-coupled chloride cotransporters. Together, these findings allowed suggesting that senes-cent END-MSCs can restore depleted intracellular K+ lev-els more effectively than senescent A549 cells and thus can manage with ouabain-induced ionic disbalance.
Senescent END-MSCs display elevated
apoptosis-resistance compared to the control ones, while senescent A549 does not
The role of intracellular K+in in the regulation of cell death is well established [46]. Revealed differences between senes-cent END-MSCs and A549 cells in the ability to maintain intracellular K+homeostasis prompted us to compare the overall stress resistance acquired during senescence of both cell types. It is known that exogenous stresses are accompanied by the disturbance in ionic homeostasis, leading to a rapid exchange of various ions, including K+ between the cell and its environment [46]. The outcomes of stressful influences are largely dependent on the ability of cells to restore the appropriate intracellular ionic balance. As shown in Fig.10e,f, senescence of END-MSCs was accompanied by the upregulation of K+ import suggesting that during

Fig.10 Ouabain-resistant senescent END-MSCs are able to restore disturbance of K+/Nat homeostasis caused by ouabain via effective K+ import, while ouabain-sensitive senescent cells lack this ability. a Membrane potential determination of control and senescent END-MSCs and A549 cells before and 24 h after ouabain treatment, using fluorescent probe DiBAC4(3). b Mitochondrial membrane potential evaluation of control and senescent END-MSCs and A549 cells before and 24 h after ouabain treatment, using fluorescent probe JC-1. c Design of the comparative bioinformatic analysis of three independent RNA-Seq datasets.d Principal Component Analysis for the subset of genes related to the GO term "cellular senescence"(GO 0090398) based on the combined data from three independent RNA-Seq data sets. e GSEA results for the differentially expressed genes between the cells resistant to ouabain-mediated analysis (END-MSCs) and ones sensitive to ouabain-mediated analysis (A549 and IMR-90) for 'Positive regulation of cation transmembrane transport' and 'Potassium ion import biological processes. f Heatmap reflecting the core enrichment genes for 'Positive regulation of cation transmembrane transport' and 'Potassium ion import' biological processes. Values are mean±SD. Statistical significance was assessed by the Student's t-test: ns not significant,*p<0.05,**p<0.01,***p<0.001
senescence development END-MSCs enhance their ability to restore K+ levels. Interestingly, we were not able to reveal a similar tendency for A549 senescence, thus, A549 seemed not to acquire any specific features related to the K* import during senescence progression. According to the literary data, stress-induced decrease in intracellular K* and inability to restore its level favor activation of caspases and nucleases and thus are proposed to be a pro-apoptotic factor [46].
In line with this suggestion, using bioinformatic analysis of the above RNA-seq datasets we revealed significant down-regulation of the processes related to apoptosis during senescence of END-MSCs as compared to A549 and IMR-90 (Fig. lla). Moreover, senescence of A549 and IMR-90 is characterized by significant up-regulation of pro-apoptotic genes, including BAD, BAX, BOK, BAK-1, and NOXA (Fig.11b). These data indicate that END-MSCs acquire apoptosis-resistant phenotype during senescence, while senescent IMR-90 and A549 became apoptosis-prone (Fig.11b). To strengthen this observation, we additionally analyzed microarray dataset for control and replicative senescent AD-MSCs that also pre-served ouabain-resistance during senescence as indicated above (Fig.4g). When analyzing for the expression levels of the apoptosis-related genes, we observed distinct apoptosis-resistant phenotype of senescent AD-MSCs similar to that of senescent END-MSCs(Supplemental Fig.S8a,b).

To verify the results of transcriptomic analysis, we estimated expression levels of genes related to apoptosis and K+ import in both types of senescent cells—END-MSCs and A549 by RT-PCR (Supplemental Fig. S9). Additionally, we analyzed the expression of the same gene list using other ouabain-resistant and ouabain-sensitive cellular models, doxorubicin-treated senescent DP-MSCs, and etoposide-treated senescent SK-Help, respectively (Supplemental Fig. S9). As shown in Supplemental Fig. S9, both types of ouabain-resistant cells-senescent END-MSCs and DP-MSCs displayed similar gene expression patterns, specifically KCNJ2, SLC12A and WNK4 participating in K+restoration were upregulated, proapoptotic BAX was downregulated, while antiapoptotic MCL-I was upregulated. Moreover, the expression pattern of the same genes differed from those observed for ouabain-sensitive A549 and SK-Hep1. These results confirm our transcriptomic data and reinforce the conclusion regarding varied apoptosis profiles acquired during the senescence of different cell types.
We next tested whether the observed distinctions in apoptosis background during senescence of END-MSCs and A549 may have an impact on the overall stress resistance of senescent cells. To do so, we assessed cell viability upon oxidative stress and staurosporine, typically applied to induce apoptosis. Indeed, senescent END-MSCs turned out to be significantly more resistant to both stresses compared to their control counterparts (Fig. 1lc). Contrarily, A549 cells demonstrated a vice versa situation, as senescent cells displayed a tendency to be more sensitive towards stress-induced cell death (Fig. 11d). This observation shifts the existing paradigm that enhanced death resistance is a common feature of any kind of senescent cells, demonstrating its obvious cell specificity.
Based on the above data, we assumed that reducing apoptosis "defense" of senescent END-MSCs should create favorable conditions for further analysis. In this regard, we focused on a well-known antiapoptotic member of BCL-2 family proteins—MCL-1, since its expression was upregulated in apoptosis-resistant senescent END-MSCs (Fig. 1lb and Supplemental Fig. S9a). To this end, we pretreated control and senescent END-MSCs with A-1210477 the specific MCL-1 inhibitor, and then applied cardiac glycosides to induce analysis. Of note, A-1210477 itself had no effect on the proliferation capacity of control cells and on the viability of senescent ones as indicated on the growth curves (Fig. 1le). Subsequent application of cardiac glycosides led to the decrease in the number of viable cells, how-ever the effect was more pronounced in senescent ESCs (Fig.11f). Therefore, inhibiting MCL-1 with A-1210477 predisposed senescent ESCs towards ouabain-induced death indicating for analysis. The latter evidence for apoptosis-resistance acquired during END-MSCs senescence to be an intrinsic barrier for effective analysis triggered by cardiac glycosides.
Discussion
Within the present study, we tested senolytic effects of cardiac glycosides, namely ouabain and bufalin, on hMSCs. Both compounds were recently identified to have selective cytotoxic effect on senescent cells (oncogene-/ther-apy-induced senescence models) of different origins, including primary cells IMR-90,HUVEC, ARPE-19,T/C-28 and cancer cells SK-Help,A549,SK-Mel-5,MCF7,HCT116, HaCat,H1299,U373-MG,H1755[25,26].Unexpectedly, we were not able to reveal any preferential cytotoxic effects neither of ouabain nor of bufalin on senescent END-MSCs compared to the control ones. Importantly, the absence of analysis was verified using hMSCs obtained from adipose tissue, dental pulp and Warton jelly, and various senescence models. The latter led us to the suggestion that resistance to cardiac glycoside-induced analysis might be the common feature of hMSCs. At the same time, we were able to induce apoptosis preferentially in senescent A549 and SK-Help, thus reproducing senolytic effect of cardiac glycosides described in the relevant articles [25,26]. With cells resistant and sensitive to ouabain-induced analysis (ouabain-resistant/ouabain-sensitive) available, we tried to elucidate fundamental causes underlying this difference. The common molecular mechanism of cardiac glycosides action is binding to the Nat/K+-ATPase and blocking its activity. By hydrolyzing ATP this enzyme ensures pump-ing Na+ out of the cell and importing K+ into the cells and thus maintaining physiological electrochemical gradient, ionic homeostasis, cellular pH, and cell volume that are essential for cell survival and functioning [47]. Therefore, we speculated that the senolytic ability of ouabain might depend on the severity of K+/Nat disbalance in the treated cells. In line with this suggestion, we revealed that ouabain induced more pronounced depolarization of plasma membrane and drop of MMP in ouabain-sensitive senescent A549.

Generally speaking, the impact of Nat/K+-ATPase inhibiting on cell survival was shown to be cell-type specific. For example, ouabain was shown to potentiate apoptosis in lymphocytes, Jurkat cells, and canine epithelial cells, while it failed to induce death in epithelial cells from rat aorta, rat cerebral granule cells, and porcine renal proximal tubular LLC-PK1 lymphocytes despite similar modulation of the cationic ratio [48-54]. One of the proposed mechanisms of proapoptotic ouabain action is depletion of intracellular that favors apoptotic shrinkage, activation of caspases, and initiation of apoptotic programming [46, 47]. Considering comparable K+ loss but opposite effect on death induction obtained for ouabain-resistant and ouabain-sensitive models, we suggested the differences in cation's compensatory systems. The drop in intracellular K+ level should essentially lead to the activation of the K+ import from the extracellular space to compensate for the lack of this cation. Therefore, the effectiveness of K+ restoration might underlie apoptosis predisposition upon ouabain treatment. To test this suggestion, we applied complex bioinformatic analysis comparing alterations in transcriptomic signatures during senescence of ouabain-resistant cells (END-MSCs) and ouabain-sensitive cells (A549 and IMR-90). We observed strong upregulation of the processes related to potassium ion import during END-MSCs senescence, while during senescence of oua-bain-sensitive A549 and IMR-90 these processes remained unchanged or even decreased. Based on that we can speculate that senescent END-MSCs can effectively cope with ouabain-induced K+ depletion via active cation importing systems, which prevents apoptosis induction. Contrarily, our-bain-sensitive senescent A549 and IMR-90 seem to be less able to overcome K+ loss, and thus are apoptosis-prone.In favor of this assumption, ouabain induced more significant plasma and mitochondrial membranes depolarization in senescent A549 cells, demonstrating pronounced ionic disbalance typical for dying cells.
Rather than being the unique feature of ouabain-induced apoptosis, a drop of intracellular K+ content is probably the common characteristic of apoptotic death triggered by various stresses, e.g. staurosporine treatment and oxidative stress [46]. Accordingly, decreased ability of senescent A549 to restore cytoplasmic K+ should ultimately lead to a decline in the overall stress resistance. However, this notion contradicts the modern definition of cell senescence, which states that senescent cells are highly resistant to apoptosis [16]. It should be specifically highlighted that enhanced apoptosis resistance formed the basis for analytics development [16, 17].
The first mention of the increased stress resistance of senescent cells dates back to 1995 when replicative senescent fibroblasts were found to be more resistant to serum withdrawal compared to the control ones [55]. This initial study was followed by a limited number of investigations indicating that senescent cells are more resistant to UV-light, staurosporine, thapsigargin, and other stresses than their proliferating counterparts [56, 57]. However, we are far from the first to raise the question of the enhanced apoptosis resistance as the general feature of senescent cells. Thus, in the review published in 2003, it was reported that "... apoptosis resistance is not a general feature of senescent cells, which may also be apoptotic prone depending on cell type and apoptotic stimuli..."[58]. Indeed, several data demonstrated higher sensitivity of senescent cells to stress-induced apoptosis than of control cells [59-61]. For example, senescent HUVECs were more prone to apoptosis induced by oxidized LDL or TNFa compared to control cells[61]. Despite the obvious controversy, the last doubt appeared in 2015 and sounded as follows: "It seems doubtful that global apoptosis resistance is a general feature of senescence cells"[62].In the same year, the first study uncovering analysis was published [16]. Within this study, the authors highlighted the increased expression of pro-survival gene networks in senescent cells that contributed to their enhanced resistance to apoptosis. Thus, initially, senolytics represented the drugs to target proteins that protected senescent cells from apoptosis. The discovery of analytics together with the impressive results of senes-cent cells' removal using INK-ATTAC transgenic mice triggered a flurry of studies searching for the new com-pounds with hemolytic activity [16-26,63]. Over the last two years, plenty of reviews on analytics was published [13-15,64-66]. However, since 2015 enhanced apoptosis resistance of senescent cells was not a matter of debate, though, in fact, it was not tested within these studies.
Interestingly, when comparing transcriptomic signatures of ouabain-resistant cells (END-MSCs) and ouabain-sensitive cells (A549 and IMR-90), we found that only senescence of END-MSCs was accompanied by the acquisition of notice-able anti-apoptotic profile. On the contrary, senescence of A549 and IMR-90 is characterized by significant up-regulation of pro-apoptotic genes, including BAD, BAX, BOK, BAK-1, NOXA, and so on. Importantly, these results were extended using other ouabatin-resistant cells DP-MSCs. On the one hand, these results provide an additional molecular explanation for the absence of ouabain-induced analysis in hMSCs, and, on the other hand, clearly demonstrate that A459 and IMR-90 become predisposed to apoptosis during senescence. Of note, our data regarding transcriptomic alterations in genes related to apoptosis in senescent IMR-90 coincided with the results presented, yet not discussed, in the article published by Guerrero et al., as the datasets for control and senescent IMR-90 cells originated from this study [25]. The precise analysis of the heatmap provided within this study revealed upregulation of BAX, BAD, BAD, BAKI, and NOXA in senes-cent IMR-90 as compared to the control cells. Furthermore, Baar et al. also revealed "unexpected" upregulation of proapoptotic and downregulation of antiapoptotic genes in senescent IMR-90, mentioning that senescent IMR-90 should be primed to undergo apoptosis [21].
In order to strengthen our observation, we compared the resistance of control and senescent END-MSCs and A459 to the most common apoptosis-inducing stimuli-oxidative stress and staurosporine. According to the results of our bioinformatic analysis, senescent END-MSCs turned out to be far more stress-resistant than control cells. At the same time, senescent A549 demonstrated the opposite reaction being slightly more susceptible than control cells to both types of stressful stimuli. Therefore, senolytic action of cardiac glycosides, at least for A549, might be mediated by the predisposition of senescent cells of this origin to apoptosis compared to their control counterparts, whereas the absence of ouabain-induced senolysis in hMSCs might be explained by the increase in overall apoptotic resistance during senescence of these cells. Moreover, weakening "antiapoptotic defense" by MCL-1 inhibition predisposed END-MSCs to senolysis induced by cardiac glycosides. Although we were able to achieve the desired senolysis of END-MSCs by inhibiting MCL-1, the strategy to modulate the expression of the concrete anti-apoptotic genes to predispose senescent hMSCs to senolysis does not seem very promising, since the expression dynamics of the concrete anti- or pro-apoptotic genes might differ between cells of various origins. We believe that acquisition of the so-called "anti-apoptotic" profile during cell senescence, rather than alterations in the concrete anti- or pro-apoptotic genes, is responsible for ouabain resistance, in this regard the possibility to switch the whole profile might seem more reasonable.
Our conclusions highlighting apoptosis-resistance acquired during hMSCs senescence to be an intrinsic barrier for effective analysis might be extended far beyond cardiac glycosides. Notably, senescent hMSCs were found to be resistant to other senolytic compounds. Indeed, by analyzing the existing evidence we found that various compounds with the stated senolytic activity, including navitoclax, nicotinamide riboside, danazol geldanamycin, ganetespib, fisetin, BCL-XL inhibitors, quercetin, etomoxir, antimycin A, FOXO4-DRI,17-DMAG turned to be ineffective for the targeted removal of senescent human preadipocytes (fibroblast-like precursor cells derived from human adipose tissue) and hMSCs from bone marrow [17-19, 36, 40]. The data regarding the absence of analysis in hMSCs somewhat contradict the inspiring results obtained using in vivo mice models, demonstrating positive outcomes upon systemic application of analytics [37, 38]. For example, it was shown that clearance of senescent salivary gland stem cells using ABT263 may prevent radiotherapy-induced xerostomia [38]. Furthermore, the removal of senescent bone marrow MSCs by quercetin improved bone marrow formation [37]. Importantly, it was shown that senescent mice and rat MSCs unlike hMSCs are responsive to analytics. For example, quercetin, quercetin-tin+dasatinib, and ABT263 significantly removed senescent mouse MSCs [16]. Also, 17-DMAG has been shown to greatly reduce senescent bmMSCs in a progeroid mouse model [19]. Therefore, the improvements in various organs and systems functioning described within these studies are speculated to be the consequence of the targeted removal of senescent mouse stem cells. However, these analytics have not yet shown any functional rejuvenation of hMSCs. Such an obvious distinction between the responsiveness of mice and human MSCs towards analysis suggests that improvements from senolytic therapies may not be conserved across the species. The results obtained for the seno-lytic compound UBX0101 can be considered a good illustration confirming the last statement. This agent has been shown to clear senescent cells in a mouse model of osteoarthritis, which significantly reduced pain and promoted repair of the damaged cartilage [20]. Unfortunately, UBX0101 failed to attenuate disease progression and pain in patients with moderate-to-serve painful osteoarthritis during phase II clinical trials [67-69]. Similar concerns are raised about senolytic combination dasatinib+querce-tin[68].
To sum up, two important conclusions flow out from the obtained data. The former demonstrated that cardiac glycosides are unable to clear senescent hMSC. The absence of analysis might be mediated by effective K+ cellular import and increased apoptosis resistance in senescent hMSCs. The latter is more fundamental and reveals that apoptosis resistance is not a general feature of senescent cells, as during senescence some cells acquire an 'apoptosis-resistant' phenotype, while others do not. Importantly, only apoptosis-prone senescent A549 cells could be effectively cleared by cardiac glycosides. Based on that we can speculate that the effectiveness of other senolytic approaches might depend on whether senescent cells are indeed apoptosis-resistant compared to their proliferating counterparts. Finally, though 'analytics' is a hot topic and a lot of inspiring data are published, conclusions regarding the effectiveness of analysis should be taken with caution as heterogeneity of cell senescence still remains a 'puzzle’.
This article is extracted from Cellular and Molecular Life Sciences (2021) 78:7757–7776 https://doi.org/10.1007/s00018-021-03980-x






