Selective targeting of cancer and senescence via shared metabolic shifts extends lifespan of old mice

DMA is a new, effective, and safe senolytic, while DCA, metformin, their mix, or ABT-263 alone are not

To validate the reported dose-limiting toxicity of ABT-263, namely, thrombocytopenia [23, 44, 45], mice were administered conventional doses of the drug, and platelet numbers were quantified 18 hours later by flow cytometry. Mice treated with the conventional published dose of 50mg/kg of ABT-263 showed a significant reduction in platelet count as compared to controls (Figure 1A, 1B) [6].

Figure 1. Development of DMA. (A) Representative flow cytometry plot of platelet populations from mice treated with 50 mg/kg ABT-263 and stained with CD41. (B) Normalized platelet counts showing that ABT-263 caused a significant reduction in platelet number (mean ± SEM, unpaired t-test). (C) Representative image of Sa-β-gal of non-senescent (NS) and damage-induced senescent (DIS) IMR-90 cells (Scale bar 200μm), and quantification of Sa-β-gal+ (mean ± SD, unpaired t-test). (D) Normalized cell ablation from remaining nuclei with increasing concentrations of metformin (Met), DCA, and ABT-263 (see methods). (E) Nuclear stain of the remaining fibroblasts after 24-hour treatment of DMA, 1μM ABT-263, and vehicle control. Apoptotic nuclei are highlighted with red arrows (Scale bars 500μm). (F) Normalized platelet counts after administration of 50 mg/kg ABT-263, DMA, or solvent control, showing that DMA did not cause a significant reduction in platelet numbers (mean ± SEM, one-way ANOVA, Bonferroni’s multiple comparisons test). (Statistical significance is denoted as: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

We hypothesized that the addition of Met and DCA could allow lowering ABT-263 doses and thereby mitigate thrombocytopenia, while still achieving senolytic capacity. Accordingly, senolytic effectiveness of individual drugs, ABT-263, Met, and DCA, as well as their different combinations and concentrations was examined. Damage-induced senescent cells (DIS) were established by UV irradiating IMR-90 human lung fibroblasts, and senescence was confirmed via Sa-β-gal staining (Figure 1C). The DIS and non-senescent IMR-90 Cells were exposed for 24 hours to the three drugs, combined at eight different concentrations: 0 to 50 mM for DCA and Met, and 0 to 50 μM for ABT-263.

First, these studies demonstrated that neither the individual drugs alone nor DCA + Met were effective senolytics at the doses examined (Supplementary Figure 1A). Secondly, individually DCA, Met, and ABT-263 were non-toxic to dividing cells at those doses (Figure 1D). Lastly, the three-drug combination containing a low, non-cytotoxic dose of ABT-263 had robust senolytic capacity: DMA: 5 mM DCA, 5 mM Met, and 1μM ABT-263 (Supplementary Figure 1B). DMA ablated about 60% of DIS cells, without compromising the viability of healthy control cells (Supplementary Figure 1C). Moreover, DMA treatment resulted in higher numbers of apoptotic senescent nuclei, suggesting a direct effect on senescent cells (Figure 1E). Importantly, mice that were administered with DMA in vivo had significantly lower thrombocytopenia, compared to the animals receiving the conventional dose of ABT-263, (Figure 1F). At roughly one-tenth dose of ABT-263, i.e., its concentration in DMA, this drug does not induce thrombocytopenia [46] but when used alone at this dose, it has poor senolytic efficacy [47].

To confirm the robustness of combining DCA + Met with a BCL-2 family inhibitor, the combination was tested with an ABT-263 analog, ABT-737, on two types of senescent cells: replicative and damage-induced IMR-90, (Supplementary Figure 2A–2D). Replicative senescence was confirmed via SA-β-gal (Supplemental Figure 2A), and cell viability was assessed using the CellTiter-Blue assay. As published, ABT-737 (10 μM) markedly reduced the viability of both DIS and RS cells (Supplementary Figure 2B) [22]; DCA alone, Met alone, and their combination did not diminish the viability of DIS, RS, or healthy cells as compared to untreated control, except in the case where 4 days of DCA led to a small statistically significant decrease in numbers of DIS cells only (Supplementary Figure 2C). However, the combination of DCA, Met, and ABT-737 had a much greater effect on decreasing senescent cell viability compared to ABT-737 alone; this effect was seen for both types of senescent cells, DIS and RS, but not healthy cells (Supplementary Figure 2D).

To confirm and extrapolate these data, we tested the primary combination of DMA: DCA, Met, ABT-263, for its senolytic effects on IMR-90 cells induced into senescence by etoposide, a genotoxic carcinogen (Supplementary Figure 3A, 3B). Furthermore, we explored senolytic properties of DMA under different oxygen environments. Laboratory cell culture is typically performed in atmospheric oxygen (~21%), a much higher concentration of oxygen compared to in vivo conditions. Thus, to test the effects of DMA under more physiologic conditions, we added it to “physoxic” cell cultures at 3% oxygen. Under these conditions, there was a slight decrease in healthy control cell viability with DMA, but a much more substantial drop, ~90%, in viability of senescent cells (Supplementary Figure 3C).

To rigorously test whether DMA selectively targets senescent cells while sparing healthy cells, in addition to the non-senescent IMR-90 cell line, the cocktail was tested on primary human neural precursor cells (NPCs), primary human myoblasts, and primary human hepatocytes (Supplementary Figure 3D–3F). DMA caused no reduction in NPC viability and a slight reduction in the viability of human myoblasts (~20%) and human hepatocytes (5%). In myoblasts, a similar effect was observed with DCA alone, in agreement with the effects of this drug on reducing myoblast proliferation [48].

The senescent phenotype often propagates to more cells through the release of SASPs, a phenomenon known as paracrine senescence [49, 50]. To test whether DMA could reduce factors driving paracrine senescence, SnCs were treated with DMA, ABT-263 at varying concentrations, or vehicle control, and then the conditioned medium from these cultures was applied to healthy, proliferating IMR-90 cells. Treatment of SnCs with DMA or ABT-263 alone failed to suppress senescence transmission to healthy recipient cells. However, interestingly, even at a 10-100-fold lower concentration than in the DMA mix this paracrine senescence was significantly reduced (Supplementary Figure 3G). These results suggest that at doses too low for SnCs ablation, DMA can still limit the deleterious bystander effect of senescent cells onto healthy cells.

In summary, DMA exhibited higher senolytic effectiveness than its individual components, or DCA + metformin, moreover, with a better drug safety profile in vitro and in vivo as compared to conventional doses of ABT-263.

DMA eliminates a variety of cancer cell lines by inducing their apoptosis and inhibiting proliferation

Next, we investigated whether DMA would also be effective against cancer cells, as some cancers are resistant to ABT-263 alone and share some of metabolic defects with senescent cells. First, to better understand cancer cell responses to ABT-263, we treated three different cancer cell lines with ABT-263 for two days. Published doses of ABT-263 (10μM) had vastly different effects on cervical (HeLa), breast (MCF-7), and colorectal (SW480) cancer cells (Figure 2A) [22]. SW480 cell viability dropped by 60%, whereas MCF-7 cancer cultures continued to thrive. HeLa cells showed a moderate response, with about a ~25% reduction in viability. Based on literature [24, 25], we hypothesized that resistance to ABT-263 in certain cancers might be due to higher levels of MCL-1, a member of the BCL-2 family that ABT-263 does not target. Western blotting for BCL-2, BCL-W, BCL-XL, and MCL-1 confirmed that levels of MCL-1 protein indeed correlated positively with the resistance to ABT-263 (Supplementary Figure 4A, 4B). Expression levels of these BCL-2 family genes were studied by qRT-PCR, which confirmed the MCL-1 correlation, albeit not always at the same level as the protein, as expected from transcriptional and translational regulation (Supplementary Figure 4C).

Figure 2. DMA eliminates various types of cancer cells. (A) Relative cell viability of HeLa, MCF-7, and SW480 cancer cell lines measured by the CellTiter-Blue assay, showing variable responses to treatment with 10 μM ABT-263 for two days. (mean ± SD, one-way ANOVA, Bonferroni’s multiple comparisons test). (B) Relative cell viability over 7 days showing that the combination of DCA, metformin, and 1 μM ABT-263 produced a significantly greater reduction in viability than 1 μM ABT-263 alone (mean ± SD, two-way ANOVA, Holm-Šídák test,). (C) Percentage of TUNEL-positive MCF-7 cells, showing significantly greater apoptosis in DMA-treated samples compared with ABT-263 or solvent control (mean ± SD, one-way ANOVA, Bonferroni’s multiple comparisons test). (D) Representative flow cytometry plots of MCF-7 cells treated with DMA or solvent control after EdU staining. (E) Percent EdU-positive MCF-7 cells by flow cytometry, showing significantly less proliferating cells when treated with DMA vs. solvent control (mean ± SD, unpaired t-test). (Statistical significance is denoted as: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Next, we examined whether DMA would overcome the resistance associated with elevated MCL-1 and diminish cancer cell viability. Following seven days of DMA in culture, cancer cell viability was significantly reduced in all studied cells, with the most significant impact on the ABT-263-resistant MCF-7 breast cancer cell line (Figure 2B). DMA was necessary and sufficient for ablating the different types of cultured human cancer cells, while DCA + Met, ABT-263 + DCA, or each individual drug failed (Supplementary Figure 4D).

Several studies have reported that a combination of Met with DCA reduces MCL-1 expression [51]. However, when MCF-7 cells were treated with Met and DCA at the concentrations used in DMA and at the published doses (4-fold higher than those used in DMA), no significant change in MCL-1 protein levels were detectable by ELISA (Supplementary Figure 4E). These results demonstrate that DMA acts independently of reduction in MCL-1 levels, albeit they do not address potential effects on MCL-1 function.

The drop in cancer cell numbers could have resulted from an increase in cell death, a decrease in proliferation, or both. To investigate this, TUNEL and EdU assays were performed to assess cell death and proliferation, respectively. From the TUNEL assay, we confirmed that DMA treatment resulted in far more apoptotic cells than 1 μM ABT-263 alone or the vehicle control (Figure 2C). To explore proliferation rates, DMA and EdU were added to MCF-7 cells for 12 hours, after which the cells were analyzed by flow cytometry. The results demonstrated that DMA significantly reduced the proliferation rate of MCF-7 cells (Figure 2D, 2E). Thus, DMA lowers MCF-7 viability by simultaneously promoting apoptosis and inhibiting proliferation.

Summarily, these data suggest that DMA can eliminate cancers of different tissues, including those resistant to BCL-inhibitors, and the observed drop in cancer cell numbers following DMA treatment was due to both increased cell death and reduced cell proliferation.

Mechanisms of DMA selectivity for ablating cancer and senescent cells but not healthy cells

The effects of DMA on decreasing viability of select cancer and senescent cell lines, but not healthy cells, combined with the known activities of DCA and metformin and the metabolic differences between healthy, cancer, and SnCs, suggested the mechanisms behind DMA selectivity. We hypothesized that unhealthy mitochondria and impaired ATP production in both SnCs and cancer cells make them similarly susceptible to DMA. Namely, since SnCs and cancer cells already have compromised mitochondria, the effects of DCA and metformin are expected to be severe. Further, we hypothesized that DCA and Met would fatally deplete ATP in SnCs and cancer cells, making them more susceptible to the low dose of ABT-263. At the same time, the same doses of these drugs in healthy cells would not be deleterious due to their more efficient and resilient mitochondrial function [26–28].

To examine this mechanism, first we treated SnCs, cancer cells, and healthy cells with either DMA, DCA, or 1 μM ABT-263 for 12 hours, and compared the ATP levels using an ATP bioluminescence assay. DMA significantly decreased ATP in both SnCs and MCF-7 cells, bringing levels close to background, while healthy cells were able to maintain their ATP levels (Figure 3A, 3B). ABT-263 alone promoted a trend for reduced ATP in all tested cells including healthy cells, and significantly reduced ATP in the replicative senescent and MCF-7 cells, consistent with capacity of a BCL-2 inhibitor to decrease mitochondrial numbers [52]. As expected from its mechanisms of action, DCA increased ATP in senescent and non-senescent cell lines, but, interestingly, not in cells driven into senescence by replicative exhaustion. Of note, despite the increase of oxidative phosphorylation, DCA is considered to be relatively non-toxic to normal cells in vitro and in vivo [53] and (Supplementary Figure 1B). To test if SnCs ablation was dependent on low ATP levels, exogenous ATP was added to the media as previously published [54, 55], which partially restored SnC viability, making it similar to that with ABT-263 alone (Supplementary Figure 5A).

Figure 3. Mechanism of DMA effects on pathogenic versus healthy cells. (A) Bioluminescent ATP assay following the addition of DMA for 12 hours in NS, DIS, and RS cells showing that only DIS and RS exhibited significantly reduced ATP levels as compared to solvent control (mean ± SD, one-way ANOVA, Bonferroni’s multiple comparisons test). (B) Bioluminescent ATP assay following the addition of DMA on MCF-7 cells, showing significantly reduced ATP levels as compared to solvent control (mean ± SD, one-way ANOVA, Bonferroni’s multiple comparisons test). (C) Glycolysis stress test results from NS, DIS, and RS cells, where NS cells were better able to respond to inhibition of oxidative phosphorylation (mean ± SD, two-way ANOVA, Holm-Šídák test). (D) Glycolytic reserve and (E) glycolysis rates from the glycolysis stress tests show that DIS and RS cells cannot increase glycolysis as effectively as NS cells (mean ± SD, one-way ANOVA, Bonferroni’s multiple comparisons test). (F) Results from mitochondrial stress tests on NS and DIS cells treated with 1 μM ABT-263 or DMA, showing that DIS cells were unable to increase respiration in the presence of DMA. (mean ± SD, two-way ANOVA, Holm-Šídák test). (G) Response after FCCP (maximal respiration) in NS and DIS cells shows that NS cells increase respiration more than DIS cells (mean ± SD, unpaired t-test). (Statistical significance is denoted as: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Overall, no individual component of DMA caused a total ATP loss in SnC or cancer cells, and only complete DMA had this effect. These results suggest that the addition of DCA and Met to ABT-263 caused selective, ATP-depletion in SnCs and cancer cells, but not healthy cells.

Next, to better understand why DMA caused a decrease in ATP in SnCs but not healthy cells, we used a Seahorse XF Glycolysis Stress Test to examine glycolytic capacity. Our data showed that both DIS and RS cells had lower glycolytic reserves and glycolysis rates (with glucose saturation) than healthy cells, indicating a reduced capacity to compensate for the metformin-induced inhibition of electron transport chain Complex I and glycolysis, combined with the DCA-shift to oxidative phosphorylation (Figure 3C–3E).

Given that glycolytic capacity was diminished in SnCs, we wanted to test if respiration was also impaired. The Seahorse XF Cell Mito Stress Test demonstrated that, unlike glycolysis, there were no differences in respiration between the SnCs and healthy cells under basal conditions (Supplementary Figure 5B). However, with the addition of DMA, healthy cells but not DIS cells, were able to respond to Carbonyl cyanide p-(trifluoromethoxy) phenylhydrazone (FCCP) (Figure 3F, 3G). Since FCCP collapses the mitochondrial membrane potential, stimulating maximal respiration, these data suggest that DIS cells are unable to respond to increased energetic demand and fail to adapt to the additional stress brought on by DMA, leading to cell death.

In summary, DMA can selectively eliminate SnCs and cancer cells due to their similar vulnerabilities in energy production, by lethally diminishing their respiratory spare capacity and preventing sufficient ATP production. Crucially, SnCs could not respond to FCCP after the addition of DMA, showing a limitation in the ability to continue ATP production under stress.

Rejuvenating properties of DMA in vivo

Off-target toxicity remains a critical concern for both senolytic and cancer therapy, in general and particularly at old age, characterized by flare of cancers and tissue senescence [56, 57]. Therefore, although we had established that DMA did not cause acute thrombocytopenia, a major side effect of most BCL-2 family inhibitors, we decided to expand the evaluation of overall DMA safety in vivo, in the functional performance and lifespan of old mice.

DMA or vehicle control were administered to 18–24-month-old mice for two weeks, five days a week. The animals were randomly assigned to DMA vs. control vehicle groups and longitudinally tested in several health assays before and after the treatment (Figure 4A). The data showed that old animals treated with DMA exhibited significantly improved endurance, as measured by treadmill run time compared to vehicle controls. (Figure 4B). Further, DMA caused no adverse effect on the hang test (which measures strength and agility) nor higher overall frailty of the old animals (Figure 4C, 4D).

Figure 4. DMA effect in vivo. (A) Diagram of acute injection schedule. (B) Percent change in treadmill run time after injection of DMA or solvent control, compared with the previous month, showing that DMA-treated mice exhibited significantly greater improvement than solvent control-injected mice (mean ± SEM, unpaired t-test, Control: 7 males, 10 females, DMA = 4 males, 7 females (3 mice did not habituate)). (C) Percent change in hanging test time from mice treated with DMA or solvent control, showing no significant difference between groups (mean ± SEM, unpaired t-test, Control: 7 males, 10 females, DMA = 7 males, 7 females). (D) Percent change in frailty from mice treated with DMA or solvent control showing no significant difference between groups (mean ± SEM, unpaired t-test, Control: 7 males, 10 females, DMA 7 males, 7 females). (E) Volcano plot depicting differential serum protein expression in mice after vs. before DMA treatment, expression determined by antibody array, dotted lines at x=-0.2,0.2 and y=1.301(paired t-test, n=3 high throughput studies). (F) Change of known SASPs [6] after DMA injection, expression determined by antibody array (min to max shown, n=3). (G) Survival plot of DMA vs. Control, post treatment (log-rank (Mantel-Cox) test, Control: 3 males, 6 females, DMA = 5 males, 4 females. Median lifespan: Control: n=815, DMA; n=1002). (Statistical significance is denoted as: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

To examine potential changes in the systemic protein determinants of tissue health, we used antibody capture arrays to longitudinally assay the serum of old mice before and after DMA (Figure 4E). Based on published literature, aging is characterized by elevated levels and noise (or variation in levels) of systemic proteins [58, 59]. Thus, it was notable that some serum proteins, including pro-inflammatory and SASP factors, were reduced by the DMA treatment (Figure 4E). To provide more details to the changes in SASP proteins, quantification of their levels showed an overall downward trend (Figure 4F). Moreover, the levels of systemic proteins in DMA-treated mice were found to trend towards those found in young mice (Supplementary Figure 6A, 6B), suggesting a recalibration of the systemic proteome to a younger state.

Lastly, to better understand the sustained effect of DMA, the combination was administered to the old mice longitudinally, starting at around 18 months of age and continuing until their natural death. The mice were given DMA once daily for five consecutive days per week over two weeks, followed by an eight-week break. The lifespans were recorded. The weights of both cohorts remained stable during the injection schedule (Supplementary Figure 6C). Interestingly, despite being treated with a combination of chemotherapeutic drugs, the average lifespan was not shortened but was instead extended by an average of 102.6 days, or a 41.7% increase post-injection (Figure 4G and Supplementary Figure 6D). While similar initially, the hanging test performance of old mice dosed with DMA had a trend toward improvement as compared to vehicle-treated animals, particularly at later time points of the study (Supplementary Figure 6E). Thus, DMA not only showed potential for being safe in vivo, but it enhanced the lifespan and promoted a healthier profile of the systemic proteome in old mice.

In vivo studies with old mice

All animal experimental procedures were performed in accordance with the Guide for Care and Use of Laboratory Animals of the National Institutes of Health and approved by the Office of Laboratory Animal Care (OLAC), UC Berkeley. Aged C57B/6 mice (between 18-25 months old) were purchased from the National Institute of Aging. Blood was collected via chin puncture while the mice were under 2% isoflurane anesthesia, 1 week before injections and 1 week after injections ended. The serum was then isolated by spinning down the blood for 10 minutes at 2,000g and collecting the supernatant.

Drugs for in vivo delivery

Solid ABT-263 was dissolved using 60% Phosal 50 PG (HY-Y1903)/30% polyethylene glycol 400 (NC0814144)/ 10% Ethanol. The mixture was sonicated for 20 seconds. Metformin (ICN15169105) and dichloroacetate (347795) were both dissolved in HBSS.

For the platelet study, ABT-263 alone was administered via oral gavage (5081042) as previously described at a dose of 50 mg/kg [6]. For DMA, ABT-263 was administered via oral gavage as previously described at a dose of 5 mg/kg, while Metformin and dichloroacetate were delivered subcutaneously at a dose of 50 mg/kg.

For acute delivery, ABT-263 was delivered via oral gavage at a concentration of 5 mg/kg/day. DCA and Metformin were injected subcutaneously at a concentration of 50mg/kg/day. Both drugs were given for five consecutive days, followed by a two-day intermission, then the drugs were given for an additional five consecutive days. Control mice were given the solvent alone.

For sustained delivery, DMA was administered at the same dose as acute delivery for five consecutive days, followed by a two-day intermission. The drugs were given for an additional five consecutive days, followed by a 6-week intermission. Control was given the solvent.

Flow cytometry

Blood samples were collected in tubes with heparin (81952-112-09). 50μL of fresh blood was combined with 450 μL of HEPES-buffered Tyrode solution (AAJ67607AP). After, CD41a-FITC (50-955-3) was added at a concentration of 1 to 100 and placed on a rocker for 20 minutes. 2mL of High-Yield Lyse (HYL250) solution was added and incubated at room temperature for 10 minutes. The sample was collected and run with a ThermoFisher Attune NxT. Cells were first filtered based on size (FSC-A), then selected based on CD-41a (BH-1) intensity. For flow cytometry experiments involving nucleated cells, the ThermoFisher Attune NxT was used. Cells were first filtered based on size (FCS vs. SSC), then Doublets were removed (FSC-A vs. FSC-H).

Culturing

IMR-90 cells were cultured in DMEM (MT10017CM), 10% Fetal Bovine Serum (SH3007103), 1% Pen/Strep (15140163), and 1% Non-Essential Amino Acids (M7145-100ML) until about 70% confluency and then passaged. To passage, cells were washed with PBS, and then 0.05% Trypsin (MT25052CI) was added for 2-3 minutes. Cells were spun down and counted. The population doubling level increase was calculated every time using PDL_Increase=(1/Log(2))*(Log(Final Cell Count)-Log(Initial Cell Count)). After calculating the PDL increase, the PDL final was calculated by PDL_Final= PDL_Increase+PDL_Initial. For all experiments, Non-senescent IMR-90 cells refer to a PDL lower than 30. Human cancer cell lines were cultured using similar media and passaged using similar methods. Human neural precursor cells and human myoblasts were cultured as previously described [64, 65]. Human hepatocytes were purchased from Lonza (HUCPG) and were cultured according to the manufacturer’s protocol.

Senescence induction

Damage-induced senescence: Samples were prepared by seeding 100,000 low PDL(>30) IMR-90 fibroblast cells in 100 mm by 20 mm sterile petri plates in culture media. Cells are cultured until all petri dishes have reached 80% confluency. Damage-induced senescence was induced by exposing the cells to 16,000 μJ/cm² of ultraviolet light with Strategene’s UV Stratalinker 2400 once every other day for a total of five times. Senescence was confirmed via Sa-β-gal two days after the final UV. Replicative Senescence: IMR-90 cells were cultured until replicative exhaustion (>PDL 48). Senescence was confirmed via Sa-β-gal. Etoposide senescence induction: Etoposide senescent IMR-90 cells were generated by adding 20μM etoposide for 48 hours. Senescence was confirmed three days after via Sa-β-Gal.

SA-β-gal staining

The Cell Signaling Senescence β-Galactosidase Staining Kit (9860s) was used to detect β-galactosidase levels, and staining was performed according to the manufacturer’s protocol. Briefly, cells were fixed with 1x fixing solution and then washed twice with PBS. The 1x SA-β-gal staining solution was prepared as described by the kit; for every 1 mL, 930 μL of 1x Staining Solution, 10 μL of 100x Solution A, 10 μL of 100x Solution B, and 50 μL of the dissolved x-gal (20mg/mL in DMF) were combined. An additional solution was prepared and used to calculate the amount of either 1 M HCl or 1 M NaOH needed to adjust the pH to a range of 5.9-6.1. After adjusting the pH of the solution, it was added and incubated overnight at 37° C. Samples were viewed with a phase contrast microscope. If a cell’s body was a dark blue color, the cell was considered senescent.

Drugs for in vitro studies

Metformin and dichloroacetate solutions were prepared by dissolving solid metformin hydrochloride (C829T51) and/or sodium dichloroacetate (347795) into cell culture media. ABT-737(C832E72), ABT-263(C832E67), were dissolved in DMSO(89139-666) at 1000x the final desired concentration.

3D-Drug curve

A Multidrop™ Combi nL Reagent Dispenser was used to seed eighty 96-well tissue culture plates with 10,000 cells per well. The cells were left to incubate O/N. Stock drug plates were prepared using a Beckman Coulter Biomek NXp Automated Liquid Handling Workstation with deep 96-well plates. The drugs were transferred from the deep-well plates to the tissue culture plates with cells using the Bravo Automated Liquid Handling Platform. After 24 hours, the cells were fixed with 70% ethanol and stained with Hoechst (Cat: H1399) at a concentration of 10μg/mL in PBS. Plates were sealed with Agilent’s PlateLoc Peelable Aluminum seals and imaged with an Opera Phenix. Harmony was used to quantify the total number of remaining nuclei. After, the number of cells was normalized to untreated cell counts and graphed using MATLAB.

Viability

To assess the levels of cell viability, Promega’s CellTiter-Blue Cell Viability Assay (G8081) was used. Cells were cultured in flat-bottomed 96-well plates at 10,000 cells per well. After the desired drug exposure, the media was replaced with 80μL of Opti-MEM and incubated for 24 hours. Afterwards, 20 μL of Cell Titer-Blue was added. The cells were incubated for two hours. Fluorescence was measured using the SpectraMax M5 plate reader (ex:560/em:590). For analysis, all values were subtracted by blanks (media + CellTiter-Blue, with no cells). Relative viability was calculated by normalizing to the control.

Hypoxia

The hypoxia chamber was set to 3% O₂. Both the DIS and Non-senescent IMR-90 cells were placed in the hypoxia chamber for 24 hours before the experiment. Before any media/drugs were changed, the cells were equilibrated in the hypoxia chamber for at least 2 hours.

DMA on paracrine senescence

DIS IMR-90 cells were placed in low serum media with/without varying doses of senolytics for 9 hours. A negative control of healthy IMR-90 cells was also placed in low-serum media. The media from each dish of cells were collected and centrifuged at 300g for 5 minutes to separate any cells from the media. The supernatant was then diluted 2x with fibroblast growth media to make conditioned media. Healthy IMR-90 cells were cultured with the conditioned media for 4 days, and the conditioned media was changed every other day. The cultured cells were then fixed, stained with SA-β-gal, and imaged to determine the percentage of paracrine senescence.

Western blotting

Cells were 10x lysed with 10x Cell Signaling lysis buffer (9803) and Thermo Scientific™ Halt™ Protease Inhibitor Cocktail (1861278). Lysed cells were spun down at 13,000g for 10 minutes, and the supernatant was collected. Protein concentration was determined through the Thermo Scientific Micro BCA™ Protein Assay Kit (23235). Loading sample was prepared by combining 4x Laemmli sample buffer (1610747), 5% beta-mercaptoethanol, 20μg of protein, and PBS for a final volume of 25μL. The solution was vortexed together and heated at 97 °C for 10 minutes. The solution was loaded into SDS-Page gels and run at 120 V for 1 hr. The protein was transferred to a PVDF membrane at 15 V for 16 hrs. The PVDF Papers were placed on a rocker and blocked using 5% milk in TBST buffer for 30 mins, then exposed to primary antibodies (details listed below) in 1% TBST buffer for 2 hrs. GAPDH was detected as a loading control. The membranes were then washed three times using TBST buffer for 5 minutes and exposed to secondary antibodies in 1% milk for 2 hrs. The membranes were again washed three times with TBST buffer for 5 minutes. The chemiluminescent reaction was initiated with Advansta WesternBrightTM ECL HRP (490005-016) and the blot was visualized using UVP ChemStudio PLUS (Supplementary Figure 7).

qRT-PCR

RNA was extracted with the RNeasy Qiagen (74106) RNA extraction kit. The extracted RNA was then processed with Thermo Fisher’s SuperScript III (11752-050) according to the manufacturer’s protocol. Finally, the qRT-PCR reaction was done using Bio-Rad’s SYBR Green Supermix (1725270) and was run using the CFX Connect Real-Time System.

ELISA

The ELISA for MCL-1 levels was performed using the Invitrogen^TM Human Mcl-1/Bcl2L3 ELISA (EH317RB) kit, following the manufacturer’s protocol. A BCA assay was used to load the same protein concentration in all conditions. Absorbance was measured at 450 nm using a SpectraMax iD3 plate reader. A 4-parameter logistic (4PL) curve was used to interpolate the standard curve in Graphpad Prism.

TUNEL

The TUNEL assay was performed on MCF-7s after exposure to DMA for 12 hours. Staining was performed using a modified version of the manufacturer’s protocol. The incubation time of the TUNEL solution was increased to O/N at RT on a rocker. The stained images were imaged on the IXM Molecular Devices ImageXpress Micro. Total nuclei were counted using the provided software, and TUNEL-positive nuclei were counted blindly and manually.

EdU assay

EdU staining was performed using Thermo Fisher’s Click-iT^® EdU Imaging Kit (10338). EdU stock was diluted in warm culture media, and cells were pulsed with 10μM EdU for 2-4 hours. Then, the cells were fixed with paraformaldehyde and permeabilized for 20 minutes using Stemcell Technologies’ Intracellular Permeabilization Buffer (100-1451). On the day of the click reaction, Thermo Fisher’s EdU additive solution 10x was diluted in molecular grade water and a reaction mixture was created based on the ThermoFisher’s Click-iT^® recipe. The permeabilization buffer was aspirated, and 500 μl of reaction mixture was added to each well in a standard 6-well plate. Cells were washed twice with a 3% BSA solution and once with PBS. FITC standard filter set and 495/519 Excitation/Emission (in nm) were used for imaging.

ATP studies

The total ATP amount was quantified using the ATP Bioluminescence Assay Kit HS II (11 699 709 001) and then normalized with BCA. Cells were seeded at a density of 50,000 cells per well in a 6-well plate. 5mM DCA, 5mM Met, and 1μM ABT-263 was added for 12 hours. After, cells were washed twice with PBS, lifted with trypsin, and washed twice more with PBS. After, cells were lysed with 400 μL of Cell Lysis reagent for 5 minutes at room temperature. The insoluble material in the cell lysates was sedimented by centrifugation at 10,000 g for 2 minutes, and the supernatant was used to measure ATP levels. The standard ATP curve was created using concentrations from 1μM to .00001μM ATP. The luminescent reading was performed using a SpectraMax iD3. For the reading, 50 μL of sample/standard was added to a 96-well plate (white, clear bottom). Then, 50 μL of Luciferase Reagent was added to each well, followed by a one-second delay and a 9-second integrative read. After calculating the ATP concentration, values were normalized by the BCA results.

Glycolytic stress test analysis by Seahorse extracellular flux analyzer

Cells were plated on Agilent Seahorse XF24 Cell Culture Microplates and optimized where control cells were plated at a density of 15,000 cells/well while damage-induced senescent and replicative senescent cells were plated at 80,000 cells/well. Agilent Seahorse XF Glycolysis Stress Test Kit was used per standard protocol (https://www.agilent.com/cs/library/usermanuals/public/XF_Glycolysis_Stress_Test_Kit_User_Guide.pdf). During glycolysis, the extrusion of protons indicates the extracellular acidification rate (ECAR), which was measured using Seahorse extracellular flux XFe24 analyzers, reported by Agilent Seahorse Wave Desktop software, and the data were normalized by cell count.

Mito stress test analysis

Senescent cells were optimized at 80,000 cells per well on Agilent Seahorse XF24 Cell Culture Microplates. Cells were incubated for 24 hours, and in the last hour, media was aspirated, and the cells were incubated in Seahorse media with the drugs. Otherwise, the Agilent Seahorse XF Mito Stress Test Kit was used per standard protocol (https://www.agilent.com/cs/library/usermanuals/public/XF_Cell_Mito_Stress_Test_Kit_User_Guide.pdf). Oxidative phosphorylation, indicated by oxygen consumption rate (OCR), was measured with Seahorse extracellular flux XFe24 analyzers and reported by Agilent Seahorse Wave Desktop software. Cells were analyzed for viability using the CellTiter-Blue Cell Viability Assay, and data were normalized by cell count.

Mouse health studies

For the treadmill exhaustion test, mice were fasted for two hours prior to being placed on the treadmill during both the habituation and testing phases. During the habituation phase, mice were acclimated for a period of 2 days, for 10 minutes each day. On the third day, the testing phase, the treadmill started at a speed of 8 m/min and the treadmill increased at a rate of 1 m/min every 10 minutes for a total of 60 minutes, and increased at a rate of 1 m/min every 5 minutes beyond the 60-minute mark until the mice were exhausted. Exhaustion was defined by the mouse not voluntarily running for more than 10 seconds. If the mouse refused to run for more than a total of 10 seconds, it was considered an invalid run.

The hanging test was performed by placing mice on a 1 cm mesh, 1 mm wire screen positioned 25 cm over soft bedding. Each mouse was placed on the grid so that they could grasp it with all four paws. The grid was then inverted. Mice were recorded for a maximum of 10 minutes, or until the mouse fell into the bedding. The hanging test was repeated 3 times, and the average hang time was recorded. The frailty index was conducted and quantified as previously published [66].

Antibody array

Antibody arrays (Raybiotech, Mouse L308 Array) were performed on mouse serum. The reagents were prepared as per the manufacturer’s instructions. 10 μL of serum was combined with 90 μL PBS and cleared by centrifugation at 13,000g for 5 minutes. The supernatant was dialyzed overnight at 4° C. On the following day, the dialyzed serum was centrifuged at 13,000 g for 5 minutes. After, 35μL of the supernatant, 22μL of Labeling reagent Item B, and 155μL of Labeling buffer Item K were combined and vortexed. The mixture was placed on a rotator for 30 minutes at RT. After 30 minutes, 3 μL of Stop Solution Item D was added. The mixture was then centrifuged at 13,000 g for 5 minutes. Again, the samples were dialyzed O/N at 4° C. The array was dried for an hour and then blocked O/N using the provided blocking buffer. The following day, the dialyzed samples were then spun down at 13,000 g for 5 minutes. 40 μg of protein in 40 μL was combined with 360 μL of blocking buffer and incubated overnight at 4° C. The next day, the array was washed three times with Buffer I for 5 minutes each, followed by two 10-minute washes, and then two 5-minute washes with Buffer II. Then, 400 μL of diluted Cy3-conjugated streptavidin was added to each subarray, covered with a seal and aluminum foil, and incubated overnight at 4° C. The following day, the slides were washed three times with Buffer I for 10 minutes each, followed by two 5-minute washes with Buffer II, and a final wash with DI water for 5 minutes. Finally, the slides were dried and scanned using a Genepix 4400A scanner (Molecular Devices), and the antibody signal values were normalized to the positive and negative values (Supplementary Figure 8).

PCA plot

PCA was performed using the prcomp() function from the stats package in R (version 4.4.2). The resulting PCA scores were visualized using the ggplot2 package (version 4.0.2).

Graphing and statistical analysis

Graphing and statistical analysis was done with GraphPad Prism unless mentioned otherwise in the methods. Organization of the figures and textual edits were made in Illustrator.

Availability of data and materials

The datasets used and/or analyzed during the current study have been submitted with the manuscript.