Mitochondria are required for pro-ageing features of the senescent phenotype

Abstract

Cell senescence is an important tumour suppressor mechanism and driver of ageing. Both functions are dependent on the development of the senescent phenotype, which involves an overproduction of pro-inflammatory and pro-oxidant signals. However, the exact mechanisms regulating these phenotypes remain poorly understood. Here, we show the critical role of mitochondria in cellular senescence. In multiple models of senescence, absence of mitochondria reduced a spectrum of senescence effectors and phenotypes while preserving ATP production via enhanced glycolysis. Global transcriptomic analysis by RNA sequencing revealed that a vast number of senescent-associated changes are dependent on mitochondria, particularly the pro-inflammatory phenotype. Mechanistically, we show that the ATM, Akt and mTORC1 phosphorylation cascade integrates signals from the DNA damage response (DDR) towards PGC-1β-dependent mitochondrial biogenesis, contributing to aROS-mediated activation of the DDR and cell cycle arrest. Finally, we demonstrate that the reduction in mitochondrial content in vivo, by either mTORC1 inhibition or PGC-1β deletion, prevents senescence in the ageing mouse liver. Our results suggest that mitochondria are a candidate target for interventions to reduce the deleterious impact of senescence in ageing tissues.

Keywords: ageing; inflammation; mTOR; mitochondria; senescence.

Figure 1. Mitochondria are key factors for the development of pro‐ageing features of cellular senescence

1. Scheme illustrating the experimental design: Parkin‐expressing and control MRC5 fibroblasts were irradiated with 20‐Gy X‐ray and 2 days after were treated with 12.5 μM CCCP for 48 h. At day 4 after irradiation, CCCP was removed and cells were kept in culture with normal serum‐supplemented medium. Cells were harvested at days 10 or 20 after irradiation (6 or 16 days after CCCP treatment, respectively) for senescent phenotypes analysis.

2. Western blots showing the absence of mitochondrial proteins from the different mitochondrial complexes: NDUFB8 (complex I), SDHA (complex II), UQCRC2 (complex III) and MTCO‐1 (complex IV), in senescent (10 days after 20‐Gy X‐ray) control and Parkin‐expressing MRC5 fibroblasts. Data are representative of three independent experiments.

3. Cellular oxygen consumption rates (OCR) in senescent (10 days after 20‐Gy X‐ray) control and Parkin‐expressing MRC5 fibroblasts. Data were obtained using the Seahorse XF24 analyzer and show mean ± SD n = 4 technical repeats (representative of two independent experiments).

4. Representative 3D EM pictures of senescent (20 days after 20‐Gy X‐ray) Parkin‐expressing MRC5 fibroblasts with or without CCCP. Mitochondria are in red; the nucleus in blue.

5. Representative images of cell size, Sen‐β‐Gal activity (blue cytoplasmic staining), macro‐H2A foci and SDHA immunofluorescence in proliferating and senescent (10 days after 20‐Gy X‐ray) Parkin‐expressing MRC5 fibroblasts. Scale bar = 10 μm.

6. Quantification of ROS levels (DHE fluorescence), Sen‐β‐Gal‐positive cells and senescence‐associated heterochromatin foci (SAHF) observed by DAPI in proliferating and senescent (10 days after 20‐Gy X‐ray) control and Parkin‐expressing MRC5 fibroblasts. Data are mean ± SEM of n = 3 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

7. Representative Western blots showing p21 and p16 expression in proliferating and senescent (10 days after 20‐Gy X‐ray) control and Parkin‐expressing MRC5 fibroblasts. Data are representative of 3 independent experiments.

8. Secreted protein measured in a cytokine array (RayBiotech) of a variety of inflammatory proteins in proliferating and senescent (10 days after 20‐Gy X‐ray) Parkin‐expressing MRC5 fibroblasts. Data are mean ± SEM of n = 3 independent experiments. Asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

Figure EV1. Mitochondria are key factors for the development of pro‐ageing features of cellular senescence

1. Effects of CCCP treatment on mitochondrial depletion, immediately (left) and 16 days (right) after 48 h of 12.5 μM CCCP treatment on irradiated (IR) MRC5 fibroblasts. Cells were irradiated with 20‐Gy X‐ray and treated with 12.5 μM CCCP at day 2 after irradiation. After 48 h of CCCP treatment, cells were either immediately collected for analysis (4 days after IR) or were kept in culture with normal serum‐supplemented medium for 16 days and then harvested for analysis (20 days after IR). Western blots are representative of three independent experiments.

2. T.E.M. images of senescent Parkin‐expressing MRC5 fibroblasts at 10 and 20 days after irradiation, pre‐treated or not with CCCP. Scale bar = 1 μm; red arrows denote mitochondria. Note: when conducting 3D‐EM of 20 days after IR, one cell appeared to have lost Parkin and did not show the depletion of mitochondria (indistinguishable from controls), while those where depletion was complete (the majority) were devoid of intact mitochondria.

3. Levels of secreted IL‐6 and IL‐8 proteins (measured by ELISA) in proliferating and senescent (10 days after 20‐Gy X‐ray) control and Parkin‐expressing MRC5 fibroblasts with or without 12.5 μM CCCP treatment. Data are representative of two independent experiments.

4. (left) Representative flow cytometry histogram of mitochondrial mass staining (NAO) in parental and rho0 143B osteosarcoma cells (data are representative of three independent experiments), (middle) quantification of ROS levels (DHE intensity) and (right) mRNA abundance of the SASP factor IL‐6 in parental and rho0 cells following 10‐Gy X‐ray. Data are mean ± SEM of n = 3 independent experiments; asterisks denote a significance by one‐way ANOVA at P < 0.05.

5. Representative images of the proliferation marker Ki67 (representative of two independent experiments), quantification of population doublings (PD) and BrdU‐positive cells of proliferating and senescent (10 and/or 20 days after 20‐Gy X‐ray) control and Parkin‐expressing MRC5 fibroblasts with or without 12.5 μM CCCP treatment. Data are mean ± SEM of n = 3 independent experiments for PD analysis and mean ± SD of 10 random panes for BrdU analysis.

6. Representative Western blots of mTORC1 activity, measured by p70S6K phosphorylation (T389), in senescent (10 days after 20‐Gy X‐ray) control (C) and Parkin‐expressing (P) MRC5 fibroblasts. Data are representative of 3 independent experiments.

7. Representative Western blots showing the absence of mitochondrial proteins (NDUFB8, SDHA, UQCRC2 and TOMM20), expression of p21 and mTOR activity (measured by phosphorylation of the p70S6K (T389)) in proliferating control and Parkin‐expressing MRC5 fibroblasts (10 days after 48 h of 12.5 μM CCCP treatment). Data are representative of two independent experiments.

8. Quantification of BrdU‐positive cells (data are mean ± SD of n = 2 independent experiments), Sen‐β‐Gal activity (data are mean ± SD from 10 random planes), IL‐6 secretion measured by ELISA (data are representative of two independent experiments) and ROS levels measured by DHE intensity (data are mean ± SD of n = 3 technical repeats) of proliferating control and Parkin‐expressing MRC5 fibroblasts (after 48 h of 12.5 μM CCCP treatment).

9. Steady‐state cellular ATP levels were measured using an ATP luciferase Kit (Invitrogen) in proliferating control and Parkin‐expressing MRC5 fibroblasts (after 48 h of 12.5 μM CCCP treatment). Data are mean ± SEM of n = 3 independent experiments.

Figure 2. Global transcriptomic analysis reveals that mitochondria are significant contributors to the development of the senescent phenotype, particularly the SASP

1. Column clustered heatmap showing all genes that are differentially expressed between proliferating (Prol), senescent (Sen) and mitochondria‐depleted senescent (MDSen) Parkin‐expressing MRC5 fibroblasts (FDR ≤ 5%, absolute fold ≥ 2). Genes are by column and samples by row. The colour intensity represents column Z‐score, where red indicates highly and blue lowly expressed.

2. Venn diagram showing the overlap of the differentially expressed genes (FDR ≤ 5%, absolute fold ≥ 2) between proliferating and senescent cells and between senescent and MD senescent cells. Direction of change is indicated.

3. Table of the 10 most significantly enriched (FDR ≤ 5%) gene ontologies for the “fully reversed” genes, as defined by the tool DAVID. Ontologies that are amongst the 10 most significantly enriched for genes that are significantly up‐regulated between proliferating and senescent cells are indicated in red.

4. Column clustered heatmaps of all SASP genes and cell cycle genes that are differently expressed in senescence; genes are by column and samples by row. The colour intensity represents column Z‐score, where red indicates highly and blue lowly expressed. SASP heatmap: genes that are labelled in green represent the 60% that are fully reversed and genes labelled in pink represent the 5% that became exacerbated upon mitochondria clearance. Cell cycle heatmap: genes labelled in red are significantly changed upon mitochondria clearance.

Figure EV2. Alleviation of the senescent phenotype in mitochondria‐depleted cells does not compromise ATP generation

1. Extracellular acidification rate (ECAR) of senescent (10 days after 20 Gy) control and Parkin‐expressing MRC5 fibroblasts with or without 12.5 μM CCCP treatment. ECAR was measured using a Seahorse XF24 analyzer. Data are representative graphs from the Seahorse XF24 analyzer software. Data are mean ± SD n = 4 technical repeats (representative of two independent experiments).

2. ATP production rate in senescent (10 days after 20 Gy) control (C) and Parkin‐expressing (P) MRC5 fibroblasts with or without 12.5 μM CCCP treatment (calculated from the OCR and ECAR measured using a Seahorse XF24 analyzer). ATP production by mitochondria was calculated by multiplying the ATP turnover ((basal OCR) – (OCR with oligomycin) by the established phosphorus/oxygen (P/O) ratio of 2.3 (Brand, 2005). ATP production by glycolysis is considered to have a 1:1 ratio with lactate production. The extracellular acidification rate is mainly due to lactate and bicarbonate production and, when calibrated as the proton production rate, indicates glycolytic rate (Birket et al 2011; Wu et al, 2007). Data are mean ± SD, n = 4 technical repeats (representative of two independent experiments).

3. Steady‐state cellular ATP levels were measured using an ATP luciferase assay (Invitrogen) in senescent control and Parkin‐expressing MRC5 fibroblasts with or without 12.5 μM CCCP treatment, at 10 days (data are mean ± SEM, n = 3 independent experiments) and 20 days (data are mean ± SD from n = 3 technical repeats) after 20‐Gy X‐ray. Asterisk denotes a statistical significance at P < 0.05 using one‐way ANOVA.

4. Heatmap of the RNA‐seq analysis of glycolysis‐associated gene expression in proliferating and senescent (10 days after 20 Gy) Parkin‐expressing MRC5 fibroblasts with or without 12.5 μM CCCP treatment. Red indicates up‐regulated genes; green indicates down‐regulated genes. Data are from n = 3 independent experiments.

Figure EV3. Mitochondria are important for the development of pro‐ageing features of oxidative stress‐induced, oncogene‐induced and replicative senescence

1. (left) Scheme showing experimental design: control and Parkin‐expressing MRC5 fibroblasts were induced to senescence with 400 μM of H₂O₂ for 1 h in serum‐free medium. Two days upon the induction of senescence with H₂O_2, cells were treated with 12.5 μM CCCP for 48 h and then cultured in normal serum‐supplemented medium during 6 or 16 days for senescence markers analysis (10 and 20 days following the induction of senescence, respectively); (right) Representative images of the mitochondrial protein SDHA (immunofluorescence) in senescent control and Parkin‐expressing MRC5 fibroblasts, 10 days after the induction of senescence with 400 μM H₂O₂. Scale bar = 10 μm.

2. Quantification of Sen‐β‐Gal‐ (data are mean ± SD of n = two independent experiments) and BrdU‐positive (data are mean ± SD of 10 random planes) proliferating and senescent (10 and 20 days after the induction of senescence with 400 μM H₂O₂) control and Parkin‐expressing MRC5 fibroblasts with or without 12.5 μM CCCP treatment.

3. Levels of secreted IL‐8 and IL‐6 proteins (measured by ELISA) in proliferating and senescent (10 days after the induction of senescence with 400 μM H₂O₂) control and Parkin‐expressing MRC5 fibroblasts with or without 12.5 μM CCCP treatment. Data are representative of two independent experiments.

4. (left) Scheme showing experimental design: oncogene‐induced senescence (OIS) experiments were performed by treating Parkin‐expressing ER‐RAS/IMR‐90 fibroblasts with 12.5 μM CCCP for 48 h, followed by ER‐RAS induction with 100 nM of 4‐hydroxytamoxifen (4‐OHT). Cells were harvested 7 days after ER‐RAS induction; (right) Representative Western blots showing the expression of the mitochondrial proteins NDUFB8, SDHA, UQCRC2 and COXIV in OIS Parkin‐expressing ER‐RAS/IMR‐90 fibroblasts. Data are representative of two independent experiments.

5. Representative Western blots showing the expression of p21 and p16 proteins in OIS Parkin‐expressing ER‐RAS/IMR‐90 fibroblasts, 7 days after ER‐RAS induction. Data are representative of two independent experiments.

6. Quantification of Sen‐β‐Gal‐ and EdU‐positive proliferating and OIS Parkin‐expressing ER‐RAS/IMR‐90 fibroblasts, 7 days after ER‐RAS induction. Data are mean ± SD of n = 2 independent experiments.

7. Scheme illustrating the experimental design: Parkin‐expressing MRC5 fibroblasts were cultured until replicative senescence (RS). RS cells were treated with 12.5 μM CCCP for 48 h (until day 2) and then cultured in normal serum‐supplemented medium for 6 days.

8. Representative Western blots showing the absence of mitochondrial proteins from the different mitochondrial complexes: NDUFB8 (complex I), SDHA (complex II), UQCRC2 (complex III) and MTCO‐1 (complex IV) and expression of p21 and p16 in replicative senescent Parkin‐expressing MRC5 fibroblasts, 6 days after 12.5 μM CCCP treatment. Data are representative of two independent experiments.

9. ROS analysis (DHE intensity) in replicative senescent Parkin‐expressing MRC5 fibroblasts, 6 days after 12.5 μM CCCP treatment. Data are mean ± SD of n = 3 technical repeats and are representative of two independent experiments.

10. Quantification of Sen‐β‐Gal‐ and BrdU‐positive cells in proliferating and replicative senescent (RS) Parkin‐expressing MRC5 fibroblasts, 6 days after 12.5 μM CCCP treatment. Data are mean ± SD of 10 random planes and are representative of two independent experiments.

Figure EV4. PGC‐1β‐dependent mitochondrial biogenesis modulates cellular senescence

1. (left) Representative images and (right) histograms showing the distribution of MitoTracker green (MTG) immunofluorescence in proliferating and senescent (3 days after 10‐Gy X‐ray) wild‐type and PGC‐1β ^−/− MEFs. Scale bar = 10 μm. Dashed red lines indicate median MitoTracker green intensity (50–100 cells were quantified per condition).

2. Kinetics of mtDNA copy number after 10‐Gy X‐ray in wild‐type and PGC‐1β ^−/− MEFs. Data are mean ± SEM of n = 3 independent experiments.

3. Expression of the mitochondrial outer membrane protein VDAC in wild‐type and PGC‐1β ^−/− MEFs 1 day after 10‐Gy X‐ray. Data are mean ± SEM of n = 3 independent experiments.

4. Knockdown efficiency of PGC‐1β in human MRC5 fibroblasts infected with a control vector (shScr) and with a shPGC‐1β vector by Western blot and qPCR.

5. Quantification of mitochondrial mass (NAO intensity), mitochondrial ROS (MitoSOX intensity), mean number (N) of 53PB1 foci and Edu‐positive cells in proliferating and senescent MRC5 fibroblasts infected with a control (shScr) and shPGC‐1β vector. Mitochondrial mass and ROS measurements were performed 3 days after 20‐Gy X‐ray (data are mean ± SD of n = 3 technical replicates). 53BP1 foci and EdU incorporation were analysed 10 days after 20‐Gy X‐ray (data are mean ± SD of 80–100 cells/condition for 53BP1 analysis and 10 random planes for the EdU incorporation analysis).

6. Quantification of Sen‐β‐Gal‐positive cells in proliferating and senescent (10 days after 20‐Gy X‐ray) MRC5 fibroblasts infected with control (shScr) and shPGC‐1β vectors. Data are mean ± SD 10 random planes.

7. Representative image of FLAG‐PGC‐1β immunostaining 2 days following the transfection of a pcDNA empty vector and pcDNAf:PGC‐1β in wild‐type MEFs. Scale bar = 20 μm.

8. mRNA expression of PGC‐1β in proliferating and senescent (10 days after 10‐Gy X‐ray) wild‐type MEFs following the transfection with pcDNA empty vector and pcDNAf:PGC‐1β. Data are mean ± SEM of n = 3 independent experiments.

9. (top) Representative image of FLAG‐PGC‐1β (red) and the mitochondrial protein SDHA (green) double immunostaining and (bottom) histograms showing the distribution of SDHA fluorescence, 2 days after transfection with a pcDNA empty vector and pcDNAf:PGC‐1β in wild‐type MEFs. Scale bar = 10 μm. Dashed red lines indicate median SDHA fluorescence (100 cells were quantified per condition).

Figure 3. PGC‐1β‐dependent mitochondrial biogenesis downstream of the DDR modulates cellular senescence

1. Representative images of colony assays of wild‐type and PGC‐1β ^−/− MEFs grown at 3 or 21% O₂ (10 days after seeding 5,000 cells per well). Data are representative of 3 independent experiments.

2. Effect of 3 or 21% O₂ and X‐ray irradiation (at 3% O₂) on the percentage of Ki67 (at day 6) and Sen‐β‐Gal‐positive cells (at day 10) and the number (N) of 53BP1 foci (at day 6) in wild‐type and PGC‐1β ^−/− MEFs. Data are mean ± SEM of n = 3 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

3. Representative images of Sen‐β‐Gal activity (blue cytoplasmatic staining), Ki‐67 and 53BP1 foci in proliferating and senescent wild‐type and PGC‐1β ^−/− MEFs (scale bar = 10 μm).

4. mRNA expression of PGC‐1β,CXCL‐1 and p16 in proliferating and senescent (10 days after 10‐Gy X‐ray) wild‐type and PGC‐1β ^−/− MEFs. Data are mean ± SEM of n = 3 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

5. Effects of overexpression of PGC‐1β on percentage of Ki67‐ and Sen‐β‐Gal‐positive cells, number (N) of 53BP1 foci and percentage change in mitochondrial mass (NAO intensity) in proliferating and senescent (2 days after 10‐Gy X‐ray) MEFs cultured at 3% O₂. Data are mean ± SEM of n = 3 independent experiments. Asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA or two‐tailed t‐test.

Figure 4. mTORC1 integrates DDR signals towards mitochondrial biogenesis during cellular senescence

1. Representative Western blot of mTORC1 activity measured by phosphorylated p70S6K (T389) from 6 to 72 h after 20 Gy in MRC5 fibroblasts. Data are representative of three independent experiments.

2. Representative Western blots of the mitochondrial proteins TOMM20, NDUFB8 (complex I), SDHA (complex II), UQCR2 (complex III) and MT‐CO1 (complex IV) following 20‐Gy irradiation with or without 100 nM rapamycin treatment in MRC5 fibroblasts. Data are representative of 4 independent experiments.

3. Effect of 100 nM rapamycin on mitochondrial mass (measured by NAO fluorescence) 2–4 days following replication exhaustion (RS), genotoxic stress (generated by X‐ray irradiation, etoposide, neocarzinostatin (NCS), H₂O₂) or telomere dysfunction (TRF2^ΔBΔM) in a variety of cell lines. Data are mean from 3 independent experiments per cell line or treatment.

4. (top) Representative transmission electron microscopy (T.E.M.) micrographs of proliferating and senescent (3 days after 20‐Gy X‐ray) MRC5 fibroblasts treated with or without 100 nM rapamycin. Mitochondria are labelled in pink. Scale bar = 2 μm; (bottom left and middle) Quantification of mitochondrial volume fraction (%V_v) and mitochondrial number per cross‐section in proliferating and senescent (3 days after 20‐Gy X‐ray) MRC5 fibroblasts treated with or without 100 nM rapamycin. T.E.M. mitochondrial analysis is mean ± SEM of 24 electron micrographs per condition; (bottom right) mtDNA copy number analysis by qPCR in proliferating and senescent (3 days after 20‐Gy X‐ray) MRC5 fibroblasts treated with or without 100 nM rapamycin. Data are mean ± SEM of n = 3 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

5. Representative Western blots showing the expression of phosphorylated p70S6K (T389) and AKT (S473), the mitochondrial protein NDUFB8 and the DDR downstream target p21 in MRC5 fibroblasts treated with or without 10 μM of the ATM inhibitor KU55933 and in fibroblasts from a patient with ataxia telangiectasis (AT) at different time points after 20‐Gy X‐ray. Data are representative of three independent experiments (ATM inhibitor) and 1 experiment (AT patient).

6. Western blots showing the effect of the ATM inhibitor KU55933 on γH2A.X, AKT phosphorylation and p21 expression in MRC5 fibroblasts after 20‐Gy X‐ray. Data are from one experiment.

7. Effect of rapamycin and/or ATM inhibitor (KU55933) on mitochondrial mass (NAO intensity) in proliferating and senescent (3 days after 20‐Gy X‐ray) MRC5 fibroblasts. Data are mean ± SEM of n = 3 independent experiments. Asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

Figure EV5. mTORC1 activation contributes to the senescent phenotype via PGC1‐β‐dependent mitochondrial biogenesis and ROS‐mediated activation of a DDR

1. (left) Representative images of γH2A.X (red foci) immunofluorescence in MRC5 fibroblasts at different time points after 20‐Gy X‐ray with or without 100 nM rapamycin supplementation (scale bar = 5 μm) and (right) representative images of 53BP1 (green foci) immunofluorescence in MEFs, 3 days after 10‐Gy X‐ray with or without 100 nM rapamycin supplementation (scale bar = 5 μm).

2. Effect of genomic damage‐ or oxidative stress‐induced senescence on the mean number (N) of γH2A.X foci in MRC5 fibroblasts with or without 100 nM rapamycin. Cells were analysed 3 days after the induction of senescence and treatment with rapamycin. Data are mean ± SEM of n = 3 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

3. (top) Representative Western blot showing the inhibition of p21 protein expression with 100 nM rapamycin at different time points (days) after 20‐Gy X‐ray in MRC5 fibroblasts. Data are representative of three independent experiments; (bottom) p21 mRNA levels at 3 days after 20‐Gy X‐ray with or without 100 nM rapamycin treatment in MRC5 fibroblasts. Data are mean ± SEM of n = 3 independent experiments; asterisks denote a statistical significant at P < 0.05 using one‐way ANOVA.

4. Western blots showing the knockdown efficiency of two different siRNAs against mTOR and its effects on p70S6K phosphorylation (T389), 2 days after 20‐Gy X‐ray in MRC5 fibroblasts. Data are representative of two independent experiments.

5. Effect of 100 nM rapamycin supplementation on Sen‐β‐Gal activity in MRC5 fibroblasts induced to senescence following treatment with 20‐Gy X‐ray irradiation, 80 ng ml⁻¹ neocarzinostatin (NCS), 50 μM etoposide, 400 μM H₂O₂ and replicative exhaustion (RS). Cells were analysed 10 days after the induction of senescence and treatment with rapamycin. Data are mean ± SEM of n = 3 independent experiments (at least 80 cells were analysed per condition). Asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

6. Quantification of Ki67‐positive MRC5 fibroblasts 3 and 10 days following 20‐Gy X‐ray irradiation with or without 100 nM rapamycin treatment. Data are mean ± SEM of n = 4 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

7. Secreted protein array of a variety of inflammatory proteins following 20‐Gy X‐ray‐induced senescence in MRC5 fibroblasts with or without 100 nM rapamycin treatment (3 and 10 days after 20‐Gy X‐ray). Data are mean of three independent experiments.

8. mRNA expression of IL‐6 after 20‐Gy X‐ray (left) with or without 100 nM rapamycin treatment in MRC5 fibroblasts and (right) in human fibroblasts from an AT patient. Data are mean ± SEM of n = 3 independent experiments (for MRC5+Rap cells) and mean ± SD of n = 2 independent experiments (for AT cells).

9. Representative Western blots of mTORC1 activity, measured by p70S6K phosphorylation (T389) in wild‐type and PGC‐1β ^−/− MEFs overexpressing RhebN153T (1 day after 10‐Gy irradiation). Data are representative of three independent experiments.

10. Effect of rapamycin supplementation on Sen‐β‐Gal activity in wild‐type and PGC‐1β ^−/− MEFs, 10 days after 10‐Gy X‐ray. Data are mean ± SEM of n = 3 independent experiments (at least 100 cells were analysed per condition). Asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

Figure 5. mTORC1 activation promotes ROS‐dependent DDR and contributes to the senescent phenotype via PGC1‐β dependent mitochondrial biogenesis

1. ROS levels (DHE intensity) and mean number (N) of γH2A.X foci after mTOR knockdown (72 h) in proliferating and senescent (2 days after 20‐Gy X‐ray) MRC5 fibroblasts. Data are mean ± SEM of n = 3 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

2. ROS levels (DHE intensity) and mean number (N) of γH2A.X foci in proliferating and senescent (3 days after 20‐Gy X‐ray) MRC5 fibroblasts treated with or without 100 nM rapamycin and/or 2.5 mM of the antioxidant NAC. Data are mean ± SEM of n = 3 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

3. (top) Scheme illustrating the experimental design: human MRC5 fibroblasts were irradiated with 20‐Gy X‐ray and treated at day 3 after IR with 10 μM of the ATM inhibitor (ATMi) KU55933 and/or 100 nM rapamycin (Rap); (bottom) Effect of single or combined inhibition of ATM and mTORC1 on the mean number (N) of γH2A.X foci, Sen‐β‐Gal activity and p21 expression in proliferating and senescent (10 days after 20‐Gy X‐ray) MRC5 fibroblasts. Data are mean ± SEM of n = 3 independent experiments. Western blots are representative of three independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

4. ROS levels (DHE intensity) in proliferating and senescent (3 days after 10‐Gy X‐ray) wild‐type and PGC‐1β ^−/− MEFs (top) and PGC‐1β‐overexpressing MEFs (bottom). Data are mean ± SEM of n = 3 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

5. (top) Representative images and quantification of immunofluorescence staining against FLAG‐PGC‐1β (red) and 53BP1 (green) in PGC‐1β‐overexpressing MEFs cultured at 3% O₂, with or without 250 μM of the antioxidant Trolox (scale bar = 10 μm). Data are mean ± SEM, n = 3 independent experiments; asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

6. (top) Effect of overexpression of mutated Rheb (N153T) and (bottom) effect of 100 nM rapamycin on the mean number (N) of 53BP1 foci in proliferating and senescent (3 days after 10‐Gy X‐ray) wild‐type and PGC‐1β ^−/− MEFs. Data are mean ± SEM of n = 3 independent experiments (at least 125 cells were analysed per condition). Asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

Figure 6. mTORC1‐PGC‐1β regulates mitochondrial content and contributes to senescence in vivo

1. (top) Representative image of immunofluorescence double staining for the mitochondrial protein MT‐CO1 and the DDR marker γH2A.X (scale bar = 5 μm) and (bottom) quantification of MT‐CO1 intensity versus the number of γH2A.X foci in hepatocytes from 12‐month‐old mice. Data are mean ± SEM of n = 3 mice (at least 30 cells were analysed per mouse); asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

2. Representative Westerns blots of pS6, S6, p21, PGC‐1β, MT‐CO1 and NDUFB8 protein expression in mouse liver tissue at 3 and 12 months of age. Data are from n = 3 mice per group.

3. Effect of 4 months of rapamycin‐supplemented diet on PGC‐1β expression in the liver tissue of 16‐month‐old mice. Data are from n = 3 mice per group.

4. (top) Quantification of mitochondrial number per cross‐section and mitochondrial volume fraction (%V_v) in hepatocytes from 16‐month‐old mice with or without 4 months of rapamycin diet. Data are mean ± SEM of n = 3 mice per group (at least 20 cells were analysed per mouse); (bottom) Representative electron micrographs of hepatocytes from 16‐month‐old mice with or without 4 months of rapamycin diet (mitochondria are labelled in pink). Scale bar = 5 μm. Asterisk denotes a statistical significance at P < 0.05 using two‐tailed t‐test.

5. mtDNA copy number (measured by qPCR) in mice liver tissue at 3, 12, 16 months and at 16 months after 4 months of rapamycin diet. Data are mean ± SEM of n = 3–4 mice per group; Asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

6. Oxygen consumption rates (OCR) in isolated liver mitochondria from 16‐month‐old mice fed with or without rapamycin for 4 months, in the presence of pyruvate/malate. State III was induced by the injection of ADP. State IV was induced by the inhibition of the ATP synthase with oligomycin, and uncoupled respiration rates were determined by the injection of FCCP. Antimycin A (AA) was used to determine background, non‐mitochondrial OXPHOS, OCR. Data are mean ± SEM of n = 5 mice per group.

7. (top) Representative immuno‐TeloFISH images of hepatocytes from 3‐ and 15‐month‐old mice with or without rapamycin (12‐month diet). Co‐localizing foci are amplified in the right panel (amplified images are from single Z‐planes where co‐localization was found); (bottom) Dot plot graph of telomere‐associated foci (TAF) in 3‐, 15‐ and 16‐month‐old mice (15‐ and 16‐month‐old mice were fed with rapamycin for 12 and 4 months, respectively). Data are from n = 3 to 9 mice per group (at least 50 cells were analysed per mice). Values are the mean for individual animals, with the horizontal line representing group mean. Asterisk denotes P < 0.05 using an independent samples t‐test.

8. (top) 4‐month‐old and 15‐month‐old mice livers [control (−) or rapamycin (+)] stained with Sen‐β‐Gal solution (Sen‐β‐Gal activity is evidenced by blue staining). Data are from n = 3 mice per group; (bottom) Representative image showing Sen‐β‐Gal staining (Sen‐β‐Gal activity is evidenced by blue staining) in hepatocytes and corresponding immuno‐TeloFISH (arrows represent co‐localizing foci); asterisks denote a statistical significance at P < 0.05 using one‐way ANOVA.

9. Representative Western blot showing the effect of 4‐month rapamycin feeding on p21 expression in 16‐month‐old mice. Data are from n = 3 mice per group.

10. Effect of 4‐month rapamycin feeding on mRNA expression of the SASP components CXCL1,CXCL5 and inhibin A in liver tissue of 16‐month‐old mice. Data are from n = 5 mice per condition; asterisks denote a statistical significance at P < 0.05 using two‐tailed t‐test.

11. Mean number of TAF in hepatocytes of wild‐type and PGC‐1β ^−/− mice with 18 months of age. Data are mean ± SEM of n = 4 mice per group; asterisks denote a statistical significance at P < 0.05 using two‐tailed t‐test.

12. Scheme represents overall hypothesis: feedback loop between DDR, mTORC1 and mitochondrial biogenesis stabilizes cellular senescence, which are key factors for the development of the senescence‐associated pro‐oxidant and pro‐inflammatory phenotypes.