Level of macroautophagy drives senescent keratinocytes into cell death or neoplastic evasion

Abstract

Senescence is a non-proliferative state reached by normal cells in response to various stresses, including telomere uncapping, oxidative stress or oncogene activation. In previous reports, we have highlighted that senescent human epidermal keratinocytes have two opposite outcomes: either they die by autophagic programmed cell death or they evade in the form of neoplastic postsenescence emergent (PSNE) cells. Herein, we show that partially reducing macroautophagy in senescent keratinocytes using 3-methyl adenine or anti-Atg5 siRNAs increases the PSNE frequency, suggesting that senescent keratinocytes have to escape autophagic cell death to generate PSNE cells. However, totally inhibiting macroautophagy impairs PSNE and leads to a huge accumulation of oxidative damages, indicating that senescent keratinocytes need to achieve quality-control macroautophagy for PSNE to occur. In accordance, we demonstrate that the progenitors of PSNE cells display a level of macroautophagy slightly lower than that of the average senescent population, which is directly dictated by their level of reactive oxygen species, their level of upregulation of MnSOD, their level of activation of NF-κB transcription factors and their level of dysfunctional mitochondria. Macroautophagy thus has antagonistic roles during senescence, inducing cell death or promoting neoplastic transformation, depending on its level of activation. Taken together, these data suggest that levels of oxidative damages and ensuing macroautophagic activity could be two main determinants of the very initial phases of neoplastic transformation by senescence evasion.

Figure 1

Growth curve and characteristics of in vitro-cultivated NHEKs. (a) Growth curve of NHEKs 4F0315 showing the exponential growth phase, the senescence plateau and the PSNE phase. (b) Percentage of SA-β-Gal-positive cells manually counted in 5–28 random fields comprising 20–262 cells. Note the senescence plateau between 18.4 and 20.3 population doublings, and PSNE from 21.4 doublings, during which some SA-β-Gal-positive senescent cells persist. Each bar represents mean±S.D. (c) Cell morphologies observed by phase-contrast microscopy. The doubling number is indicated on each photograph. At doubling 12.61, cells have the typical morphology of epithelial cells growing as islet. At doubling 18.38, most cells have increased in size and display some small vesicles. The image shown at doubling 19.73 illustrates a senescent cell, recognizable by its large size and the great number of vesicles, which has generated PSNE cells by a budding mitosis mechanism. At doubling 21.36, the culture dish comprises a mixed population of senescent and PSNE cells. From 24.38 doublings, senescent cells dying by autophagic cell death are observable among PSNE cells. Sen: senescent cell; DS: dying senescent cell. Bar represents 20 μm. (d) Quantification of cell death by Trypan blue exclusion in exponentially growing (6.5 and 8 PDs) and senescent (11, 12 and 12.3 PDs) NHEKs K18FC. Non-viable cells (blue) were counted under the microscope in four independent hemocytometer chambers for a total of at least 100 cells. The results are presented as the mean±S.D. percentage of dead cells of all counts. Significant differences are indicated with *P<0.01. NS, not significant

Figure 2

PSNE cells display less macroautophagic markers than their senescent progenitors. (a) Western blotting analysis of Beclin-1, ATG5, LC3 and LAMP-1 in protein extracts from NHEKs K18FC at the exponential growth phase, at the senescent plateau and at PSNE. PCNA was used as a proliferation marker, and GAPDH (glyceraldehyde 3-phosphate dehydrogenase) as a loading control. The anti-ATG5 antibody reveals the covalent ATG5–ATG12 complex. ns: non-specific band. (b) Band intensity of LC3 I and LC3II were quantified, and the LC3II/I ratios are given. Results are normalized to the value obtained in exponentially growing NHEKs. Results are representative of three independent experiments performed on two different donors. (c) LAMP-1 immunofluorescence assays performed on the same cells. Upper panels: Representative Apotome microscopic images. Bar represents 20 μm. Lower panel: LAMP1-staining area was quantified with ImageJ. Measures were done in five independent microscopic fields for a total of at least 100 cells for each case. The histogram represents the average±S.D. of five counts. Results are representative of at least two experiments performed on two different donors

Figure 3

Invalidating atg5 in H₂O₂-induced premature senescent NHEKs favors or inhibits PSNE according to the efficacy of siRNAs. H₂O₂-induced premature senescent NHEKs K18FC were transfected with a pool of four siRNAs targeting atg5 (#pool), two individual siRNAs (#10 and #7) targeting atg5, a pool of non-target siRNAs (siCRT) or treated only with the transfectant (Mock). (a) Verification of the efficacy of siRNAs by western blotting performed 48 h posttransfection. The anti-ATG5 antibody reveals the covalent ATG5–ATG12 complex. ns: non-specific band. (b) Twenty-four hours posttransfection, cells were seeded at low density, and PSNE frequency was measured at day 1, 2 or 3 after seeding. The counts were performed in eight independent culture dishes. The given results are the mean±S.D. of all counts. P-values calculated using the Student's t-tests are given. This experiment is representative of three independent ones. (c) Representative images of PSNE clones after fixation and coloration with crystal violet. Sen points a senescent cell, and PSNE the clone of emergent cells. Bar represents 10 μm

Figure 4

Lowering macroautophagy using 3-MA in H₂O₂-induced premature senescent NHEKs favors or limits PSNE depending on the inhibitor concentration. (a) Exponentially growing NHEKs K1MC were treated by H₂O₂ to induce premature senescence. They were then treated by 3-MA or its diluent H₂O (NT), and the proteins were extracted at different times after the beginning of the treatment. The activation of ATG5 in ATG5–ATG12 covalent complex and of LC3 in LC3-II was analyzed by western blotting. (b) Exponentially growing NHEKs K18FC were treated by H₂O₂ to induce premature senescence. They were then treated by 3-MA or its diluent H₂O, seeded at low density and monitored for emergence frequency. The counts of emerging clones were performed in four independent culture dishes at days 1, 2 and 3 postseeding. The given results are the mean±S.D. of all counts. P-values calculated using the Student's t-tests are given. The results are representative of five independent experiments

Figure 5

Inhibiting macroautophagy in senescent NHEKs favors or limits PSNE depending on the rate of inhibition. (a) NHEKs 13.20 were taken at the beginning of the senescent plateau and analyzed by flow cytometry as a function of forward scatter factor value (left histogram, the x axis represents forward scatter factor value). The delineation of S and D subpopulations is indicated. Cells of the S and D subpopulations were characterized by PI staining (right histogram; the x axis represents the PI fluorescence intensity (PI-A)). (b) Cells of the S subpopulation were sorted, seeded at low density, treated with 3-MA or its diluent H₂O and monitored for their ability to generate PSNE clones. The counts of emerging clones were performed in eight independent culture dishes. The given results are the mean±S.D. of all counts. P-values calculated using the Student's t-tests are given. These experiments are representative of two independent ones. (c) NHEKs 13.20 were taken at the beginning of the senescent plateau, treated with 3-MA or its diluent H₂O, stained by PI and analyzed by flow cytometry. The given plots of PI intensity (PI-A) were extracted using FlowJo and concern only the S subpopulation

Figure 6

Maintaining an autophagic flux is indispensable for postsenescence emergence. NHEKs 13.20 were taken at the beginning of the senescent plateau, treated with Bafilomycin A1 or its diluent dimethyl sulfoxide (DMSO), stained by Lysotacker or PI and analyzed by flow cytometry. (a) Verification of the efficacy of Bafilomycin A1 on the autophagic activity of senescent NHEKs. The given values of Lysotracker intensity (FITC-A) were extracted using FlowJo. (b) Analysis of the effect of Bafilomycin A1 on the viability of senescent NHEKs. The given values of PI intensity (PI-A) were extracted using FlowJo and concern only the S subpopulation. (c) Emergence frequency of senescent NHEKs 13.20 treated by Bafilomycin A1 or its diluent DMSO. The counts of PSNE clones were performed in 4–8 independent culture dishes. The given results are the mean±S.D. of all counts. P-values calculated using the Student's t-tests are given. These experiments are representative of two independent ones. (d) Emergence frequency of H₂O₂-induced premature senescent NHEKs 13.20 treated by Bafilomycin A1 or its diluent DMSO. The counts of PSNE clones were performed in 4–8 independent culture dishes. The given results are the mean±S.D. of all counts. P-values calculated using the Student's t-tests are given. These experiments are representative of four independent ones

Figure 7

Senescent NHEKs invalidated for atg5 or treated by 3-MA accumulate altered components in a dose-dependent manner. (a) Representative Apotome microscopic images of 8-oxo-G in exponentially growing and H₂O₂-induced premature senescent NHEKs K1MC treated or not for 48 h with 1 or 5 mM 3-MA. Bar represents 20 μm. (b) Quantitative detection of 8-oxo-G by flow cytometry in H₂O₂-induced premature senescent NHEKs K1MC treated or not for 48 h with 1 or 5 mM 3-MA. (c) Representative Apotome microscopic images of aggresomes in the same cells as in panel a. Bar represents 20 μm. (d) Quantitative detection of aggresomes by flow cytometry in the same cells as in panel b. (e) Quantitative detection of aggresomes by flow cytometry in cells at the senescence plateau invalidated for atg5 as in Figure 3

Figure 8

Autophagic activity and steady-state levels of ROS in exponentially growing versus senescent NHEKs. NHEKs 13.20 at the exponential growth phase (black) and senescence plateau (red) were stained with Lysotracker and H2-DCFDA and analyzed by flow cytometry. (a) Forward scatter (FS) factor analysis. The population at the senescence plateau shows two main peaks of size; the first one corresponds to residual small growing cells; the second one (R1) corresponds to senescent cells, including the S1, S2, D1 and D2 subpopulations of Figure 9. (b) Histograms of Lysotracker (left panel) and H2-DCFDA (right panel) staining intensities of exponentially growing (dark) and R1 senescent NHEKs (red). Peak values of the Lysotracker and H2-DCFDA stainings are given

Figure 9

The probability of senescent cells to generate PSNE clones is linked to their macroautophagy and ROS levels. (a) Flow cytometric histogram of NHEKs 13.20 at the senescence plateau according to forward scatter factor (FSC-A) and showing the S and D subpopulations. (b) Flow cytometric histograms of the S (in green) and D (in blue) senescent subpopulations according to their Lysotracker staining intensity and showing the S1, S2, D1 and D2 subpopulations. (c) Mean values of the Lysotracker and H2-DCFDA staining intensities of the four subpopulations. (d) The four subpopulations were seeded in four-well plates and, every 24 h, fixed, stained with Hoechst and automatically counted as described in Materials and Methods section. (e) The four subpopulations were seeded at low density and monitored for PSNE. The counts of clones were performed in four independent culture dishes. The given results are the mean±S.D. of all counts. The indicated fold difference corresponds to the ratio of the means of S1 on S2 and D1 on D2. P-values were calculated using the Student's t-tests

Figure 10

Gradation of NF-κB activation and mitochondrial dysfunction in senescent NHEKs. (a and b) Western blotting analysis of cRel, IκBα and MnSOD in cytoplasmic and nuclear extracts or in total cell extracts of NHEKs K18FC at the exponential growth phase and at the senescent plateau. (c and d) S1, S2, D1 and D2 subpopulations of senescent NHEKs K18FC were sorted as in Figure 9, and western blotting analysis of cRel, IκBα and MnSOD was performed. GAPDH (glyceraldehyde 3-phosphate dehydrogenase) and histone H3 were used as marker and loading control for cytoplasmic and nuclear extracts, respectively. (e) Left panel: Representative confocal microscopic images of JC-1 staining assays performed in cells as in panels (c and d). Bar represents 20 μm. Right panel: quantification of the red/green ratio using ImageJ. The counts were performed in >40 cells. The given results are the mean±S.D. of all counts. P-values were calculated using the Student's t-tests