p53 enhances DNA repair and suppresses cytoplasmic chromatin fragments and inflammation in senescent cells

Abstract

Genomic instability and inflammation are distinct hallmarks of aging, but the connection between them is poorly understood. Here we report a mechanism directly linking genomic instability and inflammation in senescent cells through a mitochondria-regulated molecular circuit involving p53 and cytoplasmic chromatin fragments (CCF) that are enriched for DNA damage signaling marker γH2A.X. We show that p53 suppresses CCF accumulation and its downstream inflammatory phenotype. p53 activation suppresses CCF formation linked to enhanced DNA repair and genome integrity. Activation of p53 in aged mice by pharmacological inhibition of MDM2 reverses transcriptomic signatures of aging and age-associated accumulation of monocytes and macrophages in liver. Mitochondrial ablation in senescent cells suppresses CCF formation and activates p53 in an ATM-dependent manner, suggesting that mitochondria-dependent formation of γH2A.X + CCF dampens nuclear DNA damage signaling and p53 activity. These data provide evidence for a mitochondria-regulated p53 signaling circuit in senescent cells that controls DNA repair, genome integrity, and senescence- and age-associated inflammation, with relevance to therapeutic targeting of age-associated disease.

Fig. 1. p53 suppresses CCF formation.

A CCF staining by IF in irradiation-induced senescent IMR90 human fibroblasts. Each value represents an individual well from a culture plate, representative of n = 2 experiments. Y-axis represents total number of CCF normalized to total number of nuclei. B Western blot and CCF staining by IF in irradiation-induced senescent IMR90 human fibroblasts using timeframe described in Fig.S1K, representative of n = 2 experiments. C CCF staining by IF at indicated day after irradiation, average of 3 experiments. D CCF and FLAG staining by IF in irradiation-induced senescent IMR90 transduced with exogenous p53 or empty vector. Each value represents a separate infection in a single experiment, representative of n = 3 experiments. E Cell number as measured by number of nuclei, normalized to DMSO control for each group, representative of n = 3 experiments. F DNA replication as measured by EdU incorporation assay. Except for n = 1 proliferating control, each dot represents a separate irradiation, n = 3, representative of n = 2 experiments. G Differentially expressed genes by RNAseq, n = 3 per group with summarized KEGG ontology for each major cluster. See Supplementary Data 1 for detailed ontology. Data shown as means ± SD, asterisk(*) indicates p < 0.05 by two-tailed Student’s t-test. Prolif proliferating control, BRCT Brca1-C-terminal sequence motif, EV empty vector, NTC non-targeting control, IR ionizing radiation-induced senescence. Source data are provided as a Source Data file.

Fig. 2. p53 activation promotes DNA repair.

A CCF formation and (B) nuclear γH2A.X foci number by IF in irradiation-induced senescent IMR90 human fibroblasts, representative of n = 3 experiments. Each value represents one well of a culture plate, from a representative experiment. C WB of γH2A.X in irradiation-induced senescent IMR90 cells from Fig. S1J, n = 2 observations. D qPCR analysis of p53 target genes in cells 4 days after irradiation, treated as indicated with RG7388 or DMSO control, n = 3–6, representative of 3 experiments. E IF quantitation of CCF, nuclear γH2A.X, p21, and cell number quantified by number of nuclei in irradiation-induced senescent IMR90 cells, representative of n = 4 experiments. F Neutral comet assay in irradiation-induced senescent IMR90 cells or proliferating control cells, representative of n = 3 experiments. G NHEJ reporter assay in irradiation-induced senescent I9A human fibroblasts, n = 5 independent infections, representative of n = 3 independent experiments. Data shown as means ± SD, asterisk(*) indicates p < 0.05 by two-tailed Student’s t-test. Prolif proliferating control, NTC non-targeting control, IR ionizing radiation-induced senescence, NHEJ non-homologous end joining. Source data are provided as a Source Data file.

Fig. 3. p53 preserves genome integrity in senescent cells.

A Representative single nucleus whole-genome plots showing copy number variations and (B) predicted ploidy in irradiation-induced senescent IMR90 human fibroblasts, n = 6, 10, 6 nuclei for Prolif, Senescent, and Senescent RG7388 groups respectively. C Histogram of deletions, with each line an aggregation of all chromosomes per cell, n = 4 senescent cells. Each chromosome is divided into 18 equal bins, where bins 1 and 18 are subtelomeric and bins 4-5 are pericentromeric. D IF for CENPA in irradiation-induced senescent IMR90 cells, with arrows marking CCF, representative of n = 3 experiments and (E) ImmunoFISH for telomeres and γH2A.X, representative of n = 4 independent experiments. F Quantitation of (D and E), where each CENPA+ marker represents n = 3 separate irradiations from a representative experiment and each telomere+ marker represents n = 4 independent experiments. Data shown as means ± SD, asterisk(*) indicates p < 0.05 by two-tailed Student’s t-test. Prolif proliferating control. Source data are provided as a Source Data file.

Fig. 4. MDM2i is senomorphic in vivo.

A WB of p53 and p21 in mouse liver and (B) quantitation, n = 5, 5, 4 per group. C TAF assay in female mice showing the percentage of hepatocytes with greater than 1 TAF, n = 5, 4, 5 mice per group. D Correlation between change in gene expression as a function of age and change in gene expression as a function of HDM201 treatment, among 4912 DE genes with age, n = 5, 4, 5 mice per group with p-value calculated by simple linear regression. E Heatmap combining 776 reversed DE genes, where the change in expression of a gene is opposed by HDM201 treatment, and 58 genes not reversed, with top 5 GO biological process terms for major hierarchical clusters (see also Fig.S5E), n = 5, 4, 5 mice per group. F Ingenuity pathway analysis of DE genes showing top upstream regulators common between old vehicle vs. young vehicle (effect of age) and old HDM201 vs. old vehicle (effect of HDM201) comparisons, using a cutoff of p < 1E-14 by right-tailed Fisher’s exact test. G Representative IPA target gene heatmap of STAT1. H Flow cytometry analysis of immune cell frequencies isolated from spleen and liver, n = 5, 5, 4 mice per group. Data shown as means ± SD, asterisk(*) indicates p < 0.05 by two-tailed Mann-Whitney U test or (H) one-way ANNOVA. TAF telomere-associated DNA damage response foci, Veh vehicle control, DE differentially expressed. Source data are provided as a Source Data file.

Fig. 5. Mitochondria dampen p53 activity in senescence.

A IF representative images and (B) quantitation of CCF and proportion of the nucleus staining positive for γH2A.X in irradiation-induced senescent Parkin-overexpressing IMR90 human fibroblasts. See Fig. S6A for experimental timeline. Each value represents an individual well from a culture plate, from a representative experiment, n = 3. Numbering indicates individual siRNA sequences. C p53 target gene expression by RNAseq from ref., average of n = 3 per group shown. D qPCR of CDKN1A and corresponding IF quantitation of mitochondria content and CCF in irradiation-induced senescent Parkin-overexpressing IMR90 human fibroblasts. See Fig. S6E for experimental timeline. E Model. Data shown as means ± SD, asterisk(*) indicates p < 0.05 by two-tailed t-test, (#) in panel (B) indicates p < 0.05 vs siNTC-4. Prolif proliferating control, NTC non-targeting control, No tfn no transfection control, IR ionizing radiation-induced senescence, SASP senescence-associated secretory phenotype. Source data are provided as a Source Data file.