A stress-induced cilium-to-PML-NB route drives senescence initiation

Abstract

Cellular senescence contributes to tissue homeostasis and age-related pathologies. However, how senescence is initiated in stressed cells remains vague. Here, we discover that exposure to irradiation, oxidative or inflammatory stressors induces transient biogenesis of primary cilia, which are then used by stressed cells to communicate with the promyelocytic leukemia nuclear bodies (PML-NBs) to initiate senescence responses in human cells. Mechanistically, a ciliary ARL13B-ARL3 GTPase cascade negatively regulates the association of transition fiber protein FBF1 and SUMO-conjugating enzyme UBC9. Irreparable stresses downregulate the ciliary ARLs and release UBC9 to SUMOylate FBF1 at the ciliary base. SUMOylated FBF1 then translocates to PML-NBs to promote PML-NB biogenesis and PML-NB-dependent senescence initiation. Remarkably, Fbf1 ablation effectively subdues global senescence burden and prevents associated health decline in irradiation-treated mice. Collectively, our findings assign the primary cilium a key role in senescence induction in mammalian cells and, also, a promising target in future senotherapy strategies.

Fig. 1. Deficiency of Joubert syndrome GTPase ARL3 or ARL13B promotes DNA damage-induced senescence.

a Immunofluorescent images for changes of primary cilia in IMR-90 cells after senescence-induction by irradiation. Cilia were labeled with acetylated tubulin. Scale bar, 10 μm. Three experiments were repeated independently with similar results. b Western blot showing changes in ciliary protein levels after senescence-induction in IMR-90 cells by irradiation. Three experiments were repeated independently with similar results. c–e SA-β-gal staining (n>100 cells per experiment) (c), western blot of senescence markers (d) and relative mRNA levels of SASP genes (e) in IMR-90 cells stably expressing control shRNA or shARL3 at day 10 post-irradiation. f–h SA-β-gal staining (f), western blot of senescence markers (g) and relative mRNA levels of SASP genes (h) in IMR-90 cells stably expressing control shRNA or shARL13B at day 10 post-irradiation. Scale bar, 200 μm. Results (c-h) from n=3 independent experiments were statistically analyzed and plotted as means ± SEM. Brown-Forsythe and Welch ANOVA tests was used for c and f. Two-tailed Student’s unpaired t-test was used for analysis in d and g. Two-way ANOVA followed by Bonferroni multiple-comparison analysis was used for e and h. Source data are provided as a Source Data file.

Fig. 2. Transition fiber component FBF1 is a key player in DNA damage-induced senescence.

a–c SA-β-gal staining (n>100 cells per experiment) (a), western blot of senescence markers (b) and relative mRNA levels of SASP genes (c) in IMR-90 cells stably expressing control shRNA or shFBF1 at day 10 post-irradiation. Scale bar, 200 μm. d, e TUNEL assay (n=30 cells per experiment) (d) and viability assay (e) in IMR-90 cells stably expressing control shRNA or shFBF1 at day 3 post-irradiation. f–h, SA-β-gal staining (f), western blot of senescence markers (g) and relative mRNA levels of SASP genes (h) in IMR-90 cells overexpressed with plasmid pCDH vector or pCDH-FBF1-Myc at day 7 after irradiation. Scale bar, 200 μm. All results from n = 3 independent experiments were statistically analyzed and plotted as means ± SEM. Brown-Forsythe and Welch ANOVA tests was used for a and f. Two-way ANOVA followed by Bonferroni multiple-comparison analysis was used for c. Two-tailed Student’s unpaired t-test was used for analysis in b, d, e and h. Source data are provided as a Source Data file.

Fig. 3. ARL13B and ARL3 negatively regulate DNA damage-induced PML-NB translocation of FBF1, a prerequisite for stress-induced PML-NB upregulation.

a Immunofluorescent images in IMR-90 cells after irradiation using antibodies against FBF1 and PML. Basal bodies were labeled with HYLS1. Scale bar, 10 μm. Number of PML NBs per nucleus (n = 30 cells) and percentage of cells containing ≥3 FBF1 foci after irradiation were quantified. b Structured illumination microscopic images of the nucleus of a senescent IMR-90 cell stained for FBF1 and PML. Scale bar, 2 μm. Three experiments were repeated independently with similar results. c Localization of FBF1 and PML in control or FBF1-knockdown IMR-90 cells at day 10 after irradiation. n=30 cells. Scale bar, 10 μm. d Localization of FBF1 and PML in control, ARL3-knockdown, or ARL13B-knockdown IMR-90 cells at day 10 after irradiation. n = 27 cells. Scale bar, 10 μm. All results from n = 3 independent experiments were statistically analyzed and plotted as means ± SEM. One-way ANOVA followed by Bonferroni multiple-comparison analysis was employed for c and d. Source data are provided as a Source Data file.

Fig. 4. DNA damage induces FBF1-UBC9 association and FBF1 SUMOylation, which are negatively regulated by the ARL13B-ARL3 GTPase cascade.

a In vitro SUMOylation assay of FBF1 and western blot using a SUMO1 antibody. GST tagged 598–1148 domain FBF1 was purified using E. coli. b Endogenous UBC9 immunoprecipitates with FBF1 in RCTE cells at day 7 after irradiation. c Endogenous FBF1 immunoprecipitates with SUMO1 in RCTE cells at day 7 after irradiation. d Endogenous UBC9 immunoprecipitates with ARL3 in RCTE cells. e GST pull-down assay shows the ARL3 and UBC9 interaction. f Endogenous FBF1 immunoprecipitates with UBC9 in WT or ARL3^-/- RCTE cells at day 7 after irradiation. g Endogenous FBF1 immunoprecipitates with SUMO1 in WT or ARL3^-/- RCTE cells with or without IR treatment. h ARL3 inhibited the interaction between FBF1 and UBC9 by co-immunoprecipitation in 293T cells. i ARL3 reduced the interaction between FBF1 and UBC9 by GST pull-down assay. GST-tagged C-terminal fragment (598-1148) of FBF1, His-ARL3 and His-UBC9 were purified using E. coli. j In vitro SUMOylation assay was performed in a reaction with GST-tagged C-terminal fragment (598-1148) of FBF1 and His-ARL3 purified in E. coli. Western blot was detected by SUMO1, GST and His antibody. k Endogenous UBC9 immunoprecipitates with Myc-ARL3 or FBF1 in ARL3^WT, ARL3^DA or ARL3^DN overexpression RCTE cells at day 7 after irradiation. Three experiments were repeated independently with similar results (a–k). Source data are provided as a Source Data file.

Fig. 5. DNA damage abolishes the ciliary SUMO protease SENP1 and enhances FBF1 SUMOylation.

a Immunofluorescent images for SENP1 staining in RCTE cells. Basal bodies were labeled with HYLS1. Scale bar, 10 μm. b Structured illumination microscopic images stained for FBF1 and SENP1 at ciliary base in IMR-90 cells. Scale bar, 1 μm. c Co-immunoprecipitation of SENP1 and FBF1 in 293T cells. d Endogenous SENP1 or FBF1 immunoprecipitates with FBF1 or SENP1 in RCTE cells. e Immunofluorescent images in RCTE cells at day 7 without or with irradiation treatment using antibodies against SENP1 and FBF1. Basal bodies were labeled with HYLS1. Scale bar, 10 μm. f Endogenous FBF1 immunoprecipitates with SENP1 in RCTE cells with or without irradiation. g SENP1 diminished FBF1 SUMOylation in a co-immunoprecipitation assay in 293T cells. h Immunofluorescent images for FBF1, PML and YFP in WT or ARL3^−/− RCTE cells co-transfected with CFP-FRB-CEP170c and YFP-FKBP-SENP1 plasmids at day 7 after irradiation. n=30 cells. Scale bar, 10 μm. i Proposed model: DNA-damage stress strongly downregulates ARLs, enables FBF1-UBC9 interaction, and results in FBF1 SUMOylation. SUMOylated FBF1 translocates from the ciliary base to PML-NBs, which is required for stress-induced PML-NB upregulation and senescence initiation. Three experiments were repeated independently with similar results (a–h). One-way ANOVA followed by Bonferroni multiple-comparison analysis was employed for h. Source data are provided as a Source Data file.

Fig. 6. Fbf1 ablation protects mice from IR-induced senescence and associated frailty.

a Representative images of mice 3 months after IR. b The change in body weight of mice with 3 months after IR. n = 10 mice. c at 2 weeks after IR, liver was analyzed by immunohistochemistry using Il1α antibody or p16 antibody and quantified; Scale bar: 200 µm. n >80 cells were counted in n = 5 different fields in n = 3 mice’s Il1α staining and n = 8 different fields in n = 4 mice’s p16 staining d, e Lysates obtained from lung (d) and fat (e) tissue of WT mice, or Fbf1^tm1a/tm1a mice 2 weeks after IR were analyzed for the indicated proteins by western blot. f, g SASP-related genes were assessed by RT-qPCR in extracts of lung (f) and fat (g) tissue of WT mice or Fbf1^tm1a/tm1a mice 2 weeks after IR. n = 8 mice. Results from 3 independent experiments were statistically analyzed and plotted as means ± SEM. h–j Maximal speed (h), grip strength (i) and treadmill endurance (j) of 12-month-old WT mice or Fbf1^tm1a/tm1a C57BL/6 mice, 6 months after exposure to mock or sublethal dose of total-body IR (n = 10 for both groups, 5 males and 5 females). Values are expressed as means ± SEMs. Two-tailed Student’s unpaired t-test was used for analysis in a, c, f and g. One-way ANOVA followed by Bonferroni multiple-comparison analysis was employed for h–j. Source data are provided as a Source Data file.