p53-dependent release of Alarmin HMGB1 is a central mediator of senescent phenotypes

Abstract

Cellular senescence irreversibly arrests proliferation in response to potentially oncogenic stress. Senescent cells also secrete inflammatory cytokines such as IL-6, which promote age-associated inflammation and pathology. HMGB1 (high mobility group box 1) modulates gene expression in the nucleus, but certain immune cells secrete HMGB1 as an extracellular Alarmin to signal tissue damage. We show that nuclear HMGB1 relocalized to the extracellular milieu in senescent human and mouse cells in culture and in vivo. In contrast to cytokine secretion, HMGB1 redistribution required the p53 tumor suppressor, but not its activator ATM. Moreover, altered HMGB1 expression induced a p53-dependent senescent growth arrest. Senescent fibroblasts secreted oxidized HMGB1, which stimulated cytokine secretion through TLR-4 signaling. HMGB1 depletion, HMGB1 blocking antibody, or TLR-4 inhibition attenuated senescence-associated IL-6 secretion, and exogenous HMGB1 stimulated NF-κB activity and restored IL-6 secretion to HMGB1-depleted cells. Our findings identify senescence as a novel biological setting in which HMGB1 functions and link HMGB1 redistribution to p53 activity and senescence-associated inflammation.

Figure 1.

HMGB1 relocalizes in SEN human cells. (A) Whole-cell lysates (WC) of 10⁵ WI-38 human fibroblasts, PRE or made SEN by X-irradiation (XRA), replicative exhaustion (REP), p16^INK4a overexpression (p16) or oncogenic RAS, were analyzed for HMGB1 and actin (control) by Western blotting. (B) IMR-90 fibroblasts described in A were immunostained for HMGB1 (red). Nuclei were stained with DAPI (blue). Parallel cells were labeled with BrdU (48 h) and stained for SA-β-Gal activity; values shown are percentage of cells. Bar, 10 μM. (C) Cytotoxicity was determined by lactate dehydrogenase in CM from IMR90 cells. Error bars are the standard deviation of triplicate samples. (D) At the indicated time after XRA, we analyzed CM from IMR90 cells for HMGB1 by ELISA. Shown is the fold increase over cells made quiescent by culture in 0.2% serum for 3 d. Error bars are the SEM of triplicate samples. (E) CM and whole-cell lysates (WC) were collected from PRE and duplicate REP IMR-90 cells and analyzed by Western blotting for HMGB1 and actin. Expression relative to actin is shown below each lane.

Figure 2.

HMGB1 relocalizes in SEN mouse cells. (A) PRE or SEN (XRA) MEFs were immunostained for HMGB1 (red). Nuclei were stained with DAPI (blue). Bar, 10 μM. (B) Lysates from PRE or SEN (XRA) MEFs were analyzed by Western blotting for HMGB1 and actin (control). (C) CM from PRE or XRA cells were analyzed for HMGB1 by ELISA. Error bars are the average of duplicate samples. (D) 6–8-wk-old C57BL/6 mice were unirradiated (control) or sublethally irradiated (IR). Kidneys collected 1 wk later were immunostained for HMGB1 (red); nuclei were stained with DAPI (blue). Bar, 25 μM. (E) Kidneys from mice in B were immunostained for nuclear (Nuc), or nuclear + cytoplasmic (Nuc + Cyto) HMGB1. More than 200 nuclei were scored blinded from 2–3 control or irradiated (IR) mice; error bars are SEM. (F) Kidneys from 4- and 32-mo-old C57BL/6 mice were immunostained for HMGB1 (red); nuclei were counterstained with DAPI (blue). Bars, 10 μM. (G) Sera from mice at the indicated ages were assayed for HMGB1 by ELISA. Kruskal-Wallis test was used to determine significance (***, P < 0.001).

Figure 3.

Senescent cells secrete acetylated HMGB1. (A) WI-38 cells were given PBS (−, control) or 200 ng/ml trichostatin A (TSA) (+) for 4 h, and immunostained with pan-acetyl-lysine (green)– and HMGB1 (red)–specific antibodies. Nuclei were stained with DAPI (blue). Bar, 20 μM. (B) CM from PRE and SEN (XRA) IMR-90 cells were immunoprecipitated by mouse IgG (control) or acetyl-lysine (Ac-L) antibody. Precipitates and 5% of unprecipitated (Input) CM were analyzed by Western blotting for HMGB1. % total = amount of acetylated HMGB1 immunoprecipitated compared with total HMGB1 in CM.

Figure 4.

HMGB1 is actively exported from senescent cell nuclei. (A) SEN (XRA) HCA2 cells were cultured for 3 h with DMSO (−, control) or 20 nM leptomycin B (LMB) (+) and stained for HMGB1 (red) and 53BP1 (green). Nuclei were stained with DAPI (blue). Bar, 10 μM. (B) REP SEN HCA2 cells were cultured as in A and stained for HMGB1 (red). Nuclei were stained with DAPI (blue). Bar, 10 μM. (C) HCA2 cells were given 5 nM staurosporine or 20 nM LBM for 6 h and stained for active caspase-3 (red; % positive cells shown). Nuclei were stained with DAPI (blue). Bar, 10 μM. (D) PRE (XRA−) or SEN (XRA+) IMR-90 cells were given DMSO (−) or 20 nM LMB (+) for 3 h. Lysates were analyzed for HMGB1 and actin (control) by Western blotting. Expression normalized to actin and relative to PRE cells (1.0) is indicated. (E) CM from D were analyzed for HMGB1 by ELISA. Error bars = SEM of triplicate samples. (F) HCA2 cells were cultured without or with 0.5 mM H₂O₂ for 2 h, washed, and incubated for 16 h with DMSO or LMB. CM were analyzed by Western blotting for HMGB1. Data are representative of at least two independent experiments.

Figure 5.

HMGB1 depletion or overexpression induces senescence. (A) Lysates from IMR-90 cells infected with lentiviruses carrying no insert (Vector) or HMGB1 shRNAs (shRNA1, 2) were analyzed for HMGB1 and actin (control) by Western blotting. Shown is expression level relative to Vector. (B) Cells in A were immunostained for HMGB1 (red). Nuclei were stained with DAPI (blue). REP SEN cells are shown for comparison. Bar, 10 µM. (C) Cells in B were scored for SA-β-Gal, BrdU (24 h pulse), and SAHF. Error bars = SEM of two independent experiments. (D) Lysates from IMR-90 cells infected with lentivirus expressing no insert (Vector) or HMGB1 (OE) were analyzed for HMGB1 and actin (control) by Western blotting. Shown is expression level relative to Vector. (E) Cells in D were immunostained for HMGB1 (red). Nuclei were stained with DAPI (blue). REP SEN cells are shown for comparison. Bar, 10 µM. (F) Cells in E were scored for SA-β-Gal, BrdU (24 h pulse), and SAHFs. REP SEN cells are shown for comparison. Error bars = SEM of two independent experiments. (G) Lysates from PRE IMR-90 cells or cells induced to senesce by XRA (XRA), HMGB1-overexpression (OE), or HMGB1-depletion (KD) were analyzed for p53, p16^INK4a, and actin (control) by Western blotting. (H) CM from PRE or SEN (REP, XRA, and HMGB1 OE or KD) IMR-90 cells were analyzed for IL-6 by ELISA. Shown is a representative of two experiments. Error bars = SEM of duplicate determinations. (I) HCA2 cells expressing an NF-κB luciferase reporter were infected with lentiviruses carrying GFP shRNA (control), RelA shRNA, HMGB1 cDNA (OE), or HMGB1 shRNA (KD). 7 d after selection, shGFP and shRel A cells were given TNF for 24 h in serum-free media. Cells were lysed and luciferase activity measured (fold change over cells given BSA (control) protein). One-way ANOVA was used to analyze groups; P < 0.001. (J) Cells infected with HMGB1 OE lentivirus were stained for HMGB1 (red) and nuclei (DAPI; blue). Bar, 10 μM. (K) CM from cells infected with insertless (control) or HMGB1-expressing (OE) lentiviruses were assayed for HMGB1 by ELISA. Error bars = SEM of duplicate determinations. (L) IMR-90 cells infected with lentiviruses carrying no insert (V) or overexpressing HMGB1 (OE) were cultured for 5 d with mouse IgG or HMGB1 blocking antibody (BA) and assessed for EdU incorporation. Error bars = SEM of duplicate determinations.

Figure 6.

p16^INK4a does not regulate HMGB1 in SEN cells. (A) Lysates from IMR-90 cells infected with lentiviruses carrying no insert (V) or p16^INK4a shRNA (p16 KD) were analyzed for p16^INK4a and actin (control) by Western blotting. (B) PRE and SEN (XRA) IMR-90 cells depleted of p16^INK4a were immunostained for HMGB1 (red). Nuclei were stained with DAPI (blue). Bars, 10 µM. (C) Lysates from PRE IMR-90 cells or cells infected with lentiviruses carrying no insert (V) or p16^INK4a shRNA (p16 KD). p16 KD cells were made senescent by X-irradiation (XRA) and analyzed 10 d later. Cells were analyzed for HMGB1 and actin (control) by Western blotting. (D) IMR90 cells depleted of p16^INK4a were infected with lentiviruses carrying no insert (V), HMGB1 cDNA (OE), or HMGB1 shRNA (KD) and stained for SA-β-Gal. Error bars = SEM of two independent experiments. (E) IMR-90 cells expressing or depleted of p16^INK4a were infected with lentiviruses carrying no insert (V), HMGB1 shRNA (KD), or HMGB1 cDNA (OE). After 10 d, cells were scored for SAHFs. Error bars = SEM of two independent experiments. (F) IMR-90 cells infected with lentiviruses carrying GFP (control) or p16 (OE) were scored for SA-β-Gal and BrdU incorporation (24 h pulse) and immunostained for HMGB1 (red). Nuclei were stained with DAPI (blue). XRA-induced SEN cells are shown for comparison. Bars, 10 µM.

Figure 7.

HMGB1 relocalization and senescence caused by altered HMGB1 expression requires p53. (A) Lysates (WC) and CM from XRA SEN HCA2 cells infected with control (−) or GSE (+) expressing lentiviruses were analyzed for HMGB1 and tubulin (control) by Western blotting. The fraction of detectable protein is given below each lane. (B) CM from PRE or SEN (XRA) HCA2 cells expressing control (V) or GSE-expressing lentiviruses were analyzed for HMGB1 by ELISA. (C) Lysates (WC) or CM from XRA SEN IMR90 cells infected with control (−) or shp53 (+) expressing lentiviruses were analyzed for HMGB1 and actin (control) by Western blotting. The fraction of detectable protein is given below each lane. (D) CM from PRE or SEN (XRA) IMR90 cells expressing control (V) or shp53 (KD) lentiviruses were analyzed for HMGB1 by ELISA. (E) REP SEN HCA2 cells infected with control (GFP; −) or GSE (+) expressing lentivirus were immunostained for HMGB1 (red). Nuclei were stained with DAPI (blue). Bars, 10 µM. (F) IMR-90 and HCA2 cells expressing (+) or depleted of p53 (−) were infected with a lentivirus carrying no insert (Vector), oncogenic RAS, HMGB1 shRNA (KD), or HMGB1 (OE), and scored for SA-β-Gal 3 d later. Error bars = SEM of two independent experiments. (G) Cells in B were infected with lentiviruses carrying no insert (Vector), HMGB1 (OE), or HMGB1 shRNA (KD). 6 d after infection, cells were immunostained for HMGB1 (red). Nuclei were stained with DAPI (blue). Bars, 10 µM. (H) CM from cells infected with lentiviruses carrying no insert (V), shp53 (p53 KD), and/or HMGB1 (OE) were analyzed for HMGB1 by ELISA. Error bars = SEM of triplicate samples.

Figure 8.

ATM is dispensable for HMGB1 secretion. (A) PRE or SEN (XRA) HCA2 cells and two A-T derived fibroblast strains (AT2SF, AT5283) were immunostained for HMGB1 (red). Nuclei were stained with DAPI (blue). Bars,10 µM. (B) Cells in A were scored for SA-β-Gal and BrdU incorporation (24-h pulse). Error bars = SEM of two experiments. (C) CM from cells in A were assessed for HMGB1 secretion by ELISA. Error bars = SEM of two experiments. (D) Lysates from HCA2 cells infected with lentiviruses carrying GFP (Control) or ATM (KD) shRNA and lysates from A-T fibroblasts (AT2SF) were analyzed for HMGB1 and actin by Western blotting. HMGB1 signals were normalized to actin, and are shown relative to shGFP.

Figure 9.

HMGB1 regulates SASP factors. (A) PRE and SEN (XRA) WI-38 cells were immunostained for HMGB1 (red). DAPI stained the nuclei (blue). Images were uniformly enhanced to detect HMGB1 in SEN cells. Bar, 10 µM. (B) XRA SEN WI-38 cells were infected with lentiviruses carrying GFP or either of two HMGB1 (HMGB1-1, HMGB1-2) shRNAs. 4 d later, CM was collected and analyzed by ELISA for IL-6 and MMP-3. Shown are fold changes relative to shGFP. Error bars = SEM of two experiments. (C) PRE or SEN (XRA) HCA2 cells were incubated with 2, 4, and 8 µg/ml mouse IgG or HMGB1 blocking antibody. CM were analyzed by ELISA for IL-6. Error bars = SEM of triplicate samples. **, P < 0.02. (D) CM was collected from IMR90 cells that were mock irradiated, or 24 h or 7 d after irradiation. CM were heated in the presence (+) or absence (−) of 5 mM DTT, then analyzed by Western blotting. Positions of reduced (all thiol) and oxidized (disulfide) HMGB1 are indicated. (E) IMR90 cells were infected with control or shTLR-4 expressing lentiviruses and assessed for relative TLR-4 mRNA and IL-6 secretion. IL-6 secretion was assessed in mock (M) or X-irradiated (XRA) SEN cells expressing no insert or TLR-4 shRNA (TLR4 KD), or cultured in the presence (+) or absence (−) of CLI-95, a TLR-4 inhibitor. Shown is the fold expression relative to mock-irradiated cells without CLI-95. Error bars = SEM of triplicate samples. (F) IMR90 cells were cultured with 400 ng/ml BSA, 20 ng/ml TNF, or 400 ng/ml recombinant HMGB1 for 20 min and immunostained for NF-κB (p65 subunit; green). DAPI stained the nuclei (blue). Bar, 10 µM. (G) HCA2 cells infected with shGFP or shHMGB1-expressing lentiviruses were selected for 7 d and cultured for 24 h with indicated concentrations of BSA, recombinant HMGB1, TNF, or IL-1α. CM were analyzed by ELISA for IL-6. Error bars = SEM of triplicate samples. (H) HCA2 cells expressing an NF-κB luciferase reporter were infected with lentiviruses expressing shGFP or shHMGB1. 7 d after selection, cells were cultured for 24 h with the indicated proteins in serum free media at the concentrations in G. Cells were lysed and luciferase was measured. Luciferase levels are fold changes over cells incubated with BSA. One-Way ANOVA was used to analyze the groups; P < 0.001.