Alkylating DNA damage stimulates a regulated form of necrotic cell death

Abstract

Necrosis has been considered a passive form of cell death in which the cell dies as a result of a bioenergetic catastrophe imposed by external conditions. However, in response to alkylating DNA damage, cells undergo necrosis as a self-determined cell fate. This form of death does not require the central apoptotic mediators p53, Bax/Bak, or caspases and actively induces an inflammatory response. Necrosis in response to DNA damage requires activation of the DNA repair protein poly(ADP-ribose) polymerase (PARP), but PARP activation is not sufficient to determine cell fate. Cell death is determined by the effect of PARP-mediated beta-nicotinamide adenine dinucleotide (NAD) consumption on cellular metabolism. Cells using aerobic glycolysis to support their bioenergetics undergo rapid ATP depletion and death in response to PARP activation. In contrast, cells catabolizing nonglucose substrates to maintain oxidative phosphorylation are resistant to ATP depletion and death in response to PARP activation. Because most cancer cells maintain their ATP production through aerobic glycolysis, these data may explain the molecular basis by which DNA-damaging agents can selectively induce tumor cell death independent of p53 or Bcl-2 family proteins.

Figure 1.

Cells deficient in p53 or Bax/Bak are susceptible to DNA-alkylating agents. MEFs from wild-type, p53^-/-, and bax^-/- bak^-/- mice were treated with staurosporine (A), etoposide (B), nitrogen mustard (C), and MNNG (D) as described in Materials and Methods. Drug concentrations are indicated in the individual panels. Cell survival was determined for triplicate samples by PI exclusion at 20 h following treatment. Data are presented as mean ± S.D. and are representative of three independent experiments.

Figure 2.

Alkylating DNA damage results in PARP activation and bioenergetic collapse. (A) MNNG activates PARP in both wild-type and bax^-/- bak^-/- cells. Wild-type and bax^-/- bak^-/- MEFs were treated with MNNG (0.5 mM) for the indicated periods of time. Cells were lysed and immunoblotting was performed using an antibody against poly(ADP-ribose) (PAR). The asterisk marks a nonspecific band. (B,C) Depletion of NAD and ATP in response to PARP activation. Wild-type and bax^-/- bak^-/- cells were treated with MNNG at indicated concentrations for 30 min, alone or together with PARP inhibitor DPQ. The cellular NAD (B) and ATP (C) levels were determined. Concentrations of NAD and ATP were normalized with that of untreated cells.

Figure 3.

PARP inhibition results in resistance to MNNG-induced cell death. (A) PARP-1 shRNA in bax^-/- bak^-/- cells. Stable clones were selected from bax^-/- bak^-/- MEFs transfected with vector or PARP-1 hairpins. Cell lysates were made from a vector cell line or PARP-1 hairpin cell lines. Immunoblotting was performed using an anti-PARP-1 antibody, and an anti-Tom20 antibody as a control for equal loading. Note that PARP-1 expression was suppressed significantly in Clone HP17, moderately in HP11, and not affected in HP23. (B) PARP activity was determined using triplicate samples by immunoblotting using an anti-PAR antibody. (C) ATP levels were measured after 30 min of treatment with MNNG at the indicated concentrations. (D) Cell survival was determined by PI exclusion 20 h after MNNG treatment. (E,F) Wild-type, bax^-/- bak^-/-, parp-1^-/-, and Clone HP17 (bax^-/- bak^-/-, shPARP-1) MEFs were treated with MNNG (0.5 mM) alone or together with PARP inhibitor DHIQ, or in the presence of staurosporine (STS). Twenty-four hours later, cells were photographed under a phase-contrast filter (E), and cell survival was determined by PI exclusion (F). (G) Wild-type, bax^-/- bak^-/-, parp-1^-/-, and Clone HP17 MEFs were treated with MNNG (0.25 mM). Cell survival was measured over time by PI exclusion.

Figure 3.

PARP inhibition results in resistance to MNNG-induced cell death. (A) PARP-1 shRNA in bax^-/- bak^-/- cells. Stable clones were selected from bax^-/- bak^-/- MEFs transfected with vector or PARP-1 hairpins. Cell lysates were made from a vector cell line or PARP-1 hairpin cell lines. Immunoblotting was performed using an anti-PARP-1 antibody, and an anti-Tom20 antibody as a control for equal loading. Note that PARP-1 expression was suppressed significantly in Clone HP17, moderately in HP11, and not affected in HP23. (B) PARP activity was determined using triplicate samples by immunoblotting using an anti-PAR antibody. (C) ATP levels were measured after 30 min of treatment with MNNG at the indicated concentrations. (D) Cell survival was determined by PI exclusion 20 h after MNNG treatment. (E,F) Wild-type, bax^-/- bak^-/-, parp-1^-/-, and Clone HP17 (bax^-/- bak^-/-, shPARP-1) MEFs were treated with MNNG (0.5 mM) alone or together with PARP inhibitor DHIQ, or in the presence of staurosporine (STS). Twenty-four hours later, cells were photographed under a phase-contrast filter (E), and cell survival was determined by PI exclusion (F). (G) Wild-type, bax^-/- bak^-/-, parp-1^-/-, and Clone HP17 MEFs were treated with MNNG (0.25 mM). Cell survival was measured over time by PI exclusion.

Figure 4.

PARP-mediated cell death is necrotic. (A) Wild-type and bax^-/- bak^-/- cells were treated with 0.5 mM MNNG, or 2 μM staurosporine (STS). Transmission electron microscopy was performed 9 h later. (B) Transformed wild-type baby mouse kidney (BMK) cells were treated with MNNG alone or together with DPQ. Immunofluorescence was performed 6 h later using an antibody against cytochrome c. Cells were counterstained with DAPI to show the nuclear morphology. (C) Wild-type and bax^-/- bak^-/- MEFs were treated with indicated agents. Cells were lysed after 8 h, and 20 μg of protein was separated on a 4%-12% gradient NuPAGE gel. PARP and Lamin B1 were detected using respective antibodies.

Figure 4.

PARP-mediated cell death is necrotic. (A) Wild-type and bax^-/- bak^-/- cells were treated with 0.5 mM MNNG, or 2 μM staurosporine (STS). Transmission electron microscopy was performed 9 h later. (B) Transformed wild-type baby mouse kidney (BMK) cells were treated with MNNG alone or together with DPQ. Immunofluorescence was performed 6 h later using an antibody against cytochrome c. Cells were counterstained with DAPI to show the nuclear morphology. (C) Wild-type and bax^-/- bak^-/- MEFs were treated with indicated agents. Cells were lysed after 8 h, and 20 μg of protein was separated on a 4%-12% gradient NuPAGE gel. PARP and Lamin B1 were detected using respective antibodies.

Figure 5.

PARP-mediated cell death is proinflammatory. (A) HMGB1 translocates from nucleus into cytosol upon MNNG treatment. bax^-/- bak^-/- MEFs were treated with 0.5 mM MNNG. Six hours later, immunofluorescence was performed using an antibody against HMGB1. Nuclei were visualized by DAPI staining. (B) HMGB1 is released into extracellular environment during MNNG-induced necrosis. Wild-type and bax^-/- bak^-/- MEFs were treated with 0.5 mM MNNG alone or together with DPQ. Culture media were collected 16 h later, and cells were lysed in RIPA buffer. HMGB1 was detected by immunoblotting in both cell lysates and culture media. (C) Inflammatory response triggered by necrosis. Wild-type, bax^-/- bak^-/-, or parp-1^-/- MEFs were treated with MNNG (0.5 mM) for 15 min or STS (2 μM) for 2 h. Drugs were washed away and cells were refed with fresh media. Twenty hours later, cell culture media were collected and added to 1 × 10⁵ macrophages. Concentration of TNFα was measured 24 h later.

Figure 6.

Vegetative cells are more resistant to PARP-mediated necrosis. (A) IL-3-dependent hematopoietic bax^-/- bak^-/- cells were cultured in media with or without IL-3 for 2 d. Cells were treated with MNNG for 15 min at indicated concentrations, and cell survival was measured by PI exclusion 24 h later. (B) Wild-type and bax^-/- bak^-/- MEFs were plated at 2 × 10⁴/well (subconfluent) or 2 × 10⁵/well (confluent) in 12-well plates. Cells were cultured for 36 h. Confluent cells were then cultured in the absence of serum for 12 h, and subconfluent cells were cultured in the presence of serum. Cells were treated with 0.5 mM MNNG for 15 min. Cell survival was determined 24 h later by PI exclusion. (C,D) IL-3-dependent bax^-/- bak^-/- cells were cultured in the presence or absence of IL-3 for 2 d. Cells were then treated with MNNG for 15 min. PARP activity was determined by immunoblotting of PAR, and NAD and ATP levels were measured and shown as the percentage of the levels in untreated cells. Cellular NAD pool was decreased in both proliferating and vegetative cells (C), whereas the cellular ATP pool was preserved in vegetative cells but not in proliferating cells (D). (E) PARP activation preferentially depletes cytosolic NAD. IL-3-dependent cells were treated with 0.5 mM MNNG for 15 min. Cells were fractionated and the NAD levels were measured in total cell lysates, as well as in the cytosolic and mitochondrial fractions. Immunoblotting of tubulin and COX IV were performed to assure the purity of the fractionation.

Figure 6.

Vegetative cells are more resistant to PARP-mediated necrosis. (A) IL-3-dependent hematopoietic bax^-/- bak^-/- cells were cultured in media with or without IL-3 for 2 d. Cells were treated with MNNG for 15 min at indicated concentrations, and cell survival was measured by PI exclusion 24 h later. (B) Wild-type and bax^-/- bak^-/- MEFs were plated at 2 × 10⁴/well (subconfluent) or 2 × 10⁵/well (confluent) in 12-well plates. Cells were cultured for 36 h. Confluent cells were then cultured in the absence of serum for 12 h, and subconfluent cells were cultured in the presence of serum. Cells were treated with 0.5 mM MNNG for 15 min. Cell survival was determined 24 h later by PI exclusion. (C,D) IL-3-dependent bax^-/- bak^-/- cells were cultured in the presence or absence of IL-3 for 2 d. Cells were then treated with MNNG for 15 min. PARP activity was determined by immunoblotting of PAR, and NAD and ATP levels were measured and shown as the percentage of the levels in untreated cells. Cellular NAD pool was decreased in both proliferating and vegetative cells (C), whereas the cellular ATP pool was preserved in vegetative cells but not in proliferating cells (D). (E) PARP activation preferentially depletes cytosolic NAD. IL-3-dependent cells were treated with 0.5 mM MNNG for 15 min. Cells were fractionated and the NAD levels were measured in total cell lysates, as well as in the cytosolic and mitochondrial fractions. Immunoblotting of tubulin and COX IV were performed to assure the purity of the fractionation.

Figure 7.

Susceptibility to PARP-mediated necrosis is controlled by cellular metabolic status. (A) Cellular glycolysis rate in the presence or absence of IL-3. IL-3-dependent bax^-/- bak^-/- cells were cultured in the presence or absence of IL-3 for 2 d. One million cells were harvested and their glycolysis rate was determined. (B) Effect of inhibition of oxidative phosphorylation on ATP levels in cells cultured with or without IL-3. Cells cultured in media with or without IL-3 were treated with oligomycin (5 μg/mL) for 30 min, and ATP levels were determined. (C) IL-3-dependent bax^-/- bak^-/- cells cultured in the presence or absence of IL-3 for 2 d. An additional population was cultured in the presence of IL-3 in complete medium made without glucose (-glucose), whereas a fourth population was cultured in the presence of IL-3 and supplemented with 10 mM methyl-pyruvate (+pyruvate) immediately prior to MNNG treatment. Cells were then treated with MNNG, and cell survival was determined by PI exclusion as described in Materials and Methods. The data presented are representative of three independent experiments. (D) Methyl-pyruvate (10 mM) was added to the IL-3-dependent wild-type cells immediately prior to MNNG treatment. Cells were treated with MNNG and cell survival was determined by PI exclusion.