DNA damage or unprotected telomeres can trigger apoptosis via signaling pathways that directly sense abnormal DNA structures and activate the p53 transcription factor. We describe a p53-independent mechanism that acts in parallel to the canonical DNA damage response pathway in Drosophila to induce apoptosis after exposure to ionizing radiation. Following recovery from damage-induced cell cycle arrest, p53 mutant cells activate the JNK pathway and expression of the pro-apoptotic gene hid. Mutations in grp, a cell cycle checkpoint gene, and puc, a negative regulator of the JNK pathway, sensitize p53 mutant cells to ionizing radiation (IR)-induced apoptosis. Induction of chromosome aberrations by DNA damage generates cells with segmental aneuploidy and heterozygous for mutations in ribosomal protein genes. p53-independent apoptosis limits the formation of these aneuploid cells following DNA damage. We propose that reduced copy number of haploinsufficient genes following chromosome damage activates apoptosis and helps maintain genomic integrity.
EUKARYOTIC cells employ diverse mechanisms to preserve the structure and function of their genome following chromosome damage. Unicellular organisms rely on DNA repair systems and cell cycle arrest to prevent propagation of genome damage, while multicellular organisms can also activate programmed cell death pathways to eliminate cells following damage (Sancar et al. 2004; Roos and Kaina 2006; Bartek and Lukas 2007; Harper and Elledge 2007). In response to double-strand DNA breaks (DSBs), the Mre11, Rad50, and Nbs1 (MRN) complex helps process the damage and activates the ATM and ATR kinases. ATM and ATR phosphorylate many substrates including the downstream kinases Chk1 and Chk2, which regulate cell cycle arrest and apoptosis. The p53 transcription factor plays an evolutionarily conserved role, connecting the DNA damage signaling pathway to the core apoptotic machinery (Murray-Zmijewski et al. 2006; Helton and Chen 2007). Mammalian p53 is directly phosphorylated by ATM and Chk2 and activates many targets genes including pro-apoptotic Bcl-2 family members and the cell cycle regulator p21 (Riley et al. 2008). The p53 paralogs p63 and p73 also contribute to p53-dependent apoptosis by helping p53 bind the promoters of pro-apoptotic genes in some, but not all cell types (Flores et al. 2002; Senoo et al. 2004).
While p53 is required for normal induction of cell death following DNA damage, p53-independent mechanisms can also activate a reduced or delayed response ( Yuan et al. 1999; Irwin et al. 2003; Urist et al. 2004; Ozaki and Nakagawara 2005; reviewed in Roos and Kaina 2006). Several studies have demonstrated that p53-independent death can be enhanced by inactivating checkpoints that block cell cycle progression in the presence of unrepaired DNA damage. In one example, cancer cells lacking p53 required ATM, ATR, Chk1, and p38MAPK/MK2 signaling for cell cycle arrest; loss of this response led to caspase-3 activation and mitotic catastrophe following DNA damage (Reinhardt et al. 2007). In another case, zebrafish embryos lacking p53 and Chk1 function required ATM and ATR to activate an unusual apoptotic response requiring caspase-2, but not caspase-9 or caspase-3 (Sidi et al. 2008).
In Drosophila, rapid induction of apoptosis by ionizing radiation (IR) or unprotected telomeres requires homologs of ATM, Chk2 (TEFU, MNK, respectively), and p53 (Brodsky et al. 2000a, 2004; Ollmann et al. 2000; Xu et al. 2001; Peters et al. 2002; Oikemus et al. 2004; Silva et al. 2004; Song et al. 2004). G2 arrest requires ATR, ATRIP, and Chk1 (MEI-41, MUS304, and GRP, respectively) (Hari et al. 1995; Fogarty et al. 1997; Brodsky et al. 2000b, 2004; De Vries et al. 2005). Transcriptional targets of p53 include the pro-apoptotic genes reaper (rpr), head involution defective (hid), and sickle (skl ) (Brodsky et al. 2000a, 2004; Christich et al. 2002; Akdemir et al. 2007). hid is essential for the rapid induction of apoptosis, while rpr may play a less significant role (Brodsky et al. 2004; Moon et al. 2008). rpr, hid, and skl encode proteins that induce apoptosis by directly binding and inhibiting DIAP1, leading to activation of an initiator caspase, Dronc, and effector caspases (Bergmann et al. 2003; Chew et al. 2004; Daish et al. 2004; Waldhuber et al. 2005; Kondo et al. 2006; Xu et al. 2006; Steller 2008). Regulation of mitochondrial ultrastructure by HID and RPR may also help induce cell death after DNA damage (Holley et al. 2002; Olson et al. 2003; Abdelwahid et al. 2007; Goyal et al. 2007).
While rapid induction of apoptosis following IR requires Drosophila p53, there is also p53-independent apoptosis following chromosome damage. Mutations in the Drosophila ATM/ATR/MRN damage signaling pathways lead to loss of telomere protection and high levels of spontaneous apoptosis (Bi et al. 2004, 2005; Ciapponi et al. 2004, 2006; Oikemus et al. 2004, 2006). The apoptosis induced by unprotected telomeres is only partly suppressed by loss of p53 or MNK (Oikemus et al. 2004, 2006). No additional p53 family members contribute to this response since Drosophila has only a single p53 homolog. Similarly, in the absence of p53, a delayed or reduced induction of apoptosis is observed following IR or telomere loss (Wichmann et al. 2006; Titen and Golic 2008). Double-mutant combinations of tefu, mus304, and nbs also exhibit spontaneous apoptosis, suggesting that the activation of apoptosis by chromosome damage is independent of the central upstream components of the DNA damage response pathway (Oikemus et al. 2006).
Here, we characterize the regulation and function of p53-independent apoptosis following IR. We show that p53-independent apoptosis requires hid, which is a target of IR-induced JNK signaling. Mutations in grp or the JNK phosphatase puc sensitize p53 mutant cells to IR-induced apoptosis, suggesting that cell cycle progression and JNK signaling are critical for this response. We find that this response acts to maintain genomic stability by reducing the number of aneuploid adult cells induced by IR. We propose that activation of JNK and HID in response to aneuploidy induces p53-independent apoptosis following IR.
MATERIALS AND METHODS
Drosophila melanogaster genetics:
All flies were raised at 25°. Genes and alleles are described in http://flybase.org. w1118 was used as the wild-type strain. The following mutant genotypes were used: p5311-1b-1 (this allele is used for all experiments except in Figure S3B); p535A-1-4; mnkP6; grpfs1; p5311-1b-1; grpfs1 mnkP6; actin-Gal4/+; UASp35/+; en-Gal4/+; UASp35/+; en-Gal4/+; UASp35 p5311-1b-1/p5311-1b-1; en-Gal4/+;UASpuc p5311-1b-1/p5311-1b-1; en-Gal4/+; UASpuc UASp35 p5311-1b-1/p5311-1b-1; en-Gal4/+; pucA29lacZ p5311-1b-11/+UASp35 p5311-1b-1; pucA29-lacZ p5311-1b-1/+p5311-1b-1; grpfs1; pucA29-lacZ p5311-1b-1/+p5311-1b-1; brk38-20-lacZ/+; p5311-1b-1; brk38-20-lacZ/+; en-Gal4/+;UASp35 p5311-1b-1/p5311-1b-1; brk38-20-lacZ/+; en-Gal4/+;UASp35/+; droncI29/droncI24; hidX14/hid05014; hidX14 p5311-1b-1/ hid05014 p5311-1b-1; mwh1; and mwh1 p5311-1b-1.
Analysis of X-irradiation-induced changes in apoptosis, cell cycle, and gene expression:
For 4- and 8-hr time points, wandering third instar larvae were mock treated or X-irradiated with 4000 rad using a Faxitron RX650 X-ray cabinet system (Faxitron X-ray Corporation). For 12-, 16-, and 24-hr time points, larvae were treated as early third instar in Drosophila media and dissected as wandering third instar larvae.
Immunostaining and acridine orange staining were largely performed as described previously (Abrams et al. 1993; Brodsky et al. 2000b; Oikemus et al. 2004). Imaginal wing discs from wandering third instar larvae were dissected in 1× PBS and fixed in 4% formaldehyde for 30 min at room temperature. Samples were washed in 250 μl of PBS + 0.3% Triton X-100 five times for 5 min each and incubated in blocking solution (PBS + 0.3% Triton X-100 + 5% normal goat serum) for 1 hr. Samples were incubated in primary antibody diluted in blocking solution overnight at 4°. The posterior expression domain in en-Gal4 experiments was marked by expression of endogenous En protein. Primary antibodies were used at the following dilutions: rabbit anti-cleaved caspase-3 (Cell Signaling Technology), 1:100; mouse anti-phospho-H3 (Cell Signaling Technology), 1:500; mouse anti-Beta-Gal (Santa Cruz Biotechnology), 1:1000; mouse anti-EN (Developmental Studies Hybridoma Bank at the University of Iowa, DSHB), 1:10; mouse anti-WG (DSHB), 1:50; mouse anti-PTC (DSHB), 1:100; and guinea pig anti-HID (Ryoo et al. 2004) (gift from Hyung Don Ryoo, New York University Medical Center), 1:500. Following primary antibody incubation, samples were washed five times for 5 min each and incubated with secondary antibody in blocking solution for 2 hr at room temperature. Secondary antibodies were used at the following dilutions: anti-mouse Alexa 488 (Molecular Probes, Eugene, OR), 1:2000; anti-rabbit Alexa 555 (Molecular Probes), 1:2000; anti-guinea pig Alexa 488 (Molecular Probes), 1:2000; anti-mouse FITC ( Jackson Laboratories, for staining with anti-phospho-histone H3 primary), 1:250; and anti-mouse Cy5 ( Jackson Laboratories), 1:250. Imaginal wing discs were stained with 4′,6-diamidino-2-phenylindole (DAPI) and mounted in Vectashield (Vector Laboratories, Burlingame, CA). Wide field images of wing discs were acquired with a Zeiss Axioplan fluorescence microscope, an ORCA-ER digital camera (Hamamatsu), and Axiovision 4.5 software. Confocal images were acquired with a Leica TCS SP2 confocal microscope. At least five wing discs were scored for each genotype and time point.
Quantitative analysis of cleaved-caspase staining:
A confocal z-series was taken through the entire disc at intervals of 1.42 μm with a 20× objective. A three-dimensional reconstruction was generated using Imaris image analysis software (version 5.0.1, Bitplane). A high-intensity threshold was used to create isosurfaces of regions that stained positive for cleaved caspase-3. A low-intensity threshold was used to create an isosurface of the background staining observed throughout the entire disc. Examples of a maximum-intensity projection, low-threshold isosurface, and high-threshold isosurface are shown in Figure 1E. The anti-caspase-3 index is the total volume of high-intensity isosurfaces divided by the volume of the low-intensity isosurface. P-values were calculated with a two-tailed Student's t-test.
Newly laid eggs were collected for 48 hr. Seven days later, larvae were X-irradiated with 0, 500, 1000, or 2000 rad at two different stages, either wandering late third instar or second to early third instar larvae still in the media. Late third instar larvae were transferred to new vials following IR. For larvae irradiated in the media, wandering third instar larvae were picked out of the vials over the next 48 hr. After 1 week, the numbers of normal and Minute bristles were scored in the resulting adults as previously described (Brodsky et al. 2000b). For each genotype and treatment, bristle phenotypes were scored in at least 60 flies unless otherwise noted. Data from three different irradiation experiments were pooled. P-values were calculated with a two-tailed Student's t-test.
Wandering third instar larvae were irradiated with 0, 250, or 1000 rad and transferred to new vials. After 1 week, adult wings were mounted in 1:1 methyl salicylate/Canada balsam (Sigma, St. Louis) and individual cells were scored for the mwh−/− phenotype (Brodsky et al. 2000b). For each genotype and dose, five wings were scored. P-values were calculated using a two-tailed Student's t-test.
p53-independent apoptosis requires the pro-apoptotic gene hid and the apical caspase gene dronc:
Cell death in Drosophila wing discs was examined following X-irradiation of third instar larvae. At least five discs were examined for every genotype and time point shown. Mutant alleles are described in materials and methods. Four hours after irradiation, high levels of apoptosis are induced in wild-type discs (Figure 1A; Figure S1A and Figure S2A). p53 or mnk mutant discs exhibit no change at 4 hr, but a significant increase in apoptosis between 16 and 24 hr (Figure 1, B and F; Figure S2, B and C). A previous study indicated that one or more of the pro-apoptotic genes in the Df(3L)H99 region are likely to contribute to p53-independent apoptosis following IR (Wichmann et al. 2006). In hid single-mutant discs, apoptosis is blocked at 4–8 hr following IR and is reduced at later time points (Figure 1, A and C; Figure S2, A and F). Twenty-four hours following IR, hid, p53 double-mutant discs exhibit a 7- and an 11-fold reduction in the caspase staining compared to p53 or hid single-mutant discs (Figure 1, B–F), demonstrating that hid contributes to p53-independent apoptosis following IR (Figure 1, B–F).
All IR-induced acridine orange and anti-C3 staining was blocked at both early and late time points in wing discs mutant for the caspase-9 homolog dronc (Figure S1B and Figure S2D), indicating that dronc is required for both p53-dependent and p53-independent apoptosis. Similarly, in discs expressing the anti-apoptotic baculovirus protein p35, which inhibits the effector caspases Drice and DCP-1 (Xue and Horvitz 1995; Kondo et al. 2006; Xu et al. 2006; Baum et al. 2007; Lannan et al. 2007), no acridine orange staining was observed at any time point (Figure S2E). Because p35 inhibits effector caspase activity after proteolytic cleavage, cleaved caspase-3 staining can still be observed once these caspases are activated ( Yu et al. 2002). When p35 is expressed in the posterior half of wild-type or p53 mutant wing discs, IR-induced caspase staining is induced in both halves, but the staining is delayed in the p35-expressing region (Figure S1, C and D).
Induction of HID by IR was examined in p53 mutant discs expressing p35, which allows HID to accumulate without causing cell death. HID induction is observed 16 and 24 hr following irradiation of p53 mutant cells expressing p35 (Figure 2A, posterior cells). Overall, our analysis of hid and caspase function indicates that p53-independent apoptosis following IR utilizes the same core apoptotic machinery activated in response to most developmental and cell stress signals in Drosophila.
Regulation of p53-independent apoptosis by the JNK pathway:
Following IR, JNK signaling is activated in a p53-dependent manner and inhibition of JNK signaling reduces induction of apoptosis (McEwen and Peifer 2005). The Drosophila JNK signaling pathway also regulates apoptosis in response to other cellular stresses, including UV irradiation ( Jassim et al. 2003; Luo et al. 2007), abnormal expression patterns of the DPP or WG morphogens (Adachi-Yamada et al. 1999; Adachi-Yamada and O'Connor 2002), or reduced dosage of ribosomal protein genes (Moreno et al. 2002).
The role of JNK signaling in p53-independent induction of HID by IR was examined using puc, a JNK target gene encoding a JNK phosphatase. puc acts in a negative feedback loop to limit JNK signaling and can be used to efficiently block JNK activity when overexpressed in a specific cell type or region (Martin-Blanco et al. 1998). puc overexpression in p53 mutant discs reduced HID induction after IR (compare posterior compartments in Figure 2, A and B). puc overexpression also reduces anti-C3 staining following IR (Figure 2, C and D); the ratio of posterior to anterior anti-C3 staining is 1.4 in p53 mutant discs and is reduced to 0.3 in p53 mutant discs overexpressing puc in the posterior. The incomplete inhibition of apoptosis in these discs may indicate that a JNK-independent pathway also contributes to this response.
A puc-lacZ reporter, puc A29, was used to monitor JNK signaling in irradiated discs (Martin-Blanco et al. 1998; Kanda and Miura 2004). As previously reported (McEwen and Peifer 2005), p53-dependent expression of the puc reporter is observed 4 hr following IR (data not shown). However, p53 mutant discs begin expressing the reporter at later time points following IR (Figure 3A). In irradiated p53 mutant wing discs expressing p35, HID and the puc reporter are expressed in an overlapping set of cells (Figure 2E).
Cells heterozygous for puc mutations exhibit hyperactivation of the JNK pathway (Martin-Blanco et al. 1998), indicating that puc gene dosage is rate limiting for JNK activity. If the level of JNK activation determines the amount of p53-independent apoptosis, then p53 mutant discs heterozygous for puc mutations should exhibit increased apoptosis following irradiation. At 16 and 24 hr after IR, there is a 3- and 4-fold increase in caspase staining in p53 mutant discs that are also heterozygous for puc compared to p53 single-mutant discs (Figure 3, C, D, and F). The overall correlation of puc expression levels and apoptosis suggests that the JNK pathway helps regulate p53-independent apoptosis following IR.
Drosophila Chk1 is a negative regulator of p53-independent apoptosis following IR:
The induction of p53-independent apoptosis following IR correlates with cell cycle progression. IR-induced G2 delay in the wing disc lasts up to 8 hr and requires the Drosophila chk1 kinase homolog grp (Fogarty et al. 1997; Brodsky et al. 2000a), even in the absence of p53 or MNK function (Figure 4, A–D). While IR-induced apoptosis in wild-type discs begins during the cell cycle arrest period (within 3–4 hr), apoptosis in p53 mutant discs is largely observed after mitosis has resumed. In grp p53 double-mutant discs, which fail to arrest in G2, increased levels of apoptosis are observed at 12 and 16 hr compared to p53 single-mutant discs (Figure 4, E, F, and I). Similarly, apoptosis is increased in grp mnk double-mutant wing discs compared to mnk single-mutant discs (Figure 4, G, H, and J). Previous studies have established that mutations in mnk and grp do not affect the frequency of chromosome breaks following IR (Jaklevic and Su 2004; Oikemus et al. 2006), indicating that the increased apoptosis is due to a defect in cell cycle arrest, not simply an increase in unrepaired DNA breaks.
In mammalian cells that lack both p53 and Chk1 function, DNA damage can induce a type of mitotic catastrophe in which the mitotic marker phosphohistone H3 and cleaved caspase-3 are present simultaneously in the dying cells (Reinhardt et al. 2007). However, in irradiated grp p53 mutant wing discs, these markers label two separate cell populations: mitotic cells stained with anti-phosphohistone H3 are predominantly in the apical region of the disc, while apoptotic cells stained with anti-cleaved caspase-3 are largely found in the basal region (Figure 5, A and B). These results suggest that these cells do not undergo apoptosis while in M-phase.
The relationship between grp and puc during p53-independent apoptosis was also examined. IR-induced expression of the puc reporter is faster and stronger in grp p53 double-mutant discs compared to p53 single-mutant discs (Figure 3, A and B), suggesting that cell cycle arrest normally helps delay induction of JNK signaling. The effect of both inactivating cell cycle arrest and reducing the level of negative feedback signaling by puc was also examined: grp puc/+ p53 triple-mutant discs have higher levels of apoptosis than either grp p53 or puc/+ p53 double-mutant discs (Figure 3, D–G). Thus, while the inactivation of cell cycle delay accelerates the formation of cells with elevated JNK signaling, decreasing the negative feedback signal from puc can further increase the number of cells that undergo p53-independent apoptosis following IR.
p53-independent apoptosis reduces the number of aneuploid cells recovered following IR:
The role of p53-dependent and p53-independent apoptosis in maintaining genomic integrity was examined using an assay based on haploinsufficiency of ribosomal protein genes. In animals exposed to IR or with mutations in DNA repair loci, somatic mutations scored by loss of heterozygosity (LOH) are typically accompanied by loss of multiple genetic markers and the Minute phenotype, caused by haploinsufficiency of ribosomal protein genes (Baker et al. 1978). These results indicate that most LOH was due to segmental aneuploidy (loss of large chromosomal regions) rather than induction of point mutations or mitotic recombination. To probe genetic loss across the entire genome, we scored the frequency of Minute cells following IR. Because the 65 Minute loci are spread throughout the euchromatic genome (Marygold et al. 2007), reduced copy number of most large genomic regions is likely to result in the Minute phenotype. Previous studies have used this phenotype to assay the genetic consequences of diverse sources of chromosome damage, including mutations in telomere protection genes (Oikemus et al. 2004, 2006), telomere loss due to dicentric chromosomes (Ahmad and Golic 1999), mutations in DNA damage response and repair genes (Brodsky et al. 2000b; Johnson-Schlitz et al. 2007), and induction of DNA breaks by P-element mobilization (Engels et al. 1987). These studies confirm that induction of Minute bristles following IR is due to chromosome damage rather than other types of cellular damage. It is possible that a small percentage of defective bristles are not due to induction of aneuploidy, but these events would probably reflect other types of genetic damage, such as Minute gene point mutations.
Minute cells were scored in adults following X-irradiation of early third instar larvae. If apoptosis reduces the number of IR-induced aneuploid cells, then inhibiting apoptosis should increase the number of these cells that survive to adulthood. Following irradiation of wild-type larvae with 1000 rad, 2% of bristles exhibit the Minute phenotype (Figure 6, A and C). No difference in the number of defective bristles was seen in p53 mutant compared to wild-type animals at either dose, indicating that p53-dependent apoptosis is not essential to eliminate Minute cells following IR (Figure 6C). This result is inconsistent with a previous study, in which an increase in Minute bristles was reported following IR of late third instar larvae (Lee et al. 2003). When we repeated this experiment using the later stage, there was still no significant increase in Minute bristles in p53 mutant animals (Figure S3A). To help control for differences in genetic background, additional wild-type and p53 mutant lines were tested. Again, no increase in Minute bristles was observed in p53 mutants compared to wild-type lines (Figure S3B). It is possible that the previously reported increase was due to a second mutation on the p53 mutant chromosome.
Elimination of aneuploid cells by p53-independent apoptosis could compensate for the apoptotic defect in p53 mutant animals. To test if p53-independent apoptosis helps limit IR-induced aneuploidy, the frequency of IR-induced Minute bristles was also examined in hid p53 double-mutant animals. At 1000 rad, the frequency increased from 2% in p53 single-mutant animals to 5% in double mutants (Figure 6, B and C). Thus, in a p53 mutant background, the p53-independent apoptosis induced by IR is required to limit genomic instability as measured by the appearance of adult cells heterozygous for a Minute mutation. If bristle cells are representative of the ∼50,000 cells in the developing wing, then ∼2500 (5% of 50,000) Minute cells are induced by IR per disc and ∼1500 (3% of 50,000) are eliminated by apoptosis in a p53 single-mutant disc, but not in a p53 hid double-mutant wing disc.
hid is a target of both p53-dependent and p53-independent signaling following DNA damage. In hid single mutants, IR-induced apoptosis is eliminated at early time points, but only partly reduced at later time points (Figures 1 and Figure S2F), a phenotype that is intermediate between p53 single- and hid p53 double-mutant animals. The increased frequency of Minute cells in IR-treated hid p53 double mutants compared to hid single mutants indicates that p53 helps limit the appearance of aneuploid cells when p53-independent apoptosis is also blocked (Figure 6C).
Altered brk expression following IR:
To explore if IR broadly disrupts patterned gene expression in the developing wing, we examined the expression of several key regulators of wing patterning. Following DNA damage, missegregation of rearranged or broken chromosomes during mitosis generates cells with segmental aneuploidy. Reduced copy number of haploinsufficient genes, such as the Minute genes, could lead to altered gene expression and apoptosis. Phenotypes associated with Minute cells in the wing disc include increased apoptosis, JNK pathway activation, and ectopic expression of the transcriptional repressor brinker (brk) due to reduced signaling by the dpp signaling pathway (Moreno et al. 2002; Tyler et al. 2007). Whereas both JNK and apoptosis are activated by many different types of cellular stresses, including altered gene expression patterns (Umemori et al. 2009), brk may be more specific for dpp signaling. brk expression is normally repressed in the center of developing wing discs and gradually increases toward the anterior and posterior edges of the disc (brk-lacZ in Figure 7, A–C). At 24 hr following IR, individual cells in the center of p53 mutant discs ectopically express brk and expression in the lateral regions becomes more irregular (Figure 7, A–C). Inhibition of apoptosis by p35 expression increases the number of cells with ectopic brk expression. An average of 9 cells ectopically express high levels of brk in the center of the p53 mutant wing discs expressing p35 (SEM = 2.6, n = 5 discs, P-value <0.01, treated compared with untreated), while an average of 2.6 cells ectopically express brk without p35 (SEM = 1.7, n = 5, P-value <0.01, treated compared with untreated). (These numbers significantly underestimate the total number of cells with abnormal brk expression since we excluded cells within or adjacent to the normal brk expression domain in our analysis.) Ectopic brk expression following IR is also observed in wild-type discs, suggesting that induction of this phenotype by IR can occur in parallel to p53-dependent apoptosis (Figure S4). In wild-type discs, p53-dependent apoptosis is induced within 4 hr of X-irradiation, but ectopic expression of brk is not observed at this time point (data not shown), indicating that altered brk expression is not a secondary consequence of induced apoptosis. Because ectopic brk expression is sufficient to activate JNK-dependent apoptosis (Martin et al. 2004), the observed changes in brk expression could contribute to the elimination of Minute cells (Moreno et al. 2002).
Although ectopic expression of brk is consistent with the generation of Minute cells, it could alternatively reflect altered expression of genes that act earlier in the anterior–posterior patterning hierarchy or reflect a widespread disruption of patterned gene expression following cellular damage by IR. The expression patterns of three additional gene products, two with restricted expression along the anterior–posterior axis (EN and PTC) and one with restricted expression along the dorsal–ventral axis (WG), were also examined in irradiated p53 mutant wing discs; no change in expression was observed for any of these proteins (Figure 7, D–F). This small-scale survey suggests that irradiation does not generally disrupt developmental networks and that the effect on brk expression is not due to loss of the anterior–posterior compartment (reflected by EN expression) or aberrant activity of the Hedgehog pathway along the compartment boundary (reflected by PTC expression).
While p53 plays a conserved role connecting the DNA damage response pathway to the core apoptotic machinery, there are also p53-independent mechanisms to induce cell death following chromosome damage. We find that this p53-independent response in Drosophila requires the pro-apoptotic gene hid, the apical caspase dronc, and downstream effector caspases. Proper regulation of the JNK phosphatase puc is required for the full induction of hid expression and apoptosis following irradiation of p53 mutant discs. puc and the cell cycle checkpoint gene grp are negative regulators of this response and mutations in these genes act additively to sensitize p53 mutant cells to IR. This response plays an important in vivo role in limiting the accumulation of aneuploid cells following IR. We propose that p53-independent apoptosis reflects a second, indirect mechanism that acts in parallel to the canonical DNA damage response pathway to eliminate cells with altered genomes following IR (Figure 8): in this model, incorrect repair of IR-induced chromosome breaks generates cells with segmental aneuploidy, resulting in haploinsufficiency of genes required for cell survival.
IR-induced apoptosis without components of the canonical DNA damage response pathway:
The ATM/Chk2/p53 signaling module is activated by local disruptions in chromatin structure following DNA damage and plays a central role in IR-induced apoptosis. In mammalian cells lacking p53, components of the DNA damage response pathway, such as ATR and p73, have been implicated in damage-induced apoptosis (Roos and Kaina 2006; Sidi et al. 2008). In this study and previous work (Oikemus et al. 2004, 2006; Wichmann et al. 2006; Titen and Golic 2008), genetic analysis of single and double mutants in components of the Drosophila DNA damage response pathway suggests that chromosome damage can induce apoptosis through a mechanism that does not utilize this pathway. Consistent with these results, activation of ATM and ATR kinases following IR (as measured by phospho-H2Av staining) largely subsides within a few hours of IR in embryos (Kusch et al. 2004) and discs (our unpublished results), well before increased p53-independent apoptosis, JNK target gene expression, or HID. Furthermore, in a recent study using dicentric chromosomes to induce chromosome breaks and telomere loss, all dicentric chromosomes could induce p53-dependent apoptosis, but only dicentric chromosomes that led to aneuploidy could induce p53-independent apoptosis (Titen and Golic 2008).
In p53 or mnk mutant discs, IR-induced apoptosis is observed only after cells have recovered from damage-induced cell cycle delay. This response is accelerated in double-mutant discs that fail to arrest due to mutations in the Drosophila chk1 homolog grp. Similarly, p53-independent induction of puc by IR is observed after resumption of cell cycle progression in single-mutant discs and is accelerated by loss of grp. Inhibition of JNK signaling by overexpression of puc substantially reduces p53-independent induction of HID and apoptosis, suggesting that this response is partly dependent on the JNK pathway. The remaining apoptosis may either reflect incomplete inhibition of JNK signaling by puc overexpression or suggest that an alternative pathway acts in parallel to JNK to promote a lower level of p53-independent apoptosis. While the only described function of puc to date is as an inhibitor of JNK signaling, we cannot rule out the possibility that inhibition of other pathways by puc contributes to the regulation of p53-independent apoptosis. Overall, these results suggest that cell cycle arrest helps delay p53-independent activation of JNK signaling and apoptosis following IR.
Regulation of genomic stability by p53-dependent and -independent responses:
p53-dependent apoptosis is hypothesized to help preserve genome stability by rapidly eliminating cells with severe chromosome damage. In mouse models with elevated genomic instability, loss of p53 function further increases tumorigenesis (Attardi 2005). However, temporal analysis of mouse p53 function following irradiation suggests that the DNA damage response function of p53 may not be critical for tumor suppression (Christophorou et al. 2006).
In Drosophila, the role of p53 in limiting genomic instability appears to depend on the source of DNA damage or the type of genetic alterations assayed. In one study, telomere loss following dicentric chromosome formation led to an extremely high frequency of larval neuroblasts with abnormal karyotypes; p53 helped reduce the number of abnormal cells over time (Titen and Golic 2008). Another study using a genetic assay for LOH at the multiple wing hair (mwh) locus reported increased LOH following IR in p53 mutant animals (Sogame et al. 2003), a result we have confirmed with an independently derived p53 allele (Figure S5). However, when using the Minute assay to examine genomic instability, we found that p53 function was not essential to limit IR-induced aneuploidy, despite the very high levels of p53-dependent apoptosis normally induced following IR. Several factors might account for the different conclusions reached using these assays. In the mwh assay, two types of genetic damage not associated with aneuploidy, small deletions or mitotic recombination, can be recovered that may not be scored in the Minute assay. p53-independent apoptosis may be able to compensate in the absence of p53 function to eliminate aneuploid cells, while p53-dependent apoptosis may play a more critical role in eliminating cells with genetic damage that is not associated with aneuploidy. Consistent with this interpretation, we find that p53 does contribute to genomic stability in the Minute assay when the p53-independent response is blocked (compare hid single-mutant to hid p53 double-mutant animals).
We find a significant increase in the number of aneuploid cells recovered following IR when both p53-dependent and p53-independent apoptosis are blocked. Inhibition of both pathways results in a 2.5-fold increase in adult bristle cells with the Minute phenotype, resulting in 5% of bristle cells exhibiting this severe morphological defect. These results demonstrate that p53-independent apoptosis plays an important functional role following IR by substantially reducing the eventual number of adult aneuploid cells. Because hid mutations affect both p53-dependent and p53-independent apoptosis, we cannot determine whether this response is required for genomic stability in cells with the normal p53 signaling pathway.
Aneuploidy as a possible apoptotic signal following IR:
A key unresolved question in this study is what signal activates p53-independent puc and hid expression and apoptosis following IR. An attractive possibility is that aneuploidy itself is the signal. In strong support of this hypothesis, induction of p53-independent apoptosis following telomere loss is observed only if the resulting broken chromosome can lead to aneuploidy (Titen and Golic 2008). Our analysis of IR-induced Minute bristles indicates that many aneuploid cells are generated following IR and then eliminated by apoptosis. In larvae, chromosome rearrangements appear shortly after IR (Gatti et al. 1974). IR-induced changes in the expression of brk, puc, and hid are consistent with the production of aneuploid cells exhibiting the Minute phenotype, although these markers are not specific for aneuploid cells. Since haploinsufficiency of the widely distributed Minute genes is sufficient to induce apoptosis (Moreno et al. 2002; Coelho et al. 2005) and induction of spontaneous apoptosis in Minute heterozygous animals does not require p53 (L. M. McNamee, unpublished results), this mechanism should eventually be activated in aneuploid cells induced by IR. While Minute genes represent the primary source of loci that are haploinsufficient for normal growth (Lindsley et al. 1972; Marygold et al. 2007), it is also possible that the cumulative effect of reduced dosage of many genes in aneuploid cells disrupts normal gene expression and cell survival. On the basis of these observations, we propose that DNA damage indirectly induces apoptosis by creating chromosome aberrations leading to segmental aneuploidy and activating JNK-dependent and p53-independent apoptosis (Figure 8). However, we cannot easily rule out the possibility that an additional response to IR could kill these cells before apoptosis due to aneuploidy is activated.
Haploinsufficiency of ribosomal genes could contribute to damage-induced apoptosis in vertebrates as well. In mice heterozygous for ribosomal protein gene mutations, both p53-dependent and p53-independent signaling are induced, leading to increased apoptosis and reduced growth (Panic et al. 2006; Danilova et al. 2008). Diamond Blackfan anemia (DBA) is linked to haploinsufficiency of human ribosomal protein genes (Draptchinskaia et al. 1999; Gazda and Sieff 2006; Gazda et al. 2006; Cmejla et al. 2007; Choesmel et al. 2008) and is associated with increased cell death (Flygare and Karlsson 2007). As in Drosophila, mammalian ribosomal protein genes are widely distributed throughout the genome. Thus, genetic or environmental changes resulting in aneuploidy should frequently induce the cellular responses associated with ribosomal protein gene haploinsufficiency.
In summary, we have identified a p53-independent pathway that limits the formation of aneuploid cells through JNK- and HID-dependent apoptosis. We hypothesize that rather than detect damaged chromosomes directly, this mechanism is activated by the reduced dosage of critical loci, such as ribosomal protein genes, in aneuploid cells.
Inactivation of this response causes an increase in visibly defective bristle cells following DNA damage. Thus, in Drosophila, this mechanism helps eliminate cells that would reduce the fitness of the adult animal. We have shown that mutations in a cell cycle checkpoint gene and a JNK phosphatase enhance this response following IR. These and other components of this pathway may provide useful targets to reduce the proliferation of aneuploid cancer cells or to sensitize cancer cells to the effects of DNA-damaging agents.
We especially thank Hyung Don Ryoo for providing polyclonal antibody to HID. We also thank Bruce Hay, Hermann Stellar, Andreas Bergmann, Gerard Campbell, Takashi Adachi-Yamada, and the Bloomington Stock Center for providing Drosophila stocks. We thank Neal Silverman, Steve Grossman, and Eric Baehrecke for comments on the manuscript. We thank Simon Titen and Kent Golic for sharing their results prior to publication. This work was supported by a New Scholar in Aging Award from the Ellison Medical Foundation and by a grant from the American Cancer Society (RSG-05-026-01-CCG).
Supporting information is available online at http://www.genetics.org/cgi/content/full/genetics.109.102327/DC1.
Communicating editor: E. Alani
- Received February 28, 2009.
- Accepted March 18, 2009.
- Copyright © 2009 by the Genetics Society of America