Dual contributions of Xrp1 to genome integrity through the DNA damage response and cell competition

  1. Chaitali Khan
  2. Nasser M Rusan
  3. Nicholas E Baker

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Model organisms may help understand how p53 suppresses tumorigenesis in mammals. In Drosophila , the primary transcriptional target of p53 is the gene encoding the bZip AT-hook protein Xrp1, which is another transcription factor. We report that Xrp1 mediated multiple functions of p53 in the DNA damage response (DDR), contributing to p53-dependent gene transcription and DNA damage-induced apoptosis. In addition to this role as a p53 effector, a p53-independent role for Xrp1 in cell competition has been described. Cell competition can remove cells whose genome has been altered by DNA damage and repair. During cell competition, Xrp1 is induced by RpS12, which acts as a sensor of defective ribosome biogenesis. RpS12-dependent cell competition began as the DDR wound down, and was even more prominent if p53 function was reduced. Such p53 inhibition resulted in persistence of DNA damage revealed by γH2Av accumulation. Thus, Xrp1 limited the accumulation of abnormal cells resulting from genotoxicity through both the acute, p53-dependent DDR, and also from subsequent cell competition that removes cells where DNA repair did not restore the normal genome. Both these processes might contribute to the tumor suppressor function of p53 in mammals.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    we thank the reviewers for their close reading of the manuscript and detailed comments.

    __Reviewer #1 __

    The idea that Xrp1 induction switches around 16 h post-IR, becomes RpS12-dependent, and subsequently engages cell competition is interesting and potentially important. However, the evidence supporting RpS12-dependence of Xrp1 induction is currently not sufficiently convincing. For example, based on the images in Figure 6F-supplement 1, the conclusion that Xrp1 is induced in an RpS12-dependent manner appears difficult to support. The authors should strengthen and quantify this result or provide the raw image data. In addition, because this point is central to the authors' model, they should move the key supporting data from the supplementary figures to the main figures to ensure that this critical claim is clearly supported and readily accessible to readers.

    We apologize for confusing all three reviewers with this figure. Actually, Figure 6F supplement 1 does not compare RpS12-dependent and -independent Xrp1-HA expression. Instead, it shows that the rps12-independent Xrp1-HA expression is only mildly p53-dependent, which is consistent with our idea. We had not compared RpS12-dependence or Xrp1 expression in this manuscript because we had published that previously and found a substantial dependency (Fig 1N-P of Ji et al 2021). Because that previous paper used an anti-Xrp1 antibody, and the present paper measures an HA-tagged Xrp1 protein, it is probably a good idea to include the RpS12-dependence of late Xrp1 expression again, using the Xrp1-HA reagent. We have this data, which shows ~75% dependence, which is highly significant statistically. We will include this data in the revised manuscript, within one of the main figures.

    The authors suggest a model in which Xrp1 executes two qualitatively distinct "modes"(pro-repair/acute DDR and elimination of aneuploid cells), but this remains only partially convincing as currently presented. The authors should at least (i) provide quantitative evidence that could explain how Xrp1 might produce distinct outcomes across phases(e.g., comparing Xrp1-HA levels and/or the fraction of Xrp1-HA-positive cells at 2-4 h versus 16-24 h post-IR), and (ii) explicitly discuss plausible mechanisms in the Discussion. Even if the molecular "switch" is not fully resolved experimentally, a clearer, data-grounded discussion of how Xrp1 could mediate these temporally distinct functions is needed. In addition, since ISR signaling (e.g., eIF2α phosphorylation) has been implicated as a single feature associated with Xrp1-dependent loser elimination, the authors should consider assessing p-eIF2α levels in Xrp1-HA positive cells at early versus late time points after IR(e.g., 4 h vs 24 h).

    We thank the reviewer for highlighting the need for this discussion. We will clarify these issues in the revised manuscript but do not think further experiments are necessary.

    1. It was well established previously and confirmed here that little DNA damage remains ~24h after IR. This is sufficient to explain why there is little DDR at this stage. We will make this clear in the revision.
    2. We did not intend to claim that no cell competition happens during the acute DDR ~4h after IR. We are not aware of experiments showing the DDR is strictly cell autonomous and not influenced by neighboring cells. If the acute DDR is indeed cell autonomous, or mostly so, this could be due to the additional genes induced directly by p53 that are not induced by Xrp1 ~24h after IR. The cell death gene Rpr is one example reported in our paper. We will discuss this in the revision.
    3. The reference to ISR as the single feature inducing Xrp1 expression is referring to two Nature Cell Biology papers published in 2021 (Baumgartner et al 2021; Recasens-Alvarez et al 2021). This idea has not stood the test of time. The ISR reporter activities shown in these papers were later shown to be downstream of Xrp1, not upstream (Langton et al 2021; Kiparaki et al 2022). Langton et al argued that there could be an initial ISR that was too small to be detectable, but this is hypothetical. There are now multiple papers and preprints showing that it is long isoforms of Xrp1 are ISR responsive, but that short isoforms of Xrp1 initiate cell competition, and that RpS12-dependent alternative splicing produces the short isoform. The short Xrp1 isoforms lack the uORF that responds to ISR (Elife 2021 Oct 4:10:e74047; bioRxiv 06.15.659587; bioRxiv 2025.10.29.685279). This is not consistent with the ISR initiating cell competition idea. Because we and others have shown that it is Xrp1 activity that induces eIF2α phosphorylation (Ochi et al 2021, Langton et al 2021, Kiparaki et al 2022), eIF2α phosphorylation in Xrp1 expressing cells would not prove a role for ISR and we do not propose to make these measurements. We are undecided whether to include this discussion of the ISR in the paper. It would lengthen the paper and we do not think it is directly relevant.

    The idea that aneuploid cells-or cells with altered ribosomal gene dosage-could be removed via Xrp1-mediated cell competition is intriguing. However, the manuscript does not currently provide any evidence that such cells are, in fact, being eliminated. The authors should therefore (i) quantify cell-level overlap metrics, such as the fraction of γH2Av-positive cells that are Xrp1-HA-positive (and vice versa), as well as the fraction of γH2Av-positive cells that are cleaved Dcp-1-positive (and vice versa) at 24 h post-IR. These quantitative analyses would clarify whether the late Xrp1-HA-positive population corresponds to persistently damaged cells and whether it is enriched for cells undergoing apoptosis/clearance. The authors should also (ii) directly assess aneuploidy/segmental copy-number imbalance in the late Xrp1-HA-positive clusters (e.g., by DNA FISH targeting one or two chromosome arms/regions), and if these experiments cannot be completed within a reasonable revision timeframe, the authors should temper their wording and present aneuploidy and selective elimination as a plausible interpretation supported byRpS12 dependency and prior literature, rather than as a demonstrated conclusion in the current study.

    We agree that aneuploidy is not demonstrated in the current study. Elimination of aneuploid cells with altered Rp gene dose was already established by previous papers. We cited previous work in the manuscript but did not summarize the evidence explicitly, so we are not sure whether the referee was fully aware. Ji et al (2021) created 17 different segmental aneuploidies using Flp/FRT recombination including or abutting 10 different Rp genes, together covering >20% of the euploid genome. The results showed that segmental aneuploidies are largely removed by Rp gene dose-dependent cell competition using the RpS12 and Xrp1 genes. Others have since confirmed that aneuploidies are removed by cell competition and that the effects of Rp gene dose depend on Xrp1 (Fusari et al Cell Genomics 2025). Therefore, we consider it established that aneuploid cells with altered Rp gene dosage are removed by this mechanism. We will discuss this explicitly in the revised manuscript.

    The question of whether cells dying in a p53-independent manner ~24h after irradiation are aneuploid cells undergoing cell competition was also addressed previously. Ji et al 2021 already showed that most of these cells are eliminated by RpS12 and Xrp1, consistent with altered Rp gene dosage, and that preventing cell competition leads to persistence into adulthood of cells that can be recognized at Rp+/- from their bristle phenotype. Evidence was shown that most such cells are segmental aneuploids, consistent with earlier studies of DNA repair mutants (Baker, 1978). We will summarize this in the revised manuscript so that it is not necessary to read the cited references to appreciate the evidence. The only new observation being made in this paper about the ~24h cell death stage is that loss of p53 increases the number of these cells, which could be because inadequate DNA repair leads to more aneuploid cells.

    It is important to appreciate that we do not claim that cells labeled by the DNA damage marker γH2Av are aneuploid, or being removed by cell competition. On the contrary, γH2Av labels cells with unrepaired DNA damage, whereas segmental aneuploidy can only occur as a consequence of completed DNA repair. Thus γH2Av-labeled cells are not generally expected to be Xrp1 positive or undergoing cell competition. Some may be, if they are cells that have both unrepaired DNA damage and repaired DNA damage that led to aneuploidy. We cannot quantify overlap in the existing data, since mouse antibodies for γH2Av and HA-tag were used in separate experiments. Repeating the experiments with different antibodies to measure the overlap would not address any outstanding questions.

    We doubt FISH would be effective at measuring aneuploidy because only gene dose corresponding to the probes would be detected. Only small portions of the genome could be assessed at a time so the frequency at which aneuploidy could be detected would be low. We will make it clear in the revised manuscript that cell competition of aneuploid cells is not a new claim of this paper but something that has been studied before.

    Regarding the statistical analysis, revisions are warranted. In multiple panels, Student's t-tests are repeatedly performed against the same control, which inflates the family-wise error rate and increases the risk of false-positive findings. In such cases, an overall ANOVA (one-way) followed by an appropriate multiple-comparison procedure-such as Dunnett's-test would be more appropriate.

    This concern applies in particular to:

    Figure 1A- Supplement 1

    Figure 2M-R

    Figure 3Q, R

    Figure 5D

    Figure 5J- Supplement 1

    Figure 6G- Supplement 1

    Figure 6I- Supplement 2

    We agree and will apply Anova with multiple comparison procedures in the revised manuscript.

    Minor comments:

    Figure 2E is not cited in the text, and it is difficult to tell from the images as presented whether p53DN overexpression suppresses the Gstd-lacZ signal at 4 h post-IR.

    We will replace Fig 2E with a clearer example, and add a quantification of all our data, with statistics, as a supplemental figure. Note that the conclusion is already substantiated by qRT-PCR data (Figure 2M)

    In Figure 4, rpr150-lacZ does not appear to be upregulated by Xrp1 overexpression. Therefore, the authors should revise the figure title to avoid misleading readers, because rpr, a well-known p53-responsive pro-apoptotic gene, is not induced under this condition.

    We will change the Figure title. Failure to induce rpr150-LacZ here is a control to show that Xrp1 overexpression does not induce p53 activity.

    In Figure 6E, based on the data as presented, it is difficult to determine whether cleaved Dcp-1 (cDCP1)-positive cell counts are reduced upon Xrp1 knockdown. The authors should provide clearer representative images and/or include the underlying raw images as supplementary source data to support the conclusion.

    We will replace Fig 6E with a clearer example, and add a quantification of all the data.

    The authors should (i) show raw data points overlaid on summary plots (e.g., dot plots on top of bar graphs/box plots) to convey data distribution and (ii) include higher-magnification insets and/or quantitative localization/overlap analyses where colocalization is central to the interpretation (e.g., Xrp1-HA relative to γH2Av).

    We agree regarding the data display. As discussed later, colocalization is not relevant to the interpretation.

    __Reviewer #2 __

    First, authors present evidence that Xrp1 is induced in wing discs exposed to ionizing radiation (IR, known to cause DSBs) and that this induction relies on p53 regulating Xrp1transcription (Figure 1 and S1). Data are clear but there is a puzzling result. Xrp1-lacZ (a reporter of Xrp1 transcription) is induced by IR but independently of p53. These results need attention as they appear to be contradictory (why Xrp1-mRNA but not Xrp1-lacZ relies on p53). Nicely, authors show that Xrp1-lacZ induction relies on Xrp1/Irbp18 autoregulatory feedback. Is the lacZ insertion somehow interfering with the capacity of p53 to bind and regulate Xrp1 expression?

    We agree that it is a puzzling result. We have also noted elsewhere that Xrp1-LacZ does not always reflect Xrp1 mRNA and protein expression (Kumar and Baker 2022). We can add the reviewer's hypothesis to the manuscript, although it does not explain why Xrp1-LacZ is induced by IR

    Second, authors use a collection of reporter genes and show that Xrp1 regulates, most but not all, Dp53 target genes. It is really unclear whether the reaper-lacZ used in Figure 3L-P recapitulates the induction of reaper by p53. I know this reporter was claimed by other do so, but NOT in the wing disc. I would then remove it as mRNA data are clear.

    rpr150-lacZ was used as a p53 reporter in wing imaginal discs by Wells et al. 2011 (PMC3296280). We will cite this in the revised manuscript. We prefer not to remove it as we also use this reporter for the experiment shown in Fig 4.

    3 Third, authors show that Xrp1, as expected from the previous data in Figure 2 and 3, also mediated the role of Dp53 in inducing cell death, although only partially, and these differences are attributed to the gene reaper (p53 but not Xrp1 target). Dcp1 should be cDcp1 and clones should be magnified in Fig 5E-G.

    We will follow this advice in the revised manuscript

    First, the impact of Xrp1 on the levels of DNA damage and cell death after 24h of IR are shown in a p53 mutant background (6E1-6E3). Authors should present the data in a clean +/+ background. Quantification of 6F should also be done in the same background.

    This data was presented in a the p53 mutant background to focus on the p53-independent removal of cells by cell competition. We can perform an experiment in the presence of wild type p53 for completeness if desired, but a mixture of DDR and cell competition effects may result.

    Second, hid-GFP is being induced by IR already at 4 h after IR and this induction and this induction relies on p53 and Xrp1 activities as shown in previous figures. Thus, the data presented in 6G-J could be a trivial consequence of the strong perdurance of the GFP protein.

    hid-GFP is not expressed at 4 hours in p53DN and Xrp1 K/D (Fig 3D,E), so the expression in 6G-J cannot be explained by GFP perdurance from the earlier timepoint.

    Third, the role of cell competition (driven by Minute aneuploids) is not demonstrated and relies simply on the potential role of Xrp1 in the late wave of cell death, proposal that has not been demonstrated in this paper either. Indeed, the no-role of RpS12 in the late induction (24 h wave) of Xrp1 (Figure 6 S1-F) reinforces my doubts. Authors should reflect in the introduction and discussion sections the most recent literature in the field.

    The role of Xrp1 in the late wave of p53-independent cell death is shown in Fig 6D-F. As discussed above (reviewer 1 point 1), Fig 6S1-F shows the limited role of p53 in rpS12-independent Xrp1 induction, not the role of RpS12. We will add a figure to the revised manuscript showing the strong RpS12 dependence of the late induction of Xrp1-HA and explain this more clearly. We did not include this in the first manuscript version because we had already published this result, albeit with an anti-Xrp1 antibody (Ji et al Fig 1 N-P). As also discussed above (reviewer 1 point 3), we agree that the role of cell competition in removing aneuploid cells is not demonstrated in the present manuscript, but we considered this had been demonstrated previously (Ji et al 2021), and parts of that study recently confirmed by others (Fusari 2025 Cell Genomics), so it is not necessary to add further experimental support here, although it will be useful to explain the published literature more fully.

    Reviewer #3

    Figure 2E. Based on the text, I think the authors are claiming that the expression of GStD-LacZ is reduced in the posterior compartment of panel 2E compared to 2D. This is unconvincing. If at all, the expression along the DV boundary in the posterior compartment is stronger in E than in D. Am I missing something?

    We will replace Fig 2E with a clearer example, and add a quantification of all our data, with statistics, as a supplemental figure. Note that the conclusion is already substantiated by qRT-PCR data (Figure 2M)

    Figure 3I - K. The expression in the posterior compartment is supposed to be reduced compared to the anterior compartment. Once again, these differences are not easily apparent to me. Perhaps these images need to be quantified to illustrate the supposed difference.

    We are sorry that the reviewer found the images unconvincing. We will replace these figures with other examples, and add quantifications of all data, with statistics, as a supplemental figure. Note that the conclusions are already substantiated by qRT-PCR data (Figure 3R)

    • . *

    Line 286. The heading "Xrp1 is sufficient for the expression of p53-dependent DDR genes" is misleading. As stated in the final sentence of paragraph 2 of this section, the authors show that Xrp1 functions downstream of p53 and is sufficient for expressing a subset of p53-dependent DDR genes.

    We apologize for misleading the reviewer. We will change the heading to "Xrp1 is sufficient for the expression of many p53-dependent DDR genes", which is the meaning we intended.

    Figure 5, panels F and G could be made much easier for the reader to follow. The labels in these two panels are very difficult to see and understand. It might be better to show some high magnification regions (e.g. insets) that show the differences in the prevalence of cell death in regions with different genotypes. Also, why is Xrp1 +/- not quantified in panel H since the authors claim that cell death is reduced even in the heterozygous cells?

    It is a good idea to add enlarged figures, and we will do so. We can quantify the Xrp1+/- genotype as well.

    Line 363 and Figure 6D, E. The authors argue that the increase in H2Av in the posterior compartment implies that cells with damaged DNA are not being eliminated when Xrp1 function is reduced. An alternative explanation is that the p53 mutation together with the Xrp1 knockdown impairs the DDR even more resulting in increased H2Av staining. I don't know how that authors' data can exclude this possibility.

    We agree with the reviewer and did not intend to exclude this possibility. We will rewrite this text to make both explanations clear.

    Line 365. Is the resolution of the "double labeling" sufficient to conclude that some of the H2Av cells upregulate Xrp1-HA? A more conservative interpretation would be that in these regions that have increased H2Av, that there is more expression of Xrp1-HA.

    We apologize for a mistake in the submitted manuscript. In fact the anti-H2Av and anti-HA primary antibodies used were both raised in mouse, and Fig 6G,H show distinct wing discs, not double labels. We will replace line 365 with the sentence suggested by the reviewer.

    Figure 6 - supplement 1. The expression of Xrp1-HA is reduced in the p53DN cells when they are a loss mutant for rps12. Although statistically significant, this reduction is modest. If this induction were due to a cell competition like phenomenon, would you not expect the induction to be completely abolished since rpS12 mutations abolish cell competition completely? Please explain.

    We apologize for confusing all three reviewers with Figure 6F supplement 1. This figure does not compare RpS12-dependent and -independent Xrp1-HA expression. Instead, it shows that the rps12-independent Xrp1-HA expression is only mildly p53-dependent, which is consistent with our conclusions. We will add a figure to the revised manuscript showing the strong RpS12 dependence of the late induction of Xrp1-HA and explain this more clearly. We did not include this in the initial manuscript version because we had already published this result, albeit with an anti-Xrp1 antibody (Ji et al Fig 1 N-P).

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Previous work has shown that when Drosophila imaginal discs are irradiated with X-rays that there are two phases of cell death. Within a few hours of irradiation, cells die in a p53-dependent manner. There is a much later phase of cell death that occurs approximately 20 hours after irradiation which seems to be mostly p53-independent. There is evidence that this latter phase of cell death might occur due to a phenomenon resembling cell competition where aneuploid cells are eliminated. In this manuscript, Chaitali Khan and colleagues explore the mechanistic basis of these two waves of cell death, focusing on the key regulator of cell competition Xrp1 and its relationship to p53. They make several conclusions: 1) Xrp1 appears to function downstream of p53 in activating the transcription of a number of genes involved in the DNA damage response. 2) Some pro-apoptotic genes but not others seem to be regulated via Xrp1 3) When p53 function is inhibited, cells with damaged DNA accumulate and Xrp1 expression is increased at the late time points. 4) Xrp1 contributes to the late death in the absence of p53 function consistent with its role in elimination of these cells by a mechanism that resembles cell competition. Overall the data are clean and the conclusions are mostly justified. Some conclusions appear a little overstated (see below). The authors could address most of these issues by more careful presentation of data and by more conservative interpretations of some of their experiments.

    1. Figure 2E. Based on the text, I think the authors are claiming that the expression fo GStD-LacZ is reduced in the posterior compartment of panel 2E compared to 2D. This is unconvincing. If at all, the expression along the DV boundary in the posterior compartment is stronger in E than in D. Am I missing something?

    2. Figure 3I - K. The expression in the posterior compartment is supposed to be reduced compared to the anterior compartment. Once again, these differences are not easily apparent to me. Perhaps these images need to be quantified to illustrate the supposed difference.

    3. Line 286. The heading "Xrp1 is sufficient for the expression of p53-dependent DDR genes" is misleading. As stated in the final sentence of paragraph 2 of this section, the authors show that Xrp1 functions downstream of p53 and is sufficient for expressing a subset of p53-dependent DDR genes.

    4. Figure 5, panels F and G could be made much easier for the reader to follow. The labels in these two panels are very difficult to see and understand. It might be better to show some high magnification regions (e.g. insets) that show the differences in the prevalence of cell death in regions with different genotypes. Also, why is Xrp1 +/- not quantified in panel H since the authors claim that cell death is reduced even in the heterozygous cells?

    5. Line 363 and Figure 6D, E. The authors argue that the increase in H2Av in the posterior compartment implies that cells with damaged DNA are not being eliminated when Xrp1 function is reduced. An alternative explanation is that the p53 mutation together with the Xrp1 knockdown impairs the DDR even more resulting in increased H2Av staining. I don't know how that authors' data can exclude this possibility.

    6. Line 365. Is the resolution of the "double labeling" sufficient to conclude that some of the H2Av cells upregulate Xrp1-HA? A more conservative interpretation would be that in these regions that have increased H2Av, that there is more expression of Xrp1-HA.

    7. Figure 6 - supplement 1. The expression of Xrp1-HA is reduced in the p53DN cells when they are alos mutant for rps12. Although statistically significant, this reduction is modest. If this induction were due to a cell competition like phenomenon, would you not expect the induction be be completely abolished since rpS12 mutations abolish cell competition completely? Please explain.

    Minor issue:

    Line 152: I assume you mean "p53-dependent apoptosis" and not p-53-dependent DDR".

    Significance

    Overall this manuscript clarifies the role of Xrp1 in DNA-damage repair and cell death following X-ray irradiation. Since mammals do not have an Xrp1 ortholog and mammalian p53 seems to function in cell competition similar to Xrp1 in Drosophila, this raises the interesting possibility that the tumor-suppressive function of p53 could, at least in part, be due to its role in cell competition that eliminates aneuploid cells.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    In this ms, authors utilize the Drosophila wing epithelium as model system to analyze the role of Xrp1 in DNA-damage induced cell death. The p53 gene is well known to have a conserved role (in mammals and flies) in driving cell death (and DNA repair) upon DNA damage induction (double stranded breaks, DSBs, in particular). Xrp1, a transcription factor mostly known for its role in cell competition induced by haploinsufficiency of ribosomal encoding genes (Minute genes), was indeed identified as a target of p53, but its role in the DNA damage response pathway was not addressed. In this ms, Baker and colleagues fill this gap. The paper is subdivided into the following chapters/figures: First, authors present evidence that Xrp1 is induced in wing discs exposed to ionizing radiation (IR, known to cause DSBs) and that this induction relies on p53 regulating Xrp1 transcription (Figure 1 and S1). Data are clear but there is a puzzling result. Xrp1-lacZ (a reporter of Xrp1 transcription) is induced by IR but independently of p53. These results need attention as they appear to be contradictory (why Xrp1-mRNA but not Xrp1-lacZ relies on p53). Nicely, authors show that Xrp1-lacZ induction relies on Xrp1/Irbp18 autoregulatory feedback. Is the lacZ insertion somehow interfering with the capacity of p53 to bind and regulate Xrp1 expression? Second, authors use a collection of reporter genes and show that Xrp1 regulates, most but not all, Dp53 target genes. It is really unclear whether the reaper-lacZ used in Figure 3L-P recapitulates the induction of reaper by p53. I know this reporter was claimed by other do so, but NOT in the wing disc. I would then remove it as mRNA data are clear. Third, authors show that Xrp1, as expected from the previous data in Figure 2 and 3, also mediated the role of Dp53 in inducing cell death, although only partially, and these differences are attributed to the gene reaper (p53 but not Xrp1 target). Dcp1 should be cDcp1 and clones should be magnified in Fig 5E-G. The last figure (6 and the two supplementary figures) are devoted to address the impact of Xrp1 in the well-known p53 independent second wave of cell death caused by IR (24 h later) induced by JNK and attributed by the Brodsky lab to the induction of aneuploid karyotypes (as a result of mistakes in DNA repair). Many of the results this section might be an artefactual consequence of GFP perdurance, some of the genetic tests are not clean enough, and lastly, the role of cell competition in this process relies on correlation (Xrp1 induction) but not clear functional data has been provided so far. I will go point by point

    (1) First, the impact of Xrp1 on the levels of DNA damage and cell death after 24h of IR are shown in a p53 mutant background (6E1-6E3) Authors should present the data in a clean +/+ background. Quantification of 6F should also be done in the same background

    (2) Second, hid-GFP is being induced by IR already at 4 h after IR and this induction and this induction relies on p53 and Xrp1 activities as shown in previous figures. Thus, the data presented in 6G-J could be a trivial consequence of the strong perdurance of the GFP protein.

    (3) Third, the role of cell competition (driven by Minute aneuploids) is not demonstrated and relies simply on the potential role of Xrp1 in the late wave of cell death, proposal that has not been demonstrated in this paper either. Indeed, the no-role of RpS12 in the late induction (24 h wave) of Xrp1 (Figure 6 S1-F) reinforces my doubts.
    Authors should reflect in the introduction and discussion sections the most recent literature in the field.

    Significance

    Overall, data presented in Figures 1-5 fills an important gap (a role of Xrp1 in mediating the activity of p53 in tissues subjected to IR and DNA damage) and data are convincing and need minor revision.

    However, the proposed role of Xrp1 and/or cell competition in this mysterious second wave of cell death attributed to the generation of aneuploid karyotypes is not demonstrated. Genetics are not clean, impact on reporters might by affected by GFP perdurance and the contribution of cell competition is correlative and lacks solid functional validation.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    This study reports that the Drosophila transcription factor Xrp1 plays two temporally distinct and crucial roles in maintaining genome integrity: a p53-dependent acute DNA damage response (DDR) and subsequent p53-independent cell competition. The authors show that immediately following ionizing radiation (IR), Xrp1 is induced in a p53-dependent manner. During this acute phase, Xrp1 acts as an effector of the p53-driven DDR by promoting the expression of target genes, including the pro-apoptotic gene hid and DNA repair genes such as rad50, mus205, lig4, and Ku80. Approximately 16 hours after IR, as the acute DDR winds down, Xrp1 is induced in a largely independent manner of p53 but in a RpS12-dependent manner. This second phase of Xrp1 induction serves to eliminate affected cells by cell competition, possibly through acquired aneuploidy (specifically segmental monosomies affecting Rp gene dose) due to defective DNA repair. Furthermore, the authors show that reducing p53 function increases the persistence/accumulation of γH2Av-positive cells at 24 h post-IR, supporting a model in which Xrp1 contributes both to early DDR outputs and to later tissue-level quality control after irradiation.

    Major comments

    1. The idea that Xrp1 induction switches around 16 h post-IR, becomes RpS12-dependent, and subsequently engages cell competition is interesting and potentially important. However, the evidence supporting RpS12-dependence of Xrp1 induction is currently not sufficiently convincing. For example, based on the images in Figure 6F- supplement 1, the conclusion that Xrp1 is induced in an RpS12-dependent manner appears difficult to support. The authors should strengthen and quantify this result or provide the raw image data. In addition, because this point is central to the authors' model, they should move the key supporting data from the supplementary figures to the main figures to ensure that this critical claim is clearly supported and readily accessible to readers.
    2. The authors suggest a model in which Xrp1 executes two qualitatively distinct "modes" (pro-repair/acute DDR and elimination of aneuploid cells), but this remains only partially convincing as currently presented. The authors should at least (i) provide quantitative evidence that could explain how Xrp1 might produce distinct outcomes across phases (e.g., comparing Xrp1-HA levels and/or the fraction of Xrp1-HA-positive cells at 2-4 h versus 16-24 h post-IR), and (ii) explicitly discuss plausible mechanisms in the Discussion. Even if the molecular "switch" is not fully resolved experimentally, a clearer, data-grounded discussion of how Xrp1 could mediate these temporally distinct functions is needed. In addition, since ISR signaling (e.g., eIF2α phosphorylation) has been implicated as a single feature associated with Xrp1-dependent loser elimination, the authors should consider assessing p-eIF2α levels in Xrp1-HA positive cells at early versus late time points after IR (e.g., 4 h vs 24 h).
    3. The idea that aneuploid cells-or cells with altered ribosomal gene dosage-could be removed via Xrp1-mediated cell competition is intriguing. However, the manuscript does not currently provide any evidence that such cells are, in fact, being eliminated. The authors should therefore (i) quantify cell-level overlap metrics, such as the fraction of γH2Av-positive cells that are Xrp1-HA-positive (and vice versa), as well as the fraction of γH2Av-positive cells that are cleaved Dcp-1-positive (and vice versa) at 24 h post-IR. These quantitative analyses would clarify whether the late Xrp1-HA-positive population corresponds to persistently damaged cells and whether it is enriched for cells undergoing apoptosis/clearance. The authors should also (ii) directly assess aneuploidy/segmental copy-number imbalance in the late Xrp1-HA-positive clusters (e.g., by DNA FISH targeting one or two chromosome arms/regions), and if these experiments cannot be completed within a reasonable revision timeframe, the authors should temper their wording and present aneuploidy and selective elimination as a plausible interpretation supported by RpS12 dependency and prior literature, rather than as a demonstrated conclusion in the current study.
    4. Regarding the statistical analysis, revisions are warranted. In multiple panels, Student's t-tests are repeatedly performed against the same control, which inflates the family-wise error rate and increases the risk of false-positive findings. In such cases, an overall ANOVA (one-way) followed by an appropriate multiple-comparison procedure-such as Dunnett's-test would be more appropriate. This concern applies in particular to: Figure 1A- Supplement 1 Figure 2M-R Figure 3Q, R Figure 5D Figure 5J- Supplement 1 Figure 6G- Supplement 1 Figure 6I- Supplement 2

    Minor comments

    1. Figure 2E is not cited in the text, and it is difficult to tell from the images as presented whether p53DN overexpression suppresses the Gstd-lacZ signal at 4 h post-IR.
    2. In Figure 4, rpr150-lacZ does not appear to be upregulated by Xrp1 overexpression. Therefore, the authors should revise the figure title to avoid misleading readers, because rpr, a well-known p53-responsive pro-apoptotic gene, is not induced under this condition.
    3. In Figure 6E, based on the data as presented, it is difficult to determine whether cleaved Dcp-1 (cDCP1)-positive cell counts are reduced upon Xrp1 knockdown. The authors should provide clearer representative images and/or include the underlying raw images as supplementary source data to support the conclusion.
    4. The authors should (i) show raw data points overlaid on summary plots (e.g., dot plots on top of bar graphs/box plots) to convey data distribution and (ii) include higher-magnification insets and/or quantitative localization/overlap analyses where colocalization is central to the interpretation (e.g., Xrp1-HA relative to γH2Av).

    Significance

    This study puts forward an appealing conceptual framework in which Xrp1 exhibits temporally distinct activation patterns after ionizing radiation and may connect cell-autonomous DDR outputs with non-cell autonomous tissue-level quality control. In particular, the idea that Xrp1 can function both as a downstream effector of p53-associated DDR programs and as a mediator of the subsequent elimination of damaged cells is potentially important for understanding how epithelia maintain homeostasis under genotoxic stress.

    The study should be of interest to DDR researchers because it dissects p53 downstream outputs in a genetically tractable in vivo tissue context and provides a temporal framework for how p53-linked programs are coordinated after irradiation.

    The manuscript will be of particular interest to the cell competition community. By proposing that RpS12-dependent Xrp1 induction engages a damaged-cell elimination program after IR, the study raises the possibility that tissues exposed to genotoxic stress might exploit cell competition as a quality-control machinery. Even if some mechanistic aspects require stronger support, this framework could broaden the contexts in which cell competition is thought to contribute to tissue homeostasis.

    The reviewer's expertise: mechanism of tissue growth control in Drosophila.

  5. Version published to 10.1101/2022.06.06.494998 on bioRxiv