Centriolar subdistal appendages promote double strand break repair through homologous recombination

Article activity feed

1. 

   Review Commons

   Dec 7, 2022

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

   Learn more at Review Commons

   ---

   Reply to the reviewers

   The authors do not wish to provide a response at this time

   Read the original source
2. 

   Review Commons

   Dec 7, 2022

   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

   Summary

   In this manuscript, Rodríguez -Real and colleagues investigate how the centrosome may influence the repair of DNA double-stranded breaks (DSBs), building on the initial finding that relative HR frequencies (as measured using a standard split-GFP gene conversion reporter assay) are reduced in centrinone treated centrosome-depleted cells relative to mock treated controls cells. Such defects are found correlate to concordant reductions in immunofluorescence proxies for resection (RPA recruitment into foci) and upstream and downstream events in the HR cascade (BRCA1 and RAD51 recruitment, respectively), and a correlating …

   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

   Summary

   In this manuscript, Rodríguez -Real and colleagues investigate how the centrosome may influence the repair of DNA double-stranded breaks (DSBs), building on the initial finding that relative HR frequencies (as measured using a standard split-GFP gene conversion reporter assay) are reduced in centrinone treated centrosome-depleted cells relative to mock treated controls cells. Such defects are found correlate to concordant reductions in immunofluorescence proxies for resection (RPA recruitment into foci) and upstream and downstream events in the HR cascade (BRCA1 and RAD51 recruitment, respectively), and a correlating increase in NHEJ repair of I-SceI induced repair in EJ5-like reporter assay. Taking a candidate approach to identifying which centrosome proteins link the centrosome to DSB repair regulation, the authors reveal cells depleted for subdistal appendage proteins show equivalent deviations in DSB repair reporter assays and show concordant defects in RPA recruitment, leading to the proposal that subdistal appendage proteins regulate DNA resection and thus optimal HR. Experiments are then used to show CEP170 (a subdistal appendage protein) may be phosphorylated by DDR kinases and some rescue experiments are used to support hypothesis that this phosphorylation may be involved in centrosome-DSB repair cross-talk signalling. Figure 3 experiments then show centrosome-depleted and heterozygous losses of CEP170 result in moderate sensitivities across a number of DSB-inducing treatments. Lastly meta-analyses of cancer datasets correlate low CEP170 expression to differences in cancer mutations signatures (Fig 4) and altered patient outcomes across a number of cancers (Fig 5), and propose that CEP170 - via a DSB repair repair function - may be causal in these alterations. Ultimately, the authors propose that the centrosome acts as a signalling node or 'centrosomal processing unit' (CPU) via distal appendage proteins to coordinate the signalling of DNA damage and its repair, and speculate this may link to the clinical phenotypic overlap between centrosome-related ciliopathies and DDR signalling disorders (e.g. ATR-Seckel).

   Major comments

   1. Concerning Figs 1-3, it is argued that the presented skews in pathway choice are not an indirect consequence of cell-cycle effects that accompany centrosome depletion (i.e. following centrinone treatments) or depleted centrosome factors. Indeed, S1B shows centrione depleted cell show reduced S-phase indices (where HR is most active) are concordant with increased G2(/M) cell indices, significant effects that may contribute (at least in part) to some of the reported. In the case of the reporter assays it will be difficult/impossible to normalise data vs cell cycle skew, however in the case of RAD51 IRIF frequencies and RPA recruitment, this can be done easily by monitoring the relative frequencies of these events specifically S-phase (BrDU/EdU positive) cells. This should be done if the case for indirect cell-cycle effects is to be dismissed.
   2. Related to point (1): RPA/RAD51/BRCA1 measurements made quantitatively (i.e. by QIBC or equivalent) given % IRIF positive cells can be misleading given it is completely subjective to user defined thesholds.
   3. Fig 3 - The fact that CEP170 KD decreases BRCA1 IRIF but does not increase RIF1 IRIF, is not indicative of a lack of NHEJ stimulation, nor does it infer the existence of a/some distinct mechanism stimulating NHEJ, or an 'undiscovered factor', as is stated. This is important as RIF1 IRIF are not an accepted, nor accurate surrogate marker of NHEJ pathway activity, only an indicator of RIF1 recruitment downstream of 53BP1, whose role in resection control is clear, yet whose contribution to NHEJ is highly context-specific.
   4. Is CEP170 Ser-637 an evolutionarily conserved ATM/ATR site? - Conservation, at least in mammals/vertebrates would be expected if a regulatory event in DSB pathway choice. This should be commented on with supplementary alignment included to demonstrate whether this is likely to be a universally conserved mechanism of repair regulation.
   5. Fig 3F-G: Important to show appendage localisation of wild-type and mutant CEP170 S637A/D proteins to inform whether these are functional, expressed at equivalent levels and support equal centrosome localisation intensities. Immunoblot data in support of CEP170 siRNA depletion and CEP170 transgene complementation efficiencies is missing, and needs to be included to reassure a reader the results are specific to defects in the phosphorylation (not stability/expression level/other).
   6. Do the CEP170 P'n nmutations affect its physiological centrosome functions? If separation of function is not experimentally defined, it should be at least discussed.

   Comments on interpretation and accuracy of stated conclusions:

   1. P12. - The manuscript is lacks the necessary evidence to support the section title: "CEP170 Ser647 phosphorylation is critical for HR double strand break repair", and as such I find this and related textual conclusions in the manuscript body to be inaccurate and misleading. To make this claim would require generating a cell-line knockin of the S647A mutation, preferably at the endogenous CEP170 locus (or a robust complementation system), and its utilisation to establish that standard measures of HR e.g. RAD51 recruitment, PARPi sensitivity, and/or SCE frequencies are all affected as expected in cells bearing this mutation.
   2. Abstract reads: "we identify a centriolar structure, the subdistal appendages, and a specific factor, CEP170, as the critical centrosome component involved in the regulation of recombination and resection... " - I disagree with this statement given that the study has not excluded other centrosome components/features of the centrosome in regulation of resection. Can the authors perform experiments to exclude a role for other centrosome components and substantiate the conclusion that this is a specific function of the subdistal appendages as is stated?
   3. Based on the marginal sensitivity phenotypes shown in Fig 4 for heterozygous cell-lines, it seems unlikely that CEP170 is a central player in the DSB response.
   4. The CPU model for DDR-centric role of the centrosome is premature based on the provided data, likewise the fact that a centrosome-regulated resection could explain the clinical overlap between seckel and and this model should be toned down. We probably don't need another acronym for the DDR.

   Minor comments

   • Abstract, lasts sentence needs correction: "suggesting this protein can act as a driver mutation but also..." - a protein cannot act as a driver mutation.
   • Information regarding biological replicates, sample sizes, error bars should be made more clear throughout to better represent reproducibility; e.g. n=3 {plus minus} Dt. Dev, biological replicates consisting >500 cells/nuclei per condition

   Significance

   General assessment

   In exploring for functional links between DSB repair and the centrosome, the results encompass a series of corelating results that collectively hint at a potential role for the centrosome in repair regulation. The indirect and perhaps boring explanation for the presented DSB repair imbalances is these are an indirect consequence of the inevitable cell cycle defects that accompany centrosome depletion. In S1 the authors make some effort towards dispelling this less interesting (indirect) explanation for the presented results, yet not really far enough to dismiss it as the unifying explanation. A major consequence of centrosome-loss is prolonged time spent in G2/M dues to sub-optimal spindle nucleation and assembly kinetics, and an extended transit through mitosis, defects that occur independently of the p53-dependent checkpoint to centrosome loss (in fact the defects have long been speculated precede and perhaps propagate p53 activation). Indeed, supplementary data indicates that in centrosome-depleted cells a reduction in S-phase index (when HR activity is highest) correlates to greater proportion of cells with DNA with G2(/M) content. While I agree that these cell-cycle skews are unlikely to be great enough fully account for the reductions in HR reporter and IF proxies, more targeted approaches to control for indirect cell cycle effects (one suggestion below) could strengthen the case for a direct role in repair regulation. The manuscript also falls short of a identifying a discrete mechanism that explains centrosome-repair crosstalk, and on this basis I feel some of the conclusions are too preliminary and speculative and thus the authors would benefit from being more nuanced in their conclusions. One clear example is the authors's oversimplistic attribution of DSB regulation to distal appendage components of the centrosome/cilia, yet doing so having only tested the appendage proteins on the basis of literature based exercise of protein segregation of DDR and centrosome proteins (S2A). I also find it premature to propose "CPU" models of DDR regulation, the results (while interesting) haven't gone far enough to rigorously challenge this hypothesis, and define its mechanistic basis. I also question the importance and relevance of the analyses in Figs 4-5: in the absence of scientific evidence to establish causation for low CEP170 expression in tumour mutation signature burden or patient prognosis, the presented remain correlates that might equally result from a number of phenomena unrelated to DSB repair. As such, I feel the manuscript does encompass results worthy of report that would be of interest to cell cycle and DNA repair biologists, it would be greatly improved by being more rigorous, objective and nuanced in its interpretation.

   Read the original source
3. 

   Review Commons

   Dec 7, 2022

   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 manuscript by Rodriguez-Real et al, the authors address the contribution of the centrosome to cellular process unrelated to organizing the microtubule cytoskeleton, namely DNA repair. As many proteins contributing to the DNA damage response physically associate with centrosomes, this appears a relevant question that has been neglected so far and led to a number of studies that appeared controversial. To do so, the authors exploit a variety of tissue culture models that are well established in the fields of centrosomes and DNA repair, including U2OS and RPE1 cells, exposed to perturbations promoting DNA damage (such as …

   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 manuscript by Rodriguez-Real et al, the authors address the contribution of the centrosome to cellular process unrelated to organizing the microtubule cytoskeleton, namely DNA repair. As many proteins contributing to the DNA damage response physically associate with centrosomes, this appears a relevant question that has been neglected so far and led to a number of studies that appeared controversial. To do so, the authors exploit a variety of tissue culture models that are well established in the fields of centrosomes and DNA repair, including U2OS and RPE1 cells, exposed to perturbations promoting DNA damage (such as ionizing radiation or pharmacologic perturbation of DNA stability) in conjunction with siRNA mediated depletion of candidate centrosomal proteins., followed by the visualization of repair events either using fluorescent reporters, or visualizing endogenous repair foci by immunofluorescence. On this basis, the authors propose that a discrete centrosomal sub-structure, namely sub-distal appendages and the CEP170 protein therein concur to promote a particular nuclear DNA repair process, namely homologous recombination.

   The manuscript suffers of two main limitation:

   1. the authors provide no mechanistic understanding of how CEP170, a protein that resides at centriolar subdistal appendages and shows no nuclear translocation upon DNA damage, concurs to regulate processes in the nucleus. The fact that all reported phenomena appear to be independent of microtubules suggests that neither the LINC complex nor the precise position of the centrosome in the vicinity of nuclear pore complexes contribute to the reported phenomena.
   2. some of the experimental perturbations performed in the manuscript might elicit the reported phenotypes due to spurious effects on cellular processes that have not been considered with sufficient caution.

   Given that uncovering the mechanism underlying the contribution of CEP170 to homologous recombination might prove very demanding, my comments will focus primarily on the second point.

   Major comments:

   The centriolar depletion using centrinone is known to impinge on cell proliferation in p53 WT cells. Thus, I am not convinced that the data shown in Figure S1B and S1C will sufficiently document that the observed unbalance between HDR and NHEJ are not simply reflecting a different cell cycling speed/behavior. Moreover, it would be important to address whether centrinone or depletion of CEP170 (an essential gene, according to the authors!) will trigger DNA damage by themselves. In fact, even a small extent of chronic genotoxic stress caused by the perturbations used in the manuscript might explain the reported differential proficiency of HDR.

   Minor comments:

   It is a pity that CEP170 is not amenable to functional dissection using a complete knockout. The fact that in PMID: 27818179 a complete knockout of CEP128 has been achieved, suggests however that subdistal appendage mediated DNA repair is not the essential process in itself. As the authors employ other cell lines stemming from the same laboratory, they could consider acquiring CEP128 KO to complement their own experiments.

   The proposal that CEP170 phosphorylation of by ATM/ATR upon DNA damage might require SDA localization of the protein is plausible, yet not circumstantiated by any experimental evidence. If the authors could monitor the phosphorylation of the endogenous CEP170 protein in WT vs CEP128 KO cells (phosphor-specific antibody, MS-based proteomics or simply "phos-tag" gels), this could provide a first spark towards a mechanistic understanding of the reported phenomenon.

   The entire Figure 4 is based on quantifications of clonogenic potential.

   1. it would be helpful if the data were accompanied by images displaying representative crystal violet stained dishes.
   2. clonogenic potential potential is discussed as a mere readout of cell survival, yet a combination between survival and proliferation concur to the reported differential clonogenic potential

   Odf2 contribution to both DAs and SDAs: while Odf2 has been initially proposed to be necessary for the assembly of both types of appendages, its contribution to distal appendages has been disputed by Tanos et al using siRNA (PMID: 23348840), also confirmed by our group using CRISPR (unpublished). Thus, the role of Odf2 in SDA assembly appears more crucial than for DA assembly.

   CEP164 contribution to ATM/ATR activation: this has been disputed in this paper by the Morrison lab (PMID: 26966185). Thus, a cautionary note should be mentioned when referring to this concept.

   Significance

   Taken together, this manuscript addresses the contribution of the centrosome to DNA repair. This is in itself a very interesting topic with the potential to attract the interest of both cell/molecular biologists as well as cancer researchers. The major advance strength is represented by pinpointing a specific centriolar substructure, namely subdistal appendages, in the control of HDR. CEP170 had been previously shown to be target of phosphorylation by ATM/R and the present study highlights that the abovementioned phosphorylation is not a mere passenger event during DNA repair, but that potentially reflects a decisive event informing the repair pathway of choice. However, several experiments have alternative explanations/interpretations and no understanding of the underlying mechanism is provided.

   The expertise of this reviewer is the study of cell cycle regulation and on the centrosome structure/function.

   Read the original source
4. 

   Review Commons

   Dec 7, 2022

   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

   Summary

   Rodríguez-Real, Huertas and colleagues here explore the roles of centrosomes in DNA damage responses, focussing on DNA repair activities. They show that centrosome depletion by PLK4 inhibition leads to reduced levels of homologous recombination and increased nonhomologous end-joining, along with altered level of nuclear focus formation by DNA repair proteins. Knockdown of genes that encode components of centriolar subdistal appendages (SDAs) cause reduced levels of RPA foci, with CRISPR-generated CEP170 heterozygotes also showing defects in focus formation. Knockdown of CEP170 impairs homologous recombination, although NHEJ …

   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

   Summary

   Rodríguez-Real, Huertas and colleagues here explore the roles of centrosomes in DNA damage responses, focussing on DNA repair activities. They show that centrosome depletion by PLK4 inhibition leads to reduced levels of homologous recombination and increased nonhomologous end-joining, along with altered level of nuclear focus formation by DNA repair proteins. Knockdown of genes that encode components of centriolar subdistal appendages (SDAs) cause reduced levels of RPA foci, with CRISPR-generated CEP170 heterozygotes also showing defects in focus formation. Knockdown of CEP170 impairs homologous recombination, although NHEJ activities are unaffected. Some increase in sensitivity to DNA damaging agents is seen in CEP170- or centriole-deficient cells, albeit with a modest effect size. CEP170 status is shown to affect mutational signatures and patient prognosis in different cancer samples.

   While the experiments are generally well-presented and controlled, the effects seen are not large, so that the the conclusions that the authors draw are not entirely substantiated by the data presented, even without the suggestion of a mechanism. There are several additional experiments and clarifications that I consider necessary to provide appropriate support for the phenomenon.

   Major points

   1. The lack of cell cycle arrest or phenotype in the U2OS cells after a week's treatment with centrinone is somewhat surprising, given their p53 status. The initial description of centrinone showed a distinct impact on U2OS proliferation, albeit after 2 weeks' treatment (although the present paper shows robust impact on centriole numbers after only 1 week in centrinone). It would be useful to know the percentage of mitotic cells, or if there is any increased cell death observed at this stage of treatment.
   2. In the I-SceI assays, were transduction efficiencies or apoptosis within the experiment impacted by centrinone treatment? If not, it would be useful to state that this was examined and that there were no confounding effects; having only normalised data does not allow the reader to exclude these potential confounding factors.
   3. The authors present binary data for a given type of nuclear focus (positive or negative for RPA/ BRCA1/ RAD51), while the supporting images show altered numbers/ intensities. For example, the BRCA1 signals shown in Fig. 3D are less readily distinguished than they are in Fig. 1D. These data should be reconsidered: it is possible that these observations reflect different kinetics of focus formation, rather than a change in IRIF formation capacity. Numbers and a timecourse should be provided, with details of how these are quantitated provided in the Methods.
   4. Are the BRCA1 and RAD51 results seen with centrinone treatment of U2OS cells recapitulated in the Saos-2 and RPE1 lines?
   5. Some additional analysis is needed of the extent to which cells are sensitised to DNA damaging treatments by CEP170 deficiency or centrinone treatment. It should be confirmed that these experiments were performed in biological triplicate, rather than a technical triplicate (within a single experiment); if this is not the case, these experiments should be done in triplicate. Analysing p53-deficient hTERT-RPE1 clones, Kumar et al. (NAR Cancer 2020 PMID: 33385162) showed <10% survival with 100 ng/ml NCS. Hustedt et al. (Genes Dev 2019 PMID: 31467087) showed just over 50% survival with 10 nM CPT treatment, although their data for IR were comparable to the current study. Given the wide variation that these assays seem to incur, the extent to which a ≈20% difference in clonogenic survival is biologically significant may be limited. A rescue of the CEP170 siRNA, and/ or washout in the centrinone experiment would make these data more convincing. The knockdown of CEP170 in Figure 4 should be correctly labelled (not as CEP170+/-); given that the authors have generated CEP170 heterozygotes in Figure 2, this is potentially confusing.
   6. Direct data for the (centrosomal) phosphorylation of CEP170 are limited; it has not been demonstrated that the S637A mutants are fully functional in terms of the centrosome functions of CEP170, so that the conclusion regarding a requirement for centrosomal CEP170 phosphorylation is not sufficiently supported by the available data. The CEP170-dependent changes in RPA focus positive cell percentages shown in Figure 3 are not very marked. The relevant sections should be revised, or the authors should include additional experiments showing directly a phosphorylation of CEP170.
   7. It is difficult to interpret the mutational spectrum data and their significance. These should be compared with data for mutations in NDEL1 mutant cells, and/or other SDA components.
   8. The Kaplan-Meier curves data are intriguing, but their interpretation is highly speculative, given that there are no data on treatment groups included in this study. It is unclear whether other genes that affect SDAs might also impact survival (in the same, or different cancers), so the presentation of those patient groups where CEP170 status impacted survival seems selective, given the ubiquity of HR and centrosomes. These data would be better included as Supplemental information.
   9. The independence of p53 status/ responsiveness of the system is a crucial aspect of this study. Sir et al. (JCB 2013 PMID: 24297747) showed no DNA repair defect in centrosome-deficient chicken DT40 cells. This paper is very relevant to the current study and should be discussed. Similarly, the work by Lambrus et al (JCB 2015 PMID: 26150389) should also be considered.

   Minor points

   1. References for the RPE1 TP53/ SAS6 mutant cell lines should be provided (or controls for their generation presented).
   2. Fig S1K should correct its x-axis to reflect the time intervals correctly.
   3. Fig 2D should show blow-ups of the centrosomes.
   4. To avoid any potential confusion, it would be helpful to indicate in the Figure proper which cells are used for the various analyses.
   5. The 'basal side' of the centriole is not a standard term- this should be clarified. This may be confusing, given the role of centrioles in the basal body.
   6. The consideration of Seckel syndrome seems somewhat speculative at this stage in the exploration of this phenomenon.

   Referees cross commenting I think the comments from Reviewers #2 and #3 are reasonable and justified; there is good convergence between the comments that we all made and I have no issues to raise in this cross-commentary.

   Significance

   Strengths: Much previous work linking centrosomes and DNA damage responses has addressed cell cycle and checkpoint roles of the centrosome, so that a direct role in (nuclear) DNA repair is intriguing. Limitations:The present study shows a relatively moderate impact of centrosome defects on DNA repair, without a clear mechanism. There are some technical details that should be addressed. The relatively limited sensitization to DNA damaging treatments caused by centrosome deficiency questions the biological significance of the phenomenon.

   Advance: The current study presents some new findings that potentially show DNA repair defects resulting from the loss of centrioles (or SDA proteins). This has not been demonstrated to date.

   Audience: The idea of subdistal appendage components contributing to homologous recombinational repair of DNA damage is of potential interest to several fields, ranging from basic centrosome biology through translational to clinical cancer research.

   Reviewer's expertise: basic/ cell biology.

   Read the original source
5. 

   Version published to 10.1101/2022.10.19.512819 on bioRxiv

   Oct 19, 2022