Gcn5 – mTORC1 – TFEB signalling axis mediated control of autophagy regulates Drosophila blood cell homeostasis

  1. Arjun Aji Reeja
  2. Mincy Kunjumon
  3. Suraj Math
  4. Laxmi Kata
  5. Rohan Jayant Khadilkar

  • Curated by eLife

    eLife logo

    eLife Assessment

    This valuable study identifies Gcn5 as a regulator of blood cell development in the Drosophila lymph gland, with links to autophagy and nutrient-sensing mTORC1 signalling. The evidence is solid that altering Gcn5, autophagy genes and mTORC1 activity perturbs blood cell homeostasis, and the revised manuscript adds helpful genetic and quantitative analyses. However, the evidence for a clean linear Gcn5-mTORC1-TFEB/autophagy pathway is insufficient, because several cell-type-specific phenotypes remain difficult to reconcile and the pathway logic relies on different genetic tools, cell populations and pharmacological perturbations.

Reviewed by

Read the full article

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Blood progenitors are regulated by a variety of systemic and nutritional cues from their environment. In the Drosophila lymph gland (LG), the Posterior Signalling Center (PSC) acts as a stem cell niche striking a balance between progenitors and differentiated blood cells. Blood progenitors maintain homeostasis by fine tuning intrinsic and extrinsic cues.. Autophagy is one such cellular process that maintains homeostasis by removing unnecessary or dysfunctional cell components through autophagic degradation and recycling. Here, using genetic perturbation analysis we show that autophagy plays a critical role in regulating LG blood cell homeostasis. General control non-derepressible 5 (Gcn5), a histone acetyltransferase is expressed in the primary LG lobe and modulation of Gcn5 levels perturbs LG homeostasis. Our results demonstrate thatGcn5 through its known non-histone acetylation target, TFEB controls autophagic flux in the hemocytes.. Additionally, we show that modulation of mTORC1 activity can perturb hematopoiesis. Our results indicate that Gcn5 acts as a nutrient sensor and modulation of mTORC1 activity regulates Gcn5. Chemical intervention shows that mTORC1 over-rides the effect exerted by Gcn5 in regulating LG hematopoiesis. Taken together, our findings indicate that Gcn5 – mTORC1 – TFEB signaling axis mediated control of autophagy is required for maintaining blood cell homeostasis in Drosophila.

Article activity feed

  1. eLife

    eLife Assessment

    This valuable study identifies Gcn5 as a regulator of blood cell development in the Drosophila lymph gland, with links to autophagy and nutrient-sensing mTORC1 signalling. The evidence is solid that altering Gcn5, autophagy genes and mTORC1 activity perturbs blood cell homeostasis, and the revised manuscript adds helpful genetic and quantitative analyses. However, the evidence for a clean linear Gcn5-mTORC1-TFEB/autophagy pathway is insufficient, because several cell-type-specific phenotypes remain difficult to reconcile and the pathway logic relies on different genetic tools, cell populations and pharmacological perturbations.

  2. eLife

    Reviewer #1 (Public review):

    In their manuscript Arjun et al. investigate the role of the histone acetyl transferase Gcn5 in controlling drosophila blood cell homeostasis in the larval lymph gland. Using gcn5 zygotic mutants as well as targeted knock-down and over-expression of Gcn5 in various lymph gland cell populations, they show that these manipulations impact (but in a rather haphazard manner) niche cell number, blood cell progenitor maintenance, plasmatocyte differentiation, crystal cell differentiation, DNA damage accumulation. Their results suggest that Gcn5 controls autophagy and show that reducing the expression of the autophagy machinery affect blood cell differentiation. By using drugs as well as genetic approaches to modulate the mTOR pathway, they conclude that Gcn5 levels are regulated by mTOR, but that the impact of this pathway on blood cell homeostasis can override Gcn5 function.

    Overall, the main conclusions are sound but interpreting several lines of experiments and results remain complicated. Consequently, the overall picture of the role of Gcn5 in Drosophila larval lymph gland development, and its relationship to mTOR and autophagy, remains unclear.

  3. eLife

    Reviewer #2 (Public review):

    Summary:

    Drosophila haematopoiesis has been shown to be governed by a number of signalling pathways such as JAK/STAT and Dpp. This important study shows a role for nutrient sensing and autophagy in determining blood cell differentiation. The authors show that General control non-derepressible 5 (Gcn5), a histone acetyltransferase affects blood cell differentiation. Gcn5 also negatively regulates autophagy through its effector TFEB which directly regulates autophagy genes. The authors also show that mTORC1 modulates Gcn5 levels and through it TFEB activity thus acting as a fine-tuning mechanism which maintains optimal levels of autophagy.

    Strengths:

    The main strength of the work lies in the interesting finding that cellular metabolic processes such as autophagy has a direct role in blood cell differentiation and has the potential to be of interest to those working on vertebrate haematopoiesis as well. The report has generated intriguing data, using promoters specific for sub sections of the lymph gland, that different cellular subsets of the lymph gland contribute differently towards haematopoiesis, but this is not followed up in detail and the final conclusions are derived from a combination of whole lymph gland perturbations as well as those from specific promoters.

    Weakness:

    (1) Gc5 seems to be expressed throughout the lymph gland but modulating it in the subsections do not have the same result. It is very striking that the knockdown of Gcn5 in the prohemocyte population does not have an effect on differentiation whereas overexpression does. And the modulations of Gcn5 in PSC also has variable effects across hemocyte subpopulations which is not explored in the manuscript. Interestingly, also the domain deletion constructs show differential effect on blood cell differentiation when altered solely in the prohemocytes which is not explained. While Gcn5 can be seen in all sections of the lymph gland in the first figure, under the HHLT-Gal4 and Hml-Gal4, Gcn5 looks cytoplasmic and almost completely excluded from the nucleus strikingly unlike Gcn5 expression under the Collier-Gal4 and Dome-Gal4. The rest of the experiments in the manuscript are done with multiple promoters, with autophagy flux measured by modulating Gcn5 with a pan hemocyte promoter, but the mTORC1-Gcn5 axis is explored using chemical modulators which affect the whole of the lymph gland (Fig7) or using two pro-hemocyte promoters (Fig8).

    (2) The knockdown of Gcn5 seems to affect the gland size (A compared to B and C). Since mTORC1 is a central regulator of cell size, it is possible that some of the effects seen in these knockdowns are potentially through mTORC1 affecting size suggesting that the signalling axis between mTORC1 and Gcn5 might not be a one-way axis as suggested in Figure 9. Also, this would mean that in experiments where absolute cell counts of crystal cells or niche cells are used to assess blood cell differentiation, further analysis to consider total cell numbers in the lymph gland would strengthen the manuscript.

    (3) A genetic manipulation of mTORC1 specifically in the pro hemocytes would strengthen the role of mTORC1 in the pathway rather than the chemical modulation which affects the whole of the lymph gland.

    Comments on the revised manuscript:

    Overall, the revisions make the narrative more coherent. The authors have also added data which substantiates their conclusions.

    However, in some instances, the authors are not clearly able to explain the discrepancies in the data (Gen-5 depletions under the Hml-Gal4 in the whole larval lysates remove p62 completely) which is not ideal.

    A query regarding the discrepancies in the immunofluorescence data: The authors have removed the IF data which suggested that there could be differences in the shuttling of Gcn5 between the nucleus and cytoplasm. The authors suggest that immunofluorescence issues are at the root of these variable results, but the reviewer wonders whether there could be further unexplored mechanisms re: shuttling that is unexplored here and would have been potentially novel.

  4. eLife

    Author response:

    The following is the authors’ response to the original reviews.

    Public Reviews:

    Reviewer #1 (Public Review):

    In their manuscript, Arjun et al. investigate the role of the histone acetyltransferase Gcn5 in the control of drosophila blood cell homeostasis in the larval lymph gland. They use gcn5 zygotic mutants as well as targeted knock-down and over-expression of Gcn5 in various lymph gland populations to show that these modulations impact (in a rather haphazard manner) niche cell number, blood cell progenitor maintenance, plasmatocyte differentiation, crystal cell differentiation or DNA damage accumulation. Their results suggest that Gcn5 controls autophagy and they show that decreasing the expression of the autophagy machinery increases blood cell differentiation. Using drugs to modulate the mTOR pathway, they conclude that Gcn5 levels are regulated by mTOR but that the impact of this pathway on blood cell homeostasis can override Gcn5 function.

    While the authors did a lot of experiments and good quantifications of the blood cell phenotypes, many results do not make much sense or do not bring valuable information about Gcn5 mode of action. Several conclusions of the manuscripts are not backed by solid data (e.g. that Gcn5 action is mediated by TFEB and the autophagy machinery) and different aspects of the literature are not well taken into consideration. Some results (such as the validation of the knockdown and overexpression of Gcn5) seem flawed. There are some concerns about the results obtained with gcn5 zygotic mutants and an interpretation of the phenotypes observed upon manipulation of Gcn5 expression in different cell types is missing.

    We have now performed several experiments to address the comments raised by the reviewer and have also provided possible explanation of the phenotypes in cases where it was lacking.

    Important revisions are needed to improve the quality of the manuscript and confirm the authors' findings.

    Reviewer #2 (Public Review):

    Summary:

    Drosophila hematopoiesis has been shown to be governed by a number of signaling pathways such as JAK/STAT and Dpp. This important study shows the role of nutrient sensing and autophagy in determining blood cell differentiation. The authors show that General control non-derepressible 5 (Gcn5), a histone acetyltransferase affects blood cell differentiation. Gcn5 also negatively regulates autophagy through its effector TFEB which directly regulates autophagy genes. The authors also show that mTORC1 modulates Gcn5 levels and through it, TFEB activity thus acting as a fine-tuning mechanism that maintains optimal levels of autophagy.

    Strengths:

    The main strength of the work lies in the interesting finding that cellular metabolic processes such as autophagy have a direct role in blood cell differentiation and has the potential to be of interest to those working on vertebrate haematopoiesis as well. The report has generated intriguing data, using promoters specific for sub-sections of the lymph gland, that different cellular subsets of the lymph gland contribute differently towards haematopoiesis, but this is not followed up in detail and the final conclusions are derived from a combination of whole lymph gland perturbations as well as those from specific promoters.

    Weaknesses:

    (1) Gc5 seems to be expressed throughout the lymph gland but modulating it in the subsections does not have the same result. It is very striking that the knockdown of Gcn5 in the prohemocyte population does not have an effect on differentiation whereas overexpression does. The modulations of Gcn5 in PSC also have variable effects across hemocyte subpopulations which is not explored in the manuscript.

    We have now explained and discuss why Gcn5 modulation could be affecting the PSC size. Please check Discussion section Paragraph 1 line 10 onwards.

    Interestingly, also the domain deletion constructs show a differential effect on blood cell differentiation when altered solely in the prohemocytes which is not explained.

    Currently, with our observations all that we can comment about that data is that expression of domain deletion mutants causes aberrant hematopoiesis indicating a dominant negative phenotype since they are expressed in the wild type genetic background. Beyond this, we will be exploring mechanistically how these domains are functioning during hematopoiesis in future studies. We have already described the dominant negative effect in the text: Discussion Section Paragraph 3.

    While Gcn5 can be seen in all sections of the lymph gland in the first figure, under the HHLT-Gal4 and Hml-Gal4, Gcn5 looks cytoplasmic and almost completely excluded from the nucleus strikingly unlike Gcn5 expression under the Collier-Gal4 and Dome-Gal4.

    We have now revised Figure 1 and have only included the images with Collier-Gal4 and Dome-Gal4 which clearly shows both the niche cells, Dome-positive progenitors and Dome-negative cells of the primary LG lobe essentially showing that Gcn5 is expressed throughout the primary LG lobe. In Fig. 1C-F’, Gcn5 expression is both in the nucleus and cytoplasm as this molecule shuttles between cytoplasm and nucleus. The staining pattern with the other Gal4 could be due to problems in the immunofluorescence protocol and acquisition parameters. We have now removed those images from Figure 1. Please check revised Figure 1.

    The rest of the experiments in the manuscript are done with multiple promoters, with autophagy flux measured by modulating Gcn5 with a pan hemocyte promoter, but the mTORC1-Gcn5 axis is explored using chemical modulators which affect the whole of the lymph gland (Fig7) or using two pro-hemocyte promoters (Fig8).

    We have used a pan-hemocyte promoter for the autophagy analysis to investigate if Gcn5 regulation over autophagy is a hemocyte specific effect which we indeed see. We have removed the western blot data now in the revised manuscript where we looked at Atg8 and p62 levels in whole larval lysates when Gcn5 was perturbed using hemocyte driver as the results were puzzling and difficult to comprehend given the complete absence of a p62 band in Gcn5 knockdown conditions. Also, it’s worth noting that Hml-Gal4 is also active in the LG hemocytes. We did 2 alternate promoters for prohemocytes to cross-validate some of our results and the chemical modulators experiment was done since effects like mTOR inhibition/nutrient sensing effects are systemic and hence such modalities were employed.

    (2) The knockdown of Gcn5 seems to affect the gland size (A compared to B and C). Since mTORC1 is a central regulator of cell size, it is possible that some of the effects seen in these knockdowns are potentially through mTORC1 affecting size suggesting that the signalling axis between mTORC1 and Gcn5 might not be a one-way axis as suggested in Figure 9. Also, this would mean that in experiments where absolute cell counts of crystal cells or niche cells are used to assess blood cell differentiation, further analysis to consider total cell numbers in the lymph gland would strengthen the manuscript.

    It is a possibility that Gcn5 perturbation could be affecting lymph gland size although we have not seen any consistent trend that would point towards this phenotype either upon knockdown or over-expression. We believe Gcn5 controls blood cell differentiation phenotypes strongly via mTORC1. But in order to answer reviewer’s comment we have now re-analyzed our crystal cell differentiation data particularly and quantitated it and represented it as crystal cell differentiation index for dome-Gal4 specific Gcn5 modulation and for the data with genetic modulation of mTORC1 pathway. Please see Fig 3P and S10J for the revised analysis.

    (3) A genetic manipulation of mTORC1 specifically in the pro hemocytes would strengthen the role of mTORC1 in the pathway rather than the chemical modulation which affects the whole of the lymph gland.

    We thank the reviewer for their useful critique. We have now addressed this concern and we have genetically perturbed the mTORC1 pathway in the progenitors using both abrogation of TORC1 via depletion of Tor or Raptor or by activation using over-expression of Rheb. We have now included this data as Supplementary figures – Fig S10 and S11 and have described it in the results section. Please see results section “Chemical or genetic modulation of mTORC1 activity controls blood cell differentiation” in the revised manuscript.

    Recommendations for the authors:

    Reviewer #1 (Recommendations For The Authors):

    The abstract could clearly be improved. It does not make a clear presentation of what is new in the manuscript. The conclusions that Gcn5 function in the lymph gland is mediated by the autophagy machinery and the acetylation of its non-histone target TFEB are not grounded and purely circumstantial. The implication of mTOR and nutrition in drosophila larval blood cell homeostasis has already been studied but not mentioned here. Most of the time the authors do not provide any possible explanation about the phenotypes they observe and how they fit with the current literature. Several pieces of results are of serious concern.

    We would like to thank the reviewer for their feedback. We have revised the abstract and have incorporated the insights obtained from our study. We have now included relevant literature that talks about the implication of mTOR and nutrition in Drosophila larval blood cell homeostasis (Please see Introduction section Paragraph 2 in the manuscript). We have also noted the input of the reviewer on many phenotypes lacking any description of a possible explanation. We have worked on results section to provide possible explanation and speculation wherever relevant.

    In the introduction, the authors do not provide an up-to-date and accurate presentation of the field. For example, they could use much more recent and comprehensive reviews since Evans et al. 2003. (eg. MID: 30733377 or 35887113). Their choice for signaling pathways involved in Drosophila blood cell progenitors seems very much biased for lead author self-citation rather than more directly related citations. It is surprising too that the authors failed to mention a series of publications on Akt/mTOR and nutrient sensing impact on drosophila larval blood cells (PMID: 22951642 ; 22911822 ; 22407365 ; 22510984). Along the same line, there are already several reports on autophagy genes implicated in Drosophila hematopoiesis and blood cell functions (PMID: 23406899; : 33560224 ; 20498061 ; 37623416). The introduction on GCN5 is a bit of a catalogue and should be streamlined- citing a recent review would be useful (PMID: 32735945). Again, the authors fail to cite publications showing that Gcn5 levels can be modulated by nutrition (PMID 27022023; 27874008) and they do not mention that amino acids starvation or mTOR inhibition leads to a decrease in GCN5 activity / TFEB acetylation (ref 40). Taking into account all the missing information, the novelty of the present manuscript is strongly decreased.

    We would like to thank the reviewer for the detailed suggestions on including the relevant literature that are appropriate and relevant to be mentioned in the context of the observations in our manuscript. We have now included these references and have cited them as per reviewer’s suggestions. Please check Introduction section paragraph 2.

    Results

    While there is little doubt that Gcn5 is expressed in the entire primary lobes based on Fig 1C-F, the quality of the staining in G-J (especially H, J) is really poor and essentially looks like non-specific background with no clear signal in the nuclei. Better images should be presented. The conclusion of the paragraph ("all cellular populations of the LG") and title of Fig.1 are not fully accurate as the authors do not provide evidence that Gcn5 is also expressed in posterior lobes.

    As per reviewer’s suggestions, since the images in Fig 1A-D’ clearly show that Gcn5 is expressed in the entire primary LG lobe in PSC cells, MZ and CZ; we have removed panels E-H’ which lacked clear nuclear signal. Fig1A-D’ clearly show the nuclear staining pattern of Gcn5. We have also modified the conclusion of the paragraph to say that Gcn5 is expressed in cellular populations of the primary lymph gland lobe accordingly.

    Concerning, Fig S1 and Fig 2, while the analysis seems technically sound, the results are puzzling. The lack of P1 differentiation in gcn5 null heterozygotes is very surprising. The authors should check that this stock does not carry a mutation in nimC1 (for details see: PMID: 23899817) and use other plasmatocyte differentiation markers to confirm their observation (also with the different allelic combinations). I'm also concerned by the levels of plasmatocyte differentiation and crystal cell number in the control line (notably in S1H), which seem very low (and quite variable for P1 as there is a notable difference between S1H and Fig 2H). Moreover, the analysis of the allelic combinations gives rather incoherent results: PCSC cell numbers are affected only in null/hypomorph, whereas differentiation (NimC1 and Hnt), as well as DNA damage, was only increased in hypomorph homozygotes. The authors propose no hypothesis to explain these observations.

    We have now repeated these experiments with the E333st null allele by placing it on a different balancer and we observe homozygotes that are alive till late third instar/early pupal stage as shown before by Carre et al., 2005. We have now included these revised results on the plasmatocyte differentiation status of the E333st heterozygotes and homozygotes (See Fig 2 and Fig S1). We do find P1 positive cells in the E333St heterozygotes unlike earlier. Plasmatocyte and crystal cell numbers in the control line always shows some level of heterogeneity. We have included the wild type control individually with those respective mutants during the experiment hence drawing a cross comparison across two different experiments would not be appropriate. We have now explained the observations obtained on PSC cell numbers (Discussion section paragraph 1). Experiments to check all hematopoietic aspects of the gcn5 null have been done after changing the balancer line and the null mutants overall show a decrease in PSC size and a widespread increase in hemocyte differentiation which could be due to a systemic effect due to various signalling pathways being affected which needs to be investigated and is beyond the scope of this study. This has also been discussed in the Discussion section Paragraph 1.

    Although a side-by-side comparison would have been better suited, it seems that the homozygotes or trans-heterozygotes do not have stronger phenotypes than the heterozygotes as far as crystal cell and DNA damage are concerned, which is rather unexpected. Besides the authors should introduce why they look at DNA damage.

    We agree with the reviewer that for the crystal cell and DNA damage phenotype the homozygotes or trans-heterozygotes do not have a stronger phenotype as compared to the heterozygotes alone but since these are whole animal mutants there could activation/inactivation of various signalling pathways and systemic effects that would be difficult to account for and comprehend here which needs to be investigated further. The only conclusion that we draw from these observations is that Gcn5 is required for maintaining blood cell homeostasis. Regarding DNA damage, we have now included the rationale and supporting literature for why we have studied DNA damage in the context of Gcn5. Please see result section 2 paragraph 1.

    Importantly too, the authors failed to obtain gcn5 E333st/E333st (null) larvae, whereas Carre et al. originally reported that E333st/E333st individuals are viable until the late third instar larvae. I suspect that the stock they use carries additional mutations that need to be eliminated by back-crossing it to control flies for several generations. Of note too, a recent report showed that a deletion of gcn5 (generated by CRISPR) does not prevent adult emergence, challenging the conclusion that gcn5 expression is absolutely required for fly development (PMID: 37545086).

    The reviewer is right in pointing out that E333st homozygotes survive until late third instar as reported by Carre et al.,2005. We have procured the null allele again and used another balancer to obtain homozygotes and we were able to get homozygotes that survived till late third instar as reported earlier. We have now included new data from these homozygotes for all hematopoietic aspects and heterozygotes particularly for plasmatocyte differentiation Please see Fig 2 and Fig S1 and corresponding results section 2 of the manuscript.

    Concerning the validation of Gcn5 knock-down and overexpression: the results are highly dubious. In Fig S2B (hml>Gcn5 RNAi), there is virtually no Gcn5 signal in the primary lobes but hml is normally expressed only in the cortical zone. How is it possible? Similarly, the western blot (which is really too much cropped around the bands of interest) does not show any signal in the hml>Gcn5 RNAi lane (not even some background. According to the Methods section, the western was performed on whole larvae extracts; hml-mediated knock-down can not wipe out its expression in all the tissues. As for the overexpression, flag immunostaining in hml>Gcn5-flag is mostly cytoplasmic (S2E), which doesn't make sense and does not fit with S2C (Gcn5 immunostaining).

    Hml-Gal4 is a pan hemocyte driver and its expression is not limited to the CZ of the primary lymph gland lobe (Banerjee et al., 2019) and recent single cell sequencing data corroborate this that Hml domain is not limited to the cortical zone (Yarikipati and Bergmann, 2026). GFP driven by Hml-Gal4 is spread out across the primary LG lobe which could explain the phenotype of no Gcn5 signal obtained in the immunofluorescence experiment. Regarding the western blotting experiment which was performed on whole larval extracts, we were also puzzled by lack of Gcn5 bands in these lysates upon depleting Gcn5 using Hml-Gal4. We need to systematically probe further to understand expression of Gcn5 in other tissues and organs. We have now removed the western blot data as the data obtained cannot be comprehended at the moment. Regarding the FLAG staining experiment – the staining gave us a cytoplasmic pattern and since Gcn5 is known to shuttle between the cytoplasm and nucleus it is possible that the anti-FLAG staining detected the Gcn5 localizing in the cytoplasm. It is difficult to draw a direct comparison here between the images S2C and S2E as both are different antibodies.

    The initial analysis of Gcn5 level modulation in the prohemocytes, PSC or Hml+ cells is mainly descriptive and the authors do not elaborate on possible explanations based on the current literature.

    We have added a possible explanation wherever required for these respective results on Gcn5 modulation in prohemocytes, PSC and Hml positive hemocytes. Please see result section 3 where we elaborate on possible explanation for the phenotypes observed.

    The structure/function analysis of Gcn5 is based on overexpression of truncated mutants in the prohemocytes using the tep4-GAL4 driver and monitoring PSC cell, prohemocyte maintenance, plasmatocyte and crystal cell differentiation as well as DNA damage. As the overexpression of the full-length protein was made with a different driver (Dome), it is difficult to interpret the data. Nevertheless, no clear message emerges from this analysis and the authors do not reach any conclusion. Thus, the interest of these experiments remains limited.

    The structure-function analysis was largely done to understand which of the domains of Gcn5 upon over-expression results in a dominant negative like phenotype and our analysis shows that expression of some of these domain mutants results in a dominant negative phenotype in the wild type genetic background which we have now stressed upon in the text. However, further mechanistic understanding and in-depth analysis of each of these domains of Gcn5 warrants further separate investigation and is beyond the scope of this study. Please see the end of result section 4 for conclusion and possible explanation.

    The authors then analyze autophagy markers (in hml>Gcn5 LOF or GOF). Contrary to their say, hml-GAL4 is not a pan-hemocyte marker. It would have been interesting to ensure that the effects observed on Atg8 and Ref(2)P in the lymph gland are cell-autonomous- as expected for a direct role of Gcn5 on this pathway. Again, it is very surprising that p62 is not detected in the western blot on whole larval extracts when Gcn5 is knocked down in Hml+ cells only (Fig 5D). Moreover, quantifications on multiple samples will be needed to validate the increase/decrease of p62 and Atg8 as detected by western blot. As for the RT-qPCR (Fig S5), according to the Methods sections, they were made on adult blood cells but this is not explicit in the result section.

    We have corrected the text and mentioned Hml-Gal4 as a hemocyte specific Gal4 shown earlier as Gal4 marking both embryonic and larval hemocyte population (Goto et al., 2003, Yarikipati and Bergmann, 2026). Regarding the Atg8 and Ref (2)P blots – yes, it is surprising to us too that the p62 is not detected in the larval lysates when Gcn5 is depleted using Hml-Gal4. However, this result was consistent over the replicates performed and needs to be further studied. Since this phenotype of complete absence of p62 in larval lysates upon Gcn5 depletion cannot be comprehended and explained, we have removed the western blot data from the figure and have just retained the immunofluorescence data and have also quantified the Atg8 and p62 puncta per cell and included this data in Figure 5, Graphs D and E. For the qRT-PCR we have now included a description in the corresponding results section. Please see result section – result 5 under “Autophagic flux in the Drosophila blood cells is negatively regulated by Gcn5”.

    The knock-down of TFEB or several autophagy genes in the prohemocytes (tep4-GAL4) leads to a rather convincing increase in plasmatocyte and crystal cell differentiation. It would have been interesting though to quantify prohemocyte maintenance, PSC cell number, and DNA damage. Also, the authors should have performed Gcn5 GOF/LOF experiments with the same driver (they present tep>Gcn5 RNAi in Fig 8 but without the proper controls).

    We have now included data for prohemocyte index (Figure S8M) upon knockdown of TFEB and other autophagy genes along with PSC cell number, DNA damage (Supple Fig S8) in the revised manuscript. Please see corresponding results section titled “Genetic and chemical ablation of autophagy boosts blood cell differentiation in the primary lymph gland lobe” for the description of the results.

    The use of chloroquine should be better described. How long was the treatment? Did the authors observe an effect on autophagy in the lymph gland? Chrorloquine also affects lysosomal pH, so it remains to be demonstrated that the effects observed here are only autophagy-related.

    We have now written a detailed protocol for the treatment in the methods section and also mentioned the treatment time which is 16 hours in the results. We have included data to validate the effect of Chloroquine on autophagy by p62 and Atg8 staining in the LG and have quantitated the data (Refer Supple Fig S9) and the corresponding results section titled “Genetic and chemical ablation of autophagy boosts blood cell differentiation in the primary lymph gland lobe”

    Similarly, the use of drugs to activate (3BDO) or inhibit (Rapamycin) mTOR should be better controlled. More generally, given the promiscuous roles of mTOR (and autophagy) in the larvae, tissue-specific manipulations would be better suited.

    We have now perturbed mTOR pathway genetically by activation and in-activation and have studied the effect on blood cell differentiation. Please see Figure S10 and the corresponding result section titled “Chemical or genetic modulation of mTORC1 activity controls blood cell differentiation” where we discuss the results of genetic perturbation of mTOR pathway.

    Actually, as pointed out above, it has already been shown that modulation of Akt/TOR in hemocytes or amino-acid deprivation affects blood cell homeostasis (see above). The authors should definitely discuss how their results fit with the literature on this subject.

    We have added relevant literature in the introduction section and have also discussed how Gcn5 could fit into this context of nutritional sensing and control of hematopoiesis. Please check revised Introduction section paragraph 2. Also, check discussion section in last paragraph where we have discussed role of Gcn5 in nutrient sensing.

    Again, Gcn5 levels need to be quantified using multiple samples (Fig 7M, N) before concluding.

    Sorry for not including the quantitation earlier but we have now included the quantitation for the blots presented in Fig. 7 M and N.

    Finally, the authors show that 3BDO still induces an increase in blood cell differentiation when gcn5 is knocked-down in tep4+ cells and that Rapamycin still represses differentiation when Gcn5 is overexpressed in Dome+ cells. They conclude that mTORC1 overrides the effect of Gcn5. This seems a far-reaching conclusion given the available evidence.

    We have now toned down the conclusion that we make to accommodate other possibilities which we have been unable to test here currently.

    In particular, in the conditions used, the authors do not necessarily assess the activity/requirement for Gcn5 and mTORC1 in the same cell population.

    Other comments and suggestions:

    The discovery of the SAGA complex is not Grant 1999 but 1997 (PMID: 9224714).

    Ref 30 is not appropriate -nothing to do with HAT.

    GCN5 not only acetylates TFEB but also Atg7 (PMID: 28594263) to limit autophagy.

    Thank you so much for these suggestions. We have made the necessary amendments in the references.

    In the results section, the first paragraph is largely a repetition of the introduction. The same is true for most paragraphs in this section. A shorter (hypothesis-driven) introductory sentence would be more adequate.

    We have now taken the suggestion into consideration and made the necessary change in the results section throughout the manuscript.

    Fig 1: it seems that there is a higher accumulation of Gcn5 in a few cells in the cortical zone. This may correspond to crystal cells and could be easily confirmed.

    We have now checked this aspect. Please see supple fig S5 where we co-stain lozenge-GFP cells containing LG with Gcn5 to check for the accumulation. However, we do not see any accumulation in the Lozenge-positive crystal cells.

    Figure 3: the authors should also quantify the proportion of progenitors (dome>GFP+) in the different conditions.

    We have now done this and added it to the Figure. Please see panel N in Figure 3 and Figure S8M.

    Figure S3: how do the authors explain that Gcn5 knockdown in the PSC reduces plasmatocytes differentiation (but does not affect PSC cell number or crystal cell differentiation)? What could be the origin of the increase in DNA damage (essentially in CZ)? How do they explain that Gcn5 over-expression increases PSC size but does not affect (reduce?) blood cell differentiation?

    These observations need to be investigated further. We currently have no answer to these comments. The signals that are produced by the PSC could be affected due to which we observe these phenotypes like an effect on plasmatocyte differentiation and an increase in DNA damage whereas no effect on PSC cell numbers or crystal cell numbers which needs to be studied further. Also, in the case of Gcn5 over-expression in PSC we do not know how the increased size of PSC controls differentiation. This would need further experimentation and since this paper is not about the role of Gcn5 in PSC exclusively, we will look into this in our future studies. These aspects will be studied in our future follow-up studies as it is beyond the scope of the current manuscript.

    Figure S4: how do the authors explain the non-cell autonomous increase in PSC cell number upon Gcn5 KD/GOF in hml+ cells? How do they explain the increase in crystal cell number in Gcn5 GOF? Is it really cell-autonomous (i.e. all the Hnt+ cells are Hml+?)?

    We have discussed how Gcn5 depletion or over-expression in HmlΔ cells could affect PSC cell numbers. Please see discussion section, paragraph 1. Regarding the crystal cell phenotype - We have now tested if the increase in crystal cell numbers is cell autonomous by driving Gcn5 over-expression using a crystal cell specific driver and we find that the increase is cell-autonomous. Please refer to Supple Fig S5.

    The discussion is lengthy and should be reduced. It does not appropriately consider the current literature.

    We have tried to reduce the length of the discussion and have also added relevant references as per recommendations of the reviewer.

    Reviewer #2 (Recommendations For The Authors):

    (1) In general, it is not clear why in some of the experiments Tep-Gal4 is used to modulate proteins in prohemocytes while in others Dome-Gal4 is used.

    There is no particular reason. These Gal4’s have been used interchangeably as both label the hematopoietic progenitor population. Although recent single cell sequencing data has identified subsets within the progenitors namely core progenitors marked by tep4 largely and dome being a distal progenitor marker (Cho et al.,2020, Girard et al.,2021), in our study perturbations in Gcn5 using either of the Gal4’s results in a similar phenotype.

    (2) Considering alteration in lymph gland size (Figure 2), the number of positive cells should be analysed in relation to total cell numbers or s4ize.

    Although we do not find any visible differences or defects in the overall LG size in various genetic conditions discussed in this manuscript, we have done so for the plasmatocyte differentiation where we have represented it as plasmatocyte differentiation index (relative to the size of primary LG lobe) throughout the manuscript. We have now done this for crystal cell numbers too for critical genotypes in this manuscript and have represented it is as crystal cell index for example please see Figure 2O, 3P, S5G, S10J where these graphs have now been added.

    (3) Figure 1A G-I' does not look like mCD8 GFP expression, but rather cytoplasmic GFP.

    We have made the change in the figure and the corresponding text accordingly.

    (4) One of the main conclusions in the manuscript is that Gcn5 affects autophagy (Figure 5). Here, the puncta need to be quantified (relative to total cell numbers).

    Thank you for the suggestion. We have now quantitated the p62 and Atg8 positive puncta per cell and have represented it as panel D and E in Figure 5.

    (5) Figure 5 D and E show p62 and Atg8 total protein levels in the larvae when Gcn5 is modulated only in the hemocytes. It is surprising that there is a complete reduction in p62 levels across the whole larvae when Hml gal4 is used for the knockdown.

    Yes, we observe a complete absence of p62 in whole larval lysates when Gcn5 is depleted using Hml-Gal4 and we see this across replicates. This result is indeed puzzling to us and difficult to comprehend as to why a hemocyte specific driver would result in such a dramatic change hence we have decided to remove the western blot data as it is difficult to draw a solid conclusion from. We have retained the immunofluorescence data which shows a consistent alteration in autophagy upon Gcn5 perturbation using Hml-Gal4 and we have now included the quantification for the number of p62 and Atg8 positive puncta per cell for the IF data.

    (6) The beta-actin levels in the western blots in Figure 5 are highly oversaturated and do not represent loading control adequately. Also, it looks like there is substantially more total protein in 5D 3rd lane where Gcn5 is overexpressed.

    Thank you for pointing this out. We have loaded equal amount of protein in all the wells so we are unsure why the actin bands look over-saturated. We have now removed the western blot data from this figure as the data is puzzling and difficult to comprehend given a total absence of p62 in whole larval lysates in Gcn5 depletion conditions using Hml-Gal4. Hence, we are just retaining the immunofluorescence data.

  5. Version published to 10.7554/elife.91615.2 on eLife
  6. Version published to 10.7554/elife.91615 on eLife
  7. Version published to 10.7554/elife.91615.1 on eLife
  8. eLife

    eLife assessment

    This manuscript shows that manipulating the expression of the histone acetyltransferase Gcn5 affects blood cell homeostasis in the Drosophila larval hematopoietic organ. The data suggest a link between autophagy and the mTOR pathway, as could be expected from the literature. The authors use several genetic manipulations as well as some chemical modulators to generate solid evidence supporting most of their conclusions, but some of the analyses are inadequate and would benefit from improvement.

  9. eLife

    Reviewer #1 (Public Review):

    In their manuscript, Arjun et al. investigate the role of the histone acetyltransferase Gcn5 in the control of drosophila blood cell homeostasis in the larval lymph gland. They use gcn5 zygotic mutants as well as targeted knock-down and over-expression of Gcn5 in various lymph gland populations to show that these modulations impact (in a rather haphazard manner) niche cell number, blood cell progenitor maintenance, plasmatocyte differentiation, crystal cell differentiation or DNA damage accumulation. Their results suggest that Gcn5 controls autophagy and they show that decreasing the expression of the autophagy machinery increases blood cell differentiation. Using drugs to modulate the mTOR pathway, they conclude that Gcn5 levels are regulated by mTOR but that the impact of this pathway on blood cell homeostasis can override Gcn5 function.

    While the authors did a lot of experiments and good quantifications of the blood cell phenotypes, many results do not make much sense or do not bring valuable information about Gcn5 mode of action. Several conclusions of the manuscripts are not backed by solid data (e.g. that Gcn5 action is mediated by TFEB and the autophagy machinery) and different aspects of the literature are not well taken into consideration. Some results (such as the validation of the knockdown and overexpression of Gcn5) seem flawed. There are some concerns about the results obtained with gcn5 zygotic mutants and an interpretation of the phenotypes observed upon manipulation of Gcn5 expression in different cell types is missing.

    Important revisions are needed to improve the quality of the manuscript and confirm the authors' findings.

  10. eLife

    Reviewer #2 (Public Review):

    Summary:
    Drosophila hematopoiesis has been shown to be governed by a number of signaling pathways such as JAK/STAT and Dpp. This important study shows the role of nutrient sensing and autophagy in determining blood cell differentiation. The authors show that General control non-derepressible 5 (Gcn5), a histone acetyltransferase affects blood cell differentiation. Gcn5 also negatively regulates autophagy through its effector TFEB which directly regulates autophagy genes. The authors also show that mTORC1 modulates Gcn5 levels and through it, TFEB activity thus acting as a fine-tuning mechanism that maintains optimal levels of autophagy.

    Strengths:
    The main strength of the work lies in the interesting finding that cellular metabolic processes such as autophagy have a direct role in blood cell differentiation and has the potential to be of interest to those working on vertebrate haematopoiesis as well. The report has generated intriguing data, using promoters specific for sub-sections of the lymph gland, that different cellular subsets of the lymph gland contribute differently towards haematopoiesis, but this is not followed up in detail and the final conclusions are derived from a combination of whole lymph gland perturbations as well as those from specific promoters.

    Weaknesses:
    1. Gc5 seems to be expressed throughout the lymph gland but modulating it in the subsections does not have the same result. It is very striking that the knockdown of Gcn5 in the prohemocyte population does not have an effect on differentiation whereas overexpression does. The modulations of Gcn5 in PSC also have variable effects across hemocyte subpopulations which is not explored in the manuscript. Interestingly, also the domain deletion constructs show a differential effect on blood cell differentiation when altered solely in the prohemocytes which is not explained. While Gcn5 can be seen in all sections of the lymph gland in the first figure, under the HHLT-Gal4 and Hml-Gal4, Gcn5 looks cytoplasmic and almost completely excluded from the nucleus strikingly unlike Gcn5 expression under the Collier-Gal4 and Dome-Gal4. The rest of the experiments in the manuscript are done with multiple promoters, with autophagy flux measured by modulating Gcn5 with a pan hemocyte promoter, but the mTORC1-Gcn5 axis is explored using chemical modulators which affect the whole of the lymph gland (Fig7) or using two pro-hemocyte promoters (Fig8).

    2. The knockdown of Gcn5 seems to affect the gland size (A compared to B and C). Since mTORC1 is a central regulator of cell size, it is possible that some of the effects seen in these knockdowns are potentially through mTORC1 affecting size suggesting that the signalling axis between mTORC1 and Gcn5 might not be a one-way axis as suggested in Figure 9. Also, this would mean that in experiments where absolute cell counts of crystal cells or niche cells are used to assess blood cell differentiation, further analysis to consider total cell numbers in the lymph gland would strengthen the manuscript.

    3. A genetic manipulation of mTORC1 specifically in the pro hemocytes would strengthen the role of mTORC1 in the pathway rather than the chemical modulation which affects the whole of the lymph gland.

  11. Version published to 10.1101/2023.07.04.547637 on bioRxiv