Sexually dimorphic ATF4 expression in the fat confers female stress tolerance in Drosophila melanogaster

  1. Lydia Grmai
  2. Melissa Mychalczuk
  3. Aditya Arkalgud
  4. Deepika Vasudevan

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Metabolic differences between males and females have been well documented across many species. However, the molecular basis of these differences and how they impact tolerance to nutrient deprivation is still under investigation. In this work, we use Drosophila melanogaster to demonstrate that sex-specific differences in fat tissue metabolism are driven, in part, by dimorphic expression of the Integrated Stress Response (ISR) transcription factor, ATF4. We found that female fat tissues have higher ATF4 activity than their male counter parts under homeostatic conditions. This dimorphism was partly due to a female bias in transcript abundance of specific ATF4 splice isoforms. We found that the canonical sex determinants transformer ( tra ) and doublesex ( dsx ) drive such dimorphic ATF4 transcript abundance. These differences persist in a genetic model of nutrient deprivation, where female animals showed greater resistance to lethality than males in an ATF4-dependent manner. These results suggest that higher ATF4 activity confers higher tolerance to stress in females. Together, our data describe a previously unknown facet of ISR signaling wherein sexual identity of adipose tissue confers differential stress tolerance in males and females. Since energy storage mechanisms are known to be dimorphic and have been linked to ATF4 regulation, our studies provide a mechanistic starting point for understanding how sexual identity influences metabolic disease outcomes.

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

    Referee #1 Major concerns:

    1. One major concern that I have about the sexual dimorphism in tolerance to nutrient deprivation is that the papers cited by the authors, and paradigms that are used broadly in the field, all use adult flies. The authors must show that in larvae, a completely different life stage from their citations, there is a sexual dimorphism in tolerance to nutrient deprivation.

    In our descriptions of previous literature that describes tolerance to nutrient deprivation, we have added language that specifies that the results from nutrient deprivation mentioned therein were performed in adults (lines 82, 91, 96, highlighted in the preliminary revision).

    In response to the concern from this reviewer that our data do not assay for nutrient deprivation in larvae, we would like to clarify that our “stress tolerance assay” more specifically demonstrates that developmental nutrient deprivation compromises male survival through pupariation to adulthood. While the effects of acute nutrient deprivation on developmental delay can be assayed in foraging or earlier larval stages, we have not tested whether ATF4 signaling is present and dimorphic in these stages and believe it to be beyond the scope of this study. In the revision, we will edit the text to be more precise in our conclusions with these data.

    Interestingly, Diaz et al 2023 (Genetics) show that male larvae have greater fat stores than female larvae. Considering fat is the main determinant of tolerance to nutrient deprivation it's not clear that females will actually survive nutrient deprivation longer as larvae. This is an essential test of whether female larvae do have increased tolerance to nutrient deprivation, which is the basic foundation of the authors' model.

    We thank the reviewer for making this clarifying point about the relationship between fat stores and nutrient deprivation. ____In response to the concern our data do not assay for nutrient deprivation in larvae (major point #1), we would like to clarify that our “stress tolerance assay” more specifically demonstrates that developmental nutrient deprivation compromises male survival through pupariation to adulthood. While the effects of acute nutrient deprivation on developmental delay can be assayed in foraging or earlier larval stages, we have not tested whether ATF4 signaling is present and dimorphic in these stages and believe it to be beyond the scope of this study. In the revision, we will edit the text to be more precise in our conclusions with these data.

    1. Another concern is the way that the authors "genetically induce nutrient deprivation using methioninase overexpression". As they acknowledge in the discussion (Line 381-390), methioninase overexpression will have many cellular effects. While there is no doubt that methionine levels would be lower in their model, it is less certain whether this is the main driver of the male-specific lethality.

    There are two potential solutions to this problem. First, the authors could change the text throughout the paper to more accurately describe their paradigm as "methioninase-induced lethality" rather than "nutrient deprivation". This would limit the scope of their scientific question and the conclusions they draw, but would eliminate the need for more experiments.

    The second solution would be to complete experiments to establish the following points: i) methioninase overexpression causes all the classical features of nutrient deprivation (e.g. changes to canonical signaling pathways such as TOR); ii) using other genetic means of nutrient deprivation such as slimfast-RNAi to see if those manipulations phenocopies the male-specific lethality they see with methioninase overexpression; iii) testing a role for ATF4 in mediating sex differences (if any) in other contexts such as slimfast-RNAi. This will take 2-3 months but is essential to draw any conclusions about whether their paradigm is truly a model for nutrient deprivation.

    We agree that methionine depletion is not the only cellular change effected by methioninase over-expression. For example, a molecular byproduct of methioninase metabolism via methioninase is the production of ammonia, which has recently been shown to indue ISR signaling in the context of____ alcohol-associated liver disease (Song et al. 2024, PMID 37995805). We believe our experimental controls and genetic rescues account for this and other possible effects in the interpretation of our data. ____To further establish the utility of methioninase overexpression as a genetic means of methionine deprivation (first described in Parkhitko et al. 2021, PMID 34588310), we will perform ____slimfastRNAi____ in the fat (another genetic means of reducing intracellular amino acid levels) per the reviewer’s suggestion. In these animals we will evaluate 1) ATF4 activity in L3 adipocytes using 4E-BPintron-GFP (1.5 months) , and 2) male vs. female lethality (as determined by counting eclosed adults) (2 months. If male lethality is observed with ____*UAS-slimfastRNAi ____as with ____methioninase *____expression, we will test the requirement for dimorphic ATF4 signaling in the fat for such male susceptibility to lethality/female resistance to lethality. (3 months)

    1. Another important point is that the authors state that sexually dimorphic ATF4 activity in the fat body is instructed by sexual identity in a cell-autonomous manner. Despite a clear decrease in ATF4 reporter levels in tra mutants, the fat body-specific tra-RNAi effect on the ATF4 reporter was less convincing. Together with the fact that changes to tra in the fat body affect insulin secretion from the insulin-producing cells, it is possible that the effect on ATF4 is not cell-autonomous. To conclusively test if sexual identity regulates ATF4 in a cell-autonomous manner the authors should use the flp-out system to make Tra-expressing or tra-RNAi-expressing clones in the fat body. This would take approximately 1.5 months to make the strain and test this.

    We thank the reviewer for making the astute observation that the effect of fat body-specific ____tra____ knockdown on female ATF4 reporter activity was more modest than whole-animal ____tra____ mutants. We ascribe this to RNAi knockdown efficiency rather than non-autonomous effects of sexual identity on ATF4 expression in the fat. This is underscored by our data showing fat body knockdown of ____spenito____ (____nito____), a sex determinant upstream of ____tra____ that is shown to instruct female sexual identity in the larval fat (Diaz et al. 2023, PMID 36824729), does indeed reduce ATF4 levels in female fat to that of control male fat (Fig. 2K).

    1. As the authors show for the UAS-methioninase, other UAS lines used in the paper such as UAS-traF, UAS-tra-RNAi, UAS-dsx-RNAi may have leaky effects on gene/reporter expression. The authors must include a UAS only control to establish that the tra-RNAi, UAS-traF, UAS-dsx-RNAi do not affect gene/reporter expression.

    We thank the reviewer for suggesting that we evaluate the “leakiness” of all UAS lines used in this study (major point #4). To do this, we will quantify ATF4 reporter activity in the fat (4E-BPintron-GFP) in the presence of UAS lines but in the absence of GAL4 for ____UAS-traRNAi____, ____UAS-traF____, and ____UAS-dsxRNAi____ (1.5 months)

    1. I have concerns about the statistics used. In the methods and legends only t-tests are mentioned; however, when three groups are compared a one-way ANOVA with post-hoc tests must be used to correct for multiple comparisons. To compare differential responses to genetic/environmental manipulations between the sexes, a two-way ANOVA must be used. For example, to conclude that males and females have different responses in the two-way ANOVA, there must be a significant genotype:sex interaction. The p-values for comparisons between genotypes in either the one-way or two-way ANOVA must be derived from post-hoc tests within the ANOVA analysis.

    __We thank the reviewer for carefully assessing our usage of statistical analyses to interpret the data in the study. To the best of our understanding, such ANOVA analyses are helpful in evaluating significance when comparing multiple sample groups simultaneously. However, in all our analyses we are only ever comparing two samples at a time, making a two-tailed Student’s t-test with Welch’s correction (assuming unequal variance) to be the best statistical method. __

    Referee #1 Minor points

    1. Please ensure to make the reader aware of which life stage was tested in the literature cited supporting sexually dimorphic tolerance to nutrient deprivation.

    We thank the reviewer for pointing out this ambiguity in our description of previous and current work on nutrient deprivation tolerance. We address this minor point in tandem with major point #1 above ____by adding language that specifies that the results from nutrient deprivation mentioned therein were performed in adults (lines 82, 91, 96, highlighted in the preliminary revision).

    1. Published data about sex-specific mechanisms of metabolic regulation mean that the introduction should be more fully cited than it is. Even in the introduction "the molecular basis of these differences and how they impact tolerance to nutrient deprivation is still under investigation" is inaccurate, as there are published studies identifying some mechanisms (work on gut hormones and sex-specific effects on starvation resistance and body fat, role of ecdysone on body fat and feeding, sex-specific roles for brummer and Akh in regulating body fat, intestinal transit and gut size and feeding). Please adjust the paper to acknowledge this growing body of knowledge.

    __We thank the reviewer for appropriately highlighting that there are other relevant studies in the context of sex-specific mechanisms of metabolic regulation in addition to those referenced in the original manuscript. Specifically, we will include additional citations and appropriate descriptions of previous work, such as those that report on sex-specific effects of starvation (i.e. Millington et al. 2022, PMID 35195254) and sex-specific roles for metabolic regulators such as Brummer/ATGL (Wat et al. 2020, PMID 31961851) and Adipokinetic hormone (Wat et al. 2021, PMID 34672260) in ____Drosophila____ fat storage.____ __

    1. Please list the ingredients per L so that individuals can replicate the diet easily.

    __We thank the reviewer for requesting additional details on the diet fed to animals in this study, which will improve the reproducibility of our findings. In the Methods section, we have now included additional details on the specific diet fed to animals used in this study (lines 465-468 in the preliminary revision).____ __

    1. Please cite grant numbers for all the community resources (e.g. Bloomington, DSHB), and please acknowledge FlyBase and its grants as well. For example, here are the instructions for citing BDSC https://bdsc.indiana.edu/about/acknowledge.html and similar instructions are available for the other resources.

    We thank the reviewer for underscoring the importance of citing grant numbers for all community resources used. We have added to the Acknowledgements section statements and grant numbers regarding use of community resources such as FlyBase, Bloomington Drosophila Stock Center, and DSHB (lines 533-538 in the preliminary revision).

    Referee #2:

    Figure 4 is an important part of this study, where the authors show a male-specific vulnerability to methioninase expression. They show that ATF4 RNAi confers vulnerability to methioninase expression even in females. An obvious question is whether ATF4 overexpression is sufficient to enhance resistance to methionine deprivation in males.

    We thank the reviewer for pointing out that the ability of increased ATF4 in male fat to enhance resistance to methionine deprivation was not interrogated. To examine this, ____we will quantify survival rates of males and females following dual over-expression of methioninase and ATF4 (3 months). We would like to state here that experimental over-expression of ATF4 at the levels induced by GAL4 activity is sometimes lethal, so this experiment may be difficult to execute/interpret due to technical limitations.

    Methioninase expression results (Figure 4) are interesting. Are the levels of methioninase expression similar between males and females?

    We thank the reviewer for asking for clarification on whether methioninase induction is similar between males and females. Whether methioninase induction is sexually dimorphic is likely a function of whether there is sexual dimorphism in the strength of the GAL4 driver used. While the drivers employed in this study are widely used for fat body expression, to our knowledge relative expression of ____Dcg-GAL4____ in males versus females has never been reported. Thus, we will perform qPCR to compare GAL4 and methioninase transcript levels in ____Dcg-GAL4; UAS-methioninase____ male and female fat bodies (1 month).

    1. This manuscript focuses on ATF4, but there could be additional possible reasons for the sexually dimorphic ISR activity. For example, the degree of physiological stress that activates ISR could be different between males and females. I suggest comparing the levels of Phospho-eIF2alpha (or any other markers upstream of ATF4) in both sexes.

    We thank the reviewer for suggesting additional checks for sexual dimorphism in ISR activity in the fat, such as degree of eIF2α phosphorylation, which is directly upstream of ATF4 induction. Per their suggestion, we will compare p-eIF2α staining in male and female larval adipocytes (1.5 months).

    In Figures 1 to 3, the authors examine the intensity of ATF4 signaling after perturbing the sexual determination pathway. The methioninase experiments in Figure 4 are interesting, but there is nothing in this Figure linking male-specific vulnerability to sex determination genes. Examining the vulnerability to methioninase expression after perturbing the sexual determination genes would make Figure 4 integrate better with the rest of the manuscript.

    We thank the reviewer for highlighting that the role of male sexual identity in vulnerability of males to methioninase expression was not interrogated. Similar to our genetic interaction study proposed in point #1 from this reviewer, we will test whether feminizing male fat bodies (using UAS-traF over-expression) will change survival rate of males in our methioninase-expression paradigm (3 months).

    1. The authors write that they generated 4EBP intron-GFP because the 4EBP intron-DsRed signal was frequently observed in the cytoplasm (line 122). They seem to suggest that the DsRed reporter is less reliable than the GFP reporter. However, they continue to mix results using 4EBP intron-GFP (Fig. 4A) and 4EBP intron-DsRed (Fig. 4F). The two figures examine slightly different conditions (Fig 4A shows tra1 KO females, while Fig. 4F shows traF males). If the DsRed reporter is less reliable due to the signal from the cytoplasm, the authors should show results with the GFP reporter in traF males.

    We thank the reviewer for raising the legitimate concern that the ____4EBPintron-DsRed____ reporter used for some of the included quantifications in Fig. 3 might be less reliable then ____4EBPintron-GFP____ that was generated for this study. We have updated the manuscript text (in the Results section) to more accurately describe the justification for building the ____4EBPintron-GFP____ line (lines 122-127 in the preliminary revision).

    1. In Figure S1, the authors label 4EBP intron-GFP as Thor2p-GFP, which is confusing. There are other parts in the methods section referring to Thor2p. I suggest using consistent terminology throughout the manuscript.

    We thank the reviewer for pointing out this typo. We have modified the text accordingly in Figure S1.

    Referee #3 Major concerns:

    1. Sexually dimorphic ATF4 activity (Figure 1 and associated supplemental figure) as evidenced by reporter expression is the basis of this study, yet a detailed description of the immunofluorescence quantification is lacking. The methods sections needs to include information on how a) images were acquired (Were the same acquisition settings used across all images?), b) the intensity measurements were taken (What software was used? Does each data point in the distribution represent a single nucleus (the assumption is yes)? Is nuclear size adjusted for? Panels A' and B' have obvious differences in nuclear size which would in turn affect total intensity measurements), c) the sample size (How many fat images taken per animal per sample/genotype? How many trials were performed?)

    We thank the reviewer for requesting additional information describing the immunofluorescence quantification methods. ____We have now added an additional paragraph to the Methods section detailing image acquisition for quantifying reporter activity (lines 483-494 in the preliminary revision).

    1. While the authors nicely address the lack in specificity for two of the Gal4 driver lines used in the study limitation section, the fact that the one driver that is fat body-specific, 3.1Lsp2-Gal4, shows a modest, not statistically significant decrease in Figure 4C still raises some concern. There is another Lsp2-Gal4 line described in Lazareva et al., 2007 (PLoS Genetics) that drives expression in larval fat, perhaps to combat the issue of 3.1Lsp2-Gal4 have low activity, as mentioned by the authors. Alternatively, this phenotype could be assessed using Gal4 lines that only drive expression in the other tissues (if available). Otherwise, the conclusion that ISR/ATF4 signaling specifically in the fat mediates the starvation response needs to be toned down.

    We thank the reviewer for carefully analyzing our data showing survival during methioninase over-expression using different GAL4 drivers. ____The reviewer raises a valid concern that the GAL4 driver with highest specificity for the fat body (that is, with the least off-target tissue expression), ____3.1 Lsp2-GAL4____, induces the most modest methioninase-induced lethality (major point #2). We attribute this to the fact that ____3.1 Lsp2-GAL4____ is reportedly (and in our hands) a weaker driver than ____Dcg-GAL4____ in the larval fat body. We will demonstrate this experimentally by performing UAS-nucGFP expression using both ____Dcg-GAL4____ and ____3.1 Lsp2-GAL4____ side by side and quantifying nuclear GFP intensity in the larval fat (2 months).

    The reviewer also mentions that the other drivers with more statistically significant effects on male lethality (____Dcg-GAL4____ and ____r4-GAL4____, Fig. 4) are not restricted to the fat body. Importantly, both these drivers are also expressed in the blood lineage (hemocytes). To examine whether ISR activation in hemocytes contributes to the female stress tolerance (and/or male lethality) observed upon methioninase induction, we will quantify male and female survival rate following methioninase induction in the blood lineage using a blood-specific driver, ____HHLT-GAL4____ (Mondal et al 2014, PMID 25201876). (2.5 months)

    1. Several analyses rely on RNAi, and this is understandably important for tissue-specific knockdown of gene expression. At least one of the two following issues needs to be addressed: a) the efficiency of knockdown for each gene are not provided or reported on and b) only single RNAi lines were used for each gene targeted for knockdown.

    We thank the reviewer for pointing out that the original manuscript does not report on knockdown efficiencies of the RNAi lines used in the study. The RNAi lines from the Harvard Transgenic RNAi Project (TRiP) collection (traRNAi, dsxRNAi, nitoRNAi) have been verified in Yan & Perrimon 2015 (PMID 26324914). The ATF4RNAi line was verified in Grmai et al. 2024 (PMID 38457339). We have included all citations for these validation studies in Table S1 in the preliminary revision.

    Referee #3 Moderate concerns:

    1. Lines 137-141: It would be nice to see a gel that confirms that these newly designed primers detect the expected isoforms (supplemental perhaps).

    We thank the reviewer for requesting confirmation of isoform specificity of the primers used to detect ATF4 transcript in the fat body in Fig. 2B-C. Because these are qPCR primers, they were all designed to produce amplicons of nearly equal size. There is currently no reliable method to specifically deplete one ____ATF4____ isoform at a time, which would be the only way to experimentally demonstrate isoform specificity of each primer set. However, we have designed each primer pair to specifically detect isoform-specific regions of ____ATF4____ mRNA and have verified specificity (and lack of off-target products in the ____D. melanogaster____ genome) ____in silico____ using Primer-BLAST (NCBI).

    1. Lines 278-282 and Figure 4D: Shouldn't the second and fourth bars be compared? Based on the hypothesis and conclusion, second bar females can resist nutrient stress because they have ATF4, but fourth bar females can't because they don't have ATF4 - is this difference statistically significant?

    We thank the reviewer for pointing out this missing statistical report that compares the second and fourth bars in Figure 4D ____(females expressing methioninase, with and without ATF4 knockdown). We have now performed this analysis and reported the p-value in text (lines 282-285 in the preliminary revision).

    1. For all scatter plot graphs, figure legends should indicate what the horizontal line represents (is this the average?). Also, error bars and what they represent (SD or SEM) are not included or described.

    We thank the reviewer for asking for additional details on our graph annotations. We have added language to explain that 1) horizontal lines on ATF4 reporter quantification graphs denote mean intensity (Fig. 1 legend, lines 567-568 in the preliminary revision) and 2) error bars on qPCR graphs represent SEM (Fig. 2 legend, line 583 in the preliminary revision).

    Referee #3 Minor concerns:

    1. Line 27: "counter parts" should be one word
    2. Line 33: should the word "nutrient" be included before "stress"
    3. Line 42: It would be nice to see a couple of examples of the "well documented across species" statement
    4. Line 44-45: Add in the word "human" before population and use "women" instead of "females"
    5. Line 53: There seems to be an issue with comma placement or word usage in the section of the sentence that reads "coincident with, or a comorbidity, for"
    6. Lines 82-83: Mention of a couple examples would be nice
    7. Line 104: Perhaps add the word "cellular" before "sexual"
    8. Line 204: Delete the word "and" after "expression"
    9. Line 234: Delete "a" before "significantly"
    10. Line 276: Should "adult" be "adulthood"
    11. For the discussion, a model schematic would nicely depict the findings as a whole
    12. Line 330: May consider incorporating the following studies - Stobdan et al., 2019 and De Groef et al., 2021
    13. Related to the point above: It would be great to see discussion/speculation of potential ATF4 targets that might be mediating this effect
    14. Line 374: The placement of "yet unidentified" makes it seem like other ATF4 target genes aren't known, but really what is meant is that their sexually dimorphic expression is not known
    15. Line 535: (beta-gal) "protein" instead of "gene"?
    16. Figure S2: Please indicate what the two horizontal dotted lines are supposed to point out

    We thank the reviewer for carefully pointing out these minor yet critical text concerns. ____We have addressed all minor concerns raised by the reviewer in text edits to the preliminary revision, which are each highlighted in yellow in lines 27, 33, 44, 53, 105, 204, 236, 279, 375,554, 624 in the preliminary revision. The exceptions are points 3, 6, 11, 13, which we will address in the subsequent revision as described in the previous section.

  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

    Summary: Using a combination of genetic and molecular tools, Grmai and colleagues present data showing the sexually dimorphic expression of ATF4, a transcription factor that mediates the integrated stress response, in larval fat tissue. Moreover, they find that higher basal ATF4 activity in female larvae supports the stronger resistance to nutrient deprivation that females exhibit compared to male larvae. The data are clearly described and nicely laid out in well-organized figures. Some major, moderate, and minor concerns, delineated below, regarding the approach and conclusions should be addressed prior to acceptance for publication.

    Major concerns:

    1. Sexually dimorphic ATF4 activity (Figure 1 and associated supplemental figure) as evidenced by reporter expression is the basis of this study, yet a detailed description of the immunofluorescence quantification is lacking. The methods sections needs to include information on how a) images were acquired (Were the same acquisition settings used across all images?), b) the intensity measurements were taken (What software was used? Does each data point in the distribution represent a single nucleus (the assumption is yes)? Is nuclear size adjusted for? Panels A' and B' have obvious differences in nuclear size which would in turn affect total intensity measurements), c) the sample size (How many fat images taken per animal per sample/genotype? How many trials were performed?)
    2. While the authors nicely address the lack in specificity for two of the Gal4 driver lines used in the study limitation section, the fact that the one driver that is fat body-specific, 3.1Lsp2-Gal4, shows a modest, not statistically significant decrease in Figure 4C still raises some concern. There is another Lsp2-Gal4 line described in Lazareva et al., 2007 (PLoS Genetics) that drives expression in larval fat, perhaps to combat the issue of 3.1Lsp2-Gal4 have low activity, as mentioned by the authors. Alternatively, this phenotype could be assessed using Gal4 lines that only drive expression in the other tissues (if available). Otherwise, the conclusion that ISR/ATF4 signaling specifically in the fat mediates the starvation response needs to be toned down.
    3. Several analyses rely on RNAi, and this is understandably important for tissue-specific knockdown of gene expression. At least one of the two following issues needs to be addressed: a) the efficiency of knockdown for each gene are not provided or reported on and b) only single RNAi lines were used for each gene targeted for knockdown.

    Moderate concerns:

    1. Lines 137-141: It would be nice to see a gel that confirms that these newly designed primers detect the expected isoforms (supplemental perhaps).
    2. Lines 278-282 and Figure 4D: Shouldn't the second and fourth bars be compared? Based on the hypothesis and conclusion, second bar females can resist nutrient stress because they have ATF4, but fourth bar females can't because they don't have ATF4 - is this difference statistically significant?
    3. For all scatter plot graphs, figure legends should indicate what the horizontal line represents (is this the average?). Also, error bars and what they represent (SD or SEM) are not included or described.

    Minor concerns:

    1. Line 27: "counter parts" should be one word
    2. Line 33: should the word "nutrient" be included before "stress"
    3. Line 42: It would be nice to see a couple of examples of the "well documented across species" statement
    4. Line 44-45: Add in the word "human" before population and use "women" instead of "females"
    5. Line 53: There seems to be an issue with comma placement or word usage in the section of the sentence that reads "coincident with, or a comorbidity, for"
    6. Lines 82-83: Mention of a couple examples would be nice
    7. Line 104: Perhaps add the word "cellular" before "sexual"
    8. Line 204: Delete the word "and" after "expression"
    9. Line 234: Delete "a" before "significantly"
    10. Line 276: Should "adult" be "adulthood"
    11. For the discussion, a model schematic would nicely depict the findings as a whole
    12. Line 330: May consider incorporating the following studies - Stobdan et al., 2019 and De Groef et al., 2021
    13. Related to the point above: It would be great to see discussion/speculation of potential ATF4 targets that might be mediating this effect
    14. Line 374: The placement of "yet unidentified" makes it seem like other ATF4 target genes aren't known, but really what is meant is that their sexually dimorphic expression is not known
    15. Line 535: (beta-gal) "protein" instead of "gene"?
    16. Figure S2: Please indicate what the two horizontal dotted lines are supposed to point out

    Significance

    Study novelty: This work begins to shed light on the underlying molecular mechanisms that mediate differential responses to nutrient deprivation in male and female larvae. The knowledge gained from Drosophila studies will very likely have implications for human adipose physiology given the known sex differences in adipose associated physiology and pathophysiology in men and women.

    General assessment: This study makes excellent use of the Drosophila melanogaster genetic toolkit to better understand the involvement of the ISR in mediating sexually dimorphic responses to nutrient deprivation. In addition, carefully thought-out figure layouts make the data easy to visualize. Limitations of the study include lack of specificity of fat body-specific driver lines and thus a potentially overstated conclusion.

  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

    It is now well-established that the Integrated Stress Response (ISR) mediated by ATF4 plays important roles in metabolism and proteostasis. This manuscript by Grmai and colleagues reports that the sex determination genes tra and dsx allow higher levels of ATF4 expression in Drosophila. They further show that female flies depend on ATF4 to survive under conditions of metabolic stress.

    The presented data are technically sound, and the manuscript is generally very well written. It is a concise study with four Figures. The authors could have chosen to expand the scope: For example, they have shown the requirement, but not the sufficiency, of ATF4 in the sexually dimorphic nature of vulnerability to nutrient deprivation. They also demonstrate that ATF4 affects male-specific survival upon metabolic stress, which could be improved with additional experiments. These and other technical points are outlined below:

    1. Figure 4 is an important part of this study, where the authors show a male-specific vulnerability to methioninase expression. They show that ATF4 RNAi confers vulnerability to methioninase expression even in females. An obvious question is whether ATF4 overexpression is sufficient to enhance resistance to methionine deprivation in males.
    2. Methioninase expression results (Figure 4) are interesting. Are the levels of methioninase expression similar between males and females?
    3. This manuscript focuses on ATF4, but there could be additional possible reasons for the sexually dimorphic ISR activity. For example, the degree of physiological stress that activates ISR could be different between males and females. I suggest comparing the levels of Phospho-eIF2alpha (or any other markers upstream of ATF4) in both sexes.
    4. In Figures 1 to 3, the authors examine the intensity of ATF4 signaling after perturbing the sexual determination pathway. The methioninase experiments in Figure 4 are interesting, but there is nothing in this Figure linking male-specific vulnerability to sex determination genes. Examining the vulnerability to methioninase expression after perturbing the sexual determination genes would make Figure 4 integrate better with the rest of the manuscript.
    5. The authors write that they generated 4EBP intron-GFP because the 4EBP intron-DsRed signal was frequently observed in the cytoplasm (line 122). They seem to suggest that the DsRed reporter is less reliable than the GFP reporter. However, they continue to mix results using 4EBP intron-GFP (Fig. 4A) and 4EBP intron-DsRed (Fig. 4F). The two figures examine slightly different conditions (Fig 4A shows tra1 KO females, while Fig. 4F shows traF males). If the DsRed reporter is less reliable due to the signal from the cytoplasm, the authors should show results with the GFP reporter in traF males.
    6. In Figure S1, the authors label 4EBP intron-GFP as Thor2p-GFP, which is confusing. There are other parts in the methods section referring to Thor2p. I suggest using consistent terminology throughout the manuscript.

    Significance

    Overall, the authors report a novel and interesting observation because the sex determination pathway was not previously associated with ISR signaling. As many metabolic diseases show sex-specific outcomes, the main findings of this study will draw broad interest.

  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

    Summary

    This study aims to explore sexual dimorphism in tolerance to nutrient deprivation using Drosophila larvae as a model. In particular the authors focus on sex differences in the larval fat body. They show that ATF4, an ISR transcription factor, has higher mRNA levels in female fat bodies. ATF4 transcriptional activity is also higher based on a reporter of ATF4 function, where this female bias in expression is influenced by sex determination factors. When the authors

    Overall, this study is interesting, as it identifies previously unrecognized sex-specific regulation of ATF4, an important transcription factor that mediates cellular stress responses. The study also shows that sex determination genes regulate ATF4. However, I have concerns about the paradigms of nutrient deprivation used in the study, and about data interpretation and statistical analysis that should be addressed prior to publication to support the authors' conclusions.

    Major concerns

    1. One major concern that I have about the sexual dimorphism in tolerance to nutrient deprivation is that the papers cited by the authors, and paradigms that are used broadly in the field, all use adult flies. The authors must show that in larvae, a completely different life stage from their citations, there is a sexual dimorphism in tolerance to nutrient deprivation.

    Interestingly, Diaz et al 2023 (Genetics) show that male larvae have greater fat stores than female larvae. Considering fat is the main determinant of tolerance to nutrient deprivation it's not clear that females will actually survive nutrient deprivation longer as larvae. This is an essential test of whether female larvae do have increased tolerance to nutrient deprivation, which is the basic foundation of the authors' model.

    1. Another concern is the way that the authors "genetically induce nutrient deprivation using methioninase overexpression". As they acknowledge in the discussion (Line 381-390), methioninase overexpression will have many cellular effects. While there is no doubt that methionine levels would be lower in their model, it is less certain whether this is the main driver of the male-specific lethality.

    There are two potential solutions to this problem. First, the authors could change the text throughout the paper to more accurately describe their paradigm as "methioninase-induced lethality" rather than "nutrient deprivation". This would limit the scope of their scientific question and the conclusions they draw, but would eliminate the need for more experiments.

    The second solution would be to complete experiments to establish the following points: i) methioninase overexpression causes all the classical features of nutrient deprivation (e.g. changes to canonical signaling pathways such as TOR); ii) using other genetic means of nutrient deprivation such as slimfast-RNAi to see if those manipulations phenocopies the male-specific lethality they see with methioninase overexpression; iii) testing a role for ATF4 in mediating sex differences (if any) in other contexts such as slimfast-RNAi. This will take 2-3 months but is essential to draw any conclusions about whether their paradigm is truly a model for nutrient deprivation.

    1. Another important point is that the authors state that sexually dimorphic ATF4 activity in the fat body is instructed by sexual identity in a cell-autonomous manner. Despite a clear decrease in ATF4 reporter levels in tra mutants, the fat body-specific tra-RNAi effect on the ATF4 reporter was less convincing. Together with the fact that changes to tra in the fat body affect insulin secretion from the insulin-producing cells, it is possible that the effect on ATF4 is not cell-autonomous. To conclusively test if sexual identity regulates ATF4 in a cell-autonomous manner the authors should use the flp-out system to make Tra-expressing or tra-RNAi-expressing clones in the fat body. This would take approximately 1.5 months to make the strain and test this.
    2. As the authors show for the UAS-methioninase, other UAS lines used in the paper such as UAS-traF, UAS-tra-RNAi, UAS-dsx-RNAi may have leaky effects on gene/reporter expression. The authors must include a UAS only control to establish that the tra-RNAi, UAS-traF, UAS-dsx-RNAi do not affect gene/reporter expression.
    3. I have concerns about the statistics used. In the methods and legends only t-tests are mentioned; however, when three groups are compared a one-way ANOVA with post-hoc tests must be used to correct for multiple comparisons. To compare differential responses to genetic/environmental manipulations between the sexes, a two-way ANOVA must be used. For example, to conclude that males and females have different responses in the two-way ANOVA, there must be a significant genotype:sex interaction. The p-values for comparisons between genotypes in either the one-way or two-way ANOVA must be derived from post-hoc tests within the ANOVA analysis.

    Minor points

    1. Please ensure to make the reader aware of which life stage was tested in the literature cited supporting sexually dimorphic tolerance to nutrient deprivation.
    2. Published data about sex-specific mechanisms of metabolic regulation mean that the introduction should be more fully cited than it is. Even in the introduction "the molecular basis of these differences and how they impact tolerance to nutrient deprivation is still under investigation" is inaccurate, as there are published studies identifying some mechanisms (work on gut hormones and sex-specific effects on starvation resistance and body fat, role of ecdysone on body fat and feeding, sex-specific roles for brummer and Akh in regulating body fat, intestinal transit and gut size and feeding). Please adjust the paper to acknowledge this growing body of knowledge.
    3. Please list the diet ingredients per L so that individuals can replicate the diet easily.
    4. Please cite grant numbers for all the community resources (e.g. Bloomington, DSHB), and please acknowledge FlyBase and its grants as well. For example, here are the instructions for citing BDSC https://bdsc.indiana.edu/about/acknowledge.html and similar instructions are available for the other resources.

    Significance

    This study identifies for the first time the sex-specific regulation of ATF4, and reveals the sex determination genes that mediate this effect. A strength of the study is the characterization of sex-specific ATF4 regulation. Limitations of the study include the paradigm for nutrient deprivation, need for additional controls, and statistical analysis. If the concerns above are addressed, this study will be of interest to researchers studying organismal and cellular stress responses, stress signaling, and builds upon a growing body of knowledge of sex differences in stress responses (e.g. autophagy, infection responses).

  5. Version published to 10.1101/2024.12.27.630478v1 on bioRxiv