Sex-dimorphic gene regulation in murine macrophages across niches

  1. Cassandra J. McGill
  2. Olivia S. White
  3. Ryan J. Lu
  4. Nirmal K. Sampathkumar
  5. Bérénice A. Benayoun

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Macrophages are a key cell type of the innate immune system and are involved at all steps of inflammation: (i) they present antigens to initiate inflammation, (ii) they clear up foreign bodies through phagocytosis, and (iii) they resolve inflammation by removing or deactivating mediator cells. Many subtypes of macrophages have been identified, classified by their niche and/or embryonic origin. In order to better develop therapies for conditions with macrophage dysfunction, it is crucial to decipher potential sex-differences in key physiological mediators of inflammation so that treatment efficacy can be ensured regardless of biological sex. Here, we conduct a meta-analysis approach of transcriptomics datasets for male vs . female mouse macrophages across 8 niches to characterize conserved sex-dimorphic pathways in macrophages across origins and niches. For this purpose, we leveraged new and publicly available RNA-sequencing datasets from murine macrophages, preprocessed these datasets and filtered them based on objective QC criteria, and performed differential gene expression analysis using sex as the covariate of interest. Differentially expressed (DE) genes were compared across datasets and macrophage subsets, and functional enrichment analysis was performed to identify sex-specific functional differences. Consistent with their presence on the sex chromosomes, three genes were found to be differentially expressed across all datasets (i.e. Xist , Eif2s3y , and Ddx3y ). More broadly, we found that female-biased pathways across macrophage niches are more consistent than male-biased pathways, specifically relating to components of the extracellular matrix. Our findings increase our understanding of transcriptional similarities across macrophage niches and underscore the importance of including sex as a biological variable in immune-related studies.

Highlights

  • Across 17 independently transcriptomic macrophage datasets, three genes are always differentially expressed between males and females: Ddx3y , Eif2s3y , and Xist .

  • Jaccard analysis reveals female macrophages have more stereotypical transcriptional profile than male macrophages across niches.

  • Functional enrichment analyses show that female-biased pathways across macrophage subtypes are enriched for extracellular matrix-related genes.

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

    Manuscript number: RC-2024-0284z

    Corresponding author(s): Bérénice, Benayoun A

    1. General Statements [optional]

    This section is optional. Insert here any general statements you wish to make about the goal of the study or about the reviews.

    2. Point-by-point description of the revisions

    This section is mandatory. *Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. *

    Reviewer #1 (Evidence, reproducibility and clarity (Required)):

    This paper by McGill and colleagues explores sex differences in murine macrophages from different niches. They use a combination of publicly available, and newly developed datasets, and combine these using meta-analysis approaches. They explore DEGs between sexes - both common across niches, and specific to certain niches - and use enrichment analyses to identify pathways linked to these genes. Their overall conclusions are that gene expression changes in females are more consistent across niches, than for males, and are enriched in extracellular matrix-related genes. The paper is easy to follow and very well written.

    Major Comments:

    1. I would suggest Figure 1 be moved to a supplemental figure. We agree that the Xist and Ddx3y is QC and can be removed. However, we believe that the separation of macrophage transcriptomes based on sex in the Multidimensional Scaling plot is an important result. Thus, we have revised Figure 1 to only include the MDS plots and have moved the Xist/Ddx3y plots to the supplement (new Supplemental Figure S1) in line with the reviewer’s suggestion.

    Line 106 - It should be clarified why 50 DEGs was selected as the cut off for exclusion.

    We apologize that our cut off criteria was not explained clearly enough. Because these are publicly available datasets, every lab used different numbers of biological replicates, methods, and sequencing depths, impacting the power of the assay to detect differences in gene expression robustly. Since we were interested in functions that were sex-dimorphic, and that requires running functional enrichment analysis, we needed to have a minimum gene set size to be able to run these analyses, which, in the field, is usually accepted to be 50 genes for robustness. Thus, we used 50 DEGs and have updated the methods to explain our reasoning: “Applying a cutoff for the number of differentially expressed genes (DEGs) helps ensure data consistency and comparability across datasets with varying methodologies and sequencing depths. This prevents datasets with excessively low DEG counts from disproportionately influencing downstream analyses. A cutoff also reduces noise from spurious findings, prioritizing datasets with robust transcriptional changes that are more likely to be biologically meaningful. The excluded microglia dataset contained only 11 DEGs (whereas all other microglia datasets had hundreds of DEGs), the pleural macrophage dataset had 37 (whereas all other lung-related macrophage datasets had above 50), and the spleen macrophage dataset had only 30.” (page 12, lines 381-388).

    Optional - would suggest sex chromosome-linked genes are excluded and the analysis redone to see if there are other autosomal genes that are statistically shadowed by the X and Y linked genes.

    We thank the reviewer for this great suggestion, and we now added this point to the discussion (page 9, lines 260-268). However, we think that genes on the X and Y chromosomes will impact overall function of the macrophages and that they are necessary to understand how macrophages from males and females may support differences in immune function throughout life. We now add this in the discussion as a potential future direction: “We find that a majority of genes similarly differential across sexes among the macrophage niches are sex chromosome linked. X-linked genes like Tlr7, Cxcr3, and Kdm6a enhance immune responses in female macrophages, potentially increasing inflammation with age (Feng et al., 2024). Meanwhile, Y-linked genes such as Uty and Sry influence transcriptional regulation and inflammatory signaling in male macrophages, which may contribute to chronic low-grade inflammation (Lusis, 2019). These genetic differences affect macrophage activity, tissue-specific immune responses, and susceptibility to age-related diseases, highlighting the importance of sex-specific factors in immune research. Future research should also explore how non-sex chromosome-linked genes interact with these sex-specific mechanisms to further shape macrophage and immune function.” (page 9, lines 260-268).

    More metadata about the included studies should be included eg mouse ages, strains, experimental manipulations etc. I can't seem to access all of the Supplemental tables so this may already be included in Table S1.

    We agree that this information is important to take into consideration and have now included this information in Supplemental Table S1A, along with the accession numbers to each dataset. All mice were aged between 2 to 24 weeks and all on variations of the C57BL/6 background.

    How relevant the findings in mice are for humans should be explained further in the discussion.

    We agree that our discussion needs to better explain broader implications. Our findings are relevant for human health because macrophages play key roles in immunity, inflammation, and tissue homeostasis, and their functions are known to differ between sexes. Understanding these sex-specific transcriptional differences in mice can provide insights into how male and female immune systems respond differently to infections, autoimmune diseases, and aging in humans. Since macrophage phenotypes are influenced by both systemic factors (e.g., hormones) and tissue-specific environments, studying multiple macrophage subtypes from different organs helps identify conserved and context-dependent sex differences. Indeed, our findings suggest the ECM may be a potential mechanism underlying sex-biased diseases, such as higher autoimmune prevalence in females or increased susceptibility to certain infections in males. We have added this detail to the discussion (page 10, lines 269-275).

    Minor Comments:

    1. Lines 63-66 - need references here. This mirrors Reviewer 2’s major point #2. We agree with the reviewer that references are needed and now cite PMID: 31541153, PMID: 29533975, PMID: 37863894, PMID: 33415105, and PMID: 37491279 (page 4 line 68-69).

    Line 61 and 69 - repeated.

    We thank the reviewer for catching this oversight and have deleted the first instance of the sentence.

    Reviewer #1 (Significance (Required)):

    Although this study is primarily descriptive, it adds to the current knowledge about sex differences in macrophages, an important and relatively understudied area. Those interested in sex differences and in the innate immune system generally, plus those who study macrophages in any context, should be interested in this work.

    We thank the reviewer for their interest in our work and their helpful suggestions.

    Reviewer #2 (Evidence, reproducibility and clarity (Required)):

    Summary: The study investigates sex-specific differences in macrophage gene expression across various tissue niches by analyzing both newly generated and publicly available datasets of varying quality. The key finding is the identification of three consistently differentially expressed genes (DEGs) across all macrophage niches: the Y-chromosome-encoded genes Ddx3y and Eif2s3y, and the X-chromosome-specific gene Xist. However, the number of sex-dimorphic DEGs varied significantly between macrophage niches, with female-biased genes showing more consistency across datasets. To further explore these sex-specific differences, the authors performed an overrepresentation analysis of the DEGs across datasets. They found enriched gene sets associated with specific biological terms in female-biased macrophages from peritoneal macrophages, bone marrow-derived macrophages (BMDMs), and osteoclast progenitors (OCPs), while male-biased enrichment was observed in microglia, exudate macrophages, OCPs, and BMDMs. Notably, extracellular matrix (ECM)-related genes were specifically enriched in female peritoneal macrophages and OCPs, whereas the term "nucleic acid binding" was more prominent in male samples from microglia, BMDMs, and OCPs, driven by the Y-chromosome genes Uty and Kdm5d. A gene set enrichment analysis (GSEA) using Gene Ontology (GO) and Reactome databases further confirmed the enrichment of sex-biased pathways. Based on these findings, the authors conclude that three sex chromosome-associated genes are consistently differentially expressed across all datasets and that female-associated gene expression appears to be more stable, particularly in relation to ECM-associated processes.

    Major Comments:

    Are the key conclusions convincing?

    1. The study provides valuable insights into sex-dimorphic gene expression in macrophages across different niches. However, some conclusions appear overinterpreted due to the limited number of differentially expressed genes (DEGs) driving specific terms in the overrepresentation analysis. The reliance on only a few recurring genes (e.g., Kdm5d, Eif2s3y, Uty, and Ddx3y) raises concerns about the biological significance of some enriched terms. A clearer discussion on the limitations of such findings is necessary. We apologize for the confusion. Although the Venn Diagram may give the impression that our comparisons are limited to those few genes, we only highlight them with bold text because they are a good quality control mechanism for our analyses.

    Importantly, methods like gene set enrichment analysis [GSEA] use whole-transcriptome ranking, which means the results we obtain are driven by the entire transcriptome and not just a few genes (GSEA results are reported in Figure 5). We agree that further explanation of these methodologies would improve interpretation of our findings for readers unfamiliar with these analytical techniques. To address this, we have now added the following to the methods: “GSEA relies on whole-transcriptome ranking, ensuring that the results reflect global transcriptomic patterns rather than being influenced by only a few genes.” (page 13, lines 415-417).

    Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?Some claims, particularly those regarding the role of macrophages in diseases such as AD, histiocytosis, and osteoporosis, lack relevant references.

    This mirrors minor point #1 from Reviewer #1. We apologize for not originally including references for this statement and have now updated the introduction and discussion with appropriate references: “Excessive macrophage activation is associated with numerous conditions, including neurodegeneration, atherosclerosis, osteoporosis, and cancer, many of which exhibit sex-biased tendencies (Chen et al., 2020; Hou et al., 2023; Li et al., 2023; Mammana et al., 2018)” (page 4, lines 67-69) and “Thus, investigating female and male-biased processes in macrophages, including the contribution of the ECM, will be an important step in developing treatments for diseases including, but not limited to, AD, histiocytosis, and osteoporosis(Chen et al., 2020; Cox et al., 2021; Hou et al., 2023; Li et al., 2023; Mammana et al., 2018)” (page 10, lines 285-288).

    Would additional experiments be essential to support the claims of the paper? While additional wet-lab experiments are not strictly necessary, a deconvolution analysis of the datasets could be highly beneficial. This would allow the identification of enriched macrophage subtypes and help assess whether differences between datasets are driven by specific macrophage populations rather than global sex differences. Since peritoneal macrophage origin is influenced by age and inflammation status, deconvolution could also clarify dataset comparability.

    The reviewer makes an interesting point. We apologize for the confusion regarding the purity and origin of these datasets. All the datasets we curated from public repositories for our analysis are from purified populations of macrophages. To clarify this, we now include a column with the purification method used for each of the datasets based on the original manuscript in revised Supplemental Table S1A.

    Since all the used datasets were derived from pure macrophage populations, deconvolution (which is used to identify cellular proportions in heterogeneous contexts) would not accomplish much, predicting that all the cells in the data are macrophages. While some people have argued that deconvolution may be used to identify different cell states, this is very controversial, especially since the “pure” reference and the heterogeneous query are subject to batch effects (i.e. either from differences in bench processing, sex of provenance for target/query datasets, transcriptional impact of sorting methods, differences in transcriptomic quantification methods, etc.) which overshadow most differences beyond cell types. Thus, due to the known batch sensitivity of deconvolution methods and the fact that we only selected pure macrophage transcriptomic profiling datasets, using deconvolution to identify macrophage subtypes would not be informative/feasible. Importantly, we focused our analyses on datasets derived only from young, healthy, naïve animals (2 to 24 weeks), without any interference from age-related inflammation.

    To make this caveat clearer, we have added sentences to the results section indicating the age range of the animals (page 6, lines 100-101), as well as in the discussion to discuss how inflammation states and age may change some of our findings (page 10, lines 295-299).

    Are the suggested experiments realistic in terms of time and resources? Performing cell-type deconvolution using established computational tools (e.g., CIBERSORT, BisqueRNA, or single-cell deconvolution methods) would be a realistic approach within a few weeks and would significantly strengthen the study. This analysis would not require additional experimental work but could refine the interpretation of the dataset. Additionally, a PCA of all datasets could help identify potential similarities among macrophages from different niches and between sexes.

    As explained in our response to point #4, the use of only datasets from purified macrophages from young animals (before any influence of age or disease) makes deconvolution analysis meaningless, especially due to batching concerns. Specifically, it would require us to generate paired single-cell and bulk datasets on all macrophage subtypes in house to remove batch-inducing experimental biases, which we believe is outside of the scope of this small bioinformatics study.

    To the second point, doing a PCA of all the datasets together would not provide much new information beyond cell type of origin due to batching concerns that could not be corrected, which are a known problem in transcriptomics analyses (PMID:20838408, PMID:28351613). Since datasets come from different labs, using different isolation methods, RNA capture choices, library construction kits and sequencing platforms, the main separating effects overall will be batch/dataset, not biology (PMID:20838408, PMID:28351613). Indeed, this is what we observe (Reviewer Figure 1), with broad separation of datasets by tissue of origin, then dataset of origin. Additionally, the top 10 loadings for PC1 and PC2 are primarily associated to autosomal genes (i.e. not on the sex chromosomes; Reviewer Table 1).

    Reviewer Figure 1. (A) PCA of all samples across datasets. Read counts were processed together through R package sva v.3.46.0 for surrogate variable estimation, and surrogate variables were removed using the removeBatchEffect function from ‘limma’ v.3.54.2. DESeq2 normalized counts were used to make the PCA. (B) Zoomed in PCA excluding three outlier sample to enable easier visual discrimination of samples.

    Principal Component – Gene

    Loading

    Chromosome

    PC1- Srcin1

    0.013601

    11

    PC1- Cacna1c

    0.013593

    6

    PC1- Pclo

    0.01357

    5

    PC1- Tro

    0.013547

    X

    PC1- Ppp4r4

    0.013541

    12

    PC1- Ppp1r1a

    0.01354

    15

    PC1- Homer2

    0.013538

    7

    PC1- Caskin1

    0.013535

    17

    PC1- Arhgef9

    0.013527

    X

    PC1- Slc4a3

    0.013499

    1

    PC2- Gm15446

    0.017978

    5

    PC2- 1810034E14Rik

    0.017897

    13

    PC2- Gm19557

    0.017871

    19

    PC2- Pkd1l2

    0.017792

    8

    PC2- H60b

    0.017274

    10

    PC2- Appbp2os

    0.01723

    11

    PC2- Mir7050

    0.017221

    7

    PC2- Nkapl

    0.017166

    13

    PC2- Tmem51os1

    0.017083

    4

    PC2- Dpep3

    0.016962

    8

    Reviewer Table 1. Top 10 loadings for principal component 1 and principal component 2 with their respective chromosomal location.

    Thus, since batch effects can only be accounted for rigorously when they are not confounded by biology (and in our case since each dataset only looks at one type of macrophage), this cannot be corrected in a rigorous manner to yield the desired results.

    We have added a sentence to the discussion to highlight how future work where macrophages from diverse niches would be profiled in parallel may give greater insights into niche-specific sex-dimorphic effects (page 10, line 295-296).

    Are the data and the methods presented in such a way that they can be reproduced? Some methodological details are missing, particularly regarding:

    The isolation of mouse peritoneal macrophages (details on injection and harvesting procedure needed). Quality control of isolated macrophages (How were contaminating cells excluded? Was additional validation performed beyond using the kit?)

    The age of mice used for bone marrow-derived macrophages (BMDMs) is not provided, which is important given that immune responses can be age-dependent.

    We appreciate the reviewer’s request for additional methodological details. We apologize for not being clear with our details and have updated the methods to be clearer (page 11, lines 320-346), as well as added this information in revised __Supplemental Table S1A __(e.g. age of animals and purification method as described in the original papers). For all our in house datasets, mice were 4-months old, and the text is now updated to reflect this: “Long bones (tibia and femur) of young (4-months-old) from both sexes were collected and bone marrow was flushed into 1.5mL Eppendorf tubes via centrifugation (30 seconds, 10,000g) (Amend et al., 2016)” (page 11, lines 334-336).

    While we couldn’t check the purity post hoc for published datasets we identified for meta-analysis, we performed a purity check on our isolated peritoneal macrophages using Cd11b-F4/80 staining by flow cytometry and have now included this data (including gating strategy) in Supplemental Figure S4. For BMDMs, no purity check was performed, as there is extensive literature on the efficiency of this differentiation protocol which consistently yields > 90% of macrophages. This has been added to the methods: “We used a protocol that is expected to yield ~90% Cd11b+ F4/80+ cells (Mendoza et al., 2022; Toda et al., 2021)” (page 11, lines 336-337).

    Are the experiments adequately replicated and statistical analysis adequate? The statistical analysis appears generally appropriate, but there are concerns about dataset inconsistencies that should be addressed. Some datasets were not used across all analyses, which is not clearly indicated in figures or text. This should be explicitly mentioned to avoid misleading interpretations.

    We appreciate the reviewer’s careful evaluation of our statistical analysis and the concern regarding dataset inconsistencies.

    We believe that the reviewer is referring to the omission of the exudate dataset from the Venn Diagram analysis (Figure 2C), as this is the only time that we did not report the results from all datasets. We originally chose not to include the exudate dataset in the shared differentially expressed gene (DEG) analysis, because it contained over 1,300 DEGs, whereas all other datasets had between 4–30 DEGs, resulting in an unreadable figure.

    However, we agree that it is important to include for the readers, and while we have decided to still exclude the exudate dataset from Figure 1C for readability purposes, we now include the overlap analyses for all datasets in Supplemental Figure S2 using an upset plot (an alternative visualization method) showing all 6 niches, as well as a table panel that lists the shared genes across niches “Three genes were found to be differentially expressed across all six niches: Xist, Ddx3y, and *Eif2s3y *(Figure 2C, Supplemental Figure 2A,B)” (page 6, lines 124-126). We thank the reviewer for drawing our attention to this and making our analysis clearer for future readers.

    Minor Comments

    1. Figures are included twice in the manuscript. We apologize for this, and figures are now only included once.

    The use of stereotypic colors in figures (e.g., blue for male, pink for female) could be reconsidered for better readability and to avoid reinforcing gender stereotypes.

    While we understand that this color choice might feel gender normative, we respectfully disagree with the reviewer, as we believe that for the expediency of scientific communication it is important to choose a color palette that is easily understandable without confusion without even needing to consult a legend.

    Importantly, we have been using the same color palette in all publications from the lab on sex-differences for consistency (Lu et al, Nat aging 2021 PMID: 34514433; McGill et al, PLoS ONE, 2023 PMID: 38032907; Kang et al, J Neuroinflammation, 2024 PMID: 38840206; McGill et al, STAR Protocols, 2021 PMID: 34820637), which is crucial for scientific rigor and communication consistency.

    Results - Section 1

    Line 92: The word 'identified' may not be the most appropriate choice here, as it implies discovery rather than selection. Consider rephrasing to 'compiled' or 'gathered' to more accurately reflect the process of assembling the datasets. Additionally, the sentence structure could be refined for clarity, such as specifying that the datasets include both newly generated and publicly available data.

    We have changed two instances of using the word identified to “collected” and “gathered” (page 4, line 83 and page 6, line 98). We also adjusted the sentence to say, “Although we initially collected 21 datasets, both newly generated and publicly available, for our study, only 18 datasets were retained after various quality filtering steps for downstream analysis” (page 4, lines 83-85).

    Line 95: Specify the source of exudate-derived macrophage data.

    We have updated Supplemental Table S1A to make sure it was comprehensively describing the datasets we used in our analysis and double checked that it was complete (including for the exudate data). We have updated the text to reflect this: “All accession numbers and corresponding manuscripts are found in __Supplemental Table S1A” __(page 6, lines 103-104).

    Figure 1/2A: The scheme overview lacks clarity-its purpose is unclear. The two identical boxes are redundant and do not provide additional insight. Consider illustrating the origins of different macrophage subtypes instead. The cutoff of >50 DEGs should be included in the schematic to improve clarity. Overrepresentation and GSEA analysis should not be illustrated multiple times across different figures-it is redundant.

    In Figure 1A, we included the identical boxes to indicate that no datasets were excluded for incorrect labeling of males/females. However, we agree that this is unnecessary and have removed the second box as suggested.

    In Figure 2A, we agree the identical boxes are unneeded as the Xist/Ddx3y quality control step was listed in Figure 1A, and we have modified the figure accordingly.

    We also agree that including the DEG cutoff and removing the GSEA mention will streamline the figures and have updated them accordingly as well.

    Line 100: The mention of R software should be moved to the Methods section instead of appearing in the Results section.

    We have now updated the text to say, “Expression levels of male-specific Ddx3y and female-specific Xist genes across all samples were examined to ensure proper sex labeling of samples (Supplemental Figure 1A-U)” (page 6, lines 111-112).

    Figure 1B-V: The current figure layout is visually cluttered. Consider plotting male and female datasets together in a single graph with different point shapes instead of separate panels for each specific niche.

    This seems to echo the above request for a global PCA in Reviewer 2’s Major Point #4, which unfortunately cannot be included due to the disproportionate impact of batch effects that has been well documented in the literature (Reviewer Figure 1; PMID:20838408, PMID:28351613). However, to make the figure clearer and less cluttered, and to address related Reviewer 1’s Major Point #1, we have moved the Xist/Ddx3y plots to Supplemental Figure S1 and only include the Multidimensional Scaling plots in Figure 1 to showcase the sex separation in each dataset.

    Text-Figure alignment: The text describes male/female-specific gene expression levels first, while the figure starts with MDS analysis. The order should be consistent.

    We agree and have adjusted the text accordingly (lines 109-112).

    Figure 2C: Exudate data is missing-explain why.

    This point echoes major point #6. As explained above, we have clarified this and included new data panels for clarity (New Supplemental Figure S2).

    Results - Section 2

    Line 151: Use consistent terminology-either "DEGs" or "DE genes", not both.

    We replaced all instances of “DE genes” with DEGs (lines 132, 137, 141, 147, 149, 163, and 397).

    Figure 3A: The text suggests not all datasets were included in this analysis-this should be explicitly indicated in the figure.

    We apologize for the confusion. All datasets were included in this analysis; however, some niches did not have any GO terms passing the FDR

    Show the number of DEGs used for analysis.

    We apologize for the confusion. For the ORA analyses (Figures 3 and 4), we indicate the number of DEGs used for analysis in the panel header. For the GSEA analysis (Figure 5, Supplemental Figure S3), all expressed genes are ranked based on effect size without any prior filter (see response to major point #1), so DEGs are irrelevant for these analyses.

    Figure 3B: Smaller pale dots in the bubble plot are difficult to distinguish-consider using a darker outline.

    We have now added outlines to all the bubbles in the plots to help improve visibility.

    Line 158: The term "phagocytosis" appears inconsistent with the figure, where it is labeled "phagocytosis, recognition".

    We have updated the text accordingly (page 7, line 170).

    Figure 4B, D, E: The overrepresentation analysis is based on very few genes (often only 1-2 genes per term), which may lead to overinterpretation.

    We apologize for the lack of clarity of our previous manuscript. The number of genes used for DEG analysis is in the panel titles of Figure 3 and 4. While the overlap is small, this is unlikely to be spurious since all of the pathways we discuss show significant enrichment with FDR

    Consider explicitly naming these genes and discussing their biological role instead of assigning terms based on minimal evidence.

    We now discuss these genes in the results: “Male-biased GO terms for microglia, OCPs, and BMDMs derived from four genes: Kdm5d, Uty, Ddx3y, and *Eif2s3y. *All of these are Y-linked genes and play crucial roles in regulating innate and adaptive immune responses (Meester et al., 2020). Kdm5d and Uty influence adaptive immunity through chromatin remodeling and histone modification, while Ddx3y and Eif2s3y shape innate immune responses by modulating macrophage activation and cytokine production via translation initiation and RNA processing (Bloomer et al., 2013; Hamlin et al., 2024; Meester et al., 2020) “(page 8, lines 195-200).

    Figures S3G and S3H seem to be switched.

    We are puzzled by this comment, as our original manuscript did not include a Supplemental Figure S3. Out of an abundance of caution, however, we checked that Supplemental Table S3G and H were correctly labelled, and independently confirmed that they are not switched.

    Results - Section 3

    Figure 5A does not add significant new insights. Consider refining its content to highlight key findings more effectively.

    We respectfully disagree and believe that schematic overviews help readers understand what is accomplished in any specific figure and have thus decided to keep it.

    Number of genes included in the analysis is not provided-this is important to assess significance and should be stated in methods and figure legends.

    We apologize for the lack of clarity. As explained above, GSEA uses all the genes in rank order (PMID: 16199517), we now explain GSEA more explicitly in the text “GSEA relies on whole-transcriptome ranking, ensuring that the results reflect global transcriptomic patterns rather than being influenced by only a few genes” (page 13, lines 415-417).

    Discussion

    1. Line 201-203: Missing reference.

    We have now updated the text with the proper reference: “Tissue-resident macrophages are crucial to proper immune system function (Guilliams et al., 2020). While all macrophages share the responsibility of clearing cellular debris and foreign bodies, tissue-resident macrophages also have unique responsibilities that facilitate homeostasis throughout the body (Guilliams et al., 2020; Varol et al., 2015)” (page 9, lines 227-230).

    Reference 23 (1999) is outdated. Newer literature should be cited to reflect modern insights into sex differences in macrophages.

    We have now updated the text with an updated reference for two outdated references: (i) “Sex differences have previously been reported in macrophages, with female macrophages having higher phagocytic activity than males (Scotland et al., 2011)” (page 9, lines 232-233) and (ii) “Dysfunctional OCPs are associated with development of osteoporosis, a disease that is four times more prevalent in women (Alswat, 2017)” (page 10, lines 284-285).

    Peritoneal macrophages and OCPs originate from monocytes. Would deconvolution help identify enriched subtypes and assess dataset comparability?

    As noted in Reviewer 2’s Major Points #3 and #4, deconvolution analysis is not meaningful for subtype analysis without paired isolated/bulk datasets, which are outside of the scope of this study to generate.

    The 'more consistent' pathways found for female datasets are not discussed.

    We now discuss pathways found among the female datasets: “In addition, GSEA analysis of REACTOME gene sets showed male-biased expression for cell cycle related pathways (average set size 499), and female-biased expression for G protein-coupled receptor (GPCR) signaling (average set size 122) and extracellular matrix organization (average set size 127) (Figure 5C, Supplemental Table S4S-AJ; consistent with our ECM observation, Supplemental Figure S3A). Macrophages express a wide variety of GPCRs that allow them to respond to different stimuli. The expression of specific GPCRs influences macrophage polarization toward either a pro-inflammatory or anti-inflammatory state (Wang et al., 2019). A manual review of the genes contributing to this GPCR enrichment reveals the presence of several chemokine-related genes (such as Ccl4, Ccr4, Cxcl1, and others) (Supplemental Table S4). This suggests that females may have an increased abundance of chemokine GPCRs, potentially contributing to heightened autoimmune activity, among other factors.” (page 8, lines 212-222).

    Methods

    • Peritoneal macrophage isolation:

    Details on injection and harvesting are missing.

    We apologize for not being clear with our details and have modified the methods to be clearer (page 11, lines 320-331).

    How was contamination from other cell types assessed? F4/80 selection may not be fully macrophage-specific, and contamination could occur due to insufficient washing or the presence of non-macrophage F4/80+ cells.

    For the peritoneal macrophage datasets we generated, the macrophages were checked for purity through flow cytometry using Cd11b and F4/80 antibodies. We considered double positive Cd11b+ F4/80+ cells to be macrophages, which represents >95% of cells using our methodology (Supplemental Figure S4), without a difference between sexes.

    For the BMDMs, we utilize a protocol that is expected to yield ~90% Cd11b+ F4/80+ cells (PMID: 35212988 and PMID: 33458708).

    Finally, we now include the purification method for all publicly available datasets according to their original manuscript in Supplemental Table S1A and explicitly discuss the information for our in-house datasets in the methods (page 11, lines 321-346).

    • Bone marrow macrophages:

    Mouse age is not provided in the results part.

    We now provide this information in the methods (page 11, line 334). All ages for all datasets are now included in Supplemental Table S1A.

    Figure Legends

    Figure 2: Peritoneal macrophages are abbreviated as PeriMac-consider using this abbreviation consistently in the text.

    We respectfully disagree with the reviewer and choose to keep Peritoneal Macrophages spelled out in the text for clarity. We use the shorthand “PeriMac” in Figure 2 and Figure 5 solely for spacing purposes, but these are explained in the figure legend.

    Reviewer #2 (Significance (Required)):

    The study's strengths include the integration of multiple datasets, the use of both overrepresentation and GSEA, and the exploration of tissue-specific macrophage niches. These findings have relevance for diverse communities, including immunologists, sex-difference researchers, and those studying macrophage-driven diseases such as osteoporosis, neurodegeneration, and chronic inflammation. The work provides a foundation for further studies on sex-specific macrophage biology and may have implications for sex-specific therapeutic strategies. However, the study has limitations. The conclusions regarding enriched pathways rely heavily on a small number of DEGs, raising concerns about overinterpretation. Additionally, dataset variability and missing data for some analyses (e.g., exudate macrophages) could affect the robustness of the results.

    Despite these limitations, the study makes a meaningful but incremental advance by highlighting stable sex-dimorphic patterns in macrophage biology. It provides insights for both fundamental and translational research, particularly for audiences focused on immune regulation, sex-specific gene expression, and tissue-specific macrophage function.

    We thank the reviewer for understanding the importance of our work.

    Reviewer #3 (Evidence, reproducibility and clarity (Required)):

    Summary: McGill et al. explore sex-based differences in macrophage gene expression across various tissues. Using a meta-analysis of publicly available and newly generated datasets, they identify conserved and divergent sex-dimorphic genes and pathways between tissues. Overall, the report is easy to follow and guides the reader through the analysis. The authors highlight the relevance of the report by noting sex differences in immune responses to infection, autoimmunity, and chronic diseases. The inclusion of 17 independent transcriptomic datasets provides a robust and extensive analysis of sex-based transcriptional differences. The authors explore potential biological implications of sex-based transcriptional differences using pathway analysis. Despite the overall strengths, there are some points for which further clarification and analysis would improve the manuscript. Detailed comments are listed below.

    Major comments:

    1. A comparison of the overall transcriptomic profiles of macrophages regardless of sex would be additive. Knowing the degree of similarities and differences among macrophages from different niches would help the reader determine what genetic programs vary by compartment. If macrophages are very different by niche, it is not surprising that they share few sex-dimorphic patterns. This mirrors Reviewer 2’s Major Point #4. While this approach may seem valuable, it would only be feasible if all datasets were generated simultaneously by the same lab using identical sequencing and library preparation protocols to avoid batch effects. In this case, biology and batch effects are confounded, making any global analysis misleading. Although the reviewer may find the limited overlap unsurprising, given that macrophages are generally considered to be the same cell type, our goal was to explore the extent of shared versus distinct features across datasets, which we believe to be an invaluable question for the field.

    Although it would not be possible to do this rigorously with the data we curated, the question of niche specific gene regulation of macrophages has been studied, showing extensive niche-specific regulation: “While the question of niche-specific gene regulation has been studied, showing extensive niche-specific regulation (Gosselin et al., 2014; Lavin et al., 2014), a comprehensive and systematic study of sex-differences across macrophage subtypes has not yet been performed” (page 4, lines 78-81).

    It is unclear what age and strain the mice were and the number of samples that were included (n) for each dataset. This information should be included in S1A. If different ages or strains were used, how might this impact findings?

    This mirrors Reviewer 1’s Major Point #4. We agree that this information is important to take into consideration and have now included this information in Supplemental Table 1A, along with the accession numbers to each dataset. Because there is no aging effect (all mice are aged between 2 to 24 weeks) and all mice are on a variation of the C57BL/6 background, we don’t expect this to be a major problem impacting our findings.

    The authors used a Jaccard index to examine similarities in sex-based differences across tissue compartments. They claim that there are more similarities in females. However, the male are female graphs (Fig. 1E,D) do not look that different. Is there a better way to display this?

    We apologize for the lack of clarity. We clustered the Jaccard matrices using hierarchical clustering to determine patterns of sharing. Thus, in these figures, the samples cluster based on the degree of similarity in sex-biased genes. In the females, there is clear separation by macrophage origin (yolk sac or circulating monocytes); whereas males have some separation but also have some mixing (e.g. Peritoneal Macrophage 2 clustering with the yolk-sac derived macrophage datasets). Additionally, four microglia datasets are together in the females with only one separate, whereas in the males they are split into three. We included colored bars by the dataset names to help highlight clear separation by niche of origin.

    We have added this detail to the text to better explain the similarities: “Our results indicate that female-biased genes were more consistent among the cell types compared to male-biased genes (Figures 2D,E). In females, there is clear separation by macrophage origin (yolk sac or circulating monocytes), with all the peritoneal macrophages clustering together, followed by bone-related macrophages, then microglia and lung macrophages. In the males, the five microglia datasets are split into three groups, and Peritoneal Macrophage 2 clusters with the yolk-sac derived macrophage datasets” (page 7, lines 155-160).

    In the Gene Ontology analysis, it is unclear what type of GO pathways were included (biological process, cellular component, molecular function). Also, some of the GO analyses were done with very few genes (as little as 4).

    This echoes Reviewer #2’s Major Comment #1. For the Overrepresentation analysis (ORA) using Gene Ontology, we use the “ALL” option to include biological process, cellular component, and molecular function terms. We used ORA to look at shared DEGs across datasets of the same niche which is why some have very low input. For this reason, we also performed Gene Set Enrichment Analysis that uses all genes, not just those differentially expressed at FDR 5%, to examine gene changes at a broader level. In the methods we have added this information: “The differentially expressed genes shared within each niche were divided into up and down-regulated based on the sign of the DEseq2 log2 fold change. These gene lists were used as the shared genes and all expressed genes across datasets in that specific niche were used as the universe for the clusterProfiler function ‘enrichGO’, using the “ALL” option to include biological process, cellular component, and molecular function terms” (page 13, lines 405-410) and “GSEA relies on whole-transcriptome ranking, ensuring that the results reflect global transcriptomic patterns rather than being influenced by only a few genes.” (page 13, lines 415-417)”.

    Is it possible to combine datasets by tissue to remove potential batch effects before downstream analyses? At the very least, PCA on combined data may help determine if some biological (e.g., age, strain) or technical (batch) differences are contributing to identifying few common sex differences.

    This mirrors Reviewer #2’s Major Point #4. Unfortunately, since every dataset only examined a single niche, biology and batches are confounded, and thus performing a PCA on all datasets together will be driven by technical rather than biological drivers. Batch effects are a well-documented issue in genomics (PMID:20838408, PMID:28351613) Indeed, this is largely observed when we attempt this analysis, with datasets clustering by batch (Reviewer Figure 1). Due to the issue of uncorrectable batch effects, we do not believe this analysis meets the rigor required to be included in the revised manuscript and have chosen to not include it.

    Validation of key results would further strengthen the manuscript.

    We agree that future validation is important but is beyond the scope of this purely bioinformatic analysis. We have included text in the revision to highlight the importance of future validation studies: “Thus, investigating female- and male-biased processes in macrophages, including the contribution of the ECM, will be an important step in developing treatments for diseases including, but not limited to, AD, histiocytosis, and osteoporosis, and future research will be essential to validate these findings and further refine therapeutic strategies (Chen et al., 2020; Cox et al., 2021; Hou et al., 2023; Li et al., 2023; Mammana et al., 2018)” (page 10, lines 285-289).

    Further contextualization of key results would enhance the discussion. For example, ECM-related differences in female macrophages could have broader roles in wound healing, fibrosis, and migration.

    We agree with the reviewers and have added this detail to the discussion: “ECM components are emerging as key regulators of innate immune responses (García-García & Martin, 2019). Macrophages contribute to ECM remodeling by producing and degrading collagens (Sutherland et al., 2023), and ECM-related differences in female macrophages may impact wound healing, fibrosis, and migration. In lung and kidney tissues, macrophages recruit and activate fibroblasts, influencing fibrosis through direct interactions and ECM-degrading enzymes (Nikolic-Paterson et al., 2014). The balance between ECM deposition and degradation is crucial for tissue homeostasis, as excessive fibrosis leads to pathology (Nikolic-Paterson et al., 2014; Ran et al., 2025). Mechanical properties of the ECM, such as stiffness and collagen crosslinking, enhance macrophage adhesion, migration, and inflammatory activation (Hsieh et al., 2019). These ECM cues direct macrophage behavior during injury response, influencing their ability to reach inflammation sites and promote repair. Thus, female-biased expression of ECM-related genes may contribute to phenotypes such as enhanced wound healing or even fibrosis(Balakrishnan et al., 2021; Harness-Brumley et al., 2014; Rønø et al., 2013) “ (page 9, lines 248-259).

    Minor comments:

    1. Line 51: In the introduction, the authors state that macrophages produce chemokines. There are other signaling molecules produced by macrophages (e.g., cytokines) that also contribute to immune responses. We apologize for this and have updated the text to say: “Macrophages are a key component of the mammalian immune system and are responsible for producing a diverse array of signaling molecules including (but not limited to) cytokines, chemokines, and interferons that activate the rest of the immune system to combat infection (Shapouri-Moghaddam et al., 2018)” (page 4, lines 49-52).

    Line 53: The authors state that after birth the primary source of new macrophages come from differentiation of monocytes. However, some tissue resident macrophages are self-renewing.

    We apologize for this oversight and have adjusted the text to say: “After birth, the primary source of new macrophages comes from the differentiation of monocytes, which can be recruited to tissues throughout life. However, some tissue resident macrophages can self-renew, including those from the pleural and peritoneal cavities (Röszer, 2018)” (page 4, lines 53-56).

    Line 123: "spermatogenial" should be "spermatogonial"

    We have updated the text accordingly (page 6, line 130).

    Reviewer #3 (Significance (Required)):

    Significance: • General assessment: The study provides a novel and comprehensive analysis of sex-dimorphic gene expression in macrophages, with key findings that emphasize the importance of ECM remodeling in female macrophages. The strengths include the broad dataset inclusion, rigorous quality control, and methodological rigor. However, consideration of potential confounding variables (e.g., age, strain) should be included and validation of key results would strengthen the manuscript. • Advance: This study advances knowledge by analyzing sex differences across multiple macrophage niches rather than focusing on a single tissue type. It extends findings from previous immune studies. • Audience: This report would be of interest to immunologists and researchers studying sex differences. Expertise: Immunology, sex differences in disease, macrophage biology, transcriptomics, and inflammation research.

    We thank the reviewer for their positive comments on the impact of our work and for their useful feedback.

    __ __

    References

    Alswat, K. A. (2017). Gender Disparities in Osteoporosis. J Clin Med Res,* 9*(5), 382-387. https://doi.org/10.14740/jocmr2970w

    Amend, S. R., Valkenburg, K. C., & Pienta, K. J. (2016). Murine Hind Limb Long Bone Dissection and Bone Marrow Isolation. J Vis Exp(110). https://doi.org/10.3791/53936

    Balakrishnan, M., Patel, P., Dunn-Valadez, S., Dao, C., Khan, V., Ali, H., El-Serag, L., Hernaez, R., Sisson, A., Thrift, A. P., Liu, Y., El-Serag, H. B., & Kanwal, F. (2021). Women Have a Lower Risk of Nonalcoholic Fatty Liver Disease but a Higher Risk of Progression vs Men: A Systematic Review and Meta-analysis. Clin Gastroenterol Hepatol,* 19*(1), 61-71.e15. https://doi.org/10.1016/j.cgh.2020.04.067

    Bloomer, L. D., Nelson, C. P., Eales, J., Denniff, M., Christofidou, P., Debiec, R., Moore, J., Zukowska-Szczechowska, E., Goodall, A. H., Thompson, J., Samani, N. J., Charchar, F. J., & Tomaszewski, M. (2013). Male-specific region of the Y chromosome and cardiovascular risk: phylogenetic analysis and gene expression studies. Arterioscler Thromb Vasc Biol,* 33*(7), 1722-1727. https://doi.org/10.1161/atvbaha.113.301608

    Chen, K., Jiao, Y., Liu, L., Huang, M., He, C., He, W., Hou, J., Yang, M., Luo, X., & Li, C. (2020). Communications Between Bone Marrow Macrophages and Bone Cells in Bone Remodeling. Front Cell Dev Biol,* 8*, 598263. https://doi.org/10.3389/fcell.2020.598263

    Cox, N., Pokrovskii, M., Vicario, R., & Geissmann, F. (2021). Origins, Biology, and Diseases of Tissue Macrophages. Annu Rev Immunol,* 39*, 313-344. https://doi.org/10.1146/annurev-immunol-093019-111748

    Gosselin, D., Link, V. M., Romanoski, C. E., Fonseca, G. J., Eichenfield, D. Z., Spann, N. J., Stender, J. D., Chun, H. B., Garner, H., Geissmann, F., & Glass, C. K. (2014). Environment drives selection and function of enhancers controlling tissue-specific macrophage identities. Cell,* 159*(6), 1327-1340. https://doi.org/10.1016/j.cell.2014.11.023

    Hamlin, R. E., Pienkos, S. M., Chan, L., Stabile, M. A., Pinedo, K., Rao, M., Grant, P., Bonilla, H., Holubar, M., Singh, U., Jacobson, K. B., Jagannathan, P., Maldonado, Y., Holmes, S. P., Subramanian, A., & Blish, C. A. (2024). Sex differences and immune correlates of Long Covid development, symptom persistence, and resolution. Sci Transl Med,* 16*(773), eadr1032. https://doi.org/10.1126/scitranslmed.adr1032

    Harness-Brumley, C. L., Elliott, A. C., Rosenbluth, D. B., Raghavan, D., & Jain, R. (2014). Gender differences in outcomes of patients with cystic fibrosis. J Womens Health (Larchmt),* 23*(12), 1012-1020. https://doi.org/10.1089/jwh.2014.4985

    Hou, P., Fang, J., Liu, Z., Shi, Y., Agostini, M., Bernassola, F., Bove, P., Candi, E., Rovella, V., Sica, G., Sun, Q., Wang, Y., Scimeca, M., Federici, M., Mauriello, A., & Melino, G. (2023). Macrophage polarization and metabolism in atherosclerosis. Cell Death Dis,* 14*(10), 691. https://doi.org/10.1038/s41419-023-06206-z

    Lavin, Y., Winter, D., Blecher-Gonen, R., David, E., Keren-Shaul, H., Merad, M., Jung, S., & Amit, I. (2014). Tissue-resident macrophage enhancer landscapes are shaped by the local microenvironment. Cell,* 159*(6), 1312-1326. https://doi.org/10.1016/j.cell.2014.11.018

    Li, M., Yang, Y., Xiong, L., Jiang, P., Wang, J., & Li, C. (2023). Metabolism, metabolites, and macrophages in cancer. J Hematol Oncol,* 16*(1), 80. https://doi.org/10.1186/s13045-023-01478-6

    Mammana, S., Fagone, P., Cavalli, E., Basile, M. S., Petralia, M. C., Nicoletti, F., Bramanti, P., & Mazzon, E. (2018). The Role of Macrophages in Neuroinflammatory and Neurodegenerative Pathways of Alzheimer's Disease, Amyotrophic Lateral Sclerosis, and Multiple Sclerosis: Pathogenetic Cellular Effectors and Potential Therapeutic Targets. Int J Mol Sci,* 19*(3). https://doi.org/10.3390/ijms19030831

    Meester, I., Manilla-Muñoz, E., León-Cachón, R. B. R., Paniagua-Frausto, G. A., Carrión-Alvarez, D., Ruiz-Rodríguez, C. O., Rodríguez-Rangel, X., & García-Martínez, J. M. (2020). SeXY chromosomes and the immune system: reflections after a comparative study. Biol Sex Differ,* 11*(1), 3. https://doi.org/10.1186/s13293-019-0278-y

    Rønø, B., Engelholm, L. H., Lund, L. R., & Hald, A. (2013). Gender affects skin wound healing in plasminogen deficient mice. PLoS One,* 8*(3), e59942. https://doi.org/10.1371/journal.pone.0059942

    Röszer, T. (2018). Understanding the Biology of Self-Renewing Macrophages. Cells,* 7*(8). https://doi.org/10.3390/cells7080103

    Scotland, R. S., Stables, M. J., Madalli, S., Watson, P., & Gilroy, D. W. (2011). Sex differences in resident immune cell phenotype underlie more efficient acute inflammatory responses in female mice. Blood,* 118*(22), 5918-5927. https://doi.org/10.1182/blood-2011-03-340281

    Shapouri-Moghaddam, A., Mohammadian, S., Vazini, H., Taghadosi, M., Esmaeili, S. A., Mardani, F., Seifi, B., Mohammadi, A., Afshari, J. T., & Sahebkar, A. (2018). Macrophage plasticity, polarization, and function in health and disease. J Cell Physiol,* 233*(9), 6425-6440. https://doi.org/10.1002/jcp.26429

    Wang, X., Iyer, A., Lyons, A. B., Körner, H., & Wei, W. (2019). Emerging Roles for G-protein Coupled Receptors in Development and Activation of Macrophages. Front Immunol,* 10*, 2031. https://doi.org/10.3389/fimmu.2019.02031

  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:

    McGill et al. explore sex-based differences in macrophage gene expression across various tissues. Using a meta-analysis of publicly available and newly generated datasets, they identify conserved and divergent sex-dimorphic genes and pathways between tissues. Overall, the report is easy to follow and guides the reader through the analysis. The authors highlight the relevance of the report by noting sex differences in immune responses to infection, autoimmunity, and chronic diseases. The inclusion of 17 independent transcriptomic datasets provides a robust and extensive analysis of sex-based transcriptional differences. The authors explore potential biological implications of sex-based transcriptional differences using pathway analysis. Despite the overall strengths, there are some points for which further clarification and analysis would improve the manuscript. Detailed comments are listed below.

    Major comments:

    • A comparison of the overall transcriptomic profiles of macrophages regardless of sex would be additive. Knowing the degree of similarities and differences among macrophages from different niches would help the reader determine what genetic programs vary by compartment. If macrophages are very different by niche, it is not surprising that they share few sex-dimorphic patterns.
    • It is unclear what age and strain the mice were and the number of samples that were included (n) for each dataset. This information should be included in S1A. If different ages or strains were used, how might this impact findings?
    • The authors used a Jaccard index to examine similarities in sex-based differences across tissue compartments. They claim that there are more similarities in females. However, the male are female graphs (Fig. 1E,D) do not look that different. Is there a better way to display this?
    • In the Gene Ontology analysis, it is unclear what type of GO pathways were included (biological process, cellular component, molecular function). Also, some of the GO analyses were done with very few genes (as little as 4).
    • Is it possible to combine datasets by tissue to remove potential batch effects before downstream analyses? At the very least, PCA on combined data may help determine if some biological (e.g., age, strain) or technical (batch) differences are contributing to identifying few common sex differences.
    • Validation of key results would further strengthen the manuscript.
    • Further contextualization of key results would enhance the discussion. For example, ECM-related differences in female macrophages could have broader roles in wound healing, fibrosis, and migration.

    Minor comments:

    • Line 51: In the introduction, the authors state that macrophages produce chemokines. There are other signaling molecules produced by macrophages (e.g., cytokines) that also contribute to immune responses.
    • Line 53: The authors state that after birth the primary source of new macrophages come from differentiation of monocytes. However, some tissue resident macrophages are self-renewing.
    • Line 23: "spermatogenial" should be "spermatogonial"

    Significance

    Significance:

    • General assessment: The study provides a novel and comprehensive analysis of sex-dimorphic gene expression in macrophages, with key findings that emphasize the importance of ECM remodeling in female macrophages. The strengths include the broad dataset inclusion, rigorous quality control, and methodological rigor. However, consideration of potential confounding variables (e.g., age, strain) should be included and validation of key results would strengthen the manuscript.
    • Advance: This study advances knowledge by analyzing sex differences across multiple macrophage niches rather than focusing on a single tissue type. It extends findings from previous immune studies.
    • Audience: This report would be of interest to immunologists and researchers studying sex differences.

    Expertise: Immunology, sex differences in disease, macrophage biology, transcriptomics, and inflammation research.

  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

    Summary:

    The study investigates sex-specific differences in macrophage gene expression across various tissue niches by analyzing both newly generated and publicly available datasets of varying quality. The key finding is the identification of three consistently differentially expressed genes (DEGs) across all macrophage niches: the Y-chromosome-encoded genes Ddx3y and Eif2s3y, and the X-chromosome-specific gene Xist. However, the number of sex-dimorphic DEGs varied significantly between macrophage niches, with female-biased genes showing more consistency across datasets. To further explore these sex-specific differences, the authors performed an overrepresentation analysis of the DEGs across datasets. They found enriched gene sets associated with specific biological terms in female-biased macrophages from peritoneal macrophages, bone marrow-derived macrophages (BMDMs), and osteoclast progenitors (OCPs), while male-biased enrichment was observed in microglia, exudate macrophages, OCPs, and BMDMs. Notably, extracellular matrix (ECM)-related genes were specifically enriched in female peritoneal macrophages and OCPs, whereas the term "nucleic acid binding" was more prominent in male samples from microglia, BMDMs, and OCPs, driven by the Y-chromosome genes Uty and Kdm5d. A gene set enrichment analysis (GSEA) using Gene Ontology (GO) and Reactome databases further confirmed the enrichment of sex-biased pathways. Based on these findings, the authors conclude that three sex chromosome-associated genes are consistently differentially expressed across all datasets and that female-associated gene expression appears to be more stable, particularly in relation to ECM-associated processes.

    Major Comments:

    Are the key conclusions convincing?

    The study provides valuable insights into sex-dimorphic gene expression in macrophages across different niches. However, some conclusions appear overinterpreted due to the limited number of differentially expressed genes (DEGs) driving specific terms in the overrepresentation analysis. The reliance on only a few recurring genes (e.g., Kdm5d, Eif2s3y, Uty, and Ddx3y) raises concerns about the biological significance of some enriched terms. A clearer discussion on the limitations of such findings is necessary.

    Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether? Some claims, particularly those regarding the role of macrophages in diseases such as AD, histiocytosis, and osteoporosis, lack relevant references.

    Would additional experiments be essential to support the claims of the paper?

    While additional wet-lab experiments are not strictly necessary, a deconvolution analysis of the datasets could be highly beneficial. This would allow the identification of enriched macrophage subtypes and help assess whether differences between datasets are driven by specific macrophage populations rather than global sex differences. Since peritoneal macrophage origin is influenced by age and inflammation status, deconvolution could also clarify dataset comparability.

    Are the suggested experiments realistic in terms of time and resources?

    Performing cell-type deconvolution using established computational tools (e.g., CIBERSORT, BisqueRNA, or single-cell deconvolution methods) would be a realistic approach within a few weeks and would significantly strengthen the study. This analysis would not require additional experimental work but could refine the interpretation of the dataset. Additionally, a PCA of all datasets could help identify potential similarities among macrophages from different niches and between sexes.

    Are the data and the methods presented in such a way that they can be reproduced?

    Some methodological details are missing, particularly regarding: The isolation of mouse peritoneal macrophages (details on injection and harvesting procedure needed). Quality control of isolated macrophages (How were contaminating cells excluded? Was additional validation performed beyond using the kit?) The age of mice used for bone marrow-derived macrophages (BMDMs) is not provided, which is important given that immune responses can be age-dependent.

    Are the experiments adequately replicated and statistical analysis adequate?

    The statistical analysis appears generally appropriate, but there are concerns about dataset inconsistencies that should be addressed. Some datasets were not used across all analyses, which is not clearly indicated in figures or text. This should be explicitly mentioned to avoid misleading interpretations.

    Minor Comments

    Figures are included twice in the manuscript. The use of stereotypic colors in figures (e.g., blue for male, pink for female) could be reconsidered for better readability and to avoid reinforcing gender stereotypes.

    Results - Section 1

    • Line 92: The word 'identified' may not be the most appropriate choice here, as it implies discovery rather than selection. Consider rephrasing to 'compiled' or 'gathered' to more accurately reflect the process of assembling the datasets. Additionally, the sentence structure could be refined for clarity, such as specifying that the datasets include both newly generated and publicly available data.
    • Line 95: Specify the source of exudate-derived macrophage data.
    • Figure 1/2A: The scheme overview lacks clarity-its purpose is unclear. The two identical boxes are redundant and do not provide additional insight. Consider illustrating the origins of different macrophage subtypes instead. The cutoff of >50 DEGs should be included in the schematic to improve clarity. Overrepresentation and GSEA analysis should not be illustrated multiple times across different figures-it is redundant.
    • Line 100: The mention of R software should be moved to the Methods section instead of appearing in the Results section.
    • Figure 1B-V: The current figure layout is visually cluttered. Consider plotting male and female datasets together in a single graph with different point shapes instead of separate panels for each specific niche.
    • Text-Figure alignment: The text describes male/female-specific gene expression levels first, while the figure starts with MDS analysis. The order should be consistent.
    • Figure 2C: Exudate data is missing-explain why.

    Results - Section 2

    • Line 151: Use consistent terminology-either "DEGs" or "DE genes", not both.
    • Figure 3A: The text suggests not all datasets were included in this analysis-this should be explicitly indicated in the figure.
      • Show the number of DEGs used for analysis.
    • Figure 3B: Smaller pale dots in the bubble plot are difficult to distinguish-consider using a darker outline.
    • Line 158: The term "phagocytosis" appears inconsistent with the figure, where it is labeled "phagocytosis, recognition".
    • Figure 4B, D, E: The overrepresentation analysis is based on very few genes (often only 1-2 genes per term), which may lead to overinterpretation.
    • Consider explicitly naming these genes and discussing their biological role instead of assigning terms based on minimal evidence.
    • Figures S3G and S3H seem to be switched.

    Results - Section 3

    • Figure 5A does not add significant new insights. Consider refining its content to highlight key findings more effectively.
      • Number of genes included in the analysis is not provided-this is important to assess significance and should be stated in methods and figure legends.

    Discussion

    • Line 201-203: Missing reference.
    • Reference 23 (1999) is outdated. Newer literature should be cited to reflect modern insights into sex differences in macrophages.
    • Peritoneal macrophages and OCPs originate from monocytes. Would deconvolution help identify enriched subtypes and assess dataset comparability?
    • The 'more consistent' pathways found for female datasets are not discussed.

    Methods

    • Peritoneal macrophage isolation:
      • Details on injection and harvesting are missing.
      • How was contamination from other cell types assessed? F4/80 selection may not be fully macrophage-specific, and contamination could occur due to insufficient washing or the presence of non-macrophage F4/80+ cells.
    • Bone marrow macrophages:
      • Mouse age is not provided in the results part.

    Figure Legends

    • Figure 2: Peritoneal macrophages are abbreviated as PeriMac-consider using this abbreviation consistently in the text.

    Significance

    The study's strengths include the integration of multiple datasets, the use of both overrepresentation and GSEA, and the exploration of tissue-specific macrophage niches. These findings have relevance for diverse communities, including immunologists, sex-difference researchers, and those studying macrophage-driven diseases such as osteoporosis, neurodegeneration, and chronic inflammation. The work provides a foundation for further studies on sex-specific macrophage biology and may have implications for sex-specific therapeutic strategies.

    However, the study has limitations. The conclusions regarding enriched pathways rely heavily on a small number of DEGs, raising concerns about overinterpretation. Additionally, dataset variability and missing data for some analyses (e.g., exudate macrophages) could affect the robustness of the results.

    Despite these limitations, the study makes a meaningful but incremental advance by highlighting stable sex-dimorphic patterns in macrophage biology. It provides insights for both fundamental and translational research, particularly for audiences focused on immune regulation, sex-specific gene expression, and tissue-specific macrophage function.

  4. Review Commons

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

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    This paper by McGill and colleagues explores sex differences in murine macrophages from different niches. They use a combination of publicly available, and newly developed datasets, and combine these using meta-analysis approaches. They explore DEGs between sexes - both common across niches, and specific to certain niches - and use enrichment analyses to identify pathways linked to these genes. Their overall conclusions are that gene expression changes in females are more consistent across niches, than for males, and are enriched in extracellular matrix-related genes. The paper is easy to follow and very well written.

    Major Comments:

    1. I would suggest Figure 1 be moved to a supplemental figure.
    2. Line 106 - It should be clarified why 50 DEGs was selected as the cut off for exclusion.
    3. Optional - would suggest sex chromosome-linked genes are excluded and the analysis redone to see if there are other autosomal genes that are statistically shadowed by the X and Y linked genes.
    4. More metadata about the included studies should be included eg mouse ages, strains, experimental manipulations etc. I can't seem to access all of the supplementary tables so this may already be included in Table S1.
    5. How relevant the findings in mice are for humans should be explained further in the discussion.

    Minor Comments:

    1. Lines 63-66 - need references here.
    2. Line 61 and 69 - repeated.

    Significance

    Although this study is primarily descriptive, it adds to the current knowledge about sex differences in macrophages, an important and relatively understudied area. Those interested in sex differences and in the innate immune system generally, plus those who study macrophages in any context, should be interested in this work.

  5. Version published to 10.1101/2024.12.20.629843v1 on bioRxiv