CoChat
Sign Up
Cited by 463Neuroinflammation and Neurodegeneration MechanismsSingle-cell and spatial transcriptomicsNeurogenesis and neuroplasticity mechanisms

Cross-Species Single-Cell Analysis Reveals Divergence of the Primate Microglia Program

Open access
Geirsdottir, Laufey, David, Eyal, Keren-Shaul, Hadas, Weiner, Assaf, Bohlen, Stefan Cornelius, Neuber, Jana, Balic, Adam, Giladi, Amir +20 more·2019·Cell
DOI 10.1016/j.cell.2019.11.010
Open this paper in CoChat — free accountOpen PDF

Abstract

Microglia, the brain-resident immune cells, are critically involved in many physiological and pathological brain processes, including neurodegeneration. Here we characterize microglia morphology and transcriptional programs across ten species spanning more than 450 million years of evolution. We find that microglia express a conserved core gene program of orthologous genes from rodents to humans, including ligands and receptors associated with interactions between glia and neurons. In most species, microglia show a single dominant transcriptional state, whereas human microglia display significant heterogeneity. In addition, we observed notable differences in several gene modules of rodents compared with primate microglia, including complement, phagocytic, and susceptibility genes to neurodegeneration, such as Alzheimer's and Parkinson's disease. Our study provides an essential resource of conserved and divergent microglia pathways across evolution, with important implications for future development of microglia-based therapies in humans.

AI snapshot

Written from the paper’s full text.

TL;DR

Geirsdottir et al. profile microglia morphology and transcriptomes across evolutionarily distant species and find a conserved core microglial program alongside primate-specific divergence. Human microglia are notably heterogeneous and express more neurodegeneration susceptibility genes than rodent microglia.

Key findings

  • A conserved microglia core program was identified across species: 8,890 genes were expressed in microglia of at least one species, clustering into 17 prominent gene clusters, with core signature clusters 1–10 expressed in all species but less in zebrafish and chicken; 163 genes were conserved and unique to microglia versus other tissue macrophages.
  • Most species showed a single dominant microglial transcriptional state, whereas human microglia displayed significant heterogeneity; a putative senescent-like subset comprised roughly 20% of all human microglia, was consistent across all six sampled individuals, and co-expressed CDKN1, CCL3, CCL4, CCL3L3, and CCL4L2 with higher TNF and IL1B.
  • Human PD and AD susceptibility gene expression correlated most with macaque (PD, r = 0.73; AD, r = 0.59), while mice, rats, and hamsters showed moderate to low correlation (r = 0.16–0.40); no significant enrichment or depletion was seen for schizophrenia or Huntington’s disease genes.
  • Histological analysis across 18 species found the highest microglial density in leech ganglia (299.0 ± 41.4 Iba1+ microglia/mm²) and lowest in axolotls (11.1 ± 0.7/mm²); mouse frontal cortex had 65.8 ± 1.09 cells/mm² versus human frontal cortex 35.5 ± 2.48 cells/mm².

Methods

  • Microglia morphology was assessed by immunohistochemistry and 3D Imaris reconstruction across 18 species, using Iba1, mPEG1-GFP for zebrafish, and CSF1R-mApple for chicken.
  • For transcriptomics, brains from eight species with high-quality reference genomes (human, macaque, marmoset, sheep, mouse, hamster, chicken, zebrafish) were processed by FACS: CD45+ cells for MARS-seq2.0 scRNA-seq and CD11b+CD45lo microglia (or GFP+ zebrafish, CD45+CSF1RmApple+ chicken) for bulk RNA-seq; 4,458 QC-positive microglia cells and 3–6 bulk samples per species were collected, with rat and blind mole rat later added for single-cell comparisons.
  • Analysis used an ortholog “metagene” strategy, metacell clustering, scRNA-seq-based deconvolution of bulk RNA-seq to remove contaminating cells, rank-based normalization, k-means clustering, and GWAS enrichment with NHGRI-EBI gene lists and hypergeometric tests (p < 0.01); human brain tissue came from healthy regions of individuals undergoing epilepsy surgery.

Limitations

  • Stated: the authors caution that human CNS tissue studies cannot exclude that patients’ microglia do not accurately recapitulate the in situ states of healthy brains, despite careful sample acquisition and processing and similar results in comparable studies.
  • Stated/evident: human samples were from epilepsy surgery (healthy regions with absence of pathology confirmed histologically), so they may not represent healthy brain microglia; sample sizes were small (3–6 individuals per species, human n = 6, with some species using pooled brains).
  • Evident from text: cross-species bulk RNA-seq normalization is challenging, deconvolution depends on scRNA-defined contaminating clusters, chicken and zebrafish were removed from mammalian comparisons, and microglial heterogeneity may be influenced by tissue source or processing.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.cell.2019.11.010

Cites

103 of 103 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

32 of 45 cited sources checked · 10 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    Spi1 and Irf8 are core orchestrators for microglia development.

    “The transcription factors Spi1 and Irf8 have been shown to be the core orchestrators for microglia development”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Hoeffel et al., 2015 source

      The Introduction and Results describe the yolk-sac macrophage origin of microglia, but do not identify Spi1 or Irf8 as regulators of their development. Searches for Spi1, Irf8, PU.1, and IRF8 found no passage in our copy addressing that claim; page numbers were unavailable. (LLM, conf=0.75)

      Experimental — accuracy not yet measured
    • cites Kierdorf et al., 2013 source
  • Premisecitation

    Tgfbr2 and Csf1r play an essential signaling role in microglia development, with absence or inhibition causing marked reduction or complete absence of microglia.

    “Tgfbr2 and Csf1r have also been demonstrated to play an essential signaling role in microglia development, where the absence of these genes or inhibition of their signaling shows marked reduction or complete absence of microglia”
    Quote in the paperNot a substantial claim3 of 3 sources checked
    • cites Cohen et al., 2014 source

      The Results section on long TGFβ1 exposure reports Tgfbr2 expression in adult microglia, but does not test its role in microglial development. The Discussion addresses inflammatory phenotype switching, not developmental loss of microglia. Searches for Csf1r, gene absence, inhibition, and microglial depletion found no passage supporting the claimed reduction or absence of microglia. (LLM, conf=0.75)

      Experimental — accuracy not yet measured
    • cites Cronk et al., 2018 source

      In Results (‘Partial microglia depletion leads to beMφ engraftment’; page unavailable), the paper calls Csf1r critical for microglia survival and reports that inducible Csf1r deletion caused an approximately 25% reduction in microglia. That supports a role for Csf1r and partial depletion, but the experiment does not establish a role in microglia development or complete absence of microglia. No passage found addresses Tgfbr2 (searched: Tgfbr2 and microglia development). (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites Elmore et al., 2014 source

      The Introduction (page not provided) says microglial development depends on CSF1R and that CSF1R-knockout mice are devoid of microglia. The Discussion reports that CSF1R inhibition eliminates virtually all microglia in the adult CNS. Our copy does not support the claim about Tgfbr2; searches for Tgfbr2 and transforming growth factor beta receptor 2 found no relevant passage. (LLM, conf=0.90)

      Experimental — accuracy not yet measured
  • Premisecitation

    Standard antibody-based cell type purification across species remains challenging because epitope and antibody differences and impurities restrict accurate cross-species analysis.

    “Standard antibody-based cell type purification across species remains challenging because epitope and antibody differences and impurities of the cell population captured with these antibodies across different species significantly restrict accurate cross-species analysis”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Giladi and Amit, 2018 source

      The supplied passage and further searches for antibody-based purification, epitope differences, and cross-species analysis return only reference URLs, with no substantive text or section/page information. Our accessible copy therefore cannot establish whether the paper makes the claimed point. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
  • Premisecitation

    Microglia-specific pathways such as Trem2-Tyrobp and ApoE are critical determinants in Alzheimer's disease and other neurodegenerative diseases.

    “microglia-specific pathways, such as Trem2-Tyrobp and ApoE, are critical determinants in AD and other neurodegenerative diseases”
    Quote in the paperSubstantial claim3 of 3 sources checked
    • cites Jansen et al., 2019 source

      The Results section (‘Gene-sets implicated in AD and AD-by-proxy’; page not supplied) reports suggestive microglial associations with AD risk, partly dependent on APOE, but says they did not survive the stringent Bonferroni correction. The paper identifies APOE as a strong AD risk locus, but the passages do not establish a TREM2–TYROBP pathway as a critical determinant or extend the finding to other neurodegenerative diseases. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites Lambert et al., 2013 source

      The Abstract (page not given) identifies APOE as a major genetic risk factor for Alzheimer’s disease and reports an AD-associated TREM2 variant. It does not establish a TREM2–TYROBP pathway, describe these pathways as microglia-specific, or show that they are critical determinants of other neurodegenerative diseases. (LLM, conf=0.87)

      Experimental — accuracy not yet measured
    • cites Keren-Shaul et al., 2017 source

      The Results and Discussion (pages unavailable) identify Trem2–Tyrobp signaling and Apoe expression in disease-associated microglia, and show that full activation of these cells in an Alzheimer’s mouse model depends on Trem2. The Discussion also reports a similar microglial population in an ALS mouse model, but does not establish Trem2–Tyrobp and Apoe as critical determinants across other neurodegenerative diseases. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
  • Premisecitation

    CSF1R mutations reduce microglia numbers in the human brain.

    “CSF1R mutations also reduce microglia numbers in the human brain”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Colonna and Butovsky, 2017 source

      The Development of CNS Mononuclear Phagocytes section (page not provided) says genetic defects in CSF1R drastically reduce microglia numbers. It does not establish that this reduction occurs in the human brain; the cited studies on CSF1R defects are in mice. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
  • Classified as probably not a substantial claim. 36
  • May combine two claims. 36
  • The quote only partly states this claim. 1

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Microglia are the primary resident immune cells of the CNS parenchyma and are seeded during development from mesodermal progenitors.

    “Microglia are the primary resident immune cells of the CNS parenchyma and are seeded during development from mesodermal progenitors”
    Quote in the paperNot a substantial claim1 of 3 sources checked
    • cites Alliot et al., 1999 source
    • cites Ginhoux et al., 2010 source
    • cites Kierdorf et al., 2013 source

  • May combine two claims.

    Premisecitation

    Yolk-sac-derived microglia marker genes are highly enriched in microglia and are not upregulated in monocytes that engraft the brain in microglia-depleted mouse models.

    “These genes are highly enriched in microglia and are not upregulated in monocytes that engraft the brain in microglia-depleted mouse models”
    Quote in the paperSubstantial claim3 of 3 sources checked
    • cites Bennett et al., 2018 source

      The Results section on ontogeny markers reports that Gpr56 is expressed by yolk-sac-derived cells and microglia but not by engrafted HSC-lineage cells. However, the Results section on CNS-induced gene expression says engrafted cells of diverse origins express many other microglial signature genes at near-microglial levels. The claim is therefore supported for a specific ontogeny marker, not for microglial marker genes generally; page numbers are unavailable in our copy. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites Cronk et al., 2018 source
    • cites Shemer et al., 2018 source

      The Results section (‘Engrafted cells have transcriptomes distinct from microglia’; page not provided) reports that several microglia signature genes are reduced or absent in engrafted cells. But it also reports expression of other signature genes, including Fcrls and Tgfbr, in those cells. The claim therefore overgeneralizes when it says the marker genes are not upregulated in engrafted monocytes. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    These cells are implicated in sterile inflammation, wound healing, and age-related processes, including neurodegeneration.

    “These cells are implicated in sterile inflammation, wound healing, and age-related processes, including neurodegeneration”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Bussian et al., 2018 source

      The paper’s opening section (page not provided) links senescent cells to age-related tissue deterioration and reports a role for senescent glial cells in a mouse model of tau-dependent neurodegeneration. It does not address wound healing or implicate these cells in sterile inflammation; the Methods describe experimentally induced inflammatory responses, which do not establish that part of the claim. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim.

    Premisecitation

    scRNA-seq has been used to deconvolve immune cell type heterogeneity by identifying novel distinct immune cell subsets in health and disease.

    “scRNA-seq has been used to deconvolve immune cell type heterogeneity by identifying novel distinct immune cell subsets in health and disease”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites Jaitin et al., 2014 source

      The indexed text (section labeled Abstract; page unavailable) reports that single-cell RNA-seq distinguished immune-cell populations in mouse spleen and revealed new transcriptional subpopulations after LPS exposure. But LPS was used to mimic infection; the paper does not report studying disease, and it cautions that transcriptional subpopulations need not be distinct, committed cell subtypes. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites Keren-Shaul et al., 2017 source
    • cites Papalexi and Satija, 2018 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Adgrg1 (Gpr56) is conserved in human, mouse, and zebrafish and is associated with development of oligodendrocytes.

    “Among them is Adgrg1 (Gpr56), demonstrated previously to be conserved in human, mouse, and zebrafish, which is associated with development of oligodendrocytes”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Giera et al., 2018 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Several lysosomal hydrolases are ubiquitously expressed in macrophages, but others have more restricted tissue expression, and absence of some (e.g., Ctsa and Hexa) can cause severe neurological phenotypes.

    “Several of them are ubiquitously expressed in macrophages (e.g., Ctsb, Ctsh, and Ctsc), but others have more restricted tissue expression, and their absence (e.g., Ctsa and Hexa) can cause severe neurological phenotypes”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Caciotti et al., 2013 source

      The Background (page unavailable) says CTSA mutations cause galactosialidosis and describes psychomotor delay and neurological deterioration among its manifestations, supporting the CTSA example. The cited paper does not establish the claimed macrophage expression patterns or tissue restrictions, and the passages retrieved do not address Hexa. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    The inflammatory signature of potentially senescent-like microglia has been identified in most tissues of all ages in human and mice, but their number increases with age.

    “This inflammatory signature of potentially senescent-like microglia has been identified in most tissues of all ages in human and mice, but their number increases with age”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites He and Sharpless, 2017 source

      The Introduction describes pro-inflammatory cytokine secretion by senescent cells, and the Abstract says senescent cells occur throughout life and accumulate with age. The Wound Healing section discusses macrophages with some senescence features, but our copy does not identify an inflammatory signature in microglia across most tissues, ages, and humans and mice, or show that the number of those microglia increases with age. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Tardbp is protective against synapse loss in an Alzheimer's disease mouse model, whereas its absence promotes higher amyloid clearance by enhancing phagocytosis.

    “Tardbp has been shown recently to be protective against synapse loss in an AD mouse model, whereas its absence promotes higher amyloid clearance by enhancing phagocytosis”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Paolicelli et al., 2017 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Inflammatory cytokines have been seen previously in a sub-set of microglia in aged C57BL/6 mice.

    “inflammatory cytokines have been seen previously in a sub-set of microglia in aged C57BL/6 mice”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites Hammond et al., 2019 source
    • cites Sierra et al., 2007 source
    • cites Mrdjen et al., 2018 source

      The Aging and Neurodegenerative Disease section describes a reactive subset of microglia in geriatric mice, and Figure 5 (p. 389) identifies the mice as C57BL/6. The reported features are surface and activation markers, not inflammatory cytokines detected in those microglia. The Discussion mentions cytokines as possible environmental cues, which does not establish the claimed observation. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Vsir expression has been shown to be increased in several neurological diseases, such as Alzheimer's disease.

    “Its expression has been shown to be increased in several neurological diseases, such as AD”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Borggrewe et al., 2018 source

      The Results and Discussion report elevated VISTA protein expression in microglia near plaques in one Alzheimer’s disease patient. They also report elevated expression in normal-appearing white matter from an MS patient, but reduced expression in chronic MS lesions. The evidence therefore supports the AD example, not a general increase across several neurological diseases; it measures VISTA protein rather than directly establishing increased Vsir gene expression. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The level of spurious UMIs in the data was estimated using statistics on empty MARS-seq wells as previously described.

    “We estimated the level of spurious UMIs in the data using statistics on empty MARS-seq wells as previously described (Jaitin et al., 2014).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Jaitin et al., 2014 source

      The indexed text describes MARS-Seq’s molecular barcoding, molecule counting, and filtering in the section labeled Abstract, but does not describe estimating spurious UMIs from empty wells. Searches for “empty MARS-seq wells spurious UMIs” and “empty wells background noise molecular barcodes false molecules” found no passage addressing that procedure. (LLM, conf=0.72)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    DNA repair pathways have been implicated previously in longevity.

    “which have been implicated previously in longevity”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Ungvari et al., 2008 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Previous reports show high constraints of brain gene expression variance across evolution.

    “This is in accordance with previous reports showing high constraints of brain gene expression variance across evolution”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites Chan et al., 2009 source
    • cites Brawand et al., 2011 source
    • cites Chen et al., 2019 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Genes expressed by microglia have been implicated in Alzheimer's disease in human GWASs.

    “Previously, genes expressed by microglia have been implicated in AD in human GWASs”
    Quote in the paperNot a substantial claim3 of 3 sources checked
    • cites Gandal et al., 2018 source

      The indexed text describes a microglial gene module associated with autism and reports GWAS enrichment for other psychiatric disorders, but it does not link microglia-expressed genes to Alzheimer’s disease in human GWASs. Searches for “Alzheimer,” “AD GWAS,” and microglial Alzheimer’s risk variants found only an Alzheimer’s-related reference, not a passage making the claimed point; page numbers were unavailable. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
    • cites Jansen et al., 2019 source
    • cites Keren-Shaul et al., 2017 source

      The Discussion says that genes expressed by disease-associated microglia include known AD risk-factor genes, and it describes human genetic variants associated with disease risk. It does not say that those genes were implicated by human GWASs, so the specific study type in the claim is unsupported. (LLM, conf=0.84)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Msr1 mediates endocytosis of low-density lipoproteins and is involved in uptake and degradation of amyloid beta.

    “Msr1 mediates endocytosis of low-density lipoproteins (LDLs) and is involved in uptake and degradation of amyloid b (Ab)”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Frenkel et al., 2013 source

      The Results section ('Scara1 deficiency is associated with reduced sAβ uptake'; page unavailable) identifies Scara1 as a major receptor for soluble amyloid-β uptake, and another Results section links reduced uptake to reduced amyloid-β degradation. The Discussion says Scara1 mediates adhesion to oxidized-LDL-coated surfaces, but does not establish that it mediates LDL endocytosis. (LLM, conf=0.87)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data

    Human microglia display significant heterogeneity and are organized into several microglia types.

    “whereas human microglia display significant heterogeneity”
    • Also stated in: § Single-Cell Transcriptomic Analysis Identifies Human Microglia Subsets · plain
      “Cell-to-cell correlation analysis revealed that microglia from all human individuals are organized into several microglia types”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    There is a considerably smaller number of differentially expressed genes between human and macaque microglia, with most related to metabolic pathways such as NADK and BCO2.

    “we observed a considerably smaller number of differentially expressed genes between human and macaque microglia, with most of the differentially expressed genes related to metabolic pathways such as NAD kinase (NADK; soluble vitamin metabolism) and beta-carotene oxygenase 2 (BCO2; carotenoid oxidization)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The highest microglial density was detected in leech ganglia (299.0 ± 41.4 Iba1+ microglia/mm2) and the lowest in axolotls (11.1 ± 0.7/mm2).

    “The highest microglial density was detected in leech ganglia with 299.0 ± 41.4 Iba1 + microglia/mm 2 and lowest microglial numbers in axolotls (11.1 ± 0.7/mm 2 )”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingnone given

    Although microglia display similar dendritic morphology, the degree of ramification and cell size differs considerably in many species examined.

    “although microglia display similar dendritic morphology, the degree of ramification and cell size differs considerably in many of the species examined”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The 17 clusters include a microglia core signature expressed in all species (clusters 1-10), but to a lesser extent in zebrafish and chicken, as well as clade- and species-specific programs (clusters 11-17).

    “These clusters include a microglia core signature expressed in all species (clusters 1-10), but to a lesser extent in the zebrafish and chicken, as well as clade and species-specific programs (clusters 11-17”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Putative senescent-like microglia are not specific to a single individual but are consistent across all six sampled individuals, comprising roughly 20% of all microglia.

    “Further investigation of these putative senescent-like microglia revealed that they were not specific to a single individual but consistent across all six sampled individuals, comprising roughly 20% of all microglia”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    There is no detectable correlation between microglial and neuronal density or between microglia/neuron ratio and microglia process length.

    “we did not observe any detectable correlation between microglial and neuronal density (Figure S1E) or microglia/neuron ratio and microglia process length”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Human microglia heterogeneity is in sharp contrast to mouse, macaque, marmoset, hamster, and sheep, which show high intra-cluster correlation and predominantly one dominant microglia type.

    “The human microglia heterogeneity identified was in sharp contrast to mouse (Figure 4C), macaque (Figure S5A), marmoset (Figure S5B), hamster (Figure S5C), and sheep (Figure S5D), which show high intra-cluster correlation and predominantly one dominant microglia type.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Most species contained some metacells that were identified as contaminating cells.

    “most species contained some metacells that were identified as contaminating cells”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Hierarchical clustering grouped species based on evolutionary distance and displayed two major groups, with zebrafish and chicken microglia forming an outgroup from mammals.

    “Hierarchical clustering across samples grouped the species based on their evolutionary distance and displayed two major groups, with the zebrafish and chicken microglia forming an outgroup from the mammals”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    This microglia subset consistently co-expressed CDKN1, CCL3, CCL4, CCL3L3, and CCL4L2, along with higher expression of inflammatory cytokines TNF and IL1B.

    “This microglia subset consistently co-expressed CDKN1, CCL3, CCL4, CCL3L3, and CCL4L2 (Figure 4F), along with higher expression of inflammatory cytokines such as TNF and IL1B”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    A sub-population of human microglia has increased expression of several inflammatory genes linked with a senescence-associated secretory phenotype, without any change of homeostatic gene expression.

    “we observed a sub-population with increased expression of several inflammatory genes that have been linked with a senescence-associated secretory phenotype (SASP) (von Kobbe, 2018; Figure 4D), without any change of homeostatic gene expression”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Macaque showed the highest similarity in expression patterns to human microglia, whereas laboratory mouse strains clustered together with wild mice and hamster.

    “Macaque showed the highest similarity in expression patterns to human microglia, whereas laboratory mouse strains clustered together with wild mice and hamster”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Grn (progranulin) is highly expressed and conserved in microglia.

    “Another highly expressed and conserved lysosomal gene, Grn (or progranulin)”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The microglia expression program of all mouse strains showed high similarity to each other, and only a few genes were differential between mouse strains.

    “the microglia expression program of all mouse strains showed high similarity to each other (Figure S5F), and only a few genes were identified as differential between the mouse strains”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Human microglia subtypes are unrelated to sex differences and are observed in both male and female microglia.

    “These subtypes are unrelated to sex differences and observed in both male and female microglia”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    P2Y12 and PU.1 are conserved across microglia evolution.

    “Conservation of P2Y12 and PU.1 across Microglia Evolution”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Enriched pathways included phagocytosis, more specifically apoptotic cell clearance, and the complement pathway.

    “Additionally, enriched pathways included phagocytosis (Figure S6E)-more specifically, apoptotic cell clearance (Figure S6F; Table S6)-and the complement pathway”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    A group of 26 genes (e.g., Cdkn1, Ccl2, Tnf) exhibited a strong Pearson correlation across the Metacells' log2 footprint expression of the 200 most variable genes excluding genes associated with mitochondria and stress.

    “identified a group of 26 genes (e.g., Cdkn1, Ccl2, Tnf) that exhibited a strong Pearson correlation across the Metacells' log2 footprint expression of the 200 most variable genes excluding genes associated with mitochondria and stress that were filtered from these lists in advance.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Human brain tissue from epilepsy surgery patients' healthy regions is representative of normal human microglia.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    Significant biological processes or pathways were identified using Metascape software analysis of all genes from clusters 1-3 using all genes as a background or microglia specific genes identified in a comparative analysis of tissue macrophages and microglia in Figure 3D, using all genes as a background.

    “Significant biological processes or pathways identified using Metascape software analysis of all genes from clusters 1-3 using all genes as a background (F) or microglia specific genes identified in a comparative analysis of tissue macrophages and microglia in Figure 3D, using all genes as a background (G).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Human brain tissue from epilepsy surgery patients' healthy regions is representative of normal human microglia.

  • May combine two claims.

    Findingown data

    Several pathways were enriched in all tested mammals except mice; in particular, negative regulation of the ferroptosis pathway.

    “Several other pathways were enriched in all tested mammals except mice; in particular, negative regulation of the ferroptosis pathway”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Human microglia express the great majority of genes related to susceptibility to Alzheimer's disease and Parkinson's disease, but not Huntington's disease or schizophrenia, whereas rodent microglia only express a fraction of these genes.

    “Our analysis additionally revealed that human microglia express the great majority of genes related to susceptibility to AD and PD (but not to Huntington's disease or schizophrenia), whereas rodent microglia only express a fraction of these genes.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    RNA-seq expression of representative highly conserved and highly expressed genes was identified from clusters 1-3.

    “RNA-seq expression of representative highly conserved and highly expressed genes identified from clusters 1-3 (as shown in Figure 3A).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Human brain tissue from epilepsy surgery patients' healthy regions is representative of normal human microglia.
    • Rests on an assumption: Orthologous genes across the compared species are functionally comparable for cross-species microglia transcriptome comparison.

  • May combine two claims.

    Findingown data

    There are species-specific and disease-associated pathways in microglia.

    “Characterization of Microglia Species-Specific and Disease-Associated Pathways”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    There is significant over-representation of Parkinson's disease genes and, to a lesser extent, Alzheimer's disease genes in the human-specific cluster and clusters shared by human and macaques.

    “We observed significant over-representation of PD genes and, to a lesser extent, AD genes in the human-specific cluster and clusters shared by human and macaques”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    The clustering result was not affected by the number of clusters or analysis method, and similar results were observed with PCA.

    “This was not affected by the number of clusters or analysis method, and similar results were observed when applying principal-component analysis (PCA”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Laboratory mouse models have been described to show considerable differences regarding immune system regulation.

    “Laboratory mouse models have been described to show considerable differences regarding the immune system regulation”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Sellers et al., 2012 source

  • May combine two claims.

    Interpretationown data

    Single-cell transcriptomic analysis identifies major microglia heterogeneity in humans, in contrast to other mammals, which mostly display a single microglia type in the steady state under non-pathological conditions.

    “single-cell transcriptomic analysis identifies major microglia heterogeneity in humans, in contrast to other mammals, which mostly display a single microglia type in the steady state under non-pathological conditions”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Vsir (VISTA) is an immune checkpoint gene that inhibits T cell response.

    “Vsir (or VISTA) is an immune checkpoint gene that inhibits T cell response”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Xu et al., 2018

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Grn has been implicated in frontotemporal dementia in humans and mice.

    “has been implicated in frontotemporal dementia in humans and mice”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Baker et al., 2006 source

  • May combine two claims.

    Interpretationown data

    scRNA-seq can define, with minimal biases, conserved cell type transcriptional signatures across species, enabling marker-free molecular comparison of the microglia program across evolution.

    “Together, these findings highlight the potential of scRNA-seq to define, with minimal biases, conserved cell type transcriptional signatures across species, enabling marker-free molecular comparison of the microglia program across evolution.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Cells with high (> 32 UMI) expression of hemoglobin genes were discarded.

    “Cells with high (> 32 UMI) expression of hemoglobin genes were discarded (e.g., Hbb, Hba1, Hba2).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    All experiments were conducted in accordance with the institutional review board of the University of Freiburg Medical Center.

    “All experiments were conducted in accordance with the institutional review board of the University of Freiburg Medical Center.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    In post mortem sample collection, samples were included in the study only if no indication of a tumor or inflammatory disease was examined by histological analysis.

    “In post mortem sample collection, samples were included in the study only if no indication of a tumor or inflammatory disease was examined by histological analysis.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Reads with multiple mapping positions were excluded.

    “reads with multiple mapping positions were excluded.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    Expression libraries were produced with a minimum of three replicates per population.

    “developed for scRNA-seq to produce expression libraries with a minimum of three replicates per population.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Keren-Shaul et al., 2019 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    ERCC genes were discarded.

    “ERCC genes were discarded.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    It cannot be excluded that patients' microglia are not accurately recapitulating the in situ states of healthy brains.

    “it cannot be excluded that patients' microglia are not accurately recapitulating the in situ states of healthy brains.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    All experimental procedures were either approved by a local ethical review committee and conducted in accordance with personal and project licenses under the UK Animals (Scientific Procedures) Act (1986) or performed in accordance with institutional animal welfare guidelines and were approved by the government of Baden-Wurttemberg, Germany.

    “All experimental procedures were either approved by a local ethical review committee and conducted in accordance with personal and project licenses under the UK Animals (Scientific Procedures) Act (1986) or performed in accordance with institutional animal welfare guidelines and were approved by the government of Baden-Wurttemberg, Germany.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Cells with less than 500 total UMI were discarded as low-quality cells.

    “Cells with less than 500 total UMI were discarded as low-quality cells.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Sample acquisition and processing time were carefully calibrated.

    “we took great care and carefully calibrated sample acquisition and processing time (STAR Methods)”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Bootstrapping was used to derive robust clustering (500 iterations; resampling 70% of the cells in each iteration, and clustering of the co-cluster matrix with minimal cluster size set to 20).

    “We used bootstrapping to derive robust clustering (500 iterations; resampling 70% of the cells in each iteration, and clustering of the co-cluster matrix with minimal cluster size set to 20).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). The quote only partly states this claim.

    Study design and conductown data

    Absence of pathology in the human epilepsy surgery samples was confirmed by histological analysis in the respective area by a neuropathologist.

    “Absence of pathology in the sample was confirmed by histological analysis in the respective area by a neuropathologist.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Many hallmarks of human neurodegeneration specifically involve increased presence of double-strand DNA damage and senescent cells as well as accumulation of iron and peroxidation of polyunsaturated fatty acids in CNS cells.

    “Many hallmarks of human neurodegeneration specifically involve an increased presence of double-strand DNA damage and senescent cells as well as accumulation of iron and peroxidation of polyunsaturated fatty acids in CNS cells”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Fielder et al., 2017 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Mitochondrial genes and ribosomal genes were removed from the list of features used for the Metacell analysis.

    “Mitochondrial genes and ribosomal genes were removed from the list of features used for the Metacell analysis.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Human brain tissue was collected from healthy brain regions of individuals that underwent epilepsy surgery.

    “Human brain tissue was collected from healthy brain regions of individuals that underwent epilepsy surgery.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    For each organism, Metacell analysis was used to discard outlier cells.

    “For each organism, we used Metacell analysis to discard outlier cells.”
    Quote in the paperNot a substantial claim
62 of 62

  • The paper relies on this without support.

    Large differences exist in multiple characteristics of disease manifestation between human and animal models.

    “Despite immense efforts over the past decades to find a cure for neurodegenerative disorders using transgenic animal models, large differences are observed in multiple characteristics of disease manifestation in human versus animal models.”

  • The paper relies on this without support.

    Effective development of therapies for neurodegenerative diseases is limited.

    “Effective development of therapies for neurodegenerative diseases is limited.”

  • The paper relies on this without support.

    Bulk RNA-seq data can contain a mixture of cell types due to sample acquisition impurity and existence of outlier and/or 'contaminating' cells.

    “Bulk RNA-seq data can contain a mixture of cell types due to sample acquisition impurity and existence of outlier and/or ''contaminating'' cells.”

  • The paper relies on this without support.

    Comparing bulk RNA-seq data between species is especially challenging due to differences in the total number of genes, total read number, housekeeping genes expression profiles, and possible other factors that affect conventional RNA-seq normalization methods.

    “Comparing bulk RNA-seq data between species is especially challenging due to differences in the total number of genes, total read number, housekeeping genes expression profiles, and possible other factors that affect conventional RNA-seq normalization methods.”

  • The paper relies on this without support.

    Human brain tissue from epilepsy surgery patients' healthy regions is representative of normal human microglia.

    “Human brain tissue was collected from healthy brain regions of individuals that underwent epilepsy surgery.”
    “Absence of pathology in the sample was confirmed by histological analysis in the respective area by a neuropathologist.”

  • The paper relies on this without support.

    Post mortem samples with no indication of tumor or inflammatory disease are representative of normal human microglia.

    “In post mortem sample collection, samples were included in the study only if no indication of a tumor or inflammatory disease was examined by histological analysis.”

  • The paper relies on this without support.

    Orthologous genes across the compared species are functionally comparable for cross-species microglia transcriptome comparison.

    “All genes from microglia core gene module (Figure 3A) were plotted unless an orthologous gene did not exist in one of the two compared species.”
    “Genes with no orthologs are marked with a diagonally striped bar.”
7 of 7

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Complement and apoptotic-clearance Figure S6 references are reversed

    Minor

    The results cite S6F for apoptotic cell clearance and S6G for the complement pathway. The supplementary legend assigns S6F to complement and S6G to apoptotic cell clearance.

  • PCA methods point to the wrong supplementary figure

    Minor

    The methods place the PCA in Figure S4, whereas the supplementary legend identifies the PCA plot as Figure S3B; Figure S4 is described as histology of P2Y12 and PU.1.

  • Macaque and marmoset Figure S5 references are reversed

    Minor

    The results assign macaque to Figure S5A and marmoset to Figure S5B, but the supplementary legend assigns S5A to marmoset and S5B to macaque.

  • Non-human-primate CD45 antibody has conflicting catalogue numbers

    Minor

    The flow-cytometry antibody table and key resources table list the same non-human-primate CD45-PE antibody from Miltenyi Biotec with different catalogue numbers.

  • Single-cell comparison is incorrectly cited as Figure S4A

    Minor

    The methods cite Figure S4A for a single-cell expression comparison, but the supplementary legend describes S4A as P2Y12 histology. It describes the cross-species single-cell expression heatmap under Figure S3A.

5 of 5

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Geirsdottir, L., David, E., Keren-Shaul, H., Weiner, A., Bohlen, S. C., Neuber, J., Balic, A., Giladi, A., Sheban, F., Dutertre, C.-A., Pfeifle, C., Peri, F., Raffo-Romero, A., Vizioli, J., Matiasek, K., Scheiwe, C., Meckel, S., Mätz-Rensing, K., van der Meer, F., Thormodsson, F. R., Stadelmann, C., Zilkha, N., Kimchi, T., Ginhoux, F., Ulitsky, I., Erny, D., Amit, I., & Prinz, M.
Affiliations
  • Weizmann Institute of Science (IL)
  • University of Freiburg (DE)
  • Roslin Institute (GB)
  • Agency for Science, Technology and Research (SG)
  • Max Planck Institute for Evolutionary Biology (DE)
  • University of Zurich (CH)
  • Inserm (FR)
  • Ludwig-Maximilians-Universität München (DE)
  • Innovation Center Iceland (IS)
  • Universitätsmedizin Göttingen (DE)
Corresponding author
  • Erny, Daniel
  • Amit, Ido
  • Prinz, Marco
ORCID
Venue
Cell
Publisher
Elsevier BV
Published
2019-12
Volume
179
Issue
7
Pages
1609-1622.e16
ISSN
0092-8674 (ISSN-L)
DOI
10.1016/j.cell.2019.11.010
OpenAlex
W2996729595
Semantic Scholar
f1088abd77215a8a0ff7c2586e1119fc5d78bde5
Cited by
461
References
93
Citations by year
2026: 49 · 2025: 48 · 2024: 78 · 2023: 62 · 2022: 92 · 2021: 79 · 2020: 53
FWCI
20.71
Type
journal article
Language
en
Update notices
erratum (2020-04-30)
Open access
Licence
Locations
  • Cell (publishedVersion)
  • PubMed (publishedVersion)
  • Edinburgh Research Explorer (University of Edinburgh) (submittedVersion, cc-by-nc-nd)
  • HAL (Le Centre pour la Communication Scientifique Directe) (submittedVersion)
  • GoeScholar The Publication Server of the Georg-August-Universität Göttingen (Georg-August-Universität Göttingen) (submittedVersion, cc-by)
  • MPG.PuRe (Max Planck Society) (submittedVersion, other-oa)
  • Zurich Open Repository and Archive (University of Zurich) (submittedVersion)
Funders
  • Chan Zuckerberg Initiative
  • HHMI
  • European Research Council (ERC-COG 724471-HemTree2.0)
  • MRA (509044)
  • DFG (SFB/TRR167, SFB 992, SFB1160, Reinhart Koselleck, 390939984)
  • Ernest and Bonnie Beutler Research Program for Excellence in Genomic Medicine
  • Helen and Martin Kimmel
  • Wolfson Foundation and Family Charitable Trust
  • Thompson Family Foundation Alzheimer’s Research Fund
  • Adelis Foundation
  • Eden and Steven Romick Post-Doctoral Fellowship Fund
  • International Progressive MS Alliance (PA-1604-08459)
  • Berta Ottenstein Programme for Clinician Scientists
  • Sobek Foundation
  • Ernst-Jung Foundation
  • Ministry of Science, Research and Arts, Baden-Wuerttemberg (Sonderlinie “Neuroinflammation”)
  • ISF (1324/15)
  • Minerva Foundation
  • Biotechnology and Biological Sciences Research Council of the United Kingdom
  • Institute Strategic Programme (BBS/E/D/10002071)
  • EMBO
  • Singapore Immunology Network
  • Singapore National Research Foundation (NRF2016NRF-NRFI001-02)
Topics
  • Life Sciences › Neuroscience › Neurology › Neuroinflammation and Neurodegeneration Mechanisms
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
  • Life Sciences › Neuroscience › Developmental Neuroscience › Neurogenesis and neuroplasticity mechanisms
Keywords
microglia, primate evolution, single-cell transcriptomics, microglial heterogeneity, microglial morphology, microglia-neuron interaction, complement system, phagocytosis, Alzheimer's disease, Parkinson's disease, cross-species comparison
MeSH terms
Animals, Chickens, Humans, Primates, Reptiles, Rodentia, Sheep, Swine, Zebrafish, Microglia, Evolution, Molecular, Neurodegenerative Diseases, Genetic Predisposition to Disease, Gene Expression Profiling, Gene Regulatory Networks, Single-Cell Analysis, Transcriptome

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-14.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-14.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review