CoChat
Sign Up
Cited by 1,129Neuroinflammation and Neurodegeneration MechanismsSingle-cell and spatial transcriptomicsNeurogenesis and neuroplasticity mechanisms

Developmental Heterogeneity of Microglia and Brain Myeloid Cells Revealed by Deep Single-Cell RNA Sequencing

Open access
Li, Qingyun, Cheng, Zuolin, Zhou, Lu, Darmanis, Spyros, Neff, Norma F., Okamoto, Jennifer, Gulati, Gunsagar, Bennett, Mariko L. +7 more·2019·Neuron
DOI 10.1016/j.neuron.2018.12.006PMC PMC6336504

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Microglia are increasingly recognized for their major contributions during brain development and neurodegenerative disease. It is currently unknown whether these functions are carried out by subsets of microglia during different stages of development and adulthood or within specific brain regions. Here, we performed deep single-cell RNA sequencing (scRNA-seq) of microglia and related myeloid cells sorted from various regions of embryonic, early postnatal, and adult mouse brains. We found that the majority of adult microglia expressing homeostatic genes are remarkably similar in transcriptomes, regardless of brain region. By contrast, early postnatal microglia are more heterogeneous. We discovered a proliferative-region-associated microglia (PAM) subset, mainly found in developing white matter, that shares a characteristic gene signature with degenerative disease-associated microglia (DAM). Such PAM have amoeboid morphology, are metabolically active, and phagocytose newly formed oligodendrocytes. This scRNA-seq atlas will be a valuable resource for dissecting innate immune functions in health and disease.

AI snapshot

Written from the paper’s full text.

TL;DR

Deep single-cell RNA sequencing of 1,816 mouse brain myeloid cells across embryonic, early postnatal, and adult stages shows that adult homeostatic microglia are transcriptionally similar across brain regions, whereas early postnatal microglia are more heterogeneous. The authors identify proliferative region-associated microglia (PAM), a transient white-matter subset that shares a gene signature with degenerative disease-associated microglia, is TREM2/APOE-independent, and phagocytoses newly formed oligodendrocytes.

Key findings

  • Adult homeostatic microglia show limited regional transcriptomic heterogeneity: fewer than 20 genes were differentially expressed in cluster 0 adult microglia between regions (FDR < 0.05), and bulk RNA-seq of TMEM119+ microglia from cortex, cerebellum, hippocampus, and striatum was highly correlated (R > 0.99) with no differentially expressed genes between regions (FDR < 0.05).
  • Early postnatal microglia are more heterogeneous; after cell cycle regression, P7 microglia resolved into three clusters, including P7-C1/PAM, which expressed DAM-associated genes such as Spp1, Gpnmb, Clec7a, Lpl, Cd9, Cd63, Lgals3, Fabp5, Itgax, Apoe, and Tyrobp.
  • PAM appeared mainly in developing corpus callosum and cerebellar white matter, peaked around P7, were amoeboid, and phagocytosed newly formed oligodendrocytes; PAM were still present in P7 Trem2−/− or Apoe−/− white matter, unlike DAM.
  • Cell cycle pseudotime identified 315 periodically expressed genes for P7-C0 and 347 for P7-C1, with 183 overlapping genes; many newly identified genes were not previously described as periodically expressed.

Methods

  • Smart-seq2 scRNA-seq was performed on index-sorted innate immune cells from mouse brain at E14.5, P7, and P60; P7 and P60 cells were isolated from cortex, cerebellum, hippocampus, striatum, olfactory bulb, and choroid plexus. In total, 1,922 cells were sequenced to at least 1 million raw reads per cell, and 1,816 cells (94.5%) passed QC: 142 E14.5, 978 P7, and 696 P60 cells.
  • Unsupervised Seurat clustering yielded 15 clusters, including 7 microglia clusters; analyses included differential expression, cell cycle regression, pseudotime reconstruction, gene network construction, and bulk RNA-seq of TMEM119+ adult microglia.
  • Validation used RNAscope, immunohistochemistry, FACS, a pH-sensitive bead phagocytosis assay on acute brain slices, and Trem2−/− and Apoe−/− mice.

Limitations

  • The paper states that additional brain regions beyond the 4 major regions tested may uncover transcriptomic differences, and that microglial heterogeneity may also be represented by electrophysiological, epigenetic, or functional measurements; future work is needed to distinguish whether PAM are transcriptionally pre-disposed to phagocytosis or represent a transient phenotypic conversion.
  • Evident from the text: all experiments used male mice; the IEG+ microglia cluster is likely an experimental artifact from cell extraction/sorting, since Fos/Egr1 were not detected in adult microglia on tissue sections.
  • Evident from the text: PAM phagocytosis evidence is largely histological and correlative, although the slice phagocytosis assay supports preferential phagocytic capacity in corpus callosum and cerebellar white matter.

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

Integrity signals

Registry

Crossref · 10.1016/j.neuron.2018.12.006

Cites

54 of 54 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 39 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Early postnatal microglia are distributed among clusters 0-5.

    “early postnatal microglia, which were distributed among clusters 0–5”
    Quote in the paperNot a substantial claim
  • Findingown data

    Cell cycle phase-specific gene sets for microglia were reconstructed, identifying 315 periodically expressed genes for P7-C0 and 347 for P7-C1.

    “In total, we identified 315 periodically expressed genes for P7-C0 cells, and 347 genes for P7-C1 cells”
    Quote in the paperSubstantial claim
  • Findingown data

    Choroid plexus macrophages separate into two distinct clusters that are largely confined to developmental stages.

    “CP MΦ were separated into two distinct clusters (cluster 7 and cluster 9) largely confined to their developmental stages”
    Quote in the paperSubstantial claim
  • Findingown data

    P7-C2 is a small cluster mainly containing cells that clustered with embryonic E14.5 microglia.

    “P7-C2 was a small cluster that mainly included cells clustered together with embryonic E14.5 microglia (cluster 5) in the initial analysis”
    Quote in the paperSubstantial claim
  • Findingown data

    Early postnatal microglia display greater transcriptomic heterogeneity than adult microglia.

    “By contrast, early postnatal microglia are more heterogeneous.”
    • Also stated in: § Introduction · plain
      “we observed much higher heterogeneity in early postnatal microglia.”
    Quote in the paperSubstantial claim
  • May combine two claims. 47
  • The quote only partly states this claim. 1
  • Classified as probably not a substantial claim. 24

  • May combine two claims.

    Findingown data

    A gene network analysis identified three interconnected gene modules explaining P7 microglia identities, with many DAM signature genes at hub positions.

    “We found that three interconnected gene modules explained P7 microglia identities and, more importantly, many DAM signature genes occupied the “hub” positions in the network.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Many genes not previously described as periodically expressed in dividing cells were identified, and more than half have possible links to cell proliferation.

    “we identified many genes that have yet to be described as periodically expressed in dividing cells (“Novel”), although more than half of these new genes have possible links to cell proliferation based on gene function annotations”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    P7-C1 microglia have higher Igf1 and Itgax expression, with co-expression in Gpnmb-positive cells.

    “scRNA-seq data suggested that this P7-C1 population also had higher levels of Igf1 and Itgax expression. Indeed, we detected co-expression of both transcripts in Gpnmb-positive cells”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    H2-Eb1 and Lilra5 are markers for adult choroid plexus macrophages, and Clec4n is a marker for postnatal choroid plexus macrophages, validated by RNA in situ and immunohistochemistry.

    “We identified H2-Eb1 and Lilra5 as molecular markers for adult CP MΦ and Clec4n for their postnatal counterparts, and validated this selective gene expression by RNA in situ and immunohistochemistry”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    scRNA-seq of microglia sorted by GPNMB and CLEC7A co-expression validated the original P7-C1 gene expression pattern and identified a double-positive cluster expressing P7-C1 markers at higher levels.

    “scRNA-seq of microglia freshly isolated based on GPNMB and CLEC7A co-expression validated the original gene expression pattern observed with the initial P7 clustering results; the clustering algorithm also reported another GPNMB+CLEC7A+ concentrated cluster which overlapped with a portion of P7-C1 cells”
    • Also stated in: § Validation of early postnatal proliferative region-associated microglia · plain
      “This new double positive cluster expressed all the markers identified in P7-C1, just at higher levels on average”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    A subset of proliferative region-associated microglia (PAM) exists mainly in developing white matter and shares a gene signature with degenerative disease-associated microglia (DAM).

    “We discovered a proliferative region-associated microglia (PAM) subset, mainly found in developing white matter, that share a characteristic gene signature with degenerative disease-associated microglia (DAM).”
    • Also stated in: § Heterogeneity of early postnatal microglia revealed by scRNA-seq · plain
      “P7-C1 microglia exhibited characteristic expression of many genes that have recently been shown to be up-regulated in degenerative disease-associated microglia (DAM) (Keren-Shaul et al., 2017; Krasemann et al., 2017).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Keren-Shaul et al., 2017 source

      Not checked in the free run.

    • cites Krasemann et al., 2017 source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Spp1 was detected in close to 100% of double positive cells by scRNA-seq, while SPP1 labeled only a portion of CLEC7A+ cells on tissue sections.

    “Spp1 was detected in close to 100% of the double positive cells by scRNA-seq, and SPP1 only labeled a portion of CLEC7A+ cells on tissue sections”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Bulk RNA-seq of purified adult homeostatic microglia shows high correlation (R>0.99) across cortex, cerebellum, hippocampus, and striatum, with no differentially expressed genes between regions.

    “samples from cortex, cerebellum, hippocampus and striatum were highly correlated (R>0.99), and individual samples did not cluster according to tissue origins”
    • Also stated in: § Limited transcriptomic heterogeneity of adult homeostatic microglia · hedged
      “we could not detect any differentially expressed genes (FDR < 0.05) between regions from the bulk samples”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    PAM and periventricular myeloid cells specifically express CLEC7A protein.

    “early postnatal PAM and periventricular myeloid cells specifically expressed CLEC7A protein”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Cluster 6 differs from cluster 0 by 27 differentially expressed genes, about half of which are immediate early genes.

    “Compared with cluster 0, only 27 total genes (including P2ry12) were differentially expressed in cluster 6, out of which half were immediate early genes (IEG)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The 15 clusters were annotated as 7 microglia clusters, 2 choroid plexus macrophage clusters, 3 monocyte clusters, and 3 other immune or epithelial cell clusters.

    “This included 7 microglia clusters, 2 choroid plexus macrophage clusters, 3 monocyte clusters and 3 other immune or epithelial cell clusters”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The three cell-cycle-regressed P7 clusters largely correspond to the three dominant cell cycle-independent clusters, with the two original cycling clusters redistributed among them.

    “these three clusters largely corresponded to the three dominant cell cycle-independent clusters (cluster 1, 2 and 5) found in the original analysis, and the two original cycling clusters (cluster 3 and 4) were simply redistributed among P7-C0, P7-C1 and P7-C2 following cell cycle regression”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    P7-C1 microglia express many DAM-associated genes and down-regulate homeostatic genes.

    “P7-C1 microglia exhibited characteristic expression of many genes that have recently been shown to be up-regulated in degenerative disease-associated microglia (DAM)”
    • Also stated in: § Heterogeneity of early postnatal microglia revealed by scRNA-seq · plain
      “up-regulation of the disease genes was accompanied by down-regulation of the homeostatic gene cassette”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Keren-Shaul et al., 2017 source

      Not checked in the free run.

    • cites Krasemann et al., 2017 source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Spp1 and Gpnmb expression in P7 brain overlaps almost exclusively in corpus callosum and cerebellar white matter, and Spp1+Gpnmb+ cells are positive for Cx3cr1.

    “RNA in situ labeling demonstrated that expression of these two genes overlapped almost exclusively in the corpus callosum and cerebellar white matter (marked by Mbp+ oligodendrocytes), and that Spp1+Gpnmb+ cells were positive for the microglia marker Cx3cr1”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    P7-C2 cells express lower levels of microglial signature genes and up-regulate genes involved in metal homeostasis, actin cytoskeleton dynamics, and ribosomal components.

    “Cells belonging to P7-C2 expressed microglial signature genes at lower levels (e.g. Tmem119, Selplg, P2ry12, Tgfbr1), and displayed up-regulation of genes in metal homeostasis (Fth1, Ftl1, Mt1), actin cytoskeleton dynamics (Tmsb4x, Pfn1, Cfl1), and ribosomal components (e.g. Rps14, Rps18, Rpl35, Rps29)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Adult choroid plexus macrophages express high levels of MHC-II and immunoregulatory genes, while postnatal choroid plexus macrophages express genes involved in endocytosis and intracellular trafficking.

    “Only in the adult CP MΦ cluster (cluster 7) did we detect high levels of MHC-II gene expression”
    • Also stated in: § Postnatal transcriptional alteration of choroid plexus macrophages · plain
      “By contrast, postnatal CP MΦ (cluster 9) expressed many genes that function in endocytosis and intracellular trafficking”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Starting at P4, CLEC7A+ microglia populate corpus callosum and cerebellar white matter, expand and peak around P7, and are barely detectable by P14.

    “Starting at P4, CLEC7A+ microglia populated the corpus callosum and cerebellar white matter (Figure 7A), and they continued to expand and peaked around P7 (Figures 6E and 7B). By P14, these white matter PAM were barely detectable”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    CLEC7A+ microglia are transiently associated with developing white matter and found in regions of neurogenesis.

    “we discovered that CLEC7A+ microglia were associated transiently with the developing white matter, and found in regions of neurogenesis”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In acutely sectioned postnatal brain slices, preferential phagocytosis of pH-sensitive beads occurred only in corpus callosum and cerebellum white matter microglia (mostly CLEC7A+); this was independent of opsonizing factors or other serum components.

    “We found that only microglia in the corpus callosum and cerebellum white matter (mostly CLEC7A+) preferentially phagocytosed beads (Figures 8C, 8D, S7A, and S7B), and this phenomenon was independent of opsonizing factors or other components in serum (Figure 8D).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In scRNA-seq, Mbp and Gfap transcripts were detected in 9.9% and 1.5% of P7-C1 microglia and 13.5% and 8.9% of P7-GPNMB+CLEC7A+ microglia, respectively, while Pdgfra counts were zero in all P7 microglia.

    “From the scRNA-seq data, Mbp and Gfap transcripts were detected in certain percentages of P7-C1 (32/323=9.9% for Mbp; 5/323=1.5% for Gfap) and P7-GPNMB+CLEC7A+ microglia (26/192=13.5% for Mbp; 17/192=8.9% for Gfap), while Pdgfra counts were zero in all P7 microglia (Figure 8I).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Developmental pseudotime analysis of P7 and P60 microglia identified one trajectory toward PAM and another toward homeostatic mature cells.

    “We discovered that one trajectory pointed towards the development of early postnatal PAM, while the other pointed towards homeostatic, mature cells”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Early postnatal PAM do not represent a general intermediate stage of microglial development; they appear mainly transiently in developing white matter.

    “We showed here that early postnatal PAM are not a general intermediate step for microglia development, but mainly transiently appear in the developing white matter.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Early postnatal PAM display unique cytokine/chemokine profiles and express trophic factors including Igf1, Spp1, Lgals1, and Lgals3.

    “We showed that early postnatal PAM display unique cytokine/chemokine profiles and express many trophic factors including Igf1, Spp1, Lgals1, and Lgals3 (Figure S5D).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    At E17.5, CLEC7A+ microglia appear almost exclusively in neurogenic niches near lateral ventricles, while CLEC7A−IBA1+ cells are present throughout the brain parenchyma.

    “During the late embryonic (E17.5) stage, CLEC7A+ microglia appeared almost exclusively in the neurogenic niches near the lateral ventricles (Figure 7A), while CLEC7A−IBA1+ cells were present throughout the brain parenchyma.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    The appearance of PAM coincides with the onset of myelination, when many newly formed oligodendrocytes die.

    “The timing of their appearance coincides with the onset of myelination, during which large numbers of newly formed oligodendrocytes die (Barres et al., 1992; Trapp et al., 1997).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Barres et al., 1992 source

      Not checked in the free run.

    • cites Trapp et al., 1997 source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    CLEC7A+ microglia in P7 white matter are amoeboid with thicker primary branches and larger cell bodies, whereas CLEC7A− microglia in cortex are more ramified.

    “CLEC7A+ microglia in the P7 white matter were amoeboid in shape and had thicker primary branches with larger cell bodies (Figures 7E–7G). By contrast, CLEC7A−microglia in the cortex were much more ramified.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In developing white matter, CLEC7A+ microglia had amoeboid morphology and often contained pyknotic nuclei.

    “we followed up on our histological observation that CLEC7A+ microglia in the developing white matter had amoeboid shapes and frequently contained pyknotic nuclei (Figures 8A and 8B; Movie S1).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Early postnatal PAM contained inclusions positive for cCASP3; these inclusions were MBP-positive and had pyknotic nuclei.

    “More importantly, we observed early postnatal PAM containing cleaved Caspase-3 (cCASP3)-positive inclusions, which were MBP positive and possessed pyknotic nuclei (Figures 8E, 8F, and S7E; Movie S2).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    No cCASP3+ Aldh1l1-GFP-labeled astrocytes were found, despite substantial interactions between astrocytes and early postnatal PAM.

    “In addition, we could not find cCASP3+ Aldh1l1-GFP labeled astrocytes, although substantial interactions between astrocytes and early postnatal PAM existed (Figures 8H, S7D, and S7F).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Detection rates of Mbp and Gfap in P7-C0 microglia (2.9% and 0.7%) were more than twofold lower than in P7-C1/P7-GPNMB+CLEC7A+ microglia (P<0.001 for each gene by chi-squared test), making mRNA contamination during cell isolation unlikely.

    “Because the detection rates for P7-C0 microglia (8/274=2.9% for Mbp; 2/274=0.7% for Gfap) were more than 2 folds lower for both genes (P<0.001 for either gene, Chi-squared test), these differences were unlikely due to mRNA contamination during cell isolation.”
    Quote in the paperSubstantial claim

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

    Findingown data

    Using higher overlap thresholds for defining interactions produced similar results to the 30% threshold.

    “Similar results were observed with higher percentages.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Immediate early gene expression in sorted microglia is an experimental artifact; adult microglia in tissue are negative for Fos and Egr1.

    “Indeed, adult microglia stained negative for Fos/Egr1 transcripts or proteins, whereas their expression was readily detected in the surrounding non-microglial cells”
    • Also stated in: § Limited transcriptomic heterogeneity of adult homeostatic microglia · hedged
      “if we consider the contribution of IEGs to the clustering result as an experimental artifact”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Ayata et al., 2018 source

      Not checked in the free run.

    • cites Haimon et al., 2018 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Choroid plexus macrophages undergo a postnatal transcriptional alteration or developmental phenotypic switch.

    “suggesting a developmental phenotypic switch for these particular brain resident macrophages.”
    • Also stated in: § Postnatal transcriptional alteration of choroid plexus macrophages · boosted
      “CP MΦ are a second example of brain resident macrophages that undergo a postnatal transcriptional alteration.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Machine-reported cell cycle gene lists were compared with four published datasets for experimentally characterized cell cycle genes.

    “We compared these machine-reported gene lists with four published datasets for experimentally characterized cell cycle genes (including the HeLa data we used as seeds) (Grant et al., 2013)”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Grant et al., 2013 source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductcitation

    Region-specific bulk RNA-seq was performed on highly purified homeostatic microglia sorted with TMEM119 antibodies, which label over 90% of microglia in the adult brain.

    “we performed region-specific bulk RNA-seq of highly purified homeostatic microglia sorted with TMEM119 antibodies, which label over 90% of microglia in the adult brain (Bennett et al., 2016).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Bennett et al., 2016 source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Amoeboid morphology in microglia or macrophages is associated with a phagocytic state.

    “Amoeboid morphology has indeed been associated with a phagocytic state in microglia or macrophages (Bohlen et al., 2017; Ling and Wong, 1993).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Bohlen et al., 2017 source

      Not checked in the free run.

    • cites Ling and Wong, 1993

      Not checked in the free run.

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

    Study design and conductnone given

    Animal procedures received approval from the Stanford Administrative Panel on Laboratory Animal Care and complied with institutional and national rules.

    “All animal studies were performed with approval from the Stanford Administrative panel on Laboratory Animal Care in accordance with institutional and national regulations.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Phagocytes might actively induce cell death to shape tissues and form functional circuits.

    “Alternatively, these phagocytes may promote cell death to sculpt tissue structures and create functional circuits (Fuchs and Steller, 2011).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Fuchs and Steller, 2011 source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    The Smart-seq2 platform had nearly single molecule detection limit and high precision (R=0.98) for ERCC spike-in RNA.

    “we analyzed the sequencing results for External RNA Controls Consortium (ERCC) Spike-In RNA, and found that our Smart-seq2 platform had nearly single molecule detection limit and high precision (R=0.98)”
    Quote in the paperSubstantial claim

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

    Interpretationreasoning

    In PAM, phagocytosis might induce or strengthen genes linked to lipid metabolism, lysosomal activity, and phagocytosis.

    “Alternatively, the act of phagocytosis may drive or reinforce the expression of genes associated with lipid metabolism, lysosomal functions and phagocytosis.”
    Quote in the paperNot a substantial claim

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

    Study design and conductcitation

    A recently developed cell extraction protocol was used to minimize microglial activation.

    “We used a recently developed protocol for cell extraction to minimize microglial activation (Bennett et al., 2016).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bennett et al., 2016 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    PAM have high energy turnover, particularly lipid metabolism, which may be required for phagocytosis of lipid-rich oligodendrocytes.

    “This indicates high levels of energy turnover, particularly lipid metabolism, in PAM that may be required for phagocytosis of lipid-rich oligodendrocytes.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Postnatal immature microglia can either turn on adult homeostatic genes or assume the PAM phenotype similar to DAM.

    “These data suggest that postnatal immature microglia can either turn on genes specific for classical microglia found in the adult brain (P60-homeostatic) or assume the PAM phenotype (P7-C1 and P7-GPNMB+CLEC7A+) similar to DAM in aging and neurodegeneration”
    Quote in the paperSubstantial claim

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

    Interpretationreasoning

    Efficient engulfment of newly formed oligodendrocytes by PAM may be necessary to generate evenly spaced oligodendrocyte tiling and make room for myelination.

    “It is thus possible that efficient engulfment of newly formed oligodendrocytes by PAM in the developing white matter is necessary to generate evenly spaced oligodendrocyte tiling and to make room for myelination.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Midbrain microglia have distinct gene expression profiles compared with cortical microglia.

    “For example, microglia in the midbrain have been shown to possess distinct gene expression profiles compared with cortical microglia (De Biase et al., 2017).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites De Biase et al., 2017 source

      Not checked in the free run.

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

    Study design and conductnone given

    Only male mice were used in all experiments.

    “Male mice were used in all experiments.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    DAM are phagocytic and involved in phagocytosis of Aβ plaques in AD models.

    “Indeed, while early postnatal PAM engulf newly formed oligodendrocytes and possibly newborn astrocytes during development, DAM are also phagocytic and involved in phagocytosis of Aβ plaques in AD models (Keren-Shaul et al., 2017; Krasemann et al., 2017).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Keren-Shaul et al., 2017 source

      Not checked in the free run.

    • cites Krasemann et al., 2017 source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    The study detected about 3 times as many genes per cell and higher detection rates for over 80% of genes compared with two published scRNA-seq datasets.

    “our study detected about 3 times as many genes per cell, and produced higher detection rates for over 80% of genes”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Keren-Shaul et al., 2017 source

      Not checked in the free run.

    • cites Matcovitch-Natan et al., 2016 source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    In degenerative disease models, microglial transcriptional changes and DAM signature genes (Spp1, Gpnmb, Clec7a) depend on the TREM2-APOE pathway and are suppressed when Trem2 or Apoe is absent.

    “It has been shown that transcriptional changes in degenerative disease-associated microglia depend on the TREM2-APOE pathway, and the DAM signature genes such as Spp1, Gpnmb and Clec7a are suppressed in disease models that also lack Trem2 or Apoe (Keren-Shaul et al., 2017; Krasemann et al., 2017).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Keren-Shaul et al., 2017 source

      Not checked in the free run.

    • cites Krasemann et al., 2017 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    PAM phagocytic activity mainly targets newly formed oligodendrocytes rather than OPCs.

    “suggesting that this phagocytic activity mainly targets newly formed oligodendrocytes but not their precursors.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Beyond gene transcription, microglial heterogeneity may be captured by other measures, including electrophysiological, epigenetic, or functional readouts.

    “Microglial heterogeneity can also be represented by measurements other than gene transcription, such as electrophysiological properties, epigenetic features or functional readouts (Ayata et al., 2018; De Biase et al., 2017).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Ayata et al., 2018 source

      Not checked in the free run.

    • cites De Biase et al., 2017 source

      Not checked in the free run.

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

    Interpretationcitation

    Transcriptomic differences among adult microglia may exist in brain regions beyond the four major regions tested.

    “Furthermore, additional brain regions may uncover transcriptomic differences not present in the 4 major regions tested here.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Early postnatal PAM phagocytose newly formed oligodendrocytes and, to a lesser extent, astrocytes.

    “Taken together, these data suggest that early postnatal PAM phagocytose newly formed oligodendrocytes and perhaps to a lesser extent, astrocytes.”
    Quote in the paperSubstantial claim

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

    Study design and conductnone given

    RNA-seq data processing and clustering were performed blinded, and genotype was blinded when analyzing PAM presence in wild-type, Trem2−/−, or Apoe−/− brains.

    “Processing of RNA-seq data and clustering analysis were performed blinded, and cell identities were unblinded when combining the meta data into the analysis pipeline. Genotype information was blinded when analyzing PAM presence in the wild type, Trem2−/− or Apoe−/− mutant mouse brains.”
    Quote in the paperNot a substantial claim

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

    Study design and conductcitation

    Sensitivity and accuracy of scRNA-seq data were estimated using ERCC spike-ins with known concentrations and a published algorithm.

    “We used a previously published algorithm to estimate sensitivity and accuracy of scRNA-seq data based on ERCC spike-ins with known concentrations (Svensson et al., 2017).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Svensson et al., 2017 source

      Not checked in the free run.

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

    Study design and conductown data

    Bulk RNA-seq included three biological replicates per brain region; histology experiments were performed on at least three animals and repeated at least twice.

    “For the bulk RNA-seq data, we included 3 biological replicates for each brain region in our analysis. All of the histology experiments (RNA in situ or immunohistochemistry) were performed on at least 3 individual animals, and repeated at least twice.”
    Quote in the paperNot a substantial claim

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

    Study design and conductreasoning

    Cells were excluded from scRNA-seq analysis if they had few reads, few detected genes, or poor ERCC correlation.

    “For scRNA-seq data analysis, we applied 3 statistical criteria in the quality control workflow to exclude cells that had low sequencing reads (poor RNA recovery), low numbers of detected genes (low complexity of libraries), or low ERCC correlation coefficients (insufficient accuracy) using the methods described below.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    To make phase labels applicable to microglia, each gene's expression was correlated with the average expression of each gene set, the gene was assigned to the set with the highest correlation, and this reassignment was repeated until no further changes occurred.

    “To ensure the phase label for each remaining gene is applicable to microglia, correlations between its expression and the average expression of each gene set were calculated, and the gene was assigned to the gene set that had the maximal correlation with it. The correlation examination and re-assignment were run iteratively until no more genes were re-assigned.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    A permutation test with 100,000 repetitions generated a null distribution of expression differences along the cell order, against which the observed difference was compared to compute a p-value.

    “Permutation was repeated for 1.0e+05 times to generate a distribution of permutation differences, with which the observed difference was compared to calculate a p-value.”
    Quote in the paperSubstantial claim

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

    Study design and conductown data

    For bulk RNA-seq, all three replicates per region were prepared identically and included in the analysis; they had similar read counts and detected genes.

    “For bulk RNA-seq, all 3 replicates for each region were prepared identically and included in the analysis. They generated sequencing results with similar read counts and numbers of detected genes.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    Sequencing and histology used randomly selected wild-type or mutant mice kept under identical housing conditions.

    “For sequencing or histology experiments, randomly chosen wild type or mutant mice that were housed under the same condition were used.”
    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

    For scRNA-seq, samples from multiple mice were pooled to generate the gene expression atlas, followed by histology validation.

    “For the scRNA-seq data, we pooled samples from multiple mice to generate the gene expression atlas followed by histology validation.”
    Quote in the paperNot a substantial claim

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

    Study design and conductown data

    After quality control filtering, 1,816 of 1,922 cells were retained.

    “After filtering, 1816 cells, out of 1922, were retained.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductcitation

    The number of cells for scRNA-seq and replicates for bulk RNA-seq were empirically determined based on previous studies, and the sample size allowed robust identification of cell populations as small as 20 cells.

    “The number of cells for scRNA-seq and the number of replicates for bulk RNA-seq were empirically determined based on previous studies (Darmanis et al., 2015; Tabula Muris et al., 2018; Zhang et al., 2014). The used sample-size in scRNA-seq allowed us to robustly identify cell populations as small as 20 cells.”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Darmanis et al., 2015 source

      Not checked in the free run.

    • cites Tabula Muris et al., 2018

      Not checked in the free run.

    • cites Zhang et al., 2014 source

      Not checked in the free run.

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

    Study design and conductreasoning

    During iterative cell cycle gene discovery, the initial gene seeds were retained in the updated list for each round to prevent random drift caused by noise.

    “During the iteration, the initial gene seeds were always kept in the updated list for the next round to avoid random drifting by noise.”
    Quote in the paperNot a substantial claim
65 of 65

  • The paper relies on this without support.

    Little is known about the molecular heterogeneity of microglia under physiological conditions, especially during development.

    “Despite the importance of these multitasking cells, little is known about their molecular heterogeneity under physiological conditions and especially during development when they perform many critical non-immune functions.”

  • The paper relies on this without support.

    It is unclear whether there are molecularly defined subtypes of microglia in the adult brain and how they are distributed across brain regions.

    “it remains unclear whether there are molecularly defined subtypes of microglia in the adult brain and, if so, how they are distributed across brain regions.”

  • The paper relies on this without support.

    How extensive microglial heterogeneity is, and what it does in the developing and mature healthy brain, is still unclear.

    “The extent of microglial heterogeneity and its functional significance in the developing and mature homeostatic brain, however, remains elusive.”

  • The paper relies on this without support.

    Professional phagocytes must remove dying or dead cells to keep tissue homeostasis.

    “Dying or dead cells need to be cleared away by professional phagocytes in order to maintain tissue homeostasis.”

  • The paper relies on this without support.

    During development of multicellular organisms, excess cells are commonly produced and then removed by controlled apoptosis.

    “Overproduction of cells, balanced by controlled apoptosis, is a common theme during development of multicellular organisms, although the biological significance behind this is often unknown.”
5 of 5

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Several subsets of microglia with unique molecular and functional phenotypes exist in the healthy, aged, and degenerating brain.

    “Recently, it has been shown that several subsets of microglia with unique molecular and functional phenotypes exist in the healthy, aged, and degenerating brain3–5.”

    From Lipid-droplet-accumulating microglia represent a dysfunctional and proinflammatory state in the aging brain (2020) · cites this paper as [5]

    The Results describe distinct microglial populations in the healthy developing mouse brain, and the Discussion compares one of them with disease-associated microglia. But the paper does not establish several distinct subsets in aged brains; its Discussion instead reports little transcriptomic variation among classical adult microglia. The claim extends the paper’s findings beyond the populations it studied. (LLM, conf=0.85)

    Experimental — accuracy not yet measured

  • Single-cell deep sequencing methods revealed regional-, age- and sex-dependent differences of the microglial population and various expression profiles defining microglial subpopulations.

    “Recent progress in single-cell deep sequencing methods allowed for more refined differentiation and revealed regional-, age- and sex-dependent differences of the microglial population, and a growing number of studies demonstrate various expression profiles defining microglial subpopulations.”

    From Beyond Activation: Characterizing Microglial Functional Phenotypes (2021) · cites this paper as [108]

    The Results describe developmental changes in microglial gene expression and distinct microglial clusters; the Discussion reports discovery of the PAM subset. But the Results and Discussion find little transcriptomic variation among classical adult microglia across the regions tested, and STAR Methods state that only male mice were used. This paper therefore does not establish the claimed sex-dependent differences, and its regional finding is more qualified than the claim suggests. (LLM, conf=0.90)

    Experimental — accuracy not yet measured

  • Deep sequencing experiments revealed microglial differences depending on brain region, age, gender, and pathology.

    “Recent experiments using deep sequencing revealed differences not only depending on the brain region, but also on age, gender, and pathology [28,39,99,102,108,109].”

    From Beyond Activation: Characterizing Microglial Functional Phenotypes (2021) · cites this paper as [108]

    The Abstract and Results (page numbers not supplied) report developmental-stage differences and a microglial subset concentrated in developing white matter. The Discussion reports little overt regional transcriptomic difference among classical adult microglia, so the regional claim needs that qualification. The paper compares its findings with disease-associated microglia, but does not report differences by gender or experiments comparing microglia across pathologies. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

  • Clusters were identified based on expression of known cell type-specific markers: microglia (clusters 0 to 6 and 10), T cells (cluster 7), macrophages (clusters 8 and 9), neutrophils (cluster 12), and B cells (cluster 14).

    “These clusters were identified based on the expression of known cell type–specific markers: microglia (clusters 0 to 6 and 10), T cells (cluster 7), macrophages (clusters 8 and 9), neutrophils (cluster 12), and B cells (cluster 14) (Fig. 8A and table S5) (40, 41).”

    From Deletion of <i>Abi3</i> gene locus exacerbates neuropathological features of Alzheimer’s disease in a mouse model of Aβ amyloidosis (2021) · cites this paper as [40]

    In Results, under “Clustering of brain myeloid cells across developmental stages by deep scRNA-seq” (page unavailable), the paper describes using enriched genes to assign cell identities and identifies microglial and macrophage clusters. It does not establish the citing study’s cluster numbering or provide support in these passages for identifying its T-cell and B-cell clusters. Its own cluster 7 is a macrophage cluster, and cluster 14 is natural killer cells, so those numbers should not be read as the cited paper’s assignments. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • The microglia population is highly heterogeneous, comprised of cells in discrete transcriptional dynamic substates that appear in defined contexts and brain regions.

    “Indeed, the microglia population is highly heterogeneous, comprised of cells in discrete transcriptional dynamic “substates” that appear in defined contexts and brain regions [333,334,335,336].”

    From Revisiting the Pathogenesis of X-Linked Adrenoleukodystrophy (2025) · cites this paper as [336]

    The Introduction (page unavailable) reports greater heterogeneity in early postnatal microglia and identifies a transient population in specific developing brain regions, supporting context- and region-associated transcriptional states. But the Results section, “Limited transcriptomic heterogeneity of adult homeostatic microglia” (page unavailable), finds little variation across adult brain regions. The unqualified claim that the microglia population is highly heterogeneous overstates this source. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Li, Q., Cheng, Z., Zhou, L., Darmanis, S., Neff, N. F., Okamoto, J., Gulati, G., Bennett, M. L., Sun, L. O., Clarke, L. E., Marschallinger, J., Yu, G., Quake, S. R., Wyss-Coray, T., & Barres, B. A.
Affiliations
  • Stanford Medicine (US)
  • Virginia Tech (US)
  • Stanford University (US)
Corresponding author
Li, Qingyun
ORCID
Venue
Neuron
Publisher
Elsevier BV
Published
2019-01
Volume
101
Issue
2
Pages
207-223.e10
ISSN
0896-6273 (ISSN-L)
DOI
10.1016/j.neuron.2018.12.006
PMC
PMC6336504
OpenAlex
W2949996247
Cited by
1,129
References
54
Citations by year
2026: 102 · 2025: 158 · 2024: 163 · 2023: 174 · 2022: 153 · 2021: 185 · 2020: 122 · 2019: 67 · 2018: 1 · 2017: 1 · 2016: 1 · 2012: 2
FWCI
42.06
Type
journal article
Language
en
Open access
Licence
Locations
  • Neuron (publishedVersion)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion)
  • PubMed (publishedVersion)
  • Europe PMC (PubMed Central) (acceptedVersion)
  • PubMed Central (acceptedVersion)
Funders
  • JPB Foundation
  • Adelson Medical Research Foundation
  • Vincent J. Coates Foundation
  • NIH (R01 DA015043, R01 MH110504)
  • Veterans Administration
  • NOMIS Foundation
  • Helen Hay Whitney Foundation
  • National Multiple Sclerosis Society
  • Life Science Research Foundation
  • Stanford Medicine Dean’s Fellowship
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, microglial heterogeneity, brain development, white matter development, amoeboid microglia, microglial transcriptome, mouse brain, innate immunity
MeSH terms
Algorithms, Animals, Animals, Newborn, Brain, Choroid Plexus, Computer Simulation, Embryo, Mammalian, Mice, Inbred C57BL, Mice, Transgenic, Oligodendroglia, Phagocytosis, Antigens, CD, Cluster Analysis, Sequence Analysis, RNA, Microglia, Gene Expression Regulation, Developmental, Myeloid Cells, Cell Proliferation, Mice, Gene Regulatory Networks, High-Throughput Nucleotide Sequencing, 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 2026-10-09.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.

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