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

In the paper

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

What the sources say

• ?Not checked — no full text of the sourcecites Alliot et al., 1999 source
• Supported

  The Abstract identifies microglia as resident macrophages of the CNS and reports that adult microglia derive from primitive myeloid progenitors arising before embryonic day 8. The developmental account describes these progenitors in the yolk sac and their subsequent seeding of the brain, supporting the claim's developmental-origin paraphrase. (LLM, conf=0.88)

  cites Ginhoux et al., 2010 source
• ?Not checked — no full text of the sourcecites Kierdorf et al., 2013 source

In the paper

“These genes are highly enriched in microglia and are not upregulated in monocytes that engraft the brain in microglia-depleted mouse models”

What the sources say

• Partly supported

  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)

  cites Bennett et al., 2018 sourceExperimental — accuracy not yet measured
• Supported

  In Results (‘beMφs have a predictable genetic signature’; page unavailable), the paper reports that its microglia gene signature was enriched in microglia across three models of brain engraftment, while engrafted macrophages had a distinct signature. It also identifies P2ry12 as a microglia-specific marker absent from the engrafted cells. The Introduction identifies homeostatic microglia as originating from the yolk-sac lineage. (LLM, conf=0.90)

  cites Cronk et al., 2018 source
• Partly supported

  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)

  cites Shemer et al., 2018 sourceExperimental — accuracy not yet measured

In the paper

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

What the sources say

• Partly supported

  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)

  cites Bussian et al., 2018 sourceExperimental — accuracy not yet measured

In the paper

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

What the sources say

• Partly supported

  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)

  cites Jaitin et al., 2014 sourceExperimental — accuracy not yet measured
• Supported

  The Introduction describes single-cell RNA-seq as a tool for resolving immune-cell heterogeneity across normal and disease states. The Results report that analysis of cells from wild-type and Alzheimer’s disease mice identified distinct immune subpopulations, including rare ones. (LLM, conf=0.92)

  cites Keren-Shaul et al., 2017 source
• ?Not checked — no full text of the sourcecites Papalexi and Satija, 2018 source

In the paper

“Among them is Adgrg1 (Gpr56), demonstrated previously to be conserved in human, mouse, and zebrafish, which is associated with development of oligodendrocytes”

What the sources say

• Supported

  The Introduction (page not provided) explicitly calls ADGRG1 an evolutionarily conserved regulator of oligodendrocyte development in zebrafish, mice, and humans. It also describes effects of Adgrg1 loss on oligodendrocyte precursor proliferation and myelination. (LLM, conf=0.95)

  cites Giera et al., 2018 source

In the paper

“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”

What the sources say

• Partly supported

  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)

  cites Caciotti et al., 2013 sourceExperimental — accuracy not yet measured

In the paper

“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”

What the sources say

• Partly supported

  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)

  cites He and Sharpless, 2017 sourceExperimental — accuracy not yet measured

In the paper

“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”

What the sources say

• Supported

  The Results section on the AD mouse model reports that microglia lacking TDP-43 showed enhanced amyloid clearance alongside significant synapse loss. The paper identifies TDP-43 as encoded by Tardbp, and its Results link TDP-43 depletion to increased phagocytic uptake of Aβ. The supplied passages do not give page numbers. (LLM, conf=0.93)

  cites Paolicelli et al., 2017 source

In the paper

“inflammatory cytokines have been seen previously in a sub-set of microglia in aged C57BL/6 mice”

What the sources say

• Supported

  The Results section reports that a small microglial population enriched in aged mice expresses inflammatory signals, including Il1b. The Discussion also identifies Il1b and Tnf in this subpopulation, and the STAR Methods specify that the mice used were C57BL/6J. Page numbers are not available in our copy. (LLM, conf=0.92)

  cites Hammond et al., 2019 source
• ?Not checked — no full text of the sourcecites Sierra et al., 2007 source
• Partly supported

  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)

  cites Mrdjen et al., 2018 sourceExperimental — accuracy not yet measured

In the paper

“Its expression has been shown to be increased in several neurological diseases, such as AD”

What the sources say

• Partly supported

  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)

  cites Borggrewe et al., 2018 sourceExperimental — accuracy not yet measured

In the paper

“We estimated the level of spurious UMIs in the data using statistics on empty MARS-seq wells as previously described (Jaitin et al., 2014).”

What the sources say

• ?Could not verify

  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)

  cites Jaitin et al., 2014 sourceExperimental — accuracy not yet measured

In the paper

“which have been implicated previously in longevity”

What the sources say

• Supported

  The ‘Repair of oxidative DNA damage’ section reports that earlier studies proposed a correlation between DNA repair efficiency and maximum species lifespan, and describes evidence linking nucleotide excision repair to lifespan. It also cautions that DNA damage’s contribution to longevity remains debated. That supports the claim’s modest wording, ‘implicated.’ (LLM, conf=0.92)

  cites Ungvari et al., 2008 source

In the paper

“This is in accordance with previous reports showing high constraints of brain gene expression variance across evolution”

What the sources say

• Supported

  The Results section reports that brain gene expression was the most similar across the vertebrate species studied compared with other tissues, and describes its relatively low divergence as consistent with evolutionary constraints. This supports the claim as a paraphrase, although the proposed reason for those constraints is explicitly a hypothesis. (LLM, conf=0.91)

  cites Chan et al., 2009 source
• ?Not checked — no full text of the sourcecites Brawand et al., 2011 source
• Supported

  In Results, under “Evolutionary distribution of gene expression levels predict gene function,” the paper reports that brain had the most genes with low evolutionary variance, which it identifies as the most constraint. It also says this finding is consistent with previous estimates for brain expression evolution. (LLM, conf=0.94)

  cites Chen et al., 2019 source

In the paper

“Previously, genes expressed by microglia have been implicated in AD in human GWASs”

What the sources say

• ?Could not verify

  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)

  cites Gandal et al., 2018 sourceExperimental — accuracy not yet measured
• Supported

  The Results section reports associations between AD genetic signals and microglia-expressed gene sets in both mouse and independent human single-cell expression datasets. The associations did not survive the study’s stringent Bonferroni correction, but the claim’s cautious wording, “implicated,” is consistent with the reported signals. (LLM, conf=0.88)

  cites Jansen et al., 2019 source
• Partly supported

  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)

  cites Keren-Shaul et al., 2017 sourceExperimental — accuracy not yet measured

In the paper

“Msr1 mediates endocytosis of low-density lipoproteins (LDLs) and is involved in uptake and degradation of amyloid b (Ab)”

What the sources say

• Partly supported

  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)

  cites Frenkel et al., 2013 sourceExperimental — accuracy not yet measured

In the paper

“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”

In the paper

“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)”

In the paper

“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 )”

In the paper

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

In the paper

“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”

In the paper

“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”

In the paper

“we did not observe any detectable correlation between microglial and neuronal density (Figure S1E) or microglia/neuron ratio and microglia process length”

In the paper

“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.”

In the paper

“most species contained some metacells that were identified as contaminating cells”

In the paper

“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”

In the paper

“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”

In the paper

“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”

In the paper

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

In the paper

“Another highly expressed and conserved lysosomal gene, Grn (or progranulin)”

In the paper

“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”

In the paper

“These subtypes are unrelated to sex differences and observed in both male and female microglia”

In the paper

“Conservation of P2Y12 and PU.1 across Microglia Evolution”

In the paper

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

In the paper

“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.”

In the paper

“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).”

In the paper

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

In the paper

“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.”

In the paper

“RNA-seq expression of representative highly conserved and highly expressed genes identified from clusters 1-3 (as shown in Figure 3A).”

In the paper

“Characterization of Microglia Species-Specific and Disease-Associated Pathways”

In the paper

“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”

In the paper

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

In the paper

“Laboratory mouse models have been described to show considerable differences regarding the immune system regulation”

What the sources say

• ?Not checked — no full text of the sourcecites Sellers et al., 2012 source

In the paper

“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”

In the paper

“Vsir (or VISTA) is an immune checkpoint gene that inhibits T cell response”

What the sources say

• ?Not matched

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

  cites Xu et al., 2018

In the paper

“has been implicated in frontotemporal dementia in humans and mice”

What the sources say

• ?Not checked — no full text of the sourcecites Baker et al., 2006 source

In the paper

“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.”

In the paper

“Cells with high (> 32 UMI) expression of hemoglobin genes were discarded (e.g., Hbb, Hba1, Hba2).”

In the paper

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

In the paper

“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 the paper

“reads with multiple mapping positions were excluded.”

In the paper

“developed for scRNA-seq to produce expression libraries with a minimum of three replicates per population.”

What the sources say

• ?Not checked — no full text of the sourcecites Keren-Shaul et al., 2019 source

In the paper

“ERCC genes were discarded.”

In the paper

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

In the paper

“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.”

In the paper

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

In the paper

“we took great care and carefully calibrated sample acquisition and processing time (STAR Methods)”

In the paper

“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).”

In the paper

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

In the paper

“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”

What the sources say

• ?Not checked — no full text of the sourcecites Fielder et al., 2017 source

In the paper

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

In the paper

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

In the paper

“For each organism, we used Metacell analysis to discard outlier cells.”