CoChat

Identification and therapeutic modulation of a pro-inflammatory subset of disease-associated-microglia in Alzheimer’s disease

Open access

The DOI matches its Crossref record; 70 of the 70 references it cites were found in a registry or an index; 41 of the 109 sources its claims cite were checked for support.

Abstract

BACKGROUND: Disease-associated-microglia (DAM) represent transcriptionally-distinct and neurodegeneration-specific microglial profiles with unclear significance in Alzheimer's disease (AD). An understanding of heterogeneity within DAM and their key regulators may guide pre-clinical experimentation and drug discovery. METHODS: Weighted co-expression network analysis (WGCNA) was applied to existing microglial transcriptomic datasets from neuroinflammatory and neurodegenerative disease mouse models to identify modules of highly co-expressed genes. These modules were contrasted with known signatures of homeostatic microglia and DAM to reveal novel molecular heterogeneity within DAM. Flow cytometric validation studies were performed to confirm existence of distinct DAM sub-populations in AD mouse models predicted by WGCNA. Gene ontology analyses coupled with bioinformatics approaches revealed drug targets and transcriptional regulators of microglial modules predicted to favorably modulate neuroinflammation in AD. These guided in-vivo and in-vitro studies in mouse models of neuroinflammation and neurodegeneration (5xFAD) to determine whether inhibition of pro-inflammatory gene expression and promotion of amyloid clearance was feasible. We determined the human relevance of these findings by integrating our results with AD genome-wide association studies and human AD and non-disease post-mortem brain proteomes. RESULTS: WGCNA applied to microglial gene expression data revealed a transcriptomic framework of microglial activation that predicted distinct pro-inflammatory and anti-inflammatory phenotypes within DAM, which we confirmed in AD and aging models by flow cytometry. Pro-inflammatory DAM emerged earlier in mouse models of AD and were characterized by pro-inflammatory genes (Tlr2, Ptgs2, Il12b, Il1b), surface marker CD44, potassium channel Kv1.3 and regulators (NFkb, Stat1, RelA) while anti-inflammatory DAM expressed phagocytic genes (Igf1, Apoe, Myo1e), surface marker CXCR4 with distinct regulators (LXRα/β, Atf1). As neuro-immunomodulatory strategies, we validated LXRα/β agonism and Kv1.3 blockade by ShK-223 peptide that promoted anti-inflammatory DAM, inhibited pro-inflammatory DAM and augmented Aβ clearance in AD models. Human AD-risk genes were highly represented within homeostatic microglia suggesting causal roles for early microglial dysregulation in AD. Pro-inflammatory DAM proteins were positively associated with neuropathology and preceded cognitive decline confirming the therapeutic relevance of inhibiting pro-inflammatory DAM in AD. CONCLUSIONS: We provide a predictive transcriptomic framework of microglial activation in neurodegeneration that can guide pre-clinical studies to characterize and therapeutically modulate neuroinflammation in AD.

Overview

CoChat Summary

TL;DR

WGCNA of microglial transcriptomes resolved disease-associated microglia (DAM) into pro-inflammatory and anti-inflammatory subsets in Alzheimer’s disease (AD) mouse models, validated by flow cytometry. LXRα/β agonism and Kv1.3 blockade promoted anti-inflammatory DAM, inhibited pro-inflammatory DAM, and augmented Aβ clearance, with human AD relevance.

Key findings

  • WGCNA of 43 microarray samples (16,721 genes) identified 19 co-expression modules; AD-associated modules included a pro-inflammatory Magenta module (Ptgs2, Il12b, Tlr2, Il1b; CD44, Kv1.3; NFkB/Stat1/RelA), an anti-inflammatory/phagocytic Yellow module (Igf1, Apoe, Myo1e; CXCR4; LXRα/β/Atf1), and a homeostatic Blue module (Tmem119, P2ry12). Validation in 64 Nanostring samples with 369 common genes replicated these modules.
  • In 6–8 mo 5xFAD mice, CD11c+ DAM were ≈50% of CD11b+ CNS immune cells versus <10% in WT; flow cytometry confirmed CD44+, CXCR4+, double-positive, and double-negative DAM subsets with age-dependent trajectories that were augmented in 5xFAD.
  • LXR agonist T0901317 (30 mg/kg i.p. twice weekly for 2 weeks) produced nearly two-fold increases in CXCR4+CD44− anti-inflammatory DAM and increased fibrillar Aβ42 phagocytosis. ShK-223 (100 nM in vitro; 100 μg/kg i.p.) inhibited LPS-induced Ptgs2/Il1b, augmented Kcnj2/Nceh1/Timp2, promoted Aβ phagocytosis and Lamp1 co-localization, and after 3–6 mo treatment lowered hippocampal Aβ plaque burden/size and peri-plaque Iba1 in 5xFAD mice.
  • Human relevance: AD GWAS hits were enriched in the homeostatic Blue module (p = 0.005; Bin1, Cnn2, Picalm, Sorl1), and pro-inflammatory DAM proteins were positively associated with neuropathology and preceded cognitive decline in 47 BLSA post-mortem brains (20 AD, 14 asymptomatic AD, 13 controls).

Methods

  • WGCNA was applied to existing GEO microarray datasets of purified CD11b+ microglia: 43 arrays from WT, TREM2−/−, 5xFAD, TREM2−/−/5xFAD, APP/PS1, and LPS/IL4-treated cells; data were batch-corrected and analyzed for 16,721 genes. External validation used 64 Nanostring microglial samples and 369 common genes.
  • Flow cytometry and functional assays used acutely isolated CNS immune cells from WT, 5xFAD, and aging mice; markers included CD11b, CD45, CD11c, CD44, CXCR4, and Kv1.3 detected with ShK-F6CA. Ex vivo fluorescent Aβ42 phagocytosis, RT-qPCR, immunohistochemistry, BV2 microglial assays, and HEK293 Kv1.3 transfection were also performed.
  • In vivo studies used female 5xFAD and C57BL/6 mice; acute neuroinflammation was induced with i.p. LPS, LXR agonist T0901317 was given at 30 mg/kg twice weekly for 2 weeks, and ShK-223 was given at 100 μg/kg i.p. twice weekly from 3 to 6 mo (n = 10/group) with behavior and Aβ quantitation. Statistics used ANOVA with Tukey’s HSD and p ≤ 0.05.

Limitations

  • Stated: existing RNAseq microglial transcriptomes were not included because of lower depth of coverage and inability to sufficiently overcome batch effects when combining microarray and RNAseq data.
  • Stated: most DAM did not express high levels of either CD44 or CXCR4, suggesting additional complexity within DAM may not have been revealed by the flow cytometry panel.
  • Evident: therapeutic and mechanistic conclusions rely largely on mouse models and in vitro assays; human validation is associative (GWAS enrichment and post-mortem proteomics), and the human BLSA proteomics methods are noted as “available online … and will be published subsequently.”

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

Integrity signals

Registry

Crossref · 10.1186/s13024-018-0254-8

Cites

70 of 70 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 WGCNA of mouse microglial transcriptome identifies co-expression modules with distinct AD pathology-associated inflammatory profiles. a Three transcriptomic microarray datasets containing 43 arrays derived from CD11b+ mouse microglia were normalized, batch-corrected and used in a WGCNA meta-analysis. Datasets included microglia polarized in-vitro to M1-like (LPS) or M2-like (IL4) states, acutely isolated microglia from WT, 5xFAD, Trem2−/− and 5xFAD/Trem2−/− mice, and from WT and APP/PS1 transgenic mice. b WGCNA identified 19 distinct modules (networks) of highly co-expressed genes (See Additional file 1: Table S1). c Cluster dendrogram showing the 19 microglial modules and trait correlations with inflammatory status (LPS or IL4) and AD pathology (5xFAD or APP/PS1 vs. wild-type). d Comparison of expression of module eigengene for 4 modules of interest, across treatment conditions/traits (Hub gene members are shown in descending order of module membership or kME)
Fig. 1 WGCNA of mouse microglial transcriptome identifies co-expression modules with distinct AD pathology-associated inflammatory profiles. a Three transcriptomic microarray datasets containing 43 arrays derived from CD11b+ mouse microglia were normalized, batch-corrected and used in a WGCNA meta-analysis. Datasets included microglia polarized in-vitro to M1-like (LPS) or M2-like (IL4) states, acutely isolated microglia from WT, 5xFAD, Trem2−/− and 5xFAD/Trem2−/− mice, and from WT and APP/PS1 transgenic mice. b WGCNA identified 19 distinct modules (networks) of highly co-expressed genes (See Additional file 1: Table S1). c Cluster dendrogram showing the 19 microglial modules and trait correlations with inflammatory status (LPS or IL4) and AD pathology (5xFAD or APP/PS1 vs. wild-type). d Comparison of expression of module eigengene for 4 modules of interest, across treatment conditions/traits (Hub gene members are shown in descending order of module membership or kME)CC BY · PubMed Central
Fig. 2 Microglial co-expression modules recapitulate homeostatic and DAM profiles and predict heterogeneity within DAM. a 2D-representation of trait-associated WGCNA modules using viSNE. Space between two clusters of genes is an arbitrary representation of dissimilarity or anti-correlation between expression patterns. b Mapping of DAM and homeostatic signature genes to WGCNA modules shows clustering of DAM genes to Yellow and Magenta modules and of Homeostatic genes to the Blue module. Genes shown in Grey represent those that were either not classified as DAM or homeostatic or were not identified by single cell RNAseq [9]. c Key transcriptional regulators, membrane-associated drug targets and predicted functional phenotypes of Blue, Magenta and Yellow AD-associated microglial modules (also see Additional file 7)
Fig. 2 Microglial co-expression modules recapitulate homeostatic and DAM profiles and predict heterogeneity within DAM. a 2D-representation of trait-associated WGCNA modules using viSNE. Space between two clusters of genes is an arbitrary representation of dissimilarity or anti-correlation between expression patterns. b Mapping of DAM and homeostatic signature genes to WGCNA modules shows clustering of DAM genes to Yellow and Magenta modules and of Homeostatic genes to the Blue module. Genes shown in Grey represent those that were either not classified as DAM or homeostatic or were not identified by single cell RNAseq [9]. c Key transcriptional regulators, membrane-associated drug targets and predicted functional phenotypes of Blue, Magenta and Yellow AD-associated microglial modules (also see Additional file 7)CC BY · PubMed Central
Fig. 3 Replication co-expression analysis of microglial transcriptomic datasets across neuroinflammatory and neurodegenerative disease models. a Cluster dendrogram summarizing results of validation WGCNA performed using 64 existing Nanostring purified microglial transcriptomic datasets (369 genes in common across all conditions). Datasets included aging WT and APP/PS1 models (age range 2–17 mo, n = 24), WT and SOD1-G93A transgenic mice (n = 13), control and various stages of EAE in mice (n = 21) and WT and intra-cerebral LPS-treated WT mice (n = 6). Expression data were first batch-corrected followed by WGCNA to identify modules of highly co-expressed genes. b Module over-representation analysis (ORA) of derivation (43 microglial arrays, y-axis) and validation WGCNA (64 microglial Nanostring transcriptomic datasets, x-axis) studies. Module overlap was assessed by Fisher-exact test and the negative log10 transformed false discovery rate (FDR) for overlap is indicated by color intensity. Validation modules are numbered (M1–19) while derivation modules are indicated by the original color scheme. c Module eigengene expression and trajectories of change in module expression across various traits in the Nanostring co-expression dataset. Validation modules highlighted here were chosen based on strength of overlap from the over-representation analysis. M6 and M7 are representative of the Blue homeostatic derivation module. M4 is representative of the Yellow anti-inflammatory derivation module while M3 is representative of the Magenta pro-inflammatory derivation module. M1 is representative of LPS-induced Midnightblue derivation module while M19 is representative of the IL4-upregulated Cyan derivation module
Fig. 3 Replication co-expression analysis of microglial transcriptomic datasets across neuroinflammatory and neurodegenerative disease models. a Cluster dendrogram summarizing results of validation WGCNA performed using 64 existing Nanostring purified microglial transcriptomic datasets (369 genes in common across all conditions). Datasets included aging WT and APP/PS1 models (age range 2–17 mo, n = 24), WT and SOD1-G93A transgenic mice (n = 13), control and various stages of EAE in mice (n = 21) and WT and intra-cerebral LPS-treated WT mice (n = 6). Expression data were first batch-corrected followed by WGCNA to identify modules of highly co-expressed genes. b Module over-representation analysis (ORA) of derivation (43 microglial arrays, y-axis) and validation WGCNA (64 microglial Nanostring transcriptomic datasets, x-axis) studies. Module overlap was assessed by Fisher-exact test and the negative log10 transformed false discovery rate (FDR) for overlap is indicated by color intensity. Validation modules are numbered (M1–19) while derivation modules are indicated by the original color scheme. c Module eigengene expression and trajectories of change in module expression across various traits in the Nanostring co-expression dataset. Validation modules highlighted here were chosen based on strength of overlap from the over-representation analysis. M6 and M7 are representative of the Blue homeostatic derivation module. M4 is representative of the Yellow anti-inflammatory derivation module while M3 is representative of the Magenta pro-inflammatory derivation module. M1 is representative of LPS-induced Midnightblue derivation module while M19 is representative of the IL4-upregulated Cyan derivation moduleCC BY · PubMed Central
Fig. 4 Flow cytometric phenotyping of microglia using pro- and anti-inflammatory DAM markers. a Selection of potential flow cytometric markers for Magenta (pro-inflammatory) and Yellow (anti-inflammatory) DAM modules in acutely isolated WT and 5xFAD microglia. Criteria included kME ≥ 0.75 for one module, ≤0.5 for the other module and predominant cell membrane localization. Yellow module: Cxcr4 (CD184); DAM-phenotype: CD11c and Pdcd11 (PD1); Magenta module: CD44, CD274 (PDL1), CD45 and Kv1.3. Cell surface functional Kv1.3 channel expression was detected by ShK-F6CA, a fluorescein-conjugated Kv1.3 channel blocker that selectively binds to Kv1.3 channels. b Acutely isolated CNS immune cells were gated for live cells (FSC/SSC) and then for single cells (using FSC-A/FSC-H) followed by CD11b positivity. CD11c was used as a marker of DAM based on prior RNAseq studies. CD11c+ DAM were gated as shown and proportions of CD11c+ DAM in CD11b+ CNS immune cells were compared between WT and 5xFAD mice (n = 6 mice/group, age 6–8 mo). c Comparison of pro-inflammatory DAM (CD44, CD45 and CD274) and anti-inflammatory DAM (CXCR4) specific markers within CD11c+ DAM in 5xFAD and WT mice (n = 6/group, age 6–8 mo). Grey histogram represents isotype control. d Flow cytometric phenotyping of CD11b+ cells in mouse brain based on CD44 and CXCR4 expression. e Changes in proportions of DAM subsets in aging WT and in aging 5xFAD mouse brains (n = 4 mice/group/time point). Pairwise and group-wise comparisons and significant differences are highlighted. Levels of significance: *p < 0.05, **p < 0.01, ***p < 0.005. Colored markers indicate within group comparisons (Red: 5xFAD, Blue: WT) while Black markers indicate inter-group differences (WT vs. 5xFAD)
Fig. 4 Flow cytometric phenotyping of microglia using pro- and anti-inflammatory DAM markers. a Selection of potential flow cytometric markers for Magenta (pro-inflammatory) and Yellow (anti-inflammatory) DAM modules in acutely isolated WT and 5xFAD microglia. Criteria included kME ≥ 0.75 for one module, ≤0.5 for the other module and predominant cell membrane localization. Yellow module: Cxcr4 (CD184); DAM-phenotype: CD11c and Pdcd11 (PD1); Magenta module: CD44, CD274 (PDL1), CD45 and Kv1.3. Cell surface functional Kv1.3 channel expression was detected by ShK-F6CA, a fluorescein-conjugated Kv1.3 channel blocker that selectively binds to Kv1.3 channels. b Acutely isolated CNS immune cells were gated for live cells (FSC/SSC) and then for single cells (using FSC-A/FSC-H) followed by CD11b positivity. CD11c was used as a marker of DAM based on prior RNAseq studies. CD11c+ DAM were gated as shown and proportions of CD11c+ DAM in CD11b+ CNS immune cells were compared between WT and 5xFAD mice (n = 6 mice/group, age 6–8 mo). c Comparison of pro-inflammatory DAM (CD44, CD45 and CD274) and anti-inflammatory DAM (CXCR4) specific markers within CD11c+ DAM in 5xFAD and WT mice (n = 6/group, age 6–8 mo). Grey histogram represents isotype control. d Flow cytometric phenotyping of CD11b+ cells in mouse brain based on CD44 and CXCR4 expression. e Changes in proportions of DAM subsets in aging WT and in aging 5xFAD mouse brains (n = 4 mice/group/time point). Pairwise and group-wise comparisons and significant differences are highlighted. Levels of significance: *p < 0.05, **p < 0.01, ***p < 0.005. Colored markers indicate within group comparisons (Red: 5xFAD, Blue: WT) while Black markers indicate inter-group differences (WT vs. 5xFAD)CC BY · PubMed Central
Fig. 5 CMAP analysis pro-inflammatory and anti-inflammatory DAM genes reveals genetic and chemical perturbagens as selective modulators. a Venn diagram representing exclusive members and dual members of Magenta and Yellow microglial modules. b Distribution of perturbagens from connectivity map (CMAP) analysis based on a median summary score (range − 100 to + 100) that represents likelihood of the perturbagen resulting in the desired transcriptional profile. c, d For these analyses, the desired transcriptional profile was upregulation of anti-inflammatory and down-regulation of pro-inflammatory module expression. A summary score ≥ + 90 typically indicates a high likelihood of achieving the desired profile (c, ↓Magenta/pro-inflammatory ↑Yellow/anti-inflammatory) while a score ≤ − 90 indicates likelihood of the opposite transcriptional profile (d, ↑Magenta/pro-inflammatory ↓Yellow/anti-inflammatory). The most significant chemical perturbagens (existing drugs) and genetic perturbagens (gene suppression or over-expression) are shown
Fig. 5 CMAP analysis pro-inflammatory and anti-inflammatory DAM genes reveals genetic and chemical perturbagens as selective modulators. a Venn diagram representing exclusive members and dual members of Magenta and Yellow microglial modules. b Distribution of perturbagens from connectivity map (CMAP) analysis based on a median summary score (range − 100 to + 100) that represents likelihood of the perturbagen resulting in the desired transcriptional profile. c, d For these analyses, the desired transcriptional profile was upregulation of anti-inflammatory and down-regulation of pro-inflammatory module expression. A summary score ≥ + 90 typically indicates a high likelihood of achieving the desired profile (c, ↓Magenta/pro-inflammatory ↑Yellow/anti-inflammatory) while a score ≤ − 90 indicates likelihood of the opposite transcriptional profile (d, ↑Magenta/pro-inflammatory ↓Yellow/anti-inflammatory). The most significant chemical perturbagens (existing drugs) and genetic perturbagens (gene suppression or over-expression) are shownCC BY · PubMed Central
Fig. 6 Agonists of liver-X-receptor (LXR) α/β signaling promote anti-inflammatory DAM and Aβ phagocytic activity in microglia. a Experimental design: 6–7 mo 5xFAD mice (females, n = 4/group) received either vehicle (20% DMSO) or a LXR α/β agonist T0901317 (30 mg/kg i.p. daily × 2 weeks) after which acutely isolated CNS immune cells were assessed for flow cytometric immuno-phenotyping as well as flow cytometric assays of fluorescent fibrillar Aβ. b, c CD11b+CD45low CNS immune cells were assessed for CD11c expression. A comparison of CD11c expression in 5xFAD mice (blue) and WT mice (red) is shown. Gating threshold for CD11c is shown in (c). d, e A comparison of CD44 and CXCR4 expression profiles within CD11c+ DAM acutely isolated from vehicle-treated or T0901317-treated 5xFAD mice. Relative proportions of each subset (of all CD11b+CD45lowCD11c+ DAM) are shown. Yellow represents anti-inflammatory DAM, Magenta represents pro-inflammatory DAM and Brown represents double positive cells. A quantitative analysis of various subsets of CD11c + DAM in vehicle-treated and T0901317-treated mice is shown in (e). f Comparison of ability of acutely isolated CD11b+CD45low microglia (from vehicle-treated and T0901317-treated mice) to phagocytose fluorescent Aβ42 fibrils. *p < 0.05, **p < 0.01, ***p < 0.005
Fig. 6 Agonists of liver-X-receptor (LXR) α/β signaling promote anti-inflammatory DAM and Aβ phagocytic activity in microglia. a Experimental design: 6–7 mo 5xFAD mice (females, n = 4/group) received either vehicle (20% DMSO) or a LXR α/β agonist T0901317 (30 mg/kg i.p. daily × 2 weeks) after which acutely isolated CNS immune cells were assessed for flow cytometric immuno-phenotyping as well as flow cytometric assays of fluorescent fibrillar Aβ. b, c CD11b+CD45low CNS immune cells were assessed for CD11c expression. A comparison of CD11c expression in 5xFAD mice (blue) and WT mice (red) is shown. Gating threshold for CD11c is shown in (c). d, e A comparison of CD44 and CXCR4 expression profiles within CD11c+ DAM acutely isolated from vehicle-treated or T0901317-treated 5xFAD mice. Relative proportions of each subset (of all CD11b+CD45lowCD11c+ DAM) are shown. Yellow represents anti-inflammatory DAM, Magenta represents pro-inflammatory DAM and Brown represents double positive cells. A quantitative analysis of various subsets of CD11c + DAM in vehicle-treated and T0901317-treated mice is shown in (e). f Comparison of ability of acutely isolated CD11b+CD45low microglia (from vehicle-treated and T0901317-treated mice) to phagocytose fluorescent Aβ42 fibrils. *p < 0.05, **p < 0.01, ***p < 0.005CC BY · PubMed Central
Fig. 7 Kv1.3 channels are expressed by pro-inflammatory DAM and regulate pro-inflammatory DAM genes. a Confirmation of specificity of flow cytometric assay of functional Kv1.3 channels: HEK293 cells were transiently transfected with 1 μg of pRC/CMV (empty vector) or pRC/CMV-mKv1.3 (mouse Kv1.3) after which ShK-F6CA was added (final concentration 100 nM) × 30 min, followed by flow cytometry (3 independent experiments were performed). Electrophysiological confirmation of Kv1.3 current expression was performed by whole-cell patch clamp (data not shown). b Expression of Magenta module markers Kv1.3 and CD45 in subsets of CD11c+ DAM in 5xFAD mice (Homeostatic: CD44−CXCR4−, Magenta/pro-inflammatory-DAM: CD44+CXCR4−, Yellow/anti-inflammatory-DAM: CXCR4+CD44− and double-positive CD44+CXCR4+ DAM) in adult 5xFAD mice (n = 5 mice, age 6–8 mo). Grey histogram represents isotype control (for CD45) or negative control (unlabeled cells for ShK-F6CA). c Gating of CNS immune cells based on CD11b and CD45 into CD11b+CD45low (resident microglia) and CD11b+CD45high subpopulations (Top). Kv1.3 surface expression in CD45high and CD45low subsets of CD11b+ cells, is shown below. d Comparison of Kv1.3 expression in CD45low and CD45high subsets of CD11b+ CNS immune cells in 3 age groups of WT and 5xFAD mice (n = 3 mice/group/time point. Post-hoc statistical tests: *p < 0.05, **p < 0.01, ***p < 0.005. e Comparison of gene expression in acutely isolated CNS immune cells isolated from 6 mo WT mice treated with saline, ShK-223, LPS or LPS + ShK-223 (4 daily i.p. doses, n = 3/group). Selected module markers: Homeostatic: Tmem119, Pro-inflammatory DAM: Il1b, Ptgs2; Anti-inflammatory DAM: Kcnj2, Nceh1, Timp2
Fig. 7 Kv1.3 channels are expressed by pro-inflammatory DAM and regulate pro-inflammatory DAM genes. a Confirmation of specificity of flow cytometric assay of functional Kv1.3 channels: HEK293 cells were transiently transfected with 1 μg of pRC/CMV (empty vector) or pRC/CMV-mKv1.3 (mouse Kv1.3) after which ShK-F6CA was added (final concentration 100 nM) × 30 min, followed by flow cytometry (3 independent experiments were performed). Electrophysiological confirmation of Kv1.3 current expression was performed by whole-cell patch clamp (data not shown). b Expression of Magenta module markers Kv1.3 and CD45 in subsets of CD11c+ DAM in 5xFAD mice (Homeostatic: CD44−CXCR4−, Magenta/pro-inflammatory-DAM: CD44+CXCR4−, Yellow/anti-inflammatory-DAM: CXCR4+CD44− and double-positive CD44+CXCR4+ DAM) in adult 5xFAD mice (n = 5 mice, age 6–8 mo). Grey histogram represents isotype control (for CD45) or negative control (unlabeled cells for ShK-F6CA). c Gating of CNS immune cells based on CD11b and CD45 into CD11b+CD45low (resident microglia) and CD11b+CD45high subpopulations (Top). Kv1.3 surface expression in CD45high and CD45low subsets of CD11b+ cells, is shown below. d Comparison of Kv1.3 expression in CD45low and CD45high subsets of CD11b+ CNS immune cells in 3 age groups of WT and 5xFAD mice (n = 3 mice/group/time point. Post-hoc statistical tests: *p < 0.05, **p < 0.01, ***p < 0.005. e Comparison of gene expression in acutely isolated CNS immune cells isolated from 6 mo WT mice treated with saline, ShK-223, LPS or LPS + ShK-223 (4 daily i.p. doses, n = 3/group). Selected module markers: Homeostatic: Tmem119, Pro-inflammatory DAM: Il1b, Ptgs2; Anti-inflammatory DAM: Kcnj2, Nceh1, Timp2CC BY · PubMed Central
Fig. 8 Kv1.3 channel blockade promotes Aβ phagocytosis and limits amyloid β burden in 5xFAD mice. a Comparison of ex-vivo phagocytosis of fluorescent Aβ42 fibrils by acutely isolated CNS CD11b+ cells from WT (sham-treated or LPS-treated for 4 days) and 5xFAD mice (age 6–8 mo, n = 4 mice/group). Cells were loaded with fluorescent fibrillar Aβ42 in the presence of either sham or ShK-223 (100 nM) for 1 h prior to labeling with anti-CD11b and anti-CD45 antibodies. b Comparison of ex-vivo phagocytic capacity for fluorescent Aβ42 fibrils by acutely isolated CNS CD11b+ cells from sham-treated or ShK-223-treated 5xFAD mice (n = 4/group). ShK-223 was administered twice a week (i.p., 100 μg/kg) in this study. c-g In this study of ShK-223 in 5xFAD mice, mice were treated with PBS or ShK-223 (100 μg/kg) between 3 and 6 mo (n = 10 mice/group) and brains were fixed for immunohistochemistry. c Comparison of hippocampal Aβ plaques in PBS-treated (top) and ShK-223-treated (bottom) mice (Right: Higher magnification image). d Quantification of Aβ plaque burden in the hippocampus and frontal cortex. e Comparison of hippocampal (subiculum) Aβ plaque size in PBS- and ShK-223-treated mice. f, g, h Immunofluorescence images showing Iba1 (Green) and Aβ (Blue) immunoreactivities in the hippocampus of PBS- and ShK-223-treated mice. g Quantitative analyses of Iba1 expression (% area) and Aβ plaque burden. h Quantitative analysis of peri-plaque Iba1 immunoreactivity comparing PBS- and ShK-223-treated mice. 5–8 plaques of relatively similar size were selected from each mouse and Iba1+ area was normalized to the plaque area and compared across groups. *p < 0.05, **p < 0.01, ***p < 0.005
Fig. 8 Kv1.3 channel blockade promotes Aβ phagocytosis and limits amyloid β burden in 5xFAD mice. a Comparison of ex-vivo phagocytosis of fluorescent Aβ42 fibrils by acutely isolated CNS CD11b+ cells from WT (sham-treated or LPS-treated for 4 days) and 5xFAD mice (age 6–8 mo, n = 4 mice/group). Cells were loaded with fluorescent fibrillar Aβ42 in the presence of either sham or ShK-223 (100 nM) for 1 h prior to labeling with anti-CD11b and anti-CD45 antibodies. b Comparison of ex-vivo phagocytic capacity for fluorescent Aβ42 fibrils by acutely isolated CNS CD11b+ cells from sham-treated or ShK-223-treated 5xFAD mice (n = 4/group). ShK-223 was administered twice a week (i.p., 100 μg/kg) in this study. c-g In this study of ShK-223 in 5xFAD mice, mice were treated with PBS or ShK-223 (100 μg/kg) between 3 and 6 mo (n = 10 mice/group) and brains were fixed for immunohistochemistry. c Comparison of hippocampal Aβ plaques in PBS-treated (top) and ShK-223-treated (bottom) mice (Right: Higher magnification image). d Quantification of Aβ plaque burden in the hippocampus and frontal cortex. e Comparison of hippocampal (subiculum) Aβ plaque size in PBS- and ShK-223-treated mice. f, g, h Immunofluorescence images showing Iba1 (Green) and Aβ (Blue) immunoreactivities in the hippocampus of PBS- and ShK-223-treated mice. g Quantitative analyses of Iba1 expression (% area) and Aβ plaque burden. h Quantitative analysis of peri-plaque Iba1 immunoreactivity comparing PBS- and ShK-223-treated mice. 5–8 plaques of relatively similar size were selected from each mouse and Iba1+ area was normalized to the plaque area and compared across groups. *p < 0.05, **p < 0.01, ***p < 0.005CC BY · PubMed Central
Fig. 9 Identification of AD-risk genes as key members of homeostatic, pro-inflammatory DAM and anti-inflammatory DAM modules. a Enrichment analysis of GWAS-identified AD risk genes in mouse microglial modules (1234 genes identified by MAGMA were used for this analysis [34]). An enrichment score of ≥1.96 (red line) corresponds to significant enrichment (p < 0.05). b-d AD GWAS hits specific to (b) Blue (homeostatic), (c) Yellow (anti-inflammatory DAM) and (d) Magenta (pro-inflammatory DAM) modules. Top 5 genes based on strength of disease association (X-axis: -log10 p-value) are highlighted by a gray silhouette and top 5 genes based on module membership (kME, Y-axis) are shown (also see Additional file 5: Table S5, Additional file 6: Table S6). e-h Analyses performed using quantitative proteomic data obtained from 47 post-mortem human frontal cortices (BLSA, n = 20 confirmed AD cases with cognitive decline, n = 13 controls without any AD pathology and n = 14 cases with AD pathology without cognitive symptoms). e Overlap of mouse modules [Blue/homeostatic, Yellow/anti-inflammatory DAM, Magenta/pro-inflammatory DAM, Purple (homeostatic), Cyan (IL4-upregulated module) and Midnightblue (LPS-upregulated module)] with human protein modules by hypergeometric one-tailed Fisher-exact test. Color intensity indicates strength of significance after Benjamini-Hochberg FDR correction. Strongest FDR-corrected p-values (0.075) were seen for overlap between Yellow and Magenta mouse microglial modules with M2 (unadjusted p = 0.0013), M4 (unadjusted p = 0.0014) and M35 (unadjusted p = 0.0014) human proteomic modules. Y-axis: Mouse WGCNA modules; X-axis: Protein BLSA modules. Correlations between BLSA modules with CERAD and Braak neuropathological grades of AD are shown. f Distribution of mouse microglial module hub genes (kME ≥ 0.90) in human BLSA AD proteome. g Comparison of aggregate expression of Magenta, Yellow and Blue mouse microglial modules in control, asymptomatic AD and clinical AD. Modules with ≥5 hub genes (kME ≥ 0.8) were included. Median module expression was calculated and compared across groups. h Normalized protein expression data of top three proteins from the pro-inflammatory Magenta module (CD44, Bst2 and Nampt) are shown
Fig. 9 Identification of AD-risk genes as key members of homeostatic, pro-inflammatory DAM and anti-inflammatory DAM modules. a Enrichment analysis of GWAS-identified AD risk genes in mouse microglial modules (1234 genes identified by MAGMA were used for this analysis [34]). An enrichment score of ≥1.96 (red line) corresponds to significant enrichment (p < 0.05). b-d AD GWAS hits specific to (b) Blue (homeostatic), (c) Yellow (anti-inflammatory DAM) and (d) Magenta (pro-inflammatory DAM) modules. Top 5 genes based on strength of disease association (X-axis: -log10 p-value) are highlighted by a gray silhouette and top 5 genes based on module membership (kME, Y-axis) are shown (also see Additional file 5: Table S5, Additional file 6: Table S6). e-h Analyses performed using quantitative proteomic data obtained from 47 post-mortem human frontal cortices (BLSA, n = 20 confirmed AD cases with cognitive decline, n = 13 controls without any AD pathology and n = 14 cases with AD pathology without cognitive symptoms). e Overlap of mouse modules [Blue/homeostatic, Yellow/anti-inflammatory DAM, Magenta/pro-inflammatory DAM, Purple (homeostatic), Cyan (IL4-upregulated module) and Midnightblue (LPS-upregulated module)] with human protein modules by hypergeometric one-tailed Fisher-exact test. Color intensity indicates strength of significance after Benjamini-Hochberg FDR correction. Strongest FDR-corrected p-values (0.075) were seen for overlap between Yellow and Magenta mouse microglial modules with M2 (unadjusted p = 0.0013), M4 (unadjusted p = 0.0014) and M35 (unadjusted p = 0.0014) human proteomic modules. Y-axis: Mouse WGCNA modules; X-axis: Protein BLSA modules. Correlations between BLSA modules with CERAD and Braak neuropathological grades of AD are shown. f Distribution of mouse microglial module hub genes (kME ≥ 0.90) in human BLSA AD proteome. g Comparison of aggregate expression of Magenta, Yellow and Blue mouse microglial modules in control, asymptomatic AD and clinical AD. Modules with ≥5 hub genes (kME ≥ 0.8) were included. Median module expression was calculated and compared across groups. h Normalized protein expression data of top three proteins from the pro-inflammatory Magenta module (CD44, Bst2 and Nampt) are shownCC BY · PubMed Central
Fig. 10 A cohesive model of homeostatic and distinct pro- and anti-inflammatory subsets of DAM in neurodegeneration. Transcriptional regulators, cell surface markers and key drug targets are shown. Each dot in this figure represents a single gene. Colors represent module membership (Blue: Homeostatic module, Red and Midnightblue: M1-like modules not associated with AD, Cyan: M2-like module not associated with AD, Black: intermediate stages, Yellow: anti-inflammatory DAM module upregulated in AD, Magenta: pro-inflammatory DAM module upregulated in AD). Upward and downward arrows indicate direction of change in gene/protein expression. Double arrow (for CD45) indicates higher degree of change in pro-inflammatory DAM as compared to anti-inflammatory DAM subsets
Fig. 10 A cohesive model of homeostatic and distinct pro- and anti-inflammatory subsets of DAM in neurodegeneration. Transcriptional regulators, cell surface markers and key drug targets are shown. Each dot in this figure represents a single gene. Colors represent module membership (Blue: Homeostatic module, Red and Midnightblue: M1-like modules not associated with AD, Cyan: M2-like module not associated with AD, Black: intermediate stages, Yellow: anti-inflammatory DAM module upregulated in AD, Magenta: pro-inflammatory DAM module upregulated in AD). Upward and downward arrows indicate direction of change in gene/protein expression. Double arrow (for CD45) indicates higher degree of change in pro-inflammatory DAM as compared to anti-inflammatory DAM subsetsCC BY · PubMed Central

Verification

Claims the paper makes

16 Valid · 5 Not valid · 33 Not checked · 122 Own results and interpretations

103 more claims · 7 with their cited sources checked

Create a free account to see all 176 claims and to verify the 68 references still open.

See all claims — free account
  • It may state two separate claims as one. 57
  • Probably not a substantial claim. 19

Create a free account to see these claims.

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?

1 more claim about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

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.

5 of 5

Citations

Cites

Show all 70 references

Cited by

See all 472

Metadata

Authors
Rangaraju, S., Dammer, E. B., Raza, S. A., Rathakrishnan, P., Xiao, H., Gao, T., Duong, D. M., Pennington, M. W., Lah, J. J., Seyfried, N. T., & Levey, A. I.
Affiliations
  • Emory University (US)
  • Peptides International (United States) (US)
Corresponding author
Rangaraju, Srikant
Venue
Molecular Neurodegeneration
Published
2018-05-21
Volume
13
Issue
1
Pages
24-24
ISSN
1750-1326 (ISSN-L)
OpenAlex
W2805185070
Cited by
430
References
70
Citations by year
2026: 34 · 2025: 56 · 2024: 71 · 2023: 60 · 2022: 64 · 2021: 73 · 2020: 39 · 2019: 21 · 2018: 9
FWCI
15.93
Type
journal article
Language
en
Open access
Licence
Locations
  • Molecular Neurodegeneration (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, other-oa)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • National Institute on Aging (P50 AG025688)
  • American Brain Foundation (28301)
  • Alzheimer's Association (37102)
  • National Institute of Neurological Disorders and Stroke (K08-NS099474-1, P30 NS055077)
  • National Institutes of Health (AG046161)
Topics
  • Life Sciences › Neuroscience › Neurology › Neuroinflammation and Neurodegeneration Mechanisms
  • Life Sciences › Neuroscience › Cellular and Molecular Neuroscience › Alzheimer's disease research and treatments
  • Life Sciences › Neuroscience › Neurology › Neurological Disease Mechanisms and Treatments
Keywords
disease-associated microglia, Alzheimer's disease, microglial heterogeneity, pro-inflammatory microglia, anti-inflammatory microglia, neuroinflammation, amyloid-beta clearance, microglial transcriptome, 5xFAD mice, CD44, CXCR4
MeSH terms
Alzheimer Disease, Animals, Humans, Inflammation, Microglia, Mice, Genome-Wide Association Study, 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.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-07, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-07 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (qwen3.8-omni-flash and 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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Identification and therapeutic modulation of a pro-inflammatory subset of disease-associated-microglia in Alzheimer’s disease. https://app.cochat.ai/papers/10.1186/s13024-018-0254-8
The paper
Rangaraju, S., Dammer, E. B., Raza, S. A., Rathakrishnan, P., Xiao, H., Gao, T., Duong, D. M., Pennington, M. W., Lah, J. J., Seyfried, N. T., & Levey, A. I. (2018). Identification and therapeutic modulation of a pro-inflammatory subset of disease-associated-microglia in Alzheimer’s disease. Molecular Neurodegeneration. https://doi.org/10.1186/s13024-018-0254-8

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.