. 2024 May;27(5):873-885.

doi: 10.1038/s41593-024-01610-w. Epub 2024 Mar 27.

TREM1 disrupts myeloid bioenergetics and cognitive function in aging and Alzheimer disease mouse models

Edward N Wilson^{1

2}, Congcong Wang¹, Michelle S Swarovski¹, Kristy A Zera¹, Hannah E Ennerfelt¹, Qian Wang¹, Aisling Chaney^{1

3}, Esha Gauba¹, Javier A Ramos Benitez¹, Yann Le Guen¹, Paras S Minhas¹, Maharshi Panchal¹, Yuting J Tan¹, Eran Blacher¹, Chinyere A Iweka¹, Haley Cropper^{1

3}, Poorva Jain^{1

3}, Qingkun Liu¹, Swapnil S Mehta¹, Abigail J Zuckerman¹, Matthew Xin¹, Jacob Umans¹, Jolie Huang¹, Aarooran S Durairaj¹, Geidy E Serrano⁴, Thomas G Beach⁴, Michael D Greicius^{1

2}, Michelle L James^{1

2

3}, Marion S Buckwalter^{1

2

5}, Melanie R McReynolds^{6

7

8}, Joshua D Rabinowitz^{6

7}, Katrin I Andreasson^{9

10

11}

Affiliations

¹ Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA, USA.
² Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA, USA.
³ Department of Radiology, Stanford University School of Medicine, Stanford, CA, USA.
⁴ Civin Laboratory for Neuropathology, Banner Sun Health Research Institute, Sun City, AZ, USA.
⁵ Department of Neurosurgery, Stanford University School of Medicine, Stanford, CA, USA.
⁶ Lewis-Sigler Institute for Integrative Genomics, Princeton University, Princeton, NJ, USA.
⁷ Department of Chemistry, Princeton University, Princeton, NJ, USA.
⁸ Department of Biochemistry and Molecular Biology, Huck Institutes of the Life Sciences, Pennsylvania State University, University Park, PA, USA.
⁹ Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA, USA. kandreas@stanford.edu.
¹⁰ Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA, USA. kandreas@stanford.edu.
¹¹ Chan Zuckerberg Biohub, San Francisco, CA, USA. kandreas@stanford.edu.

PMID: 38539014
PMCID: PMC11102654
DOI: 10.1038/s41593-024-01610-w

TREM1 disrupts myeloid bioenergetics and cognitive function in aging and Alzheimer disease mouse models

Edward N Wilson et al. Nat Neurosci. 2024 May.

. 2024 May;27(5):873-885.

doi: 10.1038/s41593-024-01610-w. Epub 2024 Mar 27.

Authors

Affiliations

¹ Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA, USA.
² Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA, USA.
³ Department of Radiology, Stanford University School of Medicine, Stanford, CA, USA.
⁴ Civin Laboratory for Neuropathology, Banner Sun Health Research Institute, Sun City, AZ, USA.
⁵ Department of Neurosurgery, Stanford University School of Medicine, Stanford, CA, USA.
⁶ Lewis-Sigler Institute for Integrative Genomics, Princeton University, Princeton, NJ, USA.
⁷ Department of Chemistry, Princeton University, Princeton, NJ, USA.
⁸ Department of Biochemistry and Molecular Biology, Huck Institutes of the Life Sciences, Pennsylvania State University, University Park, PA, USA.
⁹ Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA, USA. kandreas@stanford.edu.
¹⁰ Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA, USA. kandreas@stanford.edu.
¹¹ Chan Zuckerberg Biohub, San Francisco, CA, USA. kandreas@stanford.edu.

PMID: 38539014
PMCID: PMC11102654
DOI: 10.1038/s41593-024-01610-w

Abstract

Human genetics implicate defective myeloid responses in the development of late-onset Alzheimer disease. A decline in peripheral and brain myeloid metabolism, triggering maladaptive immune responses, is a feature of aging. The role of TREM1, a pro-inflammatory factor, in neurodegenerative diseases is unclear. Here we show that Trem1 deficiency prevents age-dependent changes in myeloid metabolism, inflammation and hippocampal memory function in mice. Trem1 deficiency rescues age-associated declines in ribose 5-phosphate. In vitro, Trem1-deficient microglia are resistant to amyloid-β₄₂ oligomer-induced bioenergetic changes, suggesting that amyloid-β₄₂ oligomer stimulation disrupts homeostatic microglial metabolism and immune function via TREM1. In the 5XFAD mouse model, Trem1 haploinsufficiency prevents spatial memory loss, preserves homeostatic microglial morphology, and reduces neuritic dystrophy and changes in the disease-associated microglial transcriptomic signature. In aging APP^Swe mice, Trem1 deficiency prevents hippocampal memory decline while restoring synaptic mitochondrial function and cerebral glucose uptake. In postmortem Alzheimer disease brain, TREM1 colocalizes with Iba1⁺ cells around amyloid plaques and its expression is associated with Alzheimer disease clinical and neuropathological severity. Our results suggest that TREM1 promotes cognitive decline in aging and in the context of amyloid pathology.

PubMed Disclaimer

Conflict of interest statement

COMPETING FINANCIAL INTRESTS STATEMENT

M.L.J and K.I.A are co-founders and Scientific Advisory Board members of Willow Neuroscience, Inc. The remaining authors declare no competing interests.

Figures

**Extended Data Fig. 1.. low cytometry gating for quantification of TREM1 expression in blood, spleen and brain myeloid cells**
a. Gating strategy for live CD45+Cd11b+TREM1+ blood myeloid cells b. Gating strategy for live CD45+Cd11b+TREM1+ spleen myeloid cells c. Gating strategy for live CD45loCd11b+TREM1+ microglia; also includes gating for P2RY12, CD206, CX3CR1 and Tmem119 microglial markers.

**Extended Data Fig. 2.. TREM1 amplifies levels of inflammatory factors but does not alter phagocytosis**
Data are mean ±} s.e.m. unless otherwise indicated. a. Quantification of immune factors in lysates of peritoneal MF isolated from 8.5 mo WT and Trem1−/− mice treated with vehicle or 100 ng/ml LPS for 20 hr. Two-way ANOVA with Tukey’s multiple comparisons shown for WT+LPS vs Trem1−/−+ LPS (n=5–6 male mice per genotype as shown). b. Trem1 deficiency does not alter phagocytosis in peritoneal MF. Cells were stimulated with vehicle or E.coli +/− LPS 100 ng/ml for 20h. 2-way ANOVA, effect of LPS ***P < 0.001, no effect of Trem1 genotype basally or with LPS stimulation. 2–3 mo male mice with 6–11 technical replicate wells as shown.

**Extended Data Fig. 3.. TREM1 promotes a pro-inflammatory polarization state in aged peritoneal MF**
Data are mean ±} s.e.m. unless otherwise indicated. a. Gating strategy for live Cd11b+Ly6G-CD45+Ly6CHi and Cd11b+Ly6G-CD45+Ly6CLo peritoneal MF from young WT and Trem1−/− (2 mo) and aged WT and Trem1−/− (22–23 mo) male mice. b. Representative gating for MHC II, CD86, and CD71 positive Ly6CHi peritoneal MF. c. Representative gating for MHC II, CD86, and CD71 positive Ly6CLo peritoneal MF. d. Multianalyte Luminex quantification of immune factors in young WT (2 mo) and aged WT (23–24 mo) and aged Trem1−/− (23–24 mo) cerebral cortex. One-way ANOVA with Tukey’s multiple comparisons shown (n=5–6 male mice per group as shown). e. Total time exploring (in seconds) during the training phase of the NOR task; ANOVA with Tukey’s multiple comparisons show no significant differences (n=6–8 female mice as shown). f. Speed (pixels/second) of mice during the final trial of the Barnes Maze; ANOVA with Tukey’s multiple comparisons (n=5–6 male and female mice as shown).

**Extended Data Fig. 4.. Trem1 deficiency alters gene expression in aged peripheral MF with minimal changes in aged brain microglia**
a. Venn diagram of DEGs (all genes q < 0.05) in pairwise comparisons of microglia isolated from aged (18 – 20 mo) vs young (3 mo) mice and pairwise comparisons of microglia isolated from aged Trem1−/− vs aged WT mice (18–20 mo). Blue indicates number of downregulated genes while red indicates the number of upregulated genes (n=3 microglial samples (2 pooled male mice per sample) per group). b. KEGG Pathway enrichment analysis of upregulated genes in aged compared to young microglia; there were no enriched pathways in the comparison of aged WT vs aged Trem1−/− microglia. c. Venn diagrams showing the number of DEGs (all genes q < 0.05) in pairwise comparisons of primary mouse MF from young WT (2 mo), aged WT (25 mo) and aged Trem1−/− (25 mo) mice. Blue indicates number of downregulated genes while red indicates the number of upregulated genes (n=3–4 male mice per group). d. Pathway enrichment analysis of upregulated genes in aged MF vs young and in aged Trem1−/− vs aged WT MF. e. Heatmap of DEGs (q < 0.05) comprising the Coordinated Lysosomal Expression and Regulation (CLEAR) network. Lysosomal functional categories are indicated for each gene. Aged Trem1−/− mice cluster with the young WT mice. Scale represents z-score values from FPKM. n=3–4 male mice per group.

**Extended Data Fig. 5.. TREM1 deficiency restores youthful metabolism in aged macrophages Data are mean ±} s.e.m. unless otherwise indicated.**
a. PCA of metabolites from young (2 mo) WT and Trem1−/− peritoneal MF (n=3 male mice per group). b. PCA of significantly regulated metabolites from primary peritoneal MF isolated from young WT (2mo), aged WT (25 mo) and aged Trem1−/− (25 mo) male mice. n=3–5 male mice per group. c. Volcano plot comparing aged WT vs young WT (left) and aged Trem1−/− vs aged WT (right) peritoneal MF with Log2 fold change (FC) and -Log10(P) values of metabolites. Significantly regulated metabolites with -log10(P) > 2 and log2 fold change > 1 are shown in blue. d. KEGG metabolic pathway enrichment analysis of differentially regulated metabolites between aged WT and young WT macrophages and between aged Trem1−/− vs aged WT MF. X-axis shows -Log10(q) value. No pathways were significant following FDR correction for the comparison of aged Trem1−/− vs young WT. Pathway analysis was performed using MetaboAnalyst 5.0; n=3–5 male mice per group. e. Total ion count of Ribose-5P, the precursor of purines and pyrimidines and total ion counts for nucleobases and derivatives, ribo-nucleosides, deoxynucleosides, and nucleoside monophosphates. Data are analyzed by one-way ANOVA with Tukey’s multiple comparisons (n=3–5 male mice per group). Abbreviations: AMP: adenosine monophosphate, GMP: guanosine monophosphate, UMP: uridine monophosphate, CMP cytidine monophosphate, IMP: inosine monophosphate. n=3–5 male mice per group as shown. f. Transcription factor (TF) enrichment analysis revealing TFs enriched for differentially expressed metabolite enzymes. g. FKPM of NRF2. ANOVA followed by Tukey’s multiple comparison (n=3–4 male mice per group as shown). h. FKPM of the 3 pyruvate dehydrogenase complex (PDH) subunits: pyruvate dehydrogenase (PDHA1), dihydrolipoyl transacetylase (DLAT) and dihydrolipoyl dehydrogenase (DLD); one-way ANOVA followed by Tukey’s multiple comparison (n=3–4 male mice per group as shown).

**Extended Data Fig. 6.. TREM1 expression does not significantly change in context of accumulating amyloid.**
Data are mean ±} s.e.m. unless otherwise indicated. a. (Left) Immunofluorescent staining of TREM1 in 9 mo WT and APPSwe-PS1ΔE9 hippocampal stratum lacunosum in IBA1+microglia. 5–7 hippocampal sections per mouse were imaged, n=5 10 male mice/genotype. Scale bar = 10 μm. TREM1 colocalizes with Iba1+ cells (white arrows). (Right) Quantification of MFI. b. Thioflavin-S (ThioS) fluorescent staining of hippocampus in 9–10 mo 5XFAD and 5XFAD;Trem1+/− female mice. Scale bar = 500 μm. c. Soluble levels of cerebral cortical As40 and 42 and the ratio of As42/40 in 5X FAD and 5XFAD;Trem1+/− mice. Student’s two tailed t-test (n=4–6 males/group as shown, 10–13 mo) d. Insoluble levels of cerebral cortical As40 and 42 and the ratio of As42/40 in 5X FAD and 5XFAD;Trem1+/− mice; Student’s two tailed t-test (n=4–6 males/group as shown, 10–13 mo) e. Quantification of hippocampal immune factors in WT, 5XFAD, and 5XFAD;Trem1+/− 10 mo mice. One-way ANOVA with Tukey post-hoc comparisons (n= 5–6 female mice per group as shown). f. Microglial number from Fig. 4i. One-way ANOVA with Tukey’s multiple comparisons (n=5 male mice per condition).

**Extended Data Fig. 7.. Disease-associated microglial (DAM) signature in 5XFAD;Trem1+/− mice.**
Data are mean ±} s.e.m. unless otherwise indicated. a. PCA of significantly regulated DAM signature genes from primary microglia isolated from WT, 5XFAD and 5XFAD/Trem1+/− mice. Individual microglial samples were pooled from 2 male mice (n=3 samples per condition). b. Hierarchical clustering of DAM signature gene set showing homeostatic, DAM Stage 1 and DAM Stage 2 genes. Scale represents z-score values from FPKM. c. Hierarchical clustering of top 170 DEGs (FDR-corrected) belonging to the full DAM signature gene set. Scale represents z-score values from FPKM. Scale represents z-score values from FPKM. d. (Left) Percent of CX3CR1+ microglia in young (3 mo) and 13–17 mo WT, 5xFAD and 5xFAD;Trem1+/− mice. (Right) MFI of CX3CR1+ microglia. One-way ANOVA with Tukey’s multiple comparison, not significant (n= 5–8 male and female mice per group as shown). e. (Left) Percent of Tmem119+ microglia in young (3 mo) and 13–17 mo WT, 5xFAD and 5xFAD;Trem1+/− mice. (Right) MFI of Tmem119+ microglia. One-way ANOVA with Tukey’s multiple comparisons (n=4–5 mice per group). f. (Left) Total time exploring (in seconds) during the training phase of the NOR task in APPSwe mice; ANOVA with Tukey’s post-hoc test (n=14–17 male and female mice per condition as shown). g. Motor speed (pixels/sec) during the final training session of the Barnes maze; ANOVA with Tukey’s post-hoc test (n=6 male and female mice per group). h. Distance traveled during the final training session of the Barnes maze; ANOVA with Tukey’s post-hoc test (n=6 male and female mice per group). i. Coupling assay tracings of synaptic mitochondria oxygen consumption rates (OCR). Shown over time are the rates of basal Complex II respiration, State III (ADP stimulated respiration), State IV (oligomycin) and State IIIu (State III uncoupled, FCCP) that were consecutively measured over the course of the assay (n=5–9 male mice per condition). j. Diagram of model of action of TREM1 in aging and transgenic mice with amyloid accumulation. In aging, peripheral TREM1 activity contributes to declines in myeloid metabolism and immune functions that lead to age-dependent cognitive decline. In models of amyloid accumulation, both peripheral and microglial TREM1 contribute to cognitive deficits associated with local accumulation of amyloid.

**Extended Data Fig. 8.. TREM1 expression in development of AD**
a. Immunofluorescent staining of human frontal cortex without addition of TREM1 antibody and secondary only control (red), Iba1 (green), and X-34 (blue) signal. This antibody validation experiment was performed three times. Scale bar = 20 μm. b. Immunoblot of TREM1 protein in postmortem human mid frontal gyrus. Clinicopathological diagnoses: non-demented Braak I-II, demented non-AD Braak stages I-II, AD Braak III-IV and AD Braak V-VI. Primary antibody detected human TREM1 band at the molecular weight of positive control (human liver lysates, Cat# HT-314, Zyagen, San Diego, CA). The TREM1 band used for analysis are indicated by blue box. Also shown is the band for s-actin at 42 kDa. This antibody validation experiment was performed once. c. Immunoblot of TREM2 protein in postmortem human mid frontal gyrus. Primary antibody detects human TREM2 band at the molecular weight of positive control (human liver lysates, Cat# HT-314, Zyagen, San Diego, CA). TREM2 bands used for analysis are indicated by blue box. Also shown is the band for s-actin at 42kDa. This antibody validation experiment was performed once. d. Mendelian Randomization (MR) analyses with Alzheimer’s disease (AD) risk level as exposure and sTREM1 plasma protein level as outcome. Blue lines are estimated MR-median weight effects, and the dashed lines indicate the 95% confidence interval for the MR effects. e. Mendelian Randomization (MR) analyses with AD risk level as exposure and TREM2 plasma protein levels as outcome. Blue lines are estimated MR-median weight effects, and the dashed lines indicate the 95% confidence interval for the MR effects.

**Figure 1.. TREM1 deficiency prevents age-associated inflammation and memory decline**
Data are represented as mean +/− SEM. a. TREM1 expression on myeloid cells is low under basal conditions, however with stimulation by PAMPs or DAMPs, TREM1 synergizes with innate immune receptors to amplify the pro-inflammatory response. Its functional counterpart TREM2 promotes phagocytosis and cell survival. b. Percent TREM1 surface expression across blood, spleen and brain CD45⁺CD11b⁺ myeloid cells in young (3 mo) and aged (13–17 mo) male mice. Student’s t-test, one-tailed (n=6–8 mice per group as shown). c. Multi-analyte quantification of plasma immune factors from young (3 mo) and aged (18.5 mo) WT and *Trem1−/−* male mice. Two-way ANOVA, effect sizes shown for genotype (Gen) and interaction (Int); post-hoc Tukey comparisons: * P < 0.05 (n=3–9 male mice/group as shown). d. Flow cytometric analysis of anti-inflammatory CD71 and pro-inflammatory MHCII and CD86 in Ly6C^Hi and Ly6C^Lo peritoneal MΦ from young 2 mo WT and aged 22–23 mo WT and *Trem1*^−/− male mice. Two-way ANOVA, effect sizes shown for genotype and marker; post-hoc Tukey comparisons: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (n=3–5 male mice/group as shown). e. Unsupervised hierarchical clustering of significantly regulated immune factors by one-way ANOVA in cortical lysates from young wild-type (2 mo) and aged WT (25 mo) and *Trem1*^−/− (23.5 mo) male mice, (n=5–6 male mice/group as shown). f. The Novel Object Recognition (NOR) task with habituation to testing arena, training phase with exploration of two identical objects and testing phase where object is replaced with a novel object. g. Discrimination index for the NOR task for young WT (5 mo), aged WT (18 mo) and aged *Trem1−/−* (18 mo) mice; paired two tailed t-test (n = 6–8 male and female mice/group as shown). h. Representative movement tracings for final training trial of the Barnes maze. Escape hole is labelled with green arrow. i. Escape latency on the Barnes maze. 2-way repeated measures ANOVA: interaction between training and genotype (F(_6,120) = 7.38, P < 0.0001); Tukey’s post hoc: ***P < 0.001, ****P < 0.0001 (n= 16 young WT, 14 aged WT, and 13 aged *Trem1−/−* male and female mice/group).

**Figure 2.. TREM1 is activated in peripheral macrophages in aging.**
Data are mean ± s.e.m. unless otherwise indicated. a. Volcano plot of DEGs from aged (18 – 20 mo) vs young (3 mo) CD45^lowCD11b⁺ microglia with −log₁₀(P) > 1 after FDR adjustment and log₂ fold change > 1. n=3 samples/age (each sample pooled from 2 male mice). b. Volcano plot of DEGs from aged *Trem1−/−* vs aged WT microglia with −log₁₀(P) > 1.25 after FDR and log₂ fold change > 1. n=3 samples/genotype (each sample pooled from 2 male mice). c. Unsupervised hierarchical clustering of DEGs from mouse peritoneal MΦ from young WT (2 mo), aged WT (25 mo) and aged *Trem1−/−* (25 mo) mice. Scale represents z-score values of FPKM (n=3–4 male mice/group). d. Unsupervised hierarchical clustering of cytokine and chemokines DEGs (q < 0.05). Scale represents z-score values from FPKM (n=3–4 mice/group). e. Pathway enrichment analysis of DEGs in the Mouse MitoCarta3.0 gene list. Scale represents Log(p) × fold enrichment (FE). f. Representative oxygen consumption rate (OCR) traces with SeaHorse for young (4.5 mo) and aged (22–23 mo) wild-type (WT) and *Trem1*^−/− peritoneal MΦ (n= 5 male mice/group). Arrowheads indicate application of electron transport-chain inhibitors. Abbreviations: olig: oligomycin, FCCP: carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone, rot/ant: rotenone/antimycin A (n=5 male mice/group). g. Quantification of basal respiration and glycolysis (extracellular acidification rate, or ECAR) for young (4.5 mo) and aged (22–23 mo) WT and *Trem1*^−/− peritoneal MΦ. Two-way ANOVA, Tukey’s multiple comparisons shown; effect of age, P<0.0001 and genotype, P=0.0001 (OCR, n=3 male mice/group; ECAR, n=3 male mice/group). h. Transmission electron microscopy (TEM) images of young (4.5 mo) and aged (24–26 mo) *Trem1*^−/− and WT peritoneal MΦ. Yellow arrows indicate mitochondria. Abbreviations: M: mitochondrion; N: nucleus. Scale = 1 μm. i. Numbers of MΦ mitochondria and percent abnormal mitochondria from (h). Two-way ANOVA, Tukey’s multiple comparisons shown; effect of age, P=0.0021; genotype, P=0.0005; interaction, P=0.0008 (n= 3–5 male mice/group). j. Mitochondrial superoxide production in peritoneal MΦ, normalized to MitoTracker signal, in young (4 mo) and aged (20 mo) male WT and *Trem1*^−/− mice, n=8 mice/group. Two-way ANOVA with Tukey’s multiple comparisons test; effect of age, P<0.0001; genotype, P=0.0012; interaction, P=0.0289.

**Figure 3.. TREM1 suppresses pentose phosphate pathway (PPP) generation of Ribose 5P and purine/pyrimidine synthesis in aged macrophages**
a. Unsupervised hierarchical clustering of significantly regulated metabolites from peritoneal MΦ isolated from young WT (2 mo), aged WT (25 mo) and aged *Trem1*^−/− (25 mo) male mice. TREM1-deficient MΦ cluster with young WT MΦ. Data were analyzed and clustered using MetaboAnalyst 5.0; n=3–5 male mice per group as shown. b. Metabolic pathway map of glycolysis, pentose phosphate pathway (PPP), TCA cycle, and purine and pyrimidine synthesis demonstrates decrease of Ribose 5-Phosphate (R5P), the precursor of purine and pyrimidine synthesis, in aged WT MΦ that is restored to young WT levels with *Trem1* deletion. Colored circles represent z-scored fold change levels for significantly modified metabolites. Yellow circles indicate no significant difference. Small gray circles indicate metabolites that were not measured. c. Integration of transcriptomic and metabolomic features demonstrates enrichment in purine and pyrimidine metabolism. Included in the joint pathway analysis were differentially regulated metabolites (q < 0.05) and the highest regulated DEGs (q-values < 0.05 and Log2FC > +/−2) and fold change for each feature. No pathways were enriched in the comparison between aged *Trem1*^−/− and young WT. Data were analyzed using the Joint Pathway Analysis function in MetaboAnalyst 5.0. d. Glycolysis and PPP enzymes (blue) with FPKM quantification. 1-way ANOVA with Tukey’s multiple comparisons shown (n=3–4 male mice/group as shown). Abbreviations: 6PGD: 6-phosphogluconate dehydrogenase; G6PD: glucose-6-phosphate dehydrogenase; HK1: hexokinase-1, PK: pyruvate kinase; RPI: ribose-5-phosphate isomerase. Enzymes *G6pdx*, *6Pgd*, *Rpia*, *Taldo1* and *Pkm* are known NRF2 target genes. Data are shown as mean ± s.e.m. e. Immunofluorescent localization of NRF2 in CD11b⁺ peritoneal MΦ isolated from (4.5 mo) and aged (24–26 mo) WT and *Trem1*^−/− male mice. f. Quantification of total NRF2 levels in MΦ from (e). 3–5 confocal images were processed per cell using ImageJ. 2-way ANOVA, Tukey’s multiple comparisons test shown (n=3–6 male mice/group as shown). Data are shown as mean ± s.e.m.

**Figure 4.. TREM1 deficiency preserves hippocampal function in 5XFAD mice**
Data are shown as mean ± s.e.m. a. OCR traces from postnatal mouse microglia from WT and *Trem1*^−/− mice stimulated with vehicle or Aß₄₂ oligomers (100 nM) for 20 hours (n=5 wells per genotype/condition). b. Quantification of basal respiration and glycolysis (ECAR). Two-way ANOVA with Tukey’s multiple comparisons; for basal respiration, effect of genotype and interaction, P=0.0006; for ECAR, effect of genotype, Aß, and interaction, P<0.0001 (n=5 wells per genotype/condition). c. Percent TREM1⁺CD45^loCD11b⁺microglia in WT vs 5XFAD mice (n=6–8 male and female mice/group as shown, 13–17 mo) d. Novel Object Recognition (NOR) of 6–7 mo WT, 5XFAD and 5XFAD;*Trem1*^+/− mice. Training (clear circles) and testing (green circles) discrimination index differences were assessed with paired t-tests. Discrimination index of 0.5 indicates chance preference (n=11–12 male and female mice/condition). e. Escape latency of 5XFAD cohorts on the Barnes maze. 2-way repeated measures (RM) ANOVA with Tukey’s multiple comparisons. Interaction between training and mouse genotype [F(_6,90) = 4.14, P < 0.001)] (n=11 WT, n=12 5XFAD, n=10 5XFAD;*Trem1*^+/− male and female mice/group). f. Mean integrated density of 6E10 signal from hippocampal CA1 of 9–10 mo 5XFAD and 5xFAD;*Trem1*^+/− mice (n=6 female mice/group). g. Mean integrated density of ThioS signal, a marker of fibrillar amyloid, from CA1 of 9–10 mo 5XFAD and 5xFAD;*Trem1*^+/− mice (n=5–7 female mice/group). h. Unsupervised hierarchical clustering of significantly regulated immune proteins in hippocampus from 9–10 mo WT, 5XFAD, and 5XFAD;*Trem1*^+/− mice, one-way ANOVA (n=5–6 female mice/group). i. Immunofluorescent images of Iba1+ microglia and 6E10 staining of amyloid in CA1 of 9–10 mo WT, 5XFAD, and 5XFAD;*Trem1*^+/− female mice and 3-dimensional rendering of representative microglia. Scale bar is 25μm (top row) and 10 μm (bottom row). j. Quantification of average microglial branch length and maximum branch length. One-way ANOVA with Tukey’s multiple comparison (n=5 female mice/group). k. Heatmap depicting group means of significantly regulated (one-way ANOVA) DAM signature genes. Microglia samples were pooled from two 6–7 mo male mice (n=3 samples per genotype). Scale represents z-score values from FPKM. l. Neuritic dystrophy visualized with anti-BACE1 immunostaining at X34+ amyloid plaques in WT, 5XFAD, and 5XFAD;*Trem1*^+/− 9–10 mo female mice; scale bar=10 μm m. Quantification of BACE1-positive percent area around amyloid plaques. One-way ANOVA with Tukey’s multiple comparisons (n= 5–6 female mice/group). n. Quantification of plasma immune factors in WT, 5XFAD, and 5XFAD;*Trem1*^+/− 9–10 mo mice. One-way ANOVA with Tukey’s multiple comparisons (n= 4–6 female mice/group).

**Figure 5.. M1 deficiency preserves hippocampal spatial memory and brain glucose uptake in *APP*^Swe mice**
Data are shown as mean ± s.e.m. a. Discrimination index during the NOR task for aged (18–21 mo) *APP*^Swe, *APP*^Swe/Trem1^+/− and *APP*^Swe/Trem1^−/− mice; paired t-test of Training vs. Testing (n=11–17 male and female mice/group as shown). b. Escape latency of *APP*^Swe cohorts in the Barnes maze. 2-way repeated measures ANOVA with Tukey’s multiple comparisons. Interaction between training and mouse genotype [F(_6,132) = 12.24, P < 0.0001)] (n=17 *APP*^Swe, 14 *APP*^Swe/Trem1^+/− and 16 *APP*^Swe/Trem1^−/− male and female mice/group). c. Percent area positive for 6E10 in hippocampus of 18–21 mo *APP*^Swe and *APP*^Swe;*Trem1*^−/− mice. Student’s two tailed t-test (n=5–7 male/female mice/group as shown). d. Average integrated density for ThioS⁺ signal in hippocampus of 18–21 mo APP^Swe and *APP*^Swe;*Trem1*^−/− mice. Student’s two tailed t-test, n=5–8 male/female mice/group as shown. e. Hippocampal immune factors in 20–23 mo in WT, *APP*^Swe, *APP*^Swe/Trem1^+/− and *APP*^Swe/Trem1^−/− mice. One-way ANOVA with Tukey’s multiple comparisons (n=4–9 female mice/group as shown). f. Plasma cytokines in 20–23 mo in WT, *APP*^Swe, *APP*^Swe/Trem1+/− and *APP*^Swe/Trem1−/− mice. One-way ANOVA with Tukey’s multiple comparisons (n=4–9 female mice/group as shown). g. Respiratory control ratio (RCR) in synaptic mitochondria fractions from 18–24 mo mouse brain. High RCR indicates mitochondria with a high capacity for substrate oxidation and ATP turnover and a low proton leak. One-way ANOVA with Tukey’s multiple comparisons (n=5–9 male mice/group as shown). h. Synaptic mitochondria OCR (pmol/min) calculated basally and for state III, state IVo, and state IIIu in 18–24 mo WT, *APP*^Swe and in *APP*^Swe;Trem1^−/− mice. State III reflects maximal ADP-stimulated respiration and State IV reflects the return to a basal state of respiration after addition of ATP synthase inhibitor oligomycin. One-way ANOVA with Tukey’s multiple comparisons (n=5–9 male mice/group as shown). i. Representative coronal brain [¹⁸F]FDG-PET/CT images of cerebral glucose metabolism in 17–19 mo female mice. SUV: standardized uptake values. Static 20 min PET images were acquired at 75–95 min following [¹⁸F]FDG injection. j. Quantification of [¹⁸F]FDG-PET signal in hippocampus and thalamus. SUVs were normalized to individual blood glucose concentrations. One-way ANOVA with Tukey’s multiple comparisons (n=5–9 female mice/group).

**Figure 6.. Myeloid TREM1 expression positively associates with increasing AD pathology.**
Data are shown as mean ± s.e.m. unless otherwise indicated. a. Immunofluorescent staining of human frontal cortex for TREM1 (red), Iba1 (green), and X-34 (blue). TREM1+/Iba1+ cells are defined as plaque associated (within 15 μm of an X-34+ amyloid plaque) or non-plaque associated. Scale bar = 20 μm. White arrows indicate colocalization of Iba1 and TREM1. b. Quantification of TREM1+/Iba1+ plaque-associated and non-plaque-associated cells. One-way ANOVA with Tukey’s multiple comparisons. n=5 AD donors and n=3 Control donors. c. Immunoblot of TREM1 and TREM2 in middle frontal gyrus across clinicopathological diagnoses (n=11–12 donors/group). d. Quantification of TREM1 and TREM2 protein normalized to ß-actin and expressed as fold-change over non-demented control. ANCOVA with Tukey’s multiple comparisons. Age and sex were included as covariates (n=11–12 donors/group). e. Linear regression analyses between amyloid plaque and neurofibrillary tangle (NFT) density and TREM1 (left plot) and TREM2 (right plot) levels from quantitative immunoblotting. Model was adjusted for age and sex. Plotted are the 95% confidence bands of the best-fit line from the linear regression. The standardized regression coefficients (ß) and P-values from the linear model are shown. n=43 donors (amyloid plaque density vs TREM1) and n=42 donors (amyloid plaque density vs TREM2). f. Linear regression analyses between neurofibrillary tangle (NFT) density and TREM1 (left plot) or TREM2 (right plot) levels from quantitative immunoblotting; n=41 donors (NFT density vs TREM1) and n=40 donors (NFT density vs TREM2). g. Mendelian Randomization (MR) analysis of relationship between plasma levels of soluble TREM1 (sTREM1; exposure) and Alzheimer’s disease risk (outcome). Blue lines are estimated MR-inverse variance weighted effects, and dashed lines indicate the 95% confidence interval for MR effects. Increased plasma sTREM1 level was associated with increased AD risk (β = +0.0929, [95% CI = 0.019 to 0.166] P = 0.013). h. Mendelian Randomization (MR) analysis showing relationship between plasma levels of soluble sTREM2 (as exposure) and Alzheimer’s disease risk (outcome). Increased plasma sTREM2 level was associated with decreased AD risk (β = −0.15 [95% CI = −0.223 to −0.077] P = 5.61×10⁻⁰⁵).

See this image and copyright information in PMC

Cited by

TREM1 is the new kid on the block for the aging-associated innate immune response and Alzheimer's disease.
Heneka MT. Heneka MT. Nat Immunol. 2024 Jun;25(6):938-940. doi: 10.1038/s41590-024-01854-5. Nat Immunol. 2024. PMID: 38831103 No abstract available.
Fear bypasses the amygdala.
Lewis S. Lewis S. Nat Rev Neurosci. 2024 May;25(5):287. doi: 10.1038/s41583-024-00808-y. Nat Rev Neurosci. 2024. PMID: 38459142 No abstract available.

References

1. Lindenberger U Human cognitive aging: corriger la fortune? Science 346, 572–578 (2014). 10.1126/science.1254403 - DOI - PubMed
1. Cunningham C & Hennessy E Co-morbidity and systemic inflammation as drivers of cognitive decline: new experimental models adopting a broader paradigm in dementia research. Alzheimers Res Ther 7, 33 (2015). 10.1186/s13195-015-0117-2 - DOI - PMC - PubMed
1. Franceschi C et al. Inflammaging and anti-inflammaging: a systemic perspective on aging and longevity emerged from studies in humans. Mech Ageing Dev 128, 92–105 (2007). 10.1016/j.mad.2006.11.016 - DOI - PubMed
1. Nikolich-Zugich J The twilight of immunity: emerging concepts in aging of the immune system. Nat Immunol 19, 10–19 (2018). 10.1038/s41590-017-0006-x - DOI - PubMed
1. Andrews SJ, Fulton-Howard B & Goate A Interpretation of risk loci from genome-wide association studies of Alzheimer’s disease. Lancet Neurol 19, 326–335 (2020). 10.1016/s1474-4422(19)30435-1 - DOI - PMC - PubMed

Methods-Only References

1. Weber B et al. TREM-1 deficiency can attenuate disease severity without affecting pathogen clearance. PLoS Pathog 10, e1003900 (2014). 10.1371/journal.ppat.1003900 - DOI - PMC - PubMed
1. Oakley H et al. Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: potential factors in amyloid plaque formation. J. Neurosci. 26, 10129–10140 (2006). 10.1523/JNEUROSCI.1202-06.2006 - DOI - PMC - PubMed
1. Yousef H et al. Aged blood impairs hippocampal neural precursor activity and activates microglia via brain endothelial cell VCAM1. Nat. Med. 25, 988–1000 (2019). 10.1038/s41591-019-0440-4 - DOI - PMC - PubMed
1. Wang L et al. Peak Annotation and Verification Engine for Untargeted LC-MS Metabolomics. Anal. Chem. 91, 1838–1846 (2019). 10.1021/acs.analchem.8b03132 - DOI - PMC - PubMed
1. Adusumilli R & Mallick P Data Conversion with ProteoWizard msConvert. Methods Mol. Biol. 1550, 339–368 (2017). 10.1007/978-1-4939-6747-6_23 - DOI - PubMed

Publication types

Actions
Actions

MeSH terms

Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

TREM1 disrupts myeloid bioenergetics and cognitive function in aging and Alzheimer disease mouse models

Affiliations

TREM1 disrupts myeloid bioenergetics and cognitive function in aging and Alzheimer disease mouse models

Authors

Affiliations

Abstract

Conflict of interest statement

Figures

Similar articles

Cited by

References

Methods-Only References

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources

Full Text Sources

Medical

Abstract

Conflict of interest statement

Figures

Similar articles

Cited by

References

Methods-Only References

Publication types

MeSH terms

Substances

Related information

Grants and funding

LinkOut - more resources

Full Text Sources

Medical