Background
One of the hallmarks of Alzheimer’s disease (AD) is the extracellular deposition of amyloid-β (Aβ) peptide in the brain parenchyma as amyloid plaques. Autosomal dominant Alzheimer’s disease (ADAD) is an early-onset form of AD which is caused by rare mutations in amyloid β (A4) precursor protein (
APP), presenilin-1 (
PSEN1), or presenilin-2 (
PSEN2) that alter Aβ production [
1]. Genetic variants also influence the risk of developing the more common late onset form of AD (LOAD). To date the two strongest identified LOAD genetic risk factors are the well-studied apolipoprotein ϵ4 (
APOE4) allele and several recently identified variants in the triggering receptor expressed on myeloid cells-2 (
TREM2) gene [
2‐
4]. While these variants are not common, since
TREM2 variants strongly increase the risk of developing AD, understanding how TREM2 dysfunction affects AD pathology could yield novel therapeutic strategies.
TREM2 encodes a transmembrane protein possessing an extracellular IgG-like ligand binding domain and an intracellular region that associates with the immunoreceptor tyrosine based activating motif (ITAM)-containing signaling adaptor protein DAP12 [
5]. Individuals that are homozygous for loss of function mutations in either
TREM2 or
TYROBP (DAP12) suffer from polycystic lipomembranous osteodysplasia and sclerosing leukoencephalopathy (PLOSL) which is characterized by early onset dementia and cystic bone lesions [
6]. Within the brain, TREM2 is expressed by microglia and appears to regulate microglial-mediated phagocytic clearance of cellular debris and the inflammatory response of microglia to pathology, however the endogenous ligand(s) for TREM2 are unknown [
7‐
10]. TREM2 expression is increased in plaque-associated microglia in APP23 and TgCRND8 mice suggesting that TREM2 is involved in the microglial response to Aβ plaque deposition [
3,
11,
12]. The role of microglia in AD is complex and incompletely understood. Microglia rapidly migrate to Aβ plaque deposits and acquire an amoeboid “activated” morphology [
13,
14]. Pro-inflammatory M1-like microglial activation is generally considered neurotoxic, while pro-phagocytic M2-like activation can lead to microglial clearance of Aβ in murine AD models [
15]. Since TREM2 is implicated in regulating the phagocytic and inflammatory function of macrophages, TREM2 dysfunction could conceivably increase Aβ plaque burden through decreased phagocytic clearance of Aβ and/or promote a neurotoxic, inflammatory microglial phenotype in response to Aβ deposition.
In this study we tested whether loss of a single
trem2 allele affected Aβ plaque burden in APPPS1-21 mice[
16]. To facilitate analysis of microglia we took advantage of the CX3CR1-GFP mice which in the CNS express GFP specifically within microglia [
17]. Although we did not observe a significant difference in Aβ plaque deposition between TREM2
+/+ and TREM2
+/− mice, there was a substantial decrease in plaque-associated microglia in TREM2
+/− mice compared to TREM2
+/+ mice. These data suggest that TREM2 function may affect the microglial response to Aβ pathology.
Discussion
TREM2 variants, particularly the R47H mutation, strongly increase the risk of developing AD, however how TREM2 affects AD and AD pathology is unknown [
3,
4]. Here, we report a decrease in the number and size of plaque-associated microglia in 3-month old TREM2 Het mice as compared to TREM2 WT mice, suggesting that TREM2 regulates the microglial response to Aβ plaque deposition. To the best of our knowledge this is the first report of an observable microglial phenotype in hemizygous TREM2 mice. As the resident macrophages in the brain, microglia are hypothesized to mediate both a beneficial phagocytic clearance of Aβ from the brain, and a detrimental chronic inflammatory phenotype resulting in neurotoxicity [
20]. Longitudinal
in vivo imaging studies demonstrate that microglia rapidly form clusters around Aβ deposits, although the molecular determinants of microglial migration to Aβ deposits are poorly understood [
13,
14]. Plaque-associated microglia also assume an amoeboid morphology with larger cell somas than non-plaque associated microglia [
21]. Our data indicates that plaque-associated microglia in TREM2 Het mice are smaller than in TREM2 WT mice, which may indicate a defect in microglial activation.
We observed a reduced number of plaque-associated microglia in TREM2 Het mice compared to TREM2 WT mice. The reduced microglial response in TREM2 Het mice could result from defective microglial activation, migration, survival, or proliferation. Genetic network analysis of TREM2 expressed in the brain linked TREM2 to genes involved in regulating cytoskeletal rearrangements required for phagocytosis and migration [
22]. In the periphery TREM2-DAP12 signaling is important for chemotaxic macrophage migration to the lungs following exposure to cigarette smoke, supporting the hypothesis that TREM2 can regulate macrophage migration to sites of injury [
23]. Microglial proliferation also contributes to the population of plaque associated microglia [
21,
24]. TREM2 regulates macrophage-colony stimulating factor (M-CSF)-induced osteoclast precursor cell proliferation [
25]. Microglia express colony-stimulating factor 1 receptor (CSF1R) which is regulates both microglial proliferation and viability [
26‐
28]. Therefore, one potential explanation is that TREM2 is important for CSF1R-dependent responses to pathology. Further studies will be needed to characterize the mechanistic basis for how TREM2 regulates the number of plaque-associated microglia.
Despite the reduction in plaque-associated microglia, we did not observe any statistically significant difference in the expression of inflammatory cytokines or genes associated with microglial activation in TREM2 Het and TREM2 WT mice in either 3-month or 7-month old animals. TREM2-DAP12 signaling inhibits Toll-like receptor (TLR)-dependent cytokine production and bone marrow derived macrophages from TREM2 KO mice exhibit increased expressed inflammatory cytokine production in response to microbial stimulation [
29]. Similarly, knockdown of TREM2 expression in microglia co-cultured with apoptotic neurons resulted in increased production of TNFα and NOS2 [
8]. In contrast, TREM2 KO mice exhibited decreased inflammatory cytokine production compared to TREM2 WT mice in the middle cerebral artery occlusion model of stroke concomitant with decreased localization of activated microglia within the glial scar [
10]. Thus the overall effect of TREM2 dysfunction on inflammatory signaling may depend upon the precise pathological context. It is also important to note that the effects of TREM2 on cytokine production were described in the context of a complete loss of TREM2 function, such as occurs in PLOSL. TREM2 Het mice may retain sufficient TREM2 function to properly regulate cytokine production. One caveat to our study is that although we did not detect a compensatory upregulation of TREM2 at the mRNA level, we were unable to quantify TREM2 protein expression in brain lysate by western blot using currently available reagents. Therefore, we cannot exclude the possibility that TREM2 protein expression is post-transcriptionally modified to compensate for loss of TREM2 expression.
Although we did not detect a TREM2-dependent difference in Aβ plaque burden, another microglial-associated protein genetically associated with AD, CD33, appears to substantially influence Aβ deposition [
30‐
33]. CD33 appears to inhibit microglial uptake of Aβ
in vitro and genetic deletion of
CD33 in APP
SWE/PS1
ΔE9 mice reduces Aβ plaque burden [
30]. Furthermore, individuals possessing
CD33 variants that were associated with increased odds of developing AD exhibited higher CD33 expression and protective
CD33 variants resulted in lower CD33 expression [
30,
34]. Taken together, the effects of CD33 on microglial clearance of Aβ and the TREM2-dependent effects on plaque-associated microglia reported in this study, suggest that alterations in microglial function may impact different stages of AD pathogenesis.
Although we observed a strong decrease in microglial localization near Aβ plaques at 3 months, we did not observe a significant difference in Aβ plaque burden between TREM2 WT and TREM2 Het mice at either 3 or 7 months. One hypothesized function of plaque-associated microglia is to restrict the growth of Aβ plaque, which would imply that a decrease in plaque-associated microglia could result in larger Aβ plaques [
14]. However, a previous study demonstrated that a four-week ablation of microglia had no effect on Aβ plaque burden in APPPS1-21 or APP23 mice, suggesting that, over the short term, Aβ plaque growth was not significantly impacted by microglia [
35]. TREM2 is thought to promote microglial phagocytic activity, and therefore decreased functional TREM2 expression could result in reduced clearance of Aβ and a subsequent increase in plaque deposition [
8]. Although in this study we did not test the phagocytic function of TREM2, the lack of significant effect of TREM2 hemizygosity on Aβ plaque burden does not support the hypothesis that TREM2 regulates Aβ deposition. The discovery that variants in
TREM2 strongly increase the odds of developing not only AD, but also Parkinson’s disease, amyotrophic lateral sclerosis, and frontotemporal dementia underscores the important role that the innate immune system plays in neurodegenerative disease and suggests that TREM2 subserves a beneficial microglial response in a variety of pathologies [
36,
37].
Methods
Animals
APPPS1-21 transgenic mice (APP (KM670/671NL)/PS1 (L166P), gift of Mathias Jucker) were crossed with TREM2−/− x CX3CR1GFP/GFP mice or TREM2+/+x CX3CR1GFP/GFP mice to generate APPPS1-21 x TREM2+/− CX3CR1+/GFP (TREM2 Het) and APPPS1-21 x TREM2+/+ x CX3CR1+/GFP (TREM2 WT) mice. All mice were maintained on a C57BL/6 background and all animal work was in accordance with guidelines established by the Animals Studies Committee at Washington University.
Amyloid plaque analysis
Mice underwent transcardial perfusion with PBS (pH 7.4) followed by removal of the brain. Half the brain was fixed in 4% paraformaldehyde for 24 hours (4°C) and half was either frozen on dry ice and stored at −80°C for biochemical and qPCR analysis or processed to isolate microglial cells. Fixed hemibrains were cryoprotected in 30% sucrose in PBS (pH 7.4), frozen in dry ice, and serial coronal sections (50 μm thick) from the rostral anterior commissure to the caudal hippocampus were collected using a freezing sliding microtome. Three sections, 300 μm apart, were stained for Aβ using biotinylated HJ3.4 (anti-N-terminal Aβ antibody) and developed with DAB using a VECTASTAIN ABC Elite kit (Vector Labs) per manufacturer’s directions. To stain amyloid, three sections, 300 μm apart, were stained with X-34 dye (10 μM). HJ3.4 and X-34 stained sections were imaged using a NanoZoomer slide scanner (Hamamatsu Photonics) and the percent cortical area covered by HJ3.4 or X-34 staining was quantified by an experimenter blinded to the genotype and gender of the animal.
Microglial isolation
A single cell suspension was generated from mouse hemibrains using a neural tissue dissociation kit (Miltenyi Biotec, 130-093-231) and gentleMACS Dissociator (Miltenyi Biotec) according to manufacturer recommended protocols. Microglia cells were then enriched by labeling the cells with mouse CD45 MicroBeads (MIltenyi Biotec, 130-052-301) and subsequent purification using a magnetic column. Microglia cells were then FACS sorted based on the surface markers of CD45lo, CD11bhigh and GFP expression.
Real-time qPCR analysis
RNA was extracted from frozen cortical tissue using the RNeasy kit (Qiagen) or from adult microglia using the RNeasy Micro kit (Qiagen). Reverse transcription was performed using a High-Capacity cDNA Reverse Transcription Kit (Life Technologies). Real-time qPCR was conducted with TaqMan primers (Life Technologies)[
19] and the TaqMan Universal PCR Master Mix (Life Technologies) using an ABI Prizm 7500 thermocycler. Relative gene expression levels in TREM2 WT and TREM2 Het mice were compared using the ΔΔC
t method with β-actin used as a reference.
Biochemical analysis of insoluble Aβ levels
Cortical tissue was sequentially homogenized in PBS (pH 7.4) and 5 M guanidine-Tris buffer (pH 8.0) in the presence of protease inhibitors (Roche). Aβ40 and Aβ42 levels were quantitatively measured by sandwich ELISA using either HJ2 (anti-Aβ35–40) or HJ7.4 (anti-Aβ37–42) as capture antibodies and biotinylated HJ5.1 (anti-Aβ13–28) as the detection antibody. Following incubation with poly-horseradish peroxidase-20 (Fitzgerald) ELISAs were developed using Super Slow ELISA TMB (Sigma).
Microglia quantification
Alexa568-HJ3.4-stained brain sections were imaged using a 40x water-immersion objective (Zeiss, NA = 1.2) on a Zeiss LSM5 confocal microscope. All images were acquired and analyzed by an experimenter blinded to the genotype of the animal. Z-series stack images of randomly selected plaques within the lateral half of the cortex located above the hippocampus were then sequentially acquired for Alexa568 and GFP fluorescence (~12 optical sections, 3 μm apart). All images were acquired using identical acquisition parameters as 8-bit, 1024 × 1024 arrays. Z-series stacks were then converted to maximum intensity projections and threshold adjusted to isolate specific GFP fluorescence. Plaque-associated microglial coverage was assessed by measuring the percent area covered by GFP fluorescence within 20 μm of the edge of the plaque, including the area of the plaque itself. To assess the number and size of plaque-associated microglia, thresholded images were segmented using a watershed function and the number and area of microglia assessed in ImageJ using a minimum size cut-off of 16 μm2.
Cytokine analysis
Cortical tissue from 3-month old APPPS1-21 x TREM2+/+x CX3CR1+/GFP and APPPS1-21 x TREM2+/− CX3CR1+/GFP was homogenized in 9x volumes lysis buffer (50 mM Tris–HCl (pH7.4), 2 mM EDTA, protease inhibitors). Lysates were centrifuged for 2 min at 13,000xg and analyzed using the Rodent Cytokine Multi-Analyte Profile (Myriad RBM).
Statistics
Amyloid plaque immunohistochemistry and insoluble Aβ levels between male and female TREM2 WT and TREM2 Het mice were statistically analyzed using 2-way ANOVA (α = 0.05). The number, soma size, and percent-area covered by plaque-associated microglia were compared using a Mann Whitney test. RT-qPCR results from TREM2 WT and TREM2 Het groups were compared by t-test using a Benjamini-Hochberg correction for multiple comparisons. P-values less than 0.05 were considered statistically significant.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
DMH co-founded and is on the scientific advisory board of C2N Diagnostics and currently serves as a consultant for Astra Zeneca, Eli Lilly, and Genentech.
Authors’ contributions
Tissue immunohistochemistry and X-34 staining were performed by JDU, MBF, AS, TEM, FRS, and HJ. Aβ and X-34 staining were quantified by MBF and JDU. Microglial isolation was performed by YW and JDU. Tissue biochemistry was performed by MBF and JDU. Microglial localization was quantified by JDU. RT-qPCR was performed by JDU and AS. Experiments were conceived and designed by JDU, DMH, MC, and LP. Manuscript was written by JDU and critically reviewed by DMH, MC, and LP. All authors read and approved the final manuscript.