Introduction
Multiple sclerosis (MS) is a chronic demyelinating disease of the central nervous system (CNS), whose pathogenesis involves inflammatory and neurodegenerative processes. Current MS immunomodulatory treatments target CNS inflammation, while therapies capable of regenerating myelin and halting disease progression are lacking [
28]. Oligodendrocyte precursor cells (OPCs) are multipotent progenitor cells widely distributed in the CNS that could differentiate into mature oligodendrocytes (OLs) to sustain remyelination. In MS, impaired generation of OLs from OPCs leads to persistent demyelination, myelin debris accumulation, and axonal damage which clinically manifests as neurological disability [
11]. Efficient myelin debris removal and clearance by phagocytic cells are critical to eliminate inhibitory signals interfering with OPC activation, recruitment to the site of demyelination and/or differentiation into myelinating mature OLs [
14,
26,
27].
Microglial cells and infiltrating monocytes/macrophages can have a dual role in MS lesions. They could contribute to myelin damage and lesion expansion, or they may have a protective role by clearing myelin debris, reducing inflammation and secreting regenerative factors promoting remyelination [
48]. A critical modulator of microglia functions is the triggering receptor expressed on myeloid cells-2 (TREM2), an innate immune receptor expressed by several myeloid cells including brain microglia [
23,
45]. TREM2 is a phospholipid sensing receptor known to sustain microglial cell activation and expansion in response to demyelination or amyloid plaques in Alzheimer’s disease (AD) [
8,
42,
50]. Recently, TREM2 was proposed to be a key transcriptional regulator of cholesterol metabolism during chronic phagocytic activity for myelin clearance in response to demyelination [
35].
TREM2 binds on the membrane to the DNAX-activation protein 12 (DAP12) which is an adaptor protein required for TREM2 surface expression and intracellular signaling [
47]. TREM2 engagement leads to DAP12 phosphorylation, followed by recruitment and activation of the spleen-associated tyrosine kinase (SYK), resulting in downstream signaling events leading to proliferation, survival, phagocytosis, and secretion of cytokines and chemokines [
47]. Homozygous loss-of-function mutations in TREM2 or DAP12 genes cause Nasu–Hakola disease (NHD), a rare genetic disorder characterized by fatal presenile dementia and bone cysts [
38,
39]. Neuropathological findings in NHD include loss of myelin and axons in the brain, with reactive astrocytosis and microglial activation [
25]. Heterozygous TREM2 gene variants were reported to increase the risk for AD and other neurodegenerative diseases (frontotemporal dementia, Parkinson’s disease and amyotrophic lateral sclerosis) [
7,
16,
20,
43].
Here, we explored the effect of antibody-mediated TREM2 activation on microglia in a well-established toxin-induced model of demyelination in the CNS resulting from exposure to the copper chelator cuprizone (CPZ) in the diet [
31]. In this model, oligodendrocyte degeneration in the brain is followed by a robust microglial response consisting of activation, proliferation and clearance of myelin debris [
17]. These events lead to OPC recruitment, differentiation into mature OLs, and then remyelination which is almost complete within few weeks from toxin withdrawal [
31]. We show that treatment with a TREM2-agonistic antibody was able to enhance myelin debris clearance by microglia in vivo in the CPZ-model and by bone marrow-derived macrophages (BMDM) in vitro. Most importantly, these events resulted in increased OPC recruitment and differentiation into mature OLs eventually accelerating remyelination in vivo and preserving axonal health. This provides the proof of concept that antibody-mediated TREM2 activation could promote remyelination, suggesting this as a potential novel therapeutic avenue in MS and other demyelinating disorders.
Materials and methods
Mice
Trem2+/+,
Trem2+/,
−and Trem2−/− mice (backcrossed 12 generations to the C57BL/6 background) were obtained from Professor Marco Colonna (Washington University in St. Louis). These three strains were bred in parallel. Animals were housed in accordance with University and National Institutes of Health (NIH) guidelines, and animal protocols were approved by the Washington University Animal Studies Committee (study approval number: 20180056). Mice were maintained under controlled conditions (19–22 °C and a in a 12-h light/dark cycle with unrestricted access to food and water). Experiments performed by Alector (Fig.
2 and Supplementary Fig. 1a and c) used the
Trem2−/− mouse colony originally generated by the trans-NIH KnockOut Mouse Project (KOMP). Frozen sperms were obtained from the UC Davis KOMP repository, and a colony of mice was established at UC Davis.
Antibody generation
Monoclonal antibodies targeting mouse TREM2 were generated by immunizing mice genetically deficient of TREM2 (Trem2−/−) with recombinant TREM2 protein and hybridoma generation, as well as by a yeast display campaign (performed by Adimab). AL002a was screened for TREM2 specificity by selecting for binding to wild-type (Trem2+/+), but not Trem2−/− BMDM.
Mouse model of CPZ‑induced demyelination
Six- to eight-week-old Trem2+/+, Trem2+/−, and Trem2−/− mice were fed a 0.2% Bis-(cyclohexanone) oxaldihydrazone (cuprizone) diet (5C5N: Modified PicoLab® Rodent w/0.2% Cuprizone, TestDiet. Cuprizone from Alpha Aesar, A10628) for 4 weeks (WK 4) or for 4 weeks followed by 3 days (WK 4 + 3D), 7 days (WK 4 + 7D), or 14 days (WK 4 + 14D) on regular chow (PicoLab Rodent Diet 20 #5053, Purina). For the full duration of the experiment mice were injected intraperitoneally (i.p.) once a week with the anti-TREM2 antibody or the control antibody at a dose of 80 mg/kg. The first injection with the antibodies was performed 4 days before beginning the CPZ-diet.
Quantification of antibody levels in the brain
Six- to eight-week-old Trem2+/− mice were injected i.p. with AL002a. 48 h post injection, brains were removed after perfusion with PBS, micro-dissected to isolate the corpus callosum (CC) and the cortex (Ctx), and immediately frozen in liquid nitrogen. Tissues were solubilized in N-Per lysis buffer (87792, ThermoFisher) and cell protein content was measured using the Pierce bicinchoninic acid (BCA) protein assay kit (23227, Thermo Scientific). A mouse Trem-2b/Fc Chimera (R&D Systems) was used as a capture antibody and coated overnight at 4 °C on 96-well Meso Scale Discovery (MSD) plates in PBS. After washing, wells were blocked for 1 h at 37 °C with binding buffer (3% BSA in PBS). Titration of samples and standards were incubated for 1 h at room temperature (RT) on a shaker at 500 rpm. For detection (0.5 mg/ml) of Sulfo-TAG goat anti-mouse IgG (MSD) were added to the plate. After washing, read buffer was added and the plate was read on the Sector Imager. Washes between the different steps were done three times with 0.05% Tween 20 in PBS. Antibody content was normalized to serum and protein content.
Bone marrow-derived macrophages
Bone marrow-derived macrophages (BMDM) were obtained by flushing tibial and femoral marrow cells with cold PBS 2% FBS. Red blood cells were lysed using ACK lysing buffer (Thermo Fisher), and after two washes in PBS 2% FBS, the cells were re-suspended in complete media (RPMI, 10% FBS, Pen/Strep, l-glutamine, non-essential amino acid) with (50 ng/ml) murine M-CSF (m-M-CSF) to obtain differentiated macrophages after 6 days. Adherent macrophages were detached with 1 mM EDTA in PBS.
Immunoprecipitation
Immunoprecipitation in vitro: before stimulation, BMDM were starved for 4 h in RPMI with 1% FBS. 10 × 106 cells were incubated for 15 min at 4 °C with AL002a or control antibody (1 µg for 106 cells). Cells were then washed and incubated at 37 °C in the presence of goat-anti mouse IgG (1.5 µg for 1 × 106 cells). After stimulation, cells were lysed with lysis buffer (1% n-dodecyl-β-d-Maltoside, 50 Mm Tris–HCl (pH 8.0), 150 mM NaCl, 1 mM EDTA, 1.5 mM MgCl2, 10% glycerol, protease, and phosphatase inhibitors) and immunoprecipitated with an anti-TREM2 antibody that binds a different domain (rat anti-h/m-TREM2, clone 237920, R&D system). Immunoprecipitation in vivo: 6- to 8-week-old C57BL/6 mice were injected i.p. with 3 ml of 3% thioglycollate. After 3 days, when the peritoneal cavity was enriched with CD11b+F4/80+ macrophages expressing TREM2, mice were injected with control or TREM2-specific antibodies (40 mg/kg). 24 h after antibody injection, peritoneal macrophages were collected, immediately lysed in the lysis buffer previously described, and immunoprecipitated with rat anti-h/m TREM2 antibody described above (R&D System, clone 237920). Then, for both in vitro and in vivo experiments, precipitated proteins were fractionated by SDS-PAGE in non-reducing conditions, transferred to PVDF membranes, and probed with an anti-phosphotyrosine antibody (4G10, Millipore). TREM2 is not detected in non-reducing condition. To confirm that all substrates were adequately immunoprecipitated, whole cell lysates from each sample were also fractionated by SDS-PAGE in reducing condition and immunoblotted with an anti-actin antibody (actin, sc-47778 Santa Cruz).
Myelin production
Human myelin was prepared as previously described [
34] and stored in lyophilized form at − 80 °C. Prior to use, myelin was suspended in DMEM to a final concentration of (2 mg/ml) and dissolved by vortexing and sonicating. Myelin was then irradiated with 10,000 RADS to achieve sterility. Aliquots were stored at − 80 °C for further use.
NFAT-Luciferase reporter assay
A stable BW5147.G.1.4 (ATCC
® TIB48™) (BWZ) cell line expressing both mouse TREM2 and DAP12 was generously provided by the Seaman lab [
10]. This line was infected with a Cignal PLenti NFAT-Luciferase virus (Qiagen) to generate a stable mouse TREM2 reporter cell line, able to induce luciferase signaling upon TREM2 activation. The activity of the reporter was validated using PMA (0.05 μg/ml) and ionomycin (0.25 μM). To test whether TREM2 antibody induced signaling, 5 µg/ml of soluble AL002a, or control antibody was added to each well of 96-well culture plates together with 100,000 cell/well and incubated for 4–6 h at 37 °C in Dulbecco’s Modified Eagle Medium (DMEM). Luciferase activity was measured by removing media and adding 50 μl of PBS and 50 μl of OneGlo Reagent (Promega) to each well and incubating for 3 min at room temperature on a plate shaker to lyse the cells. Luciferase signal was measured using a BioTek plate reader. Data were analyzed using GraphPad Prism.
To test myelin-induced signaling, human myelin was diluted in PBS to 200 µg/ml, titrated onto a 96-well tissue culture plate, and incubated overnight at 4 °C. The next morning, the solution was removed, and plates were washed three times with 200 µl PBS. Plates were air dried and BWZ cells (with or without antibodies) were added, incubated, and analyzed as described above.
Myelin phagocytosis and degradation assays
BMDM obtained from Trem2+/− mice were seeded in 8-well chamber slides (35,000 cells/well) (154534 Nunc, Lab-Tek) in normal medium as described before. After 1 day, media was substituted with reduced FBS media (5% FBS) and macrophages were pre-incubated with AL002a or control antibody (10 µg/ml) 3 h before starting the experiments. For phagocytosis assays, BMDM were incubated with human myelin (20 µg/ml) for 30 min, 1 h, or 3 h, and after incubation, cells were washed in PBS and fixed in 4% PFA. For degradation assays, BMDM were incubated with human myelin (20 µg/ml) for 2 h, then washed thoroughly with PBS, and again put in culture in reduced FBS media. Cells were left in culture for 1 h, 24 h, or 48 h, then washed in PBS, and fixed in 4% PFA. For intracellular staining, cells were permeabilized and blocked for 60 min at RT in 5% horse serum and 0.1% saponin in PBS, and incubated at 4 °C overnight with primary antibodies Rt anti-MBP (Abcam, ab7349, 1:100) and Gt anti-Iba1 (Novus, NB100-1028, 1:250) diluted in PBS and 5% horse serum.
Mouse tissue processing and histological analyses
Mice were perfused with 4% paraformaldehyde. Mouse brains were removed and post-fixed in 4% PFA for 24 h, followed by immersion in 30% sucrose for 48 h, then embedded in Optimal Cutting Temperature (OCT). 5-μm sections were placed on glass slides and stained with solochrome cyanine to confirm the presence of a lesion as previously described [
22]. Sections were stained with the following primary antibodies: Rb anti-dMBP (Millipore, ab5864, 1:2000), Rb anti-Iba1 (Wako, 019-19741, 1:600), Gt anti-Iba1 (Novus, NB100-1028, 1:250), Rt anti-LAMP1 (Abcam, ab25245, 1:500), Rt anti-CD68 (Invitrogen, 14-0681-82, 1:300), Rb anti-PDGFRα (ThermoFisher, PA5-16742, 1:50), Rb anti-OLIG2 (Milipore, AB9610, 1:300), Ms anti-CNPase (Abcam, ab6319, :100), Shp anti-BrdU (Abcam, ab1893, 1:250), Rt anti-GFAP (ThermoFisher, 13-0300, 1:200), and Ms anti-SMI-31 (Biolegend, 801603, 1:1000). AlexaFluor-conjugated secondary antibodies (Invitrogen, 1:1000) were used. Some of the images were acquired with a Nikon Eclipse 90
i fluorescent and bright field microscope equipped with 10 × and 20 × zoom objectives and analyzed with Metamorph 7.7 software. CNPase, dMBP, and GFAP were analyzed as the percentage area of positive staining (number of positive pixels/mm
2) within the region of interest. Iba1, PDGFRα, BrdU, and OLIG2 were quantified as the density of cells in the region of interest (number of cells/mm
2). LAMP1 and CD68 were analyzed as the percentage area of LAMP1
+Iba1
+ and CD68
+Iba1
+ staining (number of positive pixels/mm
2) and then normalized on the percentage of Iba1
+ staining (number of positive pixels/mm
2) within the region of interest. For confocal analysis, images were acquired with an Olympus FV1200 laser scanning confocal microscope (Olympus-America Inc., Waltham, MA) equipped with a PlanApoN 60 ×, 1.4 NA super corrected oil objective. The Olympus FV1200 confocal microscope was equipped with five detectors: two spectral and one filter-based and two gallium arsenide phosphide (GaAsP) photo-multiplier tubes (PMTs). The 405-, 488-, and 559-nm diode lasers and 635-nm HeNe (helium neon) lasers were used with an optimal pinhole of 1 airy unit to acquire images. Images were finally processed with ImageJ and Imaris Software (Bitplane, Switzerland).
Transmission electron microscopy and g-ratio quantification
Mice were perfused with PBS, brains were removed, and immersion fixed in 2% PFA, 2.5% Glutaraldehyde, and 0.1 M PBS. 50 μm sagittal brain sections were cut using a vibratome, then fixed in osmium tetroxide in 0.1 M PBS (EMS, 19100), followed by dehydration in ethanol and infiltration of Spurr’s resin. Tissues were embedded using Spurr’s resin and aclar film. After polymerizing, the corpus callosum was dissected from the tissue and attached to a pre-made Spurr’s resin block, then sectioned using a DiATOME ultra 45° diamond knife and a LEICA Ultracut UC7. 90-nm sections were cut and picked up onto 200 hex mesh, formvar-carbon coated copper grids (Ted Pella, 01800-F), and stained with uranyl acetate and lead citrate. Images were captured using a JEOL 1200 EX II Transmission Electron Microscope with AMT digital camera. Remyelination was analyzed by counting the number of naked axons and the number of myelinated axons per field, with a minimum of ten fields being analyzed. The g-ratio was quantified by dividing the axonal diameter by the myelinated fiber diameter. Thirty myelinated axons were randomly analyzed across multiple fields per mouse to calculate the g-ratio.
In vivo myelin engulfment quantification
Fixed brain slices were permeabilized for 45 min at RT in PBS 0.1% Triton X-100, followed by 1 h RT in blocking solution (2% BSA 0.1% Triton X-100 in PBS) and overnight incubation with primary antibody for Rb anti-dMBP (Millipore, ab5864, 1:2000), Gt anti-Iba1 (Novus, NB100-1028, 1:250) and Rt anti-CD68 (Invitrogen, 14-0681-82, 1:300) at 4 °C. Upon washing, sections were incubated 2 h at RT with Alexafluor-conjugated secondary antibodies (Invitrogen, 1:1000). Images were acquired with an Olympus FV1200 laser scanning confocal microscope (Olympus-America Inc., Waltham, MA) with 2 × digital zoom, and a z-step size of 0.33 µm. Z-stacks ranged from 4 to 5 µm in thickness. Images were processed and analyzed by Imaris Software (Bitplane, Switzerland). CD68 and Iba1 volume was quantified by applying 3D surface rendering of confocal z-stacks in their respective channels, using identical settings (fix thresholds of intensity and voxel) within each experiment. Each confocal acquisition contained an equal number of images from the CC of mice treated with AL002a and control antibody. For quantification of dMBP engulfment by microglia, only dMBP signals present within microglial CD68
+ structures were considered. To this end, a new channel for ‘‘engulfed dMBP’’ was created, by using the mask function in Imaris, masking the dMBP signal within CD68
+ structures. Quantification of volumes for ‘engulfed dMBP in CD68’ was performed following the ‘3D Surface rendering of engulfed material’ protocol previously published [
44]. To account for variations in cell size, the amount of ‘engulfed dMBP in CD68’ was normalized to the total volume of the phagocyte in each field (given by Iba1
+ total volume). Total dMBP volume per field from the same confocal z-stacks was also quantified following the same protocol.
Neurofilament light detection
Mouse blood was collected into EDTA tubes (Sarstedt 201341102) with a capillary tube (Sarstedt 201278100), spun at 15,000×g for 7 min at 4 °C, and the top plasma layer was transferred to a 1.5 ml tube and stored at − 80 °C. Frozen plasma samples were thawed at room temperature, diluted tenfold, and run on a SIMOA HD-X (Quanterix) using the Simoa NF-light advantage kit (Quanterix 103186) according to the manufacturer’s protocol.
BrdU used as a marker of proliferation
Microglia and OPC proliferation in vivo was measured by 5-bromo-2′-deoxyuridine (BrdU) incorporation (Sigma, B5002). BrdU was administered (25 mg/kg) by intra peritoneal injection every 12 h starting 4 days before collecting the brains.
Microglia isolation and flow cytometry
Microglia isolation was carried out following a published protocol [
3] with modifications. The whole procedure was done on ice with cold buffers and centrifuges at 4 °C. Briefly, anesthetized mice were perfused intracardially with ice-cold Hank’s balanced salt solution (HBSS), and the CC and hippocampi were dissected under a stereotactic microscope. The different areas were mechanically homogenized and digested in ice-cold Accutase (Millipore) on a wheel for 20 min at 4 °C. After spinning the tubes for 1 min at 2000
g, pellets were resuspended in cold Hibernate buffer (Thermo Fisher). The cell suspension was then transferred to pre-chilled 15 ml tubes and passed through a pre-wet (with Hibernate buffer) 70 µm cell strainer (PluriStrainer Mini). Cell suspensions were then spun down at 300
g for 10 min and pellets were resuspended in PBS + 2% FBS, counted, and labeled with a combination of the following conjugated antibodies: CD11b-PeCy7 (clone M1/70), CD45-Alexa 700 (clone 30-F11), P2ry12-PE (clone S16007D), CD80-FITC (clone 16-10A1), CD86-BV421 (clone GL-1), and zombie acqua (Biolegend). Dead cells were excluded by selecting the zombie aqua negative cells. FACS analysis of the microglial profile was performed by gating CD11b
+CD45
int cells. Appropriate IgG isotype control antibodies were used for all staining. FACS analysis was performed on a FACS Fortessa machine (BD Biosciences), and data were analyzed with FlowJo Software (TreeStar).
Human tissue and analysis
Twenty fresh-frozen blocks of post-mortem CNS tissue from eight MS patients and four controls with non-neurological diseases were obtained from The Neuroinflammatory Disease Tissue Repository at Washington University St. Louis. Demographic and clinical characteristics of the donors of human brain tissues at the time of collection are indicated in Table
1. 5 μm sections were stained with Solochrome Cyanine and Oil Red O to look at myelin and lipid-laden macrophages, respectively. Active lesions were characterized as tissue areas with marked demyelination and the presence of lipid-laden macrophages. Tissues were then stained with Gt anti-human TREM2 (R&D Systems, AF1828, 1:200) and Rb anti-Iba1 antibody (Wako, 019-19741, 1:600). AlexaFluor-conjugated secondary antibodies were used (Invitrogen, 1:1000). Histological images were acquired using the Nanozoomer microscope at the Hope Center for Neurological Disorders at Washington University. For confocal analysis, images were acquired with a Zeiss LSM880 Airyscan laser scanning confocal microscope (Carl Zeiss Inc., Thornwood, NY) equipped with 63X, 1.4 numerical aperture (NA) Zeiss Plan Apochromat oil objective. The system is equipped with a unique scan head, incorporating a high-resolution galvo scanner along with two PMTs and a 32-element spectral detector. ZEN 2.3 black edition software was used to obtain Z-stacks through the entire height of the cells with confocal Z-slices of 5 µm (63 ×) and an interval of 0.347 µm. Images taken were optimized for 1 airy unit using the 405-nm diode, 488 nm Argon, and 561 nm diode and 633 m HeNe (helium neon) lasers. In addition, the Airyscan unit provides sub-diffraction limited imaging down to 120 nm resolution. Quantitative PCR analysis was performed on adjacent tissue sections.
Table 1
Human CNS tissues from autopsied multiple sclerosis and healthy control subjects
38 | 39/F | NA | 4 | CNS malignant lymphoma | NA | Cerebellum | Control |
10 | 41/M | NA | 24 | Heart failure | NA | Spinal cord Spinal cord Cerebrum | Control Control Control |
52 | 69/F | NA | 43 | Sepsis | NA | Cerebrum | Control |
32 | 56/F | NA | 18 | Acute myocardial infarction | NA | Spinal cord Cerebrum | Control Control |
46 | 77/F | Unknown | 16 | Not reported | SPMS | Spinal cord | MS NAWM |
45 | 66/F | 32 | 29 | Not reported | SPMS | Brainstem Cerebrum | MS NAWM MS NAWM |
77 | 41/F | 15 | 12 | Complication from Type 1 diabetes mellitus | RRMS | Brainstem Spinal cord Cerebrum | MS NAWM MS NAWM MS NAWM |
40 | 60/M | 14 | 9 | Respiratory failure | SPMS | Spinal cord | MS active |
30 | 54/F | 17 | 7 | Pneumonia | SPMS | Spinal cord | MS active |
71 | 50/F | 13 | 20 | Not reported | SPMS | Spinal cord Spinal cord | MS active MS active |
21 | 69/F | 29 | 6 | Metastatic colon cancer | PPMS | Spinal cord Spinal cord | MS active MS active |
59 | 54/F | 22 | 8 | Pneumonia | SPMS | Spinal cord | MS active |
Quantitative PCR
RNA was extracted using the RNeasy Micro Kit (Qiagen), converted into cDNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems), and used at 20–40 ng in quantitative real-time PCR (qPCR) analysis. The DDCt method was applied to determine differences in gene expression levels after normalization to the arithmetic mean of GAPDH as internal standards. The TaqMan probes used are the following: for human studies: GAPDH (Hh_99999905_m1) and TREM2 (Hs00219132_m1); for mouse studies: Gapdh (Mm99999915_g1); Olig2 (Mm01210556_m1); Mbp (Mm01266402_m1); Cnp (Mm01306641_m1); Mog (Mm00447824_m1); and Plp (Mm01297210_m1). All data are the mean of duplicates, and the standard errors of the mean was calculated between duplicates or triplicates. Real-time PCR was performed using an ABI 7000 Real-Time PCR System (Applied Biosystems).
Human macrophage cultures from Nasu–Hakola disease (NHD) patients
Peripheral blood mononuclear cells (PBMCs) were purified from peripheral blood samples on Ficoll-Paque PLUS density gradient (Amersham Biosciences, Piscataway, NJ). Samples were from three NHD patients and three control subjects, their characteristics at the time of collection are indicated in Table
2. To generate macrophages, PBMCs were seeded in RPMI without FBS for 2 h, then PBMCs were thoroughly washed with PBS, and cultured in complete media (RPMI, 10% FBS, Pen/Strep,
l-glutamine, non-essential amino acid) supplemented with (50 ng/ml) of recombinant human M-CSF (300-25, Peprotech) for 7 days at 37 °C in 5% CO
2. Cells were then harvested and immediately lysed in RLT Buffer; RNA was extracted using the RNeasy Micro Kit (Qiagen).
Table 2
Genetic and demographic characteristics of Nasu-Hakola and healthy control subjects included in gene expression analyses of monocyte-derived macrophages
NHD1 | C97T; homozygous | F | 46 | |
NHD2 | C97T; homozygous | M | 47 | |
NHD3 | 482 + 2T → C; homozygous | F | 53 | |
CTR1 | Normal TREM2 alleles | F | 60 | |
CTR2 | Normal TREM2 alleles | M | 38 | NA |
CTR3 | Normal TREM2 alleles | M | 38 | NA |
Microarray processing and analysis
In order to analyze the Affymetrix HuGene-1_0-st-v1 microarrays, CEL files were uploaded into Thermo Fisher’s Transcriptome Analysis Console (TAC) Software. We normalized the data in TAC and did differential expression analysis to compare gene expression in 3 control samples with 3 samples carrying TREM2 homozygous mutations. The differentially expressed genes were graphed in a volcano plot, with the − log10 of the P value on the Y-axis and the fold change on the X-axis. A twofold or greater change in gene expression and a P value of 0.05 were considered significant and colored red for up-regulated genes and green for down-regulated genes.
Statistical analysis
Data are displayed as individual dots and mean ± SEM. For each graph, the number of observations indicated with ‘‘n’’ and the number of biological replicates (mice) indicated with ‘‘N’’ can be found in the figure legends. Differences between multiple groups were analyzed by one-way ANOVA and a post hoc test (Tukey’s or Sidak’s or Dunnett’s). Comparisons between two groups following a normal distribution were analyzed using an unpaired t test (two-tail distribution) or a Mann–Whitney test when the distribution was not parametric, as indicated in each figure. Statistical analysis was performed using GraphPad Prism (Graph-Pad Software). Values were considered significant if P < 0.05.
Discussion
This study provides evidence of beneficial effects of treatment with the new agonistic anti-TREM2 antibody AL002a in the CPZ model of CNS demyelination. We showed that TREM2 is highly expressed on microglia/macrophages phagocyting myelin in MS active lesions. In CPZ-induced demyelination we further validated previous findings that loss of TREM2 reduces myelin clearance and we showed that loss of one copy of TREM2 has an intermediate phenotype, suggesting that myelin removal by microglia is sensitive to TREM2 copy numbers. We then demonstrated that treatment with AL002a in vivo promotes efficient clearance of myelin debris, with increased myelin phagocytosis and intracellular degradation after CPZ-induce demyelination. This was accompanied by increased microglia activation and by a higher expression of proteins involved in phagocytic and degradative pathways. Most importantly, AL002a treatment increased OPC density and differentiation into mature OL to support remyelination, thus finally leading to axonal health preservation. Notably, macrophages from NHD individuals with loss-of-function homozygous TREM2 mutations showed a defect in the phagocytic pathways, suggesting that findings from mouse studies may translate to humans.
Growing evidences suggest the important role of TREM2 in microglia function. We previously showed that
Trem2−/− mice have a striking defect in microglia activation and myelin debris clearance in the CPZ model [
8]. Similarly, lack of TREM2 in AD experimental models led to reduced microglia number, clustering and activation around CNS amyloid plaques accompanied by decreased plaque compaction and increased neuritic damage [
50,
51,
53]. In line with these observations, TREM2 was demonstrated to mediate the switch from a homeostatic to a disease-associated microglia (DAM) phenotype [
21]. TREM2 on microglia may also act as a lipid sensor capable of binding myelin debris with subsequent engulfment and clearance [
42]. Our data demonstrate that AL002a is a potent TREM2 agonist antibody capable of activating TREM2 in vitro and in vivo and to potentiate myelin-induced activation of TREM2 signaling. We further characterized the role of AL002a in the in vivo model of CPZ-induced CNS demyelination. First, we tested AL002a in
Trem2+/+ mice, without detecting any difference in damaged myelin accumulation, number of microglia or OPC in the CC compared to the control group at 4 weeks on CPZ (data not shown). These results were not unexpected given the rapid microglia response to damage, efficient clearance of myelin debris and subsequent remyelination after CPZ in wild-type mice [
17]. Next, we tested AL002a in
Trem2+/− mice, which had less efficient clearance of damaged myelin compared to their wild-type counterparts, giving us the opportunity to unveil any effect of antibody-mediated TREM2 activation on CPZ-induced pathology. Indeed, AL002a treatment in vivo significantly enhanced myelin debris clearance in
Trem2+/− mice compared to the control
Trem2+/− group. This was likely due to a direct effect of AL002a on microglia, which consistently was shown to express TREM2 [
45]. AL002a treatment in vivo by activating TREM2-dependent intracellular pathways led to more efficient myelin phagocytosis and degradation by microglia. This was accompanied by increased microglia expression of markers of activation (Iba1 and CD80) and phagolysosomal activity (CD68 and LAMP1). Similar effects were observed in BMDM in vitro. Our results clearly suggest that TREM2 plays a key role in myelin engulfment and intracellular processing by microglia. TREM2 has often been referred to as a phagocytic receptor as it was demonstrated to be involved in microglia phagocytosis of apoptotic neurons [
46], myelin [
8] and β-amyloid [
24]. However, our current and previous findings indicate a more complex TREM2 function not only in myelin uptake, but also in promoting myelin debris degradation through the phagolysosomal pathway [
8]. These findings were corroborated by the analysis performed in macrophages derived in vitro from NHD patients revealing that genes involved in the phagosome pathway are significantly altered in NHD subjects compared to healthy controls along with changes in immune response pathways. Therefore, myelin and axonal degeneration observed in NHD underscore a critical function played by microglia and TREM2 in myelin clearance, phagocytic functions, and maintenance of neuronal integrity.
We have previously reported that
Trem2−/− mice have a significant defect compared to
Trem2+/+ in microglia numbers and proliferation at 4 weeks on CPZ and this could contribute to the defect in removing myelin debris from the tissue [
8]. Interestingly, in the current study at this time point no differences in the number of microglia were detected between
Trem2+/− and
Trem2+/+ mice, despite more myelin debris were accumulating in the CC of
Trem2+/−. This could suggest that one copy of TREM2 is sufficient to sustain microglia expansion after CPZ-induced damage, but cannot support a fully functional response by microglia to clear the conspicuous amount of myelin debris derived from massive oligodendrocyte death. Alternatively, it is possible that a defect in microglia proliferation in
Trem2+/− mice is present, but at earlier time points (before 4 weeks) and in a very narrow time window. To this end, a very robust microglia activation and increase in density have been shown to start at 2–3 weeks after CPZ diet initiation, reaching a plateau at 4–5 weeks [
15]. Future studies are needed to clarify if TREM2 activation could also directly enhance microglia proliferation in vivo.
TREM2 has been described to be intimately linked to microglia lipid metabolism [
8,
42]. Lipids have been proposed as candidates for TREM2 ligands either as free molecules, complexed in myelin or in apolipoprotein particles [
1,
2,
42,
50]. A recent report demonstrated that TREM2 is highly expressed in the adipose tissue by lipid-associated macrophages (LAM) where it drives gene expression programs involved in phagocytosis and lipid metabolism [
19]. LAM cells in adipose tissue indeed expressed a highly similar gene profile as disease-associated microglia in AD with the exception of few tissue-specific genes [
19,
21]. Chronic demyelination in vivo (12 weeks on CPZ) caused a robust accumulation of cholesteryl ester (CE) and oxidized CE in the
Trem2−/– brain, suggesting that TREM2 might be a crucial transcriptional regulator of cholesterol transport and lipid metabolism in microglia [
35]. In support of a strong link between TREM2 and lipid metabolism are also the original reports describing NHD as a lipid storage disease due to a genetic enzymatic defect leading to lipid and cholesterol accumulation in the brain and bone cysts [
37]. Therefore, it is attractive to suggest that a major role for TREM2 in microglia would be intracellular processing and degradation of lipid-rich material (e.g. myelin, cell membranes), especially after extensive demyelination. In our previous work, we provided initial evidence to support this hypothesis by showing that TREM2 is not a limiting factor in myelin engulfment, but it was clearly essential for microglia capacity to degrade myelin during chronic demyelination [
8]. In the current study, we further support this hypothesis and we expand by showing that AL002a treatment significantly increases microglia capacity to phagocyte and degrade myelin in vivo and in vitro. These results confirm TREM2 as a key regulator of phagocytic clearance of myelin debris and lipid metabolism by microglia.
The identification of therapies promoting remyelination is a new frontier to overcome in MS. Approved disease-modifying treatments for MS are targeting the inflammatory disease component by reducing attack frequency and severity. Currently, no therapies are available to regenerate myelin and to halt MS disease progression. In MS multiple factors are thought to be involved in remyelination failure, starting with a deficiency in OPCs because of impaired recruitment, or incomplete differentiation and maturation in the MS lesions. Several lines of evidence indicate a key role of microglia in these processes. Microglia regenerative expression profile involves genes related to phagocytosis, breakdown of myelin debris, as well as secretion of regenerative factors and tissue remodeling [
36,
49] that can drive OPC differentiation [
29,
32]. Here we show that TREM2 was highly expressed within active MS lesions by lipid-laden microglia/macrophages known to display an alternatively activated M2 profile and to promote the resolution of inflammation and clearance of myelin debris [
5]. AL002a, by targeting and activating mouse TREM2, seems to enhance some of these pathways. It is not fully clear if enhanced remyelination in AL002a-treated
Trem2+/− mice is due to accelerated removal of myelin debris or if there is an additional TREM2-mediated signal from the microglia to recruit OPCs, thus giving an advantage in remyelination. Successive studies will further clarify potential microglia polarization or remodeling induced by AL002a. On the other end, it is well known that
Trem2+/+ mouse microglia are very efficient in supporting complete remyelination in the acute CPZ model. This is different from what observed in MS, where a defect in TREM2 pathways has not been demonstrated, but still remyelination fails in most cases. A possible hypothesis could be that, in MS, microglia are chronically challenged by recurring bouts of demyelination and remyelination, suggesting that they might become exhausted and less able to respond over time. In this situation, microglia could be more responsive to an increase in TREM2 signaling via antibody-mediated activation. In future studies we could examine this further by either chronically feeding CPZ or looking in aged mice in which microglia functions could be defective.
Importantly, our study has also shown that AL002a treatment after CPZ is associated with reduced Nf-L plasma levels (a markers of axonal/neuronal damage), suggesting that increased remyelination also results in preserved axonal integrity, further proven by increased SMI31 staining in the CC of AL002a-treated mice. Altogether our study indicates that strategies aimed at targeting TREM2 on microglia in the CNS are feasible and might be a promising intervention in MS to promote microglia functions in clearing myelin debris, favoring the recruitment of OPCs, and their subsequent differentiation into mature myelin-generating oligodendrocytes, eventually leading to remyelination and axonal protection.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.