Engagement of TREM2 by a novel monoclonal antibody induces activation of microglia and improves cognitive function in Alzheimer's disease models

Abstract

Background: Genetic variants and mutations in triggering receptor expressed in myeloid cells (TREM2) are associated with premature and late onset Alzheimer's disease (AD).

Methods: We developed a panel of monoclonal antibodies, the selected lead of which was avidly shown to bind the extracellular domain of human and murine TREM2.

Results: By engaging membrane-bound TREM2, the selected antibody was shown to promote their cellular proliferation, uptake of oligomeric beta amyloid/apoptotic neurons, and activation in a Syk and Akt dependent manner. The antibody was shown to avidly bind soluble TREM2 in the CSF from AD patients and blunted the proinflammatory program driven by its intracerebral injection. Upon in vivo treatment, the antibody was shown to improve cognitive function in experimental amyloidopathy models and to facilitate plaque-associated microglial coverage and activation.

Conclusion: Thus, we describe a novel monoclonal antibody targeting membrane bound and soluble TREM2, that improves cognitive function by inducing microglial activation and attenuating chronic neuroinflammation.

Keywords: Alzheimer’s disease; Monoclonal antibody; Mouse model; Neuroinflammation; TREM2.

Fig. 1

Binding of Ab-T1 to TREM2. a Binding of the mouse IgG antibodies Ab-T1, Ab-T2, Ab-T3, Ab-T4, Ab-T5, and control anti human TREM2 Ab to human and mouse TREM2 expressed in HEK293T cells by dot blot assay. Control naïve HEK293T cells (control) used as negative control. Binding affinity sensograms of Ab-T1 to (b) human TREM2 and (c) mouse TREM2 as measured by surface plasmon resonance (BiaCore). d Ab-T1 binding affinity table; “Ka” refers to the association rate constant; “kd” refers to the dissociation rate constant; and “KD” refers to the affinity constant. e Confocal images of Ab-T1 immunostaining non-permeabilised HEK293T transfected with wild type hTREM2-GFP (upper panels) and parental HEK293T (lower panels). Nuclei stained with DAPI. f Human TREM2 expression levels on transfected HEK293T cells recognized by Ab-T1. The y-axis represents the relative mean fluorescence intensity (relative MFI) measured by flow cytometry. g Western blot showing human TREM2 detection on stable U937 cell line recognized by Ab-T1

Fig. 2

Recognition of TREM2 in AD models/human by Ab-T1. a Western blot showing Ab-T1 binding to human entorhinal cortex extracts from Alzheimer/control group patients. HEK293T protein lysate was used as negative control (NC). GAPDH as “housekeeping” protein loading control is shown in lower panels. b Immunohistochemistry staining showing Ab-T1 binding to human brain sample (entorhinal cortex sections) from Alzheimer’s disease patient (TREM2) with microglia and beta amyloid staining of same human brain tissue sample. Mouse IgG was used as negative control staining. c Confocal microscope scan images showing co-localization of TREM2 (mouse Ab-T1) with resident Iba1 positive cells (Microglia) in 5xFAD mice brain slices (white arrows). d Immunohistochemistry staining showing Ab-T1 binding to brain tissue from 5xFAD mice (TREM2) with microglia and beta amyloid staining of same mice tissue sample. Mouse IgG was used as negative control. e Western blots of supernatants (left panel) for soluble TREM2 detection in parental HEK293T vs. HEK293T-hTREM2 cells using mouse Ab-T1 and mouse IgG1 as control Ab. Transfected HEK293T cells with no insert DNA vector were used as sham. f Western blots of soluble hTREM2 detection in human CSF from Alzheimer patients using Ab-T1 and mouse IgG1 as control IgG antibody. TREM-ECD represents soluble TREM2 recombinant protein control

Fig. 3

Ab-T1 augments uptake of beta amyloid in a Syk-dependent manner and reduces inflammation triggered by IC delivery of soluble TREM2. Ab-T1 increases uptake of labeled oligomeric beta amyloid a in human microglia derived from differentiated human PBMC’s, b murine peritoneal macrophages, and c murine microglia in a Syk dependent manner (R406). The y-axis represents the relative geometric mean (gMFI) measured by flow cytometry. Results are mean of three repeated experiments with duplicate technical sample repetition. d Intracerebral delivery of sTREM2 with Ab-T1 attenuates the neuroinflammatory response evident by delivery of sTREM2 with control IgG (n = 7, two-tailed student’s t test). The y-axis represents the fold-change in expression relative to sham (PBS injected animals)

Fig. 4

Effects of Ab-T1 on Syk/Akt phosphorylation and the activation of microglia and macrophages. a Ab-T1 promotes Syk phosphorylation and b Akt phosphorylation. Western blot analysis of phosphorylated Syk (Tyr525/526) or Akt (Ser473) in mouse bone marrow-derived macrophages stimulated with Ab-T1. Ab-T1 upregulates (c) TNFα and (d) IL-1b mRNA and TNFα protein levels in microglia dependent on e Syk or f Akt phosphorylation as the respective inhibitor attenuates consequent Ab-T1-induced activation. Data shown in ELISA assay represent triplicate repeats. g Ab-T1 enhances the uptake of labeled apoptotic neurons by microglia as compared to control IgG. Data shown represent triplicate experimental repeats with triplicate sample repetition

Fig. 5

Ab-T1 improves cognition and attenuates neuroinflammation in AD 5xFAD Mice. a Five-month-old female mice treated with sham (PBS) and human IgG (10 mg/kg) showed cognitive impairment as indicated by the discrimination Index (DI) for the object recognition testing session (NOR) compared to Ab-T1 (10 mg/kg) treated mice (n = 9–10, two-tailed student’s t test, **P = 0.0019, *P = 0.0027) (b) and by the increase in latency to platform in the MWM probe test (two-tailed student’s t test, **P = 0.006, *P = 0.029). Error bars represent standard error of the mean. c TNFα, IL-1b, and IL-6 mRNA levels were detected using TaqMan real-time PCR in purified brain homogenates. The y-axis represents the fold-change in expression compared to no treatment (sham). GAPDH was used as “housekeeping” gene (two-tailed student’s t test, **P = 0.01, *P = 0.016). d Discrimination Index (DI) for the object recognition testing session (NOR) after 8 weeks of treatment dysfunction in young 5xFAD mice. (n = 7, two-tailed student’s t test, *P = 0.033). The result in the NOR can vary between + 100 and − 100, where a positive score indicates more time spent with the novel object, a negative score indicates more time spent with the familiar object, and a zero score indicates a null preference

Fig. 6

Effects on Ab-T1 on beta amyloid levels and target engagement in 5XFAD mice. a Ab-T1 treatment is associated with a decrease in total brain derived beta amyloid. b Beta amyloid plaques in mice brain sections were confirmed using thioflavin S fluorescent staining. c, d Images taken on microscope were analyzed using ImageJ software for mean number of plaques per mice brain section area and average plaque size (n = 9–10, two-tailed student’s t test, **P = 0.001, *P = 0.032). e Number of diffuse plaques in the cortex of Ab-T1 vs. control-treated 5xFAD mice (two-tailed student’s t test, **P < 0.005). f Ab-T1 induces a reduction of soluble TREM2 in CSF and serum (two-tailed student’s t test, *P = 0.047). g Ab-T1 levels in serum and brain of 5xFAD treated mice. (h; upper panels) Correlation of serum and brain levels of Ab-T1 with cognition tested in MWM (r = 0.46; P < 0.05 and r = 0.5; P < 0.05). (h; lower right) Association between serum and brain levels in the same treated mice (r = 0.89; P < 0.001). (h; lower left) Association of free sTREM2 levels achieved by Ab-T1 treatment, and cognition (r = 0.61; P < 0.05). Data shown in ELISA assays is the mean of triplicate sample repetition

Fig. 7

The effects of Ab-T1 on the number and activation status of plaque associated microglia. Ab-T1 induces an increase in the number of microglia activated in the vicinity of the plaques. a Hippocampus and cortex region representative micrographs of brain section of 5xFAD Alzheimer’s disease mice treated with Ab-T1 and control IgG (experiment 1). Black squares indicate an example of randomly selected 80 × 80 μm counting frame at × 20 magnification. b Enlarged images (× 20) of randomly selected frames from representative micrographs. c The average number of plaque associated-microglia in the hippocampus and cortex regions of treated mice. d A more efficient plaque coverage by a higher number of in situ associated microglia (average plaque-associated microglia; upper panels) based on different stages of microglia activation defined by their morphology (activated microglia—stage II and III) illustrated in the lower panels. (c, two-tailed student’s t test, **P = 0.0006, *P = 0.00024), (d, two-tailed student’s t test, *P = 0.0000028, **P = 0.00000065). Scale bar = 200 μm (a) and 50 μm (b). The immunohistochemistry studies were conducted employing double staining with anti-beta amyloid (4G8; red) and anti-Iba1 (brown) antibodies. e Confocal microscope images showing co-localization of CD68 with Iba1 positive cells in 5xFAD mice brain slices. f Ab-T1 but not control IgG augments BMDM induced proliferation. Cell proliferation was measured by flow cytometry (quantifying cell population growth). Results are mean of four repeated experiments with triplicate technical sample repetition