Galectin-3, a novel endogenous TREM2 ligand, detrimentally regulates inflammatory response in Alzheimer's disease

Abstract

Alzheimer's disease (AD) is a progressive neurodegenerative disease in which the formation of extracellular aggregates of amyloid beta (Aβ) peptide, fibrillary tangles of intraneuronal tau and microglial activation are major pathological hallmarks. One of the key molecules involved in microglial activation is galectin-3 (gal3), and we demonstrate here for the first time a key role of gal3 in AD pathology. Gal3 was highly upregulated in the brains of AD patients and 5xFAD (familial Alzheimer's disease) mice and found specifically expressed in microglia associated with Aβ plaques. Single-nucleotide polymorphisms in the LGALS3 gene, which encodes gal3, were associated with an increased risk of AD. Gal3 deletion in 5xFAD mice attenuated microglia-associated immune responses, particularly those associated with TLR and TREM2/DAP12 signaling. In vitro data revealed that gal3 was required to fully activate microglia in response to fibrillar Aβ. Gal3 deletion decreased the Aβ burden in 5xFAD mice and improved cognitive behavior. Interestingly, a single intrahippocampal injection of gal3 along with Aβ monomers in WT mice was sufficient to induce the formation of long-lasting (2 months) insoluble Aβ aggregates, which were absent when gal3 was lacking. High-resolution microscopy (stochastic optical reconstruction microscopy) demonstrated close colocalization of gal3 and TREM2 in microglial processes, and a direct interaction was shown by a fluorescence anisotropy assay involving the gal3 carbohydrate recognition domain. Furthermore, gal3 was shown to stimulate TREM2-DAP12 signaling in a reporter cell line. Overall, our data support the view that gal3 inhibition may be a potential pharmacological approach to counteract AD.

Keywords: Alzheimer’s disease (AD); Amyloid aggregation; Galectin-3; Inflammation; Microglia; TREM2.

Fig. 1

Galectin-3 is increased in human and mouse AD brains and marks a microglial phenotype associated with Aβ plaques. a Western blot analyses of cortex from human AD cases (n = 6) and age-matched healthy controls (n = 5). b Galectin-3 (gal3) staining mainly coincided with Iba1⁺ microglia found around Aβ plaques in human AD brains. c Immunohistochemistry showed high levels of gal3 in microglia in AD brains, as compared to Iba1-staining (low levels of gal3 was detected in association to blood vessels). d Gal3 protein was significantly upregulated in the cortex of 5xFAD mice in a time-dependent fashion (WT, 6 months old). e Gal3 expression was found in Iba1⁺ cells around Αβ plaques. Statistical significance was calculated by one-way ANOVA with Tukey’s correction (d) or Student’s t test (a). *p < 0.05; **p < 0.01. Data are shown as mean ± SD. The human AD cases are described in suppl. Tables S1 and S2 (online resources 8 and 9)

Fig. 2

Galectin-3 deficiency/inhibition reduces the microglial inflammatory response in vitro. a Reduced cytokine levels in culture medium from Gal3KO primary microglial cultures compared to WT after fΑβ treatment for 12 h. b WT primary microglial cultures increase the release of gal3 upon stimulation with fAβ. In vitro experiments represent a minimum of three independent experiments. Statistical significance was calculated by Student’s t-test (b), or one-way ANOVA with Tukey’s correction (a) *p < 0.05; **p < 0.01; ***p < 0.001. Data are shown as mean ± SD

Fig. 3

Galectin-3 reduces the microglial inflammatory response in vivo mainly through TLR and DAP12 pathways. a Heat map of the 100 most upregulated inflammatory genes in 18-month-old 5xFAD mice vs. WT and how these genes are altered in 5xFAD/Gal3KO mice. Data expressed in Log2FC. b Inflammatory pathways affected in hippocampi of 5xFAD/Gal3KO mice compared to 5xFAD mice. c Expression of homeostatic and proinflammatory microglial genes in 5xFAD and 5xFAD/Gal3KO using qPCR in aged mice (18 months). d Main TLR-associated genes and disease-associated microglia (DAM) genes affected in aged mice (18 months); red triangle is 5xFAD reference gene expression and blue square is 5xFAD/Gal3KO gene expression. Data are expressed in folds ΔCT. Statistical significance was calculated by one-way ANOVA with Tukey’s correction. *p < 0.05. Data are shown as mean ± SD

Fig. 4

Lack of galectin-3 reduces AD pathology in 5xFAD mice. Aβ load in hippocampus (CA1) and thalamus at 6 months (a) and 18 months (b) with Aβ load analyzed in 5xFAD and 5xFAD/Gal3KO mice (% in frames). Soluble fractions of Αβ40 and Αβ42 levels in cortical S1 fraction were measured (pg/ml). c Αβ40 and Αβ42 levels in CSF samples at 6 months (pg/mL). d IDE-1 levels in 5xFAD and 5xFAD/Gal3KO mice at 6 and 18 months. e Spatial memory analyzed by Morris water maze test. Center, distance traveled (meters) as an integrated distance (AUC area under the curve). Left, latency to the platform during training trials. Right, representative runs of probe trial day. Q1, quadrant where the platform was located on training trial. Statistical significance was calculated by two-way (with Bonferroni’s post hoc) (e) and Student’s t-test (a–e). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Data are shown as mean ± SEM

Fig. 5

Galectin-3 colocalizes with TREM2. a, b Iba1⁺ cells expressing galectin-3 (gal3) around Aβ plaque in 5xFAD mice. c Reduced number of Iba1⁺ microglial cells around Αβ plaques in 5xFAD/Gal3KO mice compared to 5xFAD mice (% of Αβ area). d Number of Iba1⁺ cells expressing gal3 in 5xFAD (% of Αβ area). e, f Gal3 and TREM2 in plaque-associated microglia in the brain of 5xFAD mice reveals colocalization of gal3 and TREM2. g Gal3 and TREM2 colocalization in 5xFAD mouse brain using STORM microscopy. Statistical significance was calculated by Student’s t test. *p < 0.05. Data are shown as mean ± SEM. All images were taken in 5xFAD mice at 18 months

Fig. 6

Galectin-3 interacts with TREM2 through its carbohydrate-binding domain. a Fluorescent anisotropy assay for galectin-3 (gal3)/TREM2 interaction. Data are presented as % of TREM2–gal3 binding (gal3, WT and mutant gal3 with deficient carbohydrate-binding domain, R186S) and fluorescent probe interaction, by increasing concentrations of TREM2, together with the calculated K_d values for the gal3/TREM2 interaction (n = 2). b Control and DAP12 reporter cell lines were stimulated with increasing concentrations of gal3 (250 nM–2.5 µM), ionomycin and phosphatidylserine (PS). Statistical significance was calculated by Student’s t test or one-way ANOVA with Bonferroni’s post hoc test. *p < 0.05; **p < 0.01. Data are shown as mean ± SD

Fig. 7

Galectin-3 induces the formation of insoluble Αβ aggregates following injections of Αβ monomers in the hippocampi of WT mice. Αβ monomers were injected with galectin-3 (Gal3) (Aβ + Gal3) or without (Aβ) after 1 h Aβ monomers incubation w/o gal3 into the right or left hippocampi of WT mice, respectively. a Staining for Αβ and gal3 in the left hippocampus (only Aβ monomers injected). b, c Staining for Αβ and gal3 in the right hippocampus (Aβ + gal3 injected). Dashed frames in b are magnified and shown in c. d Thioflavin-S and Αβ staining of the left hippocampus (only Aβ injected). e Thioflavin-S and Αβ staining of the right hippocampus (Aβ + gal3 injected). White arrows point to gal3 and thioflavin-S⁺ aggregates. f, g Iba1 and GFAP immunoreactivity in the right hippocampus (Αβ + gal3 were injected)