Fig 2: Reactive microglial cells expressing Gal-3 presented Aβ inclusions in human tissue samples. a–f Gal-3-positive microglial cell associated with Aβ plaques. g 3D reconstruction of microglial cells with multiple Aβ inclusions inside. Gal3 (green), Aβ (red), Iba1 (white), DAPI (blue). White arrows are pointing to Aβ inclusions (in red) (n = 3 (HC), n = 8 (AD)

Fig 3: Reactive microglial cells expressing Gal-3 interact with p-Tau in senile plaques from human tissue samples. a–f Gal-3-positive microglial cell associated with p-Tau plaques. g 3D reconstruction of microglial cell with multiple p-Tau interactions. Gal3 (green), p-Tau (red), Iba1 (white) and DAPI (blue). White arrows are pointing p-Tau Gal-3 interactions (in orange). n = 3 (HC), n = 8 (AD)

Fig 4: Sensitivity of galectin gene expression to disruptions in O-GlcNAc homeostasis: (a) Changes in the expression of six galectin genes (LGALS1, LGALS3, LGALS8, LGALS9, LGALS10, and LGALS12) in HL-60 cells grown in IMDM medium supplemented with either FBS or ITS and treated with O-GlcNAc-modulating drugs as specified in Figure 1. Significant differences (p < 0.05) between treatments are labeled by different letters as per one-way ANOVA followed by a Tukey’s multiple comparisons test for each media. The results are presented as means ± SD, n = 3–5; (b) Pairwise correlation analysis of gene expression in HL-60 cells: left, heatmap showing pairwise correlation patterns between expression of genes (log2-transformed values) encoding galectins and cell differentiation marker NCF1. Significant (p < 0.05) Pearson’s correlation coefficients are reported for each pairwise comparison while blanks cells indicate that the correlation is not significant; right, comparison of pairwise Pearson’s correlations between IMDM-FBS and IMDM-ITS samples using the ‘cocor’ R package, significant p values are shown, ns—not significant, n = 15 for each gene (3 biological replicates for each treatment including control).

Fig 5: Galectin-3 association with astrocytes and Müller glia is greater in human glaucoma. A Retinal paraffin sections of a highly preserved cohort of human donor eyes were immunofluorescently labeled for IBA1, GFAP, and Galectin-3 and imaged at ± 2000 (central retina; shown here), ± 4000 (mid-peripheral retina; Fig S2A), and ± 6000 µm (peripheral retina; Fig S2A) from the optic nerve and volumes of markers and colocalization were calculated. (B-C) The volume of IBA1 did not increase significantly at any eccentricity in the retina of glaucoma eyes compared to controls, while the GFAP volume did in the central retina. (D-E) The Galectin-3/IBA1 colocalization did not increase significantly at any eccentricity in the retina of glaucoma eyes compared to controls, while the Galectin-3/GFAP colocalization did in the central retina. F Immunofluorescent labeling of Galectin-3 in human glaucomatous tissue reflects Müller glia morphology. G Single nucleus sequencing demonstrates that Müller glia have the highest average expression and highest percentage of cells expressing LGALS3 in the normal human retina in two independent datasets. H The percentage of Müller glia expressing LGALS3 by individual retina does not change with age, but LGALS3 expression per individual Müller glia is increased in older individuals. ns = P > 0.05, * = P < 0.05, ** = P < 0.01, *** = P < 0.001. For A and F, scale bar = 50 µm. Volumes in B-C are normalized to the area of the inner retina per image and n = retina (where the value is the average of the retina either side of the optic nerve head at the set eccentricity in glaucoma, and only the side opposite the tumour in controls)