Microglial glycolytic reprogramming in alzheimer's disease: association with impaired phagocytic function and altered vascular proximity

Abstract

Background: Alzheimer's disease (AD) is characterized by chronic neuroinflammation alongside amyloid-beta plaque and phosphorylated tau (p-Tau) tangle accumulation. Microglia, as resident immune cells, undergo glycolytic reprogramming that may exacerbate inflammation and impede toxic protein clearance. Specifically, the glycolytic enzyme pyruvate kinase M2 (PKM2) drives proinflammatory microglial phenotypes linked to neurodegeneration. This study investigates how PKM2-mediated microglial glycolytic reprogramming and inflammatory responses alongside Aβ/p-Tau clearance in human AD brains.

Methods and results: Hippocampal-entorhinal cortex (HP-EC) tissues from 8 AD patients and 8 matched controls underwent multiplex immunohistochemistry and high-resolution spatial analysis. PKM2⁺Iba1⁺ microglia density significantly increased in AD versus controls (p < 0.001), predominantly displaying a disease-associated microglial (HAM-like) phenotype (ABCA7⁺) with concurrent lipid-droplet accumulation (PLIN3⁺; LDAM phenotype). Spatially, glycolytic PKM2⁺Iba1⁺ microglia accumulated near Aβ plaques, p-Tau tangles, and cerebral vasculature. Notably, their distribution around plaques/tau showed anomalous increasing density with distance (p < 0.001), suggesting impaired chemotaxis. Perivascular localization lacked clear chemotactic gradients. Functionally, overall phagocytic activity (CD68⁺) decreased significantly in AD (p = 0.001), primarily attributed to PKM2- subsets, whereas PKM2⁺Iba1⁺ microglia exhibited pronounced phagocytic exhaustion (PLIN2⁺; p < 0.001), consistent around both Aβ and p-Tau lesions (all p < 0.001).

Conclusion: Our study establishes that microglial glycolytic reprogramming via PKM2 promotes a proinflammatory HAM-like phenotype, phagocytic exhaustion, and peri-pathological accumulation with aggregates and cerebral vessels. Targeting glycolytic pathways represents a viable therapeutic strategy for alleviating microglial dysfunction and neuroinflammation in AD.

Keywords: Alzheimer’s disease; Glycolysis; Microglia; Neuroinflammation; PKM2.

Fig. 1

Study overview. (a) Age and sex composition of the AD (n = 8) and control (n = 8) groups. (b) Schematic of the experimental workflow from tissue processing to imaging and analysis. (c) An example of annotations in brain slides. (d) Quantitative analysis of total Aβ plaque counts in AD patients and CTLs. (e) Quantitative analysis of total p-Tau counts in AD patients and CTLs. (f) Example merged mIHC images from Panel (1) DAPI (blue, f₁), Iba1 (green, f₂), PKM2 (red, f₃), ABCA7 (orange, f₄), PLIN3 (white, f₅), and TMEM119 (cyan, f₆). Scale bars = 50 μm. (g) Example merged mIHC images from Panel (2) DAPI (blue, g₁), Iba1 (cyan, g₂), PKM2 (orange, g₃), p-Tau (cyan, g₄), CD31 (red, g₅), and Aβ (white, g₆). Scale bars = 50 μm. (h) Example merged mIHC images from Panel (3) DAPI (blue, h₁), Iba1 (cyan, h₂), PKM2 (red, h₃), CD68 (orange, h₄), PLIN2 (white, h₅), p-Tau (cyan, h₆), and Aβ (magenta, h₇). Scale bars = 50 μm. All the data are presented as the means ± SEMs, and unpaired t tests were performed (**p < 0.01 and ***p < 0.001)

Fig. 2

Glycolytic microglia in AD exhibit a disease-associated phenotype in AD brains. (a-b) Representative mIHC images illustrating colocalization of PKM2⁺Iba1⁺ cells with ABCA7, PLIN3, and TMEM119 in the AD (a) and CTL (b) groups. Images a₁ and b₁ show the merged mIHC images. Images a₂–a₆ and b₂–b₆ display single-channel images corresponding to a₁, including DAPI (blue, a₂ and b₂), Iba1 (green, a₃ and b₃), PKM2 (red, a₄ and b₄), ABCA7 (orange, a₅ and b₅), PLIN3 (white, a₆ and b₆), and TMEM119 (cyan, a₇ and b₇). Images a₈ and b₈ represent the original fluorescence channels for Iba1, PKM2, ABCA7, and DAPI, with arrows highlighting ABCA7⁺PKM2⁺Iba1⁺ cells. Images a₉ and b₉ illustrate the corresponding segmentation masks for a₈ and b₈. Images a₁₀ and b₁₀ show original fluorescence channels for Iba1, PKM2, PLIN3, and DAPI, with arrows indicating PLIN3⁺PKM2⁺Iba1⁺ cells. Images a₁₁ and b₁₁ present the corresponding masks for a₁₀ and b₁₀. Images a₁₂ and b₁₂ display original fluorescence channels for Iba1, PKM2, TMEM119, and DAPI. Images a₁₃ and b₁₃ show the corresponding masks for a₁₂ and b₁₂. Scale bars = 50 μm. (c) Proportions of PKM2⁺Iba1⁺ cells within Iba1⁺ microglial population. (d) Comparison of ABCA7⁺PKM2⁺Iba1⁺, PLIN3⁺PKM2⁺Iba1⁺, and TMEM119⁺PKM2⁺Iba1⁺ proportions within Iba1⁺ microglial population. (e) Abundance of ABCA7⁺PKM2⁺Iba1⁺ microglia in AD vs. CTL brains. (f) Regional comparison of ABCA7⁺PKM2⁺Iba1⁺ microglia across HP‒EC in AD vs. CTL brains. (g) Abundance of PLIN3⁺PKM2⁺Iba1⁺ microglia in AD vs. CTL brains. (h) Regional comparison of PLIN3⁺PKM2⁺Iba1⁺ microglia across HP‒EC in AD vs. CTL brains. (i) Abundance of TMEM119⁺PKM2⁺Iba1⁺ microglia in AD vs. CTL brains. (j) Regional comparison of TMEM119⁺PKM2⁺Iba1⁺ microglia across HP‒EC in AD vs. CTL brains. All the data are presented as the means ± SEMs, and unpaired t tests were performed (*p < 0.05, **p < 0.01, ***p < 0.001, and ns: no significant difference)

Fig. 3

Glycolytic microglia show increased numbers and altered spatial distribution around Aβ in patients with AD. (a-b) Representative images from human HP‒EC sections showing PKM2⁺Iba1⁺ microglia in relation to Aβ plaques in the AD (a) and CTL (b) groups. Image a₁ and b₁ depicts the merged mIHC image, with arrows highlighting PKM2⁺Iba1⁺ microglia. Image a₂ and b₂ presents the corresponding mask, with concentric circles showing the distance to Aβ plaque (white) and highlighting PKM2⁺Iba1⁺ (magenta) and PKM2^–Iba1⁺ (cyan) cell masks. Images a₃–a₆ and b₃–b₆ show the single-channel images derived from a₁ and b₁, including DAPI (blue, a₃ and b₃), Iba1 (green, a₄ and b₄), PKM2 (orange, a₅ and b₅), and Aβ (white, a₆ and b₆). Scale bars = 50 μm. (c) Proportion of PKM2⁺Iba1⁺ cells among plaque-associated Iba1⁺ microglia in AD vs. CTL brains. (d) PKM2⁺Iba1⁺ microglia–plaque spatial relationships in AD vs. CTL brains. (e) Region–specific analysis of PKM2⁺Iba1⁺ microglia–plaque spatial relationships across HP‒EC in AD vs. CTL brains. All the data are presented as the means ± SEMs, and unpaired t tests, fixed-effects and mixed-effects models were performed (*p < 0.05, **p < 0.01, ***p < 0.001, and ns: no significant difference)

Fig. 4

Glycolytic microglia show altered spatial distribution and increased numbers around p-Tau in patients with AD. (a-b) Representative images from human HP‒EC sections showing PKM2⁺Iba1⁺ microglia in relation to p-Tau aggregates in the AD (a) and CTL (b) groups. Image a₁ and b₁ depicts the merged mIHC image, with arrows highlighting PKM2⁺Iba1⁺ microglia. Image a₂ and b₂ presents the corresponding mask, with concentric circles showing the distance to p-Tau aggregates (blue) and highlighting PKM2⁺Iba1⁺ (magenta) and PKM2^–Iba1⁺ (cyan) cell masks. Images a₃–a₆ and b₃–b₆ show the single-channel images derived from a₁ and b₁, including DAPI (blue, a₃ and b₃), Iba1 (green, a₄ and b₄), PKM2 (orange, a₅ and b₅), and p-Tau (cyan, a₆ and b₆). Scale bars = 50 μm. (c) Proportion of PKM2⁺Iba1⁺ cells among aggregate-associated Iba1⁺ microglia in AD vs. CTL brains. (d) PKM2⁺Iba1⁺ microglia–aggregate spatial relationships in AD vs. CTL brains. (e) Region–specific analysis of PKM2⁺Iba1⁺ microglia–aggregate spatial relationships across HP‒EC in AD vs. CTL brains. All the data are presented as the means ± SEMs, and unpaired t tests, fixed-effects and mixed-effects models were performed (*p < 0.05, **p < 0.01, ***p < 0.001, and ns: no significant difference)

Fig. 5

Glycolytic microglia show increased numbers and preferential localization near cerebral blood vessels in patients with AD. (a-b) Representative images from human HP‒EC sections showing PKM2⁺Iba1⁺ microglia in relation to blood vessels in the AD (a) and CTL (b) groups. Image a₁ and b₁ depicts the merged mIHC image, with arrows highlighting PKM2⁺Iba1⁺ microglia. Image a₂ and b₂ presents the corresponding mask, with concentric circles showing the distance to blood vessels (red) and highlighting PKM2⁺Iba1⁺ (magenta) and PKM2^–Iba1⁺ (cyan) cell masks. Images a₃–a₆ and b₃–b₆ show the single-channel images derived from a₁ and b₁, including DAPI (blue, a₃ and b₃), Iba1 (green, a₄ and b₄), PKM2 (orange, a₅ and b₅), and CD31 (red, a₆ and b₆). Scale bars = 50 μm. (c) Proportion of PKM2⁺Iba1⁺ cells among endothelia-associated Iba1⁺ microglia in AD vs. CTL brains. (d) Quantification of the average distance between PKM2⁺Iba1⁺ microglia and the nearest blood vessel in AD vs. CTL brains. (e) PKM2⁺Iba1⁺ microglia–endothelial spatial relationships in AD vs. CTL brains. (f) Region–specific analysis of PKM2⁺Iba1⁺ microglia–endothelial spatial relationships across HP‒EC in AD vs. CTL brains. All the data are presented as the means ± SEMs, and unpaired t tests, fixed-effects and mixed-effects models were performed (*p < 0.05, **p < 0.01, ***p < 0.001, and ns: no significant difference)

Fig. 6

Glycolytic microglia show impaired phagocytic function pattern in AD patients. (a-h) Representative images from human HP‒EC sections showing phagocytic function pattern of PKM2⁺Iba1⁺ microglia around Aβ and p-Tau in the AD (a-d) and CTL (e-h) groups. Image a₁–h₁ depicts the merged mIHC image, with arrows highlighting PLIN2⁺PKM2⁺Iba1⁺ and CD68⁺PKM2^–Iba1⁺ microglia. Image a₂–h₂ presents the corresponding mask, with a circle showing 50 μm to Aβ (magenta) and p-Tau (cyan) and highlighting CD68^–PKM2⁺Iba1⁺ (red), CD68⁺PKM2^–Iba1⁺ (orange), PLIN2⁺PKM2⁺Iba1⁺ (white), and PLIN2^–PKM2^–Iba1⁺ (green) cell masks. Rest images show the single-channel images derived from a₁–h₁, including DAPI (blue, a₃–h₃), Iba1 (green, a₄–h₄), PKM2 (red, a₅–h₅), CD68 (orange, a₆ and b₆, e₆ and f₆), PLIN2 (white, c₆ and d₆, g₆ and h₆), Aβ (magenta, a₆ and c₆, e₆ and g₆), and p-Tau (cyan, b₆ and d₆, f₆ and h₆). Scale bars = 50 μm. (i-k) Quantitive comparison of phagocytic function of microglia in the whole sections (i), around Aβ plaques (j), and around p-Tau aggregates (k) in AD vs. CTL brains. Image i₁–k₁ compared CD68⁺Iba1⁺ cells among Iba1⁺ microglia. Image i₂–k₂ compared CD68⁺PKM2⁺Iba1⁺ cells among PKM2⁺Iba1⁺ glycolytic microglia. Image i₃–k₃ compared CD68⁺PKM2^–Iba1⁺ cells among PKM2^–Iba1⁺ microglia. Image i₄–k₄ compared PLIN2⁺Iba1⁺ cells among Iba1⁺ microglia. Image i₅–k₅ compared PLIN2⁺PKM2⁺Iba1⁺ cells among PKM2⁺Iba1⁺ glycolytic microglia. Image i₆–k₆ compared PLIN2⁺PKM2^–Iba1⁺ cells among PKM2^–Iba1⁺ microglia. All the data are presented as the means ± SEMs, and unpaired t tests were performed (**p < 0.01, ***p < 0.001, and ns: no significant difference)