The protective PLCγ2-P522R variant mitigates Alzheimer's disease-associated pathologies by enhancing beneficial microglial functions

Abstract

Background: Phospholipase C gamma 2, proline 522 to arginine (PLCγ2-P522R) is a protective variant that reduces the risk of Alzheimer's disease (AD). Recently, it was shown to mitigate β-amyloid pathology in a 5XFAD mouse model of AD. Here, we investigated the protective functions of the PLCγ2-P522R variant in a less aggressive APP/PS1 mouse model of AD and assessed the underlying cellular mechanisms using mouse and human microglial models.

Methods: The effects of the protective PLCγ2-P522R variant on microglial activation, AD-associated β-amyloid and neuronal pathologies, and behavioral changes were investigated in PLCγ2-P522R knock-in variant mice crossbred with APP/PS1 mice. Transcriptomic, proteomic, and functional studies were carried out using microglia isolated from mice carrying the PLCγ2-P522R variant. Finally, microglia-like cell models generated from human blood and skin biopsy samples of PLCγ2-P522R variant carriers were employed.

Results: The PLCγ2-P522R variant decreased β-amyloid plaque count and coverage in female APP/PS1 mice. Moreover, the PLCγ2-P522R variant promoted anxiety in these mice. The area of the microglia around β-amyloid plaques was also increased in mice carrying the PLCγ2-P522R variant, while β-amyloid plaque-associated neuronal dystrophy and the levels of certain cytokines, including IL-6 and IL-1β, were reduced. These alterations were revealed through [18F]FEPPA PET imaging and behavioral studies, as well as various cytokine immunoassays, transcriptomic and proteomic analyses, and immunohistochemical analyses using mouse brain tissues. In cultured mouse primary microglia, the PLCγ2-P522R variant reduced the size of lipid droplets. Furthermore, transcriptomic and proteomic analyses revealed that the PLCγ2-P522R variant regulated key targets and pathways involved in lipid metabolism, mitochondrial fatty acid oxidation, and inflammatory/interferon signaling in acutely isolated adult mouse microglia and human monocyte-derived microglia-like cells. Finally, the PLCγ2-P522R variant also increased mitochondrial respiration in human iPSC-derived microglia.

Conclusions: These findings suggest that the PLCγ2-P522R variant exerts protective effects against β-amyloid and neuronal pathologies by increasing microglial responsiveness to β-amyloid plaques in APP/PS1 mice. The changes observed in lipid/fatty acid and mitochondrial metabolism revealed by the omics and metabolic assessments of mouse and human microglial models suggest that the protective effects of the PLCγ2-P522R variant are potentially associated with increased metabolic capacity of microglia.

Keywords: Alzheimer’s disease; Lipid droplets; Microglia; PLCγ2-P522R variant; Phospholipase C gamma 2; β-amyloid pathology.

Fig. 1

PLCγ2-P522R variant increases microglia activation in the brain of 13-month-old female and male APP/PS1 mice. A Illustration of the PLCγ2 protein domain and the location of the PLCγ2-P522R variant in the protein domain. Created in https://BioRender.com. B-C Example positron emission (PET) images of the uptake (percent of the injected dose per ml of tissue) of [18F]FEPPA for both APP/PS1 (A+/P^wt/wt) and APP/PS1xPLCγ2-P522R (A+/P^ki/ki) mice at (B) 7 months and (C) 13 months of age with an anatomical CT reference image in the background. [18F]FEPPA uptake remains unaltered between the genotypes at the 7-month timepoint. A+/P^wt/wt n = 12 and A+/P^ki/ki n = 17. Compared with that in A+/P^wt/wt mice, [18F]FEPPA uptake in the whole brain (*p = 0.032), pons (*p = 0.013), and hippocampus (*p = 0.012) was significantly greater in 13-month-old A+/P^ki/ki mice. A+/P^wt/wt n = 10 and A+/P^ki/ki n = 13. The colored circles indicate data obtained from female mice, and the hollow circles indicate data obtained from male mice. The data were normalized to those of A+/P^wt/wt mice of each sex

Fig. 2

β-amyloid pathology is mitigated in female APP/PS1 mice carrying the PLCγ2-P522R variant. A X-34-positive β-amyloid plaque count (*p = 0.005) and coverage (**p = 0.006, total plaque area, % of whole analyzed area) were lower in the entorhinal cortex of APP/PS1xPLCγ2-P522R (A+/P^ki/ki) mice than in that of APP/PS1 (A+/P^wt/wt) mice. The size (µm²) of the individual plaques did not differ between the genotypes. A+/P^wt/wt n = 4 and A+/P^ki/ki n = 6. B Insoluble but not soluble Aβ40 and −42 levels are slightly but not significantly lower in the temporo-occipital cortex of A+/P^ki/ki mice than in that of A+/P^wt/wt mice. Aβ40 and Aβ42 levels were normalized to the total protein concentration in the same sample. A+/P^wt/wt n = 4 and A+/P^ki/ki = 4. C Representative Western blots and corresponding quantification showing no differences in the levels of full-length APP (APPtot, normalized to GAPDH), APP C-terminal fragments (C99 and C83, normalized to APPtot), or soluble APPα and APPβ (sAPPα, sAPPβ, normalized to APPtot) species in the temporo‒occipital cortex of A+/P^ki/ki and A+/P^wt/wt mice. A+/P^wt/wt n = 4 and A+/P^ki/ki n = 4. The scale bars in the representative immunofluorescence images are 657 µm for the whole area and 164 µm for the zoomed view. Unpaired samples t test. All the data are presented as the means ± SEMs. Each data point represents an individual mouse

Fig. 3

PLCγ2-P522R variant increases the area of microglia around β-amyloid plaques in female APP/PS1 mice. A The area (µm²) of IBA1-positive microglia was greater at distances of 20–40 µm (*p = 0.017) and 0–40 µm (*p = 0.03) from the plaque outline in the entorhinal cortex of APP/PS1xPLCγ2‒P522R (A+/P^ki/ki) mice than in that of APP/PS1 (A+/P^wt/wt) mice. Simultaneously, a trend toward an increase in the total IBA1 area is observed. A+/P^wt/wt n = 4 and A+/P^ki/ki n = 6. B GFAP-positive astrocyte area (µm²) around plaques (within 0–20 µm and 20–40 µm from the plaque outline) and total GFAP area remain unaltered between the genotypes. A+/P^wt/wt n = 5, A+/P^ki/ki n = 6. The colored disks on the x-axis of A and B indicate the analyzed area overlapping (middle circle) and surrounding β-amyloid plaques. The scale bar in the representative immunofluorescence images is 50 μm. Unpaired samples t test. All the data are presented as the means ± SEMs. Each data point represents an individual mouse

Fig. 4

PLCγ2-P522R variant decreases β-amyloid plaque-associated dystrophic neurites in female APP/PS1 mice. A Total area (µm²) of diffuse β-amyloid (6E10) and composition β-amyloid plaques, as indicated by the percentage of compact (X-34) and diffuse (6E10) β-amyloid plaques, remaining unaltered in the entorhinal cortex of the APP/PS1xPLCγ2-P522R (A+/P^ki/ki) mice compared with the APP/PS1 (A+/P^wt/wt) mice. A+/P^wt/wt n = 5 and A+/P^ki/ki n = 6. B) A 3D reconstruction showing the 6E10 and IBA1 (microglia) signals and their colocalization. Quantification of the 6E10 signal within the IBA1-positive area as a percentage of all the 6E10 signal in the entorhinal cortices of the A+/P^ki/ki mice compared with that in the A+/P^wt/wt mice. C The area (µm²) of 22C11-labeled dystrophic neurites around β-amyloid plaques strongly correlates with plaque size (r = 0.8113, ****p < 0.0001). The 22C11-positive area was decreased within the β-amyloid plaque area (p = 0.033), and a similar decreasing trend was observed within the 0–14 µm radius from the β-amyloid plaque outline in the entorhinal cortices of the A+/P^ki/ki mice compared with the A+/P^wt/wt mice when the plaque size was normalized. A+/P^wt/wt n = 4 and A+/P^ki/ki n = 6. Colored disks on the x-axis indicate the analyzed area overlapping (middle circle) and surrounding β-amyloid plaques. Scale bars in the representative immunofluorescence images are 50 μm. Pearson correlation and unpaired samples t tests were used. All the data are presented as the means ± SEMs. Each data point represents an individual mouse

Fig. 5

PLCγ2-P522R variant alleviates neuroinflammation in the brain tissue of female APP/PS1 mice. A IL-12p70 (p = 0.049) and IL-6 (p = 0.003) levels are significantly lower in the temporo-occipital cortex (CTX), whereas IL-1β (p = 0.021) and IL-8 (p = 0.026) levels are significantly lower in the hippocampus (HC) of APP/PS1xPLCγ2-P522R (A+/P^ki/ki) mice than in those of APP/PS1 (A+/P^wt/wt) mice. In addition, the IL-1β, IL-8, and TNF-α levels tended to decrease in the CTX group, whereas the IL-6, TNF-α, and TGF-β2 levels tended to decrease in the HCs of the A+/P^ki/ki mice compared with those of the A+/P^wt/wt mice. CTX A+/P^wt/wt n = 4 and A+/P^ki/ki n = 4; HC A+/P^wt/wt n = 5 and HC A+/P^ki/ki n = 6. Unpaired samples t test. All the data are presented as the means ± SEMs. Each data point represents an individual mouse

Fig. 6

PLCγ2-P522R variant does not increase the plaque-associated APOE or DAM signature in female APP/PS1 mice. A Analysis of total APOE and APOE areas (µm²) within and surrounding (within 0–10- and 10–20-µm distances from the amyloid plaque outline) β-amyloid plaques revealed no differences between the APP/PS1xPLCγ2-P522R (A+/P^ki/ki) and APP/PS1 (A+/P^wt/wt) mice. A+/P^wt/wt n = 5, A+/P^ki/ki n = 6. Colored disks on the x-axis indicate the analyzed area overlapping (middle circle) and surrounding β-amyloid plaques. The scale bar in the representative immunofluorescence images is 50 μm. Unpaired samples t test. All the data are presented as the means ± SEMs. Each data point represents an individual mouse. B Heatmap of z scored vst-normalized expression of homeostatic as well as stage 1 and stage 2 disease-associated microglia (DAM) signature genes in the bulk temporo-occipital cortex and C) CD11b+ microglia isolated from the whole brain tissue of 13-month-old wild-type (A-/P^wt/wt), A+/P^wt/wt and A+/P^ki/ki female mice. Cortex A-/P^wt/wt n = 1, A+/P^wt/wt n = 4, and A+/P^ki/ki n = 4; CD11b+ A-/P^wt/wt n = 7, A+/P^wt/wt n = 4, and A+/P.^ki/ki n = 4

Fig. 7

PLCγ2-P522R female mice exhibit avoidance responses to open places, bright environments, and sudden unpleasant stimuli. A Body weights of the PLCγ2-P522R (P^ki/ki) mice are lower than those of their wild-type (P^wt/wt) littermates on the APP/PS1-negative background (A-, *p = 0.024) but not on the APP/PS1 hemizygous background (A+). B PLCγ2-P522R does not influence the ambulatory distance traveled in the spontaneous activity test. C P^ki/ki mice spent significantly less time in the center of the novel test box than did control P^wt/wt mice, regardless of the APP/PS1 genotype. D In the passive avoidance test, the P^ki/ki mice showed increased avoidance of the illuminated chamber, which was further accentuated by the APP/PS1 hemizygous genotype. E Even after receiving a foot shock, P^ki/ki mice have a shorter latency to enter the dark chamber as compared to P^wt/wt mice. This tendency was further accentuated by the APP/PS1 hemizygous genotype. F No difference in motor skills was observed between the genotypes, as indicated by the time spent on the rotating rod. G P^ki/ki and P^wt/wt mice have similar thresholds for pain (foot shock), but H) P^ki/ki mice have a stronger behavioral response (crab-like walking backward) to electric foot shock. I) PLCγ2-P522R does not influence spatial learning or memory in the Morris water maze. The APP/PS1 hemizygous mice presented strongly impaired learning in terms of escape latency (F_1,64 = 46.8, p < 0.001). A-/P^wt/wt n = 20, A-/P^ki/ki n = 18, A+/P^wt/wt n = 14, and A+/P^ki/ki n = 11. Two-way ANOVA with Šídák's multiple comparisons test. All the data are presented as the means ± SEMs. Each data point represents an individual mouse

Fig. 8

Fatty acid metabolism and mitochondrial function-related targets are upregulated in male mouse PLCγ2-P522R microglia. A Dot plot of normalized enrichment scores for enriched and depleted gene sets in enrichment analyses by GSEA for gene (MSigDB Hallmark, FDR < 0.25) and protein (Wikipathways, FDR < 0.05) expression in CD11b+ microglia isolated from 13-month-old PLCγ2-P522R KI and wild-type (WT) mice. B Heatmap of z scored vst-normalized expression of up- and downregulated differentially expressed genes (DEGs) that were core enrichment genes in a GSEA of gene sets for calcium signaling (Calcium), cholesterol homeostasis (CHOL), fatty acid metabolism (FA), the IFN-α response (IFA), the inflammatory response (INFLAM), oxidative phosphorylation (OXPHOS), and Toll-like receptor signaling (Toll). C Heatmap of z scored vs. normalized expression of significantly up- and downregulated differentially expressed proteins (DEPs, expressed as their encoding gene symbols) derived from a proteomics study. RNA WT n = 3 and PLCγ2-P522R KI n = 4; protein WT n = 5 and PLCγ2-P522R KI n = 6

Fig. 9

PLCγ2-P522R variant changes lipid accumulation in mouse primary microglia. A The percentage of lipid droplet (LD)-positive cells was greater in mouse primary microglia after 24 h of lipopolysaccharide (LPS) and 48 h of myelin treatment than in untreated cells (UNT). The percentage of LD-positive cells did not differ between wild-type (WT) and PLCγ2-P522R knock-in (KI) microglia. PLCγ2-P522R KI decreases the size (µm²) of individual LDs under myelin-treated conditions (****p < 0.0001). n (analyzed images) WT UNT n = 11, LPS n = 22, myelin n = 42; PLCγ2-P522R KI UNT n = 6, LPS n = 12, myelin n = 35. Two-way ANOVA with Tukey’s multiple comparisons test. All the data are presented as the means ± SEMs

Fig. 10

PLCγ2-P522R variant improves mitochondrial function in human microglial models. Volcano plot of differentially expressed genes (DEGs) in blood monocyte-derived microglia-like cells (MDMi) from PLCγ2-P522R variant carriers (CG) and matched controls (CC) under A) untreated (UNT), B myelin-treated and C lipopolysaccharide (LPS)-treated conditions. FDR < 0.05. Horizontal dashed line: adjusted p value of 0.05; vertical dashed lines: |log2FC|= 0.1. D Dot plot of normalized enrichment scores for enriched and depleted gene sets in enrichment analyses by GSEA for gene (hallmarks, FDR < 0.25) expression in MDMi cells of CG and CC individuals upon UNT, LPS, and myelin treatments. E Heatmap of z scored vst-normalized expression of up- and downregulated DEGs in the LPS-treated condition that were core enrichment genes in a GSEA of gene sets for the IFN-γ response (IFG), inflammatory response (INFLAM), Kras signaling up (KRASup), MYC-target (MYC), oxidative phosphorylation (OXPHOS), and TNF-α signaling (TNFA). UNT CC n = 5 and CG n = 4; Myelin CC n = 5 and CG n = 4; LPS CC n = 5 and CG n = 4. F) The mitochondrial oxygen consumption rate (OCR, maximal respiration and spare respiratory capacity) is greater in homozygous PLCγ2-P522R (GG)-induced pluripotent stem cell-derived microglia (iMGL) than in isogenic controls (CCs) with the APOE3/3 genetic background but not the APOE4/4 genetic background. iMGL n = 1 line per group, 7–9 technical replicates per line. Two-way ANOVA with Tukey’s multiple comparisons test. The data are presented as the means ± SEMs