Microglia specific deletion of miR-155 in Alzheimer's disease mouse models reduces amyloid-β pathology but causes hyperexcitability and seizures

Abstract

Alzheimer's Disease (AD) is characterized by the accumulation of extracellular amyloid-β (Aβ) as well as CNS and systemic inflammation. Microglia, the myeloid cells resident in the CNS, use microRNAs to rapidly respond to inflammatory signals. MicroRNAs (miRNAs) modulate inflammatory responses in microglia, and miRNA profiles are altered in Alzheimer's disease (AD) patients. Expression of the pro-inflammatory miRNA, miR-155, is increased in the AD brain. However, the role of miR-155 in AD pathogenesis is not well-understood. We hypothesized that miR-155 participates in AD pathophysiology by regulating microglia internalization and degradation of Aβ. We used CX3CR1^CreER/+ to drive-inducible, microglia-specific deletion of floxed miR-155 alleles in two AD mouse models. Microglia-specific inducible deletion of miR-155 in microglia increased anti-inflammatory gene expression while reducing insoluble Aβ_1-42 and plaque area. Yet, microglia-specific miR-155 deletion led to early-onset hyperexcitability, recurring spontaneous seizures, and seizure-related mortality. The mechanism behind hyperexcitability involved microglia-mediated synaptic pruning as miR-155 deletion altered microglia internalization of synaptic material. These data identify miR-155 as a novel modulator of microglia Aβ internalization and synaptic pruning, influencing synaptic homeostasis in the setting of AD pathology.

Keywords: Alzheimer’s disease; Epilepsy; Inducible knock-out; Microglia; Mouse models; miR-155.

Fig. 1

Expression of miR-155 is not detected at 12 months of age in ex vivo microglia from the APP/PS1 mouse model of AD after Cre-recombinase induction. A Experimental and control genotypes and groups. APPswe/PS1dE9 (APP/PS1) mice were crossed with miR-155flx/flx; CX3CR1^CreER/+ mice to generate APP/PS1;miR-155flx/flx; CX3CR1^CreER/+ (with tamoxifen: APP/PS1 MG miR-155 CKO mice; with corn oil: APP/PS1 MG miR-155 WT) or non-APP littermate controls that allow for conditional miR-155 deletion (Microglia miR-155 CKO). B) Experimental timeline of study. C Gating strategy for microglia isolation (CX3CR1-YFP⁺/ CD45^low cells) using ex vivo-FACS from the adult mouse CNS. miR-155 copy number in microglia from APP/PS1 MG miR-155 CKO mice and APP/PS1 MG miR-155 WT was quantified by qPCR at D) 6 months of age E) and 12 months of age (Stats: One-Way ANOVA with Tukey’s post hoc correction for multiple comparisons (*** = p < 0.0005)

Fig. 2

Anti-inflammatory gene expression is upregulated in microglia after microglia-specific knock-out of miR-155 in vivo. qPCR analysis of total RNA extracted form ex vivo-FACS sorted microglia (CX3CR1^YFP+/CD45^int cells). In the APP/PS1 model A) cMaf B) Csf1r C) Inpp5d (SHIP1) D) Socs1 and E) Tfeb are upregulated in microglia in vivo after miR-155 CKO. We observed that Csf1r was not significantly upregulated in microglia after miR-155 CKO in the APP/PS1 mouse model of AD (Stats: two-way ANOVA with multiple comparisons, Tukey’s post hoc correction (** = p < 0.005, *** = p < 0.0005)

Fig. 3

miR-155 deletion from microglia in adult APP/PS1 mice leads to a reduction in insoluble Aβ_1–42 and total plaque area. A Analysis of the insoluble protein fractions by Luminex detected decreased levels of Aβ_1–42 in APP/PS1 microglia miR-155 conditional knock-out mice (APP/PS1 MG miR-155 CKO) compared to non-deleted controls (APP/PS1) (two-tailed unpaired t-test, * = p < 0.05), B while there was no detectable change in Aβ_1–40 levels. C and D Representative images of immunostaining for Aβ-plaques (6E10) in APP/PS1 and APP/PS1 miR-155 MG CKO brains. E Aβ-plaque area quantification based on 6E10 immunostaining in CA1 and CA3 (two-tailed unpaired t-test, * = p < 0.05)

Fig. 4

Microglia-specific inducible knock-out of miR-155 leads to increased mortality in the APP/PS1 mouse model of AD. A Survival analysis of experimental and control lines and littermate controls revealed significant mortality in APP/PS1 mice after microglia-specific deletion of miR-155 (Stats: Log-Rank test with pairwise comparisons and Bonferroni’s post hoc correction). B There was no sex-specific difference in survival observed within genotypes (Stats: Log-Rank test based on sex, p = 0.5)

Fig. 5

Microglia-specific inducible knock-out of miR-155 leads to increased seizure burden in the APP/PS1 mouse model of AD. A Experimental design summary. Mice were implanted with ECoG and EMG electrodes at 7 weeks of age and, after a baseline recording, miR-155 knock-out was induced at 8 weeks. Continuous ECoG recordings were done for 2–5 weeks or until spontaneous death. Representative trace of B baseline and C seizure that resulted in a spontaneous death event. D Spontaneous seizures were identified and manually quantified starting 1-, 2- and 3-week post-miR-155 deletion in microglia. (Stats: Mixed-effects ANOVA with Tukey’s post hoc correction: ****p < 0.0001, ***p < 0.001, *p < 0.01). E Seizure frequency was increased post miR-155 deletion in microglia in the APP/PS1 background (Stats: Wilcoxon rank sum test with continuity correction, ***p < 0.00001). F % epochs (hours) containing inter-ictal spikes, high-amplitude, synchronous spiking observable via cortical ECoG were increased after miR-155 deletion in microglia (Stats: Mixed Linear Model, p < 0.0001, R² = 0.7311)

Fig. 6

Microglia-specific inducible knock-out of miR-155 in leads to aberrant cortical excitability in the 5xFAD mouse model of AD. A Experimental and control genotypes and groups. B Experimental design summary. Mice were implanted with ECoG and EMG electrodes at 7 weeks of age, and after a baseline recording, miR-155 knock-out was induced at 8 weeks. Continuous recordings were made for 2 weeks. Representative trace of C baseline and D high-amplitude, synchronous spiking in the EEG. E %epochs (hours) containing inter-ictal spikes, high-amplitude, synchronous spiking in the EEG were increased after miR-155 deletion in microglia in 5xFAD mice (Mixed Linear Model, p = 0.027, R² = 0.1312)

Fig. 7

Inducible knock-out of miR-155 in microglia results in increased internalization of excitatory synaptic terminals marked by VGLUT1 in the 5xFAD brain. A Microglia were isolated using ex vivo, then fixed and stained for the pre-synaptic vesicular GABA transporter (VGAT; representative histograms B in and quantification in C), and the pre-synaptic vesicular glutamate transporter 1 (VGLUT1; representative histograms D in and quantification in E). Total levels of internalized VGAT and VGLUT1 were detected with flow cytometry (LSRII). Data were analyzed using FlowJo and R studio, graphs were generated using Prism 9. (Stats: two-way Ordinary ANOVA with multiple comparisons with Tukey’s post hoc corrections, ****p < 0.0001, ***p < 0.001, **p < 0.01)