Lowering Hippocampal miR-29a Expression Slows Cognitive Decline and Reduces Beta-Amyloid Deposition in 5×FAD Mice

Abstract

microRNA-29a (miR-29a) increases with age in humans and mice, and, in the brain, it has a role in neuronal maturation and response to inflammation. We previously found higher miR-29a levels in the human brain to be associated with faster antemortem cognitive decline, suggesting that lowering miR-29a levels could ameliorate memory impairment in the 5×FAD AD mouse model. To test this, we generated an adeno-associated virus (AAV) expressing GFP and a miR-29a "sponge" or empty vector. We found that the AAV expressing miR-29a sponge functionally reduced miR-29a levels and improved measures of memory in the Morris water maze and fear condition paradigms when delivered to the hippocampi of 5×FAD and WT mice. miR-29a sponge significantly reduced hippocampal beta-amyloid deposition in 5×FAD mice and lowered astrocyte and microglia activation in both 5×FAD and WT mice. Using transcriptomic and proteomic sequencing, we identified Plxna1 and Wdfy1 as putative effectors at the transcript and protein level in WT and 5×FAD mice, respectively. These data indicate that lower miR-29a levels mitigate cognitive decline, making miR-29a and its target genes worth further evaluation as targets to mitigate Alzheimer's disease (AD).

Keywords: Beta-amyloid; Cognition; Neuroinflammation; Wdfy1; miR-29a.

Fig. 1

miR-29a expression in mouse hippocampus and miR-29a sponge validation in vitro. A miR-29a was determined in the hippocampus of 2- and 12-month-old 5×FAD and WT mice by real-time RT-PCR. U6 snRNA was used as the endogenous control. Data was analyzed via two-way ANOVA (age by genotype) with Sidak's post hoc tests. Bars indicate mean ± standard deviation (SD) with n = 3 per group; NS not significant; *p < 0.05, **p < 0.01. B Relative expression level of mature miR-29a in HEK-293T cells transfected with miR-29a mimic (pAAV-miR-29a) and control vector (pAAV-MCS); ***p < 0.001. C Schematic representation of miR-29a sponge mechanism and design. Sponge sequence was shown as black color and the target miRNA was shown as red. D Renilla luciferase activity was assayed relative to firefly luciferase activity in 293T cells transfected with psiCHECK2 vector containing miR-29a sponge and miR-29a mimic (pAAV-miR-29a) or control vector (pAAV-MCS); **p < 0.01. E Cell morphology and GFP expression upon transfection of miR-29a mimic (pAAV-miR-29a) with pAAV-GFP vector containing miR-29a sponge or empty vector (pAAV-GFP), as observed by light and fluorescent microscopy. Scale bars represent 100 μm. F Western blot analysis of DNMT3A upon transfection of miR-29a mimic with or without miR-29a sponge. Experiments were performed 3 times and one representative western blot is shown. The quantification of western blots is provided. Data represents fold change relative to no treatment group ± SD. GAPDH was used as a loading control. *p < 0.05, **p < 0.01

Fig. 2

Results of Morris water maze and fear conditioning. A Timeline of experiments. B Latency to find the platform. C Water maze probe trial. D Swim speed during 5 days of training trial. E Fear training with 3 tone-shock pairings. F–G Percentage of time spent in freezing following re-exposure to the shock-associated context or tone. Data are mean ± SEM with n = 9–12 per group. Data were analyzed via two-way repeated measures ANOVA (time × group) or three-way repeated measures ANOVA (time × group × genotype). The water maze probe trial was analyzed via two-way ANOVA (genotype by group). *p < 0.05, **p < 0.01 based on post hoc pairwise comparisons

Fig. 3

miR-29a sponge attenuates amyloid deposition and decreases activated astrocytes and microglia in the hippocampus of 5×FAD mice and WT mice. A–C Representative immunofluorescence images. D–F Quantification of Aβ, GFAP, and Iba1 immunoreactivity. Data were analyzed via two-way ANOVA (genotype by group) with Sidak’s post hoc tests. Bars indicate mean ± SD. N = 3 per group. Scale bars represent 100 μm; *p < 0.05, **p < 0.01, compared with the control group

Fig. 4

Downstream effectors of miR-29a at the transcript and protein level in 5×FAD mice. A Volcano plot displaying the distribution of differentially expressed genes between FAD-Sponge and FAD-Control in the RNA-Seq analysis. Top10 DEGs were labeled. B Heatmap of DEGs between FAD-Sponge and FAD-Control. Downregulated DEGs in FAD-Sponge vs FAD-Control were depicted in blue and upregulated DEGs were in orange. C GO enrichment analysis of downregulated DEGs in FAD-Sponge. D Expression of DEGs that are predicted targets of miR-29a. E Volcano plot displaying the distribution of differentially expressed proteins between FAD-Sponge and FAD-Control. F Expression of the differentially expressed protein between FAD-Sponge and FAD-Control

Fig. 5

Downstream effectors of miR-29a at the transcript level in WT mice. A Volcano plot displaying the distribution of differentially expressed genes between WT-Sponge and WT-Control. Top 10 DEGs were labeled. B Heatmap of DEGs between WT-Sponge and WT-Control. Downregulated DEGs in WT-Sponge vs WT-Control were depicted in blue and upregulated DEGs were in orange. C GO enrichment analysis of upregulated DEGs in WT-Sponge. D Expression of DEG that is the predicted target of miR-29a