Deletion of miR-33, a regulator of the ABCA1-APOE pathway, ameliorates neuropathological phenotypes in APP/PS1 mice

Abstract

Introduction: Rare variants in ABCA1 increase the risk of developing Alzheimer's disease (AD). ABCA1 facilitates the lipidation of apolipoprotein E (apoE). This study investigated whether microRNA-33 (miR-33)-mediated regulation of this ABCA1-APOE pathway affects phenotypes of an amyloid mouse model.

Methods: We generated mir-33^+/+;APP/PS1 and mir-33^-/-;APP/PS1 mice to determine changes in amyloid pathology using biochemical and histological analyses. We used RNA sequencing and mass spectrometry to identify the transcriptomic and proteomic changes between our genotypes. We also performed mechanistic experiments by determining the role of miR-33 in microglial migration and amyloid beta (Aβ) phagocytosis.

Results: Mir-33 deletion increases ABCA1 levels and reduces Aβ accumulation and glial activation. Multi-omics studies suggested miR-33 regulates the activation and migration of microglia. We confirm that the inhibition of miR-33 significantly increases microglial migration and Aβ phagocytosis.

Discussion: These results suggest that miR-33 might be a potential drug target by modulating ABCA1 level, apoE lipidation, Aβ level, and microglial function.

Highlights: Loss of microRNA-33 (miR-33) increased ABCA1 protein levels and the lipidation of apolipoprotein E. Loss of miR-33 reduced amyloid beta (Aβ) levels, plaque deposition, and gliosis. mRNAs and proteins dysregulated by miR-33 loss relate to microglia and Alzheimer's disease. Inhibition of miR-33 increased microglial migration and Aβ phagocytosis in vitro.

Keywords: ABCA1; Alzheimer's disease; amyloid; apolipoprotein E; lipid metabolism; microRNA‐33.

FIGURE 1

Deletion of mir‐33 increases ABCA1 protein levels and increases the lipidation of apoE in APP/PS1 mice. A, miR‐33 expression was determined by qPCR from RNA isolated from cortical tissue of mir‐33 ^+/+;APP/PS1 and mir‐33 ^−/−;APP/PS1 mice. Representative western blot (B) and quantification (C) of ABCA1 protein levels from hippocampal RIPA lysate (normalized to β‐actin protein levels). D, Representative western blots probed for apoE in its native condition run without detergents. The red lines indicate the HMW apoE species. E, Quantification of the HMW apoE bands relative to the total apoE signal for each sample. F, Quantification of the cholesterol associated with apoE that was immunoprecipitated from cortical PBS lysate. G, Quantification of apoE protein levels from hippocampal PBS lysate determined by ELISA. All values are mean ± SEM. ***p < 0.001 | *p < 0.05 | NS, not significant (unpaired two‐tailed t test; n = 5–6 for mir‐33 ^+/+;APP/PS1, n = 5 for mir‐33 ^−/−;APP/PS1). apoE, apolipoprotein E; ELISA, enzyme‐linked immunosorbent assay; HMW, high molecular weight; PBS, phosphate‐buffered saline; qPCR, quantitative polymerase chain reaction; RIPA, radioimmunoprecipitation assay; SEM, standard error of the mean.

FIGURE 2

Deletion of mir‐33 decreases insoluble Aβ peptides and Aβ plaque deposition in APP/PS1 mice. Aβ levels were measured using the MSD Aβ ELISA kit. Insoluble Aβ₄₀ and Aβ₄₂ levels were measured from cortical (A) and hippocampal (B) guanidine fractions. C, Representative images of brain sections immunostained with Aβ‐specific 82E1 antibody. D, Quantification of plaque load in cortical and hippocampal areas. E, Representative images of brain sections stained with X‐34 dye that detects fibrillar plaques. F, Quantification of X34+ fibrillar plaque load in cortical and hippocampal areas. All values are mean ± SEM. Scale bars equal to 500 µm. *p < 0.05 | **p < 0.01 | ***p < 0.001 (unpaired, two‐tailed t test; n = 6 for mir‐33 ^+/+;APP/PS1, n = 5 for mir‐33 ^−/−;APP/PS1). Aβ, amyloid beta; ELISA, enzyme‐linked immunosorbent assay; MSD, Meso Scale Discovery; SEM, standard error of the mean.

FIGURE 3

Deletion of mir‐33 ameliorates gliosis in APP/PS1 mice. A, Representative images of brain sections immunostained with anti‐CD45 antibody. B, Quantification of CD45+ cell load in cortical and hippocampal areas. C, Representative image of brain sections immunostained with anti‐IBA1 antibody. D, Quantification of IBA1+ cell load in cortical and hippocampal areas. E, Representative image of brain sections immunostained with anti‐GFAP antibody. F, Quantification of cortical GFAP+ cell load. All values are mean ± SEM. Scale bars equal to 500 µm. *p < 0.05 | **p < 0.01 (unpaired two‐tailed t test; n = 6 for mir‐33 ^+/+;APP/PS1, n = 5 for mir‐33 ^−/−;APP/PS1). GFAP, glial fibrillary acidic protein; IBA1, ionized calcium binding adaptor molecule 1; SEM, standard error of the mean.

FIGURE 4

Multi‐omics analyses suggest that miR‐33 may regulate microglial migration and immune response. A, Volcano plot visualizing the 882 DEGs identified by bulk RNA‐sequencing (RNAseq) from cortical tissue between mir‐33 ^+/+;APP/PS1 (WT) and mir‐33 ^−/−;APP/PS1 (KO) mice. B, Volcano plot visualizing the 125 DAPs identified by mass spectrometry (MS/MS) from cortical lysate between mir‐33 ^+/+;APP/PS1 (WT) and mir‐33 ^−/−;APP/PS1 (KO) mice. C, Pathway Map enrichment analysis was performed with the DEGs (red) and DAPs (blue) identified between mir‐33 ^+/+;APP/PS1 and mir‐33 ^−/−;APP/PS1 mice using the MetaCore software. The top 10 Pathway Maps are shown with the vertical red line denoting significant Pathway Maps identified. D, The shared response are the common DEGs and DAPs between RNAseq and MS/MS visualized as a function of their t statistic. E, GO enrichment analysis was performed with the shared response common DEGs and DAPs using the MetaCore software. The top 10 GO Processes are shown with the vertical red line denoting significant GO Processes identified. DAPs, differentially abundant proteins; DEGs, differentially expressed genes; GO, Gene Ontology; KO, knock out; WT, wild type.

FIGURE 5

Inhibition of miR‐33 increases microglial migration and Aβ phagocytosis. A, Scratch‐wound assay was performed with BV2 cells to compare between the control (Ctrl) group and the miR‐33 inhibitor group. Twenty‐four hours after transfection, a wound was created and imaged at 0 hours, and 24 hours. The overlay demonstrates the remaining wound area (black) and the area migrated by the BV2 cells (blue). B, Quantification of the wound coverage area by BV2 cells at time 24 hours. C, Aβ phagocytosis assay performed with transfected BV2 cells comparing the Ctrl versus the miR‐33 inhibitor group. Twenty‐four hours after transfection, cells were treated with Aβ aggregates tagged with pHrodo and the 560/585 nm fluorescent signal was measured after 8 hours. D, Quantification of the relative change in pHrodo fluorescent signal compared to Ctrl. All values are mean ± SEM. Scale bars equal to 200 µm (A) and 10 µm (C). *p < 0.05 | **p < 0.01 (unpaired two‐tailed t test; n = 5 for scratch‐wound assay, n = 6 for Aβ uptake assay). Aβ, amyloid beta; SEM, standard error of the mean.