Disruption of hnRNP A2-mediated RNA dynamics by amyloid-β drives MBP increase in Alzheimer's disease

Abstract

Oligodendrocyte dysfunction, myelin degeneration, and white matter changes are critical events in the cognitive decline of Alzheimer's disease (AD). Amyloid-β peptide (Aβ), a hallmark of AD, disrupts oligodendrocyte and myelin homeostasis, through mechanisms that remain poorly understood. Here, transcriptomic profiling of Aβ-exposed oligodendrocytes revealed widespread gene expression changes, particularly in RNA-related processes. Among these, hnRNP A2, a key regulator of RNA transport and myelin protein regulation, was aberrantly upregulated in hippocampal oligodendrocytess from AD patients with high Aβ levels, from AD mouse models, and in Aβ-treated oligodendrocytes. RNA-immunoprecipitation sequencing of the hnRNP A2 interactome revealed Aβ-induced changes in mRNA interactions, particularly enriched binding to Mbp and Mobp, indicating impaired RNA metabolism of myelin components. Furthermore, Aβ, through hnRNP A2 disruption, increased the number, cargo and dynamics of Mbp- and Mobp-containing granules, enhanced MBP and MOBP synthesis, and decreased oligodendroglial voltage-gated Ca²⁺ influx in an MBP-dependent manner. These findings suggest that Aβ-induced dysregulation of hnRNP A2 impairs RNA metabolism and myelin protein synthesis, altering the intracellular Ca²⁺ homeostasis critical for oligodendrocyte function.

Keywords: Alzheimer´s disease; Amyloid-β peptide; Calcium homeostasis; HnRNP A2; MBP; Oligodendrocytes.

Fig. 1

Aβ-treated cells show alterations in RNA metabolism and translation. a Volcano plot of differentially expressed genes (DEGs). Analysed genes are scattered by their expression levels (axis X, log10 mean of normalised counts) and fold change (FC) (axis Y, log2FC). Differentially expressed genes (Padj < 0, 05) are coloured based on their FC (blue < 0; red > 0). b Gene ontology enrichment analysis of biological process (BP), cellular component (CC) and molecular function (MF) of DEGs and number of genes associated in each GO term. c Selected GO terms related to ribonucleoproteins, RNA processing and myelination. Differentially expressed genes (Padj < 0. 05) are coloured based on their FC (red < 0; blue > 0). d Venn diagram depicting the overlap of DEGs in vitro vs. human AD patients [28]. Percentage and numbers indicate the genes shared among the datasets. e Gene ontology enrichment analysis of the biological processes shared in Aβ and AD condition. f Selected GO terms related to the establishment of RNA localisation are shown. DEGs (Padj < 0. 05) are coloured based on their FC (red < 0; blue > 0)

Fig. 2

hnRNP A2 expression and localisation changes in oligodendroglial cells in AD brains, AD mouse models and primary oligodendrocyte culture treated with Aβ. a, b hnRNP A2 expression and relative quantification in human hippocampal lysates from control (n = 7) and AD patients (n = 6) (CERAD C and Thal 5) normalized to total protein content (Ponceau). c, d Representative confocal images show the Olig2 (magenta) and hnRNP A2 (green) expression in the hippocampus from controls (n = 3) and AD patients (CERAD C and Thal 5) (n = 4). Quantification of hnRNP A2 mean intensity signal in Olig2⁺ cells in the human hippocampus. Scale bars, 100 μm and 50 μm. e, f Representative confocal images of Olig2 (magenta) and hnRNP A2 (green) in the dentate gyrus of 6-month-old of WT (n = 7) and 3xTg-AD (n = 7) mice. Histogram depicts changes in the mean intensity of hnRNP A2 in oligodendroglia lineage. Scale bar, 100 μm. g, h HnRNP A2 expression and relative quantification in oligodendrocyte cell extracts normalised to β-actin (n = 7). i, j Representative confocal and binary micrographs of hnRNP A2 in control and Aβ-treated OLs. Histogram depicts changes in the area occupied by hnRNP A2 (n = 6). Data indicate means ± S.E.M and dots represent patients (b), individual cells and media from the same patient (d), individual mouse (f) or independent culture replicates (h, j). Statistical significance (*p < 0.05, **p < 0.01, ***p < 0.01, ****p < 0.001) was drawn by two-tailed ordinary one-way ANOVA followed by Dunnett’s post-hoc test (h), two-tailed unpaired (b, d, f) and paired Student t-test (j)

Fig. 3

RIP-seq analysis of hnRNP A2-associated RNAs performed in primary cultured OLs. a, b MA plot of hnRNP A2 RIP-seq data in control and Aβ-treated cells. For each transcript, the average signal (measured as log10 mean of normalised counts) against the RIP-seq log2 Fold Enrichment (RIP versus IgG) was plotted. Significantly enriched targets are highlighted in orange (Ctrl) and purple (Aβ). c Venn diagram depicting the overlap between the hnRNP A2 interactome of control (orange) and Aβ-treated (purple) OLs. The numbers indicate the genes shared among the datasets. d Classification of hnRNP A2-associated RNAs in control (orange) and Aβ-treated (purple) OLs. The majority (95%) of identified targets are protein-coding genes, but long noncoding RNAs (lncRNAs) and other ncRNAs were also present. e, f Functional annotation of enriched hnRNP A2 target genes in control (orange) and Aβ-treated cells (purple). The barplot displays the top 10 enriched biological processes. The length of each bar is proportional to the statistical significance of the enrichment. The number of genes associated in each category is displayed beside the corresponding bar. g Network representation of the biological processes GO categories is enriched in the hnRNP A2 interactome in control (orange) and Aβ (purple) OLs. The size of the node indicates the number of genes in each GO term

Fig. 4

Mbp mRNA granule content and activation is affected by Aβ. a, b Double colocalisation confocal and binary images of hnRNP A2 (green) and hnRNP F (magenta). 1 µM Aβ significantly increased hnRNP A2-F colocalisation (µm2) (n = 5). Scale bar, 10 μm. c, d Triple colocalisation confocal images of hnRNP A2 (green), hnRNP F (magenta) and hnRNP K (cyan). Graphs show that 70% of all granules contain hnRNPK (active granules) and Aβ-treated OLs contain 5% more active granules (n = 5). Scale bar, 10 μm. e Analysis of Mbp mRNAs levels by RT-qPCR (n = 4). f, g Representative confocal images showing hnRNP A2 (magenta) and Mbp (green) transcript and the colocalized image. Mbp transcripts were found in the cytoplasm and nucleus of OLs. Aβ -treated cells show a higher percentage of colocalisation in the cytoplasm (n = 5). Scale bar, 10 μm. h Quantification of Pearson´s correlation coefficient for hnRNP A2 and Mbp (n = 5). i, j Western blot of pTyr-IP and IgG to detect hnRNP A2 phosphorylation. Histogram depicting the phosphorylation of hnRNP A2 normalized to total hnRNP A2 from input (n = 4). Data indicate the means ± S.E.M and dots represent independent culture replicates. Statistical significance (*p < 0.05, **p < 0.01, ***p < 0.001) was drawn by two-tailed paired Student´s t-test (b, d, e, g, h, j)

Fig. 5

Regulation of Mbp synthesis by Aβ is dependent of hnRNP A2. a, b Photographs show MBP puro-PLA-positive puncta in the soma, primary and total processes. MBP PLA positive puncta was analysed in bins of 10 μm ranging from the soma in primary and total processes (n = 4). Scale bar, 10 μm. c, d MBP expression and relative quantification in oligodendrocyte cell extracts normalised to β-actin (n ≥ 4). e, f Representative confocal micrographs of GFP (green) and hnRNP A2 (magenta) in OLs transfected with Ctrl or Hnrnpa2-targeting siRNAs. Histogram depicting the hnRNP A2 integrated density (ID) within OLs (n = 4). Scale bar, 10 μm. g, h Representative confocal micrographs of MBP (green) and DAPI (blue) in untreated and treated OLs transfected with Ctrl or Hnrnpa2-targeting siRNAs. Scale bar, 10 μm. Histogram depicting the MBP integrated density (ID) within OLs (n = 4). Data are represented as means ± S.E.M and dots indicate independent culture replicates. Statistical significance (*p < 0.05, **p < 0.01, ***p < 0.001) was drawn by two-tailed paired Student´s t-test (b, f), one-way ANOVA followed by Dunnett’s post-hoc test (d) and two-way ANOVA followed by Sidak´s post-hoc test (h)

Fig. 6

MBP overexpression inhibits KCl-induced Ca²⁺ influx into OLs. a OLs infected with either AAV8-pMBP-GFP or AAV8-pMBP-MBP-IRES-GFP were loaded with x-Rhod-1 AM. Time course of intracellular Ca²⁺ levels were recorded before and after KCl 25 mM stimulus by confocal microscopy (n = 7). Scale bar, 10 μm. b, c Graphs show the maximum peak of KCl response. d OLs transfected with either control siRNA or Mbp-targeting siRNAs were loaded with Fluo-4 AM and exposed to 1 µM Aβ for 24 h. Time course of intracellular Ca²⁺ levels were recorded before and after KCl 25 mM stimulus by confocal microscopy (n = 6). Scale bar, 20 μm. e-g Graphs show the maximum peak and the area under the curve (AUC) of KCl response in the different conditions. Data indicate means ± S.E.M and dots represent individual cells. Statistical significance (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) was drawn by two-tailed unpaired Student´s t-test (c) and ordinary two-way ANOVA followed by Sidak´s post-hoc test (f, g)