NRF2 deficit prevents pathologic Tau seeding and spreading in an induced tauopathy mouse model

Abstract

Background: Nuclear factor erythroid 2-related factor 2 (NRF2) regulates antioxidant defenses and protects against neurodegeneration, including Alzheimer's disease (AD). Its age-related decline disrupts redox balance and increases neuronal vulnerability, but the early hippocampal effects remain unclear. Here, we tested whether NRF2 loss affects tau seeding and spreading in a PHF-tau-inoculated mouse model, contributing to accelerated aging.

Methodology: Three-month-old NRF2-knockout (Nfe2l2^-/-) and wild-type (WT) mice received hippocampal inoculations of human AD-derived PHF-tau, and tau propagation was analyzed after three months. To elucidate the molecular underpinnings of the observed changes, we performed integrative phosphoproteotranscriptomic analyses of hippocampal tissue, supported by RT-qPCR and Western blot validation.

Results: PHF-tau inoculation at 3 months of age in Nfe2l2^-/- mice, surprisingly, exhibited markedly reduced tau seeding and spreading compared to WT after 3 months of incubation. Molecular characterization of the Nfe2l2^-/- hippocampus was carried out to unravel the molecular changes associated with impaired tau propagation. Transcriptomic profiling revealed 745 deregulated genes in Nfe2l2^-/- mice, characterized by upregulation of immune and metabolic pathways but downregulation of oxidative stress and redox-related genes. RT-qPCR confirmed diminished expression of antioxidant enzymes and anti-inflammatory receptors, alongside altered astrocytic markers. Proteomic analysis identified 157 dysregulated proteins associated with mitochondrial, synaptic, and inflammatory processes, while phosphoproteomics detected 824 altered phosphosites enriched in cytoskeletal and synaptic networks. Western blot showed increased GFAP-C-term, AQP4, 8-OHdG, and MDAL, with reduced GSTM2 expression. Notably, total and 4R-tau levels were decreased, while 3R-tau was elevated in Nfe2l2^-/- mice.

Conclusion: Our findings suggest that NRF2 loss induces a hippocampal state marked by impaired antioxidant defenses, astrocytic remodeling, and disrupted tau isoform balance. This environment, while metabolically altered, paradoxically hinders tau propagation, highlighting NRF2 as a key regulator of both redox and cellular maturity programs essential for tau spread and as a potential therapeutic target in tauopathies.

Keywords: Aging; Alzheimer's disease; NRF2; Oxidative stress; Tau.

Fig. 1

Hyperphosphorylated tau-containing cells and threads following unilateral intrahippocampal injection of human p-tau into WT and Nfe2l2^−/− mice aged 3 months old and sacrificed at 6 months of age. Characterization of the sarkosyl-insoluble AD fraction inoculated into both animal models is shown in the experimental scheme by Western blotting of the phospho–Tau (Ser422) band pattern (scheme). The total p-Tau seeding and spreading across the dentate gyrus (DG) and hilum of the hippocampus (A-D) and in the corpus callosum (CC) and corpus callosum radiation (CCR) is higher in WT mice inoculated with AD homogenates than in AD-inoculated Nfe2l2^−/− mice (E-H). AT8 antibody decorates the nuclei of some neurons (arrows) and glial cells in form of coiled coils or “comas” in Nfe2l2^−/− and WT mice (C-D). Semiquantification, performed using a minimum of six slides/Z-levels per animal in hippocampal regions and adjacent areas where spreading is reported (CC, CCR, fimbria), demonstrated a significant reduction in the number of AT8-positive cytoplasmic deposits, as well as AT8-immunoreactive threads and dots, in the Nfe2l2^−/− mice (n = 11) compared with WT mice (n = 7) (I). Significance was set at ∗∗∗p < 0.001. Scale = 100 μm for A, B, E and F, and scale = 50 μm for C, D, G and H.

Fig. 2

Phospho-tau localization in WT and Nfe2l2^−/− mice after AD inoculation is not located in neurons, microglia or astrocytes in the corpus callosum. Double-label immunofluorescence and confocal microscopy for cell-type markers (GFAP, IBA1, NeuN) (green) and AT8 (red) in WT and Nfe2l2^−/− mice inoculated with sarkosyl-insoluble AD fractions into the hippocampus at 3 months of age and analyzed at 6 months (3-month survival). Tau deposits were observed along the corpus callosum (CC) in both genotypes, with no phospho-tau detected in astrocytes (GFAP) (a–h), microglia (IBA1) (i–p) or neurons (NeuN) (q–x). However, some nuclei or perineuronal spaces in the CA1 region of WT animals showed positivity, consistent with areas where tau pathology had spread (indicated with white arrows) (q–x). Nuclei were stained with DAPI (blue). Scale bar = 40 μm.

Fig. 3

Phospho-tau localization in WT and Nfe2l2^−/− mice after AD inoculation is restricted to oligodendrocyte and fibers in the corpus callosum and neurons in the hippocampus. Double-label immunofluorescence and confocal microscopy for axonal projections and oligodendrocyte markers (PLP1 and OLIG2) (green) and AT8 (red) in WT and Nfe2l2^−/− mice inoculated with sarkosyl-insoluble AD fractions into the hippocampus at 3 months of age and analyzed at 6 months (3-month survival). Tau deposits were observed along the corpus callosum (CC) in both genotypes, with phospho-tau detected in nervous fibers (a–h) and oligodendrocytes (i–t). In the CC, phospho-tau deposits were predominantly perinuclear in oligodendrocytes or distributed along cellular processes, frequently forming cap-like or comma-shaped structures reminiscent of coiled bodies observed in human tauopathies (indicated with white arrows). In the hippocampus, aggregates were also present (AT8, red), mainly in WT animals and not in Nfe2l2^−/− due its less seeding and spreading capacity, being located in neurons (NeuN, green) in the CA1 (indicated with white arrows) (u–x). Nuclei were stained with DAPI (blue). Scale bar = 40 or 20 μm.

Fig. 4

Transcriptomic profiling, functional enrichment, and network analysis of hippocampal gene expression in Nfe2l2^−/− and WT mice. (A) Heat map representing deregulated transcripts at a probability of an adjusted p-value <0.05. White boxes correspond to WT mice, and black boxes correspond to Nfe2l2^−/− mice; (B–C) Functional analysis of all DEG against the GO biological processes (BP) and Molecular Function (MF) databases (p < 0.05). (D) Functional analysis of all significant up-regulated genes against the Gene Ontology biological processes (BP) database (p < 0.05). (E) Functional analysis of all significant down-regulated genes against the Wikipathways database (p < 0.05). For all functional graphs (B–E), the blue-green color spectrum beneath each gene category is proportional to the enrichment FDR score number obtained from the differentially expressed proteins involved in the functional group. Dot size reflects the number of altered genes belonging to the category. (F) DEG clusteritzation according to cell types signatures of adult mouse hippocampus (HC) dataset (Saunders et al.). (G) Weighted GENE coexpression network analysis (WGCNA) of transcriptome identifies 48 gene modules. Modules are labeled by color and number (M1 to M48). Proteins not assigned to any particular coexpression modules were labeled M0 or gray. Dendrogram obtained by hierarchical clustering of proteins based on their topological overlap is shown at the top. (H) Modul-trait relationships between groups. Mann–Whitney-Wilcoxon test analysis revealed distinct significant changes in module eigengene expression among groups in 4 different modules. (I–L) Identification of gene enriched GO terms in each significant module was performed for modules characterization.

Fig. 5

qPCR validation of hippocampal transcripts associated with altered pathways in Nfe2l2^−/− and WT mice. Selected transcripts, identified based on potentially altered pathways and RNA-seq results, were analyzed by qPCR in the hippocampus of 3-month-old Nfe2l2^−/− (n = 6) and WT (n = 6) mice. Transcripts are grouped according to their associated pathways: (A) NRF2–Keap1 pathway-related genes; (B) antioxidant and oxidative stress response genes; (C) autophagy activity and vesicle marker genes; (D) glial neuroinflammatory genes; (E, F) inflammatory signaling and cytokine genes (sets A and B); (G, H) energetic supply and metabolic genes (sets A and B). Statistical significance between groups was defined as ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001.

Fig. 6

Proteomic profiling and functional enrichment analysis of hippocampal proteins in Nfe2l2^−/− and WT mice. (A) PCA using three principal components clearly separated the two groups (Nfe2l2^−/− and WT). (B) Volcano plot showing 67 downregulated and 90 upregulated proteins when compare WT against Nfe2l2^−/− animals. (C) Proteome heatmap of the hippocampus displaying 157 differentially expressed proteins. In both plots, red indicates upregulated expression and green indicates downregulated expression; color intensity reflects the magnitude of change. (D–I) Functional analysis of key deregulated proteins in the hippocampus of Nfe2l2^−/− mice based on GO enrichment. The blue–green color scale beneath each category is proportional to the enrichment FDR score obtained from the differentially expressed proteins in that functional group. Dot size reflects the number of altered proteins within each category. (D) All deregulated proteins based on GO cellular components. (E) All deregulated proteins based on GO biological processes. (F) Downregulated and (H) upregulated proteins associated to GO cellular components terms. (G) Downregulated and (I) upregulated proteins related to GO biological processes terms.

Fig. 7

Phosphoproteomic profiling and functional enrichment analysis of hippocampal phosphosites in Nfe2l2^−/− and WT mice. (A) PCA using two principal components partially separated the two groups (Nfe2l2^−/− and WT). (B) Volcano plot showing 401 downregulated and 423 upregulated phosphosites when comparing WT and Nfe2l2^−/− animals. (C) Phosphoproteome heatmap of the hippocampus displaying 824 differentially expressed phosphosites. In both plots, red indicates upregulated expression and green indicates downregulated expression; color intensity reflects the magnitude of change. (D, E) Functional analysis of key deregulated proteins in the hippocampus of Nfe2l2^−/− mice based on GO enrichment. The blue–green color scale beneath each category is proportional to the enrichment FDR score obtained from the differentially expressed proteins in that functional group. Dot size reflects the number of altered proteins within each category. (D) All deregulated proteins associated with altered phosphosites based on GO cellular components. (E) All deregulated proteins associated with altered phosphosites based on GO molecular function terms.

Fig. 8

Pathway-specific Western blot profiling of oxidative stress, inflammatory, and tau-related proteins in the hippocampus of Nfe2l2^−/− versus WT mice. Graphical representation of protein levels and corresponding Western blot images. Proteins are grouped according to their associated pathways: (A) oxidative stress biomarkers and related enzymes; (B) astrocytic and inflammatory markers; and (C) tau isoforms and kinases. Statistical significance was defined as ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001.

figs1

AT8 immunostaining specificity control. Representative AT8 background control staining in the dentate gyrus (DG) of the hippocampus from WT and Nfe2l2⁻/⁻ mice. Sections were processed without the primary antibody against the AT8 epitope (A, C) or with the primary antibody (B, D). In parallel, immunohistochemistry using the AT8 primary antibody was performed on human AD brain sections, which showed the expected pattern of pathological tau immunoreactivity (F) and a lack of staining when the primary antibody was omitted (E). These controls indicate that the AT8 signal observed in our study is not attributable to nonspecific secondary antibody binding or methodological artifacts. However, it should be noted that AT8 can exhibit background nuclear reactivity in mouse tissue, which is not uncommon. Scale bar = 100 or 50 μm.

figs2

Representative low-magnification AT8 immunostaining images showing the entire hippocampal structure and the injection site. Injection controls include: (A) vehicle-injected Nfe2l2⁻/⁻ mice; (B) Nfe2l2⁻/⁻ mice injected with PHFs derived from control cases; (C) Nfe2l2⁻/⁻ mice injected with PHFs derived from Alzheimer’s disease (AD) cases; and (D) WT mice injected with PHFs derived from AD cases. WT mice injected with vehicle or PHFs derived from control cases have been described previously and consistently show that neither vehicle injection nor control-derived PHFs induce tauopathy in these animals. Scale bar = 1000 μm.