Astrocytes, via RTP801, contribute to cognitive decline by disrupting GABAergic-regulated connectivity and driving neuroinflammation in an Alzheimer's disease mouse model

Abstract

Introduction: Alzheimer's disease (AD) pathogenesis involves astrocytic responses to extracellular amyloid beta deposits and phospho-tau neurofibrillary tangles, which drive inflammatory activation. RTP801, a stress-responsive protein, has been implicated in mediating neuroinflammation. Its levels are increased in AD hippocampal samples, correlating with disease severity and cognitive decline.

Methods: Using astrocyte-specific RTP801 silencing in the hippocampus of 5xFAD mice, we evaluated cognition, neuroinflammation, and hippocampal connectivity by magnetic resonance spectroscopy (MRS) and resting-state functional connectivity analyses. Histological and biochemical analyses assessed microgliosis, astrogliosis, and inflammasome-related protein levels.

Results: Astrocytic RTP801 silencing in 5xFAD mice preserved spatial memory, maintained hippocampal γ-aminobutyric acid (GABA) levels, and preserved resting-state brain networks. In addition, RTP801 silencing significantly reduced markers of microgliosis, astrogliosis, and inflammasome effectors.

Discussion: Astrocytic RTP801 contributes to AD-associated cognitive decline by disrupting GABAergic-regulated connectivity and amplifying inflammatory responses. Targeting astrocytic RTP801 may therefore offer therapeutic potential to mitigate AD progression by preserving neural connectivity and reducing neuroinflammation.

Highlights: The 5xFAD mouse model of Alzheimer's disease presents higher levels of RTP801 in hippocampal astrocytes. Normalizing the levels of astrocytic RTP801 prevents cognitive decline and restores anxiety-like behavior in the 5xFAD mouse model. Knocking down astrocytic RTP801 preserves the resting-state functional connectivity in the 5xFAD mouse model. Astrocytic RTP801 mediates the loss of Parvalbumin+ interneurons, negatively affecting the levels of γ-aminobutyric acid (GABA) in the 5xFAD mouse model. Astrocytic RTP801 contributes to astro- and microgliosis and inflammasome expression in the 5xFAD mouse model.

Keywords: Alzheimer's disease; GABA; REDD1; RTP801; astrocyte; cognitive deficits; connectivity; inflammasome; mTOR; neuroinflammation.

FIGURE 1

Higher levels of astrocytic RTP801 in 5xFAD mice. (A) Schematic representation of the experimental procedure. AAVs expressing GFP‐miCT (AAV‐miCT) or GFP‐miRNA‐RTP801 (AAV‐miRTP801) were bilaterally injected in the dorsal hippocampus of 6‐month‐old WT male mice (WT miCT or WT miRTP801 groups) or 5xFAD male mice (5xFAD miCT or 5xFAD miRTP801 groups). Four weeks later, behavioral tests were performed, followed by MRI and biochemical experiments. (B) Representative immunostaining of GFP fluoresceence, showing the specific transduction with AVV2/5 in the CA1 and DG in the four groups. Scale bar = 500 µm. (C) Representative CA1 images from the dorsal hippocampus where transduced cells (GFP⁺, green), astrocytes (GFAP ⁺, white), and RTP801 (red) are depicted. Scale bar = 50 µm. (D) Immunoreactivity quantification of the RTP801 levels in GFP⁺ astrocytes of the CA1 in the four groups is shown. (E) Representative DG images from the dorsal hippocampus where transduced cells (GFP ⁺, green), astrocytes (GFAP ⁺, white), and RTP801 (red) are depicted. Scale bar = 50 µm. (F) Immunoreactivity quantification of the RTP801 levels in GFP⁺ astrocytes of the DG in the four groups is shown. All data are shown as the mean ± SEM. All data were analyzed by two‐way ANOVA followed by Bonferroni's post hoc test: *p < 0.05, ***p < 0.001, and ^# p < 0.05. CA1, cornu ammonis‐1; DG, dentate gyrus. Each point represents data from an individual mouse (WT miCT n = 10, WT miRTP801 n = 10, 5xFAD miCT n = 8, 5xFAD miRTP801 n = 11).

FIGURE 2

Astrocytic RTP801 regulates anxiety‐like behavior and affects memory. A battery of behavioral tests was performed in the four groups: WT miCT, WT miRTP801, 5xFAD miCT, and 5xFAD miRTP801. (A) Elevated plus maze to measure anxiety‐like behavior. The distance run in the open arms was monitored for 5 min. Two‐way ANOVA and Bonferroni post hoc test was performed. ^### p < 0.001 and *p < 0.05, and ***p < 0.001. (B–D) Morris water maze. (B) The distance (in cm) to reach the visible platform was monitored in a four‐trial session to evaluate potential visual or physical impairments in 7‐month‐old mice. Two‐way ANOVA showed a significant general time effect in this procedural version of the MWM (p < 0.0001), indicating that animals significantly improved their latencies in Trial 4 compared to Trial 1. **p < 0.01 (WT miCT), ^## p < 0.01 (WT miRTP801), ⁺ p < 0.1 (5xFAD miCT), and ^$ p < 0.05 (5xFAD miRTP801). Representative pool tours during probe trials are depicted. (C) Morris water maze time curve: Time (in s) to reach the hidden platform was monitored for 6 days to evaluate spatial learning. Two‐way ANOVA showed a significant general time effect in this spatial version of the test and post hoc (Bonferroni's test) multiple comparisons indicated that WT miCT groups and 5xFAD miRTP801 significantly improved their latencies to reach the hidden platform on the day of training 6 compared to the day of training 1 (**p < 0.01 and ^$ p < 0.05, respectively), the 5xFAD miCT group showed no significant differences. (D) Representative pool tours during the Morris water maze test on Day 1 and Day 6 are depicted for the different groups. The number of mice in the behavioral test: WT miCT (n = 11), WT miRTP801 (n = 11), 5xFAD miCT (n = 10), and 5xFAD miRTP801 (n = 11). Data are means ± SEM. MWM (Morris Water Maze)

FIGURE 3

Silencing astrocytic RTP801 in 5xFAD mice changes functional connectivity. (A) Spatial maps of the independent components of the resting‐state functional magnetic resonance. Imaging data were grouped into six resting state networks. Statistical maps showing significant differences (green voxels) in the corresponding ICA connectivity in WT and FAD miCT animals (FAD miCT > WT; p < 0.01) (n = 7). (B) Cingulate–cortex. (C) Amygdala–insular cortex. (D) Hypothalamus–mammillary bodies. (E) Ventral hippocampal–ventral striatal. (F) Sensory cortices. (G) Pontine–striatum. Each point represents data from an individual mouse. Values of Z‐score in these voxels were plotted for the three experimental groups. Data are means ± SEM. In all panels, one‐way ANOVA with Bonferroni's post hoc test was performed: *p < 0.05, **p < 0.01, and ***p < 0.001. WT miCT (n = 7), 5xFAD miCT (n = 7), and 5xFAD miRTP801 (n = 7).

FIGURE 4

Changes in hippocampal metabolites in the 5xFAD mouse model after silencing astrocytic RTP801. Metabolite concentration quantification (mmol/kg w.w) was measured using MRS (magnetic resonance spectrometry), normalized by total creatine concentration. (A) Aspartate. (B) Glutamine. (C) Glutamate. (D) Glycine. (E) Glutathione. (F) NAA. (G) Taurine. (H) GABA. Data are means ± SEM. Each point represents data from an individual mouse. One‐way ANOVA with Bonferroni's post hoc test was performed: *p < 0.05 and **p < 0.01. The number of mice in the MRS experiment: WT miCT (n = 8), 5xFAD miCT (n = 9), and 5xFAD miRTP801 (n = 9). GABA, γ‐aminobutyric acid; NAA, N‐acetyl aspartate.

FIGURE 5

Astrocytic RTP801 reduces the number of parvalbumin⁺ interneurons. (A) Representative immunostaining of hippocampal PV⁺ interneurons in the different groups. Scale bar = 500 and 50 µm. (B) Number of PV⁺ interneurons in the CA1. (C) Area covered by PV⁺ interneurons in the CA1. (D) Mean perimeter of PV⁺ interneurons in the CA1. (E) Number of PV⁺ interneurons in the DG. (F) Area covered by PV⁺ interneurons in the DG. (G) Mean perimeter of PV⁺ interneurons in the DG. Each point represents data from an individual mouse. Data are means ± SEM. One‐way ANOVA with Bonferroni's post hoc test was performed *p < 0.05 and ***p < 0.001. The number of mice in this immunofluorescence experiment was WT miCT n = 5, 5xFAD miCT n = 6, and 5xFAD miRTP801 n = 6.

FIGURE 6

Silencing astrocytic RTP801 mildly recovers astrogliosis and microgliosis in 5xFAD mice. (A) Representative GFAP immunofluorescence microscopy imaging from the CA1 and DG. (B) Quantification of GFAP intensity in CA1. (C) Quantification GFAP+ cells number/field CA1. (D) Quantification of GFAP intensity in DG. (E) Quantification GFAP+ cells number/field DG. (F) Representative S100β immunofluorescence microscopy imaging from the CA1 and DG. (G) Quantification of S100β intensity in CA1. (H) Quantification S100β+ cells number/field CA1. (I) Quantification of S100β intensity in DG. (J) Quantification S100β+ cells number/field DG. (K) Representative AQ4 immunofluorescence microscopy imaging from the CA1 and DG. (L) Quantification of AQ4 intensity in CA1. (M) Quantification of AQ4 intensity in DG. (N) Representative Iba1 immunofluorescence microscopy imaging from the CA1 and DG. (O) Quantification Iba1+ cells number/field CA1. (P) Quantification Iba1+ cells number/field DG. Data are means ± SEM. Each point represents data from an individual mouse. Two‐way ANOVA with Bonferroni's post hoc test was performed: *p < 0.05, **p < 0.01, ***p < 0.001, and ^### p < 0.001. Scales bars = 50 µm.

FIGURE 7

RTP801 influences the levels of inflammasome components in the context of AD. Immunoblottings and densitometric quantifications for (A) cleaved‐NLRP1, (B) NLRP3 (C) ASC, and (D) pro‐caspase 1 and GFP as loading control for transduced astrocytes in the dorsal hippocampus of 7‐month‐old WT miCT, WT miRTP801, 5xFAD miCT, and 5xFAD miRTP801 groups of mice. (E–J) Concentration in pg/mL of cytokines was measured with the Luminex (R&D Systems). (E) IL1‐β, (F) IL‐6 (G) IL‐4, (H) OPN, (I) MCSF, and (J) CCL3 in the hippocampus of the four groups. Data are means ± SEM. Each point represents data from an individual mouse. Two‐way ANOVA with Bonferroni's post hoc test was performed: *p < 0.05, **p < 0.01, and ***p < 0.001.