Characterization of astrocytes throughout life in wildtype and APP/PS1 mice after early-life stress exposure

Abstract

Background: Early-life stress (ES) is an emerging risk factor for later life development of Alzheimer's disease (AD). We have previously shown that ES modulates amyloid-beta pathology and the microglial response to it in the APPswe/PS1dE9 mouse model. Because astrocytes are key players in the pathogenesis of AD, we studied here if and how ES affects astrocytes in wildtype (WT) and APP/PS1 mice and how these relate to the previously reported amyloid pathology and microglial profile.

Methods: We induced ES by limiting nesting and bedding material from postnatal days (P) 2-9. We studied in WT mice (at P9, P30, and 6 months) and in APP/PS1 mice (at 4 and 10 months) (i) GFAP coverage, cell density, and complexity in hippocampus (HPC) and entorhinal cortex (EC); (ii) hippocampal gene expression of astrocyte markers; and (iii) the relationship between astrocyte, microglia, and amyloid markers.

Results: In WT mice, ES increased GFAP coverage in HPC subregions at P9 and decreased it at 10 months. APP/PS1 mice at 10 months exhibited both individual cell as well as clustered GFAP signals. APP/PS1 mice when compared to WT exhibited reduced total GFAP coverage in HPC, which is increased in the EC, while coverage of the clustered GFAP signal in the HPC was increased and accompanied by increased expression of several astrocytic genes. While measured astrocytic parameters in APP/PS1 mice appear not be further modulated by ES, analyzing these in the context of ES-induced alterations to amyloid pathology and microglial shows alterations at both 4 and 10 months of age.

Conclusions: Our data suggest that ES leads to alterations to the astrocytic response to amyloid-β pathology.

Keywords: APP/PS1; Astrocytes; Aβ pathology; Early stress; GFAP; Glia.

Fig. 1

Characterizing ES effects on hippocampal GFAP expression in different developmental stages of WT mice. Representative images of GFAP expression in the HPC of P9 CTL (a) and ES (b) mice. Insets depict GFAP+ cells at 100x magnification in the SLM. At P9, ES does not affect GFAP coverage in the CA (c) but increases GFAP coverage in the SLM (d). No differences in GFAP+ cell density (e) or GFAP+ cell complexity (f) were found in the SLM region. Representative images of GFAP expression in the HPC of 6-month CTL (g) and ES (h) mice. Insets depict GFAP+ cells at × 100 in the SLM. GFAP coverage is not affected by ES at this age in the CA (i) and SLM (j). No differences in GFAP+ cell density (k) or GFAP+ cell complexity (l) were found in the SLM region. Statistical analyses were performed using independent t tests for GFAP coverage and cell density and repeated measures ANOVA for cell complexity. ^*Significant effect of condition p < 0.05; scale bars: representative images 250 μm, inset images 12.5 μm

Fig. 2

ES does not affect GFAP expression in the HPC or EC of 4-month-old-mice. Representative images of GFAP expression in the HPC of APP/PS1-CTL (a) and APP/PS1-ES (b) mice and the EC of APP/PS1-CTL (d) and APP/PS1-ES (e) mice. Both ES and APP/PS1 did not alter GFAP expression in either the HPC (c) or EC (f) of 4-month-old mice. Statistical analyses were performed using two-way ANOVA. Scale bars 250 μm

Fig. 3

APP/PS1 overexpression increases GFAP coverage and complexity in the EC of 10-month-old mice. Representative images of the EC of 10-month-old WT-CTL (a), WT-ES (b), APP/PS1-CTL (c), and APP/PS1-ES (d) mice. GFAP coverage in the EC was increased in APP/PS1 mice (e). APP/PS1 increased GFAP+ cell complexity in the EC as shown by increased number of intersections (f), area under the curve (AUC) (g), and number of primary processes (h). Dots in Sholl analysis graphs are color-coded per animal. Statistical analyses were performed using two-way ANOVA. ^$Significant effect of genotype p < 0.05, scale bars 250 μm

Fig. 4

ES and APP/PS1 overexpression affect GFAP coverage in the HPC of 10-month-old mice. Representative images of the HPC of 10-month-old WT-CTL (a), WT-ES (b), APP/PS1-CTL (c), and APP/PS1-ES (d) mice. ES modulates the effect of APP/PS1 on GFAP expression in the HPC (e). GFAP coverage in the CA was decreased by both ES and APP/PS1 (f). APP/PS1 increases hippocampal GFAP expression while ES decreases GFAP expression in the SLM (g). GFAP+ cell complexity was unaffected by either ES or APP/PS1 (h). Representative images of clustered GFAP coverage in the HPC of 10-month-old WT-CTL (i), WT-ES (j), APP/PS1-CTL (k), and APP/PS1-ES (l) mice. Clustered GFAP coverage was increased by APP/PS1 overexpression but decreased by ES in whole HPC (m), CA (n), and SLM (o) region. Statistical analyses were performed using two-way ANOVA. ^$Significant effect of genotype p < 0.05; ^*Significant effect of condition p < 0.05; ^{^}Significant post hoc effect p < 0.05; scale bars 250 μm

Fig. 5

Relationship between astrocytes, microglia, and plaque load in APP/PS1 mice. a Summary of microglia and plaque load as published in [14]. b Global GFAP is increased in the dentate gyrus (DG) of 4-month-old (4 months) ES-exposed APP/PS1 mice after normalizing to number of amyloid-β + cells. c Masked GFAP is decreased in the DG of 10-month-old (10 months) ES-exposed APP/PS1 mice after normalizing to extracellular plaque load. d, e Pearson correlation coefficients visualized from − 1 (blue) to + 1 (red), with the size of the circles representing the correlation coefficients. Circles without crosses are correlation coefficients significant at p < 0.05. d Correlation matrix of staining data from 4-month APP/PS1 mice. e Correlation matrix between staining data from 10-month APP/PS1 mice. f CD68 coverage is negatively correlated with the number of amyloid-beta positive cells in the DG at 4 months. g CD68 coverage is positively correlated to masked GFAP coverage in the CA region of 10 months APP/PS1 mice. ^*Significant effect of condition p < 0.05

Fig. 6

Overview of ES and APP/PS1-induced effects on hippocampal astrocytic GFAP. Stress exposure early in life from postnatal days 2–9 is associated with an increase of GFAP in specific hippocampal subregions at P9. No effects are observed at 4 months. At 10 months, GFAP coverage is reduced in wildtype mice. In APP/PS1 mice, no effects are observed at 4 months (early pathological stage) of either genotype or ES exposure. At 10 months (advanced pathological stage), we observe genotype-induced reduction in global and increase in localized GFAP clustering, presumably due to Aβ accumulation. Notably, amyloid pathology is modulated by ES [14], decreasing cell-associated amyloid at 4 months, and increasing amyloid load at 10 months. While ES did not affect the absolute measures of GFAP, this suggests altered astrocytic response to amyloid pathology after ES