ERβ mediates sex-specific protection in the App-NL-G-F mouse model of Alzheimer's disease

Abstract

Menopausal loss of neuroprotective estrogen is thought to contribute to the sex differences in Alzheimer's disease (AD). Activation of estrogen receptor beta (ERβ) can be clinically relevant since it avoids the negative systemic effects of ERα activation. However, very few studies have explored ERβ-mediated neuroprotection in AD, and no information on its contribution to the sex differences in AD exists. In the present study we specifically explored the role of ERβ in mediating sex-specific protection against AD pathology in the clinically relevant App ^NL-G-F knock-in mouse model of amyloidosis, and if surgical menopause (ovariectomy) modulates pathology in this model. We treated male and female App ^NL-G-F mice with the selective ERβ agonist LY500307 and subset of the females was ovariectomized prior to treatment. Memory performance was assessed and a battery of biochemical assays were used to evaluate amyloid pathology and neuroinflammation. Primary microglial cultures from male and female wild-type and ERβ-knockout mice were used to assess ERβ's effect on microglial activation and phagocytosis. We find that ERβ activation protects against amyloid pathology and cognitive decline in male and female App ^NL-G-F mice. Ovariectomy increased soluble amyloid beta (Aβ) in cortex and insoluble Aβ in hippocampus, but had otherwise limited effects on pathology. We further identify that ERβ does not alter APP processing, but rather exerts its protection through amyloid scavenging that at least in part is mediated via microglia in a sex-specific manner. Combined, we provide new understanding to the sex differences in AD by demonstrating that ERβ protects against AD pathology differently in males and females, warranting reassessment of ERβ in combating AD.

Keywords: APP knock-in; Alzheimer’s disease; Amyloidosis; Estrogen receptor beta; Microglia; sex differences; sex hormone.

FIGURE 1.. ERβ activation improves cognitive behavior in App^NL-G-F male and female mice.

(A) Treatment regime of App^NL-G-F mice. (B) Percent Y-maze alterations of male (left) and female (right) App^NL-G-F mice treated with vehicle or ERβ agonist LY500307 (LY) (n = 7–10). (C) Percent context-associated freezing time of male (left) and female (right) App^NL-G-F mice (n = 6–9) in the contextual fear conditioning test. Cued-associated freezing time in the contextual fear conditioning test of (D) male and (E) female App^NL-G-F mice before cue (baseline) and upon cue (tone) in a different cage context (n = 6–9). Female mice were either ovariectomized (OVX) or sham operated (Sham). * P < 0.05, ** P < 0.01, *** P < 0.001. Unpaired t-test was used for males and 2-way ANOVA for females followed by uncorrected Fisher’s LSD test for multiple comparisons. Overall significant main effects of treatment or OVX are indicated.

FIGURE 2.. Less Aβ pathology in App^NL-G-F male and female mice after ERβ activation.

(A) Immunohistochemical representation of amyloid plaques in frontal and motor cortex (FT/M), somatosensory and visual cortex (Ss/Vis) and hippocampus (Hippoc) of male App^NL-G-F mice after vehicle or LY treatment. (B) Quantification of number of plaques per 100 µm² (n = 4–6) and (C) percent plaque area (n = 4–6) in male App^NL-G-F mice. (D) Similar as in (A), immunohistochemical representation of amyloid plaques in different brain regions of female App^NL-G-F mice after vehicle or LY treatment. (E) Quantification of number of plaques per 100 µm² (n = 3–4) and (F) percent plaque area (n = 3–9) in female App^NL-G-F mice. (G) Linear regression analysis comparing effect size from LY treatment (vehicle vs. LY) on number of Aβ plaques in relation to average number of ERβ positive cells per 100 µm² in different brain regions of male and female mice (n = 3–6). (H) Soluble and (I) insoluble Aβ₄₂ levels in male cortex (Ctx, left) and hippocampus (Hippoc, right) (n= 3–4). (J) Soluble and (K) insoluble Aβ42 levels in female cortex (Ctx, left) and hippocampus (Hippoc, right) (n = 3). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Unpaired t-test was used for males and 2-way ANOVA for females followed by uncorrected Fisher’s LSD test for multiple comparisons. Overall significant main effects of treatment or OVX are indicated. Scale bars = 100 µm.

FIGURE 3.. ERβ activation does not alter APP processing.

Western blot analysis of full-length APP (FL-APP), β-CTF and α-CTF, Aβ peptide, and β-actin in (A) cortex and (B) hippocampus of female and male App^NL-G-F mice after vehicle (V) or LY treatment, as well as after sham surgery or ovariectomy (OVX in females). Quantification of (C) FL-APP relative to β-actin, (D) β-CTF relative to FL-APP, (E) β-CTF relative to actin, and (F) Aβ relative to FL-APP in male (left), and female (right), cortex (Ctx) (top), and hippocampus (bottom) (n = 3–4). * P < 0.05, ** P < 0.01, *** P < 0.001. Statistical significance was determined using unpaired t-test for males and 2-way ANOVA for females followed by uncorrected Fisher’s LSD test for multiple comparisons. Overall significant main effects of treatment or OVX are indicated.

FIGURE 4.. ERβ activation modulates microglia activation in a sex-specific manner in App^NL-G-F mice.

(A) Representative immunofluorescence images of male App^NL-G-F hippocampus stained with the amyloid stain AmyloGlo (magenta), Iba1 (green), and CD68 (white) after vehicle or LY treatment. Yellow dotted area (left) indicates magnified region of interest (right). Arrowheads indicate microglia with lower CD68 levels. Scale bar 100 µm (left) and 50µm (right). Quantification in male App^NL-G-F mice of (A) number of Iba1 cells per 100 µm² (n = 4–5), (B) percent CD68+, Iba1+ double positive cells (n = 3–5), and (C) percent microglia within 20 µm radius of plaque edge (n = 4–6). Representative immunofluorescence images of female App^NL-G-F hippocampus stained with AmyloGlo (magenta), Iba1 (green), and CD68 (white) after vehicle or LY treatment, as well as after sham surgery or ovariectomy (OVX). Yellow dotted area (left) indicates magnified region of interest (right). Arrowheads indicate microglia with lower CD68 levels. Scale bar 100 µm (left) and 50µm (right). Quantification in female App^NL-G-F mice of (A) number of Iba1 cells per 100 µm² (n = 4), (B) percent CD68+, Iba1+ double positive cells (n = 3), and (C) percent plaque-associated microglia (n = 3). * P < 0.05, *** P < 0.001. Unpaired t-test was used for males and 2-way ANOVA for females followed by uncorrected Fisher’s LSD test for multiple comparisons. Overall significant main effects of treatment are indicated.

FIGURE 5.. Microglial and proinflammatory markers are altered upon ERβ activation in a sex-specific manner.

Multiplex ELISA analysis of the proinflammatory markers IL-1b, CXCL1 (KC/GRO), and IL-12p70, as well as IL-10, in male (left) and female (right) APP^NL-G-F cortex (Ctx) (A, C, E, G) and hippocampus (Hippo) (B, D, F, H), after vehicle or LY treatment, as well as after sham surgery or ovariectomy (OVX in females) (n = 3–4). Expression of microglial and inflammatory markers (I) Cd68, (J) Nos2, (K) Trem2, and (L) Cx3cr1 relative to housekeeping gene Rplp0 expression in male (left) and female (right) App^NL-G-F hippocampus after vehicle or LY treatment, as well as after sham surgery or OVX in females (n = 3–7). * P < 0.05, ** P < 0.01, *** P < 0.001. Unpaired t-test was used for males and 2-way ANOVA for females followed by uncorrected Fisher’s LSD test for multiple comparisons. Overall significant main effects of treatment or OVX are indicated.

FIGURE 6.. Loss of ERβ modulates sex-specific effects in primary mouse microglia.

(A) Representative immunofluorescence of ERβ and Iba1 co-staining (arrowheads) in WT (left) and Esr2-KO (right) male cortex (dotted rectangle: magnified area), and (B) in male App^NL-G-F cortex upon vehicle or LY treatment (scale bars = 50 µm). Phagocytosis of Aβ (relative to Cytochalasin D treatment) of primary microglia originating from male (C) and female (D) WT and Esr2-KO mice. The microglia were treated with or without 10 nM LY 24h prior to and during 24h M1, M2, or no (CTL) activation stimulation (n = 3–4). Transwell migration assay of primary microglia originating from male (E) and female (F) WT and Esr2-KO mice treated as indicated above (n = 6). Expression of microglial markers Nos2, Trem2, Cx3cr1, P2ry12, Arg1, and Cd68 (G-R) relative to housekeeping gene Rplp0 expression in primary microglia originating from male (G, I, K, M, O, Q) and female (H, J, L, N, P, R) WT and Esr2-KO mice treated as indicated above (n = 3–7). * P < 0.05, ** P < 0.01. Unpaired t-test was used for male microglia and 3-way ANOVA or mixed-effects analysis for female microglia followed by Tukey’s or Dunnett’s multiple comparison test. Overall significant main effects of LY treatment or genotype are indicated.