Role of Retinoid X Receptors (RXRs) and dietary vitamin A in Alzheimer's disease: Evidence from clinicopathological and preclinical studies

Abstract

Background: Vitamin A (VitA), via its active metabolite retinoic acid (RA), is critical for the maintenance of memory function with advancing age. Although its role in Alzheimer's disease (AD) is not well understood, data suggest that impaired brain VitA signaling is associated with the accumulation of β-amyloid peptides (Aβ), and could thus contribute to the onset of AD.

Methods: We evaluated the protective action of a six-month-long dietary VitA-supplementation (20 IU/g), starting at 8 months of age, on the memory and the neuropathology of the 3xTg-AD mouse model of AD (n = 11-14/group; including 4-6 females and 7-8 males). We also measured protein levels of Retinoic Acid Receptor β (RARβ) and Retinoid X Receptor γ (RXRγ) in homogenates from the inferior parietal cortex of 60 participants of the Religious Orders study (ROS) divided in three groups: no cognitive impairment (NCI) (n = 20), mild cognitive impairment (MCI) (n = 20) and AD (n = 20).

Results: The VitA-enriched diet preserved spatial memory of 3xTg-AD mice in the Y maze. VitA-supplementation affected hippocampal RXR expression in an opposite way according to sex by tending to increase in males and decrease in females their mRNA expression. VitA-enriched diet also reduced the amount of hippocampal Aβ₄₀ and Aβ₄₂, as well as the phosphorylation of tau protein at sites Ser396/Ser404 (PHF-1) in males. VitA-supplementation had no effect on tau phosphorylation in females but worsened their hippocampal Aβ load. However, the expression of Rxr-β in the hippocampus was negatively correlated with the amount of both soluble and insoluble Aβ in both males and females. Western immunoblotting in the human cortical samples of the ROS study did not reveal differences in RARβ levels. However, it evidenced a switch from a 60-kDa-RXRγ to a 55-kDa-RXRγ in AD, correlating with ante mortem cognitive decline and the accumulation of neuritic plaques in the brain cortex.

Conclusion: Our data suggest that (i) an altered expression of RXRs receptors is a contributor to β-amyloid pathology in both humans and 3xTg-AD mice, (ii) a chronic exposure of 3xTg-AD mice to a VitA-enriched diet may be protective in males, but not in females.

Keywords: 3xTg-AD mouse; Aging; Alzheimer's disease; Amyloid; Diet; Prevention; RXRs; Sex; Vitamin A.

Figure 1 –. VitA-supplementation preserved short-term memory of 3xTg-AD mice.

Data are expressed as mean±SEM. Black dots: females, yellow dots: males. A: Protocol used for short-term spatial recognition memory assessment in the Y-maze. Animals were allowed to freely explore only two arms of the apparatus for 5 min during the acquisition phase. An hour later, access was open to the third arm for free exploration of the apparatus during the 5-min-lasting test phase. B: Discrimination ratio (DR) between the distance travelled and the time spent in the familiar and novel arms of the Y-maze. A positive DR indicates a higher exploration of the novel arm, whereas a null or negative DR means no preference or a higher exploration of the familiar arm respectively. C: Total distance travelled in the open field for 30 minutes. D: Number of nose-pokes in the light zone of the Light-Dark box. Statistics: B, One-sample T-test comparison to chance level: # p<0.05, ## p<0.01; Two-way ANOVA, genotype x VitA-supplementation interaction significant (Distance-DR: p=0.007; Time-DR: p=0.006) followed by Tukey’s post hoc test: * p<0.05. C, Kruskal-Wallis test, p=NS. D, Two-way ANOVA, genotype effect: ** p<0.01.

Figure 2 –. VitA status and hippocampal expression of RXRs, Raldh2 and Crapb-I.

Data are expressed as mean ±SEM. A: Plasmatic concentration of retinol. B-F: Hippocampal expression of genes coding for the Retinoid X Receptors (RXRs); Rxr-α (B), Rxr-β (C), Rxr-γ (D), the enzyme responsible for retinoic acid synthesis, Raldh2 (E) and the Cellular Retinoic Acid Binding protein type 1, Crabp-I (F) obtained by RT-PCR in 3xTg-AD females, F, and males, M. Statistics: A, Two-way ANOVA, genotype effect, for females: ** p<0.01, genotype x VitA-supplementation interaction significant (p=0.0003) followed by Tukey’s post-hoc test for males: ## p<0.01. B-F, Unpaired t-test or Mann-Whitney comparison test were used: Tg vs. Tg+vitA, *p<0.05.

Figure 3 –. VitA-enriched-diet reduced β-amyloid peptide levels in male 3xTg-AD mice only.

Data are expressed as mean±SEM. A: Amount of both soluble and insoluble Aβ₄₀ and Aβ₄₂ in the hippocampus of female 3xTg-AD mice with corresponding Aβ₄₂ / Aβ₄₀ ratios in B. C: Amount of both soluble and insoluble Aβ₄₀ and Aβ₄₂ in the hippocampus of male 3xTg-AD mice with corresponding Aβ₄₂ / Aβ₄₀ ratios in D. Statistics: A-D, Unpaired T-test or Mann-Whitney test: Tg vs. Tg+vitA, *p<0.05, **p<0.01.

Figure 4 –. VitA-enriched-diet reduced Tau-phosphorylation in male 3xTg-AD mice only.

A-D: Levels of proteins involved in Aβ-peptide production: Amyloid Precursor Protein, APP (A), ADAM10 (B), BACE (C) and IDE (D). E-G: Protein levels of soluble total Tau protein (E), phospho-Tau on PHF1 epitope (F), CDK5, a major kinase for Tau phosphorylation (G). H-I: Protein levels of two markers of gliosis: Iba-1 (H) and GFAP (I). All protein levels were obtained by Western blot in the hippocampus of female, F, and male, M, 3xTg-AD mice. J: Examples of Western blots assessing the protein levels of 3xTg-AD mice under VitA-control (Tg) and VitA-enriched diets (Tg+vitA). Statistics: A-G, Unpaired T-test or Mann-Whitney test: Tg vs. Tg+vitA, *p<0.05, ** p<0.01.

Figure 5 –. Relationship between RXRs expression and Aβ peptides in all 3xTg-AD mice.

A-B: Correlations between hippocampal Rxr-β mRNA expression and the amount of soluble Aβ₄₀ (A) and Aβ₄₂ (B) produced in the hippocampus of female and male 3xTg-AD mice (Tg and Tg+vitA pooled). C-D: Correlation between hippocampal Rxr-γ mRNA expression and the amount of soluble Aβ₄₂ (C) and insoluble Aβ₄₀ (D) produced in the hippocampus of female and male 3xTg-AD mice (Tg and Tg+vitA pooled). Statistics: A-D, The correlations include all male (under VitA-control and VitA-enriched diets, green line) and all female 3xTg-AD mice (pink line). The lilac and blue dots are illustrated to indicate how the animals of each group are distributed in the correlations. Pearson’s correlation coefficient was determined with a linear regression analysis.

Figure 6 –. Protein levels of RARβ and RXRγ in the parietal cortex of MCI and AD subjects.

Data are expressed as mean ± SEM. The grey and blue dots correspond to women and men, respectively. A-D: Volunteers have been assigned to No-Cognitive Impairment (NCI), Mild-Cognitive Impairment (MCI) or Alzheimer’s Disease (AD) groups according to a clinical diagnosis based on multiple clinical tests assessing cognitive function. Protein levels of the RARβ (A); the 55-kDa-RXRγ (B); the 60-kDa-RXRγ (C) and the ratio between the 60-kDa-RXRγ and the 55-kDa-RXRγ (D) obtained by Western blot in the inferior parietal cortex of the subjects. E: Examples of bands obtained in Western blots when assessing the protein levels of RARβ and RXRγ in human inferior parietal cortex samples according to the clinical diagnosis. (W, woman; M, man). F-I: The same subjects have been assigned to control (C) or Alzheimer’s disease (AD) groups according to a neuropathological diagnosis based on the ABC score evaluating the post mortem amyloidopathy and the Tauopathy in the parietal cortex of the subjects. Protein levels of the RARβ (F); the 55-kDa-RXRγ (G); the 60-kDa-RXRγ (H) and the ratio between the 60-kDa-RXRγ and the 55-kDa-RXRγ (I) obtained by Western blot in the parietal cortex of the subjects. J: Examples of bands obtained in Western blots when assessing the protein levels of RARβ and RXRγ in human inferior parietal cortex samples according to the neuropathological diagnosis. (W, woman; M, man). Statistics: A-C, One-way ANOVA followed by Tukey’s post-hoc test: * p<0.05. D, Kruskal-Wallis test (p<0.05) followed by Dunn’s post-hoc test: * p<0.05. F-G, Unpaired T-test: ** p<0.01. H-I, Mann-Whitney comparison test, * p<0.05.

Figure 7 –. Relationship between RXRγ protein, cognitive scores, insoluble Aβ₄₂ and neuritic plaques in ROS subjects.

The circles and triangles correspond to women and men, respectively. Participants have been assigned to No-Cognitive Impairment (NCI), Mild-Cognitive Impairment (MCI) or Alzheimer’s Disease (AD) groups according to a clinical diagnosis based on multiple clinical tests assessing cognitive function. A-B: Correlation between RXRγ protein levels and cognitive scores: 55-kDa-RXRγ (A); ratio between the 60-kDa-RXRγ and the 55-kDa-RXRγ (B). C-D: Correlation between RXRγ protein levels (the 55-kDa-RXRγ or the ratio between the 60-kDa-RXRγ and the 55-kDa-RXRγ) and insoluble Aβ₄₂ (C) and neuritic plaques (D) found in the parietal cortex of the subjects. Statistics: A-D, Pearson correlation coefficient (R²) was determined with linear regression analysis.