APOE4 exacerbates α-synuclein pathology and related toxicity independent of amyloid

Abstract

The apolipoprotein E (APOE) ε4 allele is the strongest genetic risk factor for late-onset Alzheimer's disease mainly by driving amyloid-β pathology. Recently, APOE4 has also been found to be a genetic risk factor for Lewy body dementia (LBD), which includes dementia with Lewy bodies and Parkinson's disease dementia. How APOE4 drives risk of LBD and whether it has a direct effect on α-synuclein pathology are not clear. Here, we generated a mouse model of synucleinopathy using an adeno-associated virus gene delivery of α-synuclein in human APOE-targeted replacement mice expressing APOE2, APOE3, or APOE4. We found that APOE4, but not APOE2 or APOE3, increased α-synuclein pathology, impaired behavioral performances, worsened neuronal and synaptic loss, and increased astrogliosis at 9 months of age. Transcriptomic profiling in APOE4-expressing α-synuclein mice highlighted altered lipid and energy metabolism and synapse-related pathways. We also observed an effect of APOE4 on α-synuclein pathology in human postmortem brains with LBD and minimal amyloid pathology. Our data demonstrate a pathogenic role of APOE4 in exacerbating α-synuclein pathology independent of amyloid, providing mechanistic insights into how APOE4 increases the risk of LBD.

Figure 1:. Increased α-synuclein pathology in αSyn-APOE4 mice.

Brain sections were prepared from αSyn-APOE mice at 9 months of age. The conformational changed pathogenic α-synuclein was determined by immunohistochemical staining with 5G4 antibody. (A) Representative images are shown for the deposition of 5G4-positive pathogenic α-synuclein in the brain regions of cerebral cortex, CA1 subfield of the hippocampus, amygdala, and thalamus from αSyn-APOE2, αSyn-APOE3, and αSyn-APOE4 mice. Scale bar, 100 μm. (B) The immunoreactivity of 5G4 staining from different brain regions and the overall immunoreactivity from all these regions was evaluated and quantified by Aperio ImageScope (n = 14 – 21 mice per group, mixed gender). Data represent mean ± SEM relative to the αSyn-APOE2 mice. Kruskal-Wallis tests with Dunn’s multiple comparison tests were used. *p < 0.05; **p < 0.01; ***p<0.001; N.S., not significant.

Figure 2:. Impaired behavioral performances in αSyn-APOE4 mice.

Behavioral performance was assessed in Ctrl-APOE and αSyn-APOE mice (n = 10 – 18 mice per group, mixed gender) at 9 months of age. (A) Exploratory behavior was evaluated in the elevated-plus maze and the ratios of the time spent in open arms to close arms are shown. (B and C) Fear conditioning test was utilized to examine associative memory. The percentage of time with freezing behavior in response to stimulus during contextual and cued tests is shown. (D) The hindlimb clasping test was performed to examine the motor coordination. The clasping scores are shown. (E and F) Hangwire tests were performed to evaluate muscle function and coordination. The latency of the first fall off and numbers of falls within two minutes were determined. (G) Rotarod performance tests were used to evaluate motor coordination and balance. The latency to fall off was assessed. Data are expressed as mean ± SEM relative to their own Ctrl-APOE mice. Mann-Whitney U tests followed by Bonferroni correction for multiple comparisons were used. P-values < 0.0167 were considered to be statistically significant. *p < 0.0167; **p < 0.01; ***p < 0.001; ****p < 0.0001; N.S., not significant.

Figure 3:. Neuronal and synaptic loss in αSyn-APOE4 mice.

Brain sections and RIPA lysates from Ctrl-APOE and αSyn-APOE mice at 9 months of age were prepared. Representative images are shown for the NeuN immunohistochemical staining (A). Scale bar, 2 mm. The immunoreactivity of NeuN staining was evaluated and quantified by Aperio ImageScope (n = 12 – 21 mice/group, mixed gender) (B). The post-synaptic makers in the cortical RIPA lysate from Ctrl-APOE and αSyn-APOE mice were evaluated by Western blotting at 9 months of age (n = 6 mice per group, mixed gender) (C). The amount of PSD95 (D), GluR2 (E), and NR2A (F) was quantified. Results were normalized to β-actin expression. Data represent mean ± SEM relative to their own Ctrl-APOE mice. Mann-Whitney tests (B) and student t tests (D-F) followed by Bonferroni correction for multiple comparisons were used. P-values < 0.0167 were considered to be statistically significant. *p < 0.0167; **p < 0.01; ***p < 0.001; N.S., not significant.

Figure 4:. Astrogliosis in αSyn-APOE4 mice.

Brain sections, RNA, and RIPA lysates were prepared from Ctrl-APOE and αSyn-APOE mice at 9 months of age. Representative images are shown for the GFAP immunohistochemical staining (A). Scale bar, 2 mm. The immunoreactivity of GFAP staining (B) was evaluated by Aperio ImageScope (n = 12 – 21 mice per group, mixed gender). The mRNA expression of Gfap was determined by qPCR (C, n = 6 mice per group, mixed gender). The GFAP expression in RIPA lysates was assessed by Western blot (D, E, n = 6 mice per group, mixed gender). The immunoblotting results were normalized to β-actin expression. Data represent mean ± SEM relative to Ctrl-APOE2 mice. Mann-Whitney tests (B) and student t tests (C-E) followed by Bonferroni correction for multiple comparisons were used. P-values < 0.0167 were considered to be statistically significant. *p < 0.0167; N.S., not significant.

Figure 5:. Transcriptomic profiling of αSyn-APOE mice.

RNA sequencing (RNA-seq) was performed using the cortical brain region from Ctrl-APOE and αSyn-APOE mice (n = 6 mice per group, mixed gender) at 9 months of age. (A) Volcano plots of differentially expressed genes (DEGs) identified between the Ctrl and αSyn mice in APOE2-TR (A), APOE3-TR (B), and APOE4-TR mice (C) backgrounds. The blue dots denote downregulated and the red dots denote upregulated DEGs (Benjamini-Hochberg adjusted P < 0.05 and |fold change (FC)| ≥ 1.5). The black circles denote the genes with significant p values (Benjamini-Hochberg adjusted P < 0.05) but |FC| are less than 1.5, and the grey dots denote the genes without marked differences (Benjamini-Hochberg adjusted P ≥ 0.05). The grey dotted lines are the reference threshold for p value (0.05) and fold change (± 1.5). (D) Numbers of DEGs in Ctrl versus αSyn mice are shown. Blue or red bars represent significantly downregulated or upregulated genes in each comparison. (E) Venn diagram shows the overlapped DEG in Ctrl versus αSyn comparison among APOE2-TR, APOE3-TR, and APOE4-TR mice. (F, G) Transcriptome-wide scatter plots demonstrate the correlation of FC in APOE4-TR mice (Ctrl versus αSyn, x axis) and APOE2-TR mice (Ctrl versus αSyn, y axis in F), or APOE4-TR mice (Ctrl versus αSyn, x axis) and APOE3-TR mice (Ctrl versus αSyn, y axis in G) for all genes (each gene corresponds to one point). The blue dots denote genes changed in APOE4-TR mice but not in APOE2-TR or APOE3-TR mice after α-synculein overexpression. (H-O) The key DEGs defined by RNA-seq were validated by qPCR. Data are expressed as mean ± SEM relative to their own Ctrl-APOE mice. Student t tests followed by Bonferroni correction for multiple comparisons were used. P-values < 0.0167 were considered to be statistically significant. *p < 0.0167; **p < 0.01; ***p < 0.001; N.S., not significant.

Figure 6:. Weighted gene co-expression network analysis (WGCNA) in αSyn-APOE4 mice.

(A) The modules associated with the comparison of ctrl versus αSyn in APOE4 mice are shown (n = 6 mice per group). Numbers in the heatmap show the correlation coefficient. Modules with positive values (orange) indicate upregulation in αSyn compared to Ctrl mice; modules with negative values (blue) indicate downregulation. (B) Gene ontology (GO) term enrichment of the turquoise module using 4149 module genes. The orange dotted line indicates the threshold of significance (p = 0.05). (C) Network plot of the top 10 genes with the highest intramodular connectivity (hub genes) in the turquoise module is shown. (D) Trajectory of the module eigengenes (MEs) in the turquoise module between Ctrl-APOE4 and αSyn-APOE4 mice is shown. (E) GO term enrichment of the royalblue module using 147 module genes. (F) Network plot of the top 10 hub genes in the royalblue module is shown. (G) Trajectory of the MEs in the royalblue module between Ctrl-APOE4 and αSyn-APOE4 mice is shown. (H) GO term enrichment of the blue module using 4107 module genes. (I) Network plot of the top 10 hub genes in the blue module is shown. (J) Trajectory of the MEs in the blue module between Ctrl-APOE4 and αSyn-APOE4 mice is shown. Student t tests were used. **p < 0.01.

Figure 7:. Increased α-synuclein pathology in the human postmortem brains of APOE4 carriers with Lewy body dementia (LBD) and minimal AD type pathology.

Human postmortem brain sections from LBD cases were prepared as described in the Materials and Methods section. The p-S129 phosphorylated α-synuclein and conformational changed pathogenic α-synuclein were determined by immunohistochemical staining with pSyn#63 and 5G4 antibodies, respectively. Representative images are shown for the deposition of p-S129-positive (A) and 5G4-positive (B) pathogenic α-synuclein in the superior temporal cortex of APOE4 carriers and non-carriers. Scale bar, 600 μm for the left panel and 50 μm for the right panel in both A and B. The immunoreactivities of p-S129 staining (C) and 5G4 staining (D) were evaluated and quantified by Aperio ImageScope (n = 22 cases per group). Data represent mean ± SEM relative to non-APOE4 carriers. Mann-Whitney U tests were used. *p < 0.05. (E) The correlation between p-S129 and 5G4 immunoreactivities in all the cases was determined by Spearman correlation tests. The black and purple circles represent non-APOE4 carriers and APOE4 carriers, respectively.