APOE3 Christchurch modulates β-catenin/Wnt signaling in iPS cell-derived cerebral organoids from Alzheimer's cases

Abstract

A patient with the PSEN1 E280A mutation and homozygous for APOE3 Christchurch (APOE3Ch) displayed extreme resistance to Alzheimer's disease (AD) cognitive decline and tauopathy, despite having a high amyloid burden. To further investigate the differences in biological processes attributed to APOE3Ch, we generated induced pluripotent stem (iPS) cell-derived cerebral organoids from this resistant case and a non-protected control, using CRISPR/Cas9 gene editing to modulate APOE3Ch expression. In the APOE3Ch cerebral organoids, we observed a protective pattern from early tau phosphorylation. ScRNA sequencing revealed regulation of Cadherin and Wnt signaling pathways by APOE3Ch, with immunostaining indicating elevated β-catenin protein levels. Further in vitro reporter assays unexpectedly demonstrated that ApoE3Ch functions as a Wnt3a signaling enhancer. This work uncovered a neomorphic molecular mechanism of protection of ApoE3 Christchurch, which may serve as the foundation for the future development of protected case-inspired therapeutics targeting AD and tauopathies.

Keywords: Alzheimer’s disease; ApoE; ApoE Christchurch; CRISPR; Presenilin; Wnt signaling; iPS cells.

Figure 1

Autosomal dominant Alzheimer’s disease (ADAD) patient in vivo neuroimaging. Patient ω is a non-protected, Paisa kindred control patient and was scanned for amyloid β (A) and tau burden (C). Patient α is an AD-protected Paisa kindred patient and was previously described and scanned for amyloid β (B) and tau burden (D) (Arboleda-Velasquez et al., 2019). Scans indicated elevated amyloid β burden in patient α (B) compared to patient ω (A). Tau burden was elevated in patient ω (C) over patient α (D), with a marked increase in medial temporal and parietal regions. PiB, Pittsburgh Compound B; FTP, Flortaucipir; DVR, Distribution volume ratio; and SUVR, Standardized uptake value ratio.

Figure 2

APOE3Ch decreases pTau S396. Cerebral organoids were formed and stained for nuclei and pTau S396 and imaged at 63X. Representative images from patient ω. (A) and patient α (C) depicted, scale bars = 50 μm. pTau S396 signal intensity was measured, and isogenic controls for each individual patient were averaged together (B,D). Quantification was performed on three to four organoids per line with n = 31–159 measurements for patient α and n = 25–47 measurements for patient ω.

Figure 3

Cell line scRNA-seq UMAP cluster characteristics. scRNAseq data were downsampled and integrated using SCTransform and Harmony. Individual cell line UMAPs show an overlap, indicating that integration was successful (A). Cell cluster transcript lists were run through GSEA C8 Cell Type for cluster identity (B). ^* represents cell cluster identities defined by the second highest hit in GSEA C8 Cell Type analysis and the post hoc transcript expression profile, while ^** represents cell cluster identities defined by post hoc transcript expression profiles. Similar cluster identities were combined to form reference groups for downstream analysis (C). Asterisk denotes cell cluster identity that was second hit in fGSEA due to its better representation of transcriptomic profile and differentiation protocol used (B,C). Pseudotime was performed using the Monocle analysis suite to represent the differentiation timing to cell identity, initiating the analysis at Cluster 9 (Progenitor) as time zero (D). PANTHER pathways analysis was performed on total cells, all clusters (E), neuronal clusters 0, 1, 8, 10, 12, 13, and 14 (F), and glia-type clusters 3, 5, and 7 (G).

Figure 4

APOE3Ch influences reelin, β-catenin, and neural rosette features. Immunofluorescence staining was performed on cerebral organoids to identify changes in protein of interest and physical data. Organoids were imaged at 63X and quantified. Representative images of patient ω (A) and patient α (C) stained for reelin and nuclei. Reelin channel intensity was averaged, with isogenic controls averaged together for patient ω (n = 28–64 measurements per line) (B) and patient α (n = 44–120 measurements per line) (D) using three to four organoids per line. Cerebral organoids were stained for β-catenin using three to four organoids per line rosettes and were identified in patient ω (n = 17–29 measurements per line) and imaged at 63X (E). Segmentation was performed, and regions of interest were defined (F). β-catenin was quantified for both rosette body (G) and rosette ribbon (H). Rosette physical features were then measured for area (I) and aspect ratio (J). Scale bars for panels A, C, and E=50 μm.

Figure 5

Increased nuclear β-catenin in neurons of protected brain regions of patient α. Representative immunofluorescence (IF) micrographs of the frontal cortex (FC), hippocampus (Hipp), and occipital cortex (OC) stained for β-catenin (red), NeuN (green), and cell nuclei (DAPI, blue). Insets present magnified images of neurons showing the degree of colocalization between the three markers. Scale bar = 100 μm. (A). Bar graphs for colocalization analysis depicting thresholded colocalization volumes (TCVs) between DAPI and β-catenin in FC, Hipp, and OC. The percentage of β-catenin colocalizing in nuclei is significantly higher in FC than in both structures, Hipp and OC (one-way ANOVA, p < 0.0001 for both) (B). Bar graphs for colocalization analysis depicting TCVs between DAPI and β-catenin in FC, Hipp, and OC. The percentage of β-catenin colocalizing with neurons is significantly higher only in Hipp when compared to OC (one-way ANOVA, p = 0.025) (C).

Figure 6

ApoE3Ch acts as a Wnt3a signaling enhancer. HEK293 cells with TCF/LEF luciferase reporter element cells were tested against ApoE3 WT, ApoE3Ch, and Wnt3a to assess Wnt signaling activation in triplicate. Cells were treated individually with each compound and in combination (A). Activation was confirmed in a secondary experiment (B). Statistical significance using a one-way ANOVA and post-hoc Tukey’s test to consider a p-value of < 0.05 significance.