Aβ impairs forebrain organoid maturation and induces stress-associated states.

(A) Schematic of the experimental design. Forebrain organoids were differentiated and exposed to Aβ42 oligomers for 8 days starting at day 90. (B–C) Immunofluorescence staining of cortical progenitor and neuronal layer markers in control (CTR) and Aβ42-treated organoids. Expression of TBR2, CTIP2, and SOX2 (B), and SATB2, CTIP2, and BRN2 (C), confirms the presence of cortical progenitor and neuronal populations in both CTR and Aβ42-treated organoids. The section sizes do not represent organoid size difference. (D) Representative immunofluorescence images showing expression of forebrain markers FOXG1, MAP2, SOX2, and PAX6 in CTR and Aβ42-treated organoids, indicating maintenance of forebrain regional identity following Aβ42 exposure. (E) Immunostaining for glial markers GFAP and S100B in CTR and Aβ42-treated organoids, demonstrating the presence of astrocyte-associated populations in both conditions. (F) Detection of Aβ using the 6E10 antibody in CTR and Aβ42-treated organoids. Quantification of 6E10 signal intensity normalized to DAPI shows significantly increased Aβ signal in treated organoids (right; violin plot, P < 0.0001, unpaired t test, n = 4–6 organoids per group). (G) UMAP visualization of integrated single-cell RNA-seq data from day 60 CTR and Aβ42-treated organoids, split by condition and colored by annotated cell types. (H) Comparison of relative cell-type proportions between CTR and Aβ42-treated organoids reveals shifts in the abundance of specific transcriptional states following Aβ treatment. (I) Volcano plot showing differentially expressed genes identified by pseudobulk analysis comparing Aβ42-treated organoids to CTR, highlighting induction of stress-associated genes and reduced expression of neuronal maturation–related genes. (J) Gene set enrichment analysis (GSEA) performed on Gene Ontology Biological Process (GO:BP) terms using differentially expressed genes from pseudobulk analysis comparing Aβ42-treated organoids to CTR. Enriched pathways highlight activation of cellular stress–related processes and depletion of neuronal maturation–associated biological processes in Aβ42-treated organoids. (K) Pseudotime trajectory analysis inferred using Monocle3 and visualized on UMAP embeddings for control (CTR; left) and Aβ42-treated (right) organoids, showing differences in the distribution of cells along the inferred neurogenic trajectory between conditions. (L) Quantification of pseudotime distributions in CTR and Aβ42-treated organoids. Aβ42-treated organoids exhibit a significant shift toward lower pseudotime values compared with CTR (Wilcoxon rank-sum test).

Development and characterization of a 2D–3D hybrid organoid model for analyzing Aβ42-induced pathology.

(A) Heatmap of the top 30 most significantly differentially expressed genes from bulk RNA-seq analysis of day 90 3D Aβ42-treated and CTR organoids. Expression values are shown as scaled normalized expression across individual samples, with hierarchical clustering applied to both genes and samples. (B) Experimental workflow for the study. Day 90 cortical organoids were either maintained as intact 3D organoids for bulk RNA-seq analysis or sectioned into 300-µm slices and cultured on Matrigel-coated plates to generate a 2D–3D hybrid organoid system for synaptic and functional analyses. Hybrid cultures were treated with 5 µM Aβ42 oligomers for 6 days prior to downstream assays. (C) Representative immunofluorescence images of co-stained Syn, MAP2 with VGLUT or GABA to assess synapse ratio. Scale bar: 10 µm. (D–G) Quantitative analysis of synaptic marker density in CTR and Aβ42-treated 2D–3D hybrid organoids. Measurements of excitatory synapses (VGLUT/synaptophysin) and inhibitory synapses (GABA/synaptophysin) reveal increased excitatory and decreased inhibitory synaptic density in Aβ42-treated organoids. (H) Representative calcium imaging traces recorded from CTR and Aβ42-treated 2D–3D hybrid organoids. Images were acquired at 7.67 s per frame. (I) Quantification of calcium imaging data showing increased spike frequency per cell in Aβ42-treated hybrid organoids compared with CTR. (J) GABA response analysis expressed as the post-/pre-treatment spike count ratio, indicating altered GABA responsiveness in Aβ42-treated hybrid organoids. Data are presented as mean ± SEM. For immunofluorescence and functional analyses, n = 4–6 organoids per group. Statistical significance was assessed using unpaired t tests.

Integration of iPSC-derived microglia modulates Aβ42-associated transcriptional states and neural activity in forebrain organoids.

(A) Experimental scheme illustrating the microglia-integrated forebrain organoid model. Forebrain organoids were treated with vehicle (0.1% DMSO; CTR) or Aβ42 oligomers (5 μM) for 6 days, followed by a 48-hour co-culture with iPSC-derived microglia (iMGLs) to generate CTR+iMGL and Aβ42+iMGL conditions. The workflow depicts iPSC differentiation through hematopoietic progenitors, immature microglia, and mature microglia stages. (B) Characterization of iPSC-derived microglia. Flow cytometry analysis (left) shows that approximately 96% of cells are CD11b+ and ∼50% exhibit a CD45^low phenotype, consistent with microglial maturation. Representative immunofluorescence images (right) show iMGLs labeled with IBA1 (green) and Aβ42 detected by 6E10 (red), demonstrating association and engulfment of Aβ42 oligomers by iMGLs. Scale bars: 100 μm (overview), 50 μm (high magnification). (C) Quantification of Aβ42 levels in culture medium by ELISA 48 hours after iMGL integration. Aβ42 levels are significantly reduced in the Aβ42+iMGL condition compared with Aβ42-treated forebrain organoids without microglia (P < 0.0001, unpaired t test, n = 3 samples per group). Data are presented as mean ± SEM. (D) UMAP visualization of integrated single-cell RNA-seq data from Aβ42-treated forebrain organoids (Aβ42) and Aβ42-treated forebrain organoids co-cultured with microglia (Aβ42+iMGL), split by condition and colored by annotated cell types. Major neural progenitor and neuronal populations are present in both conditions, with the appearance of a microglial cluster (yellow) in the Aβ42+iMGL group. (E) Comparison of relative cell-type proportions between Aβ42-treated forebrain organoids with (Aβ42+iMGL) and without (Aβ42) microglial co-culture, revealing shifts in the abundance of specific transcriptional states following microglial integration. (F) Pseudotime trajectory analysis inferred using Monocle3 and visualized on UMAP embeddings for Aβ42+iMGL condition, showing the distribution of cells along inferred developmental trajectories. Colors represent pseudotime values from early (purple/blue) to late (yellow/red) developmental states. (G) Quantification of pseudotime distributions comparing Aβ42-treated forebrain organoids with and without microglial co-culture. Aβ42+iMGL organoids exhibit a significant shift toward higher pseudotime values compared with Aβ42 alone (p = 2.17e-16, Wilcoxon rank-sum test). Box plots show median, quartiles, and individual data points. (H) Cell-cell interaction network analysis showing predicted signaling pathways between cell types in Aβ42+iMGL organoids. Node size and edge thickness represent interaction strength. Colors indicate different cell populations including microglia, various neuronal subtypes, and progenitor cells. (I, J) Microglia outgoing (I) and incoming (J) signaling with environmental cells for various signaling pathways. (K) Experimental design for functional analyses using the 2D–3D hybrid forebrain organoid system.Organoid slices were treated with Aβ42 (5 μM) for 6 days, followed by a 48-hour co-culture with iPSC-derived microglia prior to calcium imaging. (L) Representative calcium imaging field showing neuronal calcium activity labeled with Fluo-4 (green) and iPSC-derived microglia labeled with CellTracker CMTPX (red) in Aβ42-treated hybrid organoids. Scale bar: 100 μm. (M) Quantification of neuronal calcium activity (spike counts per cell) showing no significant difference between CTR+iMGL and Aβ42+iMGL conditions in the presence of microglial co-culture. Individual dots represent single cells; horizontal lines indicate mean ± SEM (n = 3–4 organoids per group). (N) Quantification of microglial calcium activity expressed as the area under the curve (AUC) of ΔF/Fo traces, showing significantly increased microglial calcium activity in Aβ42+iMGL conditions compared with CTR+iMGL (P = 0.0346, unpaired t test). Individual dots represent single microglia; horizontal lines indicate mean ± SEM (n = 3–4 organoids per group). 3K is created in BioRender (https://BioRender.com/6o4kfkx).

Aβ42-treated forebrain organoids with microglial integration promote inflammatory signaling, CD8+ T cell recruitment, and microglia–T cell functional interactions.

(A–E) Quantification of inflammatory cytokine and chemokine secretion from forebrain organoids co-cultured with microglia. Relative secretion levels of CCL4 (A), CCL5 (B), IL-1β (C), TGF-β1 (D), and CXCL10 (E) were measured in conditioned media from CTR+iMGL and Aβ42+iMGL groups. Data are presented as mean ± SEM. Statistical significance was assessed using unpaired t tests. (F) Experimental schematic of the 2D–3D hybrid forebrain organoid model used to assess CD8+ T cell migration. Forebrain organoids were treated with Aβ42 for 6 days and co-cultured with iPSC-derived microglia (iMGLs), followed by transwell-based migration assays using CD8+ T cells with or without chemokine receptor inhibitors targeting CCR5, CXCR3, and CCR1. (G) Quantification of CD8+ T cell migration in the absence (−MG) or presence (+MG) of microglial co-culture. No significant difference in migration was observed between CTR and Aβ42 conditions without microglia, whereas Aβ42-treated forebrain organoids with iMGLs exhibited significantly increased CD8+ T cell migration. (H) Effects of chemokine receptor inhibition on CD8+ T cell migration toward Aβ42+iMGL forebrain organoids. Blockade of CCR5, CXCR3, or CCR1 on CD8+ T cells reduced migration compared with the untreated Aβ42+iMGL condition. (I) Experimental design for calcium imaging analysis of microglia–T cell interactions. iMGLs were integrated into Aβ42-treated forebrain organoids on day 6, followed by addition of CD8+ T cells on day 7, with calcium imaging performed on day 8. (J) Representative multi-channel imaging showing CD8+ T cells (CellTrace Violet, blue), intracellular calcium signals (Fluo-4 AM, green), iMGLs (CellTracker CMTPX, magenta), and Aβ42 (647 nm). Scale bar: 50 μm. (K) Time-lapse calcium imaging illustrating heterogeneous microglial activation states in the presence of CD8+ T cells. Example regions show iMGLs engaging in direct T cell interactions or Aβ42-associated activity over a 13-minute imaging window. Scale bar: 50 μm. (L) Quantitative analysis of microglial calcium activity comparing Aβ42+iMGL and Aβ42+iMGL+T conditions. Metrics include spike count per cell, area under the curve (AUC), and spike amplitude. Individual dots represent single iMGLs. (M) Quantitative comparison of microglial calcium activity between CTR+iMGL+T and Aβ42+iMGL+T conditions using the same metrics as in (L). (N) Quantitative analysis of CD8+ T cell calcium activity comparing CTR+iMGL+T and Aβ42+iMGL+Tconditions, including spike count per cell, AUC, and spike amplitude. Calcium imaging data were acquired at 3.8-s intervals. (O) Schematic overview of the integrated human forebrain organoid platform. Aβ42-treated forebrain organoids are converted to a 2D–3D hybrid system enabling systematic integration of iPSC-derived microglia and CD8+ T cells. The platform supports multi-modal analysis including single-cell transcriptional profiling, real-time calcium imaging, and assessment of immune-neural cellular interactions. Key features include human-specific cellular components, multi-cellular complexity, systematic component addition, and multiple functional readouts. Data are presented as mean ± SEM. Statistical significance was assessed using unpaired t tests. For cytokine analyses, n = 3 samples per group. For migration and calcium imaging analyses, n = 3–4 organoids per group, with individual dots representing single cells where indicated. 4O is created in BioRender (https://BioRender.com/wgpmokj).