All-in-one generation and multiomic profiling of human whole brain organoid on a millifluidic plate

Abstract

Human brain organoids (hBOs) have been recently regarded as neurobiologically relevant brain models and exponentially exploited in a variety of neuroscience research. However, the current gold-standard method for generating hBOs is intricate and laborious, resulting in hBOs with morphological variability and inconsistent batch-to-batch reproducibility. Despite several studies reporting simplified hBO culture methods, few of those methods was biologically validated with multiomic profiling, which is crucial for neurobiological studies. Here, we demonstrate an all-in-one millifluidic plate (AIOMP) with individually perfusable microchambers for hBOs, simplifying the culture process, improving the uniformity and reproducibility, and enabling long-term cultivation and real-time morphogenesis observation. Additionally, our comprehensive transcriptomic and proteomic analyses revealed that AIOMP increases neurogenesis and corticogenesis of hBOs, suggesting a stronger correlation between the AIOMP-generated hBOs and human fetal brain than those generated through conventional method. Metabolomic and neurophysiological results further support the maturation-enhancing effects of AIOMP on hBOs, showing improved neurotransmitter synthesis and electrophysiological functionality. Overall, the AIOMP approach offers a simplified, reproducible, and biologically validated method for hBOs generation and maturation, with potential applications in neurobiology, neurological disease research, and central nervous system drug assessment.

Keywords: Brain organoid; Long-term culture; Millifluidic plate; Multiomic profiling; Organ-on-a-chip.

Graphical abstract

Fig. 1

Evaluation of the homogeneity and reproducibility of brain organoid formation in the AIOMP. (A) Schematic diagram of conditions used to culture brain organoid by orbital shaker (traditional method) and by AIOMP. (B) The physical map of AIOMP. (C) Partial view of AIOMP single culture unit. (D) Brightfield images of brain organoids at different days. (Scale bar = 100 μm for D2 and D7; 500 μm for D14, D21, D28, and D35). (E, F) The diameter and the ratio of length to width of the brain organoids (n = 12). (G, H) Representative images of hBOs cultured in the AIOMP and CSC group at D35, respectively. (Scale bar = 500 μm for G; 5000 μm for H). (I, J) The variation of the diameter and the ratio of length to width of the hBOs in each condition at D35 (n = 4). (K) Schematic diagrams of the morphological assessment criteria based on the distribution of the neuroepithelium like structures: Type i, neuroepithelium like structure all around the aggregate; Type ii, neuroepithelium like structure more than three quarters around; Type iii, neuroepithelium like structure more than halfway around; Type iiii, neuroepithelial structure below the quarter around. (L) Percentages of organoids with each morphological type. AIOMP, n = 33; CSC, n = 30).

Fig. 2

Numerical simulations and condition optimization for dynamic culture. (A) Computational fluid dynamic analysis of fluid velocity for the AIOMP under different rotation cycle conditions (T = 0.5 s, 5 s, 10 s, 20 s, 50 s, and 100 s). (B) The fluid dynamic analysis and statistical analysis of flow velocity under the condition of an inclination angle of 0°. (C) Bright-field images of hBOs under different rotation conditions during differentiation (D21 and D28) (Scale bar = 500 μm). (D) The percentages of growth state analysis of hBOs (n = 12). (E) The ratio of length to width of the hBOs in different rotation conditions (T = 0.5 s, 20 s, and 100 s). (F) Distribution analysis results of the velocity under different cycle conditions in organoid culture chamber and 6-well plate (n = 8). (G) Quantitative analysis of the velocity simulation results in organoid culture chamber of AIOMP and 6-well plate within a single cycle. (H) Distribution analysis results of shear stress under different cycle conditions in organoid culture chamber of AIOMP and 6-well plate. (I) Quantitative analysis of shear stress simulation results in organoid culture chamber and 6-well plate within a single cycle.

Fig. 3

AIOMP enhances corticogenesis and maturation in hBOs. (A) Pearson correlation matrix for transcriptome-wide profiles of organoids in the AIOMP and CSC group. Pearson's correlation coefficient (PCC) values are indicated in each box. (B) The top 30 enriched Gene Ontology (GO) neurodevelopment-related terms of upregulated genes in the AIOMP/CSC (shown in terms of p values). (C–D). Heatmaps of Pearson's correlation analysis of RNA-sequencing datasets of the AIOMP and CSC organoids for comparison with published Allen Brain Span human transcriptome dataset of human frontal regions across different stages and (C) different regions of the brain at postnatal stages. Pearson's Correlation Coefficient (PCC) values are displayed in a heatmap. (E, F) The expression of hindbrain marker (ISL1) was identified by immunofluorescence staining, and the mRNA of ISL1 (F) was quantified by RT-qPCR in the AIOMP and CSC group (Scale bar = 100 μm, independent replicate = 3). (G, H) The expression of forebrain marker (PAX6) was identified by immunofluorescence staining, and the mRNA of PAX6 (I) was quantified by RT-qPCR in the AIOMP and CSC group (Scale bar = 100 μm, independent replicate = 3). (I) Immunostaining for markers of deep-layer neuron markers (TBR1), and immature neurons (Tuj1) in hBOs by AIOMP or CSC (F) at D35. (J) Transcripts for TBR1 were examined by RT-qPCR of hBOs in the AIOMP and CSC group at D35 (Scale bar = 100 μm, independent replicate = 3). (K, L) Immunofluorescence image and schematic representation of ventricular zone (VZ, SOX2⁺) and cortical plate layer (CP, Tuj1⁺) measurement in cortical structures and (M) plot for relative CP thickness in the day 35 hBOs (Scale bar = 100 μm, independent replicate = 3). The relative levels of genes were normalized to GAPDH. Data represent the mean ± SEM.

Fig. 4

AIOMP improves neurogenesis in the hBOs. (A) Schematic representation of label-free quantitative proteomics of the organoids generated in the AIOMP and CSC. (B) Distribution of the peptides mass error (ppm) based on m/z of 2516 identified peptides. (C) Venn diagram showing the total numbers of proteins identified in the AIOMP/CSC organoids. (D) Volcano plot of differential protein quantification for proteomics. Red represents upregulated proteins and Purple represents downregulated proteins. (E) Top 20 neurodevelopment-related enriched Gene Ontology (GO) terms of upregulated proteins in the AIOMP/CSC organoids and ordered by p value. (F) Bubble plot showing top enriched items for KEGG pathway analysis in differential proteins. (E) The top 30 enriched gene ontology (GO) terms of upregulated genes in the AIOMP versus the CSC group (shown in terms of p values) (F) Immunostaining of the marker of neural stem cell (SOX2 and Nestin), neuron (Tuj1 and MAP2), and astrocyte (GFAP) in the AIOMP and CSC groups at D35 (Scale bar = 100 μm). (G) Immunostaining for the proliferation marker Ki67 and the progenitor marker SOX2 in the AIOMP and CSC group at D35 (Scale bar = 100 μm). (n = 3). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

AIOMP facilitates the long term cultivation and functional maturation of hBOs. (A) Schematic illustration of electrophysiological activity and neurotransmitter detection in brain organoids (D100) in the AIOMP and CSC groups. (B, C) Diameter and length-to-width ratio of brain organoids in the AIOMP and CSC groups (n = 12). (D) Heatmap comparing the relative synthesis levels of 31 neurotransmitters in the AIOMP and CSC groups. (E) Donut charts showing the proportions of detected and undetected electrical signals in the AIOMP and CSC groups. (F) Representative image of action potentials in brain organoids in the AIOMP and CSC groups. (G) Typical electrographic features in brain organoids in the AIOMP and CSC groups. (H) Statistical graph of number of spikes in 5 min. (I) Statistical graph of mean firing rates (Hz) in brain organoids in the AIOMP and CSC groups. Data represent the mean ± SEM. (n ≥ 7).

Fig. S1

Design and characterization of AIOMP. (A) Schematic diagram of the fabrication process for the AIOMP. (B) Illustration of the 3D view, top view, and cross view for organoid cultivation chamber of the AIOMP and dimensions of each part. (C) Photograph of AIOMP and cultivation chamber. (D-F) Representative bright-field images and quantification analysis of the U-shape bottom in the cultivation chamber of AIOMP (Scale bar, 1 mm; n = 12).

Fig. S2

Diffusion analysis for the AIOMP and CSC groups. (A) Distribution analysis results of diffusion under different cycle conditions in organoid culture chamber of AIOMP and 6-well plate. (B) Spatial schematic diagram of flow velocity simulation.

Fig. S3

Differentiation procedure and bright field identification of brain organoids in AIOMP and CSC groups. (A) Schematic diagram of hBOs culture procedure. (B) In situ brightfield scans of D35 brain organoids cultured in AIOMP. Scale bar, 2000 μm. (C) Images of brain organoids generated in AIOMP and CSC immersed in OCT within embedding cassette. Scale bar, 3 mm.

Fig. S4

AIOMP enhances neuron associated transcriptional gene expression in organoids. (A) The violin box plot indicates the overall dispersion of expression. The horizontal axis abscissa in the figure represents different samples, and vertical axis represents the logarithmic value of the sample expression although FPKM. (B) Heat map analysis of all gene expression levels. (C) PCA cluster analysis of transcriptome gene expression. (D) Volcano plot analysis of transcriptome DEGs. Red represents upregulated genes and green represents downregulated genes, and GO annotation information of DEGs, according to the distribution of upregulated proteins of BP (E) CC (F) and MF (G) in GO terms of level 2. (H and I) RT-qPCR showing expression of forebrain (FOXG1) and hindbrain (PAX2) genes in the cerebral organoids in AIOMP and CSC groups at D35.

Fig. S5

Profiling of LFQ proteomics in brain organoids generated in AIOMP and CSC. (A) The coverage analysis in globe proteomics. (B) Pearson correlation coefficient test for triplicate experiments. (C) Volcano plot of differential proteins. (D) Functional analysis in GO enrichment of different expressed proteins according to Biological Processes (BP), Cellular Components (CC), and Molecular Functions (MF). (E) RT-qPCR showing expression of neural progenitor (SOX2, Nestin), neuron (Tuj1, MAP2), and astrocyte (GFAP) genes in the brain organoids in AIOMP and CSC groups at D35. Data represent the mean ± SEM.

Fig. S6

The long-term cultivation of brain-like organs and neurotransmitter levels produced in AIOMP and CSC. (A) Bright-field images of brain organoids at D100 (Scale bar, 1 mm). (B-D) Statistical graphs of neurotransmitter synthesis levels. (n = 3) (E) Volcano plot of differential neurotransmitters. (F) The radar plot of differential neurotransmitters. Data represent the mean ± SEM.