Spatially Self‐Organized Three‐Dimensional Neural Concentroid as a Novel Reductionist Humanized Model to Study Neurovascular Development

Abstract Although human pluripotent stem cell (PSC)‐derived brain organoids have enabled researchers to gain insight into human brain development and disease, these organoids contain solely ectodermal cells and are not vascularized as occurs during brain development. Here it is created less complex and more homogenous large neural constructs starting from PSC‐derived neuroprogenitor cells (NPC), by fusing small NPC spheroids into so‐called concentroids. Such concentroids consisted of a pro‐angiogenic core, containing neuronal and outer radial glia cells, surrounded by an astroglia‐dense outer layer. Incorporating PSC‐derived endothelial cells (EC) around and/or in the concentroids promoted vascularization, accompanied by differential outgrowth and differentiation of neuronal and astroglia cells, as well as the development of ectodermal‐derived pericyte‐like mural cells co‐localizing with EC networks. Single nucleus transcriptomic analysis revealed an enhanced neural cell subtype maturation and diversity in EC‐containing concentroids, which better resemble the fetal human brain compared to classical organoids or NPC‐only concentroids. This PSC‐derived “vascularized” concentroid brain model will facilitate the study of neurovascular/blood‐brain barrier development, neural cell migration, and the development of effective in vitro vascularization strategies of brain mimics.


List of Tables:
Table 1: The Top-20 highly variable genes (HVGs) identified for ETV2plus NCs and NCs based on snRNAseq analysis.

Table 2:
The identified gene markers for each cluster obtained based on the UMAP plot of the merged seurat object of ETV2plus NCs and NCs.NES = normalized enrichment score; Green nodes = TFs; Purple nodes = target genes; Blue nodes = genes identified in pySCENIC but not in iRegulon analysis.

Extended Description of Methods: (1) Preparation of Neural Maintenance Medium (NMM)
Neural maintenance medium was prepared by mixing N2 medium and B27 medium as shown in table below at 1:1 ratio.The mixed medium was sterilized using 0.22 µm Bottle Top Vacuum Filter (Corning) and stored at 4 o C up to maximum 3 weeks.

Figure S1 :
Figure S1: Medium-throughput generation of neural spheroids and their progression into a mixed population of neuronal and astroglial cells.(A) The CAD model and 3D-printed micropillar array 24-well plate system as negative mold to produce 24 agarose culture inserts (each containing 137 microwells; diameter x height = 500 µm x 700 µm), and the schematic of agarose culture insert production and spheroid generation (10,000 NPCs per spheroid).In total 3288 NPC spheroids could be generated per batch of 24 agarose culture inserts.(B) Measured spheroid diameter = 273.1 µm + 9.87 on 2 days after seeding, with <10 µm variation (n = 24

Figure S2 :
Figure S2: Characterization of forebrain identity and neuronal subtypes within the neural concentroids.(A) Schematic of culture protocol to promote neural spheroid fusion into concentroids and subsequent maturation by placing ~130 spheroids in each well of a 12-well plate on an orbital shaker (at 75 round per minute).(B) Diameter of concentroids up to 42 days as compared to starting spheroids (n = 5).(C, D, E, G, H) RT-qPCR analysis of representative gene markers for forebrain, dorsal and ventral identity, glutamatergic, and motor neuronal subtypes on day 30 and 70.All data shown are mean + s.d.(n = 3, duplicate measurements; unpaired t-test compared to day 30 (*p<0.05,**p<0.01,***p<0.001).(F, I, J) Immunofluorescence images showing staining for neuronal marker NeuN, glutamatergic (PSD95), and motor (ISLET1) neurons within the neural concentroids on day 30 and 70.Nuclei were counterstained with Hoechst.

Figure S3 :
Figure S3: ETV2 overexpression in neurogenic medium produced EC and formed intertwined neuronal & endothelial network when co-cultured with NPCs.(A) Schematic of EC differentiation from the iETV2-hiPSC line by doxycycline induction.Representative

Figure S4 :
Figure S4: ETV2 overexpression in neurogenic medium produced EC with brain endothelial-like phenotypes.(A) Principal component analysis (PCA) plot of body-wide ECs, hiPSC-EC, ESC-EC, hiPSC-and hPSC-derived MEC or brain MEC (BMEC), as well as iETV2-EC.The iETV2-EC on day 8, 14 and 21 days (delineated by dotted lines) clustered among some body-wide EC isolated from human donors, the MEC and BMEC, but separated in distant from the hiPSC-EC, HSC-EC, ESC-EC, dermal lymphatic EC and fetal pancreas EC.Number in brackets next to the legend of each EC type indicates the number of sample data

Figure S5 :
Figure S5: (A) Functional enrichment analysis of the top-20 highly variable genes (HVGs) in ETV2plus NCs and NCs via g:Profiler showing ETV2plus NC was enriched with pathways mainly associated to nervous system development, whereas NC was enriched with pathways mainly associated to necrotic cell death and hypoxia responses.(B) RNA velocity analysis of top-3 gene markers specific to cluster 3, 4, 7 and 8. (Left panel) Dot plots showing the unspliced (Usp.) and spliced (sp.) ratio of each gene marker per cell per cluster.(Middle and right panels) Heatmaps showing RNA velocity and the expression of each gene marker per cell per cluster.(C, D) UMAP plot showing high velocity confidence was estimated throughout the merged Seurat object of ETV2plus NCs and NCs; UMAP plot showing cluster 4 (C4) had the highest

Figure
Figure S6: (A) Estimated correlation (corr) of each cluster identified in the merged Seurat object of ETV2plus NCs and NCs with the spatial expression of regional marker genes representing forebrain, hindbrain and midbrain at the different embryonic (E11.5, E13.5, E15.5 and E18.5) and postnatal (P4, P14, P28 and P56) stages of the mouse brain.Noted that in general the correlation with forebrain and hindbrain was higher than that of midbrain across the 10 cell clusters.The correlation was less variable (indicated by the less spreading of the data points of the bar plots) at E18.5, P4 and P14 indicating higher correlation than other stages.(B) 2D crosssectional spatial mapping of the correlation of regional marker genes across the 10 cell clusters to that of the developing mouse brain at different stages.(C) Computed 3D mouse brain models (top view; left to right = hindbrain to forebrain) showing in silico spatial mapping of the expression level (by intensity) of top-10 regional marker genes by ETV2plus NCs and NCs to that of the developing mouse brain at different stages.(D) Regional composition analysis for

( 2 )
Embedding of NC and eNC in matrigel droplets with/without iETV2-ECsThe BD Matrigel™ hESC-qualified Matrix (BD Biosciences; 1:200 dilution) was thawed on ice to prevent gelation at room temperature.In the laminar flow, a sterile Parafilm was placed on a holder obtained from 200 ul pipette tips box, and concave cavities on Parafilm were created by gently pressing the Parafilm against the holes of the holder.Each NC and eNC were placed into the concave cavities individually, and 35 ul of pre-cooled matrigel resuspended with or without 1 million day 8 iETV2-ECs was added to each NC and eNC.The matrigel-embedded samples were then incubated for 30 minutes in a cell culture incubator at 37 o C and 95% humidity to allow gelation.Subsequently, the matrigel droplets were transferred into NMM supplemented with BDNF, NT3 and bFGF for 3 days of static culture, after which they were transferred onto a shaker for dynamic culture in BDNF and NT3-supplemented NMM up to day 70.

Table 3 :
The RNA velocity pseudotime obtained via PAGA analysis.

Table 4 :
Cluster-specific putative driver genes identified via RNA velocity analysis.

Table 5 :
The ranked velocity genes per cluster obtained via PAGA analysis.

Table 6 :
Table of adjaciencies showing the identified transcription factors, their target genes and importance for ETV2plus NCs vs. NCs, and between C3, C4, C7 and C8 clusters.

Table 7 :
Table showing the regulon specificity score (RSS) for regulons identified for ETV2plus NCs vs. NCs.

Table 11 :
List of primary antibodies used for immunostaining.

Table 12 :
List of primer sequences used for RT-qPCR.fw = forward sequence; rv = reverse sequence.

Table 13 :
List of accession numbers, cell types, and source/culture conditions for the publicly available RNA sequencing data used in the DESeq2 analysis.