By L. Wang, C. Wang, J. A. Moriano, S. Chen, S. Zhang, T. Mukhtar, S. Wang, A. Cebrian-Silla, Q. Bi, J. J. Augustin, L. G. de Oliveira, M. Song, X. Ge, G. Zuo, M. F. Paredes, E. J. Huang, A. Alvarez-Buylla, X. Duan, J. Li, A. R. KriegsteinSourcebioRxiv
Fork
Clone
git clone https://openversions.com/git/index/biorxiv-10-1101-2024-01-16-575956v1.git
Cite
Version link
https://openversions.com/index/biorxiv-10-1101-2024-01-16-575956v1/commit/e8467a570f14f190ae8fc829fdfdc3dcf95e57bc
Version identifier (commit hash)
e8467a570f14f190ae8fc829fdfdc3dcf95e57bc

Fork to create your own copy with credit to the source. Clone to get a Git copy on your computer.

Latest version

Checks complete

Archive from bioRxiv Source: https://www.biorxiv.org/content/10.1101/2024.01.16.575956v1 License: https://creativecommons.org/licenses/by-nc-nd/4.0/ Manuscript unchanged.

Committed

Claims

Findings

Headline finding

A single-cell multi-omic and spatial transcriptomic atlas of the developing human neocortex across five developmental stages (first trimester to adolescence) and two cortical regions (PFC and V1) catalogs cell type-, age-, and area-specific gene regulatory networks, identifies a tripotential intermediate progenitor subtype (Tri-IPC) responsible for local GABAergic neuron production, and maps enriched ASD genetic risk to second-trimester intratelencephalic projection neurons.

Paired snMultiome profiling of 232,328 neocortical nuclei identifies 5 classes, 11 subclasses, and 33 high-fidelity cell types, demonstrating that dynamic changes in the epigenome and transcriptome are highly coordinated and that cell type proportions vary substantially across developmental stages and cortical regions.

MERFISH spatial transcriptomic analysis of 404,030 cells defines 29 cell types and 10 cortical niches, revealing lineage-specific migration route preferences for CGE- versus MGE-derived interneurons and reciprocal cell-cell communication between excitatory and inhibitory neurons.

In the second trimester, CGE-derived and MGE-derived immature interneurons exhibit distinct migration route preferences, with a marginal zone to VZ/SVZ ratio of 1:3.3 for CGE-derived interneurons and 1:5.2 for MGE-derived interneurons (P<2.2E−16P < 2.2\text{E}-16, Fisher's exact test).

Gene regulatory network analysis using SCENIC+ identifies 582 eRegulons (385 activators and 197 repressors), and trajectory analysis of excitatory neuron lineages reveals that V1-specific EN-L4-ITs diverge from common EN-L4-ITs at bifurcation point BP2 and are governed by distinct transcriptional regulators including CUX1, SMAD3, GLIS3, and POU6F2.

In the late second-trimester human neocortex, EGFRhighOLIG2+EGFR^{high}OLIG2^+ IPC-Glia (renamed Tri-IPCs) are tripotential intermediate progenitors derived from both RG-tRGs and RG-oRGs that generate astrocytes, OPCs, and MEIS2+PAX6+MEIS2^+PAX6^+ GABAergic interneurons.

Single-cell clonal analysis demonstrates that 5 out of 10 RG-tRG clones and 5 out of 10 RG-oRG clones contain both IPC-ENs and Tri-IPCs, while 7 out of 10 Tri-IPC clones contain astrocytes, OPCs, and interneurons.

Integrating snATAC-seq chromatin accessibility with GWAS data via SCAVENGE reveals that autism spectrum disorder (ASD) common risk variants are most strongly enriched in second-trimester intratelencephalic (IT) projection neurons (EN-IT-Immatures and EN-L6-ITs) and peak earlier along excitatory neuron pseudotime than MDD, BPD, and SCZ.

Methods

Paired single-nucleus RNA-seq and ATAC-seq (snMultiome) profiling of 38 human neocortical samples using 10X Genomics Chromium Next GEM Single Cell Multiome ATAC + Gene Expression, processed with Cell Ranger ARC (v2.0.0), Scrublet doublet removal, MACS2 (v2.2.7), Signac (v1.10.0), Seurat v4, and weighted nearest neighbor (WNN) clustering.

Source excerpts

“Nuclei were counted using a hemocytometer, diluted to 3220 nuclei/μL, and further processed following 10X Genomics Chromium Next GEM Single Cell Multiome ATAC + Gene Expression Reagent Kits user guide.” · Show in paper

“Weighted nearest neighbor analysis was done using Seurat v4 with 1–50 PCA components and 2–40 LSI components. The resulting nearest neighbor graph was used to perform UMAP embedding and clustering using the SLM algorithm87.” · Show in paper

Spatial transcriptomic profiling of 6 human neocortical samples using MERFISH on the Vizgen MERSCOPE platform with a custom 300-gene panel, analyzed with Seurat v5, k-means niche clustering of 50 nearest neighbors, Squidpy (v1.2.3) neighborhood enrichment, and NCEM (v0.1.4) intercellular communication modeling.

Source excerpts

“Spatial transcriptomic analysis using MERFISH was done using the Vizgen MERSCOPE platform. We designed a customized 300-gene panel composed of cell type markers (Supplementary Table 5b) using online tools at https://portal.vizgen.com/.” · Show in paper

“The graph neural network-based node-centric expression modeling (NCEM v0.1.4) method28 was used for intercellular communication modeling (Supplementary Table 7).” · Show in paper

Ligand-receptor communication inference across 232,328 neocortical nuclei using CellChat (v1.6.1) with CellChatDB.

Source excerpts

“We implemented CellChat (v1.6.1)29 to quantify the strength of interactions among cell types using default parameter settings (Supplementary Table 8).” · Show in paper

Gene regulatory network inference using SCENIC+ (v0.1.dev448+g2c0bafd) on 10,000 geometric-sketched nuclei and AUCell (v1.20.2) scoring across all 232,328 nuclei.

Source excerpts

“We implemented the SCENIC+ (v0.1.dev448+g2c0bafd) workflow30 to build gene regulatory networks of developing human neocortex based on the snMultiome data.” · Show in paper

Excitatory neuron trajectory inference using Slingshot (v2.6.0) and trajectory-based differential eRegulon activity analysis using tradeSeq (v1.12.0) beta GAM and earlyDETest.

Source excerpts

“Cells belonging to excitatory neuronal lineages, including radial glial cells, IPC-EN, and glutamatergic neurons, were selected from the whole dataset for trajectory inference using Slingshot (v2.6.0)32.” · Show in paper

“To model the activity of eRegulons along inferred trajectories, we fitted gene-based eRegulon AUC scores against pseudotimes by a generalized additive model (GAM) using tradeSeq (v1.12.0)33.” · Show in paper

FACS isolation of late second-trimester human glial progenitors (RG-tRG, RG-oRG, IPC-Glia/Tri-IPC, OPC) using EGFR, F3, CD38, PDGFRA, and ITGA2 antibodies, followed by in vitro population and single-cell clonal differentiation, scRNA-seq (10X Genomics 3' v3.1), immunostaining, and ex vivo slice transplantation.

Source excerpts

“Antibodies used for FACS include FITC anti-EGFR (Abcam, ab11400), PE anti-F3 (Biolegend, 365204), PerCP-Cy5.5 anti-CD38 (BD Biosciences, 551400), Alexa Fluor 647 anti-PDGFRA (BD Biosciences, 562798), and PE-Cy7 anti-ITGA2 (Biolegend, 359314).” · Show in paper

“For clonal analysis, samples for FACS were processed as above with the following changes: individual tRG, oRG, or Tri-IPC was sorted using a BigFoot Spectral Cell Sorter (Thermo Fisher) via single-cell precision mode into a single well of 96-well glass-bottom plates pre-coated with polyethylenimine and laminin containing 100 μL complete culture medium.” · Show in paper

Single-cell GWAS trait-relevance mapping integrating snATAC-seq with summary statistics for four cognitive traits, five psychiatric disorders, and Alzheimer's disease using SCAVENGE (v1.0.2) and gchromVAR.

Source excerpts

“We implemented SCAVENGE (v1.0.2)51 to integrate the snATAC-seq part of the snMultiome data with GWAS data of four cognitive traits (fluid intelligence, processing speed, executive function, and working memory) and five neuropsychiatric disorders (autism spectrum disorder [ASD], major depressive disorder [MDD], bipolar disorder [BPD], attention-deficit/hyperactivity disorder [ADHD], and schizophrenia [SCZ]).” · Show in paper