henhaobio
Sign in

DevelopmentDevelopmental BiologyPeer-reviewed journal article

Single cell transcriptome profiling of the human developing spinal cord reveals a conserved genetic programme with human specific features

Teresa RayóniD, Rory J. MaizelsiD, Christopher BarringtoniD, James BriscoeiD

Abstract

The spinal cord receives input from peripheral sensory neurons and controls motor output by regulating muscle innervating motor neurons. These functions are carried out by neural circuits comprising molecularly and physiologically distinct neuronal subtypes that are generated in a characteristic spatial-temporal arrangement from progenitors in the embryonic neural tube. The systematic mapping of gene expression in mouse embryos has provided insight into the diversity and complexity of cells in the neural tube. For human embryos, however, less information has been available. To address this, we used single cell mRNA sequencing to profile cervical and thoracic regions in four human embryos of Carnegie Stages (CS) CS12, CS14, CS17 and CS19 from Gestational Weeks (W) 4-7. In total we recovered the transcriptomes of 71,219 cells. Analysis of progenitor and neuronal populations from the neural tube, as well as cells of the peripheral nervous system, in dorsal root ganglia adjacent to the neural tube, identified dozens of distinct cell types and facilitated the reconstruction of the differentiation pathways of specific neuronal subtypes. Comparison with existing mouse datasets revealed the overall similarity of mouse and human neural tube development while highlighting specific features that differed between species. These data provide a catalogue of gene expression and cell type identity in the developing neural tube that will support future studies of sensory and motor control systems and can be explored at https://shiny.crick.ac.uk/scviewer/neuraltube/.

In-depth summary

Claude Opus 5.5 · xhigh effort · 7 Nov 2026 · from the full text of version 1 (12 Apr 2021) · CC BY 4.0

Human embryonic spinal cord scRNA-seq (71,219 cells, CS12-CS19) mirrors mouse patterning and neurogenesis but shows floor-plate PAX7, broad ventral NKX6.2 and early OLIG2/NKX2.2 progenitors.

Single-cell profiling has catalogued mouse neural tube and neural crest cells, but few such data exist for human embryos, whose spinal cord and sensory ganglia remain thinly characterised, leaving open how closely human and mouse cell types match molecularly. This work provides a human atlas spanning gestational weeks 4-7 and compares it directly with mouse data aligned by developmental stage.

Key results

  1. ShownKnowledge-matrix classification of 43,485 neural cells recovered the expected dorsoventral progenitor and neuronal domains, and mouse and human cells of the same type correlated best on transcription factor expression.

    Cells were assigned via a binarized marker knowledge matrix (Antler) after unsupervised clustering failed. On 1:1 orthologous transcription factors, Pearson correlation peaked for matched mouse/human types; progenitors resembled adjacent domains, neurons grouped by excitatory/inhibitory class. Human ventral pseudo-bulk CS12 best matched in vitro motor neuron differentiation Day 5, CS14 Day 9 (Spearman). · Identification of cell types in developing human…; Classification of central nervous cells; Mouse and human correlation analysis; Comparison to bulk in vitro RNA-seq data

  2. ShownHuman embryos differ from mouse in several patterning genes: PAX7 is expressed in floor plate, NKX6.2 spans most ventral progenitor domains rather than p1 alone, and PHOX2B+ visceral motor neurons express little TBX20.

    Human PAX7 was detected in floor plate at CS12, CS14 and CS17 but at no mouse timepoint; CDH7, TXLNB, CDHR3, PIFO and RRAD correlate with PAX7 in human floor plate and are not expressed in mouse floor plate. Human NKX6.2 was found in p1, p2, pMN, p3 and floor plate, declining over time in pMN and p3 (RNAscope on CS13/CS15 sections). Human PHOX2B+ neurons (n=170) express TBX2 and TBX3 but little TBX20. · Specific features of gene expression in the…; Correlation analysis for PAX7; RNAScope FISH with immunohistochemistry; Figure legends

  3. SuggestedSpinal neurogenesis follows a similar overall course and temporal transcription factor code in human and mouse, but human begins with relatively more pMN and p3 progenitors and keeps dorsal neurogenesis going longer.

    After stages were aligned by developmental events, the ratio of p3 and pMN to pI and pD progenitors was higher in human than mouse at CS12 and CS14 and comparable by CS17; dorsal neurons outnumbered ventral by CS19. In human, Onecut factors marked CS12 neurons, POU2F2 and ZFHX2-4 appeared by CS14, NEUROD2/6 and NFIA/B only by CS19 (NFI in MNs from CS14); late markers were sparse in dI3-dI6. · Dynamics of neurogenesis in the human spinal…; A conserved temporal code for the specification…

  4. SuggestedOLIG2/NKX2.2 double-positive progenitors are more common in early human than early mouse neural tube, look like a mix of pMN and p3 identities, and carry gliogenic markers, consistent with early oligodendrocyte precursors.

    In mouse, double positives were under 5% of Sox2+ cells expressing Olig2 or Nkx2.2 until E11.5 (rising to ~15% at E12.5), vs ~10% in human at equivalent stages, where the fraction fell at CS17 and CS19. Human double positives were closer to pMN than p3 at CS12-CS14, with p3 similarity rising later, and kept high FABP7, OLIG1 and OLIG2 and a higher gliogenic module score than pMN or p3. · Co-expression of Olig2 and Nkx2.2 in the…; Cell-cycle and gene module scoring

  5. SuggestedRNA velocity and CellRank reconstruct a single convergent progenitor-to-neuron path for seven spinal domains in both species, with broadly similar ordering of gene expression in latent time.

    scVelo and CellRank were run on mouse E9.5-E13.5 and human CS12, CS14 and CS19 (CS17 dropped for low counts per gene) for p3, pMN, p2, p1, dI4, dI3 and dI2. In the pMN-to-MN path of both species Olig2 came first in latent time, then Neurod4, then Elavl3 and Mnx1; latent-time heatmaps for V3, V2, V1, dI4, dI3 and dI2 were broadly alike. Human trajectories were less well resolved. · Transcriptional dynamics during neurogenesis in…; RNA Velocity; CellRank Trajectory Analysis

Methods, data and reproducibility

10x Chromium 3' scRNA-seq of microdissected cervical-thoracic trunk from four human embryos (CS12, CS14, CS17, CS19). Neural cells were integrated in Seurat, assigned to spinal cord and dorsal root ganglion subtypes with marker knowledge matrices (Antler), and compared with mouse neural tube data via orthologue correlation, cell-type proportions, RNA velocity and immunostaining/RNAscope.

  • scRNA-seq (10x Chromium 3' v3.1) of microdissected cervical-thoracic trunk at CS12, CS14, CS17 and CS19 (gestational weeks 4-7), tissue from the Human Developmental Biology Resource · human · n = 4 embryos (2 male, 2 female by SRY); 71,219 cells after QC, 43,485 neural · GEO GSE171892
  • Previously generated mouse neural tube and dorsal root ganglia scRNA-seq (E8.5 to E13.5; E9.5-E13.5 used for RNA velocity) · mouse
  • Published daily bulk RNA-seq time course of in vitro motor neuron differentiation from human embryonic stem cells · human · n = 1448 dynamic genes used for comparison
  • Published list of primate-specific protein-coding genes enriched in human cerebral neural precursors · human · n = 51 genes
  • Transverse cryosections of neural tube for immunostaining and RNAscope FISH (human CS13, CS15; mouse E9.5, E10.5) · human; mouse
Methods & tools
10x Genomics Chromium Single Cell 3' v3.1, Illumina HiSeq4000, Cell Ranger 3.1.0, Seurat 3.2.2, Antler, biomaRt (Ensembl release 93), Velocyto, scVelo, CellRank, ScanPy, RNAscope multiplex fluorescent v2, Leica SP8 confocal with ImageJ/Fiji
Reagents & environment
Cell Ranger 3.1.0 with GRCh38-3.0.0 (human) and mm10-3.0.0 (mouse) references; Seurat 3.2.2; Ensembl biomaRt release 93.
Reproducibility gaps
Versions of scVelo, CellRank, ScanPy, Velocyto and Antler are not stated and no container or requirements file is described; per-replicate QC thresholds and marker knowledge matrices are in supplementary tables (S1-S3), not in the text.
Limitations (6)
  • The CS17 sample was processed after a 24 h delay, showed signs of more ambient RNA and lower counts per gene, and was left out of RNA velocity analysis because it produced normalisation artefacts. · stated by the authors
  • Only four stages over roughly two weeks; the authors note reduced temporal resolution and less well-resolved human trajectories than mouse. · stated by the authors
  • Whether PAX7 regulates the floor-plate genes it correlates with, and the lineage and neurogenic or gliogenic potential of OLIG2/NKX2.2 double-positive cells, remain untested and need functional and lineage experiments. · stated by the authors
  • Agreement between in vivo stages and in vitro motor neuron differentiation weakens at CS17 and CS19, which the authors partly attribute to later in vitro cultures becoming more asynchronous. · stated by the authors
  • Each stage is a single embryo (CS17 and CS19 replicates come from the same embryo), so differences in proportions between stages and species rest on one sample per timepoint. · noted in this summary
  • Cell identities are assigned from knowledge matrices adapted from mouse markers, which favours known conserved types and could miss or misassign human-specific populations. · noted in this summary

Read this if you study spinal cord or dorsal root ganglion development, need a human in vivo reference for stem-cell-derived spinal neurons, or compare mouse and human neurogenesis timing.

Citations
143
FWCI
6.98
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Similarity Map
SelectedSame fieldCross-field
THISDelile '19Andersen '21Zhang '21Zhang '23Li '23Zhou '20
80% semantic matchDevelopmental Biology
Julien Delile, Teresa Rayón, Manuela Melchionda+3 more·Development·Mar 7, 2019· 356 citations