The histone demethylase Kdm6b regulates subtype diversification of mouse spinal motor neurons during development

How a single neuronal population diversifies into subtypes with distinct synaptic targets is a fundamental topic in neuroscience whose underlying mechanisms are unclear. Here, we show that the histone H3-lysine 27 demethylase Kdm6b regulates the diversification of motor neurons to distinct subtypes innervating different muscle targets during spinal cord development. In mouse embryonic motor neurons, Kdm6b promotes the medial motor column (MMC) and hypaxial motor column (HMC) fates while inhibiting the lateral motor column (LMC) and preganglionic motor column (PGC) identities. Our single-cell RNA-sequencing analyses reveal the heterogeneity of PGC, LMC, and MMC motor neurons. Further, our single-cell RNA-sequencing data, combined with mouse model studies, demonstrates that Kdm6b acquires cell fate specificity together with the transcription factor complex Isl1-Lhx3. Our study provides mechanistic insight into the gene regulatory network regulating neuronal cell-type diversification and defines a regulatory role of Kdm6b in the generation of motor neuron subtypes in the mouse spinal cord.


nature research | reporting summary
April 2020 Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design
All studies must disclose on these points even when the disclosure is negative. Unpaired two-tailed Student's t-test was used for statistical quantification. According to the standard and previous experience in the field, we picked up at least 3 mouse embryos for each genotype by testing genotypes from mice collected at indicated stages . Then at least 3 cryosections for each animal were chosen for immunostaining or in situ hybridization analyses, followed by quantification as needed. We used 9 datasets from each genotype for quantification, which is rational for statistical analysis. Increasing sample size didn't change the conclusion.
For scRNAseq analysis, sequencing data were filtered according to standard Seurat analysis criteria as follows. Raw sequencing cell barcodes were filtered to distinguish valid cell barcodes from empty cell barcodes using an algorithm in Cellranger count v2.1.1, which is an analysis pipeline for Chromium single cell 3' RNA-seq result with an expected recovered cell parameter; expect-cells set to 9,000 and 7,800 for control and Kdm6b-cKO, respectively. The gene-barcode count matrix generated was read using the Read10X function in Seurat v3.1.4. Initially, we kept all genes expressed in >3 cells and all cells with at least 200 detected genes. To exclude cells in poor quality and possible doublets, possible outliers of cells were further removed by visual inspection of the distribution of a total number of mapped genes. Cells with a high level of mitochondrial genes (>7%) were also removed, which were treated as mitochondrial cytoplasmic RNAs, which were released and sequenced due to cell lysis. Cells with a high level of hemoglobin gene expression (Hba-a1>10) were also removed, as they are likely to represent rare blood cell contaminants. For other analyses, no data were excluded.
All mouse analyses, including immunostaining and in situ hybridization assays and associated quantifications, were performed on at least three independent animals and at least three sections for each animal, with data from individual replicates and the mean presented. We repeated the animal experiments every year for three years and all results were reproducible. For scRNAseq, co-immunoprecipitation, chromatin immunoprecipitation, and luciferase experiments, at least one more independent experimental set was performed for each test, and all produced consistent results, indicating good reproducibility.
For the experiments involving cell culture work, independent experiments were randomized by choosing different days and transfecting the different cell batches for each set of experiments. Randomization is not relevant to our mouse analyses as there was no additional treatment step for the animals. Control and Kdm6b-cKO mice were identified by their genotypes and subsequently analyzed in parallel for each litter. We analyzed multiple litters for the presented data to avoid any data resulting from litter-to-litter variations.
Blinding was strictly used during data collection and analysis.