Cortex single-cell transcriptome

In our study, we aimed to comprehend the molecular features of mouse fetal brain development by analyzing single cell data from E10.5 to P4 mouse brain. The data was obtained from a publicly available database and underwent thorough quality control. Main cell types were annotated (Fig. 1A; Additional file 1: Table S1) and their correlations were analyzed (Fig. 1B), with neural precursor cells showing strong correlation and projective neurons displaying strong correlation as well. The composition of cells at different time points was found to be consistent with the stage of development (Additional file 2: Fig. S1A). Additionally, we confirmed the quality of the remaining cells by conducting quality control (Additional file 2: Fig. S1B). The annotation of cell types was validated by examining specific molecular signatures in each cluster from the differentially expressed genes (DEGs) with a log fold change greater than 0.5 (Fig. 1D).

Fig. 1

Cellular and molecular characteristics of the mouse cerebral cortex. A Visualization of mouse cortex cell clustering using UMAP. Cells are colored by cell-type assignment. B Different cell clusters correlation with the dendrogram. C UMAP visualizations colored by the expression of known marker genes for the selective cell population. Each dot represents one cell. D Heat map of top 10 DEGs between cell type clusters.

By performing Uniform Manifold Approximation and Projection (UMAP) analysis, we identified the cell types as apical progenitors (AP; Sox2, Pax6, Hes5) and intermediate progenitors (IP; Eomes, Neurog2, Btg2). During the neuron genesis stage, we detected migrating neurons (Nrp1) (Fig. 1C) and different subtypes of excitatory neurons, including up-layer neurons (Satb2) and deep-layer neurons (Bcl11b) using known markers. These subtypes included layer 2&3 neurons, layer 4 neurons, layer 5&6 neurons, corticothalamic and subcerebral projecting neurons (CThPN and SCPN), layer 6b neurons, and near-projecting Tshz2-positive neurons (Fig. 1C). Ventrally generated inhibitory interneurons (Dlx2) were also identified (Additional file 2: Fig. S1C).

During the glial cell genesis stage, we detected oligodendrocyte precursor cells (OPC; Olig2) and astrocytes (Apoe). Additionally, we identified microglial cells (Aif1), red blood cells (Car2), endothelial cells (Cldn5), pericytes (Cspg4), and vascular and leptomeningeal cells (Col1a1) (Additional file 2: Fig. S1C).

Developmental lineage of neural stem cells

Studies have shown that neurons, oligodendrocytes, and astrocytes are derived from distinct subtypes of neural stem cells. In this study, we aimed to characterize single cells grouped in the apical progenitor clusters of the mouse fetal cortex. Through t-distributed stochastic neighbor embedding (t-SNE) analysis, we identified nine distinct subclusters (Fig. 2A; Additional file 3: Table S2) and identified the correlations between these subtypes (Additional file 2: Fig. S2B). We also found that Cluster 2 consisted almost entirely of cells at the E12.5 stage (Additional file 2: Fig. S2A). Cluster 7 expressed high levels of Scl1a3, suggesting a potential role in glial cell development, while Cluster 8 expressed high levels of neurogenic-related genes Bcl11b and Neurog2, suggesting a correlation with neural development (Additional file 2: Fig. S2C and D).

Fig. 2

Lineage of neural precursor cells. A Clustering of apical progenitors from E10.5 to E17.5 visualized by t-SNE embedding. B Seven pseudotime state of apical progenitors by Monocle2. C The distribution of different cell clusters in seven pseudotime state. D Expression levels of apical progenitor genes along with pseudotime. The red and blue lines represent Lineage 2- State 4 and Lineage 2- State 6 and 7 cells. E Volcano plot of differential expressions of genes from Lineage 2- State 4 and Lineage 2- State 6 and 7 cells. . F The GSEA of GO biological processes ranked genes from high to low by values of log2 fold change. Adjusted P < 0.05

To understand the developmental trajectory of these nine subclusters, we divided them into seven states (Fig. 2B; Additional file 4: Table S3) and reconstructed a lineage tree using Monocle 2 (Fig. 2C). We found that subclusters 2 and 6 were at the top of the tree, and Cluster 4 was mainly spread out in the inflection point in a genealogical shift. These results suggest that Clusters 2 and 4 represent an earlier developmental status and the pre-transition state of different subtypes, which correlates with the main component of E12.5 cells in Cluster 2 and E12.5, E13.5, E15.5 cells' uniform distribution in Cluster 4, 5 and 6 (Additional file 2: Fig. S2A). Furthermore, we found that Clusters 7 and 8 were almost exclusively in State 4, so we analyzed the different gene expression patterns between Lineage 2- State 4 and Lineage 2- State 6 and 7. We observed increased expression of Bach2 in Lineage 2- State 4, and Prc1, Sox2, Top2a, Tubb4b, and Ube2c showed increased expression in Lineage 2- State 6 and 7 (Fig. 2D).

To further understand the discrepancy between Lineage 2-State 4 and Lineage2-State 6 and 7, we performed differential analysis and identified specific expression genes between the two cell lineages (adjusted P < 0.05 and log fold change > 0.25) (Fig. 2E). We then conducted gene set enrichment analysis (GSEA) of these genes and ranked them from high to low using values of log2 fold change. We found that Lineage 2-State 4 specifically expressed genes that were functionally related to neuron differentiation, while Lineage 2-State 6 and 7 specifically expressed genes that were functionally related to the regulation of cell migration (Fig. 2F). Our results suggest that in the anaphase of neural stem cell proliferation (Additional file 2: Fig. S2E), a precursor cell produces two seed generations, one that tends to produce neurons and the other that tends to migrate.

In conclusion, our study provides a comprehensive characterization of single cells grouped in the apical progenitor clusters of the mouse fetal cortex. We identified distinct subclusters, developmental trajectories, and specific expression genes that suggest a potential role in neuron differentiation and cell migration. Our results contribute to the understanding of neural stem cell proliferation and differentiation and may have implications for the development of new therapeutic approaches for neurological disorders.

BCAT1 is mainly expressed in layer II/III and IV neurons

In the field of neuron development, the proximity between different neuron subtypes is often visualized in a three-dimensional scatter plot (Fig. 3A), which displays the cells' positions in the MLLE space, fitted with a principal graph using STREAM [18]. In order to more intuitively and conveniently represent the trajectories of these cells in a two-dimensional plane, a Flat tree plot was constructed, preserving the lengths of tree branches in the MLLE space (Fig. 3B).

Fig. 3

Expression of Bcat1 in the developmental trajectory of neurons. A Visualization of different cell types using Stream in 3D. B Flat tree plot constructed from apical progenitors, intermediate progenitors and excitatory neurons. Labeled by cell type, branch id and pseudotime. C Stream plot of Bcat1 gene expression. D Stream plot, colored by cluster labels inferred. E Bcat1 co-stained with the NSCs markers of NESTIN. NSCs were isolated from the E12.5 mice brains and cultured in the proliferative medium for 24 h. Scale bar represents 50 μm

Our analysis revealed that apical progenitor cells and intermediate progenitor cells predominantly appear on the S4-S3 branch, while corticothalamic, subcerebral PN (CThPN and SCPN) and near-projecting neurons are primarily located on the S3-S5 branch. Layer II/III and IV neurons are primarily found on the S0-S1 branch, and layer 5 and 6 neurons are primarily located on the S0-S2 branch. Further, we identified differential genes in these different branches (Additional file 2: Fig. S3A-F).

One gene of interest that has been extensively studied in the context of cancer, BCAT1, is also involved in glutamate metabolism in excitatory neurons in the cerebral cortex. Based on our analysis, we found that BCAT1 is highly expressed in layer II/III and IV neurons, as well as in partially migrating neurons. However, it is less expressed in deep neurons (Fig. 3C and D). These results are consistent with our findings, as we observed that BCAT1 is co-expressed with the neural stem cell marker Nestin (Fig. 3E).

BCAT1 regulates neural progenitor cell proliferation

To investigate the function of the Bcat1 gene in neural stem cells (NSCs), we utilized short hairpin RNA (shRNA) to knockdown BCAT1 expression in neural progenitor cells of the developing cortex in E13.5 mouse embryos via uterine electroporation (IUE). We collected the brains at E18.5 for subsequent phenotypic analysis. The efficiency of BCAT1 knockdown was confirmed by assessing BCAT1 expression, which showed a significant decrease with the introduction of BCAT1 shRNA (Fig. 4A and B). Furthermore, we observed a decrease in the expression of the neural stem cell marker Pax6 and the proliferation marker Ki67 in vitro after BCAT1 knockdown, indicating a potential role of BCAT1 in NSCs proliferation and differentiation (Additional file 2: Fig. S4A and B).

Fig. 4

Bcat1 regulates the proliferation of the NSCs. A, B The percentage of Bcat1+GFP+ cells are decreased in Bcat1 knockdown brains. The mouse was electroporated at E13.5 and killed at E18.5. The bar graph shows the percentage of Bcat1+GFP+ cells relative to the control Bcat1+GFP+ cells (n = 3 independent experiments; ***P < 0.001; bars represent mean ± S.E.M). The scale bar represents 50 μm. C, D The GFP+ cells are decreased in Bcat1 knockdown brains. The mouse brain was electroporated at E13.5 and BrdU (50 mg/kg) was injected 24 h prior to killing at E15.5. (n = 3 independent experiments; **P < 0.01; bars represent mean ± S.E.M). The scale bar represents 50 μm. E–G The percentage of GFP+ BrdU+ cells and GFP+ Ki67+ cells are decreased in Bcat1 knockdown brains. The mouse brain was electroporated at E13.5 and BrdU (50 mg/kg) was injected 24 h prior to killing at E15.5. The bar graph shows the percentage of GFP+ BrdU+ cells and GFP+ Ki67+ cells relative to the total GFP-positive cells (n = 3 independent experiments; **P < 0.01; bars represent mean ± S.E.M). The scale bar represents 20 μm

In vivo, we found that knockdown of BCAT1 resulted in significant changes in cell distribution compared to the control group. The number of GFP positive cells in the proliferating ventricular/subventricular zone (VZ/SVZ) was significantly decreased (Fig. 4C and D). To determine whether this decrease in GFP-positive cells in VZ/SVZ was due to a decrease in neural precursor cell proliferation, we injected BrdU 24 h before brain collection at E15.5. We observed a 20% reduction in BrdU and GFP double positive cells in the BCAT1 knockdown group (Fig. 4E and F), suggesting a decrease in neural precursor cell proliferation. We also used immunostaining with the proliferative marker Ki67 to confirm this result. The proportion of double positive cells showed a significant decrease in the percentage of cells that were positive for Ki67 and labeled with GFP in the BCAT1 knockdown group (Fig. 4E and G). Finally, we measured the number of apoptotic cells and found that BCAT1 knockdown did not cause a significant increase in apoptosis (Additional file 2: Fig. S4C and D). Taken together, our findings suggest that BCAT1 is important for NSCs proliferation and may play a role in regulating the development of layer II/III and IV neurons in the cerebral cortex.

BCAT1 deletion leads to loss of neurons

To directly investigate whether BCAT1 knockdown affects the differentiation of II/III and IV layers of neurons, we collected brains at E16.5 after E13.5 electroporation of the knockdown plasmid performed immunofluorescent staining for electroperforated brain sections with the neuronal markers Satb2 and Ctip2. Our results showed that the proportion of GFP-Satb2 double-positive cells was significantly reduced in the BCAT1 knockout group (Fig. 5A and B). Similarly, we counted the number of Ctip2-labeled neurons and found that the proportion of Ctip2-GFP double-positive cells was also significantly reduced (Fig. 5A and C). These findings were consistent with the significantly reduced phenotype of GFP-positive cells (Fig. 5D and E).

Fig. 5

Bcat1 Knockdown causes the abnormal neuronal development. A–C The percentage of GFP+ Satb2+ cells and GFP+Ctip2+ cells are decreased in Bcat1 knockdown brains. The mouse was electroporated at E13.5 and killed at E16.5. The bar graph shows the percentage of GFP+ Satb2+ cells and GFP+ Ctip2+ cells relative to the control GFP+ Satb2+ cells and GFP+ Ctip2+ cells (n = 3 independent experiments; **P < 0.01; bars represent mean ± S.E.M). The scale bar represents 20 μm. D, E The GFP + cells are decreased in Bcat1 knockdown brains. The mouse brain was electroporated at E13.5 and killing at E16.5. (n = 3 independent experiments; **P < 0.01; bars represent mean ± S.E.M). The scale bar represents 50 μm. F–H The percentage of GFP+ Satb2+ cells and GFP+ Ctip2+ cells are decreased in Bcat1 knockdown brains. The mouse brain was electroporated at E13.5 and killing at E18.5. The bar graph shows the percentage of GFP+ Satb2+ cells and GFP+ Ctip2+ cells relative to the control GFP+ Satb2+ cells and GFP+ Ctip2+ cells (n = 3 independent experiments; **P < 0.01; bars represent mean ± S.E.M). The scale bar represents 20 μm. I, J The GFP+ cells are decreased in Bcat1 knockdown brains. The mouse brain was electroporated at E13.5 and killing at E18.5. (n = 3 independent experiments; **P < 0.01; bars represent mean ± S.E.M). The scale bar represents 50 μm

To further verify the impact of BCAT1 on the production of layers II/III and IV neurons, we collected brains at E18.5 after E13.5 electroporation of the knockdown plasmid. As expected, the reduction in layer II/III and IV neurons was more pronounced in the BCAT1 knockout mice than in the control group (Fig. 5F–H). Similarly, the number of GFP-positive cells was significantly reduced (Fig. 5I and J). We also tested whether BCAT1 knockdown impaired the migration of neurons by collecting brains at P0 after E15 electroporation of the knockdown plasmid. The results showed that the proportion of migration was not affected (Additional file 2: Fig. S4E and F). Additionally, we performed RT-qPCR to detect the expression of Cux1 and Dcx in cultured NPCs. Compared with the control group, BCAT1 knockdown significantly reduced the expression of Cux1 and Dcx (Additional file 2: Fig. S5A and B). These findings indicate that BCAT1 knockdown can significantly affect the development of layers II/III and IV neurons.