Single Cell Transcriptomic and Chromatin Profiles Suggest Layer Vb Is the Only Layer With Shared Excitatory Cell Types in the Medial and Lateral Entorhinal Cortex

Abstract

All brain functionality arises from the activity in neural circuits in different anatomical regions. These regions contain different circuits comprising unique cell types. An integral part to understanding neural circuits is a full census of the constituent parts, i.e., the neural cell types. This census can be based on different characteristics. Previously combinations of morphology and physiology, gene expression, and chromatin accessibility have been used in various cortical and subcortical regions. This has given an extensive yet incomplete overview of neural cell types. However, these techniques have not been applied to all brain regions. Here we apply single cell analysis of accessible chromatin on two similar but different cortical regions, the medial and the lateral entorhinal cortices. Even though these two regions are anatomically similar, their intrinsic and extrinsic connectivity are different. In 4,136 cells we identify 20 different clusters representing different cell types. As expected, excitatory cells show regionally specific clusters, whereas inhibitory neurons are shared between regions. We find that several deep layer excitatory neuronal cell types as defined by chromatin profile are also shared between the two different regions. Integration with a larger scRNA-seq dataset maintains this shared characteristic for cells in Layer Vb. Interestingly, this layer contains three clusters, two specific to either subregion and one shared between the two. These clusters can be putatively associated with particular functional and anatomical cell types found in this layer. This information is a step forwards into elucidating the cell types within the entorhinal circuit and by extension its functional underpinnings.

Keywords: ATAC-seq; cell types; chromatin; consolidation; enhancers; entorhinal cortex; memory; transcriptomics.

Figure 1

Clustering of all single cells reveals shared and sample-enriched clusters of cells. (A) Left: schematic reference of the tissue, scale bar is 1 mm, M indicates the MEC. Right: Example of microdissection of subregions of the EC and approximate level in an anatomical atlas. This section was cut horizontally to optimally allow dissection of both subregions of the EC. (B) Schematic reference and example of microdissection of the lateral subregion of the EC. L indicates the LEC. (C) UMAP projection of all cells profiled with scATAC sequencing. Coloring and labeling indicate cluster numbers of scATAC data used in the rest of this publication. The outlines indicate three different classes of cells: non-neurons (yellow), inhibitory neurons (magenta), and excitatory neurons (cyan). (D) UMAP projection labeled by sample. Note the existence of both mixed clusters and clusters consisting primarily of either MEC or LEC cells. (E) Hierarchical dendrogram of the clusters of scATAC cells with quantification of the contribution of each sample to each cluster. Contributions are given in percentages. Cluster numbers, cell class, and colors correspond with those in (C).

Figure 2

Identification of interneuron cell types. (A) There are four different clusters containing inhibitory neurons. Different marker genes can identify the identity of each of these. Here the relative accessibility score of Lhx6 labels the MGE derived interneurons. (B) Accessibility of parvalbumin labels parvalbumin positive interneurons (cluster 20). (C) Accessibility of somatostatin labels somatostatin positive interneurons (cluster 5). (D) Accessibility of vasoactive intestinal peptide (VIP), calbindin, and 5-Hydroxytryptamine Receptor 3A (Htr3a) labels the VIP+ interneurons (cluster 4). (E) Accessibility of neuropeptide Y (NPY), Reelin, and proenkephalin (Penk) labels the NPY+ interneurons (cluster 13).

Figure 3

Identification of superficial layer excitatory neurons and enriched motifs. (A) Nissl stained and reference images of a mouse brain. Used to identify gene expression in the two different subregions of the EC. Left panels are coronal (3.1 mm posterior to bregma) right panels are sagittal (3.3 mm lateral from the midline). Image taken from Lein et al. (2007). (B) Superficial neurons of the entorhinal cortex can be grouped into three broad categories: the reelin and calbindin expressing LII cells and the layer III cells. Here the Reelin expressing LII cells (clusters 7 and 10 for MEC and LEC respectively) are labeled by accessibility of the Reelin gene. Note that this gene is also expressed in various inhibitory neuron cell types. (C) Accessibility of the Calbindin gene labels the Calbindin positive cell types (clusters 3 and 15, for MEC and LEC respectively). Similar to Reelin, Calbindin is also expressed in various inhibitory cell types. (D) Accessibility of CD44 labels MEC LIII cells (cluster 8). (E) Accessibility of transcriptional repressor GATA binding 1 (Trps1) but absence of accessibility of CD44 labels LEC LIII neurons (cluster 19). (F) Detection motifs enriched in excitatory, superficial cells of either the LEC or the MEC. The left sides show the enriched motifs, while the name behind denotes the corresponding best matching transcription factor.

Figure 4

Identification of deep layer excitatory neuron clusters and enriched motifs. (A) Identification of Layer Va (LEC and MEC, cluster 2) cells by accessibility of Etv1. (B) Identification of Layer VI (LEC and MEC, cluster 16) cells by accessibility of Nxph4. (C) Identification of Layer Vb (LEC, MEC shared, clusters 1 and 14) cells by accessibility of Ths7b. (D) Identification of Layer Vb (LEC, MEC shared, clusters 1 and 14) cells by accessibility of Ptpru. (E) Identification of Layer Vb (MEC only, cluster 1) cells by accessibility of Col5a1. (F) Identification of Layer Vb (LEC only, cluster 17) cells by accessibility of Tpbg. (G) Identification of Layer Vb (LEC only, cluster 17) cells by accessibility of Ilrapl2. (H) Enriched motifs in the deep layers of the EC and their corresponding best matching transcription factor. Here selected DARs of all excitatory deep layer neurons were contrasted with selected DARs of all excitatory superficial layer neurons. (I) Enriched motifs in clusters 14 and 16. Here selected DARs of cluster 16 (LVI) were with all other non-overlapping DARs, and the same was done for cluster 14 (LVb). Note that the Tbr1 motif is the same as the one in (H), but in reverse complement. (J) Enriched motifs in clusters 1, 2, and 17. The contrasts are similar to the ones in (I), contrasting selected DARs of one cluster with all other, non-overlapping, DARs.

Figure 5

Comparison of single cell transcriptomic and chromatin data. (A) UMAP projection of a selection of the cells published by Yao et al. (2021). We identified 36 different clusters of cell types in this subset of data. The outlines indicate three different classes of cells: non-neurons (yellow), inhibitory neurons (magenta), and excitatory neurons (cyan). (B) Hierarchical dendrogram of the clusters of scRNA-seq cells. Cluster numbers, cell class, and colors correspond with those in (A). LVb clusters are indicated with an asterisk. (C) Projection of most similar transcriptomic cell types onto the scATAC data. Each transcriptomic cell is compared to each single cell ATAC cell and the connected to the most similar one. (D) Heatmap of mapping scATAC-seq on scRNA-seq cells. Each cell in the heatmap indicates the percentage of scATAC cells mapping to scRNAseq clusters. The scale bar indicates percentages.