Cell-Cell Communication Alterations via Intercellular Signaling Pathways in Substantia Nigra of Parkinson's Disease

Abstract

Parkinson's disease (PD) is a neurodegenerative movement disorder characterized with dopaminergic neuron (DaN) loss within the substantia nigra (SN). Despite bulk studies focusing on intracellular mechanisms of PD inside DaNs, few studies have explored the pathogeneses outside DaNs, or between DaNs and other cells. Here, we set out to probe the implication of intercellular communication involving DaNs in the pathogeneses of PD at a systemic level with bioinformatics methods. We harvested three online published single-cell/single-nucleus transcriptomic sequencing (sc/snRNA-seq) datasets of human SN (GSE126838, GSE140231, and GSE157783) from the Gene Expression Omnibus (GEO) database, and integrated them with one of the latest integration algorithms called Harmony. We then applied CellChat, the latest cell-cell communication analytic algorithm, to our integrated dataset. We first found that the overall communication quantity was decreased while the overall communication strength was enhanced in PD sample compared with control sample. We then focused on the intercellular communication where DaNs are involved, and found that the communications between DaNs and other cell types via certain signaling pathways were selectively altered in PD, including some growth factors, neurotrophic factors, chemokines, etc. pathways. Our bioinformatics analysis showed that the alteration in intercellular communications involving DaNs might be a previously underestimated aspect of PD pathogeneses with novel translational potential.

Keywords: CellChat; Parkinson’s disease; cell–cell communication; dopaminergic neurons; sc/snRNA-seq; signaling.

FIGURE 1

Integration of online published single-cell/single-nucleus transcriptomic sequencing (sc/snRNA-seq) datasets of postmortem human substantia nigra (SN). (A) Schematic flowchart showing the processing, integration, and generation of an integrated dataset from online published sc/snRNA-seq datasets of human SN. (B) Violin plot of the expression of canonical marker genes in the identified cell populations. (C) Uniform manifold approximation and projection (UMAP) plot of the identified cell populations. (D) UMAP feature plots of the expression of some well-known marker genes in the cell clusters. (E) UMAP plot of the identified neuron subtypes after subclustering the NEU cluster in C. (F) UMAP feature plots of the expression of well-known marker genes in the identified neuron subpopulations. ODC, oligodendrocyte cluster; AST, astrocyte cluster; OPC, oligodendrocyte precursor cell cluster; MIG, microglia cluster; NEU, neuron cluster; ENT, endothelial cell cluster; UN, unidentified cell cluster; InN, inhibitory neuron cluster; ExN, excitatory neuron cluster; DaN, dopaminergic neuron cluster; UnN1, unidentified neuron cluster 1; UnN2, unidentified neuron cluster 2; UnN3, unidentified neuron cluster 3.

FIGURE 2

The integrated dataset reveals PD-associated cell composition alteration in PD. (A) UMAP plot of the distribution of control (Ctrl) and PD donors-derived cells. (B) UMAP plot of the distribution of control and PD donors-derived neurons.

FIGURE 3

Inference of cell–cell communications by CellChat shows global alterations in signaling pathways-mediated communications between DaN and non-neuronal cells in PD. (A) Schematic diagram of cell–cell communication between DaN and non-neuronal cells. Circle plots of the interaction quantity (B) and interaction strength (C) between DaN and non-neuronal cells. Blue lines indicate that the displayed communication is decreased in PD, whereas red lines indicate that the displayed communication is increased in PD compared with healthy control. (D) Heatmaps of the interaction quantity (left panel) and interaction strength (right panel) between DaN and non-neuronal cells. Blue color indicates that the displayed communication is decreased in PD, while red color indicates that the displayed communication is increased in PD compared with healthy control.

FIGURE 4

Cell–cell communications mediated by individual signaling pathways are altered in PD between DaN and non-neuronal cells. (A) Bar plots of the ranking of signaling axes by overall information flow differences in the interaction networks between control (Ctrl) and PD sample. The top signaling pathways with red-colored labels are more enriched in the control sample, the middle ones with black-colored labels are equally enriched in control and PD sample, and the bottom ones with green-colored labels are more enriched in the PD sample. (B) Bar plot of the ranking of signaling axes between control and PD sample by pairwise Euclidean distance. (C) Heatmaps of the overall (comprising both outgoing and incoming) signaling flows of each cell population mediated by individual signaling axes in control and PD sample.

FIGURE 5

Some cell–cell communications between DaN and non-neuronal cells mediated by signaling pathways are greatly altered in PD compared with control (Ctrl) sample. Circle plots show and compare cell–cell communication alterations between DaN and non-neuronal cells mediated by some of the signaling axes, including activins (A), epithelial growth factor (EGF) (B), neuregulins (NRGs) (C), Visfatin (D), glial cell line-derived neurotrophic factors (GDNF) (E), CX3C (F), and cholecystokinin (CCK) (G).