Single-cell transcriptomics delineates the immune cell landscape in equine lower airways and reveals upregulation of FKBP5 in horses with asthma

Abstract

Equine asthma (EA) is a heterogenous, complex disease, with a significant negative impact on horse welfare and performance. EA and human asthma share fundamental similarities, making EA a useful model for studying the disease. One relevant sample type for investigating chronic lung inflammation is bronchoalveolar lavage fluid (BALF), which provides a snapshot of the immune cells present in the alveolar space. To investigate the immune cell landscape of the respiratory tract in horses with mild-to-moderate equine asthma (mEA) and healthy controls, single-cell RNA sequencing was conducted on equine BALF cells. We characterized the major immune cell populations present in equine BALF, as well as subtypes thereof. Interestingly, the most significantly upregulated gene discovered in cases of mEA was FKBP5, a chaperone protein involved in regulating the activity of the glucocorticoid receptor.

Figure 1

Initial clustering of BALF cells (a) Graphical illustration of the study design. The figure was created using the BioRender software. (b) UMAP representation of 63,022 equine BALF cells colored by the six major cell groups: alveolar macrophages (28,277 cells), proliferating macrophages (3420 cells), T cells (28,112 cells), mast cells (1257 cells), neutrophils (797 cells) and dendritic cells (1339 cells). (c) Example of expression of specific markers for the different cell types. (d) Heat map demonstrating the difference in expression levels of the top biomarker genes for each of the six major clusters. (e) Proportion of cell-types in the EA and healthy control horses, by group.

Figure 2

Re-clustering of T cells. (a) The UMAP plot illustrates the four equine BALF T cell groups. The expression of three top differentially expressed markers is mapped onto the UMAP plot, which is shown to the right. (b) The proportion of T cell subtypes in the EA vs healthy group of horses is depicted. (c) The differential expression of T cell cluster markers visualized with a heat map. (d) Iterative clustering reveals seven subpopulations of CD4⁺ cells. The subset coloured in brown in the heat map (T-ISG^Hi) expresses interferon stimulated genes, and the yellow subset (T-S100^Hi) displays higher expression of S100 genes and other genes involved in cytoskeleton organization and cell adhesion, as indicated by GO analysis. The heat map shows the genes with the highest fold changes in the T-ISG^Hi and T-S100^Hi relative to the other clusters. The green label indicates data from the horses sampled at the clinic and the blue label indicates data from the research herd. (e) Iterative clustering of the CD8⁺_TR cluster shows eight subclusters. The brown subset (T-ISG^Hi) selectively expresses interferon stimulated genes. The heat map shows the genes with the highest fold changes in the T-ISG^Hi cluster relative to the other clusters. See Supplementary Fig. S3 and Supplementary Table S3 for the proportions of case and control cells within the clusters.

Figure 3

Re-clustering of alveolar macrophages. (a) UMAP illustration of eight alveolar macrophage (AM) subsets. The top differentially expressed markers for the AM4 cluster and AM3 cluster are glycoprotein Nmb (GPNMB) and γ-adductin (ADD3), respectively. (b) Bubble plot visualizing the expression levels for genes which were among the top differentially expressed between AM clusters. (c) Proportion of AM clusters in the EA and healthy groups. (d) Genes upregulated in cluster AM3 were significantly enriched in the GO pathways colored in light blue. Genes upregulated in cluster AM4 were significantly enriched in GO pathways colored in dark blue.

Figure 4

Re-clustering of mast cells and neutrophils. (a) UMAP illustration of four mast cell clusters and one cluster comprising club or goblet cells. (b) Bubble plot visualizing the expression levels of the top differentially expressed cluster marker genes. The cluster MC1 displays significantly higher expression of FKBP5, LTC4S and RGS1. (c) Proportions of mast cell subpopulations in EA and healthy horses: The cluster MC1 (yellow) comprises a significantly higher proportion of cells in horses with asthma (p = 0.001, Kruskal–Wallis). (d) Re-clustering of neutrophils results in three subpopulations as illustrated by the UMAP plot. (e) Bubble plot visualizing the expression levels of top differentially expressed neutrophil cluster markers and the FTH1 gene.

Figure 5

Re-clustering of dendritic cells. (a) UMAP illustration of the dendritic cell clusters annotated as conventional DC1 and DC2s, migratory DCs and monocyte derived DCs. (b) Bubble plot visualizing the expression levels of the cluster marker genes. ENSECAG00000024882 = unannotated CCL15 orthologue. DC = dendritic cells, n.d. = not annotated.

Figure 6

Differential gene expression EA and control horses. (a) Volcano plots showing DE genes from pseudo-bulk (DESeq2) analysis of alveolar macrophages (AMs), T cells, and mast cells. The plots show the results from testing the mastocytic group of horses vs controls. (b) FKBP5 and CCL24 expression levels in EA and control cells visualized on the UMAP plot. (c) Violin plots showing expression levels of FKBP5 in mast cells and CCL24 in alveolar macrophages for individual horses. IDs of healthy horses are denoted in blue and asthma horses denoted in green. (IDs of the horses that did not meet the inclusion critera are marked in red).