γδ T cells control murine skin inflammation and subcutaneous adipose wasting during chronic Trypanosoma brucei infection

Abstract

African trypanosomes colonise the skin to ensure parasite transmission. However, how the skin responds to trypanosome infection remains unresolved. Here, we investigate the local immune response of the skin in a murine model of infection using spatial and single cell transcriptomics. We detect expansion of dermal IL-17A-producing Vγ6⁺ cells during infection, which occurs in the subcutaneous adipose tissue. In silico cell-cell communication analysis suggests that subcutaneous interstitial preadipocytes trigger T cell activation via Cd40 and Tnfsf18 signalling, amongst others. In vivo, we observe that female mice deficient for IL-17A-producing Vγ6⁺ cells show extensive inflammation and limit subcutaneous adipose tissue wasting, independently of parasite burden. Based on these observations, we propose that subcutaneous adipocytes and Vγ6⁺ cells act in concert to limit skin inflammation and adipose tissue wasting. These studies provide new insights into the role of γδ T cell and subcutaneous adipocytes as homeostatic regulators of skin immunity during chronic infection.

Fig. 1. Integrative overview of the murine skin infected with T. brucei infection using single cell transcriptomics.

A Representative H&E-stained images from naïve and infected mice from three independent experiments. Scale bar: 800 μm (top) and 100 μm (bottom). B Immunohistochemistry against the stumpy-specific marker PAD1 in naïve and infected samples, including insets highlighting the presence of stumpy forms in both the epidermis-subcutaneous adipose tissue and the subcutaneous adipose tissue-adipose tissue. Scale bar: 50 μm. C Uniform manifold approximation and projection (UMAP) of 56,876 high-quality cells from both naïve (left panel) and infected (right panel) samples, highlighting the major cell types detected in our dataset, including stromal cells (Keratinocytes, fibroblasts, endothelial cells, and adipocytes) and immune cells (myeloid cells, T cells, Langerhans cells, and macrophages). D Frequency plot of the major cell types identified in the murine skin over the course of infection across biological replicates (n = 2 replicates/experimental condition). E Top 10 marker genes defining all the major cell clusters detected in (C). The heatmap is colour-coded based on gene expression intensity. F Left panel: schematic representation of the murine skin highlighting the three main skin layers: epidermis, dermis, and hypodermis (encompassing the adipose tissue and muscle). Created with Biorender (Agreement number AT25PETUME). Right panel: Tissue sections and spatial UMAP plot of the samples included in the spatial transcriptomics analysis. The major skin compartments are highlighted and include the epidermis (Ep), dermis (D), adipose tissue (Ad), muscle (M), and subcutis (Sc). Scale bar, 100 μm. G Integration of single cell and spatial transcriptomics datasets for the major cell types typically found in the skin, including keratinocytes, Langerhans cells, fibroblasts, melanocytes, adipocytes, and myeloid cells using the Seurat package. Scale bar, 100 μm. keratinocytes (KC), fibroblasts (FB), endothelial cells (EC), Macrophages (MC), Langerhans cells (LC), adipocytes (Adipo), erythrocytes (Erythro).

Fig. 2. The murine skin stromal cells respond to infection by upregulating genes associated with antigen presentation and chemotaxis.

A Uniform manifold approximation and projection (UMAP) of 52,149 high-quality cells within the stromal subcluster, encompassing seven keratinocyte clusters (KC1-7), fibroblasts (FB), endothelial cells (ECs), melanocytes (MLCs), and interstitial preadipocytes (IPAs). B Dot plot representing the expression levels of top marker genes used to catalogue the diversity of skin stromal cells. The size of the dots represents the percentage of cells that express a given marker, and the colour intensity represents the level of expression. C Left panel: Frequency plot the different stromal cell types detected in the murine skin in naïve (n = 2 mice from the single cell experiment performed once) and infected (n = 2 mice from the single cell experiment performed once) samples. Module scoring for the overall expression of inflammatory cytokines. Right panel: Frequency of interstitial preadipocytes 1 (top) and 2 (bottom) in naïve and infected samples from the scRNAseq dataset. The data is presented as mean values +/− SD. Source data are provided as a Source Data file. Module scoring for Inflammatory cytokines (D) and genes associated with antigen presentation (E). Violin plot showing the expression level of several significant adipocyte-specific cytokines (F), chemokines (G), and major histocompatibility complex (MHC) I and II molecules (H). I Spatial module scoring for inflammation and antigen presentation in a naïve (top) and infected (bottom) skin section. Scale bar, 100 μm.

Fig. 3. The murine skin is colonised by a myriad of myeloid cells during chronic T. brucei infection.

A Uniform manifold approximation and projection (UMAP) of 2353 high-quality cells within the myeloid cluster were re-analysed to identify a total of four subclusters, including dendritic cells (DCs), mast cells, and two populations of macrophages (Mrc1⁺ macs and Cd14⁺ mono), and two populations of Langerhans cells (LCs 1 and LCs 2). B Dot plot representing the expression levels of top marker genes used to catalogue the diversity of myeloid cells. C Integration of single cell and spatial transcriptomics datasets for the myeloid cell identified during infection, including Cd14⁺ monocytes, Mrc1⁺ macrophages, Cd207⁺ Langerhans cells, mast cells, and conventional dendritic cells (cDCs). Scale bar, 100 µm. The corresponding histological section is included on the left, including an annotation of epidermis (Ep), dermis (D), adipose tissue (Ad), muscle (M), and subcutaneous adipose tissue (Sc). D Representative smFISH targeting Cd207 (orange), Cd4 (yellow), and Krt14 (purple) around the perifollicular space (denoted as A1; The hair follicles are highlighted with dotted lines) and in the adipose tissue in the hypodermis (denoted as A2) in skin biopsies from independent naïve and infected animals. Scale bar, 50 μm. The results presented here are representative from two independent experiments. The images have been adjusted for brightness and contrast to increase resolution. Please note that the sections contain hair with high autofluorescence. Module scoring for the overall expression of inflammatory cytokines (E) and genes associated with antigen presentation (F).

Fig. 4. Chronic T. brucei infection triggers the activation of dermal Vγ6⁺ cells.

A Uniform manifold approximation and projection (UMAP) of 1043 T cells from naïve and infected skin samples. We detected a total of five subclusters including type 2 innate lymphoid cells (ILC2s), Cd4⁺ T cells, natural killer (NK) cells, and γδT cells. B Dot Plot depicting the expression level of top marker genes for the skin T cell subcluters. C Left: Flow cytometry of CD27⁻ and CD27⁺ γδT cells in murine skin infected (n = 5 mice examined from two independent experiments) with T. brucei and naïve controls (n = 4 mice examined over two independent experiments). Right: Quantification of flow cytometry data. Statistical analysis was conducted using a parametric two-sided T test. A p value < 0.05 was considered significant. The data is presented as mean values +/− SD. Source data are provided as a Source Data file. D Dot Plot depicting the expression level of top genes associated with T cell activation and TCR engagement for the skin T cell subcluters. E Spatial module scoring for γδ T cell in naïve and infected skin biopsies used for spatial transcriptomics. Scale bar, 100 μm. F Representative smFISH (two independent experiments) targeting Trdc (yellow) and Vγ6 (orange) around the perifollicular space (denoted as A1) and in the adipose tissue in the hypodermis (denoted as A2) in skin biopsies from naïve and infected animals. The area corresponding to the hair follicle is denoted with dotted lines in A1. Please note that the sections contain hair follicles with high autofluorescence. Scale bar, 50 μm. G In silico cell-cell interaction analysis between adipocytes (“senders”) and T cells (“receivers”). The heatmap is colour coded to represent the strength of the interaction. H Spatial feature plot depicting the expression of adipose-derived ligand Tnfsf18 and T-cell specific receptor Tnfrsf18. Scale bar, 100 μm. I Representative smFISH (from two independent experiments) targeting Trdc (yellow), Tnfsf18 (orange), and Adipoq (purple) in the adipose tissue in the hypodermis in skin biopsies from naïve and infected animals. Two areas showing co-localisation of all the probes are highlighted for reference. The images have been adjusted for brightness and contrast to increase resolution. Scale bar, 50 μm.

Fig. 5. Vγ6⁺ cells are essential for controlling skin inflammation and subcutaneous adipose tissue wasting independently of skin-resident T_H1 T cells.

A In silico cell-cell interaction analysis between Vγ6⁺ cells (“senders”) and Pparg⁺ adipocytes (“receivers”) based on the upregulation of ligand-receptor pairs. The heatmap is colour coded to represent the strength of the interaction. B Expression level of Clcf1 and Areg, two of the most significant upregulated Vγ6⁺ cells-derived ligands predicted to interact with subcutaneous adipocytes. C Bodyweight of FVB/NJ and Vγ4/6^−/− mice over the course of infection (n = 3 mice/group). A non-parametric, one-way ANOVA was used to determine the level of significance. A p value < 0.05 is considered significant. Source data are provided as a source data file. D Subcutaneous adipose tissue mass from naïve and infected FVB/NJ and Vγ4/6^−/− mice normalised to whole bodyweight. A non-parametric, one-way ANOVA was used to determine the level of significance. A p value < 0.05 is considered significant. Source data are provided as a Source data file. E Representative H&E staining from skin biopsies obtained from FVB/NJ and Vγ4/6^−/− naïve and infected mice, from two independent experiments. Scale bar: 100 μm. epidermis (Ep), dermis (D), adipose tissue (Ad), muscle (M), subcutis (Sc). F Analysis of mean adipocyte area (µm²) in naïve and infected FVB/NJ and Vγ4/6^−/− mice. n = 5 biological replicates per group over two independent experiments, from two independent experiments. Lipid droplets were measured from 3 distinct areas in each image and then combined for each biological replicate. A non-parametric, one-way ANOVA was used to determine the level of significance. A p value < 0.05 is considered significant. Source data are provided as a Source data file. G Frequency plot of the adipocyte area represented in (D) for naïve and infected FVB/NJ (left panel) and Vγ4/6^−/− mice (right panel). n = 5 biological replicates per group, from two independent experiments. Lipid droplets were measured from 3 distinct areas in each image and then combined for each biological replicate. The data is presented as mean +/− SD. A non-parametric, one-way ANOVA was used to determine the level of significance. A p value < 0.05 is considered significant. Source data are provided as a Source data file.

Fig. 6. Proposed model of stromal-immune interactions in the skin during T. brucei infection.

Based on our spatially-resolved single cell atlas, we propose a model whereby Vγ6⁺ cells act concertedly with Pparg⁺ adipocytes (and potentially preadipocytes) to coordinate local immune responses, possibly viathe recruitment of immune cells (e.g., CD8⁺ T cells and natural killer (NK) cells). The Pparg⁺ adipocytes in this context provide important cues for T cell activation that we hypothesise might be involved in triggering Vγ6⁺ cells -mediated responses. Other stromal cells such as Langerhans cells and keratinocytes are also likely to be involved in this process via antigenic presentation. Created with Biorender (Agreement number WT25PETNWV).