Single cell characterization of blood and expanded regulatory T cells in autoimmune polyendocrine syndrome type 1

Abstract

Immune tolerance fails in autoimmune polyendocrine syndrome type 1 (APS-1) because of AIRE mutations. We have used single cell transcriptomics to characterize regulatory T cells (Tregs) sorted directly from blood and from in vitro expanded Tregs in APS-1 patients compared to healthy controls. We revealed only CD52 and LTB (down) and TXNIP (up) as consistently differentially expressed genes in the datasets. There were furthermore no large differences of the TCR-repertoire of expanded Tregs between the cohorts, but unique patients showed a more restricted use of specific clonotypes. We also found that in vitro expanded Tregs from APS-1 patients had similar suppressive capacity as controls in co-culture assays, despite expanding faster and having more exhausted cells. Our results suggest that APS-1 patients do not have intrinsic defects in their Treg functionality, and that their Tregs can be expanded ex vivo for potential therapeutic applications.

Keywords: Components of the immune system; Health sciences; Immunology; Proteomics; Transcriptomics.

Graphical abstract

Figure 1

Single-cell transcriptomic profiles of freshly sorted Tregs from four APS-1 patients and four healthy controls (A) Method and number of patients. (B) UMAP plot of global single-cell sequencing data for patients and controls, identifying eight distinct clusters of Tregs. (C) Cluster annotations with some of the most expressed genes in each cluster. (D) Pie charts of cluster frequencies in APS-1 patients and healthy controls. (E) Bar plots of cluster frequencies per individual. Statistical testing was performed by the Mann-Whitney non-parametric t-test. Standard deviation is included for each bar. Significance level: p < 0.05∗. (F) Volcano plot of the most differentially expressed genes (DEGs) (−0.2 > log2FC > 0.2) between APS-1 and healthy controls, where log2FC > 0 represents upregulated genes and log2FC < 0 represents downregulated genes in APS-1. (G) Heatmap of the most differentially expressed genes (DEGs) (−0.2 > log2FC > 0.2) between APS-1 and healthy controls; reported as the average, normalized SCT expression value of the genes, taking into consideration all the cells that express a specific gene. The figure was made using BioRender (Biorender.com). See alsoFigures S1 and S7andTable S5.

Figure 2

Single-cell transcriptomic profiles of in vitro expanded Tregs from eight APS-1 patients and eight healthy controls (A) Method and number of patients. (B) UMAP plot of immune panel single-cell sequencing data for patients and controls, identifying ten distinct clusters of Tregs. (C) Cluster annotations with some of the most expressed genes in each cluster. (D) Bar plots of cluster frequencies per individual. Statistical testing was performed by the Mann-Whitney non-parametric t-test. Standard deviation is included for each bar. Significance level: p < 0.05∗. (E) Pie charts of cluster frequencies in APS-1 patients and healthy controls. (F) Volcano plot of the most differentially expressed genes (DEGs) (−0.2 > log2FC > 0.2) between APS-1 and healthy controls, where log2FC > 0 represents upregulated genes and log2FC < 0 represents downregulated genes in APS-1. (G) Heatmap of the most differentially expressed genes (DEGs) (−0.2 > log2FC > 0.2) between APS-1 and healthy controls; reported as the average, normalized SCT expression value of the genes, taking into consideration all the cells that express a specific gene. The figure was made using BioRender (Biorender.com). See alsoFigures S2–S7andTables S2, S3, S6, S7, and S8.

Figure 3

TCR VDJ gene usage analysis of in vitro expanded Tregs from seven APS-1 patients and nine healthy controls (A) V gene usage. (B) J gene usage. (C) D gene usage. (D) Summary histogram showing the Shannon diversity index (all clones) of in vitro expanded Tregs from seven APS-1 patients and eight healthy controls. ∗Threshold for significance is set to p < 0.05, calculated using a Mann-Whitney test for the Shannon Index. Standard deviations are shown for the bars. Frequencies of top 30 most abundant clonotypes, where clonotype is defined as the combined TRA+TRB amino acid sequences, for (E) APS-1 patients and (F) healthy controls. None of the top 30 clonotypes were shared between the two groups. See alsoFigures S8–S12.

Figure 4

Flow cytometry analysis of FOXP3, CD25, CD31 and Helios expression in CD4+CD25+FOXP3+ cells of expanded Tregs in 17 APS-1 patients and 14 healthy controls (A) Frequency of CD4⁺CD25+FOXP3+ cells, and geometric mean of CD25 within the CD4⁺ population, (B) CD31 within CD4⁺CD25+FOXP3+ cells and (C) Helios within CD4⁺CD25+FOXP3+ cells. ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001. Statistical tests were calculated using an unpaired, parametric t-test. Standard deviations are shown for all bars. See alsoFigures S13 and S14andTable S1.

Figure 5

Expression of different T cell markers within the CD4+CD25⁺ CD127^low expanded Treg population analyzed by CyTOF in 17 APS-1 patients and 17 healthy controls (A) Frequency of CD4⁺CD25⁺ CD127^lowcells. Expression of (B) CD57, (C) CD161 and (D) CD103 within CD4⁺CD25⁺ CD127^low cells. The figures are shown for flow cytometry results of a representative APS-1 patient and healthy control. The p values were calculated using an unpaired, parametric t-test. ∗p < 0.05 and ∗∗p < 0.01. Standard deviations are shown for the bars. See alsoFigures S15–S17andTables S1 and S4.

Figure 6

In vitro Treg suppression assay for 15 APS-1 patients and 15 healthy controls CellTrace Violet labeled Tresp cells were co-cultured with Tregs at different ratios for 5 days in the presence of anti-CD3/CD28 and IL2. (A) Tresp:Treg 1:1, (B) Tresp:Treg 2:1, (C) Tresp:Treg 4:1 and (D) Tresp:Treg 8:1. The figures show flow cytometry results from a representative APS-1 patient and a representative healthy control for each ratio. P-values were determined by an unpaired, parametric t-test. A p value <0.05 was considered significant. ns; non-significant. Standard deviations are shown for the bars. See alsoFigures S18–S21andTable S1.