Characterisation of human tissue-resident T-regulatory cells in healthy tissues and in-vitro generation of tissue-like Treg cells with wound-healing capabilities

Regulatory T cells (Tregs) are characterized by the expression of FOXP3 and are important for regulating other cellular components of the immune system. Their absence leads to autoimmune diseases like the “Scurfy phenotype” in mice or the Immunodysregulation, Polyendocrinopathy, Enteropathy, X-linked (IPEX) syndrome in humans. Non-lymphoid tissue (NLT) Tregs are distinct from lymphoid tissue (LT) Tregs. Comparing the gene expression profile (transcriptome) and the DNA methylation profile (epigenome) of NLT vs LT Treg cells in mice, our lab identified transcription factors as well as other genes particularly enriched in NLT Tregs. This led to the identification of a tissue specific Treg subset in mice, characterized by the surface expression of Klrg1 and ST2, a TH2 biased gene program as well as their production of Amphiregulin, which has been shown to aid in tissue repair. In mouse, this program is driven by the Transcription factor Batf. In this thesis, we worked to identify the human counterpart of this population. We established methods to isolate T-cells from tissues, and when this was optimised, performed the Assay for Transposase Accessible Chromatin following Sequencing (ATAC-seq) on a single cell level (scATAC-seq) to identify changes in chromatin accessibility on murine and human CD4+ T-cells. We found that this was a good and reliable method to identify different populations of T-cells in various tissues, among which Tregs in Non-Lymphoid tissues. From these datasets we identified shared peaks within the human and murine datasets; open chromatin that was conserved between species. We examined BATF binding in this shared peakset and found that more than half of the shared peaks had a BATF binding site, indicating importance in both human and murine tissue Tregs. Upon closer look we found that BATF gene activity score was increased in tissue Tregs compared to blood derived Tregs. Using bulk ATAC sequencing to identify changes in chromatin accessibility, we were able to show that we could induce a tissue Treg-like profile with the combination of cytokines IL-4 and IL-33 in mouse, and that this induction is depended on the Batf, confirming that Batf is both essential in-vitro and in-vivo. Among the shared peakset, we could identify several surface receptors, among which CCR8. Upon closer look we could determine that CCR8 gene activity was increased in both blood memory Tregs and tissue derived Tregs. This could be confirmed on protein level in blood and in tissues, where we could identify a subpopulation of CCR8+ Tregs in memory Tregs in the blood. In fat and skin, the vast majority of Tregs was CCR8+. We also found that CCR8+ Tregs express a significantly higher amount of FOXP3, BATF and CD25, on both RNA and protein levels. Next we worked to identify more surface markers using LEGENDScreen™. We identified HLA-DR as a reliable positive selection marker for CCR8+ Tregs in blood, fat and skin. CD49d could be identified as a negative selection marker for these same cells. Using single cell RNA sequencing (scRNA-seq) and Single cell TCR sequencing (scTCR-seq) we were able to identify a clonal relationship between CCR8+ from blood, fat and skin, indicating that circulating CCR8+ Tregs are either potential progenitors or recirculating tissue Tregs. Through published datasets and our own RNA-seq data, we were able to show that human tissue Tregs have a Tfh-like phenotype, unlike murine tissue Tregs, which have a Th2-like phenotype. This phenotype could be induced with the cytokines IL-2, TGFβ, IL-12, IL-21, and IL-23. Supernatant produced by cells cultured with these cytokines were superior at woundhealing compared to IL-2 only Tregs. Finally, we were able to show that Tregs found in tumours share many commonalities with Tregs found in healthy tissue, like BATF, CCR8 and HLA-DR expression