Using epigenetic reprogramming through Somatic Cell Nuclear Transfer, we developed two regulatory T cell (Treg) models, T138 (Ku M., et al., PNAS 2016) and T143 (Ku M., et al., JACI 2018). While both Treg models were thymically derived, T143 displayed characteristics of agonist-selection (aTreg), which had been postulated for all Treg cells. However, T138 was quite distinct with a higher degree of interaction between TCR beta-chain and MHC-II rather than TCR and agonist. We refer to as T138 as natural-occurring Treg cell (nTreg). Strikingly, we identified FoxP3-negative T cells expressing the same TCR as nTreg and aTreg. Especially, FoxP3-negative T cells isolated from T138 were heavily poised in their ability to differentiate into FoxP3-expressing cells rather than Th1-differentation.

In order to further identify new transcription factors that mediated this epigenetically poised state in FoxP3-negative T138 (pre-nTreg) cells and to identify accessible promoters and enhancers that are associated with the pro-FoxP3 poised state, we performed ATAC-Seq (assay for transposase-accessible chromatin with high-throughput sequencing). We identified several loci that are more accessible in pre-nTreg cells than in wild-type (WT) cells. Motif analysis revealed enrichment of several lineage-determining transcription factor binding motifs, such as Atf3. Strikingly, transcriptomic analysis demonstrated that Atf3 is specifically expressed in T138 and not in WT. Interestingly; it has previously been shown that the FoxP3-locus contains an Atf-binding site, suggesting an important link of pre-nTreg transcriptional programs to its development into nTreg.

Since ATAC-Seq signals show similar enrichment at both active and poised (bivalently-marked by both H3K4me2 and H3K27me3 marks) promoters, we performed H3K4me2 and H3K27me3 chromatin-immunoprecipitation followed by high-throughput sequencing (ChIP-Seq) to further identify poised promoters, which gear e.g. pre-nTreg towards a FoxP3+ cellular fate.

In order to identify unique phosphorylation sites during activation of pre-nTreg and nTreg cells, we analyzed the phospho-proteome. We identified several differential phosphorylated peptides. Particularly, we found about 100 peptides uniquely phosphorylated in FoxP3- pre-nTreg cells. Together with the transcriptional analysis this clearly demonstrates the unique characteristics of this novel pre-nTreg population.

Several mechanisms had been shown to mediate immune regulatory function of Treg cells. Strikingly, we identified Ctla4 and Lag3 to be differentially expressed in T138, T143 and the Treg pool in WT mice. This is of particular interest since, e.g. CTLA4 is currently being explored for immunotherapy. Differential expression of Ctla4 provides evidence for selective modifications of a Treg subset through anti-CTLA4 treatment. Our data supports the concept that anti-CTLA4 treatment might function through targeting of a tumor-infiltrating Treg subset rather than check point inhibition.

Taken together our multiomics and functional data on Treg cells clearly demonstrated the existence of distinct Treg subsets, which has major implications in better understanding disease mechanisms and in devising new future therapies.


Sabouri-Ghomi:Fate Therapeutics: Employment.

Author notes


Asterisk with author names denotes non-ASH members.