Introduction. Cell therapy using Tregs has been shown to be effective for the treatment of several immune diseases in preclinical and clinical studies. However, their stability and persistence after adoptive transfer in an inflammatory environment is a matter of safety and therapeutic interest. Here, we explored the sensitivity of CD8⁺Tregs to the pro-inflammatory environment, specifically the role of FOXP3 for Treg stability.

Methods. CD8⁺CD45RC^low/- T cells were expanded in X vivo 15 medium supplemented with IL-2, IL-15 and rapamycin for 2 weeks using weekly stimulation with anti-CD3 and anti-CD28 mAbs, transduced or not with the recombinant lentiviral vector FOXP3-GFP on days +1 and +2, sorted on GFP expression on day 14 and challenged with pro-inflammatory cytokines and polyclonal stimulation for 6 days.

Results. First, we analyzed the expression of cytokine receptors on CD8⁺Tregs and set up combinations of pro-inflammatory cytokines to challenge them. While CD8⁺Tregs were resistant to a microenvironment associating IL-6 with TNFα or IFNγ, they were sensitive to IL- 1b, IL-6 and TGFβ. Indeed, after 6 days of culture with these pro-inflammatory cytokines, the suppressive activity of CD8⁺Tregs was maintained but shifted to a cytolytic mechanism, and their phenotype was disrupted, notably the expression of FOXP3 was significantly reduced. Thus, we generated a lentiviral vector encoding FOXP3 to increase and stabilize the expression of FOXP3 in CD8⁺Tregs. Indeed, the modified FOXP3_TgCD8⁺Tregs had increased expression of FOXP3, but also of CTLA-4, CD28, GITR and granzyme B, as well as increased suppressive activity in vitro. Importantly, ectopic FOXP3 expression compensated for the decrease in its endogenous expression induced by pro-inflammatory cytokine challenge. RNAseq analysis after pro-inflammatory cytokine challenge showed 18 genes differentially expressed in FOXP3_TgCD8⁺ Tregs versus 141 in unmodified cells. In addition, the increased expression in PD1 and 2B4 and the decreased expression in CTLA-4 induced by the pro-inflammatory cytokine challenge were corrected by the expression of FOXP3_Tg. Finally, we identified upstream regulators of FOXP3, whose ectopic expression increased FOXP3 expression and confered resistance to pro-inflammatory cytokine challenge.

Conclusion. Experiments are underway to assess the requirement of FOXP3 for CD8⁺Treg function, evaluate the persistence of FOXP3_Tg CD8⁺Tregs in a model of humanized acute GVHD in NSG mice, and further investigate regulators of FOXP3 expression.