Introduction. Cell therapies using CD4⁺Tregs or CD8⁺Tregs are both and separately under investigation in preclinical and clinical studies. Combining Treg subsets would increase the number of starting Tregs, reducing the volume of blood drawn from patients and thus improving their well-being; would simplify cell isolation, saving time and costing less; and could be more efficient by recognizing different antigens. Here, we explored the differences between these therapeutic products and benefits of combining them.

Methods. CD8⁺CD4^-CD56^-CD45RC^low/- T cells and CD45RC^low/- T cells were expanded in X vivo 15 medium supplemented with 25U/mL IL-2, 10ng/mL IL-15 and 50 nM rapamycin for 3 weeks using weekly stimulation with purified plastic-coated anti-CD3 and soluble anti-CD28 mAbs. CD8^-CD25⁺ T cells were expanded in X vivo 15 medium supplemented with 500U/mL IL-2 and 100 nM rapamycin for 3 weeks using weekly stimulation with anti-CD3 and anti-CD28 coated beads. NSG mice were 1.5Gy-irradiated and injected with 1.5x10⁷ human PBMCs with or without Tregs. CD4⁺CD25^- and CD8⁺CD45RC^hi Tconv were stimulated by allogeneic APC for in vitro suppression assays.

Results. Peripheral blood CD4⁺Tregs compared to CD8⁺ Tregs displayed significantly lower methylation in the FOXP3-TSDR, 253 genes differentially expressed, and higher efficiency in suppressing both naive and memory CD4⁺ Tconv responses in vitro. In contrast, after overnight polyclonal activation, transcriptomic comparison showed no DE gene between Treg subsets and similar efficacy in suppressing Tconv responses in vitro. We showed that Tregs were switchable and that combining subsets resulted in more efficient suppression than each individually. After separate expansion of CD4⁺Tregs and CD8⁺Tregs under pro-regulatory conditions, the combination of a half dose of CD4⁺Tregs and CD8⁺ Tregs in a model of humanized acute GVHD in NSG mice resulted in therapeutic effect similar to that of a full dose of CD4⁺ or CD8⁺ Tregs. Moreover, the addition of CD4⁺ and CD8⁺ Tregs improved long term mouse survival to 66%. To enrich for both FOXP3⁺ CD4⁺ and CD8⁺ Treg together, we evaluated the therapeutic properties of T cells isolated only on the CD45RC^low/- expression marker and expanded them under pro-regulatory conditions. Importantly, expanded CD3⁺CD45RC^low/- cells were able to suppress CD4⁺Tconv proliferation as efficiently as CD4⁺ or CD8⁺Tregs. Finally, the in vivo transfer of 1.5 doses of CD3⁺CD45RC^low/- cells did not induce any sign of toxicity, and was necessary and sufficient to delay GVHD as efficiently as CD4⁺ or CD8⁺Tregs, which should be largely offset by the production yield.

Conclusions. CD4⁺ and CD8⁺Tregs show similarities and their complementarity for Tconv control and is possible to isolate them from blood using only the CD45RC^low/neg marker and appropriate culture conditions.