The adoptive transfer of donor CD4+CD25high regulatory T (Treg) cells has been suggested for the prevention of graft-versus-host disease (GVHD) after allogeneic stem cell transplantation (SCT). In preparation of such trials we previously described protocols for the efficient in vitro expansion of human Treg cells (Hoffmann et al. 2004, Blood 104:895). Strong costimulation provided by immobilized anti-CD3 and anti-CD28 antibodies together with high-dose IL-2 resulted in a more than 3-log polyclonal expansion of Treg cells with strong suppressive activity. In contrast to CD4+CD25− T cells, the majority of the CD4+CD25high Treg cells maintained expression of the lymph node homing receptors CD62L and CCR7 during in vitro culture.Detailed examination of sorted subpopulations from Treg cell lines now revealed that only CD62L+CCR7+ cells combined all phenotypic and functional characteristics of natural Treg cells, such as FOXP3 expression, lack of cytokine secretion and potent suppression of responder T (Tresp) cell proliferation. To elucidate the origin of this cell population, we initiated cultures from CD45RA+ naive as well as CD45RA− effector/memory-type CD4+CD25high Treg cells. The CD45RA+ population initially contained 95 ± 2.5% CD62L+CCR7+ cells and maintained this phenotype with still over 90% CD62L+CCR7+ cells after 2 weeks and appox.70% after 3 weeks in culture (n=10). In contrast, CD62L and CCR7 expression in CD45RA− Treg cells was less stable and decreased from 58 ± 8% CD62L+CCR7+ cells after isolation to 32 ± 20% after culture for 3 weeks. Similar differences were observed with respect to cytokine production, as determined by intracellular staining: Whereas less than 5% (n=8) of expanded CD45RA+ Treg cells expressed IL-2 and/or IFN-γ, almost 40% of expanded CD45RA− Treg cells produced one or both of these pro-inflammatory cytokines upon stimulation with PMA/ionomycin. Interestingly, expanded CD45RA− Treg cells also contained a defined subpopulation of IL-10-producing cells (6.9 ± 5.6%; n=8), which was absent in expanded CD45RA+ Treg cells. Both subpopulations showed suppressive activity after polyclonal restimulation, however, suppression of Tresp cell proliferation by expanded CD45RA+ Treg cells was much more profound with only 13.6 ± 4% proliferating cells as compared to 35.4 ± 7.7% in the presence of expanded CD45RA− Treg cells and 79.6 ± 18.7% in the absence of Treg cells, as measured in a CFSE-dilution assay at a 1:4 ratio of Treg and Tresp cells (n=7). Most importantly, when determined on a single cell level by intracellular staining, 93.6 ± 1 % (n=3) of expanded CD45RA+ Treg cells still expressed FOXP3 after 3 weeks in culture, whereas only 10.8 ± 6.8% of expanded CD45RA− Treg cells remained FOXP3+. No re-expression could be induced in FOXP3− expanded CD45RA− Treg cells by short-term restimulation via CD3/CD28, whereas expanded FOXP3+ CD45RA+ Treg cells transiently upregulated FOXP3, with peak expression levels as early as 24h after stimulation and return to baseline levels by 48 to 72h. Based on these unexpected findings that only naive CD45RA+, but not memory-type CD45RA− CD4+CD25high T cells give rise to homogeneous Treg cell lines, we suggest that isolation and expansion of CD45RA+ CD4+CD25high T cells is the best strategy for adoptive Treg cell therapies.
Read full abstract