In 1995, Sakaguchi et al. [1] (CIT) described that CD4 T cells constitutively expressing the interleukin-2 receptor alpha chain (CD25) are potent suppressors of autoreactive T-cell responses. In their experimental system, they showed that the adoptive transfer of CD4 CD25 T cells inhibited the autoimmunity that is regularly observed in mice after early post-natal thymectomy. These experiments suggested that the suppressive T-cell population is thymus-derived and exported from the thymus only after birth and that there is no other peripheral tolerance mechanism compensating for the lack of these ‘regulatory’ T cells (Treg). A few years later, the transcription factor Foxp3 was identified as a master regulator for the development and peripheral function of those thymus-derived Treg cells, and it was shown that loss of function mutations in the foxp3 gene results in a lack of this natural Treg cell compartment [2 6]. In mice and humans, such mutations cause severe autoimmunity (scurfy and IPEX, respectively). Interestingly, boys with IPEX syndrome (immune dysregulation, polyendocrinopathy, eteropathy, X-linked) suffer from severe colitis and develop unusual immune responses, for example against nutritional antigens. In experimental models of colitis, Mottet et al. [7] showed that adoptively transferred Treg cells not only inhibit colitis induction but, more importantly, even cure established disease. Thus the adoptive transfer of natural Treg cells seems a promising strategy for the treatment of chronic colitis. In this issue of Cytotherapy, Sumida et al. [8] present methods for the enrichment of natural Treg cells from leukapheresis products of patients with ulcerative colitis (UC). For this purpose, they have adopted a strategy described previously for the isolation of Treg cells from healthy donors in allogeneic stem cell transplantation [9,10]. This isolation protocol was developed for the prevention of graft-versus-host disease after allogeneic stem cell transplantation, where the co-transplantation of conventional and Treg cells at a 1:1 ratio is envisaged. Sumida et al. [8] have added a CD8-depletion step in addition to the CD19-depletion and repetitive CD25enrichment cycles to get rid of the 2 5% of CD8 T cells that otherwise contaminate a Treg-enriched cell product. Overall, they demonstrate an approximately 55% enrichment of natural Treg cells from leukapheresis products of colitis patients, based on the proportion of FOXP3 cells in the target cell fraction. However, the final cell product does not contain solely Treg cells but also recently activated CD4 T cells expressing CD25 and probably also some non-regulatory T cells that transiently express FOXP3 after stimulation. Whether such transient FOXP3 expression confers suppressive activity is a matter of debate and doubted by several investigators in the field. Nevertheless, exclusive surface markers for the unequivocal identification and separation of Treg cells from activated conventional T cells have not yet been identified and the protocol presented by Sumida et al. [8] is the best currently available approach for the enrichment of Treg cells under good manufacturing practice (GMP) conditions. Although additional depletion of CD127 cells or enrichment of naive Treg cells (CD45RA ) have been suggested to improve Treg cell purities [11 13], such strategies remain to be evaluated at large scales with GMP-compatible reagents. Consequently, as long as we lack a technology that permits the generation of pure Treg cell products, we have to define the necessary degree of purity for each potential clinical application. Although leukocyte depletion from the peripheral blood of patients with UC is not yet a standard treatment world-wide [14], it