Regulatory T cells control the reactivity of potentially harmful, self-reactive T cells and prevent autoimmune diseases. Significant progress has been made in the identification, derivation, and mechanism of action of T regulatory cells, previously called suppressor T cells. Heterogeneous T regulatory subsets can be grouped into naturally occurring and those induced in the periphery. Here, we consider whether we can harness T regulatory cells to function as a therapeutic agent for patients with established autoimmune diseases. Since the principal function of thymus-derived, natural CD4+CD25+ cells is to prevent autoimmunity, this subset would be an obvious choice. Besides their contact-dependent, cytokine-independent mechanism of action, they can also induce other CD4+ cells to become suppressor cells. However, only few natural CD4+CD25+ cells circulate in human peripheral blood. Alternatively, one can use IL-2 and TGF-beta to generate large numbers of CD4+CD25+ regulatory T cells ex vivo from naive T cells. These cells have the phenotypic and functional properties similar to natural CD4+CD25+ cells, including the capacity to induce CD4+CD25- cells to develop suppressive activity. These natural-like CD4+CD25+ regulatory T cells are the product of separate effects of IL-2 and TGF-beta on both natural CD4+CD25+ and CD4+CD25- cells. The ability of natural-like CD4+CD25+ cells to induce other CD4+CD25- cells to develop suppressive activity is both contact-dependent and cytokine-dependent. Thus, the effects of IL-2 and TGF-beta on both natural CD4+CD25+ cells and CD4+CD25- cells may trigger a continuous loop which results in the renewal of antigen-specific CD4+ regulatory T cells. These studies suggest that the adoptive transfer of CD4+ T regulatory cells generated ex vivo with IL-2 and TGF-beta as a treatment for autoimmune diseases may have sustained, long-term beneficial effects.