Basis of Treg development in the thymus

Developing T-cells express antigen receptors (TCR) with diverse recognition specificities that are subjected to positive and negative selection pressures before emigrating out of the thymus.^1,2 Although negative selection deletes T-cells bearing autoreactive TCR, thymic selection is imperfect, so that some T-cells with auto-reactive TCR escape from the thymus into the periphery, where their autoreactive potential must be suppressed by regulatory T-cells (Tregs). Tregs are a specialized subpopulation of CD4 T-cells with suppressive capability that differ from conventional CD4 T-cells in their expression of the forkhead family transcription factor Foxp3 and in their dependence on γc-dependent cytokines such as IL-2.³ Curiously, thymic selection of Tregs closely resembles thymic deletion of conventional autoreactive T-cells, in that both events require high-affinity TCR signals, and both events require engagement of CD28 costimulatory ligands.^1,4 Consequently, it has not been understood why Tregs survive CD28 costimulatory signals in the thymus¹ and why Tregs, unlike all other MHC class II selected T-cells in the thymus, require γc cytokine signaling.³

To address these issues, we recently examined the effect of Foxp3 protein expression on cell viability.⁵ We found that Foxp3 induced developing CD4 T-cells to express a distinctive proapoptotic protein signature (abbreviated as Puma²⁺p-Bim²⁺p-JNK²⁺DUSP6⁻) and to repress expression of the prosurvival protein Bcl-2, with the result that Foxp3 was potentially lethal to cells that expressed it. However, we also found that Foxp3 lethality was prevented by γc cytokine signals, which upregulated Bcl-2 expression to levels that counterbalanced the lethal effects of the proapoptotic protein signature. Thus, Foxp3⁺Tregs require γc cytokines such as IL-2 to upregulate Bcl-2 expression to levels that protect Tregs from Foxp3-induced apoptosis.

The requirement for IL-2 survival signals could be circumvented by a Bcl-2 transgene, which protected Tregs from Foxp3-induced cell death in the absence of signaling by γc cytokines.⁵ This finding revealed that IL-2 was not needed to induce Foxp3 gene expression and contradicts the conventional perspective that IL-2 is required to induce Foxp3 gene expression in developing CD4 thymocytes.^6,7 In fact, expression of the Bcl-2 transgene not only obviated the requirement for IL-2, it resulted in approximately 3-fold more Foxp3⁺ T-cells than normally arise in the thymus, indicating that the thymus contained insufficient IL-2 to rescue all newly arising Tregs. This observation was not unique to Bcl-2 transgenic mice, as mice deficient in the proapoptotic proteins Puma and Bim also contained 3-fold more Foxp3⁺ T-cells in the thymus.⁵ The Foxp3⁺ T-cells that were increased in number in both Bcl-2 transgenic and Puma/Bim-deficient mice were Foxp3⁺CD25⁻ CD4 T-cells that were phenotypically distinct from functionally mature Tregs that are Foxp3⁺CD25⁺. Notably, such Foxp3⁺CD25⁻ T-cells possessed the potential to differentiate into functionally mature Foxp3⁺CD25⁺ Tregs when signaled by IL-2.⁵ Thus, Foxp3⁺CD25⁻ T-cells are precursors of Foxp3⁺CD25⁺ mature Tregs.

It has conventionally been thought that all Foxp3⁺ cells in the thymus have been IL-2 signaled, because IL-2 signaling is necessary to induce Foxp3 gene expression.^6,7 In the conventional perspective, mature Foxp3⁺CD25⁺ Tregs must arise in the thymus from IL-2-signaled Foxp3⁻CD25⁺ precursors. However, the existence of Foxp3⁺CD25⁻ Treg precursor cells requires that the conventional view of Treg differentiation be replaced with a more expanded view. We think 2 distinct developmental pathways actually exist, by which conventional CD4 T-cells are signaled to differentiate in the thymus into mature Tregs (Fig. 1). We think that most mature Tregs arise from Foxp3⁺CD25⁻ precursors, while a minority are the progeny of Foxp3⁻CD25⁺ precursors. Interestingly, both major and minor developmental pathways are initiated by CD28 costimulation of conventional CD4 T-cells and are completed by IL-2-signaled differentiation of precursors into functionally mature Tregs (Fig. 1). Importantly, the 2 pathways differ in the prominence of Foxp3-mediated apoptosis and their dependence on IL-2 survival signals. In the major pathway, Foxp3⁺CD25⁻ precursors are highly susceptible to Foxp3-induced apoptosis, with most precursors dying in the thymus, because IL-2 is present in insufficient amounts in the thymus. In the minor pathway, Foxp3⁻CD25⁺ precursors express Foxp3 only after receiving IL-2 signals, so they are not susceptible to Foxp3-induced apoptosis. It remains to be determined if important functional differences exist in the mature Foxp3⁺CD25⁺ Tregs derived from the 2 different precursor populations.

Figure 1.

Two distinct developmental pathways for Treg cell differentiation in the thymus. Developing (Foxp3⁻CD25⁻) CD4 thymocytes that receive TCR/CD28 costimulatory signals differentiate into either Foxp3⁺CD25⁻ or Foxp3⁻CD25⁺ precursors. Foxp3⁺CD25⁻ precursors represent the major developmental pathway for Tregs. Most Foxp3⁺CD25⁻ precursors die because of insufficient IL-2, while IL-2-signaled Foxp3⁺CD25⁻ precursors differentiate into mature Foxp3⁺CD25⁺ Tregs.

Finally, it is interesting to consider why developing Tregs are the only cells to survive costimulation in the thymus. It is known that Foxp3 targets Zap70 and Itk, 2 kinases involved in TCR signal transduction, so that TCR/CD28 signaling might be dampened sufficiently to allow Foxp3⁺ cells to avoid clonal deletion.⁸ If so, Foxp3 expression would be a double-edged sword that first protects Treg precursors from clonal deletion, only to then make them susceptible to Foxp3-mediated apoptosis if they fail to be signaled by pro-survival γc cytokines.