We thank Thomas et al. (1) for their insightful comments regarding the functional dynamics of CD4+Foxp3+ regulatory T-cells (Tregs) in their model of type 1 diabetes. An impressive array of studies in the literature establishes that CD4+Foxp3+ Tregs play a central, master-switch role in peripheral tolerance in the NOD mouse model of spontaneous type 1 diabetes (2,3). A central question is whether the onset of spontaneous disease in NOD mice results from a decline in regulation over time or from uncontrollable activity of self-reactive T-cells. Type 1 diabetes may reflect subtle, functional deficiencies in regulatory T-cells, thus allowing the diabetogenic process to unfold. In our study (4), we attempted to determine whether temporal, quantitative, or qualitative defects in CD4+Foxp3+ Tregs contribute to spontaneous type 1 diabetes.
The BDC2.5 mouse model contains a highly pathogenic CD4+ T-cell repertoire and represents a unique system to study CD4+Foxp3+ Treg-mediated control of self-reactivity. The spontaneous onset of type 1 diabetes in BDC2.5 mice, though lower than in wild-type NOD mice, may be attributed to a number of predisposing Treg-intrinsic and -extrinsic variables. We are in agreement with Thomas et al. (1) that the onset of type 1 diabetes cannot be merely attributed to quantitative fluctuations in CD4+Foxp3+ Tregs. Consistent with Thomas et al. (1), we also observe an age-related increase in CD4+Foxp3+ Tregs, particularly in secondary lymphoid tissues in BDC2.5 mice (4). This increased frequency of CD4+Foxp3+ Tregs in spleen and pancreatic lymph nodes is likely a response of CD4+Foxp3+ Tregs to the increased inflammation in these sites. However, this increase in the numbers of CD4+Foxp3+ Tregs in these sites is not sufficient to suppress the diabetogenic process.
Our study is in disagreement with Thomas et al. (1) as to the potency of CD4+Foxp3+ Treg function in young (age 3–4 weeks) and old (age 20 weeks) BDC2.5 mice. In our hands, older BDC2.5 regulatory T-cells (fluorescence-activated cell sorter [FACS] highly purified cells injected at physiological 1:10 Treg–to–effector T-cell ratios) have a reduced suppressive potential compared with that of young Tregs. Interestingly, infusion of increased numbers of older BDC2.5 Tregs into T-cell–deficient recipient mice (FACS highly purified cells injected at nonphysiological 1:2 or 1:1 Treg–to–effector T-cell ratios) reestablished disease protection, an observation consistent with those of Thomas et al. (1). This observation suggests that the cellular potency of CD4+Foxp3+ Tregs, while fully operative in neonatal mice, declines with age despite a stable cellular frequency of CD4+Foxp3+ Tregs in primary and secondary lymphoid tissues. This observation is consistent with a number of recently published studies (5–8).
In our study (4), we also observe that type 1 diabetes onset coincides with an age-dependent decline in the cycling of CD4+Foxp3+ Tregs, particularly within the pancreas. It is noteworthy to mention that this proliferative decline of BDC2.5 CD4+Foxp3+ Tregs within the pancreas occurs at a time point well before the clinical onset of type 1 diabetes but, nonetheless, at a time point when immune dysfunction (self-reactive T-cell activation and insulitis) is readily detectable in these young mice. We believe that NOD mice succumb to a temporal loss in CD4+Foxp3+ Treg function that coincides with a reduction in CD4+Foxp3+ Treg cycling in pancreatic sites; this, in turn, unleashes the diabetogenic potential of effector T-cells in older BDC2.5 mice. We suggest that a functional deficiency in CD4+Foxp3+ Tregs may not be visible as a sudden decline in the cellular frequency of these cells in peripheral tissues and may conceivably be related to a reduced homeostatic fitness in inflammatory sites. Evidently, these results do not exclude the possibility that time-dependent changes in Treg-extrinsic variables, like dendritic cell or effector T-cell function, may also contribute to disease onset. It remains to be seen how mouse colony/flora differences also impact these tolerogenic mechanisms.
Currently, our ability to isolate CD4+Foxp3+ Tregs in the NOD model is limited to the isolation of CD4+CD25+ T-cells, which likely constitute a phenotypic and functional heterogenous population including a variety of effector and regulatory subsets. Thus, a more discriminate characterization of CD25− and CD25low T-cells relative to CD25+ Tregs is imperative. The next step to examine the functional dynamics of Tregs will be to rigorously track and characterize CD4+Foxp3+ Treg development, function, and homeostasis in type 1 diabetes by means of NOD.Foxp3–green fluorescent protein reporter mice. These experiments are well underway.