We report here that, although mature single positive thymocytes including CD4+CD25+ thymocytes can be detected in day 1–2 neonatal mice, the FoxP3 gene is not expressed in the thymus until 3–4 days after birth and the expression dose not approach adult levels until after 6–8 days. FoxP3 expression was detected in both CD4+CD8-CD25+ and CD4+CD8-CD25- thymocytes. As expected, FoxP3-CD4+CD8-CD25+ thymocytes from day 2 newborn mice show no Treg activity. Almost all FoxP3+ thymocytes are detected in the medullary region of the thymus. In addition, we have demonstrated that TSLP enhances maturation of CD4+CD8-CD25+ thymocytes as well as FoxP3 expression, and that blocking endogenous TSLPR in the thymus inhibits FoxP3 expression. Furthermore, TSLP interacts directly with CD4+ SP thymocytes to affect FoxP3 expression and Treg development.
Our findings agree with a recent report showing that lower levels of FoxP3 mRNA were detected in CD4+
T cells from 3 day old newborn mice [30
]. This report, however, failed to clarify whether FoxP3 is expressed in 1–2 day old mice, and if all Treg cells from 3 day old mice express lower levels of FoxP3 mRNA or if a smaller fraction of these cells expresses the same level of FoxP3 mRNA as CD4+
Treg cells from adult mice. We have demonstrated here that FoxP3 expression is not detectable in Treg-like cells from 1–2 day old mice, and that a smaller fraction of day 3 neonatal CD4+
thymocytes expresses the FoxP3 protein at levels comparable to mature adult Treg cells.
As detected by immuno-staining of thymus tissues, both CD4+
cells were detected in the thymus (Fig. ). FoxP3+
thymocytes are induced in the 3–4 day old thymus, with a medullary distribution similar to that seen in the adult thymus. This is consistent with a recent report using the FoxP3-GFP knock-in mouse. In this model, expression of FoxP3 is detected by GFP [24
]. Similar to our finding using anti-FoxP3 staining, the FoxP3-GFP+
thymocytes start to be detectable in the medulla of the 3–4 day old thymus. These results suggest that CD4+
thymocytes need paracrine signals provided by cells in the medullary region to become fully mature FoxP3+
Treg thymocytes. As positive selection and negative selection of thymocytes occurs predominantly in the cortex and the CMJ, we would expect a preferential localization of emerging FoxP3+
thymocytes in the cortex/CMJ if commitment to natural Treg and FoxP3 induction is closely related to the TCR-mediated positive/negative selection. The dispersed medullary distribution of FoxP3+
thymocytes in 3–4 day old thymus organs supports the idea that induction of FoxP3 expression may be a post-selection event, involving unique cell types and/or cytokines in the medulla.
CD28-mediated signaling is required for the induction of FoxP3 and natural Treg generation [32
]. The CMJ is enriched with APC that express the B7 ligands which signal to CD28. Other factors must, therefore, be provided in the medulla to lead to functional maturation of FoxP3+
Treg thymocytes. It was recently reported that the Hassall's corpuscles in the human thymic medulla preferentially express the cytokine TSLP, which promotes maturation of DC and induces natural Treg generation [33
]. The murine thymus does not have Hassall's corpuscle-like structures although mouse TSLP is also predominantly expressed by medullary TECs in the thymus [33
]. We report here that exogenous mouse TSLP can enhance expression of FoxP3 and maturation of natural Treg cells in mouse FTOC, and that blocking TSLPR leads to reduced expression of FoxP3. We have detected TSLP gene expression by RT-PCR in 1–6 days old neonatal and adult thymus organs, but its expression kinetics is not correlated with functional FoxP3+
Treg development (data not shown). Additional paracrine factors are probably involved.
How does TSLP modulate maturation of CD4+
Treg cells in the thymus? Although TSLP is expressed predominantly by thymic medullary stromal cells, its effect on T cells in mice has not been clearly established [34
]. In transgenic mice over-expressing TSLP, development of B cells and T cells appears to be impaired, but myeloid cell numbers are increased [37
]. In human studies, TSLP has been shown to modulate DC maturation to indirectly regulate T cell responses [26
]. In contrast, mouse TSLP has been implicated in directly promoting TCR-mediated proliferation of mouse CD4+
T lymphocytes both from the thymus and spleen [29
]. Our study with purified CD4+
thymocytes after two days in culture, in the absence of TCR stimulation, suggests that TSLP induces FoxP3 in CD4+
) thymocytes, but not in CD4+
DP or CD4-
SP8 thymocytes. In addition, TSLP failed to induce FoxP3 expression in CD4+
T cells from the spleen under similar culture conditions (Fig. ). We also showed that TSLP treatment affected neither the proliferation nor the survival of CD4+
thymocytes during the two day culture in vitro (Fig. ). Therefore, TSLP may have a unique effect on CD4+
thymocytes to induce FoxP3 gene expression that is distinct from its activity in promoting CD4+
T cell proliferation.
Figure 10 TSLP does not induce Foxp3 expression in CD4+25- spleen T cells. CD4+25- T cells purified from the spleen were cultured with or without TSLP (100 ng/ml) for 2 days and analyzed for relative expression of FoxP3 (FoxP3/18S) by real-time PCR. Error bars (more ...)
The receptor for TSLP consists of a heterodimer of the interleukin 7 receptor alpha (IL-7Rα) chain and the TSLP receptor (TSLPR) that resembles the cytokine receptor common gamma chain [35
]. As suggested in previous reports with TSLPR mutant mice [29
], TSLPR and other receptors/ligands functionally interact to modulate B cell development and T cell proliferation. TSLP-mediated signaling is unique among members of the cytokine receptor family in that activation of the transcription factor Stat5 occurs without detectable Janus kinase activation [40
]. Interestingly, murine TSLPR mediates signals that are distinct from the Jak3 mediated signals seen with the common gamma chain [36
]. This potentially unique cytokine signaling activity of TSLPR may be involved in modulating FoxP3 expression and CD4+
Treg maturation in thymocytes. TSLP also appears to have different effects on pro-B cells of fetal and adult origin [42
], suggesting distinct signaling networks in developing vs. mature cell types may determine their responsiveness to TSLP. A similar distinction may exist in the thymocytes versus peripheral mature T cells as, unlike what we have seen in the thymus, treatment of CD4+
T cells from the spleen with TSLP showed no effect on FoxP3 gene expression (Fig. ).
Blocking TSLPR in FTOC only partially reduces the number of FoxP3+
thymocytes. In addition, the frequency and function of FoxP3+
Tregs in the spleen and lymph nodes of TSLPR-null adult mice are not significantly altered (Q. Jiang, G. Knudsen and L. Su, unpublished results). Thus, TSLP is involved in, but not absolutely required, for the process that regulates FoxP3 expression and Treg maturation. Other factors may compensate for the loss of TSLP signaling during embryogenesis or during Treg maturation. Paracrine factors known to be involved in Treg generation and function, including interleukin-10 and TGFβ, are also expressed in the thymus and may contribute to this process. For example, TGFβ has been reported to induce Treg and FoxP3 expression in mature CD4+
T cells [11
]. Our preliminary results, however, indicate that TGFβ and IL-10 had very low or no effect on FoxP3 expression in the FTOC model (Q. Jiang and L. Su, unpublished results). It was recently reported that the cytokine common gamma chain γc is required for the expression of FoxP3 and Treg maturation [44
]. In this context, the TSLPR and γc receptor may mediate similar down-stream signals such as STAT5 to activate FoxP3 gene expression and Treg maturation. It will be of interest to use STAT5a/b double KO mice [45
] to test weather STAT5 is required for the expression of FoxP3 and Treg generation.
T cells have been implicated in a number of pathologic processes including elevated levels of Treg cells in certain types of cancer [47
] and infectious diseases [17
], and reduced Treg levels in autoimmune diseases [9
]. Novel therapies based on modulating Treg activity have great potential in treatment of these diseases and in promoting engraftment of allogeneic organ transplant. Elucidating how FoxP3 expression is modulated during the development of natural Treg cells in the thymus will be of great importance both for understanding the ontogenic mechanism of their development, and for shedding light on how to modulate their generation and expansion for use in therapeutic applications.