Discussion
In the present study, we show that Foxp3 functions as a potent repressor of NF-κB- and CREB-dependent transcriptional activation. Furthermore, the carboxyl-terminal FKH domain appears to be dispensable for mediating these effects, at least in T cell populations. This observation may become important in light of recent reports suggesting that Foxp3 expression in thymic epithelial cells was crucial for directing development of T cells in the thymus [42]. Interestingly, the majority of the genetic mutations associated with IPEX, a severe autoimmune disorder caused by functional inactivation of Foxp3, map to the carboxyl-terminal FKH domain or the leucine zipper domain in the central region of the protein. Only one mutation associated with IPEX to date has been mapped to the amino-terminal proline-rich region [43]. It is possible that the FKH domain has a complex tertiary structure that is particularly sensitive to misfolding caused by genetic mutations and that an intact FKH domain is absolutely critical for promoting Foxp3 function in the nucleus, whereas the structure of the amino-terminal proline-rich region may tolerate certain mutations as long as the NF-κB/NF-AT binding motif remains unaltered. This motif may also include the zinc finger domain. A logical region that may be targeted by the amino-terminal proline-rich region of Foxp3 is the Rel homology domain found in both NF-κB and NF-AT family proteins. A region that may also be important with respect to Foxp3 function is the leucine zipper domain, as demonstrated by the number of mutations associated with IPEX that have been mapped in this region of Foxp3. The role of this domain in Foxp3 function remains uncharacterized, but may play a role in dimer formation as it does in other Foxp family members [44].
Because the pathogenesis of a number of retroviral-induced immunologic disorders such as HIV-1/AIDS and HTLV-I/HAM/TSP have been associated with dysregulation of Foxp3 expression [8,45], we also examined the role of Foxp3 in retroviral gene expression. HIV-1 LTR activation in CD4+ T cells is critically dependent on two tandem NF-κB sites located between nucleotide positions −102 and −81 within the HIV-1 enhancer region, whereas HTLV-I LTR activation in the presence or absence of the HTLV-I-encoded transactivator protein Tax is independent of NF-κB [18]. To our knowledge for the first time, Foxp3 was shown to have a direct effect on HIV-1 LTR transcription. Deletion of the NF-κB sites within the HIV-1 enhancer region reduced the responsiveness of the HIV-1 LTR to Foxp3-mediated suppression. In addition, the FKH domain of Foxp3 was required for this inhibitory effect in HEK 293T cells, but not in Jurkat T cells, similar to Foxp3-mediated suppression of a synthetic NF-κB reporter. The direct effect of Foxp3 down-regulating HIV-1 gene expression correlates well with recently reported evidence indicating that higher regulatory activity of CD4+CD25+ T cells from HIV-1-infected patients was associated with lower HIV-1 viral loads in these patients [46].
Foxp3 also affected two well-known functions of HTLV-I Tax: transactivation of the NF-κB pathway and, most surprisingly, transactivation of the HTLV-I LTR. Transactivation of the HTLV-I LTR by Tax involves the interaction of ATF/CREB factors with Tax in the nucleus. Binding of Tax enhances ATF/CREB dimerization and promotes assembly of Tax-ATF/CREB complexes onto specific sequences in the viral promoter known as Tax-responsive elements. This series of steps allows Tax to recruit coactivator proteins CBP/p300 to the viral promoter and facilitate a high level of viral gene expression [24–26]. Transactivation of the NF-κB pathway by Tax was inhibited by overexpression of full-length Foxp3, but not ΔFKH, as seen with basal activation of the HIV-1 LTR and a synthetic NF-κB reporter in HEK 293T cells. However, Tax-mediated transactivation of the HTLV-I LTR was inhibited by overexpression of both full-length Foxp3 as well as ΔFKH in both HEK 293T cells and CD4+ T cells. We demonstrated that Foxp3 did not directly affect the functioning of Tax, but rather Foxp3 targeted the transcription factors required for Tax transactivation (i.e., NF-κB and a then-unknown cellular factor, which we identified in this study as CREB). The negative effect of Foxp3 on HTLV-I gene expression was confirmed utilizing an HTLV-I infectious molecular clone.
Importantly, we demonstrated that HTLV-I-infected individuals with the highest levels of Foxp3 protein expression within the CD4+CD25+ T cells population exhibited lower proviral loads than did individuals with the lowest levels of Foxp3 protein expression. Previous studies have demonstrated that the HTLV-I proviral load directly correlates with HTLV-I Tax mRNA load, the frequency of immunopathogenic virus-specific CD8+ T cells, and disease severity in patients with HAM/TSP [47]. These results have important implications on the utility of Foxp3 in controlling viral gene expression and thus pathogenesis of HAM/TSP. Therefore, Foxp3 becomes an attractive target for the development of novel therapeutic applications directed at modulating the expression of this important regulatory protein, especially in light of recent observations that the expression of Foxp3 can also be down-regulated by HTLV-I Tax [8].
As the activation of the HTLV-I LTR depends primarily on ATF/CREB proteins (whether in the presence or the absence of Tax), we investigated whether Foxp3 could interact with this additional cellular signaling pathway. While the DNA-binding activity of CREB is, in most cases, constitutive, the transactivation potential of CREB is regulated by the phosphorylation of CREB and recruitment of CBP/p300 [48]. Our data demonstrate that Foxp3 interferes with the latter of these two processes and that the recruitment of the coactivator protein p300, and resulting transcriptional activation are blocked by Foxp3. This may be the result of the physical interaction we detected between Foxp3 and p300. With respect to HTLV-I LTR activity, while full-length Foxp3 inhibited both basal and Tax-dependent transcription by ~50%, ΔFKH appeared less effective in suppressing basal activation (~25% inhibition) compared to Tax-dependent activation (~50% inhibition). The effect of ΔFKH on basal activation of the HTLV-I LTR in HEK 293T cells was very similar to that shown for a synthetic CREB reporter, suggesting that the FKH domain of Foxp3 is important at some level. As observed with NF-κB activation, the Foxp3 mutant lacking the FKH domain was a stronger inhibitor of CREB activation in CD4+ T cells than in HEK 293T epithelial cells. Therefore, it appears that in CD4+ T cells, the FKH domain is dispensable for the proper functioning of Foxp3 with respect to both NF-κB and CREB activation.
In summary, this is, to our knowledge, the first direct evidence implicating a role for the Treg-specific transcription factor Foxp3 in regulating retroviral gene expression. In addition, we identify the CREB pathway as a molecular target of Foxp3. Since CREB has been shown to regulate multiple genes involved in transcription (e.g., JunD, c-Fos, signal transducer of activated T cells 3 [STAT3]), cell cycle (e.g., p15INK4b, cyclin A, cyclin D1), and immune regulation (e.g., IL-2, IL-6, T-cell receptor α) (reviewed in [48]), the findings presented in this report broaden the potential range of signaling pathways under the control of the regulatory protein Foxp3. Our evidence stresses the importance of Foxp3 expression and Treg function in the development and maintenance of protective immunity against HIV-1 and HTLV-I. Based on recent findings, Foxp3 may limit HIV-1 and HTLV-I transcription by interfering with activation of NF-κB and CREB pathways. However, observing that this inhibitory effect is not absolute, a low level of viral gene expression may persist in CD4+ T cells (in particular regulatory T cells, which are known reservoirs of HIV-1 and HTLV-I) and result in the accumulation of viral proteins that either stimulate NF-κB and/or CREB activation or directly inhibit Foxp3 expression or function. The imbalance of NF-κB and CREB activation caused by these viral gene products may be a crucial step in the pathogenesis of virus-induced immunological disorders such as AIDS and HAM/TSP. Future studies will be directed at identifying and characterizing cellular proteins that interact with Foxp3 both in the nucleus and cytoplasm, in order to better address how Foxp3 functions to guide the development and function of regulatory T cells in health and disease.