Inhibition of FOXP3/NFAT Interaction Enhances T Cell Function after TCR Stimulation

Regulatory T cells (Tregs) are a distinct lymphocyte lineage endowed with inhibitory properties that affect the activation of the immune system. They are characterized by the expression of CD25 and the Treg-specific FOXP3 transcription factor, which is required for their development and function (1). These cells can inhibit activation of other T cells (2) and are needed for protection against autoimmune diseases. However, immunoregulatory function of Tregs may hinder the induction of immune responses against cancer (3). Indeed, Tregs capable of suppressing the in vitro function of tumor-reactive T cells have been found in humans in many tumors types (4–7) and have been associated with a high death hazard and reduced survival (4, 6). Additionally, FOXP3 is expressed transiently in conventional CD4⁺ T cells upon TCR stimulation, leading to cell hyporesponsiveness (8).

The molecular basis of FOXP3 function has been poorly understood. The capacity of FOXP3 to bind DNA is critical for its functionality; however, it is clear that FOXP3/DNA interactions are assisted by FOXP3 cofactors and by multimerization (9–12). Thus, those strategies able to inhibit a particular interaction with FOXP3 or able to modify the FOXP3 interactome might have important consequences on the whole transcriptome signature of the FOXP3-expressing cell and, consequently, on its activity.

FOXP3 can regulate gene expression of a number of genes that are also targets for the transcription factor NFAT, which, in cooperation with AP-1 (Fos/Jun), can activate many genes during lymphocyte activation (13, 14). This regulatory capacity of FOXP3 is justified by its ability to interact physically with NFAT and regulate its activity (15–18). The cooperative complex between FOXP3 and NFAT is required to repress expression of the cytokine IL-2, upregulate expression of the Treg markers CTLA4 and CD25, and confer suppressor function (18). We propose in the present study that inhibition of this interaction might lead to the impairment of specific functions of FOXP3 and, thus, be beneficial in the development of vaccines and tumor therapies.

Plasmids pET45b His-NFAT1, pET20b FOXP3-His, and pET20b Runx1-His were generated to produce the human NFAT1 DNA-binding domain, human FOXP3, and Runx1 tagged with six histidines, respectively. Plasmid pDEST15-FOXP3 (provided by Dr. Ignacio Casal, Centro Nacional de Invesatigaciones Oncológicas, Madrid, Spain) was used to produce the fusion protein GST-FOXP3. Peptides FOXP3 393–403 (KCFVRVESEKG), biotinylated FOXP3 393–417 peptide (biot-KCFVRVESEKGAVWTVDELEFRKKR), NFAT1 659–673 (YVINGKRKRSQPQHF), and the indicated mutants and controls were synthesized by the solid phase method as previously described (19).

Screening of peptide binding to FOXP3, NFAT1, or to biotinylated FOXP3 393–417 peptide was performed by surface plasmon resonance using ProteOn XPR36 (Bio-Rad, Hercules, CA) optical biosensor. FOXP3/NFAT1, FOXP3/Runx1, as well as FOXP3/DNA or NFAT1/DNA interactions were analyzed by AlphaScreen technology (PerkinElmer, Benelux). Biotinylated oligonucleotides containing the putative binding sites for NFAT1 (biot 5′-AGGGACTTTCCGCTGGGGACTTTCC-3′) (20) or for FOXP3 (biot 5′-CAAGGTAAACAAGAGTAAACAAAGTC-3′) (21) and the corresponding His-tagged protein NFAT1 or FOXP3 were used. For FOXP3 dimerization assays, GST-tagged FOXP3 and His-tagged FOXP3 were coincubated in the presence or absence of the indicated peptides.

Jurkat (American Type Culture Collection), Karpas-299 (Deutsche Sammlung von Mikroorganismen und Zellkulturen, Braunschweig, Germany), TC-1 P3 (A15) cells (provided by Dr. T.C. Wu, Johns Hopkins Medicine, Baltimore, MD), and Hepa-129 (provided by Dr. M. Gonzalez-Carmona, University of Bonn, Bonn, Germany) cell lines were cultured in complete medium (RPMI 1640 containing 10% FCS, antibiotics, 2 mM glutamine, and 50 μM 2-ME). PlatE (Invivogen, Toulouse, France), B16-OVA, and 293 cell lines were cultured in DMEM supplemented with 10% FCS and antibiotics.

pGL3 basic luciferase reporter vector containing the −391 to +65 nt region of the CD40L with the putative NFAT binding site (22) and the mutant plasmid pCD40-391 containing a point mutation on the NFAT site were provided by Dr. S.A. Crist (University of Iowa, Iowa City, IA). Plasmid ph17p(-232)-Luc, expressing luciferase under the control of the minimal IL-17 promoter region (23), was purchased from Addgene. Plasmid pFOXP3 393–403 expressing the peptide FOXP3 393–403 was produced in our laboratory. The plasmid pMIG-R1 FOXP3 was provided by Dr. R.K.W. Mailer (Max-Delbrück-Center for Molecular Medicine, Berlin, Germany). As internal control, Renilla plasmid was used, provided by Dr. P. Fortes (Center for Applied Medical Research). Transfections were performed with Lipofectamine LTX and Plus reagent (Invitrogen, Carlsbad, CA).

Female BALB/c, C57BL/6, and C3H/HeN mice were purchased by Harlan (Barcelona, Spain). DO11 (TgN(TCRDO11.10)Rag2tm1) mice carrying transgenic CD4 TCR specific for the MHC class II (I-Ad)–restricted OVA peptide aa 323–339 (The Jackson Laboratory, Bar Harbor, ME) were bred at the animal facility of Center for Applied Medical Research. Female heterozygous B6.Cg-Foxp3sf/J (scurfy mice) (purchased from The Jackson Laboratory) were bred to C57BL/6 WT male mice to generate hemizygous male B6.Cg-Foxp3sf/Y (scurfy mice). C57BL/6 WT male littermates were used as controls. Mice were maintained in pathogen-free conditions and treated according to guidelines of our institution, after study approval by the Institutional Ethical Committee.

Isolation of murine CD4⁺CD25⁺, CD4⁺CD25⁻, and CD4 T cells was performed by using isolation kits (Miltenyi Biotec, Bergisch Gladbach, Germany). CD4⁺CD25⁻CD62L^high cells were purified by cell sorting (FACSAria, BD Biosciences) and cultured in the presence of TGF-β and IL-2 (5 ng/ml and 75 U/ml IL-2) for 5 d, in the presence or absence of the indicated peptides (50 μM), and analyzed by flow cytometry for the expression of FOXP3 (eBioscience, San Diego, CA).

Purified CD4⁺ T cells were stimulated with anti-CD3 and anti-CD28 for 48 h and spin infected with retrovirus containing supernatant from Plat-E packaging cells transfected with retroviral expression plasmid KMV IRES-GFP encoding empty or myc-tagged FOXP3 (18). mRNA was purified at day 5 and the expression of the indicated genes was quantified by real-time PCR.

CD25-depleted spleen cells from BALB/c mice were stimulated in vitro with anti-mouse CD3 Ab (BD Pharmingen, San Diego, CA) in the presence or absence of purified Tregs. T cell proliferation was measured 3 d later as previously described (24). IFN-γ secretion to the culture supernatant was measured by ELISA (BD Pharmingen). Alternatively, CD4⁺CD25⁻ cells were labeled with CFSE (Invitrogen) and cultured for 4 d in the presence or absence of purified Tregs and the indicated peptide. Measurement of CD4⁺ T cell proliferation (CFSE dilution) was measured by flow cytometry. T cell proliferation and cytokine production were measured in splenocytes from BALB/c, C57BL/6, or DO11 mice as described (24).

In some experiments, T cells from DO11 mice were cocultured with bone marrow–derived dendritic cells (DCs) generated from BALB/c mice femur marrow cell cultures as previously described (25). Measurement of CD86 (GL1, BD Pharmingen) or IL-12 (p40/p70, BD Pharmingen) producing DCs was carried out by flow cytometry. Human CD4⁺CD25⁻ T cells were purified by using isolation kits (Miltenyi Biotec) and stimulated with anti-CD3/anti-CD28 beads (Dynabeads, Life Technologies) in the presence/absence of peptide FOXP3 393–403. mRNA expression and cytokine production were measured after 7 or 48 h of culture, respectively. Experiments with human samples were approved by institutional Ethical Committee (reference no. 023/2010).

Naive C57BL/6 mice (n = 6) received an i.v. administration of OVA protein (1 nmol/mouse) plus polyinosinic-polycytidylic acid [poly(I:C)] (50 μg/mouse). One group received an i.p. injection with 100 μg FOXP3 393–403 peptide daily during 4 d. Mice were killed at day 7 and splenocytes were obtained for immunological analysis. Cells producing IFN-γ were enumerated by ELISPOT assays (BD Biosciences) as described (26). For T cell responses, splenocytes were stimulated with peptide OVA (257–264, 1 μg/ml) (SIINFEKL peptide) and OVA protein (10 μg/ml).

For tumor rejection experiments, different tumor models were used. In the fist model, Hepa-129 cells (10⁶ cells/mouse) were injected s.c. in C3H/HeN mice (n = 7). Ten days later, when the tumor reached 5 mm in diameter, mice were treated orally with sorafenib (0.6 mg/kg/d) from day 10 to day 14. A group of mice was also treated with peptide FOXP3 393–403 from days 10 to 20 (100 μg/mice/d). Tumor size, presented as the average of two perpendicular diameters (millimeters), was measured at regular intervals. In the second model, TC-1 P3 (A15) cells (5 × 10⁵ cells/mouse) with a downregulated MHC class I expression (27) were injected s.c. in C57BL/6 mice (n = 6). Five days later, mice were vaccinated with 3 nmol recombinant fusion protein EDA-HPVE7 in saline (vaccine) (28). A group of vaccinated mice received also peritumoral administration of the peptide FOXP3 393–403 (a daily administration of 100 μg peptide per mouse from day 5 to day 15). Tumor-bearing mice were randomly divided into different experimental groups. Mice were sacrificed when the mean tumor diameter was >20 mm. Mice were housed in appropriated animal care facilities during the experimental period and handled following the international guidelines required for experimentation with animals. The experiments were approved by the Institutional Ethical Committee.

Normality was assessed with a Shapiro–Wilk W test. Statistical analyses were performed using parametric (Student t test, one-way ANOVA, and two-tailed paired t test) and nonparametric (Kruskal–Wallis and Mann–Whitney U) tests. For all tests a p value <0.05 was considered statistically significant. Descriptive data for continuous variables are reported as mean ± SEM. SPSS 9.0 for Windows was used for statistical analyses.

The crystal structure of the ternary complex containing the NFAT1 DNA-binding domain and the FOXP3 FKH domain bound to DNA showed that a region of FOXP3, identified as Wing1 (specifically aa 399–401 from FOXP3), inserts into the CX-E′F-fg groove of NFAT, which contains a string of positively charged residues (K664, R665, K666, and R667) (18, 29, 30). We analyzed by surface plasmon resonance the capacity of synthetic peptides from FOXP3 FKH domain to bind NFAT1 protein coated to a chip. We found that peptide FOXP3 393–403 encompassing the Wing1 domain of FOXP3, but not a mutated version of the peptide [FOXP3 393(E399A)] (control peptide) interacted with the NFAT1 RHR domain (Fig. 1A). Similarly, it was found that peptide NFAT1 659–673 (encompassing the CX-E′F-fg groove of NFAT), but not any of its mutants Ala⁶⁶⁴, Ala⁶⁶⁵, Ala⁶⁶⁶, and Ala⁶⁶⁷, was able to bind to the biotinylated peptide FOXP3 393–417 (encompassing a longer peptide containing the Wing1 domain) immobilized into a neutravidin-coated sensor chip (NLC sensor chip, Bio-Rad) (Fig. 1B) (see also Supplemental Fig. 1 representing the different interaction assays).

Once we had identified a FOXP3 peptide able to bind to the fg loop peptide of NFAT1, we analyzed whether this peptide could act as a decoy molecule to inhibit the FOXP3/NFAT1 interaction. The capacity of FOXP3 to interact with the RHR domain of NFAT1 was first demonstrated using the oxygen tunneling assay platform AlphaScreen (PerkinElmer) (Fig. 1C). Using this assay, we observed that peptide FOXP3 393–403 was able to inhibit this protein/protein interaction in a dose-dependent manner (Fig. 1D). However, as expected, interaction of FOXP3 393–403 with NFAT did not affect its capacity to bind DNA oligonucleotides containing putative NFAT binding sites (20) (Supplemental Fig. 2A). We also found that peptide FOXP3 393–403 did not affect FOXP3-DNA binding or FOXP3 dimerization or the interaction of FOXP3 with other partners such as the transcription factor Runx1 (31) (Supplemental Fig. 2B–D). Thus, we concluded that FOXP3 peptide 393–403 specifically affected the FOXP3/NFAT1 interaction. Given the highly conserved nature of the interaction, the same peptide would be expected to inhibit the interaction of FOXP3 with the other NFAT proteins as well.

FOXP3-mediated suppressor function requires its interaction with NFAT (18). We studied the capacity of peptide FOXP3 393–403 to inhibit the suppressor activity of natural Tregs in vitro. Thus, CD4⁺CD25⁺ T cells from murine splenocytes were purified to analyze their immunosuppressive activity over effector T cells activated with anti-CD3 Abs. We found that FOXP3 393–403 was able to inhibit Treg suppressive function in a dose-dependent manner, restoring the proliferation of effector T cells (responders) measured by conventional assays of thymidine incorporation in the presence of Tregs (Fig. 2A). Similar results were observed when T cell proliferation was measured by CFSE dye dilution assay (Fig. 2B).

Although NFAT was originally characterized as a transcription factor important for the expression of the IL-2 gene (32), its function has been expanded to the regulation of diverse genes both within and beyond the immune system (13, 14, 33). With special interest in this work, we highlight the role of NFAT in the expression of CD40L (34), a member of the TNF ligand superfamily that plays a critical role in the modulation of humoral and cellular immunity (35). We speculated that NFAT1/FOXP3 interaction might affect NFAT1-driven CD40L expression. Indeed, the analysis of expression of a set of genes on natural Tregs and effector CD4⁺ T cells showed that in addition to IL-2, there was a significant downregulation in CD40L mRNA expression in Tregs (Supplemental Fig. 3A). Similarly, Jurkat (FOXP3⁻) cells upregulated CD40L after anti-CD3 stimulation whereas Karpas 299 (FOXP3⁺) cells did not (Supplemental Fig. 3B). Also, retroviral transduction of FOXP3 into activated CD4⁺ T cells significantly upregulates CTLA4, CD25, or GITR expression whereas it downregulates IL-2 and also CD40L (Fig. 3A). Using the pGL3 basic luciferase reporter vector containing the −391 to +65 nt region of the CD40L promoter, including the putative NFAT binding site (pCD40L-391-Luc) (22), we found that FOXP3 inhibited the expression of luciferase driven by NFAT (Supplemental Fig. 4A). When a plasmid expressing the inhibitory FOXP3 393–403 peptide (ppepFOXP3 393–403) was also cotransfected into Jurkat cells together with the pCD40L-391-Luc and pFOXP3, we found a significant restoration of luciferase activity (Fig. 3B). These data suggest that the NFAT1/FOXP3 interaction can modulate CD40L expression on T cells, and that disruption of this interaction might have a positive effect on CD40L expression. The effect of FOXP3 393–403 peptide was also observed when we analyzed the repressor capacity of FOXP3 using a luciferase reporter plasmid containing the human IL-2 promoter driven by NFAT. Indeed, as described in previous works (18, 36), FOXP3 inhibits the NFAT-inducible IL-2 promoter activity. However, the addition of the peptide to the cells restored luciferase expression (Supplemental Fig. 4B), suggesting that peptide FOXP3 393–403 is indeed inhibiting the NFAT/FOXP3 interaction.

FOXP3 is critical for the development and function of Tregs; however, FOXP3 can be transiently expressed in effector T cells after TCR stimulation (8, 24, 37) and may affect effector T cell activation through its interaction with NFAT. Indeed, we found that CD4⁺ T cells from DO11 transgenic mice stimulated with the Ag during 24 h increased Foxp3 expression when the dose of the OVA peptide added to the culture was gradually reduced (measured as Foxp3 mRNA levels [Fig. 3C] or as Foxp3 protein [Fig. 3D]). This Foxp3 expression correlates inversely with CD40L mRNA (Fig. 3C), which increases proportionally with the dose of Ag. This dose-dependent upregulation of CD40L was also observed by flow cytometry, where as low as 0.01 μg/ml OVA peptide induced a significant increase in the number of CD4⁺CD40L⁺ cells (Fig. 3E). Importantly, we found that the number of CD40L⁺ cells induced at 0.01 μg/ml OVA peptide was significantly increased when peptide FOXP3 393–403 was added to the culture (Fig. 3F). To confirm that this CD40L upregulation induced by the peptide was indeed mediated by the inhibition of Foxp3/NFAT interaction, we repeated the experiment using CD4⁺ T cells derived from male mice lacking the FKH domain of Foxp3 (scurfy mice) and their WT male littermates. Thus, CD4⁺ T cells were purified from 3-wk-old male mice and stimulated with a suboptimal dose of anti-CD3 Ab (0.5 ng/ml) in the presence or absence of the peptide Foxp3 393–403. It was found that whereas the peptide upregulates very significantly the percentage of CD4⁺CD40L⁺ cells in the WT male littermates, no significant upregulation was observed in T cells from scurfy male mice (Fig. 3G). Interestingly, note that as opposed to WT mice, T cells from scurfy mice upregulated CD40L in response to suboptimal doses of anti-CD3 stimulation.

CD40L on the surface of the CD4⁺ T cells provides an additional signal to enhance DC maturation, upregulating the expression of costimulatory molecules and the production of IL-12 (38, 39). We found that CD4⁺CD25⁻ effector T cells from DO11 transgenic mice, activated with OVA peptide in the presence of peptide FOXP3 393–403, upregulated the expression of IL-12 in CD11c⁺ DCs and the costimulatory molecule CD86 (Fig. 4A and 4B, respectively). This enhancement in DC maturation was CD40-CD40L–dependent, because addition of anti-CD40L Abs to the cocultures inhibited both IL-12 production and CD86 upregulation on DCs (Figure 4A, 4C). In agreement with this enhancement of DC maturation induced by CD40L-expressing CD4⁺ T cells, splenocytes from DO11 mice incubated with a suboptimal dose of OVA peptide (0.01 μg/ml) increased proliferation (Fig. 5A) and produced higher levels of IL-2, IFN-γ, and IL-17 when FOXP3 393–403 was added to the cultures (Fig. 5B, 5C, and 5D, respectively). However, it was found that purified effector CD4⁺ T cells stimulated with anti-CD3 Abs in the absence of APCs induced higher levels of IL-2 (measured by ELISA and by RT-PCR; Fig. 5E and 5F, respectively), suggesting that FOXP3/NFAT1 interaction may regulate immune response at different levels.

We also studied whether disruption of FOXP3/NFAT1 interaction with the inhibitory peptide might have similar effects in human CD4⁺ effector T cells. Thus, purified CD4⁺CD25⁻ T cells from six healthy donors were stimulated with anti-CD3/CD28 for 12 h in the presence or absence of the peptide. It was found that the addition of the peptide enhanced significantly the expression of mRNA for IL-2, IL-17A, IFN-γ, CD40L, IL-6, and ZAP70 whereas the expression of IL-10 and CTLA4 was downregulated (Fig. 5G). Similar results were found when cytokines IL-2, IL-6, IFN-γ, or IL-10 were quantified by ELISA in the culture supernantant (Supplemental Fig. 4C).

Regarding IL-17, it has been described that two functional NFAT binding sites located within the minimal IL-17 promoter region are the sensors of TCR signaling in the IL-17 promoter (23). We tested whether FOXP3 393–403 was able to stimulate purified effector T cells or Tregs to produce IL-17 after TCR stimulation. It was found that either CD4⁺ effector T cells or purified Tregs stimulated by anti-CD3 produce high amounts of IL-17 in the presence of FOXP3 393–403 peptide (Fig. 5H, 5I). Using a luciferase reporter plasmid ph17p(-232)-Luc (Addgene), expressing luciferase under the control of the minimal IL-17 promoter region (23) we found that FOXP3 expression was able to reduce luciferase activity driven by NFAT. However, cotransfection with a plasmid expressing peptide FOXP3 393–403 (ppepFOXP3 393–403) was able to restore luciferase activity reduced by FOXP3 (Fig. 5J).

It has been proposed that Treg differentiation is initiated by Ag stimulation via NFAT, most likely by imposing a positive feedback loop that converts low-level stochastic expression of FOXP3 into sustained upregulation (18). To explore the relevance of FOXP3/NFAT1 interaction on Treg generation, we used the in vitro system wherein the activation of CD4⁺CD25⁻CD62L^high cells in the presence of TGF-β and IL-2 results in the de novo induction of Foxp3 expression. The addition of TGF-β and IL-2 to the cultures allowed the conversion of 30% of cells to Foxp3⁺ cells. However, the presence of FOXP3 393–403 peptide reduced the number of Foxp3⁺ cells to 17.3% (Fig. 6A). We also measured the effect of the peptide on induced Treg (iTreg) conversion of human naive CD4⁺ T cells. Human blood cord–derived CD4⁺CD25⁻ naive T cells from five donors were stimulated with IL-2 and CD3/CD28-coated beads in the presence of TGF-β and the indicated peptides to measure the percentage of CD25⁺CD127⁻ T cells (iTreg) in total CD4⁺ T cells. It was found that peptide FOXP3 393–403, but not the peptide mutant FOXP3 399A, was also able to reduce significantly the percentage of iTregs induced by TGF-β (Fig. 6B).

Downregulation of Treg suppressor activity in vivo in adults may be beneficial to enhance the immunogenicity of a vaccine (24, 40). We tested whether immunization with OVA plus poly(I:C), which induces low numbers of SIINFEKL-specific CD8⁺ T cells, might be improved by the in vivo administration of peptide FOXP3 393–403. C57BL/6 mice immunized i.v. at day 0 with 1 nmol OVA plus 50 μg poly(I:C) and treated with peptide FOXP3 393–403 from days 0 to 4 (100 μg/d i.p. in saline) had a significantly higher number of IFN-γ–producing cells in response to OVA or the SIINFEKL peptide than did mice treated with a control peptide (peptide 301–316 from gp120-VIH) or with saline alone (Fig. 6C).

We studied whether FOXP3 393–403 peptide administration would exhibit a therapeutic effect in two different tumor models. Mice were injected s.c. with 10⁶ Hepa129 tumor cells per mice and at day 10, when the tumors reached ∼5 mm in diameter, they were treated with sorafenib (a drug approved for the treatment of advanced primary liver cancer) during 5 consecutive days in combination with a peritumoral administration of saline or the peptide FOXP3 393–403 (100 μg/d from day 10 to 20). Although complete elimination of the tumor was not observed, peptide administration significantly improved the antitumor efficacy of sorafenib (Fig. 6D), delaying tumor growth.

We also tested the antitumor activity of the peptide in mice bearing TC-1 P3 (A15) tumors. In a previous work, we found that repeated administration of a fusion protein containing the extra domain A from fibronectin fused to the E7 Ag from HPV16 (EDA-HPVE7 fusion protein) was able to cure established TC-1 tumors (28). In this experiment, to evaluate a potential benefit of peptide FOXP3 393–403 in the tumor outcome after the therapeutic vaccination with EDA-HPVE7 protein (vaccine), we used the TC-1 P3 (A15) cell line with downregulated MHC class I expression, developed by Cheng et al. (27) and used as a “more difficult to cure” murine tumor model expressing E7. Mice were injected s.c. with 10⁵ cells per mouse and they were treated intratumorally at days 3 and 5 after tumor challenge with 3 nmol vaccine alone or in combination with peptide FOXP3 393–403 (100 μg/d). Whereas only one of six mice treated with the protein alone rejected the tumors (16,6%), three of six mice treated with the vaccine in combination with the inhibitory peptide were cured (Fig. 6E).

FOXP3 is a multifaceted transcription factor with a major role in the control of immune homeostasis mediated by Tregs. However, FOXP3 expression has been shown to be induced transiently on CD4⁺ T cells upon TCR stimulation, leading to hyporesponsiveness of the cell (8). This finding suggests that the induction of FOXP3 serves to shut off T cell activation, and thus development of FOXP3 inhibitors might give new opportunities to modulate T cell activation for therapeutic purposes.

FOXP3 homodimerizes and likely forms supramolecular complexes that might include hundreds of proteins that constitute the FOXP3 interactome. Many of the functions of FOXP3 are clearly regulated by the interactions with these cofactors, contributing importantly to establishment of the Treg signature and also to FOXP3 functions (reviewed in Ref. 41). Among the increasing number of transcription factors able to bind to FOXP3, we have focused on NFAT/FOXP3 interaction, which has been shown to be crucial to repress expression of IL-2, upregulate expression of the Treg markers CTLA4 and CD25, and confer suppressor function (18). The crystal structure of the NFAT/FOXP3 cooperative complex reveals that the Wing1 domain from FOXP3 interacts with the CX-E′F-fg groove of NFAT (18, 29). We have found that peptide FOXP3 393–403, encompassing the Wing1 domain, interferes with FOXP3/NFAT interaction, without affecting the capacity of NFAT to bind to its DNA binding site. Interestingly, peptide FOXP3 393–403 was able to inhibit Treg activity in vitro.

Retroviral transduction of FOXP3 converts naive T cells toward a Treg phenotype similar to that occurring in natural CD4⁺ Tregs, leading to increased expression of the Treg markers CTLA-4, CD25, GITR, and CD103 (42). We found that FOXP3 also downregulated CD40L expression, a gene under the control of NFAT activity (22, 34, 43) that plays a critical role in the modulation of humoral and cellular immunity (35, 44). CD40L expression is upregulated on naive CD4⁺ T cells via TCR stimulation in an Ag dose–dependent manner. It has been recently described that Tregs do not express CD40L (45). Interestingly, and in agreement with a previous work (37), we found that suboptimal TCR stimulation of T cells resulted in a strong induction of Foxp3 mRNA expression, which was downregulated when the Ag dose was increased. The inverse correlation between CD40L and FOXP3 might suggest a regulatory activity of FOXP3 on CD40L expression. It is conceivable that FOXP3 might control T cell activation by inhibiting CD40L expression on effector CD4⁺ T cells through its interaction with NFAT. Indeed, inhibition of NFAT/FOXP3 interaction with peptide FOXP3 393–403 improved CD40L expression on T cells and the subsequent maturation of DCs.

As occurred with CD40L, it has been shown that the IL-17 promoter is regulated by NFAT (23). The two NFAT sites located between positions −232 and −159 are sufficient for inducible promoter activity. We observed that FOXP3 expression was able to reduce significantly IL-17 promoter activity and inhibition of FOXP3/NFAT interaction–upregulated IL-17 production. The molecular mechanisms underlying the differentiation of IL-17–producing Th cells (Th17 cells) are not well understood. It has been described that optimal transcription of IL-17 required at least one enhancer sequence that bounds the transcription factors retinoic acid–related orphan receptor γt and Runx1. FOXP3 can interact with Runx1 and also with retinoic acid–related orphan receptor γt through the exon 2 of FOXP3 and negatively affect Th17 differentiation (35, 45, 46). Our data might suggest that FOXP3 interaction with NFAT1 might also have a role in the regulation of Th17 differentiation.

Inhibition of NFAT/FOXP3 interaction augmented T cell proliferation and the production of cytokines such as IL-2, IL-17A, IFN-γ, or IL-6, whereas the expression of IL-10 and CTLA4 was downregulated. These data suggest that the NFAT1/FOXP3 complex can function not only as a repressor of the IL-2 or an inducer of the CTLA4 gene as described before (47), but also as a more complex modulator of the plasticity and activation program of T cells. This finding might have special relevance in the process of effector T cell activation. As mentioned above, activation of naive T cells through TCR can activate a transient FOXP3 expression, which is strongly associated with hyporesponsiveness of T cells (8). In this sense, we have found that in vivo administration of FOXP3 393–403 peptide to mice immunized with OVA plus the TLR3 agonist poly(I:C) improved the immunogenicity of the Ag, suggesting that this peptide might have a beneficial effect in vaccination strategies.

Much evidence suggests that NFAT may control Tregs. In particular, it has been demonstrated that the generation of peripherally induced Tregs by TGF-β is highly dependent on NFAT expression (48). NFAT proteins interact with Fos-Jun (AP-1) transcription factors to form cooperative NFAT/AP-1 complexes that are critical for the induction of cytokine genes and other activation-associated genes. However, it has been demonstrated very recently that the transcription factor NFAT may promote T cell anergy and exhaustion by binding at sites that do not require cooperation with AP-1 (49). The ability to participate in multiple transcriptional complexes allows NFAT to contribute to different transcriptional programs and T cell plasticity depending on the cell type and signaling context in which it is activated (13). The potential effects of NFAT/FOXP3 inhibition in effector T cell plasticity are also evidenced in Treg conversion assays in vitro, where peptide FOXP3 393–403 significantly reduced Treg conversion of effector T cells stimulated with anti-CD3 and IL-2 in the presence of TGF-β. These results might have especial relevance in tumor microenvironment (50). In this sense, it is interesting to highlight the strong upregulation of IL-6 production after NFAT/FOXP3 inhibition because it may play an important role in FOXP3 stability through the action of the E3 ubiquitin ligase Stub1, which, in collaboration with heat shock protein 70, promotes Foxp3 ubiquitination for its proteasomal degradation (51).

Finally, although synthetic peptide t_1/2 in vivo is usually short, FOXP3 393–403 peptide administration improved the antitumor efficacy of a therapeutic vaccine in a HPV-16 E6/E7-expressing tumor model and delayed tumor growth in a murine model of hepatocellular carcinoma in combination with the kinase inhibitor sorafenib.

In summary, we have shown that FOXP3 can inhibit NFAT-driven expression of CD40L and IL-17 in CD4 T cells through its interaction with NFAT1 and that inhibition of this interaction by a short synthetic peptide can modulate effector T cell activity, improving the production of cytokines such as IL-2, IFN-γ, IL-6, or IL-17 in response to an Ag. Peptide FOXP3 393–403 inhibited natural Treg activity in vitro, reduced the conversion of effector T cells into Tregs in the presence of TGF-β, and exhibited antitumor activity in vivo. Our results suggest that targeting FOXP3/NFAT1 interaction by a peptidomimetic or by a specific small molecule represents an efficient approach to enhance T cell immunity. This strategy may be considered for the design of new therapies against cancer.

We thank Elena Ciordia and Eneko Elizalde for excellent animal care and Lourdes Ortiz, Arancha Bielsa, and Ursula Bubea for assistance in DNA sequencing. We also thank Dr. Francisco Borrás-Cuesta for lessons in science and life.

The authors have no financial conflicts of interest.