WASP regulates suppressor activity of human and murine CD4⁺CD25⁺FOXP3⁺ natural regulatory T cells

To investigate whether WASP plays a role in the function of nTreg cells, we first analyzed WASP expression and localization in mouse WT nTreg cells. nTreg cells from the spleens of C57BL/6 mice express WASP at levels comparable to that of CD4⁺CD25⁻ T cells, CD8⁺ T cells (Fig. 1 a), B cells, NK cells, and granulocytes (not depicted). We previously showed that WASP promotes actin polarization at the immunological synapse of effector T cells (38). Accordingly, after activation of WT CD4⁺CD25⁻ effector T cells with WT APCs in the presence of anti-CD3 mAbs, F-actin and WASP polarized to the APC contact site in ∼50% of the conjugates (Fig. 1, b and c). In contrast, <10% of CD4⁺CD25⁺ WT nTreg cells polarized F-actin and WASP at the contact of APCs. In most nTreg cells, a ring-like distribution of F-actin and WASP beneath the plasma membrane was observed (Fig. 1 b). The reduced polarization of F-actin and WASP in nTreg cells was not due to their defective ability to form conjugates with APCs because similar proportions of nTreg cells and effector T cells were in contact with APCs (not depicted). These data indicate that murine nTreg cells express levels of WASP similar to those of CD4⁺CD25⁻ effector T cells, but they do not efficiently polarize F-actin and WASP to the APC contact site upon TCR triggering, suggesting that WASP-dependent actin polarization at the immunological synapse does not occur in nTreg cells under these in vitro conditions.

To determine whether WASP is required for nTreg cell development, Foxp3 expression was analyzed in double negative, double positive, CD4⁺ single positive, and CD8⁺ single positive thymocytes of WAS^−/− mice. Comparable proportions of Foxp3-expressing CD4⁺ single positive and double positive thymocytes were observed in WT and WAS^−/− mice. The vast majority of CD24^loFoxp3⁺ mature thymocytes were contained in the CD4⁺ single positive T cells in both WT and WAS^−/− mice (Fig. 2 a). In addition, the overall thymic cellularity was similar in WT and WAS^−/− mice (not depicted). Moreover, the relative percentages and absolute counts of CD4⁺CD8⁻CD25⁺ T cells, as well as the mean fluorescence intensity (MFI) of CD25, were also comparable (Fig. 2, b and c). A raise in the absolute count of CD4⁺CD25⁻Foxp3⁺ and CD4⁺CD25⁺Foxp3⁺ T cells was found in the thymi of WAS^−/− mice as compared with WT controls (Fig. 2, d and e). Collectively, these data indicate that thymic differentiation of nTreg cells can efficiently occur in WAS^−/− mice.

We next investigated if WASP deficiency may influence the steady-state distribution of nTreg cells into the spleen. The overall splenocyte count was comparable in WT and WAS^−/− mice (not depicted). Normal percentages and numbers of CD4⁺CD25⁺ T cells were found in the spleens of WAS^−/− mice, although the expression of CD25 was reduced as compared with WT mice (Fig. 3, a and b). As in the thymus, the absolute count of CD4⁺CD25⁻Foxp3⁺ T cells increased in the spleens of WAS^−/− mice. Conversely, the numbers of splenic CD4⁺CD25⁺Foxp3⁺ nTreg cells were comparable to those of normal mice, whereas an increased number of splenic CD4⁺CD25⁺Foxp3⁻ T cells was observed (Fig. 3, c and d). Expression of the nTreg cell markers CTLA-4 and GITR on CD4⁺CD25⁺ splenic T cells was in the normal range, indicating that WAS^−/− nTreg cells have a normal phenotype (Fig. 3 e). WAS^−/− CD4⁺CD25⁺ T cells also expressed normal levels of the activation markers CD69, CD62L, and CD45RB (Fig. 3 e). These data indicate that splenic nTreg cells are present in WAS^−/− mice in a similar amount to those present in WT mice, albeit CD25 expression is reduced.

To investigate whether WAS^−/− nTreg cells are functional, proliferation of effector T cells in the presence of WT or WAS^−/− nTreg cells was tested. As expected, WT nTreg cells suppressed proliferation of WT CD90⁺CD25⁻ effector T cells activated with anti-CD3 mAbs in a cell dose–dependent fashion (Fig. 4 a, filled bars). In contrast, WAS^−/− nTreg cells had a reduced capacity to suppress WT effector T cell proliferation at every ratio of effector T cells to nTreg cells tested (Fig. 4 a, dotted vs. filled bars). These data indicate that nTreg cells require WASP to suppress WT effector T cells in vitro. The susceptibility of WAS^−/− effector T cells, compared with WT effector T cells, to be suppressed by WT nTreg cells (Fig. 4 a, dashed vs. filled bars) or by WAS^−/− nTreg cells (Fig. 4 a, empty vs. filled bars) was also tested. WAS^−/− effector T cells proliferated at a lower extent than WT effector cells (median proliferation was 6,360 and 26,380 cpm, respectively; P < 0.05 as determined by Mann-Whitney test). Suppression of proliferation of WAS^−/− effector T cells by WT nTreg cells was comparable to the suppression observed with WT effector T cells when an effector T cell/nTreg cell ratio of 1:0.16 was used. At lower effector T cell/nTreg cell ratios, suppression of WAS^−/− effector T cells by WT nTreg cells was significantly lower compared with that of WT effector T cells (Fig. 4 a, dashed vs. filled bars). Similar results were obtained with WAS^−/− nTreg cells (Fig. 4 a, empty vs. filled bars). These data indicate that proliferation of WAS^−/− effector T cells was not suppressed by WT or WAS^−/− nTreg cells in a cell dose–dependent manner. This may be due to intrinsic defects of WAS^−/− effector T cells in providing factors required for activation of nTreg cells.

The in vivo ability of adoptively transferred WAS^−/− and WT nTreg cells to engraft and modulate an antigen-specific immune response upon OVA immunization in WT or WAS^−/− mice was investigated. 7–10 d after OVA immunization of WT recipients, donor WT nTreg cells (transferred 1 d before immunization) represented ∼2% of the CD4⁺Foxp3⁺ (Fig. 4 b) or CD4⁺CD25⁺ cells (not depicted) in the draining LNs. In contrast, donor WAS^−/− nTreg cells could not be recovered in draining LNs (Fig. 4 b). The same results were obtained when donor nTreg cells were counted in nondraining LNs, spleen, blood, liver, and lungs regardless of CD25 and Foxp3 expression (not depicted). Accordingly, the in vitro proliferative recall responses to OVA were significantly suppressed after adoptive transfer of WT nTreg cells, but not of WAS^−/− nTreg cells (Fig. 4 c). Significantly lower percentages of CD4⁺Foxp3⁺ (Fig. 4 d) or CD4⁺CD25⁺ (not depicted) WAS^−/− nTreg cells as compared with WT nTreg cells were also recovered in WAS^−/− recipients (1 and 4%, respectively). However, both WT and WAS^−/− nTreg cells modulated the OVA-specific immune response elicited in WAS^−/− mice, as demonstrated by the in vitro recall response to OVA (Fig. 4 e). It should be noted that the overall proliferation of WAS^−/− T cells to OVA was higher than that of WT T cells, even when the WAS^−/− nTreg cells were transferred in vivo. Collectively, our results indicate that WASP-deficient nTreg cells fail to engraft in WT and displayed a reduced engraftment in WAS^−/− hosts, suggesting that this defect may contribute to their dysfunction in vivo.

Activation through the TCR is required for the suppressive function of nTreg cells (8). Therefore, it is possible that defective suppressor activity of WAS^−/− nTreg cells is due to a defect in their activation. To address this question, we measured the proliferative response and the cytokine secretion profile of WAS^−/− nTreg cells after TCR stimulation. As expected, neither WT nor WAS^−/− nTreg cells proliferated (Fig. 5 a) or produced TGF-β (Fig. 5 b) in response to anti-CD3 mAbs alone. Furthermore, in contrast to WT nTreg cells, WAS^−/− nTreg cells failed to proliferate and secrete TGF-β after activation with anti-CD3 plus anti-CD28 mAbs (Fig. 5, a and b). The addition of exogenous IL-2 to anti-CD3 or anti-CD3 plus anti-CD28 mAbs stimulation induced comparable proliferation of WT and WAS^−/− nTreg cells (Fig. 5 a) and promoted TGF-β production, although at reduced levels in WAS^−/− nTreg cells (Fig. 5 b). Neither WT nor WAS^−/− nTreg cells produced IL-2, IL-4, IL-5, IL-10, IFN-γ, and TNF-α after stimulation through the TCR/CD28 (not depicted). These data show that WAS^−/− nTreg cells have a major activation defect downstream of the TCR and the CD28 receptor, and that proliferation, but not TGF-β production, is fully restored by the addition of exogenous IL-2. To determine whether WAS^−/− nTreg cells have a higher threshold for TCR-mediated activation as compared with WT nTreg cells, we measured TCR down-regulation after stimulation with increasing doses of anti-CD3 mAbs. Both WT nTreg cells and effector T cells down-regulated CD3 in a comparable dose-dependent manner (Fig. 5 c). CD3 was also down-regulated in WAS^−/− nTreg cells and effector T cells, but higher doses of anti-CD3 mAbs were required to obtain levels of down-regulation similar to those observed in WT cells (Fig. 5 c). We next investigated whether the impaired regulatory function of WAS^−/− nTreg cells could depend on the strength of TCR stimulation. WT effector T cells were suppressed by WT nTreg cells in an anti-CD3 mAb dose-dependent manner (Fig. 5 d). On the other hand, WAS^−/− nTreg cells had a reduced suppressive ability compared with WT nTreg cells at the dose of 1 μg/ml, but the suppressive capacity was enhanced when the anti-CD3 mAb dose was increased (Fig. 5 d). However, the suppressive activity of WAS^−/− nTreg cells could never reach normal levels (Fig. 5 d). Subsequently, increasing doses of anti-CD3 mAbs were used to activate WAS^−/− effector T cells to determine whether higher levels of activation render these cells more prone to nTreg cell–mediated suppression. WAS^−/− effector T cells failed to proliferate after stimulation with 1 μg/ml anti-CD3 mAbs (not depicted). At the dose of 10 μg/ml anti-CD3 mAbs, WAS^−/− nTreg cells suppressed WAS^−/− effector T cells less efficiently than WT nTreg cells (Fig. 5 e). An increase in the anti-CD3 mAbs dose (30 μg/ml) led to comparable levels of suppression of WAS^−/− effector T cells by WT and WAS^−/− nTreg cells (Fig. 5 e), although the suppressive activity was lower than that observed with WT effector T cells. Overall, these data indicate that the suppressive defect of WAS^−/− nTreg cells can depend both on the strength of TCR stimulation and on the dysfunction of WAS^−/− effector T cells, suggesting an important role of WASP in both effector and regulatory T cells.

To our knowledge, the role of WASP in the generation and function of human nTreg cells has never been studied. As observed in murine cells, levels of WASP expression in CD4⁺CD25⁺ and CD4⁺CD25⁻ T cells from the blood of healthy donors (HDs) were comparable (not depicted). To investigate the role of WASP in the generation of human nTreg cells, we analyzed CD4⁺CD25⁺FOXP3⁺ T cells from the peripheral blood of seven WAS patients of different ages and different clinical scores (Table I). Percentages of FOXP3⁺ cells among CD4⁺ T cells of WAS patients were either higher or comparable to those observed in HDs (Fig. 6 a). Overall analysis revealed a slight increase in FOXP3⁺ cells within patients' CD4⁺ T lymphocytes, which is mainly due to an increase in CD4⁺CD25⁺FOXP3⁺ T cells (Fig. 6 b). The MFI values of CD25 and FOXP3 in CD4⁺ T cells from WAS patients were comparable to those measured in HDs (not depicted). These results indicate that WAS patients have a normal or increased proportion of circulating nTreg cells.

Next, we investigated whether nTreg cells from WAS patients were able to suppress normal allogeneic effector T cells. nTreg cells isolated from three WAS patients failed to suppress proliferation of allogeneic effector T cells (Fig. 7 a), whereas suppression of IFN-γ production was reduced in two out of three WAS patients and completely defective in the third patient (Fig. 7 b). As expected, nTreg cells isolated from HDs had significant suppressor activity toward normal allogeneic CD4⁺CD25⁻ effector T cells, resulting in reduced proliferation (Fig. 7 a) and IFN-γ secretion (Fig. 7 b). Upon stimulation, nTreg cells from WAS patients and HDs were anergic and failed to proliferate (Fig. 7 a) and to produce IFN-γ (Fig. 7 b). Analysis of all experiments performed showed that nTreg cells from WAS patients were consistently impaired in their ability to suppress proliferation of allogeneic targets (Fig. 7 c), whereas a residual suppression of IFN-γ production could be detected (Fig. 7 d). We also tested the suppressor activity of nTreg cells from WAS patients toward autologous CD4⁺CD25⁻ effector T cells. Although control nTreg cells strongly suppressed the proliferation (Fig. 8 a) and IFN-γ secretion (Fig. 8 b) of autologous effector T cells, nTreg cells from two WAS patients tested had no significant suppressive effects on cell proliferation (Fig. 8 a) and IFN-γ secretion (Fig. 8 b) by autologous effector T cells. Overall, we conclude that nTreg cells from WAS patients were impaired in suppressing both proliferation (Fig. 8 c) and IFN-γ production (Fig. 8 d) of autologous targets.

Collectively, these data demonstrate that nTreg cells from WAS patients are strongly impaired in their suppressor activity toward both autologous and allogeneic targets.

In this study, we identify WASP as a key molecule for the optimal function of human and murine nTreg cells. WASP appears to be dispensable for thymic development and steady-state distribution of nTreg cells to the periphery. However, WASP is required for the TCR/CD28-triggered proliferation, the TGF-β production, and the suppressive function of nTreg cells. Given the key role played by these cells in maintaining peripheral tolerance, our data suggest that nTreg cell dysfunction might contribute to the high susceptibility of WAS patients to develop allergy and multiple autoimmune disorders.

Normal proportions of nTreg cells were found in the thymus of WAS^−/− mice, and the induction of Foxp3 expression occurred at the expected stage of thymocyte maturation. In addition, no abnormalities in thymic architecture, including the ratio between cortical and medullary areas and the number of Hassal's corpuscles, were observed in WAS^−/− mice (not depicted). The normal thymic differentiation of WASP-deficient nTreg cells suggests that mechanisms of positive and negative selection of T cells with high-affinity TCR are preserved, including the TCR-dependent conversion of autoreactive T cells into nTreg cells (5). This hypothesis is consistent with a minor role of WASP in thymic T cell differentiation, as suggested by previous studies in C57BL/6 WAS^−/− mice (45) and in patients (46).

nTreg cells generated in the thymus undergo a robust peripheral expansion that is mainly dependent on paracrine delivery of IL-2 by activated T cells and possibly DCs (47). The IL-2–IL-2R axis appears to be dispensable during thymic development, but it is required for the survival of nTreg cells in the periphery (14, 15). Although WAS is characterized by defective IL-2 production by T cells (41–44), no reduction in the relative numbers of nTreg cells in the spleens of WAS^−/− mice was observed. These findings suggest that low levels of IL-2 in WAS^−/− mice are sufficient to provide the appropriate survival signals to peripheral nTreg cells. Although normal numbers of nTreg cells were found in the spleens of WAS^−/− mice, adoptive transfer of WAS^−/− nTreg cells into WT or WAS^−/− hosts resulted in no or reduced, respectively, recovery of those cells in draining LNs after immunization. This defective engraftment of WAS^−/− nTreg cells could be due to impaired migration, survival, and/or proliferation, and it may be even more evident in the competitive environment represented by WT hosts. In a distinct WASP-deficient murine strain, Maillard et al. (reference 48; p. 381 of this issue) found decreased numbers of nTreg cells in the spleen, suggesting that the degree of defects in nTreg cell homeostasis may depend on the genetic background of the WAS^−/− mouse model.

We showed here that murine WASP-deficient nTreg cells were impaired in their regulatory function. Indeed, in vitro experiments showed that WAS^−/− nTreg cells had a significantly reduced capacity to suppress WT effector T cell proliferation at every cell ratio tested. After strong TCR stimulation (30 μg/ml anti-CD3 mAbs), a residual suppressor activity of WAS^−/− nTreg cells on WT effector T cells could be detected, suggesting that the in vitro dysfunction of WAS^−/− nTreg cells is mainly due to an activation defect. Previous studies have shown that the suppressor activity of nTreg cells directly depends on their activation through the TCR (8). Indeed, we clearly demonstrated that WAS^−/− nTreg cells have a profound defect in TCR/CD28-mediated activation. Upon stimulation with anti-CD3 and anti-CD28 mAbs, WAS^−/− nTreg cell proliferation and TGF-β production were abolished. It remains to be determined whether the defect in TGF-β production contributes to the dysfunctional nTreg cell activity.

Given the master role of WASP in immunological synapse assembly, it is tempting to speculate that WASP contributes to the assembly of an activatory synapse in nTreg cells and that the lack of this protein prevents optimal TCR activation in nTreg cells. However, we showed that, in contrast to effector T cells, murine nTreg cells failed to polarize F-actin and concomitantly WASP to the site of contact with anti-CD3 mAb-loaded APCs or beads coated with anti-CD3 and anti-CD28 mAbs (not depicted). Therefore, a role of WASP cannot be demonstrated in this process. On the other hand, it cannot be excluded that nTreg cells assemble a structurally different immunological synapse, or that its assembly follows different kinetics. It is also possible that the activation and functional defects of WAS^−/− nTreg cells are independent from immunological synapse formation and are due to defects in signaling downstream of the TCR. The observation that defective suppressive function in WAS^−/− nTreg cells was present at a dose of TCR stimulation leading to normal TCR internalization is consistent with the hypothesis that additional defects in late TCR-mediated signaling may play a role in the activation of these cells (49).

The activation state of the effector T cells may also play a role in the suppressor activity of nTreg cells. We showed that WT nTreg cells could normally suppress the proliferation of WAS^−/− effector T cells at the effector T cell/nTreg cell ratio of 1:0.16, but at lower ratios, the suppressor activity was reduced as compared with that observed in cultures with WT effector T cells. These data suggest a lack of activatory signals by WAS^−/− effector T cells. WAS^−/− nTreg cells' failure to suppress autologous WAS^−/− effector T cells was even more profound than that of WT nTreg cells (ratio 1:1), supporting a greater dependence of the former on signals delivered by effector T cells that are required for their activation. Therefore, we propose that defective activation of WAS^−/− effector T cells, which leads to impaired production of IL-2, might further amplify the intrinsic defect of WAS^−/− nTreg cells. In line with this hypothesis, a defective cytokine production by effector T cells isolated from patients with immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome was recently proposed to contribute to the dysfunction of nTreg cells in this disease (50). We found that WAS^−/− nTreg cells were responsive to TCR/CD28 stimulation in the presence of exogenous IL-2. In a complementary study, administration of exogenous IL-2 to anti-CD3 mAb–stimulated WAS^−/− nTreg cells could strongly enhance their suppressive ability in vitro (48). Collectively, these data suggest that in WAS^−/− mice, conditions leading to non-optimal activation of effector T cells and defective IL-2 production might amplify the intrinsic activation defect of nTreg cells.

Recent data demonstrated that transfer of a non-antigen–specific WT nTreg cell population in normocytic mice was able to control antigen-specific responses (51). In our study, we compared the ability of adoptively transferred WT versus WAS^−/− nTreg cells to modulate an in vivo immune response to OVA. WAS^−/− nTreg cells transferred in WT recipients were not found in LNs draining the site of OVA immunization, and subsequently, modulation of the OVA-specific immune response was not observed. Redistribution of WAS^−/− nTreg cells transferred in autologous WAS^−/− recipients was reduced, and in this condition, down-modulation of the OVA-specific immune response was still observed. It should be noted that in the immunization protocol used here, the proliferation of WAS^−/− effector T cells upon in vitro restimulation with OVA was significantly higher compared with that observed in WT mice. This could be related to a status of spontaneous immune activation in WAS^−/− mice, as illustrated by the presence of increased numbers of CD4⁺CD25⁺Foxp3⁻ splenocytes. These conditions of potent immune reaction might therefore be permissive for suppression by WAS^−/− nTreg cells, although the recall response of cells from WAS^−/− mice treated with WAS^−/− nTreg cells was still very high. These data do not exclude that functional defects of WAS^−/− nTreg cells may arise in vivo with different antigens and in different settings, such as those presented in the complementary study by Maillard et al. (48) using a different WAS^−/− mouse strain. Indeed, WAS^−/− nTreg cells displayed a defective in vivo suppressor activity toward colitis induced by transfer of WT CD45RB^high cells. WAS^−/− nTreg cell defects may therefore be exacerbated during the development of spontaneous inflammatory bowel disease (42).

Collectively, our data suggest that in WASP deficiency the impaired suppressor activity of nTreg cells may be due to multiple defects, including intrinsic defects in activation and TGF-β production, as well as defects in other cell types, which therefore determine the extent of the impairment in homeostasis and peripheral tolerance.

Importantly, the role of WASP in regulating nTreg cell function was confirmed by our study in WAS patients. In seven WAS patients of different ages and with different disease severity, we found that the percentages of circulating CD4⁺CD25⁺FOXP3⁺ nTreg cells were within or above normal range, excluding a selective defect in nTreg cell differentiation and production. Increased frequency of FOXP3⁺ cells in the blood of two patients was attributed to an increase in the CD4⁺CD25⁻FOXP3⁺ cell population, as seen in the thymuses and spleens of WAS^−/− mice. This population has been proposed to be a pool of tissue-homing nTreg cells (9). Therefore, the increase in this cell subset might reflect a reduced ability of WAS^−/− nTreg cells to migrate to peripheral nonlymphoid organs. The in vitro suppressor activity of WASP-deficient nTreg cells on proliferation was clearly impaired, both toward allogeneic and autologous effector T cells. nTreg cells from WAS patients (two out of two) also failed to suppress IFN-γ production by autologous effectors, whereas a residual regulatory activity toward IFN-γ production by allogeneic responder cells (two out of three patients) was observed. This is in line with a previous report showing that suppression of proliferation and cytokine production by nTreg cells isolated from patients with rheumatoid arthritis could occur independently (52).

Based on the results presented in this study, we hypothesize that one possible cause of autoimmunity and atopic disease in WAS patients is a functional defect of nTreg cells. Recent studies reported a compromised function of nTreg cells in several autoimmune diseases, including immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (50), multiple sclerosis (53), autoimmune polyglandular syndrome type II (54), rheumatoid arthritis (52), myasthenia gravis (55), and type 1 diabetes (56).

The similarities between our murine and human studies (normal steady-state distribution of nTreg cells in the periphery with defective suppressor activity) validate the WAS^−/− murine model for further investigation of the contribution of nTreg cell defects to autoimmunity. It will be particularly interesting to study in this model how microbial triggers may act in concert with the functional defect in nTreg cells to lead to autoimmunity.

In conclusion, we show here that WASP is specifically required for the activation and suppressive function of CD4⁺CD25⁺FOXP3⁺ nTreg cells of both human and murine origin, but not for their differentiation. We propose that the functional failure of nTreg cells contributes to the development of autoimmunity and atopic disease in WAS.

C57BL/6 WAS^−/− mice were provided by K.A. Siminovitch (Mt. Sinai Hospital, Toronto, Canada) (43). WAS^−/−-CD45.1 mice were generated in our animal facility. No spontaneous colitis was observed in these mouse models. WT C57BL/6 and WT C57BL/6-CD45.1 control mice were purchased from Charles River Laboratories. All procedures were performed according to protocols approved by the Animal Care and Use Committee of the San Raffaele Scientific Institute.

The detailed clinical status of the patients is reported in Table I. Peripheral blood samples from WAS patients and age-matched HDs were obtained following standard ethical procedures (including informed consent) and with the approval of the San Raffaele Scientific Institute Internal Review Board. PBMCs were purified on Ficoll gradient (Nycomed Pharma A/S).

The following mAbs were used for surface staining of murine cells: anti-CD4 (RM4-5), anti-CD8 (53–6.7), anti-CD24 (M1/69), anti-CD25 (PC61), anti-CD45.2 (104), anti-CD45RB (16A), anti-CD62L (MEL14), and anti-CD69 (H1.2F3), all from BD Biosciences. Intracytoplasmic staining was performed using the following mAbs: anti-WASP (B-9; Santa Cruz Biotechnology, Inc.), anti–CTLA-4 (UC10-4F10-11; BD Biosciences), and anti-GITR (108619; R&D Systems). The staining with anti-Foxp3 mAbs (FJK-16s; eBioscience) was performed according to the manufacturer's instructions. For immunofluorescence staining, purified T cells were incubated for 20 min with Orange-CMTMR–loaded CD90⁻ splenocytes as APCs (at a 1:1 cell ratio) in absence or presence of anti-CD3 mAbs (2C11; BD Biosciences). Cells were transferred onto Poly-l-lysine–coated glass slides, fixed/permeabilized, and stained with AlexaFluor⁶³³-conjugated Phalloidin (Invitrogen) and anti-WASP antibodies (H-250; Santa Cruz Biotechnology, Inc.), followed by AlexaFluor⁴⁸⁸-conjugated goat anti–rabbit antibodies. The samples were examined on a Carl Zeiss LSM 510 confocal microscope with an x63 Plan-Apochromat objective.

Human PBMCs were stained with anti-CD4 mAbs (SK3) and anti-CD25 mAbs (2A3), all from BD Biosciences. FOXP3 expression was evaluated using PHC101 mAbs from eBioscience.

CD90⁺CD25⁻ effector T cells and CD4⁺CD25⁺ nTreg cells were purified from murine splenocytes by a combination of immunomagnetic beads (Miltenyi Biotec) and FACS sorting. Purity of CD4⁺CD25⁺ nTreg cells was ≥90% in all experiments. Cells were plated at 10⁵ cells/well in the presence of the indicated dose of plate-bound anti-CD3 mAbs (17A2; BD Biosciences) either alone or in combination with either 1 μg/ml anti-CD28 mAbs (37.51; BD Biosciences) or 100 U/ml rhIL-2 (Chiron Corporation). Cell proliferation was measured after a 96-h stimulation by liquid scintillation counting. TGF-β secretion was measured by specific ELISA (R&D Systems) on supernatants conditioned for 96 h. The Bio-Plex technology using the Mouse Th1/Th2 panel (all from Bio-Rad Laboratories) was used to measure the levels of IL-10 and IL-5 (after a 96-h stimulation), and IL-2, IL-4, IFN-γ, and TNF-α (after a 48-h stimulation) in conditioned supernatants. TCR down-regulation was evaluated by FACS analysis on splenocytes previously stained with anti-CD4 and anti-CD25 mAbs, and subsequently stimulated with the indicated dose of plate-bound anti-CD3 mAbs for 5 h.

Suppression assays using murine cells were performed as follows: CD90⁺CD25⁻ effector T cells were plated at 10⁵ cells per well and cocultured with 10⁵ (1:1 effector T cell/nTreg cell ratio), 4 × 10⁴ (1:0.4 effector T cell/nTreg cell ratio), or 1.6 × 10⁴ (1:0.16 effector T cell/nTreg cell ratio) CD4⁺CD25⁺ nTreg cells in the presence of the indicated amount of plate-bound anti-CD3 mAbs. Cell proliferation was measured after a 96-h stimulation by liquid scintillation counting. Percentage of suppression was calculated as compared with effector T cells.

For suppression assays using WAS patients' cells, CD4⁺CD25⁺ nTreg cells and CD4⁺CD25⁻ effector T cells were isolated from PBMCs by FACS sorting. Cells from HDs were used as control: CD4⁺ T cells were purified either by FACS sorting or, alternatively, by immunomagnetic beads. In both cases, the purity of nTreg cells ranged from 80 to 95%, and the purity of effector cells was ≥95%. Suppression assays with purified human nTreg cells were performed as follows: 5 × 10⁴ CD4⁺CD25⁻ effector T cells were stimulated by CD3-depleted APCs (irradiated at 6,000 rad) and 1 μg/ml of soluble anti-CD3 mAbs (Orthoclone OKT3; Janssen-Cilag). Alternatively, when very low numbers of nTreg cells were recovered, 10⁴ effector T cells were cocultured with nTreg cells and stimulated with beads coated with anti-CD3 and anti-CD28 mAbs (Dynal). For each lot of beads used, the optimal responder/bead ratio and the optimal culture time were carefully determined. Suppressive activity of nTreg cells was assessed by coculture of effector T cells with nTreg cells at different ratios. Proliferation was evaluated by liquid scintillation counting, and IFN-γ secretion was evaluated on supernatants by the human Th1/Th2 cytokine cytometric bead array system (BD Biosciences). Values of IFN-γ in the supernatants were normalized to a number of 2 × 10⁵ cells.

nTreg cells from either WT or WAS^−/− mice (carrying the CD45.2 allele) were injected i.v. (1–1.5 × 10⁶ per mouse) into WT or WAS^−/− syngeneic mice. For tracking experiments, WT-CD45.1 and WAS^−/−-CD45.1 mice were used as recipients. 1 d later, mice were immunized by footpad injection of 100 μg OVA (grade V; Sigma-Aldrich) emulsified in CFA (Sigma-Aldrich). 7–10 d later, donor-derived cells were tracked in draining and nondraining LNs, blood, spleen, lungs, and liver by staining with anti-CD45.2 mAbs. In parallel, the in vitro recall response to OVA was tested as follows: total lymphocytes isolated from draining LNs were plated at 5 × 10⁵ cells per well and cultured for 72 h with 5 μg/ml OVA. Cells were then pulsed with [³H]thymidine, and cell proliferation was measured by liquid scintillation counting.

Where data followed a normal distribution (checked by Kolmogorov-Smirnov test), experimental groups were compared with a two-tailed Student's t test. Otherwise, a two-tailed Mann-Whitney test was used. p-values of <0.05 were considered significant.

We thank A. Annoni, S. Gregori, L. Passerini, E. Hauben, and S. Valitutti for helpful discussion; A. Palini for expertise in FACS sorting; K.A. Siminovitch for providing the WAS^−/− mouse strain; and M. Ponzoni for expert analysis of histological sections. We are grateful to Lucia D. Notarangelo and G. Lefranc for providing blood samples. Finally, we thank S.B. Snapper, M. Maillard, and V. Cotta-de-Almeida for sharing results before publication.

This work was supported by grants from the Italian Telethon and Fondo per gli investimenti della Ricerca di Base (FIRB) (to M.-G. Roncarolo and L. Dupré).

The authors have no conflicting financial interests.