The Wiskott-Aldrich syndrome protein is required for the function of CD4⁺CD25⁺Foxp3⁺ regulatory T cells

We have previously demonstrated that WKO mice develop colitis (6) and that WKO CD4⁺ T cells are sufficient to induce disease when transferred to RAG-2–deficient mice (unpublished data). Because nTreg cells are known to control autoimmunity in several disease states (17), we hypothesized that the colitis in WKO mice may be associated with aberrant differentiation, development, or function of nTreg cells. We first assessed the numbers of regulatory T cells in lymphoid organs from WT and WKO mice. The percentage of CD4⁺CD25⁺ regulatory T cells in the WKO spleen, mesenteric lymph nodes, and thymus was significantly reduced when compared with the percentage of nTreg cells in WT lymphoid organs (Fig. 1, A and B). In addition, WKO CD4⁺ T cells had reduced expression of cell surface CD25 (Fig. 1 C).

Regulatory T cell differentiation and function is dependent on Foxp3, a forkhead winged helix transcription factor (13, 14, 24). Foxp3⁻CD4⁺CD25⁺ T cells fail to suppress proliferation in vitro (13). The numbers of WKO CD4⁺Foxp3⁺ T cells, like WKO CD4⁺CD25⁺ cells, are also decreased in peripheral lymphoid organs (Fig. 1 D). The percentage and mean fluorescence intensity of Foxp3⁺ cells among WKO CD4⁺CD25⁺ T regulatory cells is similar to WT T regulatory cells, suggesting adequate differentiation and lineage specification of WKO T regulatory cells despite their reduced percentage (Fig. 1, D and E). Cytotoxic T lymphocyte–associated antigen 4 (CTLA-4), CD103, and glucocorticoid-induced TNF receptor (GITR), like CD25, are cell surface markers that have been also associated with a regulatory CD4⁺ T cell phenotype (25–27). Expression of CTLA-4, CD103, and GITR was indistinguishable between WT and WKO CD4⁺ T cells (Fig. 1 F).

We first sought to determine whether WASP is required for nTreg cell suppression in vivo. To this end, we used a modification of the CD45RB transfer model of colitis (Fig. 2 A) (21). This model has the advantage of assessing both effector and regulatory T cell function. SCID mice were injected intraperitoneally with CD45RB^hi T cells from either WT or WKO mice alone or in combination with WT or WKO CD45RB^lo T cells (Fig. 2 A). Mice were assessed weekly for clinical signs of colitis for 8 wk, followed by histological analysis. WT CD45RB^hi T cells, when injected alone, induced severe colitis (Fig. 2 B). Colitis induction was suppressed when WT CD45RB^lo T cells were cotransferred with WT CD45RB^hi T cells (Fig. 2, B and C). In contrast, WKO CD45RB^lo T cells were unable to suppress colitis induction when cotransferred with WT CD45RB^hi T cells (Fig. 2, B and C). When CD45RB^hi T cells from WKO mice were injected alone, mice developed colitis, although associated with later onset and decreased severity (Fig. 2 B). Not surprisingly, fewer CD4⁺ T cells were recovered from the periphery of recipient mice receiving WKO CD45RB^hi T cells (not depicted). Nonetheless, cotransfer of WKO CD45RB^lo T cells could not prevent the onset of colitis in these mice. As expected, neither WT CD45RB^lo nor WKO CD45RB^lo T cells when injected alone induced colitis (Fig. 2 B). These data demonstrate that WKO CD45RB^lo T cells were unable to protect against colitis induction by WT CD45RB^hi T cells.

To determine whether the CD4⁺CD25⁺Foxp3⁺ nTreg cells within the CD45RB^lo population were responsible for the observed suppression defects, we performed analogous cotransfer experiments in which CD4⁺CD25⁺ T cells were transferred in combination with CD45RB^hi T cells (Fig. 2 A). As expected, the transfer of WT CD45RB^hi T cells alone induced colitis, a process that was blocked by the cotransfer of WT CD4⁺CD25⁺ T cells (Fig. 2 D). Neither WT nor WKO CD4⁺CD25⁺ T cells induced colitis when transferred alone. Importantly, WKO CD4⁺CD25⁺ T cells did not suppress colitis induction when cotransferred with WT CD45RB^hi T cells (Fig. 2 D). However, WKO CD4⁺CD25⁺ T cells suppressed colitis induction by WKO CD45RB^hi. These findings, although in contrast to the lack of suppression seen with WKO CD45RB^lo on WKO CD45RB^hi T cells, correlated well with the enrichment of nTreg cells within the CD4⁺CD25⁺ population (Fig. S1) and with our in vitro experiments described below.

We further assessed whether WKO CD45RB^hi T cells could receive a suppressive signal from WT CD4⁺CD25⁺ T cells. WT CD4⁺CD25⁺ T cells effectively blocked colitis induction by WKO CD45RB^hi T cells, indicating that WASP is not required for CD45RB^hi T cells to receive a suppressive signal from nTreg cells (Fig. 2 D). In all groups examined, the histopathological assessment of colitis correlated well with the clinical scores (not depicted). Collectively, these data indicate that WKO nTreg cells are less effective at suppressing colitis induction in the CD45RB transfer model.

We have previously demonstrated in vitro and in vivo that WKO mice have global defects in directed leukocyte migration (28). We hypothesized that the failure of WKO nTreg cells to suppress colitis in the CD45RB transfer model may result, at least in part, from defects in migration of WKO nTreg cells to target organs. Consistent with this hypothesis, the numbers of WKO nTreg cells recovered from the spleen, mesenteric lymph nodes, and lamina propria was reduced compared with WT nTreg cells in mice cotransferred with WT CD45RB^hi T cells (Fig. 3 A).

To assess directly whether WKO nTreg cells had defects in migration to peripheral lymph nodes, we used an in vivo migration assay. Equal numbers of WT and WKO nTreg cells were labeled with either tetramethylrhodamine isothiocyanate (TRITC) (red) or CFSE (green). Both populations were mixed at a ratio of 1:1 immediately before intravenous injection into a WT recipient mouse. To monitor the homing characteristics, recipient mice were killed 12–15 h after adoptive transfer. Analogous experiments were performed with WT and WKO CD4⁺CD25⁻ T cells. Lymphoid organs were harvested, and single cell suspensions were analyzed using flow cytometry (Fig. 3 B). There was a statistically significant decrease in the homing of WKO nTreg cells or WKO CD4⁺CD25⁻ T cells to the spleen and peripheral and mesenteric lymph node when compared with WT cells.

In light of the relative homing defects, we asked whether the defects in suppression by WKO nTreg cells in the CD45RB transfer model of colitis might be overcome by increasing the number of nTreg cells at the time of transfer. In contrast to the failure of suppression of colitis by WKO nTreg cells when transferred at a standard 1:4 ratio with WT CD45RB^hi cells, colitis was effectively suppressed when WKO nTreg cells were transferred at a higher 1:1 ratio (clinical score: 3.3, WT CD45RB^hi alone; 0.3, WT CD45RB^hi plus WT nTreg cell [low dose]; 2.3, WT CD45RB^hi plus WKO nTreg cell [low dose]; 0.0, WT CD45RB^hi plus WKO nTreg cell [high dose]; 0.2, WKO nTreg cell alone [high dose]; n = 3–5 animals/group).

The observation that transfer of increased numbers of WKO nTreg cells (relative to WT) effectively suppresses the colitis induced by the transfer of CD45RB^hi cells is consistent with the relatively inefficient homing of WKO nTreg cells to mucosal surfaces demonstrated in Fig. 3 (A and B).

We next sought to extend our in vivo findings by using an in vitro suppression assay (29). WT CD4⁺CD25⁻ T cells were cocultured with mitomycin-treated WT APCs in the presence of anti-CD3ε antibodies and WT or WKO nTreg cells. Suppression of proliferation by WT or WKO nTreg cells was assessed by [³H]thymidine incorporation. As expected, WT nTreg cells efficiently suppressed proliferation, readily apparent at a CD25⁺/CD25⁻ cell ratio of 1:16 (Fig. 4 A) (29). In contrast, even at a 1:1 CD25⁺/CD25⁻ cell ratio, WKO nTreg cells failed to suppress the proliferation of WT CD25⁻ cells (Fig. 4 A). Moreover, consistent with these findings, preliminary data suggest that WT but not WKO nTreg cells suppress the secretion of IFN-γ by WT CD25⁻ T cells in this coculture system (unpublished data).

As in our in vivo studies, WT nTreg cells inhibited the proliferation of WKO CD4⁺CD25⁻ T cells (Fig. 4 B), demonstrating that WKO responder T cells can receive a suppressive signal coming from WT nTreg cells. However, although suppressive at maximal ratios, WKO nTreg cells were less effective than WT nTreg cells at suppressing proliferation of WKO CD4⁺CD25⁻ (consistent with our in vivo data; Fig. 2 D). Suppression of WKO but not WT naive T cells by WKO nTreg cells may be explained by evidence in vitro and in vivo supporting reduced activation/proliferation of WKO effector cells. First, as shown previously with total WKO T cells (6, 7), WKO CD4⁺CD25⁻ T cells proliferate poorly upon antigen receptor activation (Fig. 4). Second, as discussed above, WKO CD45RB^hi T cells induce colitis upon transfer with lower severity than WT CD45RB^hi T cells and are associated with decreased numbers of T cells in recipient mice.

Although remodeling of the actin cytoskeleton and formation of an immunological synapse appear to be important for TCR signaling in peripheral T cells (30–32), a role for actin remodeling and immunological synapse formation has not been described for regulatory T cells. Because WASP has been demonstrated to regulate actin remodeling and immunological synapse formation, we hypothesized that deficiencies in these processes might contribute to WKO nTreg cell dysfunction (2–5). When compared with WT CD4⁺CD25⁻ cells, both WT and WKO nTreg cells polarized their actin cytoskeleton less efficiently upon stimulation with anti-CD3/anti-CD28–coated beads (Fig. 5, A–C, and Table S1). As expected, WKO CD4⁺CD25⁻ T cells polarized their actin cytoskeleton inefficiently when compared with WT CD4⁺CD25⁻ cells (2–5).

We next analyzed T cell spreading on anti-CD3/anti-CD28–coated coverslips (Fig. 5, D and E, and Table S1). As observed for polarization, we observed a decrease in the spreading of WT nTreg cells when compared with WT CD4⁺CD25⁻ cells at both 10 and 30 min after TCR stimulation (Fig. 5 E and Table S1). The percentage of spread WKO nTreg cells was similar to WT nTreg cells. Finally, the percentage of spread WKO CD4⁺CD25⁻ cells was reduced when compared with WT CD4⁺CD25⁻ cells (Fig. 5 E and Table S1). Collectively, our data suggest that WT and WKO nTreg cells remodel their actin cytoskeleton upon TCR stimulation less efficiently than CD4⁺CD25⁻ T cells.

Previous work has demonstrated that, in vitro, the resistance of CD4⁺CD25⁺ nTreg cells to TCR-induced stimulation can be overcome by the addition of exogenous IL-2 (29). We have found that, like WT nTreg cells, the inability of WKO nTreg cells to proliferate in response to TCR-induced stimulation is overcome by the addition of exogenous IL-2 (Fig. 6 A). IL-2 has also been shown to be required for efficient generation of CD4⁺CD25⁺Foxp3⁺ cells in vivo (33–35) as well as for their maximal suppressive activity (36, 37). IL-2 secretion is also required to control self-reactive T cells, as demonstrated by pronounced autoimmune pathologies developed by IL-2–deficient mice (38). Because defective IL-2 secretion is one of the hallmarks of WASP deficiency in T cells, we reasoned that WKO nTreg cell dysfunction might result from impaired IL-2–induced signaling and inadequate nTreg cell activation. To address this issue, we precultured WT or WKO CD4⁺CD25⁺ nTreg cells with 10 μg/ml of plate-bound anti-CD3ε and 25 μg/ml IL-2 for 3 d and compared suppression to freshly isolated WT or WKO nTreg cells (Fig. 6 B). As expected, IL-2 preactivation enhanced suppression by WT nTreg cells (37). IL-2 preactivation of WKO nTreg cells enhanced suppression to levels comparable with freshly isolated WT nTreg cells. Although WKO nTreg cell suppressive function improved with exogenous IL-2, the suppressive activity of preactivated WKO nTreg cells was lower than that of the preactivated WT nTreg cells, suggesting that IL-2–independent signaling defects contribute to WKO nTreg cell dysfunction. This is reminiscent of our prior data demonstrating rescue, at least in part, of antigen receptor–mediated defects in proliferation of WKO T cells through exogenous addition of IL-2 (6).

IL-10 has been shown to play a major role in the regulation of colitis in the CD45RB transfer model (22, 23, 39). We therefore analyzed IL-10 secretion by WKO T cells. WT and WKO CD4⁺CD25⁺ nTreg cells were cultured for 3 d with plate-bound anti-CD3 and IL-2 or PMA and ionomycin, and culture supernatants were harvested to assay IL-10 concentration by ELISA. IL-10 secretion was dramatically decreased upon stimulation of WKO nTreg cells (Fig. 6 C). In contrast, stimulation with PMA and ionomycin, which bypasses proximal TCR signals, rescued this defect (Fig. 6 C). This defect in IL-10 secretion was not nTreg cell specific because the same defect was observed in WKO CD4⁺CD25⁻ cells.

Autoimmunity is one of the hallmarks of the Wiskott-Aldrich syndrome, with a majority of patients suffering from at least one autoimmune disorder (hemolytic anemia, neutropenia, arthritis, skin vasculitis, glomerulonephritis, or IBD) (1). Likewise, 100% of WKO mice develop colitis by 6 mo of age (unpublished data) (6, 11). The colitis in WKO mice is unique among murine models of IBD in that it is the only model that has a human counterpart. We sought to investigate the mechanisms responsible for the immune dysregulation leading to colitis in WKO mice and postulated that abnormalities in the homeostasis/function of regulatory T cells may play a contributing role.

CD4⁺CD25⁺Foxp3⁺ regulatory T cells are key regulators of autoreactive T cells in vivo (15–19). In mice, CD4⁺CD25⁺Foxp3⁺ nTreg cells are generated in the thymus during the first weeks of life and are dependent on the expression of Foxp3 (40). We have demonstrated that WASP is important for the homeostasis of CD4⁺CD25⁺Foxp3⁺ T cells in the thymus as well as in peripheral lymphoid organs. Furthermore, WASP is critically required for the ability of CD4⁺CD25⁺ T cells to mediate suppression in vivo and in vitro. The observed defects in suppression in vivo correlate well with the demonstrated defects in homing of WKO CD4⁺CD25⁺ T cells to target organs, findings that are consistent with the known defects in global migration by WKO leukocytes (28). Nonetheless, migration defects cannot fully explain the suppression defects because these defects are still evident in the in vitro coculture system, an assay in which migratory capacity is unlikely to play a significant role.

T cell activation by APCs is associated with reorganization of the actin cytoskeleton and the concomitant formation of an immunological synapse (31, 32). In peripheral T cells, WASP plays a critical role in the reorganization of the actin cytoskeleton and immunological synapse formation (3–7). We postulated that defects in antigen receptor–induced actin remodeling and immunological synapse formation might be responsible for the regulatory T cell defects in WKO mice. However, both WT and WKO CD4⁺CD25⁺ nTreg cells poorly remodeled their actin cytoskeleton after antigen receptor stimulation when compared with WT CD4⁺CD25⁻ nTreg cells. Recently, nTreg cells and dendritic cells have been demonstrated to interact directly in vivo (41). Although our data do not exclude a role for actin remodeling and immunological synapse formation in nTreg cell function, they clearly suggest different signaling requirements for the induction of antigen receptor–induced cytoskeletal changes in nTreg cells.

Although WASP is not required for the development of CD4⁺ T cells (6, 7), the generation/homeostasis of CD4⁺CD25⁺Foxp3⁺ T cells may be particularly sensitive to WASP deficiency. Numerous studies have demonstrated a critical role for WASP in antigen receptor–induced signaling in T cells (6, 7, 42). Studies using TCR-transgenic mice suggest that WT nTreg cells are generated after recognition of self-antigens in the thymus, with nTreg cell–selective events requiring relatively high affinity interactions (43, 44). CD28 costimulation and IL-2 are also required for efficient nTreg cell generation in thymocytes, as illustrated by decreased nTreg cell numbers in CD28 knockout and IL-2 knockout mice (33, 45, 46). However IL-2 appears to have a more critical role in nTreg cell maintenance in the periphery (33). Although CD28-mediated signals have been shown to partly rescue the proliferation defects of WKO T cells (4, 6), and potential defects in CD28 costimulation in WKO nTreg cells were not assessed in this report, a complementary report suggests that CD28 signaling does not rescue the defects in proliferation by WKO nTreg cells (see Marangoni et al. on p. 369 of this issue). Defects in IL-2 production and secretion have been observed consistently in WKO T cells (2, 6, 7, 42). Exogenous IL-2 can rescue, at least in part, antigen receptor–induced signaling events in both naive WKO T cells (6, 7) and WKO nTreg cells. We have also demonstrated a reduction in CD25 (i.e., IL-2Rα) surface expression on WKO nTreg cells. Collectively, we hypothesize that defective IL-2 production and antigen receptor activation contribute to the observed decrease in numbers and function of naturally occurring nTreg cells in WKO mice (Fig. 7). This defect in antigen receptor activation may result, in part, by an increase in the activation threshold of WKO nTreg cells that is overcome by IL-2 stimulation. This is further supported by recent complementary data directly demonstrating that proliferative defects of WKO nTreg cells can be diminished with increased stimulation through the TCR (Marangoni et al., in this issue). Defective IL-2 secretion or IL-2–induced signaling events are unlikely to be solely responsible for the nTreg cell dysfunction because CD4⁺Foxp3⁺ cells from IL-2 or CD25 knockout mice are as suppressive in vitro as WT cells (33). Because Foxp3 is expressed normally in WKO CD4⁺CD25⁺ cells, our data do not support a direct role for either WASP or antigen receptor–mediated signaling events in regulating Foxp3 expression and suggest that the suppressive defects in WKO nTreg cells are downstream and/or independent of Foxp3.

IL-10 appears to be a key player in the regulation of autoimmunity in the colon. IL-10–producing T cells (Tr1) can prevent colitis induction in the CD45RB transfer model (23). However, regulatory T cells isolated from IL-10–deficient mice do not protect against colitis induced by WT CD45RB^hi cells (22). Further, injection of an anti–IL-10–neutralizing antibody together with WT CD4⁺CD45RB^lo cells abrogates suppression (22). In this report, in vitro stimulation of WKO T cells was associated with a marked defect in IL-10 secretion, suggesting a contributing role for IL-10 in WKO nTreg cell dysfunction in vivo (Fig. 7). However, preliminary data suggest that IL-10 supplementation at the time of suppression does not rescue the defects in suppression observed in vitro by WKO nTreg cells (unpublished data).

Abnormalities in nTreg cell function have been implicated in the development of autoimmunity, immunopathology, and allergy in several human diseases. The rare X-linked recessive autoimmune/inflammatory syndrome IPEX (immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome) is also associated with an autoimmune enteropathy (47). This disorder is caused by a mutation in Foxp3 and results from a profound reduction of CD4⁺CD25⁺ nTreg cells (48, 49). Our data now implicate regulatory T cell dysfunction in the pathogenesis of the immune dysregulation associated with another human disorder, the autoimmunity associated with the Wiskott-Aldrich syndrome (Fig. 7). Lending further support to this hypothesis are recent data identifying defects in suppression in human nTreg cells analogous to those described herein with murine nTreg cells (Marangoni et al., in this issue).

WKO mice (129 SvEv background) were generated as described previously (6) and backcrossed >10 generations onto the BALB/c background. C.B-17 SCID mice were purchased from Taconic. Mice were maintained in specific pathogen-free animal facilities at Massachusetts General Hospital. WKO mice on both the 129 SvEv and BALB/c background developed colitis in our facility. All experiments were conducted after approval and according to regulations of the Subcommittee on Research Animal Care at Massachusetts General Hospital.

PE- and allophycocyanin-labeled coupled anti-CD25 antibodies; FITC-conjugated or peridinin chlorophyll protein anti-CD4; FITC-conjugated anti-CD103; PE-coupled anti–CTLA-4; PE-coupled streptavidin; and PE- and FITC-coupled anti-CD45RB antibodies were purchased from BD Biosciences. Biotinylated anti-GITR antibodies were purchased from R&D Systems, and intranuclear Foxp3 staining was performed using the FITC-coupled anti-Foxp3 intracellular staining kit (eBioscience).

Flow cytometric analyses were performed on a FACSCalibur flow cytometer (Becton Dickinson) using CELLQuest software (Becton Dickinson). More than 10,000 events were recorded through a lymphocyte gate. Analyses of CD4⁺CD25⁺ cells from the thymus and lymph nodes were performed on both the BALB/c and the 129 SvEv background, and similar findings were obtained.

For the CD45RB transfer studies, mesenteric lymph nodes were carefully dissected from 10 WT or WKO BALB/c mice, and lymphocytes were isolated by gently mincing the lymph nodes through a cell strainer in RPMI 1640 supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine (all from Invitrogen), and 50 μM β-mercaptoethanol (Sigma-Aldrich). CD4⁺ cells were then negatively selected using magnetic cell sorting according to the manufacturer's recommendations (CD4⁺ isolation kit; Miltenyi Biotec). CD45RB^hi and CD45RB^lo T cells from the CD4⁺ population were isolated using a MoFlo cell sorter (DakoCytomation). CD4⁺CD25⁺ T cells were isolated from the spleen and positively selected using magnetic cell sorting (regulatory T cell isolation kit; Miltenyi Biotec).

For the in vitro suppression assay, spleens were dissected from WT or WKO BALB/c mice, and splenocytes were isolated by gently mincing the spleens through a cell strainer in FCS-supplemented RPMI. Cells were then spun down and resuspended in ACK lysis buffer to lyse erythrocytes before proceeding to cell separation. CD4⁺CD25⁺ T cells were positively selected using magnetic cell sorting as described above (purity typically ranged between 90 and 95%). CD4⁺CD25⁻ cells were isolated using the flow-through from the previous step, followed by further purification with a CD4⁺ isolation kit (Miltenyi Biotec). Purity typically was ∼95%.

APCs were isolated by incubating total WT splenocytes with an anti-Thy1.2 antibody (Cedarlane Laboratories), followed by treatment with rabbit complement (1:10 dilution, 10⁷ cells/ml; Cedarlane Laboratories) for 1 h at 37°C. Dead cells were removed using a lympholyte-M gradient (Cedarlane Laboratories). The resulting T cell–depleted splenocytes were treated with 50 μg/ml mitomycin in PBS for 20 min at 37°C and washed thoroughly before in vitro usage.

For lamina propria cell isolation, colons from WT and WKO mice were washed with PBS and digested with 2 mg/ml dispase (Invitrogen) at 37°C for 1 h in complete media (RPMI 1640 media supplemented with 4% FCS, 2 mM l-glutamine, 10 mM Hepes, 100 U/ml penicillin, 100 μg/ml streptomycin, and 250 ng/ml amphotericin) to remove the epithelial compartment including intraepithelial lymphocytes. Samples were then minced into small pieces and further digested with 1 mg/ml collagenase (Invitrogen) and 1.5 mg/ml dispase (Invitrogen) for 1 h at 37°C. After washing, cells were passed through a glass wool column and further purified via a 40/70% Percoll gradient method to remove epithelial and dead cells. Cells at the interface were isolated and washed with PBS before flow cytometric staining.

CD45RB^hi, CD45RB^lo, or CD4⁺CD25⁺ T cells were isolated from WT or WKO BALB/c mice as described above. At day 0, C.B-17 SCID mice were injected intraperitoneally with 4 × 10⁵ CD4⁺CD45RB^hi T cells alone or in combination with 2 × 10⁵ CD45RB^lo or 10⁵ CD4⁺CD25⁺ T cells. Mice were clinically assessed every week for signs of colitis. The clinical score was defined as follows: Wasting: 0, no wasting; 1, 0.1–10% loss of initial total body weight; 2, >10% loss of initial total body weight; diarrhea: 0, none; 1, soft stool; 2, watery and/or bloody; hunching/bristled fur/skin lesions: 0, unchanged; 1, any positive sign; rectum prolapse: 0, absent; 1, present. 8–12 wk after transfer, mice were killed and colons were dissected, fixed in 10% buffered formalin, and embedded in paraffin for histological analysis.

Hematoxylin and eosin staining was performed using standard techniques. Histological grading was based on mucosal thickening (0–3), lymphocyte infiltration (0–3), and crypt abscess occurrence (0–1).

Analysis of homing of WKO nTreg cells was performed as described previously (50). Single cell suspensions of spleen, mesenteric lymph node and peripheral lymph node lymphocytes were prepared from WT and WKO mice. First, 10⁷ cells (for CD4⁺CD25⁻ cells) and 10⁶ cells (for nTreg cells) of each population was labeled for 20 min at 37°C with either CFSE or TRITC (both from Invitrogen) spun down on an FBS gradient and washed three times with cold RPMI 1640 containing 10% FCS (Life Technologies). Cells were then mixed and injected in the retroorbital plexus of anesthetized WT recipient mice. Remaining cells were analyzed to determine the exact cell input. After 12–15 h, spleen, peripheral lymph nodes, and mesenteric lymph nodes from recipient mice were analyzed by FACS after gating for viable lymphocytes by forward and light scatter characteristics. The relative frequency of the two donor cell populations was determined for each individual organ, and a homing index was calculated as described previously (50). Identical results were obtained when WT and WKO cells were stained with the alternative labeling agent.

For in vitro anti-CD3ε–induced proliferation, 5 × 10⁴ CD4⁺CD25⁻ cells were cultured in 96-well round-bottom plates together with 10⁵ mitomycin-treated T cell–depleted splenocytes and 0.5 μg/ml anti-CD3ε (BD Biosciences). CD4⁺CD25⁺ cells from WT or WKO mice were added to the culture in various ratios as indicated in Fig. 4. Proliferation was measured by pulsing the cells for 12 h with 1 μCi [³H]thymidine after 60 h of stimulation. For the preactivation of nTreg cells with IL-2, splenic nTreg cells from WT or WKO mice were cultured with 10 μg/ml of plate-bound anti-CD3ε and 25 μg/ml IL-2 (both from BD Biosciences) for 3 d and washed extensively before the in vitro suppression assays.

CD4⁺CD25⁻ or CD4⁺CD25⁺ splenocytes from WT or WKO mice were isolated and cultured in the presence of 10 μg/ml of plate-bound anti-CD3ε and 25 μg/ml IL-2, or 5 ng/ml PMA and 500 ng/ml ionomycin (both from Sigma-Aldrich). After 3 d of stimulation, culture supernatants were harvested, spun down, and stored at −20°C before ELISA analysis. IL-10 concentrations were determined by ELISA as described previously (51).

10-μm latex beads were coated with 10 μg/ml anti-CD3ε and 10 μg/ml anti-CD28 (BD Biosciences) for 1 h at 37°C, washed thoroughly with PBS, and blocked with PBS/BSA 1% for 30 min. Purified T cell subsets (2 × 10⁵ cells/ml) were then incubated with the coated beads (4 × 10⁵ beads/ml) for 10 or 30 min in supplemented medium (RPMI 1640) at 37°C to create T cell–bead conjugates. The conjugates were then plated on poly-l-lysine–coated coverslips for 10 min and fixed with 4% paraformaldehyde. Cells were then permeabilized using 0.2% Triton X-100 in PBS and actin stained with AlexaFluor-488 Phalloidin (Cambrex Bio Science Inc.). Coverslips were then mounted using Vectorshield mounting medium (Vector Laboratories) and examined by fluorescence microscopy (100× objective). Cells were defined as polarized if the intensity of the actin staining was stronger on the pole of the cell in contact with the bead, whereas unpolarized cells were defined as having a uniform staining pattern.

Coverslips were coated with 10 μg/ml anti-CD3ε and 10 μg/ml anti-CD28 (BD Biosciences) for 1 h at 37°C and washed thoroughly. 10⁵ CD4⁺CD25⁻ or CD4⁺CD25⁺ cells from WT or WKO animals were then incubated on coated coverslips for the desired time and fixed, permeabilized, stained, and mounted as described above. Spread cells were defined as cells having a flattened lamellipodia-like sheath of actin, whereas unspread cells were round in shape.

Significance was assessed for nTreg cell numbers using the Mann-Whitney test. For intergroup comparisons in the CD45RB transfers as well as in the actin polarization and spreading experiments, an unpaired two-tailed Student's t test was used. p-values of <0.05 were considered significant.

Fig. S1 characterizes CD25 and Foxp3 expression of WKO CD4⁺CD45RB^lo cells. Table S1 gives the raw data of the three independent experiments (shown in Fig. 4) assessing spreading and actin polarization of WT or WKO CD4⁺CD25⁻ cells and nTreg cells.

We thank Raif Geha, Daniel K. Podolsky, Richard Blumberg, and Lisa Westerberg for review of the manuscript; Hajime Isomoto for assistance with preliminary experiments; Svend Rietdij and Cox Terhorst for assistance with the initial CD45RB transfers; and Andrew Luster, Shannon Bromley, and Rodrigo Mora for input on the homing assays. Finally, we thank Maria-Grazia Roncarolo, Loic Dupre, Sara Trifari, and Francesco Marangoni for sharing results before publication.

This work was supported by National Institutes of Health grants HL59561 (to S.B. Snapper), AI50950 (to S.B. Snapper), DK47677 (to A.K. Bhan), and DK43351 (to S.B. Snapper, A.K. Bhan, and C. Nagler), and postdoctoral support from CAPES (to V. Cotta-de-Almeida), the SICPA foundation, and Lausanne University Hospital (to M.H. Maillard).

The authors have no conflicting financial interests.