Defective IL-10 signaling in hyper-IgE syndrome results in impaired generation of tolerogenic dendritic cells and induced regulatory T cells

We first evaluated Th1 and Th2 cell development of naive CD4⁺ T cells in STAT3 patients. Naive CD4⁺ T cells were unstimulated or stimulated with anti-CD3 and anti-CD28 (anti-CD3/CD28) mAbs under neutral, Th1, and Th2 differentiation conditions, and the development of IFN-γ– and IL-4–producing cells was evaluated by cytoplasmic staining and flow cytometry. The development of Th1 and Th2 cells was similar in control subjects and STAT3 patients (Fig. S1 A). This observation was confirmed by ELISA of the culture supernatants of naive CD4⁺ T cells, showing similar levels of IFN-γ, IL-5, and IL-13 secretion for control subjects and STAT3 patients (Fig. S1 B). We next evaluated Th1 and Th2 cytokine production from PBMCs after stimulation with anti-CD3/CD28 mAbs. The production of IFN-γ was equivalent between control subjects and STAT3 patients, but the production of IL-5 and IL-13 was increased in STAT3 patients compared with control subjects (Fig. S1 C). These results suggest that cells in PBMCs other than naive CD4⁺ T cells are likely to be responsible for increased Th2 cytokine production in STAT3 patients.

We next evaluated the number of FOXP3⁺ T_reg cells among PBMCs because STAT3 is involved in the transduction of IL-6 and IL-21 signals, which may influence the balance between nT_reg cell and Th17 cell differentiation (Harrington et al., 2005; Bettelli et al., 2006; Ivanov et al., 2006; Veldhoen et al., 2006). Similar numbers of PBMCs were obtained from control subjects and STAT3 patients, and these cells were stained for extracellular CD4 and CD25 and intracellular FOXP3 and evaluated by flow cytometry. The percentages of CD4⁺CD25⁺ cells and CD4⁺FOXP3⁺ cells did not differ significantly between control subjects and STAT3 patients (Fig. 1 A).

We then investigated the function of T_reg cells in the peripheral blood ex vivo. CD4⁺CD25⁺CD62L⁺ T_reg cells were obtained from the peripheral blood of control subjects and STAT3 patients at a purity of >99% and were co-cultured with autologous CD4⁺CD25⁻CD62L⁺ responder T cells in the presence or absence of anti-CD3/CD28 mAbs. The addition of 1.25 × 10³ control T_reg cells to the 1.25 × 10⁴ control responder T cells resulted in levels of [³H]thymidine incorporation 55% lower than those obtained after the addition of 1.25 × 10³ control responder T cells. Levels of [³H]thymidine incorporation were similarly lowered by the addition of 1.25 × 10³ patient T_reg cells to the 1.25 × 10⁴ patient responder T cells (Fig. 1 B). Modification of the ratio of T_reg cells to responder T cells from 1:1 (T_reg cell/responder T cell) to 1:100 resulted in no significant difference in the percent suppression of [³H]thymidine incorporation between control subjects and STAT3 patients (Fig. 1 C). These results indicate that the in vivo generation and ex vivo function of T_reg cells were normal in STAT3 patients.

We next evaluated the generation of MoDCs in vitro. Isolated CD14⁺ monocytes from the PBMCs of control subjects and STAT3 patients were cultured with GM-CSF and IL-4 for 5 d and then allowed to mature in the presence of LPS for 2 d. MoDC differentiation was normal, as shown by evaluations of the forward and side light scatter of the cells (Fig. S2 A), the expression levels of CD1a (Fig. S2 B), CD80, CD83, and CD86 (Fig. S2 C), and FITC-dextran uptake (Fig. S2 D). Of note, levels of CD86 expression before LPS stimulation were slightly higher in STAT3 patients than in control subjects (Fig. S2 C), suggesting that autocrine IL-10 may regulate the expression of DC maturation markers in control subjects but not in STAT3 patients (Corinti et al., 2001).

We have previously demonstrated that STAT3 plays an important role in IL-10 signal transduction in human monocyte-derived macrophages (Minegishi et al., 2007). We therefore investigated IL-10 signal transduction in MoDCs. Consistent with our previous findings, the transcriptional up-regulation of SOCS3 and VCAN (CSPG2), two genes responsive to IL-10, was impaired in MoDCs from STAT3 patients, as demonstrated by comparison with control subjects (Fig. 2 A). Intact signal transduction was observed for TGF-β1, the other critical inhibitory cytokine, in the MoDCs of STAT3 patients (Fig. S2 E). We evaluated the effect of prior treatment with IL-10 on the phenotypic maturation of MoDCs. IL-10 was added to the culture medium on day 3 of DC differentiation. The prior treatment with IL-10 did not block the differentiation of MoDCs in response to GM-CSF and IL-4 (Fig. S3, A and B) but inhibited the LPS-induced maturation of MoDCs, with inhibition of the up-regulation of co-stimulatory molecules CD80 and CD86 and defective up-regulation of the DC maturation marker CD83 in a control subject. In contrast, the maturation of MoDCs derived from STAT3 patients showed little sign of inhibition by prior treatment with IL-10 (Fig. S3 C). Up-regulation of CD80, CD83, and CD86 expression by LPS was inhibited by IL-10 pretreatment in control subjects, but the IL-10 pretreatment failed to inhibit the up-regulation of CD80, CD83, and CD86 by LPS in STAT3 patients (Fig. 2 B). Consistent with this observation, the production of inflammatory cytokines, including TNF, IL-6, and IL-12p40, was suppressed by prior treatment with IL-10 in control subjects. The suppression by IL-10 was severely impaired in the MoDCs from STAT3 patients (Fig. 2 C). Untreated and IL-10–treated MoDCs were harvested, extensively washed, and co-cultured with third-party allogeneic naive CD4⁺ T cells from control subjects. LPS-stimulated mature MoDCs induced a significant increase in the uptake of [³H]thymidine by naive CD4⁺ T cells, with IL-10–treated DCs (IL-10–DCs) from a control subject markedly inhibiting the incorporation of [³H]thymidine. In contrast, the down-regulation was very modest in the IL-10–treated MoDCs from STAT3 patients. In the absence of MoDCs or naive CD4⁺ T cells, almost no incorporation of [³H]thymidine was detected (Fig. S4 A). Production of IFN-γ, IL-5, and IL-13 followed a very similar pattern, with prior IL-10 treatment inducing significant down-regulation in control subjects and barely detectable down-regulation in STAT3 patients (Fig. 2 D). Thus, IL-10 was defective in MoDCs from STAT3 patients, impairing suppression of (a) the up-regulation of co-stimulatory molecules on MoDCs, (b) the up-regulation of cytokine production by MoDCs, (c) the proliferation of co-cultured naive CD4⁺ T cells, and (d) cytokine production by co-cultured naive CD4⁺ T cells.

Control MoDCs up-regulated the expression of inhibitory molecules, including PD-L1, PD-L2, ILT-3, and ILT-4 but not ICOS-L by IL-10 treatment. The up-regulation of these inhibitory molecules was severely impaired in the MoDCs of STAT3 patients (Fig. 3 A). We then investigated the functional consequences of the defective up-regulation of inhibitory molecules for MoDCs by co-culturing untreated and IL-10–DCs with third-party allogeneic naive CD4⁺ T cells from control subjects. FOXP3 messenger RNA (mRNA) levels in CD4⁺ T cells co-cultured with control IL-10–DCs were approximately four times higher than those for cells co-cultured with untreated control MoDCs. However, up-regulation was severely impaired when naive CD4⁺ T cells were co-cultured with IL-10–DCs from STAT3 patients (Fig. 3 B). This observation was confirmed by the cytoplasmic staining of FOXP3 protein and flow cytometric analysis of the CD4⁺ T cells (Fig. 3 C). This up-regulation of FOXP3 was not likely to be mediated by simple T cell activation because naive CD4⁺ T cells cultured with control IL-10–DCs proliferated less vigorously compared with those with control DCs and because naive CD4⁺ T cells cultured with patient DCs proliferated more vigorously compared with those with control DCs in the absence or presence of pretreatment with IL-10 (Fig. S4 B). iT_reg cells from control subjects and STAT3 patients expressed equivalent amount of CD25 on their surface, but the expression levels of CTLA-4 and GITR (glucocorticoid-induced TNFR-related) were up-regulated by the co-culture with control IL-10–DCs, but the up-regulation was impaired by the co-culture with patient IL-10–DCs (Fig. S5).

We further evaluated the consequences of defective FOXP3 up-regulation by investigating iT_reg cell activity. Purified CD4⁺CD25⁺ T cells from the co-culture were added to autologous CD4⁺CD25⁻ responder T cells, and the mixture was stimulated with anti-CD3/CD28 mAbs. CD4⁺CD25⁺ T cells cultured with control IL-10–DCs efficiently suppressed the proliferation of CFSE-labeled autologous responder T cells (Fig. 3 D, left, second panel). The suppression of proliferation was severely impaired by CD4⁺CD25⁺ T cells cultured with patient IL-10–DCs (Fig. 3 D, left, third panel). We further evaluated cytokine production by a co-culture of responder T cells and CD4⁺CD25⁺ T cells. The production of IFN-γ, IL-5, and IL-13 was suppressed by co-culture with CD4⁺CD25⁺ cells cultured with control IL-10–DCs (Fig. 3 E). The cytokine production was rather increased by the addition of CD4⁺CD25⁺ T cells cultured with patient IL-10–DCs, which might be caused by decreased iT_reg cells and increased activated T cells in this CD4⁺CD25⁺ T cell population from STAT3 patients. Thus, the IL-10 signaling defect in MoDCs results in the impaired generation of tolerogenic DCs and iT_reg cells.

The generation of FOXP3⁺ iT_reg cells is dependent on TGF-β1 (Chen et al., 2003; Coombes et al., 2007; Rubtsov and Rudensky, 2007; Zheng et al., 2007). We therefore investigated the relationship between IL-10–DCs and TGF-β1 in the generation of FOXP3⁺ iT_reg cells. TGF-β1 in the culture medium efficiently up-regulated FOXP3 expression in naive CD4⁺ T cells in the presence of untreated immature DCs (Fig. 3 F). Control IL-10–DCs up-regulated FOXP3 expression, equivalent to TGF-β1 (Fig. 3 F, fifth dataset vs. third dataset), and a combination of control IL-10–DCs and TGF-β1 (Fig. 3 F, seventh dataset) further up-regulated FOXP3 expression. TGF-β1 effectively up-regulated FOXP3 expression in naive CD4⁺ T cells when co-cultured with patient DCs, but patient IL-10–DCs were inefficient for the up-regulation of FOXP3 expression (Fig. 3 F, sixth dataset). Thus, in addition to TGF-β1, IL-10–DCs play a crucial role in the generation of FOXP3⁺ iT_reg cells. Moreover, these results demonstrated that the defect in FOXP3 up-regulation was not caused by the lack of TGF-β1 in IL-10–DCs from STAT3 patients because the addition of exogenous TGF-β1 did not rescue this defect to the level of control IL-10–DCs in the presence of TGF-β1 (Fig. 3 F, eighth dataset vs. seventh dataset). To further clarify, we evaluated TGF-β1 and IL-10 production from MoDCs from control subjects and STAT3 patients (Fig. S6). These results indicate that the production of these inhibitory cytokines from MoDCs is not impaired in STAT3 patients.

A recent study in mice demonstrated that PD-L1 plays an important role in inducing FOXP3⁺ iT_reg cells (Keir et al., 2008; Francisco et al., 2009). We investigated whether defective PD-L1 expression in IL-10–DCs from STAT3 patients played a crucial role in the defective generation of FOXP3⁺ iT_reg cells by adding a peptide neutralizing PD-L1 to the co-culture of IL-10–DCs and naive CD4⁺ T cells. The addition of this PD-L1 peptide significantly decreased the levels of FOXP3 mRNA in the naive CD4⁺ T cells co-cultured with control IL-10–DCs (Fig. 4 A). The addition of the neutralizing peptide had no detectable effect on co-cultures with MoDCs from STAT3 patients.

We next investigated whether defective ILT-4 expression in IL-10–DCs from STAT3 patients played an important role in the defective generation of FOXP3⁺ iT_reg cells with a neutralizing mAb to the co-culture of IL-10–DCs and naive CD4⁺ T cells. The addition of anti–ILT-4 mAb significantly down-regulated the levels of FOXP3 mRNA in the naive CD4⁺ T cells co-cultured with control IL-10–DCs compared with a control mAb (Fig. 4 B). The addition of the anti–ILT-4 mAb had no significant effect on the co-culture of the naive CD4⁺ T cells with patient IL-10–DCs. Thus, in addition to PD-L1, ILT-4 up-regulation in response to IL-10 plays an important role in the generation of FOXP3⁺ iT_reg cells.

We further investigated the contribution of TGF-β1 in the up-regulation of FOXP3 by IL-10–DCs because endogenous TGF-β1 may be supplied by the DCs or from the culture medium. The addition of anti–TGF-β1 mAb significantly down-regulated the levels of FOXP3 mRNA in the naive CD4⁺ T cells co-cultured with control IL-10–DCs compared with a control mAb (Fig. 4 C). The addition of anti–TGF-β1 mAb had no significant effect on the co-culture of naive CD4⁺ T cells with patient IL-10–DCs. Thus, TGF-β1 is required for the formation of FOXP3⁺ iT_reg cells in response to control IL-10–DCs.

We also investigated whether DN-STAT3 expression in naive CD4⁺ T cells plays a significant role in the generation of iT_reg cells by evaluating the up-regulation of FOXP3 mRNA levels in naive CD4⁺ T cells from STAT3 patients. The up-regulation of FOXP3 mRNA levels in response to IL-10–DCs from STAT3 patients was impaired, but naive CD4⁺ T cells from control subjects and STAT3 patients up-regulated FOXP3 mRNA levels in response to control IL-10–DCs (Fig. 4 D). Thus, DN-STAT3 expression in MoDCs plays a major role in the impairment of FOXP3 mRNA up-regulation, and DN-STAT3 expression in T cells plays, at most, a minor role in STAT3 patients.

We next investigated the development and function of primary DCs. Two DC subsets were identified in human peripheral blood on the basis of the expression of surface molecules, including CD11c and CD304 (BDCA-4). Lineage marker (Lin) negative HLA-DR⁺CD11c⁺CD304⁻ cells are conventional DCs (cDCs), whereas Lin⁻HLA-DR⁺CD11c⁻CD304⁺ cells are plasmacytoid DCs (pDCs). The number of PBMCs obtained from the peripheral blood and the percentages of cDCs and pDCs were indistinguishable between control subjects and STAT3 patients (Fig. 5 A). We next investigated IL-10 signal transduction in primary cDCs and pDCs. The transcriptional up-regulation of SOCS3 and VCAN (CSPG2) was impaired in both subsets of primary DCs from STAT3 patients, as demonstrated by comparison with control subjects (Fig. 5, B and C). We evaluated the effect of prior treatment with IL-10 on the phenotypic maturation of primary cDCs. IL-10 was added to the culture 1 d before LPS treatment, which inhibited the LPS-induced maturation by inhibiting the up-regulation of CD83 and CD86 in control subjects. In contrast, the maturation of primary cDCs derived from STAT3 patients was almost intact by prior treatment with IL-10 (Fig. 5 D). We did not detect inhibitory effects of IL-10 on CD80 up-regulation in control subjects and STAT3 patients. Furthermore, control primary cDCs up-regulated the expression of inhibitory molecules, including PD-L1, PD-L2, ILT-3 (unpublished data), and ILT-4 by IL-10 treatment. The up-regulation of these inhibitory molecules was impaired in the primary cDCs of STAT3 patients (Fig. 5 E). These results demonstrate that IL-10 signaling is defective not only in MoDCs but also in primary DCs, resulting in the defective up-regulation of surface inhibitory molecules in STAT3 patients.

We studied MoDCs from a patient with TYK2 deficiency to confirm that the IL-10 signaling defect was responsible for the defective generation of tolerogenic DCs and iT_reg cells. The absolute numbers of cDCs and pDCs in PBMCs were similar in the TYK2-deficient patient and a control subject, and no significant difference in the differentiation of MoDCs was observed on evaluations of forward and side light scatter, CD1a expression, and the expression of CD80, CD83, and CD86 of DCs before and after LPS-induced maturation (Fig. S7, A–D). No inhibition of the up-regulation of CD80, CD83, and CD86 by prior treatment with IL-10 was detectable in cells from the TYK2-deficient patient (Fig. 6 A). The up-regulation of PD-L1, PD-L2, ILT-3, and ILT-4 on MoDCs was also defective in the TYK2-deficient patient, as shown by comparisons with control subjects (Fig. 6 B). An increase in FOXP3 mRNA and protein levels was detectable in co-cultures of allogeneic naive CD4⁺ T cells with control IL-10-DCs but not in co-cultures with TYK2-deficient IL-10–DCs (Fig. 6, C and D). Consistent with these observations, no suppression of naive CD4⁺ T cell proliferation and cytokine production (including IFN-γ, IL-5, and IL-13) was detected when TYK2-deficient IL-10–DCs were used (Fig. 6, E and F). MoDCs from the TYK2-deficient patient produced an equivalent amount of TGF-β1 and reduced amount of IL-10 compared with a control subject, which might be associated with the fact that the type I IFN signal is impaired in the TYK2-deficient patient but not in STAT3 patients.(Fig. S7, E and F). Thus, the IL-10 signaling defect in HIES patients, STAT3 patients, and the TYK2-deficient patient results in the impaired generation of tolerogenic DCs and iT_reg cells.

We found that the Th1 and Th2 differentiation of naive CD4⁺ T cells and the suppressive activity of T_reg cells were normal in STAT3 patients. Recent data have shown that Ig isotype switching in B cells is normal in STAT3 patients (Avery et al., 2010). Thus, it is not likely that T cell– and B cell–intrinsic mechanisms are responsible for the allergic manifestations in HIES patients. We then investigated DCs, which can regulate the immune response and tolerance. IL-10 signal transduction was defective in the primary DCs and MoDCs of patients, despite the intact TGF-β1 signal transduction in these cells. This defect resulted in impairment of the suppression of cytokine production and T cell proliferation by IL-10–DCs. The generation and suppressive activity of FOXP3⁺ iT_reg cells cultured with IL-10–DCs was impaired in HIES patients. The defective generation of tolerogenic DCs and iT_reg cells in response to IL-10 was also observed in the other type of HIES, TYK2 deficiency. These results suggest that IL-10 signaling in DCs may be crucial for the generation of tolerogenic DCs and iT_reg cells to maintain an appropriate Th1–Th2–T_reg cell balance in HIES patients.

The exposure of the skin to allergens induces allergen-specific unresponsiveness, possibly because of the production of IL-10 by keratinocytes (Enk and Katz, 1992; Enk et al., 1993). Langerhans cells and dermal DCs receive IL-10 signal and induce allergen-specific tolerance. This defect in IL-10–mediated tolerance to innocuous environmental antigens may be one of the mechanisms underlying the allergic signs in HIES patients. In humans, we are not certain about the nT_reg cell/iT_reg cell ratio in the peripheral blood under resting conditions. Our data suggest that most of the T_reg cells in the peripheral blood are nT_reg cells, which are derived from the thymus and are independent of the IL-10 signal. iT_reg cells in the peripheral blood may be a minor population under resting conditions but may play a crucial role in the regulation of antigen-specific allergic reactions.

Human peripheral blood T_reg cells suppressed proliferation and Th2 cytokine production by responder T cells stimulated with allergens (Bellinghausen et al., 2003; Grindebacke et al., 2004; Ling et al., 2004). CD4⁺ T cells cultured with IL-10–DCs have antigen-specific iT_reg cell activity (Steinbrink et al., 2002). In vitro experiments suggested that the suppression is dependent on cell to cell contact between iT_reg cells and responder T cells and is not mediated by soluble factors. In this study, we found that the generation of FOXP3⁺ iT_reg cells by IL-10–DCs was impaired in HIES patients. Evidence is accumulating to suggest that interactions between tolerogenic DCs and T_reg cells play an important role in the maintenance of immune tolerance against self-antigens and innocuous environmental antigens (Yamazaki et al., 2006a; Hubert et al., 2007). CD4⁺CD25⁺ T_reg cell populations can expand in the presence of DCs with intact suppressive activity in vitro and in vivo (Yamazaki et al., 2006b). In addition to the IL-10 signal provided by the cells sensing innocuous environmental antigens, the IL-10–mediated positive feedback loop between tolerogenic DCs and iT_reg cells is probably impaired in HIES patients, and this may also constitute one of the mechanisms underlying the atopic signs in HIES patients.

A large number of clinical studies have demonstrated that IL-10 is involved in the molecular pathogenesis of atopic disorders in humans. The frequency of allergen-specific, IL-10–secreting T cells is significantly higher in nonatopic individuals than in atopic patients (Akdis et al., 2004). IL-10 levels are inversely correlated with the severity of human allergic diseases (Borish et al., 1996; Lim et al., 1998). Furthermore, allergen-specific immunotherapies increase IL-10 synthesis by T cells (Francis et al., 2003; Vissers et al., 2004). All of these findings suggest that IL-10 plays a key role in the control of atopic diseases in humans.

In contrast, mice lacking IL-10 or the IL-10 receptor develop spontaneous inflammation in the large intestine (Kühn et al., 1993; Davidson et al., 1996; Spencer et al., 1998). Mice with a T_reg cell–specific IL-10 deficiency also display inflammation of surfaces in contact with the environment such as the colon, lungs, and skin (Rubtsov et al., 2008). In humans, mutations in the genes encoding IL-10 receptor subunits have been found in patients with early-onset enterocolitis (Glocker et al., 2009). Thus, a lack of IL-10 signaling results in enterocolitis in both humans and mice. Interestingly, in patients with HIES, immune responses to innocuous environmental antigens are limited to the skin, with no marked increase in the frequency of enterocolitis. One possible reason for this discrepancy is the existence of a partial, as opposed to complete, IL-10 signaling deficiency in STAT3 patients, creating a situation resembling T_reg cell–specific IL-10 deficiency. An alternative nonmutually exclusive explanation is that, in addition to the IL-10 signaling defect, STAT3 patients have defective Th17 cell development (de Beaucoudrey et al., 2008; Ma et al., 2008; Milner et al., 2008; Renner et al., 2008; Minegishi et al., 2009). The combination of Th17 cell deficiency and IL-10 signaling may result in allergic signs but prevent the development of enterocolitis (Brand, 2009).

T_reg cells mediate peripheral tolerance and play a central role in determining several immunopathologies, including autoimmunity, chronic infections, tumor development, and allergies (Hawrylowicz and O’Garra, 2005). FOXP3⁺ T_reg cells are involved in protecting humans against allergic diseases, as patients with IPEX syndrome suffer from allergic symptoms (Bennett et al., 2001; Wildin et al., 2001). PBMCs from atopic patients proliferate more extensively and produce more Th2 cytokines in response to allergens than do PBMCs from nonatopic healthy individuals (Taams et al., 2002; Ling et al., 2004). However, patients with atopic dermatitis have normal numbers of T_reg cells in the periphery with normal suppressive activity (Ou et al., 2004). These results suggest that iT_reg cells may be more important than nT_reg cells in controlling atopic dermatitis. Consistent with this hypothesis, a recent study using two mouse strains, one capable of generating iT_reg cells but incapable of generating nT_reg cells and the other unable to generate either iT_reg or nT_reg cells, suggested that iT_reg cells controlled allergic inflammation against innocuous environmental allergens, whereas nT_reg cells did not (Curotto de Lafaille et al., 2008).

TGF-β1 is the other crucial inhibitory cytokine regulating lymphocyte homeostasis, inhibiting Th1 and Th2 cell responses and promoting the differentiation of iT_reg cells (Li et al., 2006). One previous study suggested that STAT3 might be involved in transduction of the TGF-β1 signal (Ohkawara et al., 2004), but we detected no impairment of TGF-β1 signaling in DCs from STAT3 patients. Unexpectedly, we found that TGF-β1 and IL-10–DCs operated synergistically to up-regulate FOXP3 expression in naive CD4⁺ T cells. This suggests that the defective generation of IL-10–DCs may have a far-reaching impact on the induction of iT_reg cells in HIES patients.

We provide in this study the first demonstration that an IL-10 signaling defect leads to the impairment of tolerogenic DC and iT_reg cell production in the HIES. These results suggest that the defect in tolerogenic DC and iT_reg cell production, even in the presence of normal nT_reg cells, may contribute to the development of complex clinical manifestations, including allergic inflammation in HIES patients. Furthermore, a unique combination of defective Th17 differentiation and iT_reg cell generation may culminate in the development of atopic dermatitis but not enterocolitis in HIES patients.

All STAT3 patients enrolled in this study had typical clinical findings associated with HIES and a National Institutes of Health score >40 points (Table I; Grimbacher et al., 1999). The diagnosis was confirmed by the identification of mutations in the STAT3 gene. The patient with TYK2 deficiency has been described elsewhere (Minegishi et al., 2006). The study was approved by the Tokyo Medical and Dental University Ethics Committee, and written informed consent was obtained from all patients. Control individuals were nonatopic, age-matched, and equivalent in sex distribution to HIES patients. All of the patients and control subjects were in a healthy state when their blood samples were collected.

We used mAbs against CD4 (RPA-T4), CD14 (M5E2), CD11c (B-ly6), CD123 (9F5), HLA-DR (TU36), CD25 (M-A251), CD62L (Dreg 56), CD1a, (HI149), CD80 (L307.4), CD86 (2331), CD83 (HB15e), PD-L1 (MIH1), PD-L2 (MIH18), FOXP3 (259D/C7), and CTLA-4 (CD152; BNI3), a Lin cocktail (antibodies against CD3 [SK7], CD14 [MΦP9], CD16 [3G8], CD19 [SJ25C1], CD20 [L27], and CD56 [NCAM16.2]), and mAbs against IFN-γ (4S.B3) and IL-4 (8D4-8), neutralizing mAbs against IFN-γ (B27), IL-4 (MP4-25D2), and isotype-matched control mAbs, all of which were purchased from BD. We obtained antibodies against ILT-3 (CD85K; 293623), ILT-4 (CD85d; 287219), LAP (latency-associated peptide; TGF-β1; 27235), and GITR (TNFRSF18; 110416) from R&D Systems. Anti–ICOS-L antibody (MIH12) was obtained from eBioscience. Anti-CD304 (BDCA-4) antibody (AD5-17F6) was obtained from Miltenyi Biotec. Recombinant human (rh) GM-CSF, IL-4, IFN-γ, IL-10, and TGF-β1 were purchased from PeproTech. Neutralizing PD-L1 peptide was obtained from Abcam, and an irrelevant peptide was used as a negative control.

PBMCs were isolated by Ficoll density gradient centrifugation (Histopaque-1077; Sigma-Aldrich). PBMCs were cultured in 96-well plates in RPMI 1640 medium supplemented with 10% fetal bovine serum, 200 mM l-glutamine, 100 mM sodium pyruvate, nonessential amino acids, minimal essential medium vitamins (all from Invitrogen), 50 U/50 µg/ml penicillin/streptomycin (Nacalai Tesque), and 50 µM mercaptoethanol. Cultures were stimulated with a 1:100 (vol/vol) dilution of anti-CD3/CD28 mAb–coated beads from Invitrogen. For some experiments, the following mAbs and cytokines were added: 10 ng/ml rhIFN-γ, 10 ng/ml rhIL-4, and neutralizing antibodies against 10 µg/ml IFN-γ and 10 µg/ml IL-4.

Total CD4⁺ T cells were isolated with the CD4⁺ T cell isolation kit (BD). The cells were stained for sorting with antibodies against CD4, CD25, and CD62L. All mAbs were used after dialysis to remove sodium azide (Baecher-Allan et al., 2006). CD4⁺CD25⁻CD62L^hi responder T cells and CD4⁺CD25⁺CD62L^hi T_reg cell populations were isolated by sorting with a cell sorter (Moflo; Beckman Coulter). In the postsort analysis, the resulting cell preparation was found to be to >99% purity. Co-culture was set up as follows: 1.25 × 10⁴ responder T cells and 1.25 × 10³ T_reg cells were co-cultured for 5 d with a 1:100 (vol/vol) dilution of magnetic beads coated with antibodies against CD3/CD28. Responder T cells were used as a negative control. Proliferation was assessed by adding 1 µCi (37 kBq) [³H]thymidine (methyl-[³H]thymidine; ICN Biomedicals) to the culture medium for the final 18 h.

Primary DCs were obtained by the enrichment using a human DC enrichment set (BD) and cell sorting with FACS Aria II (BD): cDCs as Lin⁻HLA-DR⁺ CD11c⁺CD304⁻ cells and pDCs as Lin⁻HLA-DR⁺CD11c⁻CD304⁺ cells. In the postsort analysis, the resulting cell preparation was to >99% purity.

CD14⁺ monocytes were isolated from PBMCs with immunomagnetic beads (BD) at a purity of >98%. Monocytes were cultured in the presence of 50 ng/ml GM-CSF and 10 ng/ml IL-4 for 5 d. For differentiation into mature DCs, immature DCs were stimulated on day 5 with 100 ng/ml LPS (O55:B5; Sigma-Aldrich). For the generation of tolerogenic DCs, 100 ng/ml IL-10 was added to the culture on day 3. Nonadherent DCs on day 7 were used for T cell stimulation.

Naive CD4⁺ T cells were negatively selected from PBMCs through the depletion of CD8, CD11b, CD16, CD19, CD36, CD41a, CD45RO, CD56, CD123, γδ-TCR, and glycophorin A–positive cells, with antibody-coated paramagnetic microbeads (naive CD4⁺ T cell isolation kit from BD), according to the manufacturer’s protocol. The purity of the naive CD4⁺ T cell preparation exceeded 95%. For proliferation assays, 10⁵ naive CD4⁺ T cells were co-cultured in 96-well round-bottomed plates, in triplicate, with 10⁴ allogeneic DCs. After 5 d, the cells were pulsed with 1 µCi (37 kBq) per well of [³H]thymidine for 18 h, and [³H]thymidine incorporation was evaluated with a β counter (model 1450; PerkinElmer).

Naive CD4⁺ T cells were obtained from PBMCs with the naive CD4⁺ T cell isolation kit. We obtained CD4⁺CD25⁻ responder T cells by depleting the CD25⁺ cells with magnetic beads coated with an antibody against CD25 (BD). The resulting cell preparation was >95% pure. We obtained iT_reg cells by setting up co-cultures as described for the Allogeneic naive CD4⁺ T cell proliferation assay and purifying CD4⁺CD25⁺ cells after 3 d with immunomagnetic beads. CD4⁺CD25⁺ iT_reg cells were co-cultured with CFSE-labeled autologous CD4⁺CD25⁻ responder T cells in 96-well round-bottomed plates containing a 1:100 (vol/vol) dilution of anti-CD3/CD28 mAb beads. After 5 d, the proliferation of the CFSE-labeled CD4⁺CD25⁻ T cells was assessed by flow cytometry.

Cells were analyzed on a FACSCalibur or FACSCanto II machine (BD) using CellQuest or FACSDiva software (BD).

1 mg/ml FITC–dextran (Sigma-Aldrich) was incubated with 10⁵ cells at 37°C or 4°C for 2 h. FITC-dextran uptake was stopped by adding ice-cold PBS, and the cells were then thoroughly washed in a refrigerated centrifuge. Samples were then subjected to flow cytometry. The level of antigen uptake by DCs was assessed as the difference between the test (37°C) and control (4°C) values for each sample.

For cytokine determinations, the culture supernatant was stored at −80°C until use, and the amounts of IFN-γ, TNF, IL-5, IL-6, IL-10, IL-12p40, and IL-13 present were then determined by ELISA, according to the kit manufacturer’s instructions (BD).

Naive CD4⁺ T cells were cultured with plate-bound antibodies against CD3 and CD28 in Th1 conditions, IFN-γ plus antibody against IL-4 in Th2 conditions, or IL-4 and antibody against IFN-γ, and the cells were then fixed and permeabilized (Cytofix/Cytoperm reagents; BD) and stained with mAbs against CD4, IFN-γ, and IL-4, according to the manufacturer’s instructions (BD). CTLA-4 staining was performed after Cytofix/Cytoperm treatment.

Naive CD4⁺ T cells co-cultured with untreated DCs or IL-10–DCs were fixed and permeabilized with the human FOXP3 buffer set (BD) and stained with mAb against FOXP3.

Cells were harvested for total RNA isolation with the Fastpure RNA kit (Takara Bio Inc.). Total RNA was reverse transcribed with Primescript RT (Takara Bio Inc.). An aliquot of the RT products was used as a template for real-time PCR with SYBR green Mastermix (Takara Bio Inc.) on an Mx3005P thermocycler (Agilent Technologies) with SYBR green I dye as the amplicon detector and ROX as the passive reference. The gene for HPRT (hypoxanthine phosphoribosyltransferase) was amplified as an endogenous reference. Quantification was achieved by both the standard curve and comparative ΔΔCT methods.

Data are expressed as means ± the SD. Unpaired t tests or analysis of variance was used for statistical analysis. P-values <0.05 were considered significant (*, P < 0.05; **, P < 0.01; and ***, P < 0.001).

Fig. S1 shows normal Th1 and Th2 differentiation from naive CD4⁺ T cells but increased Th2 cytokine production from activated T cells in PBMCs of STAT3 patients. Fig. S2 shows that MoDC differentiation in vitro and TGF-β1 signaling in MoDCs are intact in STAT3 patients. Fig. S3 shows that IL-10 treatment does not impair the differentiation of MoDCs, but down-regulation of CD80, CD83, and CD86 is defective in MoDCs from STAT3 patients. Fig. S4 shows that suppression of proliferation by IL-10 pretreatment is impaired in MoDCs from STAT3 patients. Fig. S5 shows that up-regulation of FOXP3, CTLA-4, and GITR is impaired in iT_reg cells co-cultured with patient IL-10–DCs. Fig. S6 shows that MoDCs from STAT3 patients produce equivalent amounts of TGF-β1. Fig. S7 shows the characterization of primary DCs and MoDCs from the patient with TYK2 deficiency.

We thank Ms. S. Miyakoshi for assistance with cell sorting with Moflo.

This work is supported by Grants-in-Aid from the Japanese Ministry of Education, Culture, Sports, Science and Technology (22021015 and 22390205), Japan Science and Technology Agency, Core Research for Evolutional Science and Technology, Research on Intractable Diseases from the Ministry of Health, Labour and Welfare, the Uehara Foundation, the Naito Foundation, the Takeda Science Foundation, and the Mitsubishi Foundation.

The authors have no conflicting financial interests.