IL-6 receptor blockade corrects defects of XIAP-deficient regulatory T cells

X-linked lymphoproliferative syndrome type-2 (XLP-2) is a primary immunodeficiency disease attributed to XIAP mutation and is triggered by infection. Here, we show that mouse Xiap−/− regulatory T (Treg) cells and human XIAP-deficient Treg cells are defective in suppressive function. The Xiap−/− Treg cell defect is linked partly to decreased SOCS1 expression. XIAP binds SOCS1 and promotes SOCS1 stabilization. Foxp3 stability is reduced in Xiap−/− Treg cells. In addition, Xiap−/− Treg cells are prone to IFN-γ secretion. Transfer of wild-type Treg cells partly rescues infection-induced inflammation in Xiap−/− mice. Notably, inflammation-induced reprogramming of Xiap−/− Treg cells can be prevented by blockade of the IL-6 receptor (IL-6R), and a combination of anti-IL-6R and Xiap−/− Treg cells confers survival to inflammatory infection in Xiap−/− mice. Our results suggest that XLP-2 can be corrected by combination treatment with autologous iTreg (induced Treg) cells and anti-IL-6R antibody, bypassing the necessity to transduce Treg cells with XIAP.

Regulatory T (Treg) cells suppress excess T cell activation and maintain peripheral T cell tolerance. Treg cells are classified into thymus-derived CD4 + CD25 + Foxp3 + regulatory T (tTreg) cells, induced regulatory T cells (iTreg cells), and peripherally generated Treg cells (pTreg cells) 23 . The development and function of Treg cells is controlled by Foxp3 expression 24 . Foxp3 is regulated at epigenetic, transcriptional, and posttranslational levels 25 . Loss of Foxp3 during the lifetime of Treg cells ablates the suppressive activity 26,27 . Within inflammatory environments and accompanied by loss of Foxp3 expression, Treg cells can be converted into effector T cells such as T helper type 1 (Th1) [28][29][30] , Th17 31, 32 , follicular B helper T (T FH ) 33 , or Th2 cells 34 . These reprogrammed Treg cells contribute to pathology and are targets of immunotherapy 35 . Treg cell plasticity is regulated by a long list of molecules, including suppressor of cytokine signaling 1 (SOCS1), which is one of the transcription factors required to maintain the inhibitory function of these cells 36 . SOCS1 deficiency leads to over-activation of signal transducer and activator transcription 1 (STAT1) and STAT3 in Treg cells [36][37][38] , resulting in excess inflammatory signaling, loss of Foxp3 expression, and spontaneous autoimmunity 36 .
Adoptive transfer of Treg cells has been explored as a potential therapy for various autoimmune diseases, graft vs host disease, as well as transplantation rejection [39][40][41] . Autologous Treg cell numbers can be expanded ex vivo to a high number and their safe application in vivo has been confirmed 42 . A further development is the genetic modification of Treg cells for expression of a chimeric antigen receptor (CAR) 43 or antigen-specific T cell receptor (TCR) 41,44 . Similarly, defective Treg cells caused by mutations in specific genes can be corrected by re-introduction of the respective wild-type gene.
In the present study, we demonstrate that XIAP is required for the suppressive function of Treg cells. XIAP-deficient Treg cells are ineffective in inhibiting inflammation. SOCS1 expression is reduced in XIAP-deficient Treg cells. XIAP promotes SOCS1 K63 ubiquitination and maintains SOCS1 protein stability. Transfer of wild-type Treg cells partly suppresses infection-induced inflammation in Xiap −/− mice. Moreover, a combination of Xiap −/− Treg cells and anti-IL-6R corrects the vulnerability of Xiap −/− mice to infection. Our results provide evidence of a mechanism underlying the generation of XLP-2 syndromes in XIAP mutant patients. Furthermore, we demonstrate the therapeutic feasibility of combining autologous Treg cells and anti-IL-6R for the treatment of primary immunodeficiency diseases such as XLP-2.

Results
Impaired inhibitory activity of Xiap −/− Treg cells. A previous study showed that XIAP-deficiency did not affect the development of tTreg cells and that the frequency of CD4 + Foxp3 + cells was comparable between Xiap −/− mice and littermate control (WT) animals ( Fig. 1a) 15 . Generation of iTreg cells was similar between WT and Xiap −/− naive CD4 + CD25 − T cells when time and dosage of TGF-β were optimized (Fig. 1a). We also determined the expression of several Treg cells-associated molecules in isolated WT and Xiap −/− tTreg cells (Supplementary Figure 1) and found that expressions of CTLA-4, GITR, LAG3, and FR4 were comparable between WT and Xiap −/− tTreg cells (Fig. 1b). Production of IL-10 and TGF-β in Xiap −/− tTreg cells was indistinguishable from that in WT tTreg cells (Fig. 1c). Despite having normal phenotypes, Xiap −/− tTreg cells were less effective than WT tTreg cells in suppressing the proliferation of CD4 + CD25 − effector T cells (Fig. 1d). Inhibition of CD4 + CD25 − T cell activation by Xiap −/− iTreg cells was also compromised (Supplementary Figure 2). Since XIAP is an anti-apoptotic and signaling protein, the proliferation and viability of Xiap −/− tTreg cells were examined and found to be comparable with WT tTreg cells (Supplementary Figure 3). We also used human XIAPknockdown Treg cells to mimic Treg cells of XLP-2 patients (Fig. 1e). The differentiation of human XIAP-deficient iTreg cells was similar to that of control iTreg cells ( Supplementary Figure 4). However, the suppressive activity of human XIAPknockdown tTreg cells was impaired relative to control tTreg cells (Fig. 1f, Supplementary Figure 5).
Xiap −/− tTreg cells also displayed a diminished capacity to suppress T cell activation in an in vivo functional assay. Colitis was induced in Rag1 −/− mice by administration of CD4 + CD25 − T cells, leading to body weight loss (Fig. 1g), rectal prolapse and diarrhea. Tissue sections of inflamed colon revealed inflammatory infiltrate, crypt cell damage and goblet cell damage (Fig. 1h, Teff). Co-administration of WT tTreg cells effectively suppressed the induction of colitis by CD4 + CD25 − T cells. In contrast, coadministration of Xiap −/− tTreg cells did not prevent the colitisassociated pathogenesis ( Fig. 1g and h), suggesting that the absence of XIAP greatly diminished the suppressive activities of tTreg cells in vivo. Therefore, even though XIAP is not involved in the development of tTreg cells or iTreg cell differentiation, XIAP is essential for the suppressive function of Treg cells in vitro and in vivo.
XIAP interacts with SOCS1 and increases SOCS1 expression. A recent report indicated that SOCS1 is essential for Foxp3 stability and its suppressive function 36 . We examined IL-2-stimulated SOCS1 expression in WT and XIAP-deficient T cells. SOCS1 was decreased in IL-2-treated Xiap −/− T-cells (Fig. 3a). XIAP-knockdown also decreased SOCS1 levels in human iTreg cells (Fig. 3b). IL-2-induced Socs1 transcript levels in T cells were not affected by XIAP-deficiency (Supplementary Figure 6a). In contrast, the protein stability of SOCS1 was dependent on the presence of XIAP (Supplementary Figure 6b), which is supported by the enhanced SOCS1 expression under increased levels of XIAP ( Fig. 3c) and that XIAP knockdown in human iTreg cells increased cycloheximide-induced SOCS1 degradation (Supplementary Figure 6c). These results suggest that XIAP regulates SOCS1 expression by maintaining SOCS1 protein stability.
We found an association between SOCS1 and XIAP. Immunoprecipitation of SOCS1-HA brought down XIAP-FLAG (Fig. 3d), and precipitation of endogenous SOCS1 pulled down endogenous XIAP in T cells (Fig. 3e)  The fractions of splenic CD4 + Foxp3 + population from control (WT) and Xiap −/− mice were determined (tTreg). WT and Xiap −/− CD4 + CD25 − T cells were treated with anti-CD3/CD28 plus TGF-β and IL-2, and Foxp3 expression at day 5 was assessed (iTreg). Experiments were independently repeated six times. b XIAPdeficiency does not affect the Treg cells phenotype. Expressions of CTLA-4, GITR, LAG3 and FR4 in WT and Xiap −/− tTreg cells were determined. Numbers indicate mean fluorescence intensities. c Normal IL-10 and TGF-β production in Xiap −/− tTreg cells. CD4 + CD25 + cells were stimulated with anti-CD3/ CD28 and IL-2 for 96 h, before generation of IL-10 and TGF-β was determined. d Impaired in vitro suppressive activity of Xiap −/− tTreg cells. CD4 + CD25cells were incubated with anti-CD3, antigen-presenting cells, and the indicated ratios of WT and Xiap −/− CD4 + CD25 + tTreg cells. [ 3 H]thymidine incorporation was determined at 80 h. Values are mean ± SD of triplicate samples in an experiment. *P < 0.05, **P < 0.01 for unpaired t-test. Experiments were reproduced independently three times with similar outcomes. e Knockdown of XIAP in human Treg cells. The levels of XIAP in control (pLKO.3) and XIAP-knockdown human iTreg cells were determined by immunoblots. f Impaired suppressive activity in XIAP-deficient human tTreg cells. Human effector T cells (Teff) were labeled with 2 μM CFSE, activated by anti-CD3/CD28 in the presence of the indicated ratio of human control and XIAP-knockdown tTreg cells, and collected at 72 h. Proliferation was determined by gating CD4 + T cells for their CFSE intensity in flow cytometry. g, h Diminished inhibitory activity of Xiap −/− tTreg cells in vivo. CD4 + CD25 − effector T cells were administered intraperitoneally into Rag1 −/− mice with WT or Xiap −/− tTreg cells. Body weight was determined at the indicated time-points (g). Data are the mean ± SD of six mice in each group. ***P < 0.001 for two-way ANOVA. Mice were killed at day 27 and colons were removed, fixed in paraformaldehyde, sectioned, and stained with H&E. Micrographs are representative of the six mice in each group. Bar indicates 100 μm a C-terminal really interesting new gene (RING)-finger domain. Using different truncated forms of FLAG-tagged XIAP, we mapped the BIR1 domain of XIAP as being the SOCS1interacting region (Fig. 3f). For SOCS1, which comprises an Nterminus, a central Src homology 2 (SH2) domain and a Cterminal SOCS-BOX domain, we found the SH2 domain to be the XIAP-binding region (Fig. 3g).
XIAP promotes SOCS1 K63 ubiquitination. Previous reports have found that SOCS1 is associated with the Elongin B/C complex, which functions as an E3 ligase. Immunoprecipitation of overexpressed Elongin B/C brought down SOCS1-HA (Fig. 4a). Notably, co-expression of full-length XIAP-FLAG increased the association of SOCS1-HA with the Elongin B/C-Myc complex (Fig. 4a). By contrast, ΔRING-XIAP did not enhance association of SOCS1 with Elongin B/C (Fig. 4a). We also determined whether XIAP promoted SOCS1 polyubiquitination. Co-expression of XIAP enhanced the addition of WT ubiquitin or K63 ubiquitin, but not K48 ubiquitin, to SOCS1 (Fig. 4b). In an in vitro ubiquitination analysis, addition of recombinant XIAP (but not XIAPΔRF) to reaction mixtures containing ubiquitin, E1, E2 (UBC13), Elongin B/C and recombinant SOCS1 increased K63 ubiquitination of SOCS1 (Fig. 4c). Together, these results suggest that XIAP binds SOCS1 and promotes SOCS1 K63 polyubiquitination, likely contributing to the increased protein stability of SOCS1.
Enhanced generation of IL-17 was also detected in Xiap −/− tTreg cells compared to WT tTreg cells after being co-stimulated with TCR and IL-6/IL-1 (Fig. 5f). We further examined the cytokine profile of tTreg cells isolated from Rag1 −/− mice after induction of colitis by effector T cells. The recovered CD45.2 + T cells, representing the transferred tTreg cells, were assessed for expression of IFN-γ and IL-17A ( Fig. 5g and h). Increased IFN-γ and IL-17A expression was observed in Xiap −/− tTreg cells isolated from Rag1 −/− mice, illustrating the enhanced plasticity of Xiap −/− tTreg cells in vivo.
Increased STAT1 and STAT3 activation in Xiap −/− Treg cells. Inflammatory cytokines including IL-12, IL-1, and IL-6 trigger JAK-STAT signaling in T cells. SOCS1 is a negative regulator of JAK-STAT signaling, so we determined whether the reduced SOCS1 in Xiap −/− iTreg cells led to an enhanced response to inflammatory cytokines. WT and Xiap −/− iTreg cells were stimulated with IL-12 or IL-1 plus IL-6, and we found increased phosphorylation of STAT1 or STAT3 in Xiap −/− iTreg cells relative to WT iTreg cells, in the context of comparable levels of STAT1 and STAT3 ( Supplementary Figure 7a and b). Therefore, XIAP deficiency leads to enhanced activation of Treg cells in  Figure 9f). Altogether, these results suggest that SOCS1 is one of the major functional targets of XIAP in Treg cells and that re-introduction of SOCS1 corrects the functional defects and plasticity of Xiap −/− tTreg cells.
To determine whether impaired Treg cells activity contributed to the sensitivity of Xiap −/− mice to infections, WT iTreg cells were transferred into Xiap −/− mice after C. albicans infection. Administration of WT iTreg cells 2 days after low-dose C. albicans infection partly rescued the survival of Xiap −/− mice; 60% of infected Xiap −/− mice that received WT iTreg cells survived 40 days after infection, whereas all untreated Xiap −/− mice had died by 18 days post-infection (Fig. 6b). WT iTreg cells transfer also alleviated kidney inflammation in infected Xiap −/− mice, as demonstrated by diminished kidney neutrophil infiltration (Fig. 6c). In addition, the elevated levels of serum inflammatory cytokines in untreated infected Xiap −/− mice were profoundly suppressed by adoptive transfer of WT iTreg cells (Fig. 6d). These results suggest that deficient inflammation control in Xiap −/− mice is partly due to impaired Treg cells functioning and that XIAPintact iTreg cells restore the ability of Xiap −/− mice to respond to infection-induced inflammation. Anti-IL-6R rescues defects in Xiap −/− Treg cells. Among the serum inflammatory cytokines analyzed in C. albicans-infected Xiap −/− mice, the high levels of IL-6 were particularly noticeable (Fig. 6d). Therefore, we examined whether the impaired functioning of Xiap −/− Treg cells was associated with IL-6. IL-6 is generated by many different types of cells. Co-stimulation with IL-12 also induced production of IL-6 from human XIAPdeficient iTreg cells (Supplementary Figure 10). Inclusion of anti-IL-6R effectively suppressed TCR/CD28-induced IFN-γ production in Xiap −/− iTreg cells with or without IL-12 (Fig. 7a), and reduced the fraction of IFN-γ-expressing Xiap −/− iTreg cells activated through TCR and IL-12 (Fig. 7b). The effectiveness of IL-6R blockage in inhibiting conversion into IFN-γ + tTreg cells was also confirmed in Foxp3-GFP-tagged Xiap −/− tTreg cells (Supplementary Figure 11). In addition, inclusion of anti-IL-6R decreased the expression and secretion of IL-6 in Xiap −/− Treg cells induced by IL-12 and/or TCR/CD28 ( Supplementary Figure 12). Inclusion of anti-IL-6R also inhibited the expression and secretion of IFN-γ in human control and XIAP-knockdown iTreg cells ( Fig. 7c and d). The production of IL-17 and IL-6 in human control and XIAP-deficient iTreg cells were similarly suppressed by anti-IL-6R (Supplementary Figure 13). Thus, anti-IL-6R effectively suppresses IFN-γ, IL-6, or IL-17 generation in XIAPdeficient Treg cells stimulated by TCR/CD28 with IL-12 or IL-6.
We also examined whether blocking other inflammatory cytokines, such as TNF or IL-1β, reversed the plasticity of Xiap −/− iTreg cells. Both anti-TNF and anti-IL-1R failed to antagonize activation-or IL-12-induced IFN-γ secretion in Xiap  Figure 14). Furthermore, inclusion of anti-IFN-γ did not effectively prevent IL-12-stimulated generation of IFN-γ in Xiap −/− iTreg cells (Supplementary Figure 15a), and exhibited no effect on the production of IL-17 promoted by IL-1 and IL-6 (Supplementary Figure 15b). Of the four antiinflammatory antibodies we examined, anti-IL-6R was the most effective in preventing production of IFN-γ in XIAP-deficient Treg cells.
Anti-IL-6R restores the function of Xiap −/− Treg cells. We further examined whether a combination of anti-IL-6R and Xiap −/− Treg cells restored the resistance of Xiap −/− mice to infectioninduced inflammation and lethality. Xiap −/− mice were infected with a lethal dose of C. albicans that did not affect WT mice. Xiap −/− iTreg cells were administered 2 days after infection, with or without by anti-IL-6R. Either Xiap −/− iTreg cells or anti-IL-6R partly increased the survival of Xiap −/− mice, though the increase was not statistically significant (Fig. 8a). Combinatory treatment of anti-IL-6R and Xiap −/− iTreg cells effectively rescued Xiap −/− mice from inflammatory infection (Fig. 8a). The elevated serum IL-6 levels in untreated infected Xiap −/− mice were suppressed by the combination of Xiap −/− iTreg cells and anti-IL-6R (Fig. 8b). We further traced the fate of CD45.1 + Xiap −/− iTreg cells after their transfer into infected CD45.2 + Xiap −/− mice. CD45.1 + Xiap −/− iTreg cells became inflammatory, with abundant IFN-γ + expression in Xiap −/− mice pre-infected with C. albicans (Fig. 8c  and d). The administration of anti-IL-6R inhibited the conversion of CD45.1 + Xiap −/− Treg cells into IFN-γ + Foxp3 + cells (Fig. 8c and d). The additive therapeutic effect of anti-IL-6R application is demonstrated by the fact that a combination of anti-IL-6R and Xiap −/− Treg cells reduced kidney neutrophil infiltration more effectively than either treatment alone (Fig. 8e). Notably, a significant decrease in kidney fungal T cells were isolated from spleen, and the frequencies of CD4 + Foxp3 + , Foxp3 + IFN-γ + and Foxp3 + IL-17 + cells were quantitated by intracellular staining. b Transfer of WT iTreg cells partially rescues Xiap −/− mice from C. albicans-induced lethality. WT and Xiap −/− female mice were intravenously administered with C. albicans (1 × 10 5 ) and a group of Xiap −/− mice also received WT iTreg cells (1 × 10 6 ) at day 2. Survival of mice was monitored and is presented as a Kaplan-Meier survival curve (n = 7 for each group). **P < 0.01. c Reduced kidney inflammation in C. albicans-infected Xiap −/− mice with WT iTreg cells transfer. Kidney was isolated from Xiap −/− mice 10 days after C. albicans injection and infiltrated neutrophil contents were determined. d Reduced inflammatory cytokine production in C. albicans-infected Xiap −/− mice into which WT iTreg cells had been transferred. Serum from mice was collected at the indicated time-points after C. albicans injection, and the levels of IL-6, TNF, MCP-1, and IL-12 were determined, n = 6. Values (a, d) are mean ± SD of samples. *P < 0.05, **P < 0.01, ***P < 0.001 for unpaired t-test titers was detected for Xiap −/− Treg cells treatment, with or without anti-IL-6R (Fig. 8f). We further investigated whether the presence of anti-IL-6R dampened the T helper 17 responses, since Th17 plays a prominent role in resolving C. albicans infection [46][47][48] . Figure 8g illustrates that IL-17 production by T cells isolated from infected mice was not affected by anti-IL-6R or anti-IL-6R plus Xiap −/− Treg cells. These results suggest that inhibition of Treg cells re-programming by anti-IL-6R significantly increased the functioning of XIAP-deficient Treg cells in vivo. Moreover, we have illustrated, likely for the first time, that in conjunction with anti-IL-6R, functionally unstable Treg cells could be used to treat inflammatory diseases.

Discussion
XIAP is a caspase-binding anti-apoptotic protein critical for cell survival. Although XIAP-knockout did not affect Treg cells development, the expression of Treg cells markers or the production by Treg cells of IL-10 and TGF-β, XIAP-deficient Treg cells were defective in their suppressive activity both in vitro and in vivo (Fig. 1). We found that SOCS1 was specifically reduced in XIAP-deficient T cells (Fig. 3). We also demonstrate that XIAP binds to SOCS1 and increases the protein stability of SOCS1, likely by promoting its K63 ubiquitination (Fig. 4). In contrast, K48 ubiquitination of SOCS1 was not affected by XIAP. Since SOCS1 is critical to maintaining the inhibitory activity of Treg cells 36 , XIAP-deficient Treg cells exhibited a compromised suppressive ability. We further demonstrate that re-introduction of SOCS1 corrects defects of Xiap −/− Treg cells (Supplementary Figure 9). Thus, the impaired Treg cells functioning in Xiap −/− Treg cells could be partly attributed to a downregulation of SOCS1. Our results reveal an unexpected requirement for XIAP in Treg cells.
We have identified two different defects that account for the diminished suppressive activity in Xiap −/− Treg cells. The stability A more profound defect of XIAP-deficient Treg cells was the increased conversion into IFN-γ-, IL-6-, and IL-17-producing Treg cells (Fig. 5). IL-12 is known to suppress Treg cells and converts Treg cells into Th1-type cells 52 . Deletion of SOCS1 elicits excess activation of the IFN-γ-STAT1 axis 36,37 . Consistent with these observations, we found that IL-12 enhances WT and XIAP-deficient Treg cells to produce IFN-γ (Fig. 5). Previous studies have shown that pathogenic Treg cells lose Foxp3 expression and produce IFN-γ 28,53 . IFN-γ promotes tissue inflammation, and IFN-γ-expressing Xiap −/− Treg cells likely albicans titres (f) were quantitated. *P < 0.05, **P < 0.01 for unpaired t-test. n.s., not significant. g Anti-IL-6R does not affect IL-17 production. Male CD45.1 + Xiap −/− mice were infected with C. albicans and treated with anti-IL-6R and anti-IL-6R plus CD45.2 + KO iTreg cells. Spleen T cells and B cells were isolated 7 days after treatments. CD45.2 + iTreg cells were depleted. T cells were incubated with B cells and heat-killed C. albicans (HKCA) for 4 days, and the production of IL-17A after stimulation with TPA/A23187 (10/100 ng ml −1 ) was determined by ELISA 47,48 . Values are mean ± SD of triplicate samples in an experiment. Experiments were independently repeated three times with similar results contribute to inflammatory pathology. XIAP-deficiency also conferred on Treg cells an enhanced production of IL-6 and IL-17 (Fig. 5f, Supplementary Figures 10, 12, 13). The increase in IFN-γ generation is more pronounced than for enhancement of IL-17 expression in XIAP-deficient Treg cells in vitro (Fig. 5c vs. 5f, Fig. 5d vs. Supplementary Figure 13b) and in vivo ( Fig. 5g and h), suggesting that XIAP-deficiency preferentially re-programs the development of Treg cells into Th1-like cells. IL-2 inhibits Th17 differentiation 54 , whether enhanced IL-2 responses in Xiap −/− Treg cells contributes to preferential Th1 over Th17 reprogramming is currently being investigated. The critical role of SOCS1 in maintaining Treg stability has been demonstrated 36,55 . In the present study, the defective activity of Xiap −/− Treg cells could be partly attributed to a loss of SOCS1. Overexpression of SOCS1 in Xiap −/− Treg cells prevented their conversion into IFN-γ-expressing Treg cells and restored the stability of Foxp3 (Supplemental Figure 9). We also demonstrate that XIAP increases the stability of SOCS1. SOCS1 is an E3 ligase known to target signaling molecules, such as VAV1 and Jak2, for ubiquitination and degradation. Our results indicate that, in the absence of XIAP, SOCS1 protein itself is subjected to degradation. Together with the capacity of XIAP to promote SOCS1 K63 polyubiquitination, our results suggest that ubiquitination of SOCS1 likely represents a new level of regulation responsible for the stability and function of the SOCS1 E3 ligase.
Despite the similar susceptibility between Socs1 f/f Lck Cre Treg cells 36 and Xiap −/− Treg cells to generate IFN-γ and their defective abilities to inhibit colitis ( Fig. 1g and h), Treg-specificknockout of SOCS1 (Socs1 f/f Foxp3 Cre ) generates Treg cells that can still suppress colitis in Rag2 −/− mice 55 . It has been suggested that environmental inflammatory cues prime Socs1 −/− Treg cells to become IFN-γ-producing cells 55 . We have previously shown that myeloid components in Xiap −/− mice become inflammatory upon infection 15 , suggesting a possibility that Xiap −/− myeloid cells prime Xiap −/− Treg cells for plasticity. We plan to investigate whether the conversion of Xiap −/− Treg cells into inflammatory ex-Treg cells involves conditioning from Xiap −/− myeloid cells using mice with Treg-specific knockout of XIAP.
Our results also illustrate that defects in XIAP-deficient Treg cells are manifested by inflammatory cytokines. TCR/CD28 activation induces low-level production of IFN-γ, IL-6 and IL-17 in XIAP-deficient Treg cells ( Fig. 5 and Supplementary Figures 10-13). Only upon co-stimulation with IL-12 or IL-1/IL-6 does the expression of IFN-γ, IL-6 or IL-17 become prominent in XIAP-deficient Treg cells. Therefore, Treg cells do not manifest pathogenic consequences in Xiap −/− mice before any inflammatory stimulation. We found that elevated serum IL-6 was the most prominent among the several inflammatory cytokines we measured upon C. albicans lethal infection in Xiap −/− mice (Fig. 6). IL-6 affects the stability of Treg cells 56 , and blockage of IL-6 increases the population of Treg cells 57,58 . We further demonstrate that anti-IL-6R effectively inhibited IL-12-induced production of IFN-γ, while maintaining Foxp3 expression in Xiap −/− Treg cells (Fig. 7). Surprisingly, other anti-inflammatory biologics, i.e., anti-TNF, anti-IL-1R, or anti-IFN-γ did not effectively prevent IFN-γ generation in Xiap −/− Treg cells (Fig. 7e, Supplementary Figures 14 and 15a), illustrating the distinct role of IL-6R in Xiap −/− Treg cells conversion. Anti-IL-6R would be expected to prevent conversion of Treg cells into Th17-like cells, but our results illustrate that anti-IL-6R is unique in maintaining Treg cells stability in a Th1-prone environment.
Among the possible complications of anti-IL-6R is the increased susceptibility to infection and reduced generation of Th17 cells. In the present experiment, anti-IL-6R was administered 2 days after C. albicans infection. Notably, anti-IL-6R decreased inflammation, evidenced by both inflammatory cytokine levels and kidney neutrophil infiltration (Fig. 8b and e). Fungal load was actually reduced in mice treated with anti-IL-6R or anti-IL-6R plus Xiap −/− iTreg cells (Fig. 8f). In addition, anti-IL-6R did not affect antifungal Th17 responses (Fig. 8g). We speculate that a single dose of anti-IL-6R acts to prevent the conversion of Xiap −/− Treg cells and to attenuate inflammation, but it is not sufficient to attenuate antifungal immunity nor to suppress Th17 responses.
XIAP-deficiency leads to XLP-2 in patients, who exhibit overactivation of macrophages and lymphocytes. Xiap −/− mice succumb to infection by various pathogens [12][13][14][15] . Our results and those of others suggest that XIAP-deficient individuals are unable to clear infection due to primary immunodeficiency and persistent inflammation, leading to XLP-2 and lethality. As a primary immunodeficiency disease, the excess activation of lymphocytes in XLP-2 is likely a consequence of myeloid cell over-activation. Here, we have identified a subtle defect in the adaptive immunity of Xiap −/− mice related to their Treg cells. Xiap −/− Treg cells were mostly normal before infection, but were converted into IFN-γ-, IL-6-, and IL-17-expressing T cells in an inflammatory environment. Normal Treg cells are known to inhibit various immune cells 59 . Our observation that transfer of WT Treg cells effectively prevented infection-induced inflammation and conferred survival after an otherwise lethal infection in Xiap −/− mice (Fig. 6b-d) supports the notion that Treg cells are defective in Xiap −/− mice and XLP-2 patients, and that correction of XIAP-deficiency restores functional Treg cells. Our findings may also explain how lymphocytes become over-activated in XLP-2 patients, as Treg cells are required to keep lymphocyte activation in check. We propose a scenario whereby Xiap −/− mice are unable to control early infection as a consequence of impaired innate immunity, and inflammation caused by persistent pathogen presence primes for aberrant inflammatory Treg cells activation, leading to further escalated lymphocyte activation in Xiap −/− mice.
As a primary immunodeficiency disease, the only curative treatment for XLP-2 patients is allogeneic hematopoietic cell transplantation (HCT) to restore expression of XIAP in hematopoietic stem cells (HSCs), yet outcomes are limited by the toxicity associated with transplantation 60 . Even though transduction of the WT gene into hematopoietic progenitor cells from patients would in theory rescue the genetic defect 61 , autologous HSC transplantation in XLP-2 patients is difficult given the treatments required to eradicate endogenous hematopoietic progenitor cells in these ill pediatric patients. As an alternative, adoptive T cell immunotherapy has been explored for treatment of primary immunodeficiency diseases linked to viral infection 62 . Our study illustrates another possibility of using Treg cells in the treatment of inflammatory diseases caused by innate immunodeficiency. Our results (Fig. 6) suggest that administration of XIAP-reconstituted Treg cells could help reduce the excess immuno-related inflammation seen in XLP-2 patients. XIAP could be re-introduced into T cells from XLP-2 patients, and these XIAP-restored T cells could then be differentiated into Treg cells in vitro in sufficient quantities. Expression of XIAP in T cells could thus be considered an improvement over transduction of XIAP in HSCs from XLP-2 patients. It may be noted that, due to the low proliferative ability of mouse tTreg cells, we used mouse iTreg cells to test such a possibility (Fig. 6). Given that tTreg cells are more stable than iTreg cells, human tTreg cells are expected to be more effective than the iTreg cells shown in Fig. 6 in suppressing infection-induced inflammation.
We further advanced our therapeutic effect in mice by using a combination of anti-IL-6R and XIAP-deficient Treg cells. We illustrate that anti-IL-6R prevents the re-programming of Xiap −/− Treg cells into inflammatory Treg cells, and that Xiap −/− Treg cells become effective inhibitors of inflammation in the presence of anti-IL-6R (Fig. 8). Inflammation-induced Treg cells reprogramming is a major cause of Treg cells inactivation in vivo. Our results suggest that even if Treg cells functioning is substantially impaired by primary mutation, suppression of inflammation by co-administration of anti-IL-6R effectively restores the functioning of Treg cells. Therefore, a combination treatment of anti-IL-6R with autologous Treg cells should be effective in treating the inflammation and pathology of primary immunodeficiency diseases, bypassing the need to re-introduce the respective WT version of the defective gene into Treg cells.
A recent study revealed that innate immunodeficiency could be rescued by adaptive immunity, evidencing that an impaired response to Staphylococcus infection caused by TLR2 adapter deficiency could be rescued by antibodies against staphylococcal lipoteichoic acid 63 . Our results may be viewed as another way of reversing innate immunodeficiency by adaptive immunity using Treg cells and anti-IL-6R. In our case, anti-IL-6R blocks the Treg cells-destabilizing action of IL-6, allowing full execution of the suppressive effect of functional Treg cells on immuno-related inflammation.
In summary, we have identified a specific defect in XIAPdeficient Treg cells that contributes to the pathogenesis of XLP-2. We further used Xiap −/− mice to demonstrate that XLP-2 could be treated through combinatory use of ex vivo-expanded Xiap −/− iTreg cells and anti-IL-6R. Our results further suggest, most likely for the first time, the possibility of treating primary immunodeficiency diseases and inflammatory diseases by simultaneous use of autologous Treg cells and anti-IL-6R. Further characterizations may help establish the protocol for clinical applications.
Methods from the Institutional Animal Care and Use Committee, Academia Sinica. All mice used in this study were 8-12 weeks old. The same sex (male or female) mice were used in the same experiment, but opposite sex mice could be used in the repeat of the given experiment. No difference was observed between male and female mice in the analyses conducted in this study. Experimental groups were assigned randomly. Five or more mice in each experimental group was planned, but four mice in some experimental groups, that have been examined in previous studies, were used due to the knockout-mice availability. No blinding was done because the readouts of the mouse experiments in this study were clear-cut (body weight loss, death). No mice were excluded from scoring.
Cell cultures. HEK293T (ATCC CRL-3216) cells were obtained from ATCC. Cell lines were examined for mycoplasma contamination using a Mycoplasma Detection Kit (R&D). Primary mouse T cells and human CD14 + cells were cultured in RPMI-1640 medium supplemented with 10% fetal calf serum (Invitrogen Life Technology), 10 mM glutamine, 100 U ml −1 penicillin, 100 μg ml −1 streptomycin, and 50 μM 2-mercaptoethanol (referred to as complete RPMI medium). Human Treg cells were cultured in X-VIVO 15 medium with supplements identical to complete RPMI medium (complete X-VIVO15 medium). HEK293T cells were cultured in DMEM medium with the same supplements as for complete RPMI medium.
Mouse tTreg cell isolation and iTreg cell differentiation. Mouse total T cells were isolated from spleen and lymph nodes using anti-mouse Ig panning. T cells were cultured in complete RPMI 1640 medium. Mouse tTreg cells were purified from total T cells by sorting on a MoFlo Astrios system (Beckman Coulter) or by using a MACS ® CD4 + CD25 + Regulatory T Cell Isolation Kit (Miltenyi Biotech, Germany). The purity of tTreg cells was confirmed by having CD25 expression of 100% and Foxp3 expression >98.6% (Supplementary Figure 1). For WT and Xiap −/− mice carrying Foxp3-GFP, tTreg cells were isolated by sorting based on GFP + . IL-10 and TGF-β secretion from tTreg cells were determined by a mouse IL-10 DuoSet ELISA System (R&D) and a TGF-β OptEIA Set (BD Bioscience) after tTreg cells were stimulated with immobilized anti-CD3 (4 μg ml −1 ) and anti-CD28 (2 μg ml −1 ) for 96 h. Naive CD4 + T cells (CD4 + CD25 -CD62L + CD44 − ) were also isolated by sorting. Naive CD4 + T cells were differentiated into iTreg cells by stimulation with plate-bound anti-CD3 (4 μg ml −1 ) plus anti-CD28 (2 μg ml −1 ) in the presence of IL-2 (20 ng ml −1 ) and TGF-β (5 ng ml −1 ). Three days after differentiation, iTreg cells were collected and sorted by CD25 expression. Purified iTreg cells were used for in vitro suppressive assays or rested in complete RPMI medium for 3 days before using in plasticity analysis.
Human peripheral blood mononuclear cell and T cell isolation. Expired human white blood cell (WBC) concentrates were obtained from the Taipei Blood Bank with approval from the institutional review boards of Taipei Blood Bank and Academia Sinica. WBC concentrates were diluted with 1 × HBSS, overlaid on Ficoll-Paque, and centrifuged at 400 × g for 20 min. The interphase cells were collected and washed with 5% complete RPMI medium. The obtained peripheral blood mononuclear cells (PBMCs) were re-suspended in MACS® buffer (PBS containing 0.5% FBS and 2 mM EDTA). Naive human CD4 + T cells (CD4 + CD25 -CD45RA + CD45RO − ) were isolated from human PBMCs by sorting on a MoFlo Astrios system. Human tTreg cells (CD4 + CD127 low CD25 + ) and Teff (CD4 + CD25 − ) cells were similarly isolated from PBMCs. Naive human T cells were cultured in complete RPMI medium. Human tTreg cells were cultured in X-VIVO 15 medium with supplements identical to complete RPMI medium (referred to as complete X-VIVO15 medium).
FACS analysis. Immune cells were labeled with specific fluorescence-conjugated antibodies. Stained cells were analyzed on a FACS LSRII (BD Biosciences), and the data collected was analyzed using FlowJo software (Flow Jo LLC, Ashland, OR).
The sorting of selective cell population was performed on either FACSAria II SORP (BD Biosciences) or MoFlo Astrios (Beckman Coulter, Brea, CA). FACS gating/ sorting strategies are presented in Supplementary Figure 21.
Recombinant XIAP, XIAPΔRing and SOCS1-HA. HEK293T cells were transfected with XIAP-Flag or XIAPΔRing-Flag or SOCS1-HA plasmids. Cells were collected and lysed by WCE buffer 48 h after transfection. Total cell lysates were incubated with protein G Mag beads and anti-Flag (M2) or anti-HA overnight at 4°C. Recombinant XIAP-Flag and XIAPΔRing were eluted from beads by Flag peptides and were concentrated by a Vivaspin500 column (GE). SOCS1-HA-Mag beads were used directly in the ubiquitination assay.
Yeast preparation and quantitation. Candida albicans (C. albicans, ATCC90028) was cultured on yeast-mold (YM) plates at 25°C for 2 days and a single colony was inoculated into 10 ml YM broth at 30°C for 18-24 h. C. albicans was harvested by centrifugation at 3000 rpm for 5 min and washed with 10 ml PBS twice. The yeast pellet was resuspended in PBS and the optical density was determined. For determination of fungal load in vivo, organs were removed from the infected mice, grinded and resuspended in water. The samples were serially diluted and the diluents were spread onto YM plates. The number of colonies on plates was counted after incubation at 30°C for 2 days. incubated with PE Detection Reagent (BD Biosciences) in the dark at room temperature. The mixtures were washed and the fluorescence of the re-suspended beads was determined in a FACSCalibur system (BD Biosciences). The acquired data were analyzed with FCAP Array software (BD Biosciences).
Statistics. Data in this study were randomly collected but were not blinded. No data were excluded in this study. Our data mostly meet the assumption of the tests (normal distribution). GraphPad Prism 5 and Microsoft Office Excel and were used for data analysis. Unpaired two-tailed Student t-tests were used to compare results from between two groups. Data were presented as mean with standard deviation (s. d.) or standard error of the mean (s.e.m.), as indicated in the figure legend. Weight loss was analyzed by two-way ANOVA for multiple comparisons. Survival curves were plotted with Kaplan-Meier survival curve and analyzed by the log rank test (Mantel-cox). P values <0.05 were considered significant.
Data availability. The authors declare that the data supporting the findings of this study are available within the article and its Supplementary files, or are available from the authors upon request.