Inability to resolve specific infection generates innate immunodeficiency syndrome in Xiap^−/− mice

Mutations in innate receptors and their signaling molecules lead to a variety of innate immunodeficiency syndromes.^1-4  Patients with primary innate immunodeficiency are characterized by normal leukocyte development and susceptibility to specific infection. For example, interleukin-1 receptor-associated kinase-4 deficiency reveals its indispensable role to control pyrogenic bacteria but not most other pathogens, whereas Toll-like receptor (TLR)-3 deficiency selectively impairs immunity to herpes simplex virus.^1,2  Recent findings suggest that Crohn disease also results from a defective innate inflammatory response.^5-7 

X-linked lymphoproliferative syndrome type-2 (XLP-2) is a lymphoproliferative disease associated with deficiency in the X-linked inhibitor of apoptosis protein (XIAP).⁸  XLP-2 manifests as sensitivity to Epstein-Barr virus infection, leading to hemophagocytic lymphohistiocytosis, an inflammatory disorder caused by excess cytokine production from overactivated lymphocytes and macrophages.^9,10  XLP-2 is also associated with recurrent splenomegaly, fever, and hemorrhagic colitis.¹⁰  The molecular mechanism describing how XIAP deficiency leads to XLP-2 remains unknown.⁹ 

XIAP is the best-characterized member of the inhibitor of apoptosis protein (IAP) family, and is the only IAP protein that directly inhibits caspase-3, -7, and -9.^11,12  However, Xiap^−/− mice appear normal and do not exhibit any phenotype reminiscent of XLP-2.^13-15  Notably, Xiap^−/− mice are sensitive to infection of Listeria monocytogenes, with impaired innate immune responses and defective activation in macrophages.¹⁶  XIAP is also known to activate nuclear factor (NF)-κB independent of the inactivation of caspase.^12,17-19  The requirement of XIAP in NF-κB activation is stimuli selective, and XIAP is involved in nonobese diabetic (NOD)2-initiated NF-κB activation but is also dispensable in tumor necrosis factor (TNF)-α- and lipopolysaccharide (LPS)-induced NF-κB activation.^20-22 

Dectin-1 is the receptor recognizing β-glucan and is required for control of fungal infection.^23-27  Ligation of dectin-1 activates protein tyrosine kinase Syk, followed by activation of protein kinase C-δ.²⁷  Protein kinase C-δ in turn phosphorylates caspase recruitment domain 9 (CARD9), leading to formation of the CARD9–B-cell chronic lymphocytic leukemia/lymphoma 10 (BCL10)-mucosa-associated lymphoid tissue lymphoma translocation 1 (MALT1) complex and activation of IκB kinase and NF-κB.^23,24  Dectin-1 signaling also activates extracellular signal-regulated kinase, c-Jun N-terminal kinase, and p38 mitogen-activated protein kinase (MAPK).²⁸ 

In the present study, we demonstrate that XIAP is specifically required for dectin-1–induced innate responses. XIAP targeted BCL10, and XIAP deficiency impaired the dectin-1–induced BCL10-mediated activation signals. Consequently, Xiap^−/− mice were unable to control infection by Candida albicans. The persistent presence of C albicans resulted in overactivation of macrophages, excess production of inflammatory cytokines, and splenomegaly in Xiap^−/− mice. Importantly, restoration of Rac1 activation and phagocytosis, but not NF-κB activation, by resolvin D1 rescued Xiap^−/− mice from C albicans lethal infection. Our results provide evidence of an innate immunity-associated mechanism that may contribute to the generation of XLP-2 syndromes in XIAP mutant patients. Furthermore, we suggest the potential therapeutic value of resolvin D1 in treating XLP-2 and, possibly, innate immunodeficiency syndromes.

Zymosan depleted (InvivoGen) is the hot alkali-treated cell wall from Saccharomyces cerevisiae lacking all its endogenous TLR-stimulating activity but retaining dectin-1–activating capacity. The human XIAP constructs were previously described.²⁹  The Xiap^−/− mouse¹⁴  was a generous gift from Dr David L. Vaux (Walter and Eliza Hall Institute of Medical Research, Parkville, Australia). Mice were maintained in the specific pathogen-free mouse facility of the Institute of Molecular Biology, Academia Sinica. All mouse experiments were conducted with approval from the Institutional Animal Care and Use Committee, Academia Sinica. Detailed reagents are listed in supplemental Methods available on the Blood Web site.

Bone marrow cells were collected from tibias and femurs flushed with cold phosphate-buffered saline through a 24-gauge needle and were cultured in complete Dulbecco’s modified Eagle medium containing 20 ng/mL granulocyte macrophage–colony-stimulating factor (R&D) for 8 days to generate bone marrow-derived dendritic cells or in Dulbecco’s modified Eagle medium containing 20% L929 cell-conditioned medium to generate bone marrow-derived macrophages. Human CD14⁺ monocytes were isolated from peripheral blood mononuclear cells and were differentiated into macrophages in the presence of macrophage–colony-stimulating factor. Rac1V12 or small interfering XIAP was introduced into human macrophages by electroporation in a Nucleofector (Amaxa, Koeln, Germany). Human peripheral blood mononuclear cells were used with approval from the Institutional Review Board of Biomedical Science Research, National Health Research Institutes.

C albicans (ATCC90028) was cultured on a yeast mold plate at 25°C for 2 days, and a single colony was inoculated into 10 mL YM broth at 30°C for 18 to 24 hours. C albicans were harvested by centrifugation. The yeast pellet was resuspended with phosphate-buffered saline and the OD was determined. For determination of fungal burden in vivo, organs were removed from the infected mice, ground, 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.

Aspirin-triggered Resolvin D1 (AT-RvD1; Cayman) (abbreviated as RvD1) was dissolved in phosphate-buffered saline and was administered intravenously (100 ng/mouse)^30-32  3 days after C albicans infection. For in vitro experiments, bone marrow-derived macrophages (BMDMs) were treated with RvD1 for 30 minutes before the addition of heat-killed C albicans (HKCA).

Additional methods are reported in supplemental Methods.

Stimulation of Xiap^−/− BMDMs with LPS (for TLR4), Pam3CSK4 (for TLR2), poly (I:C) (for TLR3), or R848 (for TLR7/8) generated comparable amounts of TNF-α and interleukin (IL)-6 as wild-type (WT) BMDMs [Figure 1A-C; data not shown for IL-6 stimulated by poly (I:C) and R848]. We used zymosan depleted to selectively activate dectin-1 in WT and Xiap^−/− bone marrow-derived dendritic cells (BMDCs). Dectin-1–induced production of TNF-α, IL-6, and IL-10 was attenuated in Xiap^−/− BMDCs (Figure 1D-F). Similarly, XIAP-knockdown human macrophages (Figure 1G) reacted normally to LPS stimulation (supplemental Figure 1A-B) and responded poorly to dectin-1 stimulation during the production of TNF-α and IL-6 (Figure 1H-I). XIAP deficiency did not affect the surface expression of dectin-1 and dectin-2 (supplemental Figure 2A-B). Dectin-1–induced activation of Syk, extracellular signal-regulated kinase, and c-Jun N-terminal kinase was comparable between control and Xiap^−/− macrophages (supplemental Figure 2C-D). XIAP deficiency specifically reduced dectin-1–induced IκBα phosphorylation, IκB degradation (supplemental Figure 2E), and p65 nuclear entry, as well as p38 MAPK activation in macrophages (Figure 1J-L). In contrast, comparable NF-κB activation was observed in WT and Xiap^−/− macrophages activated by Pam3CSK4 and LPS (supplemental Figure 3A-B) and in WT and Xiap^−/− T cells stimulated by TNF-α and LPS (supplemental Figure 3C-D). Therefore, XIAP is required for dectin-1–induced NF-κB activation but is not apparently involved in NF-κB activation stimulated by TNF-α, Pam3CSK4, LPS, poly (I:C), and R848 in T cells and macrophages.

Because XIAP participates in dectin-1–induced, but not TNF-α– and TLR-triggered NF-κB activation, the NF-κB activation stage that would be regulated by XIAP should be upstream of transforming growth factor-beta-activated kinase 1-IκB kinase activation. Dectin-1–induced NF-κB activation differs from TNF-α– and LPS-triggered NF-κB activation in the selective formation of the CARD9-BCL10-MALT1 complex.^23-28  A previous study suggested an association of XIAP with BCL10.³³  Immunoprecipitation of the overexpressed XIAP-Myc brought down FLAG-BCL10 but not FLAG-MALT1 (supplemental Figure 4). The specific interaction between endogenous XIAP and BCL10 was confirmed in curdlan-stimulated macrophages (Figure 2A). XIAP consists of the N-terminal baculovirus IAP repeat (BIR) 1, BIR2, BIR3, and C-terminal really interesting new gene finger domains.¹²  Using FLAG-tagged XIAP deletion mutants, we then mapped the domain of XIAP that binds BCL10. The BCL10-binding region was isolated to the BIR1 segment of XIAP (Figure 2B).

We examined whether XIAP promotes BCL10 polyubiquitination. Indeed, coexpression with XIAP increased the polyubiquitination of BCL10 in vivo (Figure 2C). However, deletion of the RING finger domain from XIAP (XIAPΔRF) or the removal of BIR1 eliminated the ability of XIAP to ubiquitinate BCL10 (Figure 2C). Further, the incorporation of K63 Ub into BCL10 was enhanced by XIAP, whereas the addition of K48 Ub to BCL10 was not affected by XIAP in 293T cells (supplemental Figure 4B). We also used in vitro ubiquitination analysis to show that, in reaction mixtures containing ubiquitin, E1, E2, and recombinant BCL10, the addition of recombinant XIAP, but not XIAP∆RF, induced K63 ubiquitination of BCL10 (Figure 2D).

We next examined the possible XIAP-mediated ubiquitination sites on BCL10. Ubiquitination of Lys31 and Lys63 on BCL10 has been functionally linked to the activation of NF-κB.³⁴  Mutation of Lys31 and Lys63 on BCL10 (BCL10-KKRR) prevented XIAP-mediated K63 ubiquitination (Figure 2E), suggesting that K31 and K63 on BCL10 are specifically targeted for ubiquitination by XIAP. In addition, BCL10-stimulated κB-Luc activation³⁵  was profoundly enhanced by the coexpression of WT XIAP but not by XIAPΔBIR1 or XIAPΔRF (Figure 2F). Therefore, the association of XIAP with BCL10 and the ubiquitination of BCL10 modulated the ability of BCL10 to activate NF-κB.

The CARMA1-BCL10 complex is involved in NF-κB activation triggered through the T-cell receptor (TCR) and B-cell receptor.^34,36-38  CARMA3 and BCL10 are required for NF-κB activation induced by epidermal growth factor (EGF) and G protein-coupled receptor agonists such as lysophosphatidic acid (LPA).^39-43  We therefore examined whether XIAP also participated in NF-κB activation stimulated by these receptors. In the absence of CD28 costimulation, XIAP deficiency led to attenuated CD3-stimulated interferon (IFN)-γ production (Figure 3A), deceased TCR-induced IKKα/β phosphorylation (Figure 3B), and reduced anti–CD3-induced RelA nuclear translocation (Figure 3C). It may be noted that XIAP deficiency did not affect CD3/CD28-induced IKKα/β phosphorylation and IL-2 production (supplemental Figure 5), consistent with the effect previously reported in Xiap^−/− T cells¹³  and CARMA1-knockout T cells.⁴⁴ 

We also used Xiap^−/− macrophages to confirm that LPA-triggered IL-6 production and p65 nuclear translocation were decreased in Xiap^−/− macrophages (Figure 3D-E). In addition, EGF-stimulated IκBα phosphorylation and p65 nuclear entry was reduced in XIAP-knockdown HepG2 cells (Figure 3F-G). Therefore, consistent with the suggestion that XIAP is involved in BCL10-mediated NF-κB signaling, XIAP is required for full NF-κB activation triggered by dectin-1, LPA, epidermal growth factor receptor (EGFR), or TCR alone.

C albicans is recognized by dectin-1 and dectin-2, and the fungal infection is controlled by dectin-induced responses through the CARD9-BCL10-NF-κB pathway.^23-27,45,46  We therefore tested whether deficiency in XIAP leads to defective responses toward C albicans. We found that 8- to 10-week-old Xiap^−/− mice were highly susceptible to C albicans (1 × 10⁶) infection, with all mice dying from infection by day 3 (Figure 4A). Older Xiap^−/− mice (16 weeks old) were modestly more resistant to C albicans (1 × 10⁶; Figure 4B). Immediately after C albicans administration, generation of IL-6, monocyte chemoattractant protein (MCP)-1, and TNF-α was diminished in Xiap^−/− mice relative to WT mice (Figure 4C). At a lower dose of C albicans (1 × 10⁵), Xiap^+/+ mice were fully resistant, whereas most of the Xiap^−/− mice died from infection before day 15 (Figure 4D), accompanied by progressive body weight loss. In these dying Xiap^−/− mice, scattered foci of abscesses were observed in the swollen kidneys (Figure 4E), as well as a large fraction of infiltrated neutrophils in kidney (Figure 4F). C albicans titers were also highly elevated in the kidney, liver, and spleen tissues of Xiap^−/− mice (Figure 4G). The compromised responses to C albicans in XIAP-deficient mice were comparable to those seen in mice lacking dectin-1, dectin-2, or CARD9.^23,28,45,46  Another consequence of XIAP deficiency was found in the accumulation of proinflammatory cytokines following a low dose of C albicans infection. Although the serum contents of IL-6, TNF-α, and MCP-1 were higher in WT mice than Xiap^−/− mice 6 hours after a low dose of C albicans infection (data not shown), these cytokines were no longer detectable in the peripheral blood of WT mice 1 week after infection (Figure 4H), consistent with the near clearance of C albicans in these mice (Figure 4G). In contrast, Xiap^−/− mice retained a high serum level of IL-6, TNF-α, and MCP-1 at the same time point (Figure 4H), suggesting that the continued presence of C albicans turns the defective innate responses into a proinflammatory status late in infection. A correlation was also found between the extent of the body weight loss and serum levels of IL-6, TNF-α, and MCP-1 in those Xiap^−/− mice 1 week after C albicans infection (supplemental Figure 6), illustrating that high levels of proinflammatory cytokines are associated with poor prognosis.

We also measured T-cell responses to HKCA in WT and Xiap^−/− mice. After immunization with HKCA, the production of IFN-γ and IL-4 was comparable between WT and Xiap^−/− T cells (supplemental Figure 7A-B). Because T-cell response was not affected as obviously as macrophages in Xiap^−/− mice, we focused on the dectin-1–induced innate immunity in Xiap^−/− mice during the rest of study.

To further delineate the effect of XIAP in the BCL10-dependent innate response, we used a model dectin-1 ligand, curdlan. Although WT mice did not display any abnormalities following weekly curdlan intraperitoneal immunization, Xiap^−/− mice exhibited body mass loss 1 week after curdlan priming (Figure 5A) and did not recover until the fourth week (data not shown). Serum levels of IL-6, IL-10, MCP-1, and TNF-α were decreased in Xiap^−/− mice after administration of curdlan (Figure 5B). Curdlan-induced immediate peritoneal infiltration of macrophages and neutrophils was reduced in Xiap^−/− mice (Figure 5C). The compromised early innate responses to curdlan in Xiap^−/− mice were reversed a week later. The generation of IL-6, IL-10, TNF-α, and MCP-1 were much higher in Xiap^−/− mice than the WT control (Figure 5D). Splenomegaly was found in Xiap^−/− mice 1 week after curdlan priming, with profound infiltration of mononuclear cells (Figure 5E). Also, the peritoneal presence of macrophages and neutrophils became higher in Xiap^−/− mice than in control mice 1 week after curdlan administration (Figure 5F).

As a control, immunization of Pam3CSK4 did not generate any differential response between control and Xiap^−/− mice (supplemental Figure 8). Repetitive administration of Pam3CSK4 did not produce detrimental effects in Xiap^−/− mice (data not shown). Because Xiap^−/− mice reacted normally to Pam3CSK4 in NF-κB activation and inflammatory cytokine production (Figure 1A-B; supplemental Figure 3A), we examined whether priming of Pam3CSK4 increased the resistance of Xiap^−/− mice to C albicans infection. Prior administration of Pam3CSK4 did not rescue Xiap^−/− mice from C albicans (1 × 10⁵) infection, despite an increase in the proinflammatory cytokine levels (supplemental Figure 9A-C). Instead, Pam3CSK4 accelerated C albicans-mediated mortality in both control and Xiap^−/− mice (supplemental Figure 9D-E), indicating that restoration of BCL10-dependent NF-κB activation and inflammatory cytokine expression in Xiap^−/− mice failed to compensate for the defective innate response to C albicans.

BCL10 was recently shown to participate in Rac1 activation, F-actin remodeling, and phagosome closure that are independent of NF-κB activation.⁴⁷  We found that curdlan-induced Rac1 activation was attenuated in Xiap^−/− BMDMs (supplemental Figure 10A). Similarly, dectin-1–induced Rac1 activation was impaired in XIAP-knockdown human macrophages (Figure 6A). For phagocytosis, the uptake of plain latex beads or Pam3CSK4-coated latex in Xiap^−/− BMDMs was not affected, as measured by confocal microscopy or flow cytometry (supplemental Figure 10B-C). In contrast, the phagocytosis of HKCA was impaired in Xiap^−/− BMDMs (supplemental Figure 10D). Similar to Xiap^−/− BMDMs, XIAP-knockdown human macrophages were defective in the phagocytosis of HKCA (Figure 6B). Quantitation by flow cytometry further confirmed the impaired uptake of C albicans in XIAP-deficient human macrophages (Figure 6C). We also introduced active Rac1 (Rac1-V12) into XIAP-knockdown human macrophages, shown by the constitutive Rac1 activity (Figure 6A). The defective uptake of C albicans was reversed by the expression of Rac1-V12 in XIAP-knockdown human macrophages (Figure 6B-C), suggesting that impaired Rac1 activation accounts for the ineffective uptake of HKCA in XIAP-deficient macrophages.

We then looked for reagents that may correct the defect of Xiap^−/− macrophages. Resolvin D1 is an anti-inflammatory lipid mediator biosynthesized from ω-3 fatty acid docosahexaenoic acid.^30,31  Resolvin D1 displays a specific activity to enhance phagocytosis in macrophages.^32,48-50  We found that aspirin-triggered resolvin D1 (AT-RvD1) restored dectin-1–induced Rac1 activation in Xiap^−/− BMDMs (supplemental Figure 11A). RvD1 did not further increase the phagocytosis of HKCA in control BMDMs but significantly increased the uptake of HKCA in Xiap^−/− BMDMs (supplemental Figure 11B-C). A similar effect of RvD1 was found in XIAP-knockdown human macrophages. The addition of RvD1 corrected the defective dectin-1–induced Rac1 activation (Figure 6D) and restored the phagocytosis of HKCA in XIAP-deficient human macrophages (Figure 6E-F). Notably, RvD1 did not increase dectin-1–mediated p38 MAPK phosphorylation or NF-κB activation in Xiap^−/− BMDMs (supplemental Figure 11D), indicating that the effect of RvD1 is independent of p38 and NF-κB.

We then examined the in vivo effect of RvD1 in Xiap^−/− mice. Administration of RvD1 3 days after C albicans (1 × 10⁵) infection fully rescued Xiap^−/− mice from C albicans-mediated mortality (Figure 7A). Development of the pathological morphology in the kidneys of Xiap^−/− mice after C albicans infection was completely prevented by RvD1 (Figure 7B). For reasons that are unclear, kidneys from 4 of the 6 RvD1-treated Xiap^−/− mice appeared to be smaller than those from RvD1-treated WT mice. Neutrophil infiltration in the kidney from C albicans-infected Xiap^−/− mice was abrogated by RvD1 administration (Figure 7C). The high levels of IL-6, TNF-α, and MCP-1 in Xiap^−/− mice 10 days after C albicans infection was no longer detectable after RvD1 administration (Figure 7D-F). The fungal loads in liver, spleen, and kidney were effectively reduced by RvD1 treatment (Figure 7I-K). RvD1 also increases the production of IL-10,^49,51  known to participate in the resolution of inflammation. We found that curdlan-triggered IL-10 generation in Xiap^−/− BMDCs was not altered by the presence of AT-RvD1 (Figure 7G). Application of RvD1 did not increase serum levels of IL-10 in Xiap^−/− mice with prior C albicans infection (Figure 7H). Together, these results suggest that increases in dectin-1–induced Rac1 activation and phagocytosis by RvD1 helped Xiap^−/− mice resolve inflammation and survive C albicans infection.

XIAP is a caspase-binding antiapoptotic protein critical for cell survival and has been shown to activate NF-κB. XIAP deficiency in humans leads to XLP-2, although the mechanisms remain unclear. However XIAP-knockout mice appear normal.^13-15  In the present study, we identified the involvement of XIAP in selective NF-κB activation. We showed that TNF-α/IL-6 production induced by TLR2, TLR3, TLR4, or TLR7/8 was normal in Xiap^−/− BMDMs, and XIAP deficiency did not affect TLR-induced NF-κB activation in T cells and macrophages, consistent with several recent reports.^16,20-22  XIAP deficiency impaired only NF-κB activating processes that are mediated by BCL10, such as those stimulated by dectin-1, TCR, EGF, or LPA (Figures 1 and 3). Our data also agree with the observations that TNF-α–triggered NF-κB activation is normal in BCL10-knockout cells⁴⁰  and that LPS-induced NF-κB activation is not affected by deficiency in CARD9,²³  the BCL10-binding partner. We demonstrated that XIAP promoted BCL10 K63 ubiquitination, but further work is required to delineate the exact role of XIAP-BCL10 in CARMA1-BCL10-MALT1–mediated NF-κB activation. TCR-stimulated NF-κB in Xiap^−/− T cells was indistinguishable from control T cells when stimulated through CD3/CD28 (supplemental Figure 5). A defect in T-cell activation was observed in Xiap^−/− T cells only when activated by anti-CD3 alone (Figure 3A). We also found that immunization with HKCA generated comparable T-cell responses in WT and Xiap^−/− mice (supplemental Figure 7). Despite the fact that XIAP targets to BCL10, there was a different sensitivity to TCR and dectin-1 stimulation in Xiap^−/− T cells and macrophages. Further work will be needed to delineate this discrepancy.

We demonstrated that Xiap^−/− mice died from C albicans infection. XLP-2 is known for its association with Epstein-Barr virus, and no C albicans infection has been reported in XLP-2, likely due to the rarity of XLP-2 patients. Because Xiap^−/− T-cell responses are nearly normal, XLP-2 may be attributed to abnormal innate immunity in XIAP-deficient hosts. We further illustrated that the defective BCL10-dependent uptake of C albicans plays a major role in the vulnerability of Xiap^−/− mice to C albicans infection. BCL10 participates in Rac1 activation, F-actin remodeling, and phagosome formation that are NF-κB independent.⁴⁷  XIAP-deficient macrophages displayed no defects in the phagocytosis of plain latex beads and Pam3CSK4-coated latex beads but were ineffective in the uptake of C albicans, indicating a selective deficiency in dectin-1–induced phagocytosis. Introduction of active Rac1 restored the phagocytosis of C albicans in XIAP-deficient macrophages (Figure 6A-C). Up-regulation in NF-κB activation and proinflammatory cytokine production by Pam3CSK4 failed to rescue Xiap^−/− mice from C albicans-mediated inflammation and lethality (supplemental Figure 9). Instead, restoration in the Rac1 activation and uptake of C albicans effectively led to clearance of infection (Figure 7). The inability of Xiap^−/− mice to respond to C albicans infection thus could be attributed mostly to the impaired BCL10-dependent Rac1 activation and phagocytosis necessary for yeast clearance.

Importantly, we demonstrated that XIAP-deficient human macrophages recapitulated the specific defects in Xiap^−/− mouse macrophages (Figures 1 and 6). XIAP-knockout human macrophages reacted indistinguishably from control macrophages to LPS treatment (supplemental Figure 1) but were unable to respond to dectin-1 stimulation (Figure 1H-I). Similar to Xiap^−/− macrophages, XIAP-deficient human macrophages were compromised in dectin-1–stimulated Rac1 activation and were deficient in the uptake of C albicans. The present study likely represents the first identification of the physiological defect in XLP-2 macrophages. XLP-2 patients are characterized by hemophagocytic lymphohistiocytosis.^9,10  Whether impaired Rac1 activation and defective phagocytosis lead to visible remains of erythrocytes and leukocytes in XLP-2 macrophages remains to be determined.

We further used the dectin-1 ligand curdlan to demonstrate that the reduced innate responses in Xiap^−/− mice lead to development of selected XLP-2 syndromes. The initial inability of Xiap^−/− mice to mount effective innate immunity to curdlan (Figure 5B) resulted in the persistent presence of curdlan, and its continued stimulation was accompanied by accumulation of the proinflammatory cytokines, leading to elevated accumulation of macrophages and neutrophils in the peritoneal cavity and splenomegaly (Figure 5D-F). Therefore, the inability of Xiap^−/− mice to generate effective immune responses to curdlan and C albicans results in some inflammatory syndromes seen in XLP-2 patients.

It has been shown that moderate T-cell immunodeficiency, but not severe T-cell immunodeficiency, generates immune dysregulation and autoimmunity.⁵²  We propose that XIAP deficiency is an example that partial innate immunodeficiency generates uncontrolled inflammatory consequences. In an SPF environment without pathogenic challenge, Xiap^−/− mice are indistinguishable from WT mice. However, for microbes or pathogen-associated molecular patterns that activate BCL10-dependent pattern recognition receptors, Xiap^−/− mice are unable to mount effective innate immune responses to resolve infection or eliminate pathogen-associated molecular patterns, and their continued stimulation leads to inflammation and XLP-2–like syndromes including persistent activation of macrophages, excess production of inflammatory cytokines, and splenomegaly. Our results suggest that defective BCL10/NOD2-directed innate activation represents one of the mechanisms that contribute to the pathogenesis of XLP-2 in XIAP-deficient patients. Because XIAP deficiency impairs the innate reactivity of myeloid cells, our results may explain why myeloablative conditioning regimens lead to poor survival in allogeneic hematopoietic cell transplantation on XIAP-mutated patients,⁵³  as myeloid cell depletion further aggravates pathogenesis caused by innate immunodeficiency.

Innate immunodeficiency is characterized by sensitivity to a narrow spectrum of infection.^1-3  Increasing numbers of inflammatory diseases, including cystic fibrosis, are attributed to innate immunodeficiency.⁵⁴  In the present study, we used Xiap^−/− mice to suggest a molecular basis of such selectivity. An interesting analogy is that Crohn disease-associated NOD2 variants are loss-of-function mutants.⁵⁵  How the loss of function of a proinflammatory signal molecule mechanistically contributes to an inflammatory disorder has been subjected to extensive investigations.^5-7  It has been suggested that patients with NOD2 mutants are unable to resolve infection by specific pathogens due to innate immunodeficiency.^5,7  It is likely that the persistence of microbes leads to continued inflammation and adaptive immunity activation, similar to what we observed in this study in Xiap^−/− mice.

We also demonstrated that RvD1 restored Rac1 activation and phagocytosis in XIAP-deficient macrophages, resulting in effective resolution of inflammation, elimination of C albicans, and full survival of the infected Xiap^−/− mice. Interestingly, RvD1 did not rescue the activation of NF-κB and p38 in Xiap^−/− macrophages. These results suggest that, for innate immunodeficiency-induced inflammatory syndromes such as XLP-2, after the initial failure of host responses to clear pathogen, resolution of the latter-stage inflammation is likely the key for the success in the therapy. Furthermore, our results suggest that increased uptake of pathogens and resolution of the inflammation by RvD1 are potential therapeutic interventions for inflammatory diseases caused by primary innate immunodeficiency. Further work will determine the feasibility of such an approach.

The authors thank Dr David L. Vaux for the Xiap^−/− mouse and Yamin Lin and the FACS Core, Sue-Ping Lee, and the Confocal Core of the Institute of Molecular Biology, Academia Sinica, for cell sorting and confocal microscopy. Dr AndreAna Peña, English editor of Institute of Molecular Biology, Academia Sinica, provided editorial assistance to the authors during preparation of this manuscript.

This work was supported by National Science Council grant 102-2321-B-001-045 and an Academia Sinica Investigator Award from Academia Sinica, Taiwan, Republic of China.

Contribution: W.-C.H., Y.-T.C., and I.-H.C. performed experiments; S.-C.H. and S.-C.M. contributed to generation of key materials and constructs; and M.-Z.L. conceived of, designed, and supervised the research and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Ming-Zong Lai, Institute of Molecular Biology, Academia Sinica, Taipei 11529, Taiwan, Republic of China; e-mail: mblai@gate.sinica.edu.tw.