The NOD2 defect in Blau syndrome does not result in excess interleukin-1 activity

Authors


Abstract

Objective

Blau syndrome is a rare, autosomal-dominant, autoinflammatory disorder characterized by granulomatous arthritis, uveitis, and dermatitis. Genetics studies have shown that the disease is caused by single nonsynonymous substitutions in NOD-2, a member of the NOD-like receptor or NACHT–leucine-rich repeat (NLR) family of intracellular proteins. Several NLRs function in the innate immune system as sensors of pathogen components and participate in immune-mediated cellular responses via the caspase 1 inflammasome. Mutations in a gene related to NOD-2, NLRP3, are responsible for excess caspase 1–dependent interleukin-1β (IL-1β) in cryopyrinopathies such as Muckle-Wells syndrome. Furthermore, functional studies demonstrate that caspase 1–mediated release of IL-1β also involves NOD-2. The aim of this study was to test the hypothesis that IL-1β may mediate the inflammation seen in patients with Blau syndrome.

Methods

IL-1β release was measured in peripheral blood mononuclear cells cultured in vitro, obtained from 5 Blau syndrome individuals with a NOD2 (CARD15) mutation.

Results

We observed no evidence for increased IL-1β production in cells obtained from subjects with Blau syndrome compared with healthy control subjects. Furthermore, we presented 2 cases of Blau syndrome in which recombinant human IL-1 receptor antagonist (anakinra) was ineffective treatment.

Conclusion

Taken together, these data suggest that in contrast to related IL-1β–dependent autoinflammatory cryopyrinopathies, Blau syndrome is not mediated by excess IL-1β or other IL-1 activity.

Innate immunity is largely mediated by cell-surface and intracellular sensors that specifically detect highly conserved pathogen-associated molecular patterns of microorganisms. Dysregulation of the innate immune response leads to many inflammatory diseases. Much effort has been directed toward elucidating the biologic and pathologic functions of these pattern recognition receptors. In addition to the Toll-like receptors (TLRs), another recently recognized family of proteins, the NOD-like receptors (also defined as NACHT–leucine-rich repeat [LRR] [NAIP, CIITA, HET-E, TP-1 LRRs]) (NLR), includes several members thought to function as intracellular pattern recognition receptors and to participate in inflammatory signaling pathways. Recent progress in the field has begun to impact our understanding and treatment of a number of unique inflammatory diseases mediated by pattern recognition receptors (1, 2).

Mutations in NLRP3 (NLR family, pyrin domain–containing protein 3; also called CIAS1 or NALP3) are responsible for an overlapping spectrum of autoinflammatory disease collectively referred to as cryopyrinopathies. These are represented by familial cold-induced autoinflammatory syndrome (MIM no. #120100), Muckle-Wells syndrome (MIM no. #191900), and chronic infantile neurologic, cutaneous, articular syndrome (MIM no. #607115; also called neonatal-onset multisystem inflammatory disease) (3–7). In severe cases, patients exhibit neonatal-onset fever and chronic multiorgan inflammation that frequently involves skin, joints, and the central nervous system. The pathology induced by the NLRP3 defect is largely due to aberrant activation of the caspase 1 inflammasome, leading to overproduction of the inflammatory cytokine interleukin-1β (IL-1β) (8). Furthermore, anakinra, a recombinant human IL-1 receptor antagonist, significantly alleviates the intense systemic inflammation in patients with cryopyrinopathy (9–18).

Another NLR family member, NOD-2 (also called NLRC2 [NLR family, CARD domain–containing 2]), has been implicated in inflammatory diseases. In contrast to NLRP3, the pathologic mechanisms resulting from alterations in NOD2 (CARD15) are far from clear. Three major NOD2 polymorphisms (and perhaps several minor polymorphisms, as well) are known to be associated with Crohn's disease (19–22). Furthermore, different mutations of this same gene are also responsible for a rare, autosomal-dominant disease, Blau syndrome (MIM no. #186580) (23). Blau syndrome (also known as juvenile systemic granulomatosis [familial or sporadic], Jabs disease, or, more recently, pediatric granulomatous arthritis) is characterized by granulomatous uveitis, joint inflammation, and skin inflammation (24–27). To take the emerging understanding of Blau syndrome one step further, patients with a clinical picture virtually indistinguishable from Blau syndrome except without a family history are frequently given a diagnosis of early-onset sarcoidosis, because their symptoms include granulomatous inflammation. We and other investigators have shown that these patients exhibit de novo mutations in NOD2 as well (28, 29).

How might mutations in the same gene cause 2 very distinct clinical pathologies (i.e., Blau syndrome versus Crohn's disease)? The NOD2 polymorphisms associated with Crohn's disease are thought to perturb the functional domain of the protein directly involved in detecting a core component of the bacterial cell wall, muramyldipeptide. Cells expressing NOD-2 respond to muramyldipeptide with increased NF-κB activity, which is important for inflammatory responses. The polymorphisms associated with Crohn's disease are likely to impair the ability of the cell to respond to bacteria. In contrast, all of the mutations found to cause Blau syndrome are in or near the NOD functional motif, which is thought to be involved in protein–protein interactions and nucleotide binding. There is evidence that these mutations result in increased basal NF-κB activity, possibly promoting inflammation (30, 31).

In addition to muramyldipeptide-induced activation of NF-κB via NOD-2, muramyldipeptide has been shown to trigger the release of IL-1β in an NLRP3-dependent manner (32). This IL-1β release was subsequently reported to be dependent on NOD-2 as well (33). These studies suggest that NOD-2 may participate in the inflammasome complex, which triggers release of IL-1β. Furthermore, the Blau syndrome mutations in NOD2 encode substitutions affecting a functional domain analogous to the domain of NLRP3 that is altered in the cryopyrinopathies. Last, a recent report documented elevated IL-1β, IL-6, and tumor necrosis factor α (TNFα) serum levels in 1 patient with Blau syndrome, which normalized upon treatment with anakinra (34). Together, these findings have led to the assumption by several investigators that Blau syndrome is also likely to be mediated by increased IL-1 activity.

Studies of Blau syndrome–associated NOD2 mutations would provide unique and unambiguous insights into both the pathogenesis of the disease and the biologic function of NOD2. Currently, the majority of studies to elucidate the functional effects of NOD2 mutations causing Blau syndrome have been conducted using cell culture transfection systems. In these systems, the signaling pathways could be nonphysiologically altered by overexpression of NOD2. Thus, results may not represent the actual pathogenetic mechanism of the Blau syndrome–causing NOD2 mutation.

In this study, we sought to better characterize the effect of Blau syndrome mutations of NOD2 by examining peripheral blood mononuclear cells (PBMCs) from 5 individuals with Blau syndrome. In vitro studies have shown that muramyldipeptide may synergize with the immune response of monocytes in response to TLR-2 (palmitoyl-3-cysteine [Pam3Cys]) and TLR-4 (lipopolysaccharide [LPS]) stimulation, suggesting an interaction between NOD-2 and TLR signaling pathways (35–37). Therefore, we stimulated PBMCs with muramyldipeptide, Pam3Cys, LPS, or combinations of muramyldipeptide with the TLR agonists. In particular, we tested the hypothesis that Blau syndrome is associated with excess production of IL-1β, as has been demonstrated for other autoinflammatory conditions. We provide additional insight into the role of IL-1 in this disease by presenting the clinical response in 2 patients with Blau syndrome who were treated with anakinra, an IL-1 receptor antagonist.

PATIENTS AND METHODS

Subjects.

Five subjects with the classic manifestations of Blau syndrome and 5 age- and sex-matched healthy volunteers were included in this study. Subjects 1, 2, and 3 represent a mother and her 2 sons; subjects 4 and 5 represent a mother and her son. The PBMC cytokine assays and genetics analysis were performed under protocols approved by the Oregon Health & Science University Institutional Review Board. Demographic, genotypic, and treatment information for the subjects is summarized in Table 1. NOD2 genotypes for each of the 5 subjects with Blau syndrome were previously reported as part of an international registry (27).

Table 1. Demographic characteristics, genotype, and current therapy of subjects with Blau syndrome and healthy control subjects*
 Age, yearsSexNOD2 substitutionSystemic medication
  • *

    The mean age of the control subjects was 43 years; the mean age of Blau syndrome subjects was 41 years.

Control subjects    
 156M  
 253F  
 342M  
 420M  
 545F  
Blau syndrome subjects    
 150MR334WPrednisone
 268FR334WNone
 349MR334WNone
 47MR334QEtanercept
 530FR334QEtanercept, methotrexate

Collection of PBMCs.

Single paired samples of blood were collected from subjects with Blau syndrome and healthy control subjects. PBMCs were isolated from whole blood by centrifugation on a 1-layer Ficoll-Hypaque gradient (Amersham Biosciences, Pittsburgh, PA). The PBMCs were washed 3 times in X-VIVO 15 medium (Cambrex, East Rutherford, NJ) and used immediately for further analysis.

Reverse transcriptase–polymerase chain reaction (RT-PCR) for NOD2 expression.

Total RNA from PBMCs was isolated with the RNeasy Mini Kit (Qiagen, Valencia, CA). First-strand complementary DNA (cDNA) synthesis was accomplished with oligo(dT)-primed Moloney murine leukemia virus reverse transcriptase (Gibco-BRL Life Technologies, Rockville, MD). Gene-specific cDNA was amplified by a hot-start touchdown PCR procedure, with Taq polymerase (Applied Biosystems, Foster City, CA) and specific human NOD2 primer pairs (sense 5′-AGCCATTGTCAGGAGGCTC-3′ and antisense 5′-CGTCTCTGCTCCATCATAGG-3′). PCR thermal cycler conditions were as follows: 1 × 4 minutes at 94°C, 35 cycles of denaturation (15 seconds at 94°C), annealing (60 seconds at 55°C), and extension (60 seconds at 70°C). A primer pair for a constitutively expressed gene, GAPDH, was performed in a replicate reaction. PCR products were separated by electrophoresis in a 2% agarose gel with ethidium bromide incorporation and visualized under ultraviolet light.

Cell culture and stimulation.

Fresh PBMCs were seeded in serum-free X-VIVO 15 medium containing 2 mM glutamine in an atmosphere of 95% air/5% CO2 at 37°C at a density of 2 × 105 cells/200 μl. Cells were cultured for 22 hours either without stimulation or in the presence of muramyldipeptide (10 μg/ml; Bachem, Bubendorf, Switzerland), Pam3Cys (a TLR-2 agonist) (1 μg/ml; Invitrogen, San Diego, CA), LPS from Escherichia coli 055:B5 (a TLR-4 agonist) purified by gel filtration chromatography (10 ng/ml; Sigma-Aldrich, St. Louis, MO), or combinations of muramyldipeptide and Pam3Cys or LPS, respectively.

Cytokine analysis.

Human IL-1β, IL-12/23p40, and TNFα in culture supernatants were measured by enzyme-linked immunosorbent assays, using a fluorescence detection system according to the manufacturer's protocols (R&D Systems, Minneapolis, MN). Concentrations of each cytokine were calculated based on a standard curve of known amounts of the appropriate recombinant human cytokines (R&D Systems). All determinations were performed in duplicate wells. Supernatants from both Blau syndrome subjects 1, 2, and 3 and control subjects 1, 2, and 3 were assayed separately from Blau syndrome subjects and controls 4 and 5; thus, the data were compiled from different days.

Statistical analysis.

Statistical comparisons were made between the means of data obtained from experimental groups and control groups, using analysis of variance with Fisher's post hoc analysis.

RESULTS

In order to determine whether NOD2 expression occurred at the messenger RNA level, we isolated total RNA from PBMCs and performed RT-PCR. As shown in Figure 1, NOD2 was amplified from the PBMCs of all donors, with a signal similar to or more intense than that of the constitutively expressed GAPDH transcript. Furthermore, we examined the production of IL-1β, a caspase 1–specific cytokine, TNFα, a cytokine that has elevated levels in many inflammatory diseases, and IL-12/23p40, which plays an important role in T cell activation and adaptive immunity. As illustrated in Figures 2A–C, the TLR agonists (LPS and Pam3Cys) stimulated increased release of IL-1β, TNFα, and IL-12/23p40 from PBMCs, as expected. However, cytokine release in response to TLR agonists was not elevated in PBMCs from subjects with Blau syndrome compared with those from healthy control subjects. The NOD2 agonist, muramyldipeptide, stimulated a modest increase in the level of IL-1β in healthy controls compared with unstimulated PBMCs (Figure 2A) but did not elicit increased production of TNFα compared with unstimulated cells or detectable IL-12/23p40 (Figures 2B and C). In contrast to studies on cell lines or purified subsets of cells, this study did not demonstrate a synergistic response when muramyldipeptide was used in combination with LPS or Pam3Cys.

Figure 1.

NOD2 mRNA expression in peripheral blood mononuclear cells (PBMCs) from healthy control subjects (n = 5) and subjects with Blau syndrome (n = 5). Total RNA was extracted from PBMCs for reverse transcriptase–polymerase chain reaction of NOD2, as described in Patients and Methods.

Figure 2.

Cytokine production by peripheral blood mononuclear cells (PBMCs) in response to various stimuli. PBMCs were isolated from healthy control subjects (open boxes) and subjects with Blau syndrome (shaded boxes) and stimulated as described in Patients and Methods. Culture supernatants were analyzed by enzyme-linked immunosorbent assay (ELISA) for A, interleukin-1β (IL-1β), B, tumor necrosis factor α (TNFα), and C, IL-12/23p40. For each cytokine ELISA, the lower limit of detection is indicated by a dashed horizontal line. Data are presented as box plots, where the boxes represent the 25th and 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the highest and lowest values. Values are the mean of duplicate determinations. Unstim. = unstimulated; MDP = muramyldipeptide; LPS = lipopolysaccharide; Pam3Cys = palmitoyl-3-cysteine.

In summary, muramyldipeptide and the TLR agonists failed to induce augmented production of IL-1β, TNFα, or IL-12/23p40 in PBMCs from individuals with Blau syndrome compared with healthy control subjects (no comparisons were statistically significant). In addition, no increased levels of cytokines were detected in unstimulated PBMCs from subjects with Blau syndrome compared with control subjects, and no enhanced synergistic release of cytokines was observed when muramyldipeptide was used in combination with a TLR agonist. Furthermore, there was no correlation between the levels of cytokine release and the presence (patients 4 and 5) or absence (patients 2 and 3) of biologic therapy received by patients with Blau syndrome.

Case reports.

Case 1.

The patient, a 27-year-old woman (Blau syndrome subject 5 in the cytokine synthesis study described above), presented with a long history of bilateral panuveitis, which was diagnosed when the patient was 8 years old. At age 3 years, she reportedly had a diagnosis of polyarticular juvenile arthritis. She described a history of pruritic rashes treated with corticosteroid cream. By age 12 years, the patient had bilateral lensectomies and vitrectomies and was noted to have multifocal chorioretinal scars in the periphery of each eye. She reported taking a high daily dose of prednisone for many years. She had also tried various treatments including rofecoxib, gold injections, cyclosporine, periocular corticosteroid injections, and infliximab. The infliximab was reportedly effective, but the patient could not afford continued therapy. She eventually tolerated a regimen of methotrexate injections, 25 mg weekly, and etanercept, with periodic oral or periocular corticosteroids for flares. Initially, her use of etanercept was sporadic, because she took only what was left over after treating her son, who also has Blau syndrome. Eventually, an assistance program allowed her access to etanercept routinely, and she continued this at a dosage of 25 mg twice weekly. Her baseline visual acuity at this time was 20/100 in the right eye and 20/50+ in the left eye.

After 3 or 4 years of receiving this treatment regimen, she noted a further decline in her peripheral vision. Her central visual acuity remained relatively stable, at 20/200 in the right eye and 20/60 in the left eye. She was also found to have probable liver involvement of Blau syndrome; a liver biopsy showed multiple noncaseating granulomas. Based on the worsening visual field constriction and concern regarding liver toxicity, the methotrexate was stopped. She then had a flare of uveitis and joint disease. Etanercept was discontinued, and she began receiving daily anakinra injections (100 mg/day) per the compassionate use protocol. Methotrexate was restarted at a dosage of 25 mg/week in conjunction with close monitoring of the patient's liver function.

Initially, the patient's ocular inflammation seemed to improve, but her joint disease worsened. After 3–4 weeks, she had increased ocular symptoms and continued worsening of her joint disease. Approximately 6 weeks after starting anakinra, it was evident that this treatment was ineffective, and it was discontinued. She was switched back to etanercept and continued taking methotrexate.

Her joints and eyes continued to do quite well. She decided to stop the etanercept and noted improved central and peripheral vision. Her visual acuity measured 20/50 in the right eye and 20/40 in the left eye while she received methotrexate alone. Over the next year and currently, her vision remains stable, and her eyes and joints are doing well while she receives a regimen consisting of methotrexate 25 mg/week, prednisone 5 mg daily, and naproxen 500 mg daily along with daily omeprazole for gastrointestinal protection.

Case 2.

The patient, a 17-year-old boy (who was not studied for cytokine synthesis) with a prosthetic left eye and uveitic glaucoma and inflammation in the right eye secondary to Blau syndrome, was being treated with infliximab infusions, 5–10 mg/kg every 4–6 weeks. The left eye had been enucleated several years prior due to complications of posterior uveitis, including retinal and optic nerve disease and intractable glaucoma. The patient's visual acuity was 20/50+1 in the right eye and improved with a pinhole to 20/40-2. Despite treatment with infliximab and topical prednisolone acetate 1%, 6 times daily, active inflammation remained. Topical brimonidine 3 times daily, topical timolol/dorzolamide twice daily, and acetazolamide 125–250 mg orally 3 times daily were required for control of intraocular pressure.

Because active inflammation was present, anakinra 100 mg/day subcutaneously was started in place of infliximab. Despite the use of methylprednisolone (10–25 mg/day) and cyclosporine (up to 5 mg/kg/day) along with anakinra, inflammation remained after 2 months of therapy with this regimen, and the patient experienced an acute flare of his uveitis requiring increased daily oral prednisolone.

Anakinra was discontinued, and treatment with adalimumab was substituted. Prednisolone and cyclosporine were continued. Within 1 month, inflammation was improved. His visual acuity was 20/30 in the right eye. Intraocular pressure was controlled with brimonidine, timolol/dorzolamide, and acetazolamide 500 mg twice daily.

DISCUSSION

NOD2 has been implicated in 2 well-documented inflammatory diseases, Crohn's disease and Blau syndrome. Crohn's disease is a genetically complex disease characterized by recurrent and excessive Th1/Th17-dominant immune responses in the gastrointestinal tract, whereas Blau syndrome is a rare autosomal-dominant condition that presents as granulomatous uveal, joint, and skin inflammation. Currently, NOD2 research is mainly focused on Crohn's disease–associated NOD2 polymorphisms, and few studies have characterized the cellular response in patients with Blau syndrome (38, 39). Due to the rareness of Blau syndrome and the limited availability of subjects, we included individuals taking systemic antiinflammatory medications (prednisone and etanercept) in the study. However, because all PBMCs underwent vigorous washing during the isolation process, the PBMCs were unlikely to be exposed to any unbound medication during in vitro culturing. In addition, the cellular response was observed to be comparable between the patients who received systemic medications and those who did not, indicating that the findings of this study reflect the biologic response related to NOD2 mutations and were not influenced by systemic medications taken by the patients.

Several groups of investigators have reported that HEK cells transfected with a NOD2 construct engineered to contain a Blau syndrome mutation demonstrate elevated basal activity of an NF-κB reporter and higher amounts of NF-κB reporter activation in response to muramyldipeptide compared with cells transfected with a wild-type NOD2 construct (30, 31). From these in vitro observations, a “gain of function” hypothesis has been proposed, which predicts that patients with Blau syndrome would spontaneously release more cytokines that can be transcriptionally up-regulated by NF-κB activity (such as IL-1β) and have heightened responses to muramyldipeptide. Indeed, 1 case of Blau syndrome with elevated serum levels of IL-1β has been documented (34).

In contrast to this hypothesis, PBMCs from 5 subjects with classic Blau syndrome and the most commonly observed NOD2 substitutions (R334W and R334Q) did not exhibit an augmented synthesis of cytokines in response to muramyldipeptide, LPS, or Pam3Cys. In fact, some of the IL-1β responses in PBMCs from subjects with Blau syndrome appeared to be slightly attenuated when compared with the responses in PBMCs from the control group (Figure 2A; note either the lower means or the ranges). Furthermore, no synergistic cytokine release was observed when muramyldipeptide was used in combination with either TLR agonist, even though we used agonist concentrations similar to those used in studies in which synergistic responses were detected (35–37, 40). Taken together, these findings suggest that an exaggerated NOD-2 response leading to IL-1 release is not a direct mechanism explaining the pathophysiology of Blau syndrome.

Levels of the other 2 cytokine proteins measured, TNFα and IL-12/23p40, were also not elevated in the PBMCs with NOD2 mutations. The TNFα data might seem to contradict the use of TNF inhibitors as therapy in Blau syndrome. However, it should be noted that the clinical efficacy of TNF blockade has not been rigorously tested, and many patients with Blau syndrome (including the 2 patients described in this report) experience continued inflammation while receiving these treatments.

In addition to seeking a physiologic manner in which to test functional consequences of the NOD2 mutations found in patients with Blau syndrome, we specifically chose PBMCs, because NOD2 is mainly expressed in myeloid cells such as monocytes, macrophages, and dendritic cells. However, NOD2 has also been detected in epithelial cells and vascular endothelial cells (41–43). Thus, the effect of NOD2 mutations on other cell functions needs to be explored in order to understand the pathogenesis of NOD2-related diseases.

Our results suggest that, in contrast to cryopyrinopathies, Blau syndrome is not a disease primarily mediated by IL-1. This may explain our clinical observation that 2 patients with Blau syndrome not only failed to respond to anti–IL-1 therapy, but their disease worsened. In a mouse model of NOD2-dependent eye inflammation (44), we demonstrated that locally injected muramyldipeptide induces IL-1β synthesis within the eye. However, this muramyldipeptide-induced intraocular inflammation (as quantified by leukocyte rolling and sticking within the iris microvasculature) is not reduced in either caspase 1–deficient or IL-1 receptor type I–deficient mice, suggesting that when IL-1β is present, it does not mediate muramyldipeptide-induced inflammation. Recently, a Spanish group of investigators reported the use of anakinra therapy in a patient with Blau syndrome and a novel mutation in NOD2 (8). Although a reduction in serum IL-1β levels was demonstrated, the patient's clinical improvement was not described in detail, and the persistence of ocular symptoms was reported.

In summary, this study is the first to characterize innate immune responses of PBMCs from subjects with Blau syndrome. Interestingly, these PBMCs did not display an enhanced cytokine response to a pathogen-associated molecular pattern challenge. This observation is consistent with the report here of 2 patients in whom disease failed to respond adequately to anakinra therapy. The functional effect of NOD2 mutations in Blau syndrome disease remains to be fully elucidated.

AUTHOR CONTRIBUTIONS

Dr. Martin had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Martin, Zhang, Davey, Rosenbaum.

Acquisition of data. Zhang, Chen, Lu.

Analysis and interpretation of data. Martin, Zhang, Planck, Davey, Rosenbaum.

Manuscript preparation. Martin, Zhang, Kurz, Rosé, Rosenbaum.

Statistical analysis. Martin, Zhang, Planck.

Preparation of case reports. Kurz, Rosé.

Acknowledgements

We appreciate the expert assistance of Jinnell Lewis in obtaining the blood samples.

Ancillary