Introduction

Autoinflammatory syndromes are characterized by systemic inflammation without the presence of antigen-specific T cells or high-titers of autoantibodies [1]. Many autoinflammatory syndromes are clinically characterized by recurrent or persistent features that include fever, elevation in the levels of acute-phase reactants, and organ-specific complications such as skin rashes and osteoarticular, serosal, neurologic, and ocular manifestations [2]. To date, well-known hereditary periodic fever syndromes are familial Mediterranean fever (FMF), hyperimmunoglobulinemia D with periodic fever syndrome, cryopyrin-associated periodic syndromes (CAPS), and tumor necrosis factor (TNF) receptor-associated periodic syndrome (TRAPS). These syndromes are discriminated by some characteristic phenotypes such as varying age of onset, duration of fever, development of cutaneous manifestations, and several other features.

CAPS include familial cold-induced autoinflammatory syndrome (FCAS), Muckle–Wells syndrome (MWS), and neonatal-onset multisystem inflammatory disease (NOMID), also known as chronic infantile neurologic, cutaneous, articular (CINCA) syndrome. FCAS exhibits cold-induced urticaria-like skin rash whereas MWS develops severe phenotypes, such as periodic fever, neural progressive hearing loss and renal amyloidosis. CINCA/NOMID syndrome shows additional more severe phenotypes, such as severe arthritis, patella overgrowth, aseptic meningitis, and mental retardation [3]. CAPS are caused by mutations in the Nod-like receptors (NLRs) family, pyrin domain containing 3 (NLRP3) gene, and more than 80 variants are associated with CAPS, in addition to over 50 variants of unclear significance that have been reported in the INFEVERS database (http://fmf.igh.cnrs.fr/ISSAID/infevers/) to date [4].

NLRs recognize microbial molecules such as pathogen-associated molecular patterns (PAMPs) or endogenous danger-associated molecular patterns, which trigger inflammation as well as Toll-like receptor immune responses. NLRP3 protein contains an N-terminal pyrin domain, a central nucleotide-binding site (NBS) domain, and C-terminal leucine-rich repeats (LRR) [5]. Most pathogenic mutations associated with autoinflammatory syndromes are located in exon 3 of NLRP3, which encodes the NBS domain. In addition, several mutations outside exon 3 on the LRR domain of NLRP3, such as G755R, G755A, and Y859C have been found in patients with CINCA syndrome or atypical autoinflammatory disorders [6–8].

This study reports two cases of Japanese children who presented with atypical periodic fever episodes and who had the variants in the LRR domain of NLRP3 with co-existence of Mediterranean fever (MEFV) haplotype variants. The patients showed periodic prolonged fever and erythema, but lacked symptoms typical of CAPS, FMF, and other common autoinflammatory syndromes. By genetic analysis and functional assays of these variants, the data from this study suggest that the phenotype of atypical autoinflammatory disease in patients could be modified by a synergistic effect with other autoinflammatory-associated genes.

Methods

Subjects

Case 1

The first case was a 9-year-old girl who had experienced recurrent fever episodes approximately three times a year for 6 years from onset at 3 years of age. Although she underwent a tonsillectomy at the age of 5, she still experienced recurrent fever episodes. She presented with mild abdominal pain without signs of peritoneal irritation, peritonitis or pleuritis as typically observed in FMF. High serum C-reactive protein (CRP) levels were observed in the attack phase. She presented with pigmented macules with erythema, which persisted for 6 months, and bilateral petechiae on her legs and dorsa of feet (Fig. 1a). Histological examination of the skin lesion revealed perivascular infiltrate with mononuclear cells in the upper and middle dermis, but vasculitis was not observed (Fig. 1b). Direct immunofluorescence analysis revealed deposits of complement component 3 (C3) at the capillary walls in the upper to middle dermis, but not the presence of immunoglobulin (Ig)A or IgM (Fig. 1c). Rheumatoid factor and autoantibodies were not detected. Colchicine treatment (0.5 mg per day) was effective in treating the erythema and alleviating fever with elevating CRP. Both parents had experienced lasting recurrent fever episodes during their childhood although it was likely that their symptoms were not so severe. The fever episodes of parents resolved spontaneously without specific medications such as colchicines and corticosteroids or tonsillectomy when they were about 10 years old. However, they do not remember their childhood in detail as it was over 30 years ago. Their episodes may represent autoinflammatory disease.

Fig. 1
figure 1

Presence of skin rash in patients with atypical autoinflammatory syndrome. a The clinical appearance of rash on the dorsum of foot in case 1. b The histopathological examination of a skin biopsy specimen (hematoxylin and eosin stain, original magnification × 200). Perivascular infiltrate with mononuclear cells was observed in the upper and middle dermis. c Direct immunofluorescence demonstrates C3 deposits in the capillary walls (original magnification × 50). d The clinical appearance of the skin rash on the breast in case 2

Full size image

Case 2

The second case involved a 4-year-old boy, presenting with recurrent episodes of fever of various duration from a few days to weeks, with or without mild liver dysfunction and multiple erythema without skin itch. The frequency of episodes was at least twice a year. The skin erythema was observed during the fever episodes at 18 months old and at 4 years old (Fig. 1d). The cervical lymphadenopathy and diarrhea were observed in almost all of the fever attack episodes. Although fever duration was 1 week, it resolved immediately following oral administration of 1 mg/kg prednisolone. Rheumatoid factor and autoantibodies were not detected. His parents had no symptoms like periodic fever syndromes or rheumatic diseases. The fever did not recur for a few months after the cessation of oral prednisolone treatment. From 3 years old, colchicine treatment was started because of recurrent fever attacks. However, currently this treatment is not effective.

The genotypes and the clinical profiles of these cases are summarized in Table I. This study was performed according to the Helsinki Declaration. All subjects provided informed consent to participate in the study.

Table I Genotype and clinical profiles of cases
Full size table

DNA Sequencing

Genomic DNA was extracted from leukocytes using SepaGene (EIDIA, Tokyo, Japan). DNA fragments of the NLRP3, MEFV, mevalonate kinase (MVK), and TNF receptor superfamily, member 1A (TNFRSF1A) genes were amplified by polymerase chain reaction (PCR), and analyzed using big dye terminator bidirectional sequencing (Applied Biosystems, Foster City, CA, USA).

Genotyping

Allelic frequency of NLRP3 G809S (rs141389711) was investigated on a Step One Real-Time PCR System using Custom TaqMan SNP Genotyping assays (Applied Biosystems) in 421 healthy subjects. Further, genotype was confirmed by direct sequence analysis.

Cell Culture

Peripheral blood mononuclear cells (PBMCs) were isolated from heparinized blood of control donors and from patients by gradient centrifugation using Ficoll-Paque (GE Healthcare, Uppsala, Sweden). The CD14-positive cells were cultured in medium consisting of RPMI 1640 supplemented with 10 % heat-inactivated fetal calf serum (FCS), L-glutamine (2 mmol/l), penicillin (100 U/ml), and streptomycin (100 μg/ml). Human embryonic kidney (HEK) 293 T cells and HEK293-ASC cells were cultured in high glucose Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10 % heat-inactivated FCS (Sigma-Aldrich, St. Louis, MO, USA), penicillin (100 U/mL), and streptomycin (100 μg/mL).

Analysis of Serum Cytokine Levels by Enzyme-Linked Immunosorbent Assay (ELISA)

Serum samples of patients and controls were stored at –80 °C until assayed. TNF-α concentrations were measured with an Immunoassay Kit (BioSource International, Carlsbad, CA, USA) with a detection limit of 1.7 pg/ml. Similarly, interleukin (IL)-6 and IL-1β concentrations were measured by immunoassay Kit (BioSource) with detection limits of 1.7 pg/ml and 1.0 pg/ml, respectively. IL-1ra and sTNFR1 concentrations were measured by ELISA (R&D Systems) with detection limits of 6.26 pg/ml and 0.77 pg/ml, respectively. IL-18 was assayed by ELISA (MBL, Nagoya, Japan), with a detection limit of 25.6 pg/ml. We defined serum cytokine levels of more than the mean + 2 SD as increasing. Values below the detection limit are shown as not detected.

IL-1β Production from Monocytes

CD14-positive cells were purified from PBMCs using CD14 MACS MicroBeads and MACS magnetic columns according to the manufacturer’s instructions (Miltenyi Biotec, Gladbach, Germany). The CD14 positive cells were seeded to a density of 3.0 × 105 per ml and cultured with the addition of 1.0 μg/mL LPS O127 (Sigma-Aldrich) and 20 μg/ml IFN-γ (R&D Systems, Minneapolis, MN, USA) for 24 h at 37 °C in a humidified atmosphere at 5 % CO2 and pulsed with 5 mM ATP (Sigma-Aldrich) for 30 min before harvesting. The cell-culture supernatants were harvested, and stored at –80 °C until assayed. The IL-1β was measured with ELISA. The assay was performed at two different times. The statistical significance between control and each case in the IL-1β production was analyzed using Dunnett’s multiple comparison test. P-value of <0.05 was considered statistically significant.

Vector Preparations

cDNA encoding NLRP3 tagged at the C-terminus with a FLAG-epitope (NLRP3-FLAG) was cloned into plasmid pcDNA3.1+ (Invitrogen). NLRP3 mutants (D303N, G755R, G809S and Y859C) were generated using the GeneEditor In vitro Site-Directed Mutagenesis System (Promega, Madison, WI, USA). A cDNA encoding pyrin tagged at the C-terminus with an HA-epitope (pyrin-HA) was cloned into plasmid pcDNA3.1+. Pyrin variants (P369S+R408Q) were generated using the GeneEditor in vitro Site-Directed Mutagenesis System (Promega). The apoptosis-associated speck-like protein containing a CARD (ASC) variant 1 tagged at the C-terminus with a myc-epitope (ASC1-myc) construct was cloned into pcDNA3.1+. The NF-κB luciferase reporter vector (pGL4.32-luc2P/NF-kappaB-RE/Hygro) and the Renilla luciferase reporter vector (pGL4.74-hRluc/TK) were purchased from Promega.

NF-κB Reporter Gene Activity

HEK293T cells were transfected with 16 ng per well of pcDNA3.1+ control vector or pcDNA3.1+ NLRP3-FLAG (wild type or mutant) or pcDNA3.1+ pyrin-HA (wild type or mutant) in 96-well plates using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. The pcDNA3.1+ ASC1-myc, NF-κB luciferase reporter, and Renilla luciferase reporter were co-transfected. After transfection, cells were incubated for 24 h. Cells were stimulated with R837 at a concentration of 10 μg/ml (InvivoGen, San Diego, CA, USA) or monosodium urate (MSU) at 250 μg/ml (InvivoGen) for 8 h. Luciferase reporter activity was analyzed using the Dual-Luciferase Reporter Assay System (Promega). The statistical significance of differences in luciferase activity between wild-type and mutant gene activity in the NF-κB reporter assays was analyzed using Dunnett’s multiple comparison test. A P-value of <0.05 was considered statistically significant.

Speck Quantification Assay

HEK293 cells were transfected with ASC-myc and positively selected using 1 mg/ml G418 for 4 weeks. HEK293-ASC cells (1 × 105) were co-transfected with 250 ng of each NLRP3 expression plasmid and pyrin expression plasmid using Lipofectamine LTX (Invitrogen) according to the manufacturer’s instructions. After 24-h incubation, cells were fixed with 3.7 % paraformaldehyde in PBS for 10 min, and washed with 10 mM glycine in PBS. Fixed cells were permeabilized using PBS containing 0.2 % Triton X-100 for 1 h at room temperature. Cells were then incubated with an anti-FLAG M2 monoclonal antibody (Sigma-Aldrich) and anti-myc antibody (Invitrogen). Primary antibody binding was detected by incubation with Alexa Fluor 488 goat anti-mouse IgG and Alexa Fluor 594 donkey anti-rabbit IgG (Invitrogen) secondary antibodies. Fixed cells were incubated with 4’-6-diamidino-2-phenylindole, a nuclear stain, and mounted using Vectashield Mounting Medium (Vector Laboratories Burlingame, CA, USA). The percentage of cells containing ASC specks in the cells expressing NLRP3 was calculated by randomly selecting at least 10 fields. Differences were analyzed using Dunnett’s multiple comparison test. A P-value of <0.05 was considered statistically significant.

Results

Detection of NLRP3 and MEFV Mutations in Two Patients with Autoinflammatory Syndrome

In case 1, a heterozygous c.2425G>A (p.Gly809Ser) on LRR in exon5 of NLRP3 and heterozygous P369S-R408Q in exon3 of MEFV were identified (Table I). There are 17 individuals who have the allele of G809S in 421 healthy control subjects. The allele frequency of this variant was 0.02. There were no control subjects carrying P369S-R408Q in MEFV in addition to the G809S variant. Interestingly, the same NLRP3 and MEFV haplotype variants were identified in the father of case 1. The heterozygous MEFV variant haplotype P369S-R408Q were also observed in the mother of case 1.

Case 2 expressed the same heterozygous NLRP3 variant found in case 1. In addition, heterozygous E148Q-P369S-R408Q in exon2 and exon3 of MEFV were identified (Table I). The G809S variant of NLRP3 was inherited from his asymptomatic father. His asymptomatic mother was positive for homozygous E148Q and heterozygous P369S-R408Q sequences.

MVK and TNFRSF1A mutations were not detected in either case.

The Cytokine Profile of Patients

Serum IL-1β, IL-6 and TNF-α levels were not detected in the sera of healthy control subjects. The mean concentration ± SD of serum IL-18 and IL-1ra in healthy control subjects were 169.2 ± 85.7 pg/ml and 213.4 ± 87.1 pg/ml, respectively [9]. The mean concentration ± SD of serum sTNFR1 in healthy control subjects was 1009 ± 276.4 pg/ml. Figure 2a and b show the serum cytokine profiles from the patients. The serum cytokine concentrations were measured at two different points at least during fever and inter-ictal periods respectively, and average values were calculated. In both cases, serum IL-1β, TNF-α, and IL-6 did not increase during the fever episodes, whereas serum IL-1ra and sTNFR1 levels were increased. IL-18 levels during the fever episodes were increased in case 2, not in case 1. Interestingly, the serum IL-1ra and IL-18 levels from case 2 were elevated during the inter-ictal period.

Fig. 2
figure 2

Inflammatory cytokine levels from two cases during the inter-ictal phase and attack phase. a White bars indicate serum inflammatory cytokine levels of control. Grey bars indicate serum inflammatory cytokine levels of case 1 during the inter-ictal period. Black bars indicate serum inflammatory cytokine levels of case 1 during the attack phase. b White bars indicate serum inflammatory cytokine levels of control. Grey bars indicate serum inflammatory cytokine levels of case 2 during the inter-ictal period. Black bars indicate serum inflammatory cytokine levels of case 2 during the attack phase

Full size image

Figure 3 shows the production of IL-1β from monocytes with LPS, IFN-γ and/or ATP stimulation. The mean concentration ± SD of IL-1β from monocytes of healthy control subjects (n = 5) without stimulation were 5.54 ± 4.40 pg/ml. The mean concentration ± SD of IL-1β from monocytes of healthy control subjects stimulated with 20 ng/ml IFN-γ or 1 μg/ml LPS were 7.74 ± 9.81 pg/ml and 236.0 ± 188.4 pg/ml, respectively. The mean concentration ± SD of IL-1β from monocytes of healthy control subjects stimulated with 1 μg/ml LPS added 5 mM ATP was 166.0 ± 138.3 pg/ml. The mean concentration ± SD of IL-1β from monocytes of healthy control subjects stimulated with both 20 ng/ml IFN-γ and 1 μg/ml LPS were 441.3 ± 316.5 pg/ml. The mean concentration ± SD of IL-1β from monocytes of healthy control subjects stimulated with both 20 ng/ml IFN-γ and 1 μg/ml LPS added 5 mM ATP was 549.2 ± 327.3 pg/ml. In both cases, IL-1β secretion was increased compared with the healthy controls when the monocytes were stimulated with LPS and IFN-γ. Additionally, IL-1β from monocytes in case 2 stimulated with LPS and IFN-γ was increased in response to ATP. This was not observed for monocytes from case 1.

Fig. 3
figure 3

IL-1β levels from monocytes in case 1 and 2. White bars indicate IL-1β levels in control. Grey bars indicate IL-1β levels in case 1. Black bars indicate IL-1β levels in case 2. IL-1β levels from monocytes from case 1 and 2 were significantly increased compared with controls (n = 5). * P < 0.05

Full size image

NF-κB Reporter Gene Activity of NLRP3 and Pyrin Variants

To assess the function of the NLRP3 variant G809S on NF-κB signaling, we compared the G809S sequence with those of wild-type and three NLRP3 mutations (D303N, G755R and Y859C). D303N, G755R, and Y859C were identified in CAPS patients [6, 10, 11] (Fig. 4). When ASC was co-expressed, D303N and G755R mutations increased NF-κB reporter gene activity. However, G809S and Y859C did not lead to significant activation of NF-κB. In the presence of R837, an NLRP3 inflammasome activator, NLRP3 D303N and G755R mutations showed enhanced NF-κB activation, whereas G809S and Y859C did not induce any increase in activity. Subsequently, the evaluation of G809S enhanced NF-κB activation in the presence of MSU was measured. MSU induced NF-κB activation of wild-type, D303N and G755R NLRP3. However, both G809S and Y859C mutations significantly inhibited NF-κB activation mediated by MSU.

Fig. 4
figure 4

NF-κB reporter gene activity of NLRP3 mutants. Bars represent the mean ± SD of triplicate assays. a White bars indicate the NF-κB reporter gene activity of the NLRP3 mutants without co-transfection of ASC. Black bars indicate activity with co-transfection of ASC. ASC-dependent NF-κB reporter gene activity was increased by mutants D303N and G755R. G809S and Y859C did not induce NF-κB reporter gene activity. b White bars indicate NF-κB reporter gene activity with co-transfection of ASC. Black bars indicate activity after stimulation with 10 μg/ml R837. c White bars indicate NF-κB reporter activity following co-transfection of ASC. Black bars indicate activity after stimulation with 250 μg/ml MSU. * P < 0.05

Full size image

To investigate the role of mutational effect of pyrin in the NF-κB signaling pathway, wild-type or variant pyrin (P369S+R408Q) was expressed in HEK293 cells and co-transfected with ASC (Fig. 5). Although both wild-type and variant pyrin inhibited NF-κB activation with co-transfection of wild-type or mutant NLRP3 protein, there was no significant difference in inhibitory capacity between the wild-type and variant pyrin.

Fig. 5
figure 5

Pyrin and NLRP3 mutant-induced NF-κB reporter gene activity. Bars represent the mean ± SD of triplicate assays. White bars indicate NF-κB reporter gene activity with co-transfection of ASC. Grey bars indicate activity with co-transfection of ASC and wild-type pyrin. Black bars indicate activity with co-transfection of ASC and pyrin variant P369S+R408Q. * P < 0.05

Full size image

Speck Quantification Assay

Previous studies have shown that NLRP3 LRR variants have an increased ability to induce speck formation in the presence of ASC [6, 12]. To test the effect of G809S on NLRP3-ASC interactions and speck formation, wild-type, NLRP3 variants or empty vectors and pyrin were transiently transfected with cells stably expressing ASC. Cells transfected with NLRP3 wild-type displayed speck formation (mean ± SD, 36.7 ± 6.1 %). In comparison, the NLRP3 D303N, G755R, G809S and Y859C mutants induced significantly higher numbers of speck formation (62.1 ± 8.8 %, 72.6 ± 4.8 %, 53.1 ± 10.1 % and 48.8 ± 13.2 % respectively, Fig. 6).

Fig. 6
figure 6

Effect of the G809S variant on speck formation. Transfection of HEK293-ASC cells with 250 ng each of the NLRP3 expression plasmids or an empty vector and pyrin expression plasmid was performed. Speck formation was assessed by immunofluorescence microscopy. a This panel shows examples of fields obtained by immunofluorescence microscopy. b The percentage of cells containing ASC-myc specks was calculated as the mean ± SD percentage of cells. * P < 0.05

Full size image

Discussion

The current study identified a G809S variant within the LRR domain of NLRP3 with the co-existence of MEFV haplotype variants in two unrelated patients with atypical autoinflammatory syndrome. Although we recently reported a CINCA/NOMID patient with the compound heterozygous gene mutations E688K and G809S, it is unclear whether G809S is a pathogenic mutation [9]. To confirm a functional role for the G809S variant, its effect on the NF-κB signaling pathway was investigated in vitro. Although several variants of NLRP3 show significant increases of ASC dependent NF-κB reporter gene activity in a previous report and as data presented here, no significant increase was observed owing to the NLRP3 G809S variant in this assay. Kambe et al. demonstrated that the NLRP3 G755R mutation located within the LRR domain could induce significant NF-κB activation in the presence of an NLRP3 inflammasome activator, R837 [13]. Therefore, G809S may be expected to enhance NF-κB activation in the presence of R837. However, G809S did not increase NF-κB activity like as Y859C [6] (Fig. 4b). Since the NLRP3 LRR domain plays a central role in mediating inflammation induced by another inflammasome activator, MSU crystals, we examined whether G809S affected NF-κB activation in the presence of MSU [14]. Interestingly, G809S and Y859C mutations did not show any NF-κB activity responses by MSU stimulation. In contrast, wild-type, D303N and G755R mutations significantly increased NF-κB activity following MSU stimulation. These data suggest the G809S LRR missense variant, which may diminish the responsiveness to PAMPs as NOD2 LRR variant reported in Crohn’s disease, has a pathogenic effect on these pathways [15–17].

Jéru et al. recently identified a pathogenic Y859C mutation in the LRR domain of NLRP3, which increased speck formation and pro-caspase 1 processing, but which had no direct effect on NLRP3 mediated NF-κB signaling. The G809S variant also increased speck formation relative to wild-type NLRP3. These results suggest that G809S, as well as Y859C in the LRR domain, may be a gain of function variant. It should be noted that although the assays used in this study are sensitive, our findings may provide limited evidence to prove that the G809S variant is pathogenic. However, these results indicate that the variant alters the function of NLRP3.

The two case studies presented here consistently showed elevated IL-1-related serum cytokines, IL-1ra, during the attack phase. In addition, monocytes from case 1 and 2 secreted high levels of IL-1β, which may indicate a gain of function variant in NLRP3, associated with inflammasome activation. Additionally, we previously reported a CINCA/NOMID patient positive for the compound heterozygous gene mutations, E688K and G809S [9]. This patient developed severe a phenotype compared with her mother, who carried a single mutation, E688K. This genotype-phenotype correlation suggests that the G809S variant may act as an additional genetic factor associated with the severity of CAPS.

However, in this study IL-1β was not detectable in the serum of patients, as IL-1β might be rapidly neutralized, metabolized, or captured by a plethora of IL-1 receptors in vivo. Furthermore, although elevated serum IL-18, which is activated by caspase-1 as well as IL-1β, and IL-6 levels were observed in CINCA/NOMID patient [9], the serum IL-18 levels were increased in case 2 but not case 1, and serum IL-6 levels in both cases did not increase during the fever episodes. Thus, it may be considered that the differences of cytokine profiles and disease phenotypes between case 1 and 2 and typical CINCA/NOMID patients result not only from their genetic background, but also environmental factors.

Additional mutation analysis of our patients also revealed heterozygous variant haplotype of MEFV, a gene involved in the pathogenesis of FMF, in addition to G809S in NLRP3. Case 1 was heterozygous for P369S and R408Q in cis and case 2 was heterozygous for E148Q, P369S, and R408Q in cis. Allele frequencies of P369S and R408Q in the Japanese population are 3.6 % and 4.8 %, respectively, according to the International HapMap Project (http://www.hapmap.org/). These frequent variant haplotypes were found to be in strong linkage disequilibrium in the Japanese population. In addition, P369S and R408Q variant haplotype are associated with a variable phenotype and are infrequently associated with typical FMF symptoms [18–21]. Heterozygous P369S and R408Q variant haplotype are also associated with other inflammatory diseases, such as Behçet’s disease [18], and systemic lupus erythematosus [21]. Moreover, heterozygous E148Q-P369S-R408Q variant haplotype is more rare, which is associated with chronic recurrent multifocal osteomyelitis [20]. In this report, case 1 and case 2 showed the similar phenotypes as FMF or TRAPS, respectively. Although detailed clinical features and cytokine profiles of the two cases are various, they exhibited a long duration of recurrent fever episodes compared with typical FMF. Thus, these findings suggest that P369S and R408Q variant haplotype may have effects on several inflammatory diseases, but the functional evidence of these variant haplotype remains unclear.

The MEFV gene codes for pyrin, that can interact with ASC to induce ASC oligomerization and the activation of procaspase-1, which promotes IL-1β and IL-18 processing [12, 22]. In contrast, some reports have described that pyrin inhibited NLRP3-mediated NF-κB activation by disrupting the NLRP3-ASC interaction [23, 24]. In accordance with the reports, co-expression of NLRP3 and pyrin in HEK293T cells indicated less ASC-dependent NF-κB activation than expression of NLRP3 only, whereas there was no difference in the inhibitory capacity of NF-κB activity between pyrin variants and the wild-type protein. Interestingly, a recent study using pyrin deficient and mutated pyrin knock-in mice demonstrated a gain of function with pyrin variants located in B-Box domains, which caused autoinflammatory phenotypes [22]. Thus, research using knock-in mouse experiments with MEFV exon3 variants into pyrin deficient mice would help clarify the pathogenic effects of the MEFV variant.

In general, hereditary periodic fever syndromes have been considered monogenic diseases. On the other hand, recent reports have described patients with heterozygous low penetrance variants in two recurrent fever genes [2, 25, 26]. These indicate that oligogenic inheritance has been related to pathogenesis of autoinflammatory diseases. In some cases, patients presented with specific symptoms of both diseases or with a more severe phenotypes. Although the patients in this study were positive for the NLRP3 variant, they did not present with typical symptoms of CAPS, such as deafness or cold-induced rash. In addition, variants in MEFV have been detected in both cases, but they also lacked typical FMF symptoms. However, both cases had obviously periodic fever episodes. These suggest the presence of oligogenicity and that variants in NLRP3 and MEFV synergistically modify the symptoms of the atypical autoinflammatory diseases.

There are two important limitations in this study when discussing the pathogenicity of low penetrance rare variants. The first limitation is the limited number of patients in the study. Further study using a large number of patients is necessary to confirm our results. Secondly, we only analyzed a limited number of genes. In this study, we concluded that the presence of an NLRP3 variant with the co-existence of MEFV variants contributed to atypical autoinflammatory disease. However, the patients may have had alternative genetic mutations or other rare variants of inflammasome related genes such as CARD8 [27] elsewhere in the genome, which are truly disease causing, and the two variants described in these patients may be unrelated.

Conclusions

This study describes the molecular analysis of two cases with heterozygous low penetrance variants in exon5 of NLRP3 and exon3 of MEFV. The findings provide in vivo and in vitro evidence for the effect of an NLRP3 missense variant. Importantly the mutations are within the same signaling pathway and are associated with inflammasome activation. Our observations suggest that oligogenic inheritance may occur in patients with atypical autoinflammatory syndrome. It is therefore important to consider that the phenotypes could be modified by synergistic effects with plural autoinflammatory-associated gene mutations when the patients have atypical autoinflammatory disease.