Periodic fever with aphthous stomatitis, pharyngitis, and cervical adenitis (PFAPA) is a relatively common autoinflammatory condition that primarily affects children. Although tendencies were reported for this syndrome, genetic variations influencing risk and disease progression are poorly understood. In this study, we performed next-generation sequencing for 82 unrelated PFAPA patients and identified a frameshift variant in the CARD8 gene (CARD8-FS). Subsequently, we compared the frequency of CARD8-FS carriers in our PFAPA cohort (13.9%) with a healthy local population group (3.2%) and found a significant association between the CARD8-FS polymorphism and risk for PFAPA syndrome (p = 0.012; odds ratio: 4.96 [95% confidence interval, 1.33–18.47]). Moreover, CARD8-FS carriers display a distinct PFAPA phenotype that is characterized by a higher prevalence of symptoms out of flares and oral aphthosis (both p = 0.02 compared with PFAPA patients without the frameshift variant). CARD8 encodes a protein component of the NLRP3 inflammasome, which plays an important role in inflammation and contributes to the pathology of various autoinflammatory diseases. We found that the CARD8-FS variant led to a truncated CARD8 protein lacking the FIIND and CARD domains. As a result, the mutant CARD8 protein lost the ability to interact with the NOD domain of NLRP3. In summary, these results identify a new CARD8 variant associated with PFAPA and further suggest that disruption of the interaction between CARD8 and NLRP3 can regulate autoinflammation in patients.

Periodic fever with aphthous stomatitis, pharyngitis, and cervical adenitis (PFAPA) syndrome is characterized by recurrent fever flares accompanied by pharyngitis, adenitis, and/or aphthous stomatitis without evidence of infection, relatively asymptomatic intervals between the flares, and normal growth and development. Onset before the age of 5 y used to be considered a criterion based on the first patients described; however, observations of classical PFAPA phenotype with later disease onset suggest that relying on age at onset to confirm the diagnosis might not be appropriate (13).

PFAPA syndrome represents the most common cause of recurrent fever in children in European populations, and an annual incidence of 2.3 cases per 10,000 children per year was recently reported (4). The diagnosis is based on clinical criteria and the exclusion of other causes of recurrent fever. Most of the time, these febrile episodes resolve spontaneously before adulthood. The rapid response of the fever attacks to a single dose of corticosteroid points to an inflammatory origin, and the absence of an infectious or autoimmune cause supports the hypothesis that PFAPA syndrome is an autoinflammatory disease (5). Familial susceptibility suggests a potential genetic basis for some cases (5), although no causative gene defect has been identified. We hypothesized that, as in other autoinflammatory diseases, IL-1β might play a central role in the pathogenesis of PFAPA secondary to a genetic defect that leads to a dysregulation of IL-1β secretion. This hypothesis was supported by the fact that, during PFAPA febrile attacks, an increase in IL-1β was shown at the transcript and protein levels (6, 7).

The NLRP3 inflammasome is a crucial multiprotein complex involved in IL-1β secretion. Upon activation, the sensor protein NLRP3 assembles together with the adaptor protein ASC and procaspase-1 (reviewed in Ref. 8). Formation of the inflammasome leads to the autoproteolytic maturation of caspase-1, which subsequently results in maturation and extracellular release of the proinflammatory cytokines IL-1β and IL-18. Gain-of-function mutations in exon 3 of the NLRP3 gene were implicated in hereditary inflammatory diseases, often classified as cryopyrin-associated periodic syndromes (CAPS), suggesting the presence of functionally important sites in this region (9). Interestingly, we previously found NLRP3 variants, mainly the polymorphism Q705K (rs35829419) in exon 3, (6) which is a gain-of-function variant leading to moderately increased IL-1β (10) that does not drive a typical CAPS phenotype per se but was associated with PFAPA syndrome, suggesting that even moderate dysregulation of inflammasome activation can contribute to the development of autoinflammatory diseases. Moreover, it was suggested that CAPS-like disease manifestations in patients with low-penetrance NLRP3 variants are caused by a physiopathology other than solely classical caspase-1 hyperactivation (11), highlighting the multifaceted nature of inflammatory diseases like CAPS and PFAPA. CARD8 (also called TUCAN or Cardinal) is a possible regulator of inflammasome activation (12). CARD8 is composed of an N-terminal function to find (FIIND) domain and a C-terminal CARD domain. CARD8, through its interaction with caspase-1 [via CARD–CARD interaction (13)] or with NLRP3 [via FIIND–NOD interaction (12)] was proposed to negatively regulate the NLRP3 inflammasome.

We previously showed that a familial aggregation was observed in almost 50% of patients with PFAPA syndrome, indicating a potential genetic origin. It was demonstrated recently that PFAPA results from oligogenic or complex inheritance of variants in multiple disease genes (14). These observations, together with the association of dysregulated levels of IL-1β, raised the question about which genetic variants predispose an individual to PFAPA. Therefore, we screened for variants of genes involved in pathways associated with autoinflammatory conditions, such as familial Mediterranean fever (gene MEFV), TNFR-associated periodic syndrome (gene TNFRSF1A), and hyper-IgD syndrome (gene MVK). We also extended our screening to inflammasome-related genes, such as NLRPs (1–14), and some regulator genes to gain fundamental insights into genetic alterations that are potential predisposing factors for PFAPA.

All children with PFAPA syndrome attending the pediatric rheumatology consultation of Western Switzerland at the Lausanne University Hospital and the Geneva University Hospital, from November 2009 to November 2013, were asked to participate in this study. Previously published diagnostic criteria (15) were applied to our patients. However, because PFAPA is not a well-defined disease, and there are no confirmatory tests, the power of these criteria remains limited. A new classification based on the opinion of a large panel of experts and confirmed in a large cohort of PFAPA patients is underway and should provide improvements in the diagnosis of PFAPA. Consequently, patients with minor symptoms between flares, mainly aphthous stomatitis and abdominal pain, and patients with disease onset after 5 y of age were included according to the published international multicenter cohort of 301 patients (3). Thomas et al. (15), who proposed the criteria based on a series of 94 patients, also described moderate symptoms outside of the flares in a few of them. Patients with all three constitutional symptoms were described as having complete cluster, whereas incomplete cluster was used to describe patients with one or two constitutional symptoms. All of these children had normal growth and development and did not present any other symptoms suggesting an alternative diagnosis. Exclusion of monogenic autoinflammatory diseases was performed based on clinical evaluation. Genomic DNA was extracted from EDTA blood. Mutation hotspots in the four major genes associated with monogenic periodic fevers (MEFV, TNFRSF1A, MVK, and NLRP3) were analyzed using PCR and DNA sequencing. The regions analyzed were as follows: MEFV exons 2 and 10 (detects >95% of known pathogenic familial Mediterranean fever mutations), TNFRSF1A exons 2, 3, 4, and 6 (detects close to 100% of pathogenic TNFR-associated periodic syndrome mutations), MVK exons 9 and 11 (detects ∼70% of pathogenic hyper-IgD syndrome), and NLRP3 exon 3 (detects close to 100% of pathogenic CAPS mutations).

All subjects received code numbers to ensure anonymity. Approval for the study was obtained from the Cantons’ Ethical Committees in Lausanne and in Geneva, and the parents/caregivers gave written informed consent, according to local ethical regulations.

DNA samples from 100 anonymous healthy Swiss volunteers used in the Sanger validation were obtained through Interregionale Blutspende (Bern, Switzerland).

Targeted gene capture and next-generation sequencing (NGS) were carried out at the Novartis Institutes for Biomedical Research facility to sequence 32 genes of interest (Supplemental Table I). Multiplexed sequencing libraries were prepared from genomic DNA using the NuGEN Ovation Ultralow DR Multiplex protocol. Genomic regions of interest were selectively captured from the libraries using the Agilent Technologies SureSelectXT Target Enrichment System protocol (SureSelect Design ID# 0500861; 1.561 Mbp). The multiplexed sequencing libraries were sequenced on an Illumina HiSEquation 2500 instrument with a 2 × 76-bp read length.

All sequence analyses performed for this study and genomic locations reported were based on reference human genome assembly HG19. Sequence reads were aligned to the reference human genome with the Burrows-Wheeler Aligner (16). Downstream processing was performed according to the Genome Analysis Toolkit (17) Best Practices recommendations (18, 19) (https://www.broadinstitute.org/gatk/guide/best-practices), including base quality score recalibration, indel realignment, and duplicate removal, using standard hard filtering parameters. Variant calls were made with the Genome Analysis Toolkit UnifiedGenotyper and annotated with the Ensembl Variant Effect Predictor (20). All variants with base coverage < 5× or variant quality score log-odds score < 2 were considered low quality and eliminated from subsequent analyses.

Our analysis strategy focused on identifying uncommon, protein-altering variants that are observed more frequently in our PFAPA cohort compared with healthy normal controls. We only considered variants that are novel or with a minor allele frequency (MAF) ≤ 5% in an in-house database assembled using data from various sources, such as the 1000 Genome Project (17), the Single Nucleotide Polymorphism database (dbSNP) (21), and the National Heart, Lung, and Blood Institute GO Exome Sequencing Project (http://evs.gs.washington.edu/EVS/). We also kept variants for which no MAF information is available and treated them as if MAF = 0. Furthermore, we sought variants that are predicted to alter the amino acid sequence of a resulting protein product and, thus, are likely to have a functional impact on that protein.

The frameshift variant in the CARD8 gene (CARD8-FS), identified by NGS analysis, was validated by Sanger sequencing on a 3730XL Genetic Analyzer instrument (Thermo Fisher Scientific) at the Novartis Institutes for Biomedical Research facility. Mutation Surveyor (version 4.0.0; SoftGenetics) was used to perform variant detection on Sanger sequence traces.

NLRP3, CARD8, ASC, and CASP1 were subcloned into mammalian expression vector pCR3 in frame with N-terminal vesicular stomatitis virus (VSV) or FLAG tags, as previously described (22). CARD8-FS was introduced into a CARD8-expressing pCR3 plasmid, and PCR was performed with 5′-CCTCTTATGCTTCTAAAAAGTCTGTTTTGAGATCG-3′ and 5′-CGATCTCAAAACAGACTTTTTAGAAGCATAAGAGG-3′ mutagenesis primer pair and Pfu Ultra AD enzyme (600385; Agilent). Template plasmid was digested with DpnI enzyme (R0176S; New England Biolabs). Sequencing certified correct insertion of the frameshift mutation into CARD8.

HEK293T cells were transfected with the above-described plasmids using polyethylenimine transfectant. Twenty-four hours after transfection, cells were washed with PBS and lysed by three freeze/thaw cycles in 1% Nonidet P-40, 20 mM Tris-HCl (pH 7.4), and 150 mM NaCl (lysis buffer). A fraction of cell extracts was kept as input controls, and the rest was precleared by a 1-h incubation at 4°C with Sepharose-6B beads (6B100; Sigma) before subsequent pull-down by 1 h incubation at 4°C with mouse anti-Flag M2 beads (A2220; Sigma). M2 beads were washed three times with 0.1% Nonidet P-40, 20 mM Tris-HCl (pH 7.4), and 150 mM NaCl (wash buffer). Beads and inputs were loaded on SDS-PAGE. Rabbit anti-Flag (PA1-984B; Pierce) and rabbit anti-VSV (V4888; Sigma) Abs were used for immunoblotting.

HEK293T cells were seeded in a 24-well plate at a density of 1 × 105 cells per well 1 d before transfection in DMEM supplemented with 10% FCS and penicillin-streptomycin. Cells were transfected with 100 ng of plasmid encoding for NLRP3 or empty vector, 20 ng of CASP1, 50 ng of ASC, and 200 ng of CARD8/CARD8-FS/pCR3 (empty vector). Mock transfections were done with 370 ng of pCR3. Transfections were done in Opti-MEM with a polyethylenimine/DNA ratio of 1:1 (w/w). Six hours after transfection, medium was replaced with fresh DMEM with 10% FCS and penicillin-streptomycin. Cells were harvested 48 h after transfection, and whole-cell lysates were loaded on 15% SDS-PAGE for Western blotting. Immunoblots were done using anti-human caspase-1 (Adipogen) and anti-tubulin (Sigma).

The Fisher exact test (two-tailed) was performed to investigate the association between the CARD8-FS polymorphism and risk for PFAPA syndrome. A p value <0.05 was considered statistically significant.

Differences in clinical characteristics between CARD8-FS carriers and noncarriers were analyzed using a two-sample test of proportion with 95% confidence interval and the Wilcoxon signed-rank test. Caspase-1 cleavage assays were analyzed using one-way ANOVA, followed by the Newman–Keuls multiple-comparison test.

Eighty-two unrelated PFAPA patients were included in our cohort. We performed targeted sequencing on 32 putative candidate genes that were previously found to be involved in autoinflammatory diseases or encoding components of the inflammasome, such as NLRPs and regulators of inflammasomes (Supplemental Table I). We only considered variants that are novel or with an MAF < 5%. Furthermore, we sought variants that are predicted to alter the amino acid sequence of a resulting protein product and, thus, are likely to have a functional impact on that protein. The average coverage on the targeted regions is between 40 and 125× per sample (with the exception of idv_27709, for which the coverage is 13× because of suboptimal DNA quality). In total, we identified 2613 small variant sites (<50 bp) in the targeted genes where at least one sample has a heterozygous genotype call after filtering out low-quality calls; 994 variants that are novel or with MAF < 5% were retained for further analyses.

Of the 82 unrelated PFAPA patients in our cohort, 12 were found to carry a heterozygous variant in the CARD8 gene. The 2-bp TT insertion at position chr19:48735017 was predicted to cause a frameshift in the reading frame, and the resulting transcript was predicted to be a target of nonsense-mediated decay (Fig. 1). CARD8-FS was previously observed in the 1000 Genome Project (17), estimated to have MAF = 4%, and annotated as rs140826611 in the Single Nucleotide Polymorphism database (dbSNP) (21). However, no experimental validation or functional investigation on this mutation had been done.

FIGURE 1.

Schematic exon organization of the CARD8 and CARD8-FS transcripts. Open reading frames are indicated by gray boxes. In CARD8, the N-termini undergoing alternative splicing are referred to as the V region. The open reading frames following exon 7 are more conserved and code for the FIIND and CARD domains. The CARD8-FS variant harbors the rs140826611 single-nucleotide polymorphism in exon 7. The resulting frameshift generates a unique CARD8-FS–specific C terminus (black box). The prototypical CARD8-FS amino acid sequence is shown; the mutation is outlined by a square box, and the unique CARD8-FS C terminus is underlined. The asterisk indicates a stop codon (in this instance created by the frameshift mutation).

FIGURE 1.

Schematic exon organization of the CARD8 and CARD8-FS transcripts. Open reading frames are indicated by gray boxes. In CARD8, the N-termini undergoing alternative splicing are referred to as the V region. The open reading frames following exon 7 are more conserved and code for the FIIND and CARD domains. The CARD8-FS variant harbors the rs140826611 single-nucleotide polymorphism in exon 7. The resulting frameshift generates a unique CARD8-FS–specific C terminus (black box). The prototypical CARD8-FS amino acid sequence is shown; the mutation is outlined by a square box, and the unique CARD8-FS C terminus is underlined. The asterisk indicates a stop codon (in this instance created by the frameshift mutation).

Close modal

We analyzed the samples for further validation by Sanger sequencing. The Sanger-determined genotype calls for the CARD8-FS variant confirmed the NGS-determined genotype status for all subjects (i.e., 11 [13.9%] heterozygous carriers and 68 noncarriers), with the exception of three subjects for which the Sanger genotype calls were inconclusive (Supplemental Table II). No individual had a homozygous CARD8-FS variant. Subsequent statistical analyses comparing CARD8-FS carriers and noncarriers were performed using only Sanger-confirmed results.

To compare the frequencies of CARD8-FS carriers in our PFAPA cohort with a general healthy population group, we also analyzed DNA samples of 100 healthy controls by Sanger sequencing. Because ethnicity can be an important confounding factor in genetic studies, control DNA samples were collected from Swiss volunteers to match as closely as possible the genetic backgrounds of the PFAPA cohort. Of the 100 healthy control samples, 3 (3.2%) were heterozygous carriers of CARD8-FS, and 92 were noncarriers (inconclusive for the remaining 5). Comparing the frequency distribution of the two groups (Supplemental Table III), our data show a statistically significant association between CARD8-FS polymorphism and risk for PFAPA syndrome (p = 0.012; odds ratio, 4.96 [95% confidence interval 1.33–18.47]).

The clinical characteristics of the CARD8-FS carriers are shown in Table I. The median age at onset of the disease was 1 y and 11 mo, the male/female ratio was 0.8, the median interval between fever attacks was 30 d, and the median duration of flares was 4 d. The onset of flares occurred after the age of 5 y in two patients. All patients presented with at least one constitutional symptom, and six had a complete cluster. Seven patients had a family history of recurrent fever, recurrent pharyngitis, or tonsillectomy. All patients (7/7) who had been treated with steroids responded within 12 h. One CARD8-FS patient also had the Q705K NLRP3 genetic variant, another had Q705K NLRP3 and E148Q MEFV genetic variants, and one had the c.769-7T > 6 MVK variant. All three of these patients exhibited symptoms out of flares.

Table I.
Clinical characteristics of the CARD8-FS+ PFAPA patients
Patient No.Age at Onset (y)SexDuration of Fever (d)Interval (d)Constitutional SymptomsOther SymptomsPositive Family HistoryResponse to SteroidsMinor Symptoms between FlaresPresence of Other Variants in Genes: NLRP3, MEFV, MVK, TNFRSF1A
1 (Idv_26389) 35 P, APH, AR, My, H, AbdP, V None Yesa AR, APH  
2 (Idv_26813) 8.17 30 P, APH AR, AbdP None Yesa No  
3 (Idv_27360) 15 P, AD, APH H, AbdP None NA AbdP NLRP3: Q705K het 
4 (Idv_27461) 0.25 2.5 42 P, AD, APH H, AbdP Mother and aunt: tonsillectomy NA No  
5 (Idv_27466) 2.83 30 P, AD, APH  Sister: PFAPA; mother: tonsillectomy NA AbdP NLRP3: Q705K het MEFV: E148Q het 
6 (Idv_27770) 1.67 30 P, AD, APH  Mother and father: tonsillectomy Yesa No  
7 (Idv_28014) 0.75 15 P, AD AbdP, D Father, mother, and uncle: tonsillectomy Yesa No  
8 (Idv_28152) 35 P, APH None NA APH MVK: c.769-7T > 6 het 
9 (Idv_28294) 0.83 21 P, AD, APH AbdP Father: recurrent fever Yesa No  
10 (Idv_28852) 28 P, AD, APH AbdP Father: recurrent pharyngitis during childhood, uveitis, HLAB27+ Yesa No  
11 (Idv_29769) 1.92 35  Aunt: recurrent pharyngitis Yesa No  
Patient No.Age at Onset (y)SexDuration of Fever (d)Interval (d)Constitutional SymptomsOther SymptomsPositive Family HistoryResponse to SteroidsMinor Symptoms between FlaresPresence of Other Variants in Genes: NLRP3, MEFV, MVK, TNFRSF1A
1 (Idv_26389) 35 P, APH, AR, My, H, AbdP, V None Yesa AR, APH  
2 (Idv_26813) 8.17 30 P, APH AR, AbdP None Yesa No  
3 (Idv_27360) 15 P, AD, APH H, AbdP None NA AbdP NLRP3: Q705K het 
4 (Idv_27461) 0.25 2.5 42 P, AD, APH H, AbdP Mother and aunt: tonsillectomy NA No  
5 (Idv_27466) 2.83 30 P, AD, APH  Sister: PFAPA; mother: tonsillectomy NA AbdP NLRP3: Q705K het MEFV: E148Q het 
6 (Idv_27770) 1.67 30 P, AD, APH  Mother and father: tonsillectomy Yesa No  
7 (Idv_28014) 0.75 15 P, AD AbdP, D Father, mother, and uncle: tonsillectomy Yesa No  
8 (Idv_28152) 35 P, APH None NA APH MVK: c.769-7T > 6 het 
9 (Idv_28294) 0.83 21 P, AD, APH AbdP Father: recurrent fever Yesa No  
10 (Idv_28852) 28 P, AD, APH AbdP Father: recurrent pharyngitis during childhood, uveitis, HLAB27+ Yesa No  
11 (Idv_29769) 1.92 35  Aunt: recurrent pharyngitis Yesa No  
a

Response within 12 h.

AbdP, abdominal pain; AD, cervical adenitis; APH, aphthous stomatitis; AR, arthralgia; D, diarrhea; F, female; H, headache; het, heterozygous; M, male; My, myalgia; NA, not available, patients never treated with steroids; P, pharyngitis; V, vomiting.

We also studied the clinical characteristics of the CARD8-FS PFAPA patients in our cohort. The median age at onset of the disease was 1 y and 5 mo, the median interval between fever attacks was 30 d, and the median duration of flares was 3.5 d. The onset of flare occurred after the age of 5 y in 11 patients. All patients presented with at least one constitutional symptom, and 21 had a complete cluster. Thirty patients had a positive family history. Of 58 patients treated with steroids, 57 had a good response.

A comparison of the clinical characteristics between CARD8-FS carriers and noncarriers is shown in Table II. There is no significant difference in the median age at onset or in the prevalence of onset after the age of 5 y. Furthermore, the median duration of fever attacks and the median interval between the flares are similar in both groups. However, we observed a statistically significantly higher incidence of symptoms out of flares in CARD8-FS carriers in comparison with noncarriers (p = 0.02). The prevalence of oral aphthosis was also elevated in the CARD8-FS PFAPA patients (p = 0.02). The frequencies of the other constitutional symptoms and complete clusters are similar in both groups. Almost all patients in both groups had a good response to steroids.

Table II.
Comparison of clinical characteristics between CARD8-FS+ and CARD8-FS+ PFAPA patients
CharacteristicsCARD8-FS+ (n = 11)CARD8-FS (n = 68)p Value
Median age at onset (y) 1.92 1.4 0.89 
Median duration (d) 3.5 0.49 
Median interval (d) 30 30 0.73 
Positive familial history (n [%]) 7 (64) 29 (43) 0.19 
Pharyngitis (n [%]) 11 (100) 65 (96) 0.48 
Oral aphthosis (n [%]) 9 (82) 30 (44) 0.02* 
Adenitis (n [%]) 8 (73) 39 (57) 0.34 
Symptoms out of flares (n [%]) 4 (36) 7 (10) 0.02* 
Complete cluster (n [%]) 6 (55) 21 (31) 0.12 
Onset after 5 y of age (n [%]) 2 (18) 11 (16) 0.87 
Good response to steroids (n [%]) 7/7 (100) 56/57 (98) 0.72 
CharacteristicsCARD8-FS+ (n = 11)CARD8-FS (n = 68)p Value
Median age at onset (y) 1.92 1.4 0.89 
Median duration (d) 3.5 0.49 
Median interval (d) 30 30 0.73 
Positive familial history (n [%]) 7 (64) 29 (43) 0.19 
Pharyngitis (n [%]) 11 (100) 65 (96) 0.48 
Oral aphthosis (n [%]) 9 (82) 30 (44) 0.02* 
Adenitis (n [%]) 8 (73) 39 (57) 0.34 
Symptoms out of flares (n [%]) 4 (36) 7 (10) 0.02* 
Complete cluster (n [%]) 6 (55) 21 (31) 0.12 
Onset after 5 y of age (n [%]) 2 (18) 11 (16) 0.87 
Good response to steroids (n [%]) 7/7 (100) 56/57 (98) 0.72 
*

p < 0.05.

The most extensively studied polymorphism in the CARD8 gene is the C10X nonsense mutation (rs2043211; A > T) located in exon 5 of the gene, which introduces a stop codon (Cys > Stop) that results in a truncated CARD8 protein. The C10X polymorphism is present in the V region of CARD8. Alternative splicing can bypass the C10X stop codon by changing the reading frame of exon 5 introducing a missense mutation (Phe > Ile) that does not affect expression of a full-length protein (23). The C10X polymorphism is a very common variant, with an MAF = 31.7% reported by the 1000 Genome Project, and 9% of a control population was found to be homozygous for the Stop/Phe > Ile allele (23). However, the functional consequences of the C10X polymorphism remain unclear (2325).

Because the C10X variant has a MAF well above 5%, it was not included in our initial discovery analysis, which specifically targeted uncommon variants. Nevertheless, we had good coverage on the variant position in our NGS data (average 99×); the genotype calls are listed in Supplemental Table II. In our cohort of 82 unrelated PFAPA subjects, 36 are homozygous for the reference allele, 34 are heterozygous, and 12 are homozygous for the Stop/Phe > Ile allele. The MAF derived from our cohort is 35.4%, which is close to what was reported by the 1000 Genome Project. There is no evidence to suggest that CARD8 C10X is a risk allele for PFAPA.

To study the function of CARD8-FS, we generated a construct expressing a tagged version of the protein. This variant was very unstable compared with CARD8, but its expression could be detected by Western blot (Fig. 2). The CARD8-FS gene variant does not code for the CARD8 functional domains CARD and FIIND, which were shown to regulate the inflammasome (12, 13, 22). Therefore, this variant is predicted to have lost its ability to interact with inflammasome components. Instead, CARD8-FS encodes a unique 25-aa polypeptide sequence of unknown function that is absent from all other known CARD8 isoforms. Therefore, we tested the functionality of CARD8-FS to regulate the inflammasome by interrogating its ability to interact with the inflammasome (22). We expressed Flag-tagged NLRP3, Flag-tagged CARD8, VSV-tagged CARD8, and VSV-tagged CARD8-FS in HEK293T cells. Lysates were immunoprecipitated with an anti-Flag Ab and analyzed for coimmunoprecipitating proteins. As previously reported, we showed that, in the presence of Flag-NLRP3, VSV-CARD8 was coimmunoprecipitated (Fig. 2, lane 7). Similarly VSV-CARD8 was coimmunoprecipitated with Flag-CARD8 (Fig. 2, lane 9), indicating the ability of full-length CARD8 to form dimers. In contrast, we found that VSV–CARD8-FS, expressing a truncated protein with a m.w. ∼ 10 kDa, lost the ability to associate with NLRP3 (Fig. 2, lane 8), indicating that this variant is unable to modulate NLRP3 activity. The CARD8-FS variant also lost the ability to form dimers with full-length CARD8 via CARD–CARD interactions (Fig. 2, lane 10), which is in accordance with the lack of a CARD domain in the variant.

FIGURE 2.

Evaluation of the interaction among NLRP3, full-length CARD8, and CARD8-FS. HEK293T cells were cotransfected with expression plasmids for Flag-NLRP3, Flag-CARD8, VSV-CARD8, and the VSV–CARD8-FS variant. Whole-cell lysates were immunoprecipitated with anti-Flag Ab and analyzed by Western blot 24 h after transfection.

FIGURE 2.

Evaluation of the interaction among NLRP3, full-length CARD8, and CARD8-FS. HEK293T cells were cotransfected with expression plasmids for Flag-NLRP3, Flag-CARD8, VSV-CARD8, and the VSV–CARD8-FS variant. Whole-cell lysates were immunoprecipitated with anti-Flag Ab and analyzed by Western blot 24 h after transfection.

Close modal

Furthermore, we examined the effect of CARD8 and CARD8-FS on inflammasome activation by interrogating caspase-1 cleavage after cotransfection of HEK293T cells with NLRP3 inflammasome components (100 ng of NLRP3, 50 ng of ASC, 20 ng of CASP1) in the presence of empty vector, CARD8, or CARD8-FS. We found that CARD8 inhibits inflammasome activation, but CARD8-FS has no effect on caspase-1 cleavage compared with full-length CARD8 (Fig. 3).

FIGURE 3.

CARD8-FS has no effect on caspase-1 cleavage compared with full-length CARD8. HEK293T cells cotransfected with NLRP3 inflammasome components (100 ng of NLRP3, 50 ng of ASC, 20 ng of CASP1) in the presence of empty vector, CARD8, or CARD8-FS. Whole-cell lysates were analyzed for the cleavage of caspase-1 48 h after transfection. Representative immunoblot of caspase-1 cleavage (A) and quantifications of three independent immunoblot experiments (B). §§p < 0.01 versus lane 3 (without NLRP3), *p < 0.05 versus lane 5 (NLRP3 + CARD8).

FIGURE 3.

CARD8-FS has no effect on caspase-1 cleavage compared with full-length CARD8. HEK293T cells cotransfected with NLRP3 inflammasome components (100 ng of NLRP3, 50 ng of ASC, 20 ng of CASP1) in the presence of empty vector, CARD8, or CARD8-FS. Whole-cell lysates were analyzed for the cleavage of caspase-1 48 h after transfection. Representative immunoblot of caspase-1 cleavage (A) and quantifications of three independent immunoblot experiments (B). §§p < 0.01 versus lane 3 (without NLRP3), *p < 0.05 versus lane 5 (NLRP3 + CARD8).

Close modal

These results demonstrate that the CARD8-FS polymorphism generates a loss-of-function allele.

In recent years, several studies showed that mutations in inflammasome genes are key players in the onset of classical autoinflammatory syndromes with penetrant inheritance patterns. Additionally, some studies demonstrated that more complex autoinflammatory diseases are affected by relatively frequent and less-penetrant polymorphisms, such as Q705K and C10X in NLRP3 and CARD8 genes, respectively (25). These polymorphisms are believed to enhance inflammasome activity, leading to increased IL-1β production. In this study, we identified a new CARD8 polymorphism (CARD8-FS) that is significantly associated with classical PFAPA.

CARD8 is known to interact with NLRPs proteins that assemble into inflammasome complexes; however, its physiological function has not been identified with certitude. Like most negative regulators of the inflammasome, CARD8 is not present in the mouse genome (26), precluding any relevant functional investigations in animal models. Studies in humans and the identification of disease-associated variants were instrumental in the characterization of CARD8 (12, 25). In particular, identification of the C10X mutant was linked to various inflammatory diseases, including Crohn’s disease, rheumatoid arthritis, cardiovascular diseases, and ankylosing spondylitis (24, 27, 28). However, some findings have not been replicated robustly, suggesting that the effect of the relatively frequent C10X polymorphism might be modest or requires association with additional factors to significantly drive disease outcome (25). Another explanation is that the C10X mutation is located in the region of CARD8 that is highly variable and can be bypassed by alternative splicing that restores FIIND and CARD functionality (23). This could present an adaptation mechanism to compensate for CARD8 deficiency and could explain the fact that, in this study, PFAPA was not found to be significantly associated with the C10X variant.

In contrast, the CARD8-FS variant described in this article is located in exon 7, which is conserved among most reported CARD8 isoforms (23). Therefore, the premature stop codon resulting from this variant is unlikely to be rescued by alternative splicing that restores the FIIND and CARD domains. The premature termination codon introduced by CARD8-FS is likely to direct the mutant CARD8 mRNAs for degradation via nonsense-mediated decay, leading to decreased CARD8 expression in these patients. Moreover, we showed that the protein translated from a CARD8-FS mRNA was unstable compared with CARD8. However, in an experimental set-up with forced expression of the variant, we found that CARD8-FS lost its ability to interact with NLRP3, indicating that, beyond the fact that CARD8-FS is probably expressed at very low levels, remaining protein products are unable to regulate inflammasome assembly. We also tested the ability of this variant to dimerize with CARD8. No residual interaction between CARD8 and CARD8-FS was detected, suggesting that CARD8-FS does not interfere with the normal function of the wild-type CARD8 allele in heterozygous PFAPA patients. Moreover, we show that CARD8-FS has no effect on caspase-1 cleavage compared with full-length CARD8, which could explain the excess inflammasome activation in CARD8-FS patients. Altogether, these data suggest that PFAPA patients carrying a CARD8-FS allele have an overall decreased CARD8 activity but are likely to retain one functional CARD8 allele. The absence of homozygous CARD8-FS PFAPA patients in our cohort might suggest the possibility of a much more severe inflammatory phenotype than PFAPA phenotype in homozygous individuals.

PFAPA syndrome is not a sporadic disease, as shown by increasingly reported familial susceptibility that emphasizes a probable genetic cause for the disease. A recent genetic study of patients from 14 families indicated that PFAPA was unlikely to be a monogenic condition and suggested an oligogenic or complex inheritance of variants (14). The role of low-penetrant mutations in genes that are involved in other monogenic autoinflammatory diseases was suggested by the significantly higher frequencies of three NLRP3 variants (R488K, V198M, Q705K) in PFAPA cohorts (6). The pathogenic role of the Q705K variant is a matter of debate because it is present in ∼5% of the healthy population. Interestingly, carriers of the Q705K variant with a typical CAPS phenotype were described (29), suggesting that additional genetic and/or environmental modifiers are involved.

Although IL-1 dysregulation in PFAPA is clearly supported by in vitro data on IL-1 secretion by PBMCs from PFAPA patients and the positive clinical responses observed in patients treated with anti–IL-1 biologics (7), a question that remains to be addressed is whether inflammasome abnormality is the main trigger of inflammatory flares in these patients. A decreased ability of CARD8 to bind NLRP3 is consistent with increased inflammasome activity and IL-1β production. This is further supported by the fact that mutant NLRP3 proteins that cause inherited CAPS have lost their ability to bind CARD8 (12). Moreover, a previous study showed that reduced CARD8 expression by small interfering RNA increases the production of active IL-1β in the presence of inflammasome agonists (12). Consistent with its role as a negative regulator of the inflammasome, it is likely that CARD8-FS–mediated CARD8 deficiency in PFAPA patients is not sufficient to trigger inflammasome activation per se, but it may exacerbate inflammasome responses in the presence of activating stimuli. In line with this hypothesis, we reported previously that PBMCs and monocytes from PFAPA patients have no constitutive inflammasome activity; however, they show increased IL-1β secretion, compared with controls, in the presence of crude LPS (an LPS preparation that also contains muramyl dipeptide, a potent inflammasome activator) (6).

The role of CARD8-FS in the exacerbation of the inflammatory phenotype was probed by the clinical analysis of CARD8-FS patients compared with other PFAPA patients who do not carry CARD8-FS. Patients carrying the CARD8-FS variant seem to display a more severe PFAPA phenotype (higher frequency of oral aphthosis and symptoms out of flare), supporting the hypothesis that the inflammasome dysregulation is an underlying contributor of the disease and is exacerbated by the CARD8-FS variant. Therefore, defects in negative regulation of the inflammasome may be sufficient to have an impact on autoinflammatory diseases; however, the phenotype difference between CARD8-FS+ and CARD8-FS subjects is mild. PFAPA phenotypes of patients who do not carry the CARD8-FS variant are likely to be attributed to other mechanisms that are not fully understood. Although PBMCs and monocytes of PFAPA patients showed elevated IL-1β secretion in general, we do not anticipate significantly higher IL-1β production in CARD8-FS patients compared with PFAPA patients who do not carry this variant.

Our study is limited by the lack of immediate access to a suitable independent PFAPA cohort in which this low-penetrance CARD8-FS variant can be pursued. Further studies aimed at identifying the details of inflammasome engagement in PFAPA and a better understanding of the negative-regulatory mechanisms that control the intensity and duration of inflammasome activation in humans will shed new light on this complex autoinflammatory disease.

We thank Frank Staedtler and Hermann Gram for strategic input; Stine Buechmann-Moller, Edward Oakeley, Moritz Frei, and Virginie Petitjean for NGS data generation; Martin Letzkus and Sandrine Starck-Schwertz for sample management; and Tiziana Valensise and Thomas Peters for Sanger sequencing support.

This work was supported by Swiss National Science Foundation Grant FN 310030-130085/1.

The online version of this article contains supplemental material.

Abbreviations used in this article:

CAPS

cryopyrin-associated periodic syndrome

CARD8-FS

frameshift variant in the CARD8 gene

FIIND

function to find

MAF

minor allele frequency

NGS

next-generation sequencing

PFAPA

periodic fever with aphthous stomatitis, pharyngitis, and cervical adenitis

VSV

vesicular stomatitis virus.

1
Cantarini
L.
,
Vitale
A.
,
Bersani
G.
,
Nieves
L. M.
,
Cattalini
M.
,
Lopalco
G.
,
Caso
F.
,
Costa
L.
,
Iannone
F.
,
Lapadula
G.
, et al
.
2015
.
PFAPA syndrome and Behcet’s disease: a comparison of two medical entities based on the clinical interviews performed by three different specialists.
Clin. Rheumatol.
35
:
501
505
.
2
Caorsi
R.
,
Pelagatti
M. A.
,
Federici
S.
,
Finetti
M.
,
Martini
A.
,
Gattorno
M.
.
2010
.
Periodic fever, apthous stomatitis, pharyngitis and adenitis syndrome.
Curr. Opin. Rheumatol.
22
:
579
584
.
3
Hofer
M.
,
Pillet
P.
,
Cochard
M. M.
,
Berg
S.
,
Krol
P.
,
Kone-Paut
I.
,
Rigante
D.
,
Hentgen
V.
,
Anton
J.
,
Brik
R.
, et al
.
2014
.
International periodic fever, aphthous stomatitis, pharyngitis, cervical adenitis syndrome cohort: description of distinct phenotypes in 301 patients.
Rheumatology (Oxford)
53
:
1125
1129
.
4
Ter Haar
N.
,
Lachmann
H.
,
Özen
S.
,
Woo
P.
,
Uziel
Y.
,
Modesto
C.
,
Koné-Paut
I.
,
Cantarini
L.
,
Insalaco
A.
,
Neven
B.
, et al
Paediatric Rheumatology International Trials Organisation (PRINTO) and the Eurofever/Eurotraps Projects
.
2013
.
Treatment of autoinflammatory diseases: results from the Eurofever Registry and a literature review.
Ann. Rheum. Dis.
72
:
678
685
.
5
Ombrello
M. J.
2015
.
Advances in the genetically complex autoinflammatory diseases.
Semin. Immunopathol.
37
:
403
406
.
6
Kolly
L.
,
Busso
N.
,
von Scheven-Gete
A.
,
Bagnoud
N.
,
Moix
I.
,
Holzinger
D.
,
Simon
G.
,
Ives
A.
,
Guarda
G.
,
So
A.
, et al
.
2013
.
Periodic fever, aphthous stomatitis, pharyngitis, cervical adenitis syndrome is linked to dysregulated monocyte IL-1β production.
J. Allergy Clin. Immunol.
131
:
1635
1643
.
7
Stojanov
S.
,
Lapidus
S.
,
Chitkara
P.
,
Feder
H.
,
Salazar
J. C.
,
Fleisher
T. A.
,
Brown
M. R.
,
Edwards
K. M.
,
Ward
M. M.
,
Colbert
R. A.
, et al
.
2011
.
Periodic fever, aphthous stomatitis, pharyngitis, and adenitis (PFAPA) is a disorder of innate immunity and Th1 activation responsive to IL-1 blockade.
Proc. Natl. Acad. Sci. USA
108
:
7148
7153
.
8
Strowig
T.
,
Henao-Mejia
J.
,
Elinav
E.
,
Flavell
R.
.
2012
.
Inflammasomes in health and disease.
Nature
481
:
278
286
.
9
Kuemmerle-Deschner
J. B.
2015
.
CAPS--pathogenesis, presentation and treatment of an autoinflammatory disease.
Semin. Immunopathol.
37
:
377
385
.
10
Verma
D.
,
Särndahl
E.
,
Andersson
H.
,
Eriksson
P.
,
Fredrikson
M.
,
Jönsson
J. I.
,
Lerm
M.
,
Söderkvist
P.
.
2012
.
The Q705K polymorphism in NLRP3 is a gain-of-function alteration leading to excessive interleukin-1β and IL-18 production.
PLoS One
7
:
e34977
.
11
Rieber
N.
,
Gavrilov
A.
,
Hofer
L.
,
Singh
A.
,
Öz
H.
,
Endres
T.
,
Schäfer
I.
,
Handgretinger
R.
,
Hartl
D.
,
Kuemmerle-Deschner
J.
.
2015
.
A functional inflammasome activation assay differentiates patients with pathogenic NLRP3 mutations and symptomatic patients with low penetrance variants.
Clin. Immunol.
157
:
56
64
.
12
Ito
S.
,
Hara
Y.
,
Kubota
T.
.
2014
.
CARD8 is a negative regulator for NLRP3 inflammasome, but mutant NLRP3 in cryopyrin-associated periodic syndromes escapes the restriction.
Arthritis Res. Ther.
16
:
R52
.
13
Razmara
M.
,
Srinivasula
S. M.
,
Wang
L.
,
Poyet
J. L.
,
Geddes
B. J.
,
DiStefano
P. S.
,
Bertin
J.
,
Alnemri
E. S.
.
2002
.
CARD-8 protein, a new CARD family member that regulates caspase-1 activation and apoptosis.
J. Biol. Chem.
277
:
13952
13958
.
14
Di Gioia
S. A.
,
Bedoni
N.
,
von Scheven-Gête
A.
,
Vanoni
F.
,
Superti-Furga
A.
,
Hofer
M.
,
Rivolta
C.
.
2015
.
Analysis of the genetic basis of periodic fever with aphthous stomatitis, pharyngitis, and cervical adenitis (PFAPA) syndrome.
Sci. Rep.
5
:
10200
.
15
Thomas
K. T.
,
Feder
H. M.
 Jr.
,
Lawton
A. R.
,
Edwards
K. M.
.
1999
.
Periodic fever syndrome in children.
J. Pediatr.
135
:
15
21
.
16
André
S. C.
,
Vales
F.
,
Cardoso
E.
,
Santos
M.
.
2009
.
PFAPA syndrome
.
Acta Otorrinolaringol. Esp.
60
:
208
209
.
17
McKenna
A.
,
Hanna
M.
,
Banks
E.
,
Sivachenko
A.
,
Cibulskis
K.
,
Kernytsky
A.
,
Garimella
K.
,
Altshuler
D.
,
Gabriel
S.
,
Daly
M.
,
DePristo
M. A.
.
2010
.
The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data.
Genome Res.
20
:
1297
1303
.
18
DePristo
M. A.
,
Banks
E.
,
Poplin
R.
,
Garimella
K. V.
,
Maguire
J. R.
,
Hartl
C.
,
Philippakis
A. A.
,
del Angel
G.
,
Rivas
M. A.
,
Hanna
M.
, et al
.
2011
.
A framework for variation discovery and genotyping using next-generation DNA sequencing data.
Nat. Genet.
43
:
491
498
.
19
Van der Auwera
G.A.
,
Carneiro
M.O.
,
Hartl
C.
,
Poplin
R.
,
Del Angel
G.
,
Levy-Moonshine
A.
,
Jordan
T.
,
Shakir
K.
,
Roazen
D.
,
Thibault
J.
, et al
.
2013
.
From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline.
Curr. Protoc. Bioinformatics
43
:
11.10.1
11.10.33
.
20
McLaren
W.
,
Pritchard
B.
,
Rios
D.
,
Chen
Y.
,
Flicek
P.
,
Cunningham
F.
.
2010
.
Deriving the consequences of genomic variants with the Ensembl API and SNP effect predictor.
Bioinformatics
26
:
2069
2070
.
21
Sherry
S. T.
,
Ward
M. H.
,
Kholodov
M.
,
Baker
J.
,
Phan
L.
,
Smigielski
E. M.
,
Sirotkin
K.
.
2001
.
dbSNP: the NCBI database of genetic variation.
Nucleic Acids Res.
29
:
308
311
.
22
Agostini
L.
,
Martinon
F.
,
Burns
K.
,
McDermott
M. F.
,
Hawkins
P. N.
,
Tschopp
J.
.
2004
.
NALP3 forms an IL-1beta-processing inflammasome with increased activity in Muckle-Wells autoinflammatory disorder.
Immunity
20
:
319
325
.
23
Bagnall
R. D.
,
Roberts
R. G.
,
Mirza
M. M.
,
Torigoe
T.
,
Prescott
N. J.
,
Mathew
C. G.
.
2008
.
Novel isoforms of the CARD8 (TUCAN) gene evade a nonsense mutation.
Eur. J. Hum. Genet.
16
:
619
625
.
24
Kastbom
A.
,
Johansson
M.
,
Verma
D.
,
Söderkvist
P.
,
Rantapää-Dahlqvist
S.
.
2010
.
CARD8 p.C10X polymorphism is associated with inflammatory activity in early rheumatoid arthritis.
Ann. Rheum. Dis.
69
:
723
726
.
25
Paramel
G.V.
,
Sirsjo
A.
,
Fransen
K.
.
2015
.
Role of genetic alterations in the NLRP3 and CARD8 genes in health and disease.
Mediators Inflamm.
2015
:
846782
.
26
Martinon
F.
,
Mayor
A.
,
Tschopp
J.
.
2009
.
The inflammasomes: guardians of the body.
Annu. Rev. Immunol.
27
:
229
265
.
27
Schoultz
I.
,
Verma
D.
,
Halfvarsson
J.
,
Törkvist
L.
,
Fredrikson
M.
,
Sjöqvist
U.
,
Lördal
M.
,
Tysk
C.
,
Lerm
M.
,
Söderkvist
P.
,
Söderholm
J. D.
.
2009
.
Combined polymorphisms in genes encoding the inflammasome components NALP3 and CARD8 confer susceptibility to Crohn’s disease in Swedish men.
Am. J. Gastroenterol.
104
:
1180
1188
.
28
Kastbom
A.
,
Klingberg
E.
,
Verma
D.
,
Carlsten
H.
,
Forsblad-d’Elia
H.
,
Wesamaa
J.
,
Cedergren
J.
,
Eriksson
P.
,
Söderkvist
P.
.
2013
.
Genetic variants in CARD8 but not in NLRP3 are associated with ankylosing spondylitis.
Scand. J. Rheumatol.
42
:
465
468
.
29
Vitale
A.
,
Lucherini
O. M.
,
Galeazzi
M.
,
Frediani
B.
,
Cantarini
L.
.
2012
.
Long-term clinical course of patients carrying the Q703K mutation in the NLRP3 gene: a case series.
Clin. Exp. Rheumatol.
30
:
943
946
.

M.H. received consultancy fees from Novartis and Abbvie. The other authors have no financial conflicts of interest.

Supplementary data