A somatic NLRP3 mutation as a cause of a sporadic case of chronic infantile neurologic, cutaneous, articular syndrome/neonatal-onset multisystem inflammatory disease: Novel evidence of the role of low-level mosaicism as the pathophysiologic mechanism underlying mendelian inherited diseases

Authors


Abstract

Objective

Chronic infantile neurologic, cutaneous, articular syndrome (CINCA), also known as neonatal-onset multisystem inflammatory disease (NOMID), is a severe, early-onset autoinflammatory disease characterized by an urticaria-like rash, arthritis/arthropathy, variable neurologic involvement, and dysmorphic features, which usually respond to interleukin-1 blockade. CINCA/NOMID has been associated with dominant Mendelian inherited NLRP3 mutations. However, conventional sequencing analyses detect true disease-causing mutations in only ∼55–60% of patients, which suggests the presence of genetic heterogeneity. We undertook the current study to assess the presence of somatic, nongermline NLRP3 mutations in a sporadic case of CINCA/NOMID.

Methods

Clinical data, laboratory results, and information on treatment outcomes were gathered through direct interviews. Exhaustive genetic studies, including Sanger method sequencing, subcloning, restriction fragment length polymorphism assay, and pyrosequencing, were performed.

Results

The patient's CINCA/NOMID was diagnosed based on clinical features (early onset of the disease, urticaria-like rash, knee arthropathy, and dysmorphic features). The patient has exhibited a successful response to anakinra within the last 28 months. Analysis of NLRP3 identified a novel heterozygous variant (p.D303H) that was detected in ∼30–38% of circulating leukocytes. The absence of this variant in healthy controls and in the patient's parents suggested a de novo true disease-causing mutation. Additional analyses showed that this novel mutation was present in both leukocyte subpopulations and epithelial cells.

Conclusion

Our findings identify the novel p.D303H NLRP3 variant in a Spanish patient with CINCA/NOMID as a new disease-causing mutation, which was detected as a somatic, nongermline mutation in hematopoietic and nonhematopoietic cell lineages. Our data provide new insight into the role of low-level mosaicism in NLRP3 as the pathophysiologic mechanism underlying cryopyrin-associated periodic syndrome.

Cryopyrin-associated periodic syndrome (CAPS) includes 3 different autoinflammatory entities, as follows: familial cold-induced autoinflammatory syndrome (FCAS; MIM no. #120100), Muckle-Wells syndrome (MWS; MIM no. #191900), and chronic infantile neurologic, cutaneous, articular syndrome (CINCA; MIM no. #607115), also known as neonatal-onset multisystem inflammatory disease (NOMID) (1). From a clinical point of view, the onset of these syndromes occurs neonatally or during early childhood, usually as a generalized urticaria-like skin rash associated with an intense acute-phase reaction, and different inflammatory manifestations can be observed for each clinical entity (2). As recently reported, a common genetic basis for these syndromes has been observed by means of the identification of dominantly inherited mutations in the NLRP3 gene (formerly known as CIAS1, PYPAF1, or NALP3), which encodes for the cryopyrin protein, a key component of the inflammasome (3–5). These data support the hypothesis that the 3 CAPS diseases actually represent different points along a disease severity spectrum, where FCAS represents the mildest phenotype, MWS the intermediate one, and CINCA/NOMID the severest (1). Supporting this hypothesis, some clinical cases of overlapping FCAS-MWS and MWS-CINCA/NOMID syndromes have been reported (6, 7).

Most patients with CINCA/NOMID have no familial history of the disease, and have a chronic, nonepisodic inflammatory disease characterized by a generalized urticaria-like skin rash, arthritis and/or deforming arthropathy, varying degrees of neurologic involvement, and dysmorphic features (5, 8, 9). Conventional analyses of NLRP3 have detected germline, disease-causing mutations in only ∼55–60% of CINCA/NOMID patients, which suggests the presence of genetic heterogeneity (4, 5, 10, 11). These genetic studies have also shown that most CINCA/NOMID cases result from de novo NLRP3 mutations, thus explaining the absence of familial history of the disease (1, 4, 5, 10, 11). It has been recently reported that, in some patients with CAPS, the disease resulted from somatic, nongermline NLRP3 mutations, which raises the possibility that these genetic events could explain, at least partially, the previously recognized genetic heterogeneity of CAPS (12, 13).

Herein, we report the case of a Spanish patient with CINCA/NOMID, who has been successfully treated with anakinra during the past 28 months and in whom exhaustive mutational analyses of NLRP3 revealed a novel de novo somatic missense mutation (p.Asp303His; p.D303H) on exon 3. Exhaustive genetic studies revealed that this somatic mutation was present in ∼30–38% of circulating leukocytes and was detected both in hematopoietic and nonhematopoietic lineages. All these data suggest that an early mutational event in NLRP3 probably occurred during embryonic development, thus determining the low-level mosaicism at NLRP3 responsible for the CINCA/NOMID phenotype in this patient.

PATIENTS AND METHODS

Subjects.

Clinical symptoms, hematologic and biochemical data, and outcomes of previous treatments were investigated using direct interview and were recorded using a specific questionnaire. Peripheral blood, mucosal swabs, and urine samples from the patient and his parents were obtained. For the genetic analyses, 200 anonymous unrelated Spanish blood donors were used as healthy controls. Written informed consent from the patient's parents and approval by the medical ethics committee of the Hospital Clínic were obtained, in accordance with the Helsinki Declaration.

NLRP3 mutational analysis.

Genomic DNA from whole blood, isolated leukocyte subpopulations, and epithelial cells was isolated using a QIAamp DNA Blood Mini kit (Qiagen, Hilden, Germany), and plasmidic DNA was purified using a QIAprep Spin Miniprep kit (Qiagen). All 9 exons and intronic flanking sequences of NLRP3 were amplified by polymerase chain reaction (PCR) and purified using a QIAquick PCR Purification kit (Qiagen), and bidirectional fluorescence sequencing was performed with an ABI BigDye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, Foster City, CA), using primers and PCR conditions described previously (11). A sequence-specific primers (SSP)–PCR assay was designed to discriminate mutated from wild-type alleles. The sequence of primers used was 5′-TTCCTCATGGACGGCTTCG-3′ (sense; G-allele specific), 5′-TTCCTCATGGACGGCTTCC-3′ (sense; C-allele specific), and 5′-CGGCCTTCTGCCAGTCAGT-3′ (antisense; common primer). PCR amplicons underwent electrophoresis in an agarose gel and were visualized on an ultraviolet transilluminator. A Topo TA cloning kit (Invitrogen, San Diego, CA) was used to subclone the PCR amplicons.

Restriction fragment length polymorphism (RFLP) assay.

Restriction maps of PCR amplicons from wild-type and mutated alleles were obtained. Purified PCR amplicons were digested with 5 IU of Taq I restriction endonuclease (New England Biolabs, Beverly, MA) for 4 hours at 65°C.

Purification of leukocyte subpopulations.

Freshly drawn whole blood was separated by sequential dextrane and Ficoll-Hypaque density centrifugation methods, and a magnetic-activated cell sorter (Miltenyi Biotec, Sunnyvale, CA) was used to positively select T cells, B cells, monocytes, and neutrophils using beads conjugated with anti-CD3, anti-CD19, anti-CD14, and anti-CD15 monoclonal antibodies. Purity of leukocyte isolation was determined using a FACSCalibur flow cytometer (BD Biosciences Pharmingen, San Diego, CA) with appropriate monoclonal antibodies.

Pyrosequencing of NLRP3.

For pyrosequencing, a PCR assay covering position c.907 of the NLRP3 gene was designed. The sequence of primers used was 5′-CCACCCATCCACAAGATCGT-3′ (sense biotinylated) and 5′-GCGGTCCTATGTGCTCGTC-3′ (antisense). PCR amplicons were sequenced with the Pyrosequencing PSQ96 HS system according to the instructions of the manufacturer (Biotage AB, Uppsala, Sweden), using the sequencing primer 5′-TGCTCGTCAAAGGCA-3′. PyroMark MD version 1.0 software (Biotage AB) was used for data analysis.

RESULTS

The patient we describe, a 7-year-old Spanish boy with healthy, nonconsanguineous parents, was born at 39 weeks gestation after complications related to a possible spontaneous miscarriage. At birth, he weighed 3,460 gm and had an Apgar score of 9/10. At 4 months of age, the patient exhibited a persistent inflammatory disease characterized by a generalized urticaria-like skin rash, recurrent fever, mild headache, nausea, vomiting, bilateral conjunctivitis, and lymphadenopathy (Figure 1A). No recognizable triggering or worsening factors were identified. When the patient was 11 months of age, a symmetric oligoarthritis affecting the elbow and knee joints was observed, which evolved into a chronic symmetric deforming arthropathy of the knee joints, with radiographic evidence of premature ossification and osseous overgrowth at the patella and long bone epiphysis (Figure 1B). The patient had in flexo joint contractures and delayed growth (below the third percentile), but no discrepancy in leg length was observed.

Figure 1.

A, Urticaria-like skin rash. B, Radiographs showing bilateral knee arthropathy. C, Sense and antisense electropherograms obtained by conventional genetic analyses using genomic DNA from circulating leukocytes. Arrows show the NLRP3 genotype at position c.907 in a healthy individual, the proband, and a patient with cryopyrin-associated periodic syndrome (CAPS) with the germline p.D303N mutation. Note the discrepancies in fluorescence intensity, more evident in the antisense sequences, in the heterozygous c.907 position between the proband (subtle changes) and the CAPS patient (similar fluorescence intensity for each peak). D, Genomic organization of the NLRP3 gene (top), location of common single-nucleotide polymorphisms at exon 3 (middle), and results of polymerase chain reaction (PCR)–based assay for conventional mutational analysis of exon 3 (bottom). Note that nucleotide position c.907, encoding for the p.303 codon, is located on the PCR amplicon A. E, Sequence electropherograms showing the antisense sequence of nucleotide position c.726 (arrow) in G/G-and A/A-homozygous individuals and in the proband. Note the heterozygous status at this position with identical fluorescence intensities for each peak, which rules out the possibility that a selective amplification of 1 allele occurred.

Several signs of facial dysmorphy, such as frontal bossing, delayed closure of the anterior fontanelle, increased cranial perimeter, and saddlenose deformity, were also observed. Repeated neurologic, audiometric, and ophthalmologic examinations revealed no endocranial hypertension, mental retardation, seizures, sensorineural hearing loss, papilledema, optic disc edema, or optic atrophy. An exhaustive neuropsychologic examination administered when the patient was 5 years old revealed normal cognitive functions, with an above-average IQ of 126, according to the Wechsler Preschool and Primary Scale of Intelligence.

Laboratory features included polymorphonuclear leukocytosis (18,000–20,000/mm3), thrombocytosis (706,000–750,000/mm3), and an acute-phase reaction characterized by increased erythrocyte sedimentation rate (45–60 mm/hour) and plasma levels of C-reactive protein. Findings of microbiologic and immunologic examinations were negative or normal, supporting the exclusion of an infectious or autoimmune basis for the disease. Based on clinical and laboratory features, a diagnosis of CINCA/NOMID was proposed.

Multiple therapeutic approaches, including nonsteroidal antiinflammatory drugs, methotrexate, and etanercept, were investigated in the patient, resulting in partial responses. A positive response to systemic corticosteroids was observed, but long-term use was restricted by their side effects. When the patient was 4 years of age, treatment with 25 mg/day (1.7 mg/kg/day) of anakinra (the nonglycosylated recombinant form of human interleukin-1 receptor antagonist [IL-1Ra]) was started and resulted in long-term clinical improvement and normalization of laboratory parameters.

NLRP3 mutational analysis.

Peripheral blood samples from the patient and his parents were obtained for genetic analysis. Conventional mutational analyses of all exons of the NLRP3 gene using genomic DNA from circulating leukocytes revealed the apparent absence of disease-causing mutations. However, an exhaustive search revealed a possible heterozygous G-to-C transversion at position c.907 (GenBank accession no. NM_004895.3), with an important difference in the fluorescence intensity for each nucleotide. Thus, the intensity of the wild-type G-harboring allele was nearly normal when compared with that of a healthy control, while the intensity of the variant C-harboring allele was much lower (Figure 1C). This apparent nucleotide exchange was located at the first nucleotide of codon 303 of cryopyrin, which has been previously identified as a hotspot for the severest forms of CAPS (4, 5, 10, 11). When the nucleotide exchange is present, it should provoke a histidine-to-aspartic amino acid substitution (p.D303H), which has not been described previously; consequently, it should be considered a novel missense NLRP3 genetic variant. The presence of this nucleotide transversion was verified by 3 different PCR-based methods with different detection thresholds: an RFLP assay, a subcloning method, and a mutation-specific SSP-PCR method (Figures 2A–C).

Figure 2.

A, Restriction maps of the wild-type G-harboring allele and the mutated C-harboring allele, showing the destruction of a Taq I restriction site by the nucleotide exchange (top), and Taq I restriction fragment length polymorphism analysis showing the presence of a minority undigested band (arrowhead) in the patient (P), which should correspond to the mutation-harboring allele (bottom). L = size marker; PCR = undigested polymerase chain reaction; C = Taq I–digested PCR from a control. B, Sense sequence electropherograms of subcloned wild-type and mutation-harboring alleles, showing the nucleotide G-to-C transversion at position c.907 (arrows), and consequently, the amino acid substitution p.D303H. C, Sequence-specific primer–PCR (SSP-PCR) for the mutated (top) and wild-type (bottom) alleles. An internal control for the PCR amplification is denoted by open arrowheads, and the specific SSP-PCR for the NLRP3 allele is denoted by solid arrowheads. F = father; M = mother; N = isolated neutrophils; Mo = isolated monocytes; T = isolated T cells; B = isolated B cells; S = mucosal swabs; U = urine. D, Antisense sequence electropherograms obtained by amplification of genomic DNA from isolated leukocyte subpopulations and from buccal and urinary epithelial cells. Purity of leukocyte subpopulations was determined by flow cytometry (histograms).

To exclude the possibility of a selective amplification of the wild-type allele versus the variant-harboring allele, we looked for the presence of common intragenic single-nucleotide polymorphisms (SNPs) in the PCR amplicon. The unique informative SNP identified was located at nucleotide position c.726, corresponding to codon p.242 of cryopyrin. At this position, a heterozygous G/A genotype was detected in the patient, and the fluorescence intensity for each nucleotide was identical, suggesting that both paternally and maternally derived alleles had been equally amplified during PCR (Figure 1E).

The G-to-C transversion at position c.907 detected in the CINCA/NOMID patient was found neither in his parents nor among a panel of 200 healthy Spanish controls. These data reasonably excluded the possibility that this genetic variant could be a benign polymorphism and supported the presence of a de novo disease-causing mutation in NLRP3. Paternity was confirmed by means of HLA haplotype segregation (data not shown).

Distribution and quantification of the novel p.D303H NLRP3 mutation.

To determine the cellular distribution of the novel p.D303H NLRP3 mutation, mutational analyses were performed by bidirectional conventional sequencing, PCR-SSP, and pyrosequencing methods, using genomic DNA extracted from epithelial cells and from magnetically isolated CD15+ cells (neutrophils), CD14+ (monocytes), CD3+ (T lymphocytes), and CD19+ (B lymphocytes). The novel p.D303H mutation was found in buccal and urinary epithelial cells, as well as in neutrophils, T cells, and B cells. Surprisingly, the novel mutation was not detected in isolated monocytes by any molecular methods (Figures 2C and D and Table 1). The pyrosequencing method enabled us also to quantify the mutated allele's overall frequency, which was ∼15–19% in all samples analyzed and which corresponded to ∼30–38% mutated cells; this frequency was notably different from that observed in patients with germline NLRP3 mutations (∼50% mutated alleles and 100% mutated cells) and in individuals with no mutation (0% mutated alleles and 0% mutated cells) (Table 1).

Table 1. Quantification of the mutated NLRP3 allele by pyrosequencing*
SampleClinical statusSource of genomic DNANLRP3 genotypeType of mutationPyrosequencer genotype% of mutated allele
  • *

    PB = whole peripheral blood; WT = wild type; CINCA/NOMID = chronic infantile neurologic, cutaneous, articular syndrome/neonatal-onset multisystem inflammatory disease; MWS = Muckle-Wells syndrome; FCAS = familial cold-induced autoinflammatory syndrome.

1–8Healthy controlsPBWT/WTWT/WT0
9CINCA/NOMIDPBp.D303N/WTGermlineMutated/WT46.1
10MWS-CINCA/NOMIDPBp.D303N/WTGermlineMutated/WT47.4
11FCASPBp.L305P/WTGermlineMutated/WT53.0
12FCASPBp.L305P/WTGermlineMutated/WT50.0
13CINCA/NOMID (proband)PBp.D303H/WTSomaticMutated/WT19.1
14Proband's fatherPBWT/WTWT/WT0
15Proband's motherPBWT/WTWT/WT0
16CINCA/NOMIDBuccal cellsp.D303H/WTSomaticMutated/WT17.5
17CINCA/NOMIDUrinary cellsp.D303H/WTSomaticMutated/WT15.5
18CINCA/NOMIDNeutrophilsp.D303H/WTSomaticMutated/WT15.5
19CINCA/NOMIDMonocytesWT/WTWT/WT0
20CINCA/NOMIDT cellsp.D303H/WTSomaticMutated/WT16.9
21CINCA/NOMIDB cellsp.D303H/WTSomaticMutated/WT17.0

DISCUSSION

At the beginning of the 1980s, CINCA/NOMID was first described as a new nosologic entity on the basis of certain clinical features (early onset of the disease, urticaria-like skin rash, neurologic and articular involvement, and dysmorphic features) of some patients previously diagnosed as having systemic-onset juvenile idiopathic arthritis (8, 9). Since then, several different cases of CINCA/NOMID have been reported, most of them in patients with no familial history of the disease. The presence of clinical diversity among the cases became rapidly evident, especially the differences in overall disease severity and neurologic and articular involvement (5, 10). With regard to neurologic symptoms, some patients exclusively have relatively mild to moderate chronic aseptic meningitis, whereas other patients can present with aseptic meningitis as well as with a myriad of other severe neurologic symptoms, including profound mental retardation, seizures, endocranial hypertension, papilledema, sensorineural deafness, cognitive disorders, and low IQ. The diversity among articular manifestations is slightly smaller; some patients have recurrent episodes of nondestructive arthritis, while others develop a severe chronic deforming arthropathy, prominent in the knees and ankles, due to the premature ossification of patella and long bone epiphysis, which leads to joint contractures, muscular atrophy, and notable delays in growth and bipedestation (10).

In the case described herein, the early onset of the disease, the generalized urticaria-like skin rash, and the presence of the typical knee-deforming arthropathy and some dysmorphic features associated with a chronic inflammatory status suggested to us the diagnosis of CINCA/NOMID. Nevertheless, we must highlight the fact that the neurologic symptoms observed in the current patient were not severe (only mild headache, probably secondary to mild chronic inflammatory meningitis). This case adds new evidence to the previously recognized clinical diversity of neurologic manifestations in patients with CINCA/NOMID.

Cryopyrin (or NALP3), the NLRP3-encoded protein, has been observed to be the mutated protein in patients with CAPS. It belongs to the growing family of cytosolic NOD-like receptors (NLRs) of the nonclonal, innate immunity (14) and, with the proteins ASC, Cardinal, and procaspase 1, forms a dynamic multiprotein platform termed the “NALP3 inflammasome” (15). The NALP3 inflammasome could be activated by different exogenous and endogenous ligands (bacterial toxins, small antiviral compounds, uric acid crystals, silica, etc.), thus inducing the production of biologically active IL-1β, IL-18, and IL-33 (1, 15). In 2002, Aksentijevich et al reported a constitutive increase in the production of proIL-1β in patients with CINCA/NOMID; this increase could not be counterbalanced by the production of IL-1Ra, which was increased as well (5). These results were subsequently corroborated by findings of another study using patients with MWS as a model (16). Taken together, these data supported the hypothesis that the CAPS diseases could be IL-1–mediated diseases, due to the fact that a gain-of-function mutation of cryopyrin activates the NALP3 inflammasome in a ligand-independent manner (1, 5, 16). Thus, the possibility of using IL-1–blocking agents to successfully treat CAPS became evident.

This therapeutic approach was first reported by Hawkins et al in 2 patients with MWS (17). Subsequently, similar responses in all CAPS phenotypes were observed (18–21), thus establishing unambiguously that IL-1 blockade is the first therapeutic option for patients with CAPS. On the basis of this evidence, anakinra treatment was started in our CINCA/NOMID patient, and consistent with the findings of previous studies (21), this treatment led to a rapid symptomatic improvement and a relatively rapid normalization of hematologic and biochemical inflammatory parameters, all of which remain in the normal range after 28 months of continuous treatment. However, patellar hypertrophy has remained constant during this time period despite the successful anakinra treatment.

In 2002, 2 different groups reported for the first time de novo germline NLRP3 mutations in sporadic instances of CINCA/NOMID (4, 5). However, the conventional sequencing analyses employed in these studies detected germline mutations in only ∼55–60% of patients, which has been subsequently corroborated by the findings of other studies (4, 5, 10, 11). Together, these data supported the presence of genetic heterogeneity in CINCA/NOMID and opened the possibility that mutation-negative CINCA/NOMID could be a phenocopy, due to mutations in genes other than NLRP3. It has been hypothesized that this unknown gene could be structurally and/or functionally related to NLRP3. Some researchers have explored this possibility by means of sequencing candidate genes encoding various NLR proteins, different components of inflammasome, or proteins related to IL-1β secretion. Nevertheless, these approaches have not revealed clear disease-causing mutations in any of the analyzed genes, and the genes that are potentially responsible remain undetermined (2, 22).

A different genetic approach was addressed by Saito et al, who reported for the first time the presence of low-level mosaicism at NLRP3 as the pathophysiologic mechanism responsible for the disease in a sporadic case of CINCA/NOMID, for which there had been a negative result using conventional genetic analysis (12). In 2008, the same group reported additional cases of low-level mosaicism at NLRP3 in 4 patients with CAPS (3 with CINCA/NOMID and 1 with MWS) (13). Overall, the data from these studies clearly supports the possibility that low-level mosaicism at NLRP3 could be the basis for CAPS patients exhibiting negative results in conventional NLRP3 analyses and could explain, at least partially, the genetic heterogeneity observed in CINCA/NOMID and probably in the other CAPS phenotypes. However, this hypothesis is still controversial since some investigators question whether low-level mosaicism at NLRP3 could be a general pathophysiologic mechanism underlying the disease in a majority of CINCA/NOMID patients who do not have the NLRP3 gene (23).

In the present report, we describe the clinical case of a patient with CINCA/NOMID whose disease was diagnosed on the basis of clinical characteristics, who had long-lasting successful clinical and biochemical responses to anakinra, and in whom exhaustive genetic analyses revealed the novel heterozygous p.D303H missense genetic variant in the NLRP3 gene. This newly identified genetic variant must be considered a true disease-causing mutation based on its absence among a panel of 400 control chromosomes, its de novo character, which is compatible with a dominant Mendelian inheritance pattern in a sporadic case, and its location in an amino acid residue of cryopyrin (Asp300) that is considered one of the mutational hotspots repeatedly observed to be the mutated residue in several patients with CINCA/NOMID and CINCA/NOMID overlapping with MWS (4, 5, 10, 11). Our investigations also revealed that this novel disease-causing mutation was detected in only ∼30–38% of analyzed cells. This contrasts with the percentage of mutation-harboring cells (100%) in which germline mutations have been detected and, along with the presence of low-level mosaicism at NLRP3, clearly supports the idea that the novel mutation was a somatic, nongermline mutation.

We also analyzed genomic DNA from physically isolated cells in order to address the possibility that this somatic mutation could be restricted to certain leukocyte subpopulations. Our investigations revealed that a similar proportion of neutrophils, T cells, and B cells carried the mutated allele, suggesting that a proportion of both common lymphoid and myeloid progenitors were also mutated. If this is correct, the findings of this investigation should have been the same in the monocyte subpopulation, since it shares a common hematopoietic progenitor with the neutrophil lineage. However, to our surprise, the p.D303H mutation was unexpectedly absent among isolated monocytes. One potential explanation for this result is that the presence of the novel NLRP3 mutation in a subgroup of monocytes could confer upon them some functional properties that render them nonviable, and consequently, this little subgroup could be lost during the isolation procedure. In this sense, it is interesting to note that previous studies have shown that different Toll-like receptor (TLR) agonists, such as lipopolysaccharide, induce a rapid necrosis-like death of those monocytes that harbor one NLRP3 mutated allele (13). The possible presence of trace amounts of TLR agonists among the reagents used during the isolation procedure might account for the rapid death of the mutated allele–harboring monocytes and for the mutated allele's consequent absence among the isolated and analyzed monocyte subpopulation.

Finally, we also detected the p.D303H mutation in cells from nonhematopoietic lineages, such as urinary and buccal epithelial cells. Thus, we have detected the novel NLRP3 mutation in both ectoderm- and mesoderm-derived cells, and it is probably also present in endoderm-derived cells, which were not analyzed. These findings suggest that the event that originated the novel mutation probably occurred early in embryonic development (24).

Since the discovery of the common molecular basis among the CAPS diseases, multiple efforts have been made to establish a clear genotype/phenotype correlation, which would help explain the clinical diversity observed in patients with the CAPS diseases, which could be useful in genetic counseling. However, these studies have only shown that those NLRP3 mutations detected in severe CINCA/NOMID do not cause the milder disease, FCAS, and vice versa (2, 10). Thus, the clinical phenotype in a patient with CAPS could probably be determined by both genetic and environmental factors.

From the data described herein, which are consistent with the results of previous studies, we believe that the main genetic factors that determine the phenotype are the nature of the disease-causing mutation (germline versus somatic), the amino acid residue where the mutation is located, and in the case of somatic mutations, the number of mutation-harboring cells, which determine its detection by molecular methods and its distribution among different cell lineages. CINCA/NOMID patients with somatic NLRP3 mutations usually exhibit both mild overall disease severity and mild neurologic symptoms compared with those patients with germline NLRP3 mutations (12, 13), as was the case in the patient we observed. The low-level mosaicism at NLRP3 observed in the severest form of CAPS prompted us to hypothesize that a similar pathophysiologic mechanism was probably the basis of sporadic cases of the other CAPS phenotypes (FCAS and MWS), which are probably milder than those phenotypes caused by germline NLRP3 mutations and, consequently, more difficult to identify. Further studies are needed to verify this hypothesis, which will probably require the use of sequencing methods other than conventional Sanger method sequencing.

Somatic mosaicisms have been extensively studied in cancer, but data are scarce in the context of Mendelian inherited diseases (25, 26), for which there have been 2 main types of studies describing somatic mosaicism. First, there are those studies that have shown the presence of somatic reversions from mutated to wild type in patients with a clear familial history of Mendelian inherited disease (27–29). And second, there have been a few studies that have clearly shown the presence of a somatic, true disease-causing mutation in sporadic cases of dominant Mendelian inherited diseases (12, 13, 30). The current report falls in this second group, since we have clearly demonstrated the precise causal relationship between the clinical phenotype and the presence of a somatic NLRP3 mutation. Moreover, for the current case, we report novel evidence that is consistent with previously published data from the Japanese group (12, 13) concerning the role of low-level mosaicism as a pathophysiologic mechanism in CAPS, a role which seems to be more extensive than had been thought. How a somatic mutation appears as the disease-causing mutation in a dominant Mendelian inherited disease and how the nature of that mutation can be demonstrated are difficult questions to answer.

Several mechanisms have been proposed to explain the phenomenon of somatic mosaicism, such as chimerism from cell fusion from an aborted dizygotic twin or a mutational event early in embryogenesis (31). Though we believe that the latter mechanism is the more probable explanation in the case of the current patient, we cannot formally demonstrate any mechanism, for a variety of reasons. For example, we cannot physically isolate mutation-harboring cells from wild-type cells, which might enable us to analyze polymorphic DNA markers to verify the chimerism hypothesis. Also, to verify the hypothesis of a mutational event early in embryogenesis and to narrow down when the event occurred, we would need to obtain tissue other than the tissue that was analyzed. To do so would probably involve invasive procedures that would be ethically questionable to perform in this child, since there is already a definitive diagnosis with an appropriate and successful treatment.

In conclusion, we have identified the novel p.D303H NLRP3 genetic variant in a young Spanish patient with CINCA/NOMID as a new CAPS-causing mutation. Our investigation demonstrated that this novel mutation was detected as a low-level somatic mosaicism in peripheral blood and in cells of nonhematopoietic lineage, suggesting the occurrence of a mutational event early in embryogenesis. Our data are also consistent with recent evidence of the causal role of low-level somatic mosaicism at NLRP3 in CAPS, which could explain, at least partially, the previously described genetic heterogeneity of the CAPS diseases. Furthermore, other dominant Mendelian inherited diseases, including autoinflammatory diseases other than CAPS (e.g., tumor necrosis factor receptor–associated periodic syndrome and Blau syndrome), could result from somatic nongermline mutations, which probably cannot be detected by conventional analyses and which will require alternative approaches for an appropriate genetic diagnosis.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Aróstegui 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 conception and design. Aróstegui, Lopez Saldaña, Ibañez, Yagüe.

Acquisition of data. Aróstegui, Pascal, Clemente, Aymerich, Balaguer, Goel, Fournier del Castillo, Rius, Plaza, Juan.

Analysis and interpretation of data. Aróstegui, Lopez Saldaña, López Robledillo, Juan, Ibañez, Yagüe.

Acknowledgements

We would like to thank the patient and his family for their cooperation and Sandra Martínez for her technical assistance with the isolation of leukocyte subpopulations.

Ancillary