Autoinflammatory Syndromes with Macrophage Activation (NLRC4, CDC42, DPP9 and Related)

Interleukin-18-mediated inflammasomopathies

Update of the literature and regulatory information: 31 October 2026.

0. Initial Reference Table

This article groups three genes, NLRC4, CDC42 and DPP9, whose diseases share a mechanism and a clinical problem: permanent activation of an inflammasome, excessive production of interleukin 18 and a tendency to macrophage activation syndrome. The entities caused by variants in NLRP1 and CARD8 are included in this article because they belong to the same molecular axis, in which dipeptidyl peptidase 9 is the physiological brake on the NLRP1 and CARD8 inflammasomes. XIAP deficiency, which also runs with raised interleukin 18 and recurrent haemophagocytic lymphohistiocytosis, and primary haemophagocytic lymphohistiocytosis appear here only in the differential diagnosis. The cryopyrinopathies and familial Mediterranean fever, which share the inflammasome pathway but not the risk of macrophage activation, belong to the group of classical monogenic autoinflammatory diseases.

1. Introduction

The discovery that gain-of-function variants in NLRP3 caused the cryopyrin-associated periodic syndromes established, at the beginning of the twenty-first century, the paradigm of the inflammasomopathies: diseases in which a sensor of innate immunity is locked in the active state, caspase-1 is cleaved continuously and the cytokines of the interleukin-1 family are produced without a stimulus. That paradigm explained fever and urticaria well, and also explained the striking response to interleukin-1 blockade, but it did not explain a complication that is rare in the cryopyrinopathies: macrophage activation syndrome [1, 2].

In September 2014 two groups described, independently and in the same journal, the first human diseases caused by gain of function of NLRC4. Scott Canna and colleagues, at the United States National Institutes of Health, studied a seven-year-old girl of European ancestry with recurrent episodes of fever, malaise, splenomegaly, vomiting and loose stools from six months of age, whose laboratory profile, with chronic anaemia, raised transaminases, hypertriglyceridaemia, hyperferritinaemia, leucopenia and thrombocytopenia during flares, was not compatible with a cryopyrinopathy but suggested macrophage activation syndrome; exome sequencing identified the de novo variant NM_001199138.2:c.1009A>T p.(Thr337Ser), in the nucleotide-binding domain of NLRC4, and functional studies showed spontaneous inflammasome formation with production of interleukin 1 beta and, above all, of interleukin 18 at levels higher than those seen in neonatal-onset multisystem inflammatory disease [1]. Neil Romberg and colleagues, at Yale, studied a family in which the index case presented at one week of age with secretory diarrhoea and fever and died on the twenty-third day of diffuse alveolar haemorrhage, and in which the father, two days after his son's funeral, had a severe inflammatory episode with disseminated intravascular coagulation; the causal variant, NM_001199138.2:c.1022T>C p.(Val341Ala), had arisen de novo in the father and co-segregated with the disease, and the authors proposed the acronym SCAN4 for the condition [3]. In the same year, Akiko Kitamura and colleagues described in a three-generation Japanese family a much milder phenotype, with fever, non-pruritic urticarial rash and arthralgia triggered by cold from two or three months of age, attributed to the variant p.(His443Pro) and named familial cold autoinflammatory syndrome type 4 [4, 5].

The sequence of events that followed defined the field. In 2016, an infant with p.(Val341Ala) disease refractory to high-dose corticosteroids, anakinra, infliximab, ciclosporin and vedolizumab received, under an emergency compassionate-use authorisation from the United States regulatory agency, recombinant human interleukin-18 binding protein, with improvement within two doses and weaning of all other immunosuppression; this was the first clinical demonstration that interleukin 18 was a therapeutic target [6]. In 2018, Esther Weiss and colleagues showed, in a systematic comparison of cohorts with hyperferritinaemic and autoinflammatory diseases, that chronic elevation of mature interleukin 18, and particularly of the fraction not bound to its binding protein, identified patients at risk of macrophage activation and separated these conditions from familial haemophagocytic lymphohistiocytosis [7]. In 2024, marked elevation of interleukin 18 was included in the European recommendations for the diagnosis of Still's disease, consolidating the concept, proposed by Emily Landy and colleagues, of a group of diseases of interleukin-18 excess [8, 9].

The story of CDC42 is different because the gene was already known. In 2015, Toshiki Takenouchi and colleagues described a girl with persistent macrothrombocytopenia and developmental delay caused by a de novo CDC42 variant, and in 2018 Simone Martinelli and colleagues showed, through functional characterisation of each variant, that missense changes in this gene produce a heterogeneous spectrum of developmental phenotypes, with growth dysregulation, facial dysmorphism and neurological, immunological and haematological anomalies [10–12]. In 2019, Michael Lam and colleagues identified in four unrelated patients the same de novo variant NM_001791.4:c.556C>T p.(Arg186Cys), responsible for a condition that resembled none of the previous ones: neonatal pancytopenia with dyshaematopoiesis, autoinflammation, rash and episodes of haemophagocytic lymphohistiocytosis, for which they proposed the acronym NOCARH; three of the four patients died and the survivor was saved only by interferon-gamma neutralisation followed by haematopoietic cell transplantation [13]. In the same year, Yael Gernez and colleagues described four patients with variants in the last C-terminal residues of CDC42, including two distinct from p.(Arg186Cys), with severe neonatal autoinflammation that responded to interleukin-1 blockade, and Simone Coppola and colleagues showed in 2022 that the three C-terminal variants disturb the protein in different ways and therefore cause different diseases [14, 15].

The third disease of the group stems from the discovery of the brake. Several structural biochemistry studies showed that dipeptidyl peptidase 9 forms a ternary complex with two molecules of NLRP1, or of CARD8, and sequesters the bioactive C-terminal fragment of these sensors, preventing inflammasome assembly. In 2022, Cassandra Harapas and colleagues identified in four children from three families biallelic DPP9 variants, null or hypomorphic, that no longer repressed NLRP1, with a triad of failure to thrive, cutaneous manifestations and pancytopenia with susceptibility to infection, named Hatipoglu syndrome [16, 17]. In 2023, Christine Wolf and colleagues described an infant with fulminant hyperinflammation resembling haemophagocytic lymphohistiocytosis caused by a de novo heterozygous variant in the same gene, with dominant-negative behaviour, which showed that dipeptidyl peptidase 9 deficiency can be dominant or recessive depending on the mechanism [18]. In 2026, Tianyang Xiao and colleagues explained in a humanised mouse model why the patients' pancytopenia is not reproduced in the conventional mouse: the death of human haematopoietic progenitor cells depends on the CARD8 inflammasome, which is absent from the murine genome [19].

The nomenclature of this group is unstable and should be fixed. The severe NLRC4 disease appears in the literature as autoinflammation with infantile enterocolitis, as NLRC4-associated macrophage activation syndrome and as SCAN4; the 2024 International Union of Immunological Societies classification uses the designation NLRC4-MAS in the subtable of inflammasome defects, and Orphanet brings both expressions together under code 436166 [20, 21]. NOCARH syndrome has no phenotype number in OMIM, despite having an Orphanet code, a MONDO code and a designation in the SNOMED CT clinical terminology; the International Union of Immunological Societies classification names it “CDC42 deficiency” and assigns it, for lack of an alternative, the number of Takenouchi-Kosaki syndrome [20, 22]. Together with Still's disease, XIAP deficiency and interleukin-18 binding protein deficiency, these diseases make up what the French and English-language literature has come to call interleukin-18-associated autoinflammatory diseases, with the caveat that, among these, those associated with PSTPIP1 do not carry a risk of macrophage activation [9, 23].

2. Molecular Genetics and Pathophysiology

2.1 The inflammasome as a platform and interleukin 18 as the distinctive product

Canonical inflammasomes are cytosolic complexes formed by a sensor protein, the adaptor ASC, encoded by PYCARD, and pro-caspase-1. Oligomerisation of the sensor brings pro-caspase-1 molecules together, and these autocleave; active caspase-1 converts pro-interleukin 1 beta and pro-interleukin 18 into their mature forms and cleaves gasdermin D, whose N-terminal fragments form pores in the membrane and cause the inflammatory cell death known as pyroptosis [24]. Secretion of the interleukin-1 family cytokines normally requires two signals: the first, from membrane receptors such as the Toll-like receptors, induces expression of the precursors; the second, detected by cytosolic sensors, triggers caspase-1 activation [3].

Interleukin 18 is the product that distinguishes this group of diseases from the other inflammasomopathies. It is expressed as a precursor not only by myeloid cells but also by epithelia, above all the intestinal epithelium; its receptor is most abundant on cytotoxic lymphocytes, and its best-known function is to amplify interferon-gamma production. It has an endogenous antagonist, interleukin-18 binding protein, induced by interferon gamma itself, which prevents signalling but not detection of the cytokine by conventional immunoassays; this is why measurement of the free fraction is more informative about biological activity than measurement of the total [7, 9]. Experimental proof of the importance of the free fraction came from murine models: mice with raised free interleukin 18 had more severe experimental macrophage activation, whereas mice with the Nlrc4 T337S variant, whose free interleukin 18 remained normal, did not [7].

2.2 NLRC4: structure, mechanism and variant spectrum

The NLRC4 gene, formerly named CARD12, IPAF and CLAN, is located at 2p22.3, spans coordinates 2:32,224,449–32,265,743 in the GRCh38 reference genome and has as its MANE Select transcript NM_001199138.2, which encodes the 1024-amino-acid protein NP_001186067.1, on which all NLRC4 protein variants cited in this article are numbered [25, 26]. The protein is an NLR-family receptor with five domains whose residue boundaries are: caspase activation and recruitment domain (CARD) 1 to 94, nucleotide-binding domain (NBD) 95 to 298, helical domain 1 (HD1) 299 to 355, winged-helix domain (WHD) 356 to 463, helical domain 2 (HD2) 464 to 580 and leucine-rich repeats (LRR) 581 to 1024 [24]. In cooperation with the NAIP proteins, NLRC4 detects cytosolic components of pathogenic bacteria, namely flagellin and type III secretion system proteins of Salmonella enterica serovar Typhimurium, Pseudomonas aeruginosa and Legionella pneumophila, which are injected into the cytosol by the bacterium itself; unlike Toll-like receptor 5, which recognises flagellin at the cell surface, NLRC4 recognises it inside the cell [1, 3].

In the inactive state, NLRC4 is held in an autoinhibited conformation, bound to adenosine diphosphate, with helical domain 1 acting as a lid over the nucleotide-binding pocket. Pathogenic variants are almost all missense and destabilise this autoinhibition. Val341 makes van der Waals contacts with side chains of an adjacent helix of helical domain 1; substitution by alanine reduces the hydrophobicity of this interaction, allows greater mobility of helix α12 and favours exchange of adenosine diphosphate for adenosine triphosphate, or disturbs the stabilising interaction of His443 with the beta-phosphate group, either of which leads to ligand-independent activation [3]. The p.(Thr337Ser) variant may destabilise the interactions of helical domain 1 with residues 170 and 173 of the nucleotide-binding domain or directly affect adenosine diphosphate binding [1]. In the leucine-rich repeat domain, which acts as a negative regulator, loss of regulatory function likewise results in a constitutively active inflammasome, and the first in-frame deletion of 93 base pairs in this domain was described in a patient with a cold-induced phenotype [27, 28].

The cellular consequences have been documented convergently. In HEK293 cells co-transfected with ASC and caspase-1, the p.(Val341Ala) variant produced a fourfold increase in spontaneous caspase-1 cleavage (P=0.004) and more ASC specks than the normal protein, 12.3% versus 7.6% of cells (P<0.0001) [3]. In THP-1 cells transduced with the p.(Thr337Ser) variant, spontaneous caspase-1 cleavage was observed, with interleukin 1 beta secretion 5.5 to 10.6 times that of the empty vector, against 2.4 to 5.9 times for the normal protein (P=0.01), and interleukin 18 secretion 5.2 to 11.4 times against 2.1 to 6.7 times (P=0.03) [1]. Monocyte-derived macrophages from patients, given only the first signal, secreted 76 pg/mL of interleukin 1 beta and 212 pg/mL of interleukin 18, against less than 1.4 and 22 pg/mL in controls (P<0.0001 for both), and released more lactate dehydrogenase, 12.3% versus 4.7% (P<0.0001); caspase-1 inhibition reduced cytokine secretion but not cell death, which shows that pyroptosis is promoted independently of cytokine processing [3]. In patients' monocytes and macrophages, interleukin 18 was secreted constitutively, unlike in NLRP3 disease, and stimulated macrophages produced four to five times more interleukin 18 than monocytes, which identifies the activated macrophage as a relevant source [1].

The origin of circulating interleukin 18 is more complex than haematopoietic. In the mouse with the Nlrc4 T337S variant, the systemic elevation of the cytokine came entirely from the intestinal epithelium, whose gut was histologically normal but showed increased epithelial turnover and overexpression of interferon-gamma-induced genes; on expression analysis, Il1b, Nlrp3 and Mefv predominated in neutrophils, whereas Nlrc4 and Il18 were markedly epithelial [7]. In a patient with somatic mosaicism restricted to the haematopoietic compartment, total and free interleukin 18 were raised but there was no macrophage activation, which suggests that the fraction of haematopoietic origin alone is not enough to trigger it and confirms that interleukin 18 is necessary but not sufficient [29].

The compilation available up to 2026 counts 28 pathogenic NLRC4 variants in 56 people, 23 germline and 5 somatic [24]. Genotype–phenotype correlation is weak: presentations with macrophage activation are found with variants in the nucleotide-binding domain and in the leucine-rich repeat domain, and mild phenotypes with variants in the same nucleotide-binding domain, so the location of the change does not predict the severity of the flare [27]. It has nevertheless been observed that vasoplegic shock has been described only with nucleotide-binding domain variants and that mild phenotypes have been associated preferentially with two winged-helix domain variants [27, 36]. Part of the heterogeneity arises from the cellular compartment in which the variant acts: epithelial signalling produces mainly the intestinal pathology, through local excess of interleukin 18 and barrier disruption, whereas activation in myeloid cells contributes more to the systemic manifestations [24].

Somatic mosaicism is an established mechanism and has been described in five patients. Mosaic variants cluster near the adenosine diphosphate binding site, and the allele fraction correlates with age at onset and severity: fractions of 25% and 30% were associated with congenital or neonatal disease, fatal in one case at two months, whereas fractions of 2.6% to 5% were associated with disease beginning at 21, 47 and 60 years without macrophage activation [32]. The fraction may be greater in a specific subpopulation, such as the 5.69% documented in monocytes from a patient with late-onset disease [39]. This observation has a practical consequence: in an adult with an uncharacterised autoinflammatory disease, a variant with a low allele fraction, which conventional sequencing may disregard, should be considered potentially causal, and the International Union of Immunological Societies classification expressly recommends that laboratories report such findings [20, 32].

Variant interpretation requires additional caution. Curation records indicate for NLRC4 a marked excess of variants of uncertain significance, 574 against 32 pathogenic and 12 likely pathogenic in ClinVar, which means that an incidental missense change in this gene, found in a person without autoinflammation, is far more likely to be uninformative than causal; only variants shown or predicted to be activating, supported by the interleukin-18 signature or by functional demonstration, should be considered causal [42]. The validity of the gene–disease relationship is nonetheless solid for the enterocolitis phenotype, with definitive and strong classifications in three independent submissions to the GenCC database, always with autosomal dominant inheritance [42].

2.3 CDC42: one GTPase, several mechanisms, several diseases

The CDC42 gene is located at 1p36.12 and has as its MANE Select transcript NM_001791.4, which encodes the 191-amino-acid NP_001782.1, on which the CDC42 protein variants cited are numbered [26]. It encodes a small GTPase of the Rho family that alternates between an active state bound to guanosine triphosphate and an inactive state bound to guanosine diphosphate, and acts as a signalling hub of the actin cytoskeleton. Its activity is controlled by three classes of regulators, the guanine nucleotide exchange factors, the GTPase-activating proteins and the guanine nucleotide dissociation inhibitors; it acts through more than forty effectors, including the Wiskott-Aldrich syndrome protein, p21-activated kinase and IQGAP1, controlling filopodium formation, epithelial and neuronal polarity, migration, phagocytosis, vesicular trafficking and bipolar attachment of microtubules to kinetochores [13, 43]. The reversible localisation of the protein on the cytoplasmic face of membranes is regulated by the nucleotide dissociation inhibitor RhoGDI, which controls cycling between membrane and cytoplasm, and by IQGAP1, which promotes translocation from the Golgi apparatus to the plasma membrane. CDC42 function requires post-translational processing of the C-terminus, with prenylation at Cys188, usually geranylgeranylation, proteolytic cleavage of the last three residues and carboxyl methylation [13]. There are two isoforms: isoform 1 is ubiquitous and isoform 2 predominates in the brain, and the autoinflammatory variants affect an exon specific to isoform 1 [13].

The relationship between variant and disease in this gene is determined by the exact amino acid, not by the gene. Three phenotypic groups are established. The first is Takenouchi-Kosaki syndrome, closely linked to the recurrent variant p.(Tyr64Cys), with psychomotor delay and variable intellectual disability, recognisable facial dysmorphism, macrothrombocytopenia, lymphoedema, camptodactyly, short stature of prenatal onset, microcephaly, congenital heart disease and non-specific changes on brain imaging [10, 12]. The second is a broader developmental spectrum, caused by variants distributed along the gene, which disturb the GTPase in distinct ways, altering the cycle between active and inactive states or effector binding, and whose syndromic delineation was possible only through the functional profile of each variant; it includes presentations reminiscent of Noonan syndrome, and variants that behave as inactivating are specifically associated with thrombocytopenia [11]. The third group is autoinflammatory and results from variants in the last C-terminal residues, near a di-arginine motif (Arg186 and Arg187) with an essential role in binding of the protein to membranes containing phosphatidylinositol 4,5-bisphosphate and in its lipidation [14].

The p.(Arg186Cys) variant does not reduce synthesis or stability of the protein, but it decreases binding to WASP and drastically reduces interaction with IQGAP1, with altered interaction with RhoGDI as well; the consequence is a constitutive localisation restricted to the cis and medial Golgi, which is not seen with a representative panel of other pathogenic variants, including p.(Tyr23Cys), p.(Arg68Gln), p.(Ser83Pro), p.(Ala159Val) and p.(Glu171Lys) [13]. This mislocalisation translates into measurable functional defects: absent proliferative response of CD34-positive cells to KIT ligand, reduced polarised migration and chemotaxis towards CXCL12, multipolar morphology with multiple lateral actin edges instead of canonical polarisation, filopodial changes in natural killer cell lines and decreased natural killer cytotoxicity; in the nematode Caenorhabditis elegans, the variant behaves as hypomorphic in vulval development [13]. Golgi retention, demonstrated with induced pluripotent cells from a patient, promotes assembly of the pyrin inflammasome by a mechanism independent of GTPase activity: the derived macrophages produced more interleukin 1 beta and interleukin 18 in response to Clostridioides difficile toxin A, an activator of the pyrin inflammasome; the excess was suppressed by pyrin inhibitors, including colchicine, but not by the NLRP3 inhibitor MCC950, was suppressed by MEFV silencing with no effect on interleukin 6, and disappeared with genomic correction of the variant [44]. The same entrapment blocks anterograde and retrograde transport between endoplasmic reticulum and Golgi, accumulates STING in the Golgi in a COPI-complex-dependent manner, hyperactivates the STING pathway and the type I interferon response and causes endoplasmic reticulum stress, a mechanism shared by the p.(Ter192CysextTer24) variant and absent from the p.(Tyr64Cys) and p.(Cys188Tyr) variants, which do not accumulate in the Golgi [45]. Consistently, patients carrying Golgi-retained variants had very high circulating interferon alpha at disease onset, reduced to a variable degree by treatment or after bone marrow transplantation, and an independent study showed that intrinsic type I interferon activation, triggered by sensing of cytosolic nucleic acids released from mitochondria with altered integrity and dynamics, is a driver of autoinflammation in NOCARH syndrome [45, 46].

The other two C-terminal variants behave differently. p.(Cys188Tyr) and p.(Ter192CysextTer24), the latter resulting from read-through of the stop codon with the addition of 24 amino acids, promote accelerated degradation of the protein; the former, being non-prenylated, no longer behaves as a membrane-bound protein, whereas the residual product of the latter replicates the behaviour of p.(Arg186Cys) and is targeted to the Golgi in a palmitoylation-dependent manner. Both behave as loss of function in polarised migration assays and in C. elegans development, and are associated with clinically distinct disease in dysmorphism, severity, age at onset of cytopenia and extent of autoinflammation [15]. In 2026 an additional autoinflammatory condition caused by p.(Met45Leu) was described, which promotes pyrin inflammasome activation by a different mechanism: CDC42 interacts with the PYD and B30.2 domains of pyrin, the interaction with the B30.2 domain being greatly enhanced by the variant, and the proposed model is that CDC42 nucleates pyrin through this domain until a threshold is reached that releases the PYD domain to activate the inflammasome [47, 48]. Alongside these, there are reports of variants that combine a syndromic phenotype with autoinflammation, such as p.(Cys81Tyr) in an eleven-year-old boy with facial dysmorphism, neurodevelopmental delay, immunodeficiency, haemophagocytic lymphohistiocytosis and Hodgkin lymphoma, and descriptions of systemic inflammation with myelofibrosis in a patient with the classical Takenouchi-Kosaki syndrome variant, which shows that the separation between the groups is not absolute [49–51].

2.4 DPP9, NLRP1 and CARD8: the brake and the sensors it controls

The DPP9 gene is located at 19p13.3, occupies 19:4,675,227–4,723,842 in GRCh38 and has as its MANE Select transcript NM_139159.5, which encodes the long isoform NP_631898.3, of 892 amino acids, on which the DPP9 variants cited are numbered unless otherwise stated, such as p.(Arg111Ter), p.(Ser214Ter) and p.(Gln851Ter), recorded in ClinVar as NM_139159.5:c.331C>T, c.641C>G and c.2551C>T [26, 52, 53]. It encodes dipeptidyl peptidase IX, a protease of 863 amino acids in the short isoform and about 98 kilodalton, with a GWSYG serine protease motif, without a signal sequence or transmembrane domain, ubiquitously expressed, maximally in skeletal muscle, heart and liver and minimally in the brain, with 58% identity to dipeptidyl peptidase 8 [52]. Use of an alternative start codon gives rise to a protein with additional amino acids at the N-terminus, which explains the dual residue numbering in the literature and requires attention when reading genetic reports [18, 52].

Its relevant function here is inhibitory. One monomer of the dipeptidyl peptidase 9 dimer binds two NLRP1 molecules, one in the conformation preceding autoproteolysis and the other in the form of the cleaved UPA-CARD fragment, and thus sequesters the bioactive C-terminal fragment, preventing inflammasome formation; the same occurs with CARD8 (CARD8-dependent pyroptosis in DPP9 deficiency and CARD8-associated inflammatory bowel disease belong to the group of autoinflammatory diseases). Both the catalytic activity and the structural scaffolding function are required for this repression [16, 19]. During inflammasome activation, the N-terminal peptide of the sensor undergoes functional degradation by the proteasome, which releases additional cytotoxic C-terminal fragments and saturates the complexes with dipeptidyl peptidase 9 [19].

The biallelic variants described behave as null or hypomorphic alleles. The p.(Gln851Ter) variant produces no detectable protein, through nonsense-mediated decay of the transcript; p.(Gly167Ser), situated in a flexible loop at the periphery of the substrate-binding site, reduces enzyme activity to about 15% of normal; the p.(Arg111Ter) and p.(Ser214Ter) variants truncate the protein before the catalytic domain, which lies at the C-terminus [16]. In reporter cells with double knockout of DPP8 and DPP9, re-expression of the normal protein repressed the formation of NLRP1-dependent ASC specks, whereas the patients' variants were less effective, despite retaining the ability to bind NLRP1. Primary keratinocytes from the patient spontaneously secreted active interleukin-1 beta and interleukin-18, with endogenous ASC oligomerisation and low gasdermin D cleavage, remaining viable in culture for at least four passages; cytokine release was lower than that obtained with the pharmacological inhibitor Val-boro-Pro, which is explained by partial compensation from the dipeptidyl peptidase 8 still present [16]. Across population databases the gene is intolerant to loss-of-function variants, and no damaging homozygotes have been described in the general population [16].

The dominant variant described acts by a different mechanism. The heterozygous de novo change affecting residue 252 of isoform 1, corresponding to residue 281 of the long isoform and designated c.842G>C, p.(Arg281Pro), in the MANE transcript, destabilises the protein, whose residual expression was 52% and 26% of normal in two cellular systems and was likewise reduced in induced pluripotent cells from the patient; the mutant protein no longer restrained the NLRP1 and CARD8 inflammasomes, which were constitutively activated, a behaviour interpreted as dominant-negative [18, 26]. The existence of these two mechanisms explains why the gene is associated with two patterns of inheritance, the autosomal recessive one of Hatipoglu syndrome and the autosomal dominant de novo one of the hyperinflammatory form [17, 18].

A finding from 2026 clarified the haematology of this disease. In humanised MISTRG6 mice engrafted with CRISPR-edited CD34-positive human haematopoietic progenitor cells, deletion of DPP9 reproduced the peripheral cytopenia and the loss of marrow progenitors, in a cell-intrinsic manner, with minimal transcriptional changes, which points to post-transcriptional regulation; cell death was pyroptosis dependent on activation of the CARD8 inflammasome and of caspase-1, and deletion of CARD8 or of CASP1 rescued the cytopenia, whereas the absence of NLRP1 did not. The CARD8 inflammasome is absent from the genome of most rodents, including the mouse, and is well expressed in human haematopoietic cells, which explains why conventional murine models of Dpp9 mutation maintain normal haematopoiesis although pancytopenia dominates the human picture [19].

The NLRP1 gene, at 17p13.2, causes through gain of function three distinct mendelian diseases, which should not be described as one: multiple self-healing palmoplantar carcinoma, autosomal dominant, with keratopathy, palmoplantar hyperkeratosis and keratoacanthoma-like lesions in the skin and in the conjunctival and corneal epithelia, with susceptibility to squamous cell carcinoma; autoinflammation with arthritis and dyskeratosis, whose OMIM entry records dominant and recessive inheritance, with recurrent fever, diffuse cutaneous dyskeratosis, arthritis, autoimmunity and raised transitional B lymphocytes, accompanied by raised systemic caspase-1 and interleukin-18; and congenital juvenile-onset recurrent respiratory papillomatosis, autosomal recessive and flagged as provisional, in which two Belgian brothers born to consanguineous parents developed laryngeal papillomas without detection of human papillomavirus types 6 or 11, through a homozygous variant with spontaneous oligomerisation of the sensor [54–57]. The association of the gene with susceptibility to vitiligo and to multiple autoimmunity, listed in braces in OMIM, is an association of common variants and does not predict disease in an individual [57]. For CARD8, the only mendelian relationship proposed is a form of inflammatory bowel disease with a Crohn phenotype, flagged as provisional, supported by a single family with a missense change in the T60 isoform, raised serum interleukin-1 beta, monocytes overproducing this cytokine on NLRP3 stimulation and a mutant protein unable to bind NLRP3 and behaving as dominant-negative; Orphanet recognises the entity but records only four cases worldwide, and the common variants of the gene studied in Crohn disease and in periodic fever with aphthous stomatitis, pharyngitis and adenitis syndrome are susceptibility factors, not mendelian disease [58, 59].

2.5 Penetrance, expressivity and allelic diseases

In the germline forms of NLRC4, vertical transmission is documented, with cosegregation of the p.(Val341Ala) variant in one family and of p.(His443Pro) in a three-generation Japanese family, and with a Spanish family in which six consecutive generations were affected by the p.(Ser445Pro) variant [3, 4, 40]. Expressivity is highly variable within the same family: in the family described by Romberg, the index case died in the neonatal period and the father, carrier of the same variant that had arisen de novo, had a life with recurrent fevers and a single severe episode at 43 years, whereas the digestive symptoms of the half-brother were confined to the first year [3]. In the C-terminal variants of CDC42, all published cases arose de novo, without consanguinity [13, 14]. Hatipoglu syndrome followed a recessive pattern in every family, with healthy heterozygous parents, including in a consanguineous Bedouin family from Israel and in a family of unrelated Ashkenazi parents [16].

Allelic diseases are clinically relevant in two of the genes. In NLRC4, the same p.(Ala160Thr) variant causes, in homozygosity, recessive autoinflammatory disease and, in heterozygosity, an increased risk of ulcerative colitis, with an odds ratio of 2.546 and a 95% confidence interval between 1.778 and 3.644 (P=0.01305) in an analysis of exome data from inflammatory bowel disease [33]. In CDC42, Takenouchi-Kosaki syndrome and the developmental spectrum are allelic diseases of NOCARH syndrome, with distinct pathophysiology and management [11, 15]. Besides these, interleukin-18 is raised in genetically distinct settings that belong to the same axis: deficiency of interleukin-18 binding protein, from biallelic IL18BP variants, presents as fulminant viral hepatitis; PSMB9 deficiency runs with raised interleukin-18 and severe neonatal autoinflammation; and TIM3 deficiency, from HAVCR2 variants, associates increased inflammasome activity with panniculitis and haemophagocytic lymphohistiocytosis [20].

3. Clinical Spectrum and Manifestations by System

3.1 NLRC4-associated disease: four phenotypes in a single gene

The literature recognises four presentations of NLRC4-associated autoinflammatory disease: autoinflammation with infantile enterocolitis, NLRC4-associated macrophage activation syndrome, familial cold autoinflammatory syndrome type 4 and a condition resembling neonatal-onset multisystem inflammatory disease [24, 30]. A systematic review of the literature from France, published in 2018 with the cases then known, proposed a division into two phenotypes that remains useful: a severe phenotype, corresponding to about 30% of the cases then published, dominated by multisystem inflammation in the first year of life with symptoms of inflammatory bowel disease, macrophage activation or a picture resembling chronic infantile neurological, cutaneous and articular syndrome, with a mortality of 25%; and a mild phenotype, about 70%, usually beginning after three years of age, with cold urticaria, arthralgia, ocular manifestations and fever in half of cases, without visceral failure [60].

Autoinflammation with infantile enterocolitis is the severe end. It begins in the neonatal period or in the first months, with recurrent febrile flares, secretory or bloody diarrhoea, severe enterocolitis, splenomegaly, rash and, above all, recurrent macrophage activation syndrome, which is what makes the disease potentially fatal [42, 61]. The enterocolitis is characterised by chronic secretory diarrhoea, abdominal pain, villous atrophy and transmural inflammatory infiltration of the mucosa; all described cases of NLRC4 enterocolitis began in infancy, and in survivors the gastrointestinal inflammation attenuates or resolves with age, whereas the systemic manifestations may persist into adulthood [3, 24]. In the founding cases, the enterocolitis resolved around the first year of life in all survivors, and the proposed explanation was maturation of host–microbiota interactions, since the infant gut is subject to continuous stimulation by newly acquired symbionts and has high epithelial turnover, in an organ where NLRC4 is highly expressed in macrophages and intestinal epithelial cells and NLRP3 is not [3, 24]. Weight gain is frequently compromised, owing to chronic inflammation, protein-losing enteropathy and corticosteroid dependence [1, 61].

The presentation can be dramatic and mimic other emergencies. Vasoplegic shock requiring intubation and inotropic support has been described in newborns and infants, and led to initial diagnoses of food protein-induced enterocolitis or sepsis; normal methaemoglobin and the absence of a link to a specific food help to exclude severe food allergy, and vasoplegic shock does not occur in cow's milk protein allergy [27, 38]. Massive intestinal ischaemia and necrosis requiring resection, intussusception, gallbladder hydrops, perianal abscess, ascites and gastrointestinal haemorrhage have been reported [13, 27]. In a 28-week preterm infant with mosaicism for the p.(Ser171Phe) variant, the disease was already manifest in utero, with congenital anaemia, ascites and an oedematous placenta with fetal thrombotic vasculopathy, followed by hepatosplenomegaly and systemic inflammation with laboratory criteria for haemophagocytosis, and death at two months; autopsy showed hepatic sinusoidal, marrow and lymph node haemophagocytosis and acute and chronic ischaemic bowel disease with features of necrotising enterocolitis [34]. In a term newborn, necrotising enterocolitis with multifocal necrosis over 80 centimetres of bowel, requiring ileostomy and resection, was described in association with a truncating variant in the gene; the report is weak, because the variant is not of the known activating type and the abstract itself is contradictory about its origin, so it cannot be used to assert causality [62].

Macrophage activation syndrome in this context presents as with any other cause: persistent fever, cytopenias, hyperferritinaemia, hypertriglyceridaemia, hypofibrinogenaemia, raised transaminases, coagulopathy and tissue haemophagocytosis, and it may be refractory to conventional therapy [42]. It differs from familial haemophagocytic lymphohistiocytosis in that T-lymphocyte and natural killer cell cytotoxic function is preserved, the macrophage-intrinsic inflammasome hyperactivation being the driver of the disease [1, 24]. The difference is pathophysiological and has diagnostic value, although natural killer cell lymphopenia was documented in the first patients and cytotoxic function was abnormal in chromium release assays in one of them [3].

Cutaneous manifestations are variable. Besides the urticarial rash, a macular rash progressing to a reticular livedoid pattern, erythematous nodules, recurrent panniculitis with perivascular and interstitial lymphohistiocytic infiltrates in the dermis, mixed panniculitis and psoriasis-like lesions have been described [27, 38, 40]. Hypereosinophilia of up to 2.4×10⁹/L was noted in one patient [38]. Among the less expected presentations are lupus nephritis in a girl with a gain-of-function variant, the identification of a variant in the gene in a woman with systemic lupus erythematosus and in her son with neonatal lupus complicated by macrophage activation, and central nervous system vasculitis in an adolescent with periodic fever, diarrhoea, aphthae and seizures [30, 63, 64]. These associations, described in isolated patients, require independent confirmation before being treated as an established part of the spectrum.

The cold-induced phenotype is considerably milder. In the founding Japanese family, episodes of high fever, non-pruritic urticarial rash and arthralgia began at two or three months of age, were generally triggered by exposure to cold, were not accompanied by splenomegaly or bone erosions and resolved without treatment in most cases, some patients using non-steroidal anti-inflammatory drugs for joint pain [4, 5]. In other families, joint manifestations persisted between episodes and some patients developed features resembling early-onset spondyloarthritis, and conjunctivitis and uveitis have also been described [24, 65]. In the Spanish family with the p.(Ser445Pro) variant, onset was between three months and six years and the picture was dominated by recurrent panniculitis, fever and arthralgia [40]. In patients with this phenotype, total and free interleukin 18 are raised, but interferon gamma, CXCL9 and CXCL10 are not, and there was no correlation with markers of inflammation nor any episodes of macrophage activation [29].

The condition resembling neonatal-onset multisystem inflammatory disease was described with mosaicism for the p.(Thr177Ala) variant: presentation at birth or in the neonatal period with continuous systemic inflammation, persistent fever and diffuse inflammatory rash, with prominent neurological involvement including aseptic meningitis, sensorineural hearing loss, developmental delay and cerebral atrophy on imaging, chronic inflammatory polyarthritis without erosive changes and craniofacial features classically associated with this disease, such as frontal bossing and saddle nose; these patients do not develop infantile enterocolitis or macrophage activation [24, 37].

Late onset due to somatic mosaicism constitutes a phenotype of its own. In the four patients with low allele fractions, the disease began between 21 and 60 years of age with recurrent fever, myalgia, arthralgia, large-joint arthritis, lymphadenopathy, abdominal pain with or without diarrhoea and systemic inflammation, without macrophage activation, and in some cases with scarlatiniform rash, cutaneous vasculitis, cellulitis, pleural effusion or auricular chondritis [32, 35, 39]. In one of them, the inflammatory crises coincided with decompensations of a chronic psychosis, without a causal relationship being established, and C-reactive protein reached 350 mg/L with moderate hyperferritinaemia between 482 and 662 ng/mL [32].

The recessive form described in a 60-year-old Brazilian woman shows yet another profile: episodes from six months of age with recurrent low-grade fever, chills, oral ulcers, uveitis, arthralgia and abdominal pain followed by diarrhoea with mucus and a variable rash, only partially controlled by high-dose corticosteroids and interleukin-1 beta blockade, with serum interleukin 1 beta and interleukin 18 persistently raised on treatment [33].

3.2 NOCARH syndrome and other CDC42 autoinflammatory diseases

The four patients of the original description shared an overlapping picture, beginning at birth or in the neonatal period, with trilineage pancytopenia, persistent fever, skin rash, hepatosplenomegaly and persistently raised inflammatory markers, in the absence of any neurodevelopmental involvement, which separates them from the other phenotypes of the gene [13]. All had failure to thrive, hepatomegaly, gastrointestinal symptoms, marrow dysplasia and haemophagocytic lymphohistiocytosis; three had splenomegaly, three had profound neutropenia, all had anaemia and thrombocytopenia requiring transfusions, and one had inflammatory involvement of the central nervous system with generalised seizures and white and grey matter lesions on magnetic resonance imaging. Three of the four died, at four and a half months, at six to seven months and at one and a half years, and the only survivor reached six years of age after transplantation [13]. The bone marrow showed trilineage dyshaematopoiesis with a predominance of early differentiation elements, a reduced content of clonogenic progenitors with altered clonogenicity and, in one patient, fibrosis with reduced, dysmorphic megakaryocytes and abundant CD68-positive histiocytes with focal or absent phagocytosis [13]. The gastrointestinal symptoms were severe and persistent: recurrent intestinal haemorrhage with chronic diarrhoea, small bowel inflammation, relentless enterocolitis from birth with diarrhoea, haemorrhage and intestinal infarcts, duodenal ulcers and, in one patient, perforation during colonoscopy [13].

The series of four patients with different C-terminal variants shows an overlapping but less lethal phenotype: all presented at birth or in the neonatal period with growth restriction, hepatosplenomegaly, raised transaminases, recurrent febrile episodes, urticarial rash and marked cytopenias, with anaemia in all and transfusion-dependent thrombocytopenia in two; one patient had aseptic meningitis, papilloedema and posterior retinitis, and most had mild facial dysmorphism, with frontal bossing, mild hypertelorism and a depressed nasal bridge. Skin biopsy showed perivascular lymphocytic and neutrophilic infiltration without vasculitis. Two of the four met criteria for haemophagocytic lymphohistiocytosis or macrophage activation [14]. Orphanet adds to the description of this syndrome myelofibrosis, gastrointestinal and upper respiratory tract infections, neurodevelopmental delay and facial dysmorphism with macrocephaly, mild frontal bossing, sparse hair, hypertelorism and a depressed nasal bridge [22].

Takenouchi-Kosaki syndrome is a developmental disorder, not an autoinflammatory disease. The core phenotype combines psychomotor delay with variable intellectual disability, facial dysmorphism and cardiac, genitourinary and haematological or lymphatic defects, including thrombocytopenia and lymphoedema; changes on brain imaging, skeletal anomalies and recurrent infections may be added, and some patients have a milder course reminiscent of Noonan syndrome [12]. The first patient described, of Japanese and Iranian ancestry, had a patent ductus arteriosus corrected in the first year, inguinal hernia, developmental delay with walking only at four years, lymphoedema of the right lower limb and persistent thrombocytopenia from the first year, with platelets of 52,000/µL, leucocytes of 3000/µL and normal haemoglobin at 18 years, enlarged platelets on the blood film, head circumference of −2 standard deviations, height of −1.5 and weight of −1.6, midface hypoplasia, synophrys, mild ptosis, eversion of the lateral portion of the lower eyelid, exotropia, short philtrum, thin upper lip and interphalangeal contractures of the third to fifth fingers [10, 12]. A second patient had, in addition to macrothrombocytopenia, lymphoedema and developmental delay, low visual acuity, congenital nystagmus, bilateral sensorineural hearing loss, retinal dysplasia with congenital falciform detachment, pericardial effusion, hydrothorax, ascites, hypoalbuminaemia and protein-losing enteropathy due to intestinal lymphangiectasia diagnosed at 14 years, with height of −4.3 standard deviations, weight of −3.0 and head circumference of −5.0 at 22 years, and severe intellectual disability [12]. The macrothrombocytopenia may be intermittent and very mild, with platelet counts usually above 100,000/µL, so it may escape detection entirely; without this finding, the presentation overlaps with other syndromic forms of short stature. Immunodeficiency, autoimmunity and neoplasia have been reported in carriers of the recurrent variant, which justifies early diagnosis and organised follow-up rather than treating the condition as a static dysmorphic syndrome [66]. Among disorders of cellular trafficking, this syndrome is presented as a Golgipathy with intellectual disability, epileptic encephalopathy, optic atrophy and lymphoedema, associated with macrothrombocytopenia [67].

3.3 Dipeptidyl peptidase 9 deficiency

The recessive form, named Hatipoglu immunodeficiency syndrome, presents in childhood with a triad of failure to thrive, cutaneous manifestations and pancytopenia with susceptibility to recurrent infections [17]. In the four patients of the original description, aged between 6 and 14 years, there were recurrent fevers and infections, including herpetic infections, otitis media and bronchitis, associated with pancytopenia and anaemia; two had asthma and two severe allergies. The cutaneous manifestations included infantile eczema in two, atopic dermatitis in two, hyper- or hypopigmented macules and petechiae. Three of the four had learning difficulties and speech delay, two autism spectrum disorder, two downslanting palpebral fissures and two hemivertebrae; one had febrile seizures, cryptorchidism, inguinal hernia and feeding difficulty. Three of the four required haematopoietic cell transplantation [16, 17]. The International Union of Immunological Societies classification describes dipeptidyl peptidase 9 deficiency with increased susceptibility to infection, pancytopenia with petechiae, recurrent fever, skin pigmentation changes and poor growth, with normal natural killer cell function [20]. A later report described eight-month-old twins, of non-consanguineous parents, with severe macrocytic anaemia unresponsive to vitamin B12 and recurrent sinopulmonary infections with multiple admissions and transfusions, and with a positive dihydrorhodamine reduction test, initially interpreted as chronic granulomatous disease, until exome sequencing identified the variant in the gene [68].

The de novo dominant form presented much more acutely. A German boy born at term to non-consanguineous parents had, at 12 days of life, bruising and grade 1 intraventricular haemorrhage due to thrombocytopenia of 16×10⁹/L, and at two months trilineage pancytopenia with very high lactate dehydrogenase, a marrow with increased erythropoiesis and megakaryopoiesis and left-shifted granulopoiesis, without signs of neoplasia. Splenomegaly followed and progressed to marked hepatosplenomegaly, with greatly increased transaminases, ferritin, soluble interleukin-2 receptor and triglycerides, hypofibrinogenaemia and fever, in a picture suggestive of haemophagocytic lymphohistiocytosis; immunophenotyping showed a marked reduction in neutrophils and in B, natural killer and T lymphocytes, particularly cytotoxic T cells, but perforin expression, natural killer cell degranulation and HLA-DR expression were normal, which excluded the known primary causes of haemophagocytosis. Serum pro-inflammatory cytokines, including interleukin 6, interleukin 8, tumour necrosis factor alpha, interferon gamma, interleukin 1 beta and interleukin 18, were strongly raised [18].

3.4 NLRP1 and CARD8 diseases

Multiple self-healing palmoplantar carcinoma combines keratopathy with corneal neovascularisation and opacification, palmoplantar hyperkeratosis, dyshidrosis, dystrophic nails and recurrent keratoacanthoma-like lesions in the palmoplantar skin and in the conjunctival and corneal epithelia, with susceptibility to squamous cell carcinoma. In the French family described, the mother presented at 27 years with unilateral keratopathy with neovascularisation and complete corneal opacification, and the son, at two years, with bilateral inferior corneolimbal infiltrates with limbal thickening, neovascularisation and keratin deposits, which progressed centripetally despite topical corticosteroids, with bilateral visual impairment and multiple surgical procedures, accompanied by pruritic hyperkeratotic scars, palmoplantar hyperkeratosis, chronic rhinitis, a hoarse voice, dystrophic nails with thickened nail beds and hypermobility of the finger joints. In a five-generation Tunisian family, the mean age at onset was 8.8 years, with a range of 1 to 25 years, and the lesions consisted of 8 to 12 palmoplantar tumours of 5 to 15 millimetres that grew over three months and regressed spontaneously [54].

Autoinflammation with arthritis and dyskeratosis manifested in the patients described with skin lesions in the first months of life, extensive multiple digitate hyperkeratoses on the limbs, shoulders and flanks, pigmented macules, symmetrical painful plantar keratotic lesions, comedo-like papules and vermiculate atrophic scars on the cheeks, perifollicular keratotic sheaths on the hair, diffuse xerosis, reduced sweating and xerophthalmia with punctate keratitis, accompanied by hypereosinophilia and raised immunoglobulins E and A; arthritis appeared at 12 years in one patient, affecting the wrists, knees and ankles, and at 5 years in a female cousin, with later involvement of the wrists, hands and ankles, and one of the patients had uveitis [55]. Juvenile-onset recurrent respiratory papillomatosis manifested in two brothers with hoarseness and recurrent laryngitis, with glottic and supraglottic papillomas that required eight surgical ablations in one year in the elder brother and laryngeal papillomatosis diagnosed at 20 months in the younger, accompanied by palmoplantar warts, keratosis pilaris on the lower back, buttocks and thighs and atrophoderma vermiculata on the face [56].

3.5 Treated and untreated natural history

The distinction between natural history and treated disease is difficult in these conditions, because practically all published patients received multiple drugs simultaneously. Some data nevertheless allow the two planes to be separated. Without treatment, flares of autoinflammation with infantile enterocolitis may progress to coagulopathy, organ failure and death, and this was exactly the course of the index case of the original description, who died on the twenty-third day of life [3, 21]. In NOCARH syndrome, mortality was three in four in the founding series and remained high in subsequent series despite inflammasome-directed treatments, and the only survivor depended on timely neutralisation of interleukin 1 and interferon gamma followed by transplantation [13, 46]. In the 2018 review of NLRC4 diseases, mortality of the severe phenotype was 25% [60]. Under treatment, the picture changes: with interleukin-1 blockade the patient of the original description was free of fever and clinical flares for seven months and stopped corticosteroids and colchicine, but the cytokine profile remained abnormal and she had transient rises in transaminases, lactate dehydrogenase and ferritin after a viral illness, which shows persistent subclinical inflammation [1]. In the series of CDC42 C-terminal variants, anakinra completely resolved fever and rash, reduced liver and spleen size and improved growth in all, three of the four patients went on to achieve developmental milestones and the cytopenias resolved in three, but the dysmorphism persisted, one patient continued to have intermittent flares and peripheral interleukin 18 fell little or not at all even after months of treatment [14]. Total interleukin 18 remains raised for years in surviving patients, with a stable baseline value that in one patient followed for four years was close to 60,000 pg/mL, with proportionally small rises during flares triggered by influenza vaccination [7].

4. Clinical Recognition and Warning Signs

4.1 When to suspect, by age

In the neonatal period and in the first months of life, the combination of secretory or bloody diarrhoea with fever, hepatosplenomegaly, cytopenias and raised ferritin should raise the possibility of an inflammasomopathy with macrophage activation, even when there is no family history. The association of very-early-onset severe enterocolitis with recurrent episodes of hyperferritinaemia is, in practice, the clinical signature of NLRC4 gain-of-function disease; the picture is all the more suggestive when infectious investigations are repeatedly negative and there is no response to exclusion diets [27, 61]. Trilineage cytopenia from birth, with dyshaematopoiesis in the marrow, painful rash and persistently raised inflammatory markers without neurodevelopmental involvement, points to NOCARH syndrome, and this combination is regarded by the authors of the original description as the sign that should lead to early diagnosis [13]. In a child with pancytopenia, failure to thrive, skin pigmentation changes and recurrent respiratory and herpetic infections, with unrevealing immunophenotyping, dipeptidyl peptidase 9 deficiency should be considered [17, 68].

From three years of age, the phenotype changes. Recurrent episodes of fever with non-pruritic urticarial rash and arthralgia triggered by cold, without splenomegaly or bone erosions, that resolve without treatment, in a family context compatible with dominant transmission, suggest familial cold autoinflammatory syndrome type 4, especially when testing for NLRP3 and NLRP12 variants is negative [4, 65]. Early-onset recurrent panniculitis with fever and arthralgia, with histology showing dermal lymphohistiocytic infiltrates and mixed panniculitis, is another form of presentation to recognise [40].

In the adult, the hypothesis is the same through a different mechanism. A patient who develops in adulthood recurrent episodes of fever, myalgia, arthralgia, lymphadenopathy, abdominal pain and systemic inflammation without an identified cause, especially with a good response to interleukin-1 blockade, should be investigated for somatic mosaicism of inflammasome genes, including NLRC4; conventional sequencing may not detect allele fractions of 2% to 5%, and high-depth sequencing or quantitative methods such as digital polymerase chain reaction must be requested [32, 35, 39]. The published cases had onset between 21 and 60 years of age [32].

4.2 Warning signs and emergencies

Macrophage activation syndrome should be considered in any patient with persistent fever, splenomegaly, raised or rising ferritin, inappropriately low cell counts, abnormal liver tests, intravascular activation of coagulation, or raised or rising triglycerides [8]. None of these findings is specific on its own, and it is the pattern, assessed as a whole and over time, that has value; early recognition of this pattern is what allows aetiological investigation to begin, treatment to be considered and a monitoring strategy to be established before serious complications or death occur [69]. Ferritin should be measured in all patients with new, persistent or increased suspicion of macrophage activation, even when previous measurements were normal, and in patients at risk or with suspected disease it should be repeated daily or every other day [8, 69].

Three situations demand immediate action. The first is shock, which in these diseases may be vasoplegic and mimic sepsis or a severe allergic reaction, and which has been described mainly with variants in the nucleotide-binding domain of NLRC4; the presence of active infection should not delay treatment of macrophage activation, and infectious triggers should be treated in parallel because controlling them facilitates resolution of the inflammatory picture [8, 27]. The second is intestinal haemorrhage or ischaemia, which led to resections and ileocolostomies in patients with NLRC4 and CDC42 disease [13]. The third is central nervous system involvement, which is frequent, insidious and dangerous enough to warrant specific attention: it should be suspected in all patients being assessed for macrophage activation, all should undergo a complete neurological examination, and those presenting under one year of age, those with suspected familial disease and those with symptoms or signs of neurological dysfunction, including an unreliable examination, should be assessed with cerebrospinal fluid analysis and contrast-enhanced brain magnetic resonance imaging [69].

There is a laboratory sign that points directly to this group of diseases. An infant with recurrent macrophage activation and enterocolitis whose serum interleukin 18 is enormously raised should be tested for NLRC4, and the same finding indicates which cytokine to target [42]. Marked elevation of serum interleukin 18, or of S100 proteins such as calprotectin, strongly supports the diagnosis of Still's disease and, by extension, of interleukin-18-mediated disease, so the assay should be requested when available [8].

4.3 Recognition in the patient with haemophagocytosis

In the assessment of a child with a haemophagocytic lymphohistiocytosis phenotype, an inborn error of immunity should be considered, particularly when there is a history of previous, recurrent, chronic or severe infections [70]. The features pointing to a genetic cause are young age at presentation, a positive family history, consanguinity and prominent central nervous system disease; other features point to specific causes, such as albinism, inflammatory bowel disease, isolated central nervous system involvement and immunodeficiency to Epstein-Barr virus [69]. In the inflammasomopathies of this article, the clue is precisely the absence of a cytotoxic defect: normal perforin, natural killer cell degranulation and HLA-DR expression on T lymphocytes in a patient with a haemophagocytosis phenotype should prompt a search for causes of macrophage-intrinsic hyperinflammation, rather than closing the genetic investigation [18, 24].

There is no newborn screening for any of these diseases in any known national programme, and the sources consulted do not describe any biomarker applicable to population screening.

5. Diagnosis: Criteria, Tests and Algorithms

5.1 Available formal criteria and their limits

There are no specific diagnostic criteria for any of the diseases in this article. Diagnosis of the entity is molecular, and the available formal criteria serve to characterise the complication, macrophage activation syndrome or haemophagocytic lymphohistiocytosis, not the underlying disease.

The HLH-2004 criteria remain the most widely used operational reference. The diagnosis is considered established if there is a compatible molecular diagnosis or if five of eight criteria are present: fever above 38.5 °C; splenomegaly; cytopenia in at least two of three lineages in peripheral blood, defined by haemoglobin below 9 g/dL, or below 100 g/L in infants under four weeks, platelets below 100×10⁹/L and neutrophils below 1.0×10⁹/L; hypertriglyceridaemia with fasting triglycerides above 3.0 mmol/L, equivalent to more than 265 mg/dL, and/or hypofibrinogenaemia with fibrinogen of 1.5 g/L or less; haemophagocytosis in the marrow, spleen, liver, lymph nodes or other tissues; low or absent natural killer cell activity; serum ferritin of 500 µg/L or more; and soluble CD25, that is, soluble interleukin-2 receptor, of 2400 U/mL or more. Given the high mortality of untreated active disease, directed treatment is not delayed when a patient does not meet all the criteria [70].

The joint task force of the European and American rheumatology societies concluded that each set of criteria is useful in the context in which it was developed, but that none is a sufficient diagnostic instrument in all settings and populations: the HLH-94 and HLH-2004 criteria were developed to classify infants and children with genetic causes in treatment trials, the MAS-2016 criteria to classify macrophage activation in patients with known or strongly suspected systemic juvenile idiopathic arthritis, the HScore in adults with haemophagocytosis associated mainly with neoplasia or infection, and the MS score to distinguish macrophage activation from underlying systemic juvenile idiopathic arthritis [69]. In the diseases of this article, two of the HLH-2004 criteria are particularly fallible: natural killer cell activity, which is typically normal in the inflammasomopathies, and tissue haemophagocytosis, which may not be detected even in patients with the full picture, as happened in patients of the NOCARH series [13, 24].

5.2 Interleukin 18 as a discriminating biomarker

Measurement of interleukin 18 is the laboratory test that best distinguishes this group of diseases. In a comparison of samples from active systemic juvenile idiopathic arthritis, macrophage activation, infection-associated haemophagocytosis and familial haemophagocytic lymphohistiocytosis, total interleukin 18 was about 500 times above controls in macrophage activation, whereas the binding protein and CXCL9 were raised but varied little between groups; a total interleukin 18 value above 24,000 pg/mL distinguished macrophage activation from familial haemophagocytic lymphohistiocytosis with 83% sensitivity and 94% specificity, and the ratio of total interleukin 18 to CXCL9, with an optimal cut-off of 2.3, slightly improved this distinction. In a cohort of genetically or clinically defined autoinflammatory diseases, most patients referred for active or previous macrophage activation had chronically raised total interleukin 18, typically more than 40 times normal, whereas in the NLRP3 and MEFV inflammasomopathies and in diseases mediated by interleukin 1, interferon and nuclear factor kappa B the elevation was modest and inconsistent; a value above 11,600 pg/mL separated samples from systemic juvenile idiopathic arthritis, adult-onset Still's disease and macrophage activation from all others with 88% sensitivity and 93% specificity. Values close to those of macrophage activation were never detected in patients carrying NLRP3 or MEFV variants, and in malignancy-associated haemophagocytosis total interleukin 18 and the binding protein were both very high, resulting in minimal free interleukin 18 [7].

A systematic review of 14 studies, with searches up to January 2020, concluded that serum interleukin 18 is raised in primary and secondary haemophagocytic lymphohistiocytosis, generally above 1000 pg/mL, with values consistently above 10,000 pg/mL in macrophage activation compared with other subtypes; the ability to distinguish macrophage activation from systemic juvenile idiopathic arthritis is less clear, because values above 100,000 pg/mL have been described in patients with this arthritis with and without macrophage activation, and the studies are small [71]. In the European ImmunAID cohort, with 53 patients with monogenic autoinflammatory diseases, 220 without a genetic diagnosis and 49 non-inflammatory controls, all with active disease, total and free interleukin 18 were significantly increased in Still's disease relative to most other diseases, total interleukin 18 was also raised in familial Mediterranean fever to a comparable degree, and only the free fraction was selectively high in Still's disease, with an area under the curve of 0.91 for total interleukin 18; the free fraction fell after treatment in the ten patients with a second sample [72].

Measurement of the free fraction therefore has interest of its own. Historically difficult, it became possible with assays using the recombinant binding protein as the capture protein, and an immunoassay using binding protein purified by luciferase technology has recently been described: in a series of 13 patients with systemic juvenile idiopathic arthritis and 11 controls, the free fraction was raised in 15% of patients in the active phase and in all in the macrophage activation phase, normalising in the inactive phase, and it performed better than interleukin 6 and ferritin in identifying macrophage activation [73]. In NLRC4 disease without macrophage activation, total and free interleukin 18 are raised without correlation with markers of inflammation, which means that the presence of a detectable free fraction is not enough to predict the complication [29].

There are two important caveats. The first is that interleukin 18 remains chronically raised long after clinical improvement, so it does not serve as a short-term activity marker: in one patient with NLRC4 disease a stable baseline value close to 60,000 pg/mL was observed over four years, and in an adult with Still's disease the peak value of 110,000 pg/mL improved only months after C-reactive protein, ferritin and the free fraction had normalised [7]. The second is that availability of the assay is limited and its performance depends on the method used, so published cut-off values are not transferable between laboratories without local validation; the comparability of values between studies has itself been flagged as an unresolved problem in the literature [42, 71, 74].

5.3 Other laboratory and imaging tests

Ferritin is the most sensitive test for macrophage activation. A cut-off of 500 ng/mL had 84% sensitivity in paediatric patients, served as the threshold in the Histiocyte Society trials and is part of all sets of criteria, but it has poor specificity in other contexts; the values used in paediatric studies lie between 500 and 2000 ng/mL and in adult studies are often above 10,000 ng/mL, where infectious and neoplastic contributors predominate, and conditions such as iron overload, neoplasia and hepatitis raise ferritin even without macrophage activation [69]. In a cohort of 113 patients with haemophagocytosis, a threshold above 2000 ng/mL had 70% sensitivity and 68% specificity [70].

The more specific biomarkers of the pathways involved are increasingly accessible in reference laboratories and should be used with the support of consultants: measures of T-lymphocyte activation, such as soluble interleukin-2 receptor alpha, that is, soluble CD25, and HLA-DR expression on T lymphocytes; of macrophage activation, such as CD163 and neopterin; of the inflammasome pathway, with interleukin 18; and of the interferon-gamma pathway, with interferon gamma itself and CXCL9 [69]. HLA-DR expression on 14.4% of CD8-positive T lymphocytes and on 4.8% of CD4-positive cells had more than 80% sensitivity and 90% specificity for detecting active haemophagocytosis, and the presence of more than 7% CD38-high HLA-DR-positive cells among CD8-positive cells had strong positive and negative predictive value for distinguishing active disease from early sepsis or from healthy controls [70]. These populations expand from the onset of Still's disease and peak in macrophage activation, with medians of 4.61% and 31.7% against 0.87% in healthy donors, and correlate with serum interleukin 18, which reinforces their usefulness as a marker of the axis [75].

Tests assessing cytotoxicity serve mainly to exclude the familial forms. Perforin expression is absent or greatly reduced in more than 90% of patients with familial haemophagocytic lymphohistiocytosis due to PRF1 variants, and the CD107a degranulation assay has 96% sensitivity and 88% specificity for the UNC13D, STX11 and STXBP2 forms; in the inflammasomopathies of this article, these assays are normal, which constitutes positive information and not a useless result [18, 70].

In NLRC4 disease, other tests complete the assessment: full blood count, fibrinogen, triglycerides, transaminases, lactate dehydrogenase, coagulation studies, C-reactive protein and serum amyloid A protein for inflammatory activity, bone marrow aspirate for haemophagocytosis and upper and lower endoscopy with intestinal biopsy for the enterocolitis [42]. Duodenal biopsy may show villous atrophy with intraepithelial lymphocytes and rectal biopsy an eosinophilic infiltrate with moderate interleukin-18 staining in the surface epithelium and crypts [3, 38]. In the assessment of suspected CDC42 disease, full blood count with mean platelet volume and a blood film to look for large platelets, bone marrow aspirate when there is cytopenia or dyshaematopoiesis, immunological assessment, echocardiography, brain magnetic resonance imaging and measurement of growth and head circumference are added [66]. In patients with suspected central nervous system involvement, brain magnetic resonance imaging most often shows diffuse or focal white matter changes in the cerebrum and cerebellum, and haemorrhage, nodular lesions, masses, generalised atrophy and oedema may be seen; a normal scan does not exclude involvement, especially when cerebrospinal fluid protein is raised and mononuclear cells are increased [70].

5.4 Genetic tests, limits and report content

Given the high prevalence of genetic causes in children and the clinical impact of a positive finding, genetic testing should be done early in children and in high-risk adults, preferably with multigene panels or exome or genome sequencing; single-gene sequencing remains appropriate when there is a family history of a known genetic disease, characteristic clinical features such as albinism, positive protein or functional tests, or in resource-limited settings. Genetic counselling is indicated for all patients undergoing genetic testing [69].

The limits of the tests should be made explicit. A negative panel does not exclude these diseases: in the first patients described, targeted testing of haemophagocytosis and periodic fever genes was negative and only the exome revealed the diagnosis, and in the patient with dominant dipeptidyl peptidase 9 deficiency a panel covering marrow failure, immunodeficiency, autoinflammatory syndromes and primary haemophagocytosis did not find the cause [1, 18]. Exome and genome sequencing, in turn, do not reliably detect mosaic variants with a low allele fraction, which in this class of diseases may fall to 2%, so suspected mosaicism requires targeted high-depth sequencing, digital polymerase chain reaction or analysis of separated cell subpopulations; the fraction may be greater in monocytes than in whole blood, and may be nil in non-haematopoietic tissues such as hair or detectable in urine and fibroblasts, information that helps to date the post-zygotic event [32, 39]. Detection of small structural variants, such as the 93-base-pair in-frame deletion described in the leucine-rich repeat domain of NLRC4, depends on the panel's bioinformatic analysis covering medium-sized indels [28].

The genetic report should contain the gene, the reference transcript with version, the HGVS nomenclature at cDNA and protein level, zygosity, the allele fraction when mosaicism is suspected and the classification according to the criteria of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. For the diseases of this article, the transcripts to use are NM_001199138.2 for NLRC4, NM_001791.4 for CDC42 and NM_139159.5 for DPP9 [26]. There are three documented nomenclature pitfalls that the report should avoid. The familial cold autoinflammatory syndrome type 4 variant appears in the literature, including in reviews, as c.1589A>C, a notation that does not correspond to the current reference transcript, on which the p.(His443Pro) change is written c.1328A>C [4, 26, 27]. The Hatipoglu syndrome missense variant appears as c.449G>A, but on the current reference transcript p.(Gly167Ser) is written c.499G>A, with the designation recorded in ClinVar as NM_139159.5(DPP9):c.499G>A, conflicting classifications and association with Hatipoglu syndrome [16, 53]. The dominant DPP9 variant is described on isoform 1 as c.755G>C, p.(Arg252Pro), and corresponds, in the long-isoform numbering used by the reference transcript, to c.842G>C, p.(Arg281Pro) [18, 26]. The articles' notation for the CDC42 stop-codon extension, “p.192C24”, is correctly written p.(Ter192CysextTer24) [26].

Interpretation of a variant of uncertain significance in this group of genes requires functional or biochemical support. In the case of NLRC4, the proportion of variants of uncertain significance is very high and a variant should be supported by the characteristic interleukin-18 signature or by demonstration of constitutive inflammasome activation before being considered causal [42]. The functional assays available in research include demonstration of spontaneous ASC speck formation, constitutive caspase-1 cleavage and increased interleukin-18 production in the patient's macrophages or in transduced cells, and, in the case of CDC42, assessment of the subcellular localisation of the protein and of the response to pyrin inflammasome activators [1, 44]. Family segregation has demonstrative value, as became clear in the family in which the p.(Val341Ala) variant arose de novo in the father and was transmitted to both affected children [3].

5.5 Practical algorithm

In the infant with severe enterocolitis, recurrent fever and hyperferritinaemia, the recommended sequence begins with active exclusion of infection, with blood cultures and targeted viral testing, and exclusion of neoplasia, bearing in mind that administration of corticosteroids may obscure the histopathological diagnosis of leukaemia or lymphoma and that, where possible, definitive investigation by biopsy or aspirate should be attempted beforehand [69]. In parallel, ferritin, full blood count, fibrinogen, triglycerides, transaminases, lactate dehydrogenase, soluble interleukin-2 receptor and, when available, total and free interleukin 18 and CXCL9 are requested, and cytotoxicity is assessed with perforin expression and a degranulation assay. Genetic testing by broad panel or exome, as a trio when possible, is requested immediately and not after the investigation has been exhausted, because the result changes management and prognosis. Endoscopy with biopsies documents the enterocolitis. Assessment of the central nervous system follows the indications already described.

In the adult with uncharacterised late-onset autoinflammatory disease, the sequence is the same, but the genetic investigation should explicitly include testing for mosaicism, and the yield of actionable findings is low, so the decision should weigh the context; a clear response to interleukin-1 blockade strengthens the indication [35, 69].

Prenatal diagnosis is technically possible when the familial variant is known, by DNA analysis of chorionic villi or amniocytes, and preimplantation genetic testing is applicable under the same conditions. In the diseases of this article, however, the overwhelming majority of cases result from de novo variants, so these options are of interest mainly to families with documented transmission, such as those described for the NLRC4 variants p.(Val341Ala), p.(His443Pro) and p.(Ser445Pro), and to families with the recessive DPP9 form [3, 4, 16, 40].

6. Differential Diagnosis

In practice, the differential is organised around three questions. The first is whether the patient has a primary cytotoxicity defect, and the perforin and degranulation assays answer this with good sensitivity and specificity; when they are normal in a patient with a haemophagocytosis phenotype, the diagnosis shifts towards macrophage-intrinsic hyperinflammation, which is the territory of this article [24, 70]. The second is whether the picture is a classical inflammasomopathy or an inflammasomopathy with a risk of macrophage activation, and here measurement of interleukin 18 is discriminating, because values close to those of macrophage activation have never been observed in patients with NLRP3 or MEFV variants [7]. The third is whether the picture is monogenic or corresponds to Still's disease, a question that the European recommendations resolve by active consideration of monogenic autoinflammatory diseases in the differential of any patient with suspected Still's disease, with level of evidence 5 and grade D but 83% agreement among the experts [8].

Two differentials deserve a comment of their own. NLRC4 disease shares with familial haemophagocytic lymphohistiocytosis splenomegaly, fever, gastrointestinal haemorrhage and rash at an early age, and interleukin 18 is raised in both, which complicates the distinction; the key lies in the degree of elevation, the pattern of chronic persistence and the normality of cytotoxicity assays [7, 27]. XIAP deficiency is the closest differential in pathophysiology, because it is also an inflammasomopathy with interleukin 18 that may remain raised between episodes of haemophagocytosis, with severe colitis and recurrent haemophagocytosis without Epstein-Barr virus; the distinction is molecular and by flow cytometry of protein expression, and its treatment, which includes the particularity of avoiding myeloablative conditioning for transplantation, follows the protocols for primary haemophagocytic lymphohistiocytosis [76, 77].

7. Treatments and Risk Management Strategies

7.1 Principles and regulatory framework

No drug has an approved indication, in the United States or in the European Union, for NLRC4-associated disease, NOCARH syndrome or dipeptidyl peptidase 9 deficiency. All the pharmacological options described below are used outside the authorised conditions of their respective labels, and this circumstance should be made explicit to the patient and family and recorded in the medical notes [42, 66]. The available evidence comes from case reports and very small case series, from one trial with results posted on the registry and from a phase 2 trial still without results, with the additional limitation that most published patients received several immunomodulatory drugs simultaneously, which makes it difficult to attribute improvement to a single intervention [24].

In the absence of a specific guideline, the reference for management is the joint 2024 recommendations of the European Alliance of Associations for Rheumatology and the Paediatric Rheumatology European Society for Still's disease, and the joint 2022 points to consider of the European Alliance of Associations for Rheumatology and the American College of Rheumatology for the early stages of diagnosis and management of suspected haemophagocytosis and macrophage activation; their application to these monogenic diseases is an extrapolation, acknowledged as such [8, 69]. The European recommendations establish, as a general principle, that macrophage activation should be detected promptly and treated rapidly, with 100% agreement and a mean score of 10 out of 10 among the experts, and that its treatment must include high-dose glucocorticoids, to which anakinra, ciclosporin and/or an interferon-gamma inhibitor should be considered as add-on initial therapy, with level of evidence 2b, grade B and 100% agreement [8].

7.2 Treatment of the acute episode of macrophage activation

Treatment begins with a search for triggers, most often infectious, which should be treated aggressively with appropriate antimicrobials, without the presence of active infection delaying the start of treatment for macrophage activation [8]. Supportive care should follow accepted age-appropriate standards, such as those of the Surviving Sepsis Campaign, and may require intensive care: more than one third of children and about half of adults with macrophage activation required admission to an intensive care unit, and among those admitted almost 70% required mechanical ventilation or vasopressors and inotropes and almost half required kidney replacement therapy [69].

Glucocorticoids are the mainstay. In the European recommendations for macrophage activation, they are usually given as intravenous methylprednisolone pulses of 15 to 30 mg/kg/day, with a maximum dose of 1 g per infusion, and dexamethasone should be considered when there is central nervous system involvement because it crosses the blood–brain barrier better [8]. The 2022 points to consider indicate intravenous methylprednisolone pulses of 10 to 30 mg/kg/day up to 1 g, given daily, and note that dexamethasone is used in haemophagocytosis protocols at a starting dose of 10 mg/m²/day, equivalent to about 2 to 4 mg/kg/day of methylprednisolone, with the caveat that its long half-life may be a disadvantage in rapidly evolving diagnostic scenarios [69]. High-dose glucocorticoids can achieve satisfactory results in a substantial number of patients, especially if started early [8]. Dose- and duration-dependent hyperglycaemia, hypertension, myopathy and psychosis should be monitored [69].

The recombinant interleukin-1 receptor antagonist, anakinra, is the second most used drug. In the European recommendations, for macrophage activation it should be used at doses higher than the usual 1 to 2 mg/kg/day and, where appropriate, by repeated intravenous administration, an approach that has produced satisfactory responses in several patients, although it has never been tested in a clinical trial for this indication [8]. The 2022 points to consider support its empirical use in early, evolving or undifferentiated haemophagocytosis, because it offers immunomodulation without significant immunosuppression and without compromising the investigation of neoplasia, with a rapid onset of action and short half-life that are desirable in unstable patients; even at doses up to 48 mg/kg/day in adults with bacterial sepsis there was no signal of immunosuppression, and a retrospective study in secondary haemophagocytosis supported the safety and possible efficacy of its early use [69]. In patients with NLRC4 disease, the doses reported were 10 mg/kg/day and up to 15 mg/kg/day [6, 38].

Ciclosporin, or another calcineurin inhibitor such as tacrolimus, should be considered in the event of an inadequate response to glucocorticoids. It was shown to be ineffective in primary familial haemophagocytic lymphohistiocytosis and abandoned in that context, but experience in macrophage activation is extensive and often positive; it can be given orally or intravenously, the latter mainly in intensive care, and it is of particular value in lower-resource countries [8]. High-dose intravenous immunoglobulin is an empirical immunomodulatory option that does not obstruct the investigation of neoplasia or suppress immune function, with the caveat that it constitutes a substantial colloid load that may compromise cardiac function and worsen oedema, and that it rarely causes haemolysis or aseptic meningitis; samples for serology should be taken before it is given [69].

Emapalumab, a monoclonal antibody that neutralises interferon gamma, is the only targeted therapy tested in a clinical trial in the macrophage activation of Still's disease, and in patients with severe disease who had failed high-dose glucocorticoids it was associated with remission in the great majority, with a marked corticosteroid-sparing effect and a reassuring safety profile [8]. In NOCARH syndrome, it was the drug that resolved the refractory episode in the only surviving patient of the founding series [13]. Low-dose etoposide may be considered in refractory macrophage activation; the high doses characteristic of the protocols for primary haemophagocytosis and severe Epstein-Barr virus-associated haemophagocytosis have less favourable evidence in the context of Still's disease [8, 69]. The Janus kinase inhibitors, ruxolitinib and baricitinib, have reports of efficacy in macrophage activation and are an escalation alternative in consultation with experts [8, 69].

The recommendations insist on an operational point: the above therapies should be added in patients with an unsatisfactory initial response to high-dose glucocorticoids and in patients with severe, rapidly worsening macrophage activation, and combinations of several agents on a background of high-dose glucocorticoids are often necessary and should also be considered as initial therapy; the higher mortality described in adults should weigh in this choice, and the decision should be discussed with a reference centre [8].

7.3 Gene- and variant-directed treatment

In NLRC4-associated disease, the current strategy targets the downstream cytokines. Interleukin-1 blockade is usually the first biologic and frequently controls fever and the systemic inflammatory response, but it is often insufficient in interleukin-18-dominated disease, because it does not correct the excess of this cytokine, which is the driver of macrophage activation; in the 2018 systematic review, interleukin-1 inhibitors were effective in 83% of the cases then published, which reflects the numerical weight of the mild phenotypes [24, 42, 60]. Neutralisation of interleukin 18 with recombinant human binding protein has shown striking clinical efficacy in life-threatening macrophage activation, with rapid stabilisation of systemic inflammation and sustained control [6, 24]. Inhibition of the mammalian target of rapamycin, with rapamycin, has shown benefit as an adjuvant strategy by reducing caspase-1 activation and inflammasome-dependent cytokine secretion in the patient's cells [24, 38]. Tumour necrosis factor blockade has been used empirically, especially in severe intestinal inflammation, with variable responses and an indirect mechanistic rationale, with reports of clinical remission with infliximab and good results with adalimumab, and one patient with the cold-induced phenotype well controlled with etanercept [24, 29–31]. Interferon-gamma blockade and faecal microbiota transplantation have shown benefit in selected cases [24, 78]. In a knock-in mouse model with the p.(Val341Ala) variant, still at the preclinical stage, the efficacy of interleukin-18 and tumour necrosis factor blockade was confirmed, and the pups were observed to develop profound hypoglycaemia, whose correction with glucose supplementation attenuated severity and improved survival, which suggests a metabolic component not yet explored in the clinic [24].

In CDC42 autoinflammatory disease, management depends on the substituted amino acid, and this is the most important clinical consequence of the functional characterisation of the variants. The p.(Arg186Cys) variant predicts neonatal pancytopenia and severe autoinflammation or haemophagocytosis requiring interferon-gamma neutralisation and transplantation, whereas the p.(Cys188Tyr) and p.(Ter192CysextTer24) variants predict autoinflammation responsive to interleukin-1 blockade [15, 66]. In the series of four patients with C-terminal variants, anakinra completely resolved fever and rash, reduced liver and spleen size and improved growth in all, with resolution of the cytopenias in three of the four, but with no appreciable effect on peripheral interleukin 18 [14]. In the p.(Arg186Cys) series, interleukin-1 inhibitors partially improved fever and rash with no effect on the cytopenias, chronic high doses of glucocorticoids were required and three of the four patients died; the survivor resolved the last episode, refractory to repeated glucocorticoid pulses, ciclosporin A and high-dose anakinra at 10 mg/kg/day, with emapalumab, and subsequently underwent transplantation [13]. The demonstration that Golgi retention activates the pyrin inflammasome, with suppression of excessive cytokine production by pyrin inhibitors including colchicine in cellular assays, provides the rationale for exploring colchicine in these patients, but there are no clinical data to support it [44]. The demonstration of hyperactivation of the STING and type I interferon pathway in Golgi-retained variants provides the rationale for Janus kinase inhibitors, and one patient with NOCARH syndrome achieved complete remission with ruxolitinib [45, 46]. In the new p.(Met45Leu) condition, interleukin-1 blockade was therapeutically effective [47]. In Takenouchi-Kosaki syndrome, management is supportive and organ-directed: developmental and educational support, assessment and treatment of congenital heart defects, monitoring of platelet count and bleeding and, because immunodeficiency, autoimmunity and neoplasia have been reported in carriers of the recurrent variant, immunological and oncological surveillance during follow-up [66].

In dipeptidyl peptidase 9 deficiency, the published treatment is haematopoietic cell transplantation, which was required in three of the four patients of the original description and was attempted in the patient with the dominant form, with death on day six from staphylococcal sepsis, parainfluenza reactivation and massive capillary leak during reduced-toxicity conditioning with treosulfan [16, 18]. In the same patient, anakinra at 2 to 4 mg/kg/day for one week was ineffective, methylprednisolone at 2 mg/kg/day and intravenous immunoglobulin at 1 g/kg/day produced no improvement, and ruxolitinib at 0.5 mg/kg/day, added to prednisolone because of a strong interferon signature, produced a transient rise in platelets and neutrophils and a fall in ferritin and triglycerides before a flare triggered by parainfluenza [18]. In an animal model, the observation that deletion of a single allele of the interleukin-1 receptor was enough to produce an almost normal adult mouse led the authors to propose that even partial inhibition of this pathway, with anakinra for example, could bring some therapeutic benefit, a hypothesis not yet tested in patients; interleukin-18 inhibition did not rescue lethality in the model, which suggests that in this disease interleukin 1 weighs more than interleukin 18 [16]. The demonstration that pyroptosis of human progenitor cells depends on the CARD8 inflammasome and caspase-1, and is rescued by their deletion, points to these targets as a future strategy, at a strictly preclinical stage [19].

7.4 Drugs: dosing and safety

The doses given below follow the rule of source hierarchy: when there is a guideline-recommended dose for the condition or for the complication, that is the reference; the label dose is cited alongside, for comparison, with an indication of how it differs.

7.5 Comparative regulatory status

In Brazil, the open medicines data of the National Health Surveillance Agency (ANVISA), consulted on 31 October 2026, record canakinumab (canaquinumabe in the Brazilian name; Ilaris®, Novartis Biociências) as a biological medicine with active registration since 1 March 2010 and anakinra (Kineret®, Pint Pharma) as a biological medicine with active registration since 27 July 2026; emapalumab and tadekinig alfa do not appear in these data [92]. The Brazilian professional package insert for Ilaris®, approved by ANVISA on 26 November 2025, restricts the indication, from 2 years of age, to the cryopyrin-associated periodic syndromes (including NLRP3-associated familial cold autoinflammatory syndrome), tumour necrosis factor receptor-associated periodic syndrome, hyperimmunoglobulinaemia D syndrome with mevalonate kinase deficiency, familial Mediterranean fever with contraindication to, intolerance of or inadequate response to colchicine, and Still's disease, and does not include the gout that appears on the European label [85, 93]. In Still's disease, the dose in the Brazilian package insert is 4 mg/kg every four weeks subcutaneously, up to a maximum of 300 mg, from 7.5 kg; in the cryopyrin-associated syndromes, the starting dose is 150 mg above 40 kg, 2 mg/kg between 15 and 40 kg and 4 mg/kg between 7.5 and less than 15 kg (and from 2 to 4 years), every 8 weeks, with stepwise intensification after assessment on day seven up to 600 mg or 8 mg/kg; in the other periodic syndromes, 150 mg above 40 kg or 2 mg/kg, every four weeks, with intensification up to 300 mg or 4 mg/kg [93]. The same package insert warns that macrophage activation may occur in Still's disease, requires early assessment and treatment when it is suspected and states that, in clinical experience, canakinumab does not appear to increase its incidence in systemic juvenile idiopathic arthritis, without a definitive conclusion [93]. None of the Brazilian indications covers NLRC4, CDC42 or DPP9 diseases, so the use of canakinumab and anakinra in these diseases is, in Brazil as in Europe, off-label. The Brazilian package insert for Kineret® had not been published by the registration holder at the date of consultation, and emapalumab has no Brazilian registration in the data consulted.

7.6 Clinical trials and the limits of their interpretation

These data must be interpreted with caution. The only trial with posted results is small and the primary outcome was based on five patients per arm, which makes the confidence interval very wide and prevents estimation of the median in the active arm; moreover, the registry does not break down results by disease, so it is not possible to separate the effect in patients with NLRC4 variants from that in patients with XIAP deficiency, nor to know the dose given [82]. Reports of response to a drug in patients simultaneously receiving high-dose corticosteroids, calcineurin inhibitors and other biologics do not allow the improvement to be attributed to that drug, and the authors of the first case treated with interleukin-18 binding protein themselves acknowledge that the effect may have been mediated by improved efficacy or pharmacodynamics of the other drugs, namely through reduced intestinal protein losses [6]. It is not legitimate to compare response percentages indirectly between trials with different populations and outcome definitions. Finally, the absence of published negative trials in this field does not mean an absence of failures: the literature explicitly describes minimal response to interleukin-1 blockade in patients with severe NLRC4 and CDC42 disease, no effect of anakinra on the cytopenias in NOCARH syndrome and ineffectiveness of the same drug in dominant dipeptidyl peptidase 9 deficiency [6, 13, 18].

7.7 Developments in the last twelve months

The developments of the period between October 2025 and October 2026 are mainly mechanistic and exploratory, and none has changed the authorisation of any medicine for these diseases. The 2026 review of the NLRC4 inflammasome compiled 28 pathogenic variants in 56 people and presented data from a knock-in mouse model with preclinical validation of interleukin-18 and tumour necrosis factor blockade, with the additional finding of profound hypoglycaemia whose correction improved survival [24]. In 2026 the demonstration was published, in a humanised mouse, that the pancytopenia of dipeptidyl peptidase 9 deficiency depends on the CARD8 inflammasome and caspase-1 in human haematopoietic progenitor cells, a preclinical result that identifies new therapeutic targets [19]. In August 2026 the CDC42 p.(Met45Leu) autoinflammatory condition was described, with demonstration of the mechanism of pyrin activation and efficacy of interleukin-1 blockade in the endotype studied, accompanied by an editorial commentary [47, 48]. In 2026 the characterisation of free interleukin 18 in patients with NLRC4 disease without macrophage activation was published, with the conclusion that this cytokine is necessary but not sufficient for the complication [29]. Two 2026 case reports broadened the mutational spectrum of infantile enterocolitis, one with remission on infliximab [31, 62]. In February 2026 a practical immunoassay for the free fraction of the cytokine was described [73], and in September 2026 the demonstration was published that sustained interleukin 18 expands in vitro the population of CD38-positive HLA-DR-positive T lymphocytes that expands in macrophage activation [75]. In drug development, the phase 2 trial of the bispecific antibody against interleukin 1 beta and interleukin 18 in interleukin-18-mediated monogenic diseases reached its primary completion date in October 2025 and still has no posted results, while the same drug entered phase 2 trials in Still's disease, in November 2025, and in VEXAS syndrome, in October 2026, which indicates continuity of the programme [84, 99].

7.8 Agents to avoid, special situations and transition

Live vaccines should be avoided during immunosuppressive and biological treatment, as should unnecessary exposure to infection, paracetamol in liver failure, non-steroidal anti-inflammatory drugs in thrombocytopenia, building sites or soil handling in patients with neutropenia, because of the risk of fungal infection, and transfusion of non-irradiated blood products in patients on chemoimmunotherapy or transplantation, according to the standards of each institution [70]. The patient treated with emapalumab merits additional attention, because it neutralises interferon gamma: screening for latent tuberculosis before starting is mandatory, as is monitoring for herpes zoster, adenovirus, Epstein-Barr virus and cytomegalovirus during treatment, with consideration of herpes zoster, Pneumocystis jirovecii and antifungal prophylaxis [81]. Antifungal and Pneumocystis jirovecii prophylaxis is generally recommended in patients on haemophagocytosis-directed treatment, and consultation with an infectious diseases specialist experienced in immunocompromised patients helps with planning [69].

The patient with the cold-induced phenotype should be advised to avoid the exposure that triggers the flares, since cold reproducibly worsened autoinflammation both in patients and in the corresponding murine model [4, 5].

There are no specific data on pregnancy and lactation for these diseases. Under the general principles of haemophagocytosis treatment, when a pregnant woman is diagnosed, treatment should be adapted and, if possible, incorporate non-cytotoxic drugs such as corticosteroids; etoposide, because it interferes with mitosis and inhibits DNA synthesis, has rarely been used in pregnancy, with reports of newborns exposed in the second and third trimesters with a normal course and others with anaemia, thrombocytopenia, leucopenia and neutropenia, and underlying triggers such as infection, neoplasia and autoimmune disease should be sought and treated [70]. For emapalumab there are no data in pregnant women; in an animal reproduction study, a murine surrogate antibody crossed the placenta without demonstrated fetal harm, and there are no data on passage into milk [81]. For anakinra, the European product information sets no dose adjustments in the elderly and has no data below 8 months of age, and the graduated syringe is not suitable for patients weighing less than 10 kg, a relevant practical limitation in a disease of neonatal presentation [79].

In the perioperative setting, published experience shows that abdominal surgery in these patients takes place in a context of active inflammation, with intestinal ischaemia and necrosis, perforations and the need for resection and stoma formation, and that surgery may be the moment at which the patient decompensates [13]. Inflammation should be controlled before any elective procedure and the patient should be operated on in a centre with paediatric intensive care. In haematopoietic cell transplantation, the experience available in these diseases points in the same direction as what is known for haemophagocytosis: remission of the hyperinflammatory state before conditioning is decisive, and in the only surviving patient of the NOCARH syndrome series emapalumab was continued until day +28 and anakinra until day +100 after transplantation precisely to prevent flares [13].

Transition to adult care should be planned in patients who reach adolescence. In survivors of NLRC4 disease, intestinal inflammation attenuates with age but the systemic manifestations may persist, and some patients are diagnosed in adulthood, so follow-up should be ensured in adult rheumatology or clinical immunology, with a link to a reference centre of the relevant European network [8, 24]. Patients who are difficult to treat, with severe macrophage activation or with lung disease, should be managed in collaboration with specialised centres, a recommendation that the European guidelines make with 96% agreement [8]. The European transition recommendations for inborn errors of immunity, including autoinflammatory diseases, obtained by the Delphi method in ERN RITA (93 responses from 68 centres in the first round and 66 in the second; consensus defined as agreement of 80% or more, with level of evidence 4–5 and grade D for all recommendations), propose that discussion and planning of transition begin in the paediatric centre before 16 years of age, without a time limit and adapted to the patient's development, with a period of joint paediatric and adult follow-up in the year before transfer, diagnostic reassessment in patients without a definitive diagnosis and review of the treatment plan before leaving paediatrics [100]. The programme should include education on the nature and genetic basis of the disease, inheritance with genetic counselling and family planning, treatment options and their effects, management of crises and comorbidities, assessment of treatment adherence and access to psychological support, and the paediatric and adult teams should actively monitor complications such as organ damage and chronic graft-versus-host disease in transplanted patients [100]. Transfer should be postponed during critical illness due to severe infection or inflammatory manifestations, until the disease has stabilised, when the adult service is not prepared for severe progression and frequent instability, and when the patient is taking part in a clinical trial whose follow-up cannot continue at the adult centre [100].

8. Reproductive Options and Family Counselling

8.1 Recurrence risks by gene and by scenario

In NLRC4 gain-of-function diseases, inheritance is autosomal dominant. When one parent is a carrier, the risk of transmission in each pregnancy is 50%, with the caveat that expressivity is highly variable: in the founding family, the adult carrier had recurrent fevers throughout life and a single severe episode at 43 years, whereas one of the children died in the neonatal period, so counselling should not promise a course similar to that of the parent [3]. When the variant arises de novo in the proband, the risk for siblings is low but not zero, owing to germline mosaicism in one of the parents, and the parents should be tested for the specific variant; the clearest demonstration of this scenario is the founding family itself, in which the variant arose de novo in the father and was transmitted to two affected children [3]. In families with the cold-induced phenotype, vertical transmission is well documented, with five affected individuals in a three-generation Japanese family and six consecutive affected generations in a Spanish family, which indicates apparently high penetrance for this mild phenotype [4, 40]. The recessive NLRC4 form, described in a homozygous woman, implies a 25% risk for siblings and identifies the parents as obligate carriers, with the particularity that heterozygosity confers an increased risk of ulcerative colitis, information that should be given to the family in due proportion, since it is an odds ratio of 2.546 in a common disease and not a Mendelian risk [33].

In autoinflammatory diseases due to C-terminal CDC42 variants, all published cases resulted from de novo variants, without consanguinity, and no vertical transmission has been described; the risk for siblings is therefore the low empirical risk associated with germline mosaicism, and the risk of transmission by the proband, should they reach reproductive age, would be 50% [13, 14]. Takenouchi-Kosaki syndrome follows the same dominant pattern with variants that are, as a rule, de novo [12].

In dipeptidyl peptidase 9 deficiency, there are two distinct scenarios. Hatipoglu syndrome is autosomal recessive, with a 25% risk for each sibling of a proband, a 50% probability that unaffected siblings are carriers and an increased risk in consanguineous families, such as the Bedouin family in which two cousins were affected [16, 17]. The hyperinflammatory form due to a dominant-negative variant arose de novo, with non-carrier parents, and implies a low recurrence risk for siblings [18]. This duality makes it necessary to determine the mechanism before counselling: a report identifying two variants in trans implies recessive counselling, whereas a de novo heterozygous variant with dominant-negative behaviour implies dominant counselling.

In NLRP1 diseases, the OMIM entry for autoinflammation with arthritis and dyskeratosis records dominant and recessive inheritance, so counselling depends on the genotype observed in the family; multiple self-healing palmoplantar carcinoma is dominant and juvenile-onset recurrent respiratory papillomatosis is recessive and flagged as provisional [54–56].

8.2 Cascade testing and assessment of at-risk relatives

Testing for the familial variant should be offered to the parents of any proband with an apparently de novo variant, to establish its origin and quantify the recurrence risk, and to at-risk relatives when transmission is documented. In families with severe disease, testing siblings allows diagnosis before symptoms appear, and in familial haemophagocytosis this early diagnosis speeds up transplantation and allows presymptomatic siblings to be transplanted; the usefulness of this strategy in the inflammasomopathies of this article is not established, because the indication for transplantation is not universal, but early identification allows surveillance [69, 70]. In asymptomatic carrier siblings, the reasonable approach is watchful waiting, with particular attention to febrile episodes, and any suspicious manifestation should prompt assessment of splenomegaly, anaemia, leucopenia, thrombocytopenia, transaminases, fibrinogen and coagulation, triglycerides, ferritin and soluble interleukin-2 receptor, with consultation of an experienced clinician if any parameter is abnormal [70].

When transplantation from a sibling donor is considered, potential donors should be tested for the familial variant, to ensure that only those who do not carry it in a causal state are assessed; in the recessive form, this means excluding homozygotes and compound heterozygotes [70].

Testing of asymptomatic minors in a family with a known variant is justified because the disease is of paediatric onset, treatable and potentially fatal, and because the result changes surveillance; it is a clear case in which the direct clinical benefit to the minor outweighs the usual arguments for deferring testing until adulthood. Genetic counselling is indicated for all patients undergoing genetic testing and should precede and follow the result [69].

8.3 Mosaicism and its consequences for counselling

Post-zygotic somatic mosaicism is a mechanism documented in five patients with NLRC4 variants and has two consequences for counselling. The first is that a patient with a mosaic variant restricted to the haematopoietic compartment, without germline involvement, has a very low transmission risk, but this statement can be made only when analysis of tissues of distinct embryological origin, such as hair, nails, urine or fibroblasts, documents the distribution of the variant; the literature shows examples with a fraction of 0% in hair, 3% in nails, 5% in urine and 34.1% in fibroblasts, which indicates mutational events at different times in development [32]. The second is that a parent with low-level mosaicism, possibly with few symptoms, may transmit the variant in the heterozygous germline state to a child with full-blown disease, a scenario that justifies testing the parents with sensitive methods when the proband's variant appears de novo in a dominant disease [32, 39].

8.4 Prenatal and preimplantation diagnosis

When the familial variant is known, prenatal diagnosis by molecular analysis of fetal DNA obtained by chorionic villus sampling or amniocentesis is technically possible, as is preimplantation genetic testing. The usefulness of these options is real in families with documented transmission, such as those described for the NLRC4 variants p.(Val341Ala), p.(His443Pro) and p.(Ser445Pro) and for the recessive DPP9 form, and limited in de novo variant situations, in which the recurrence risk is low [3, 4, 16, 40]. The decision should be made with information on the variability of expression, which is considerable in NLRC4 diseases, and with the note that the genotype does not allow severity to be predicted within a family. For NOCARH syndrome, in which the same variant produced an overlapping phenotype in four unrelated patients, the correlation is firmer and the prognostic information more solid [13].

The sources consulted do not describe contraceptives, assisted reproduction techniques or obstetric considerations specific to these diseases, and the management of pregnancy in an affected woman follows the general considerations already described for treatment.

9. Special Considerations: Population Diversity and Epidemiology

9.1 Known frequencies and what they measure

There is no population prevalence estimate for any of the diseases in this article. Orphanet assigns all of them the prevalence class of less than one case per million inhabitants, which is an administrative classification and not a measure obtained in an epidemiological study, and records counts of published cases: four cases or families for autoinflammation with infantile enterocolitis, based on the two original descriptions; 26 cases for familial cold autoinflammatory syndrome type 4, based on the 2018 systematic review; 15 cases for NOCARH syndrome, based on the original description and the 2023 series; two cases for Takenouchi-Kosaki syndrome, a clearly outdated number; and 19 cases for multiple self-healing palmoplantar carcinoma [21, 22, 65, 101, 102]. For coding in registries and administrative databases, the Orphanet pages (version of 21 October 2026) assign to each entity codes of the International Classification of Diseases in revisions 10 and 11 (ICD-10 and ICD-11): E85.0 and 4A60.Y for autoinflammation with infantile enterocolitis; D89.8 and 4A60.1 for familial cold autoinflammatory syndrome type 4; D76.1 and 4A01.23 for NOCARH syndrome; Q87.8 and LD2F.1 for Takenouchi-Kosaki syndrome; Q82.8 and 9B6Y for multiple self-healing palmoplantar carcinoma; and K52.8 (ICD-10) for CARD8-associated inflammatory bowel disease, for which Orphanet records autosomal dominant inheritance and onset in childhood, without a published summary [21, 22, 59, 65, 101, 102]. The ICD-10 codes are residual categories shared with other diseases, and the ICD-11 code 4A01.23 assigned to NOCARH syndrome also designates a category broader than the entity, shared with other immune disorders, so these codes do not allow the frequency of these entities to be estimated from hospital data. These counts measure publications, not patients, and underestimate the true frequency because they do not count cases described after the curation date.

The best approximation to the number of people identified with NLRC4 disease is the 2026 compilation, which counts 28 pathogenic variants in 56 people, 23 germline variants and five somatic [24]. For dipeptidyl peptidase 9 deficiency, adding up the publications consulted, five people have been described: four children from three families with the recessive form and one infant with the dominant form, to whom are added the twins of a 2025 report, whose genotype is not specified in the available abstract [16, 18, 68]. For NOCARH syndrome, the founding series described four patients, the German and French series four, and there are additional isolated reports [13, 46, 103]. None of these sums constitutes a prevalence estimate, and the literature is explicit that no population figure can be derived from the available sources [42, 66].

The absence of an estimate has a structural cause: these are diseases described little more than a decade ago, known from small families and sporadic de novo cases, without a GeneReviews chapter, without actionability curation by the ClinGen consortium and without a dedicated international registry [42]. The natural history study of the US National Institute of Allergy and Infectious Diseases, which explicitly includes NLRC4-associated macrophage activation among the eligible diseases and plans to recruit 3200 participants by 2031, is the most likely route to generating cohort data [98]. The 2018 systematic review, with the cases then published, estimated that about 80% of NLRC4 diseases had autosomal dominant transmission and that the severe phenotype accounted for about 30% of cases, with a mortality of 25%, proportions calculated on a small denominator biased by preferential publication of severe cases [60].

9.2 Geographical distribution of the descriptions

The published descriptions come mainly from North America, Europe and Japan, which reflects access to sequencing and diagnostic networks, not the true distribution of the disease. The two founding cases of NLRC4 disease were studied in the United States, in a patient of European ancestry at the National Institutes of Health and in a family at Yale [1, 3]. The cold-induced phenotype was described in a non-consanguineous three-generation Japanese family and, with another variant, in a large Dutch family and in a Spanish family with six affected generations [4, 40, 41]. Cases of NLRC4 disease have been published from France, Iran, China, Malaysia, Canada and Brazil, the last with the only recessive form described, in a 60-year-old woman [27, 30, 33, 36, 38, 39]. The four patients with the CDC42 p.(Arg186Cys) variant were studied in Rome and in the United States, with one patient born to parents from a Christian minority of the Middle East and others of Caucasian ancestry, and the patients with Golgi-retained variants were characterised in France, Germany and Japan [13, 44–46]. Dipeptidyl peptidase 9 deficiency has been described in two families from Israel, one of them consanguineous Bedouin, in a family of unrelated Ashkenazi parents, in a German family and in a pair of twins in India [16, 18, 68]. NLRP1 disease has been described in French, consanguineous Algerian, Dutch, Tunisian and consanguineous Belgian families [54–56].

No founder population has been identified for any of these diseases, and none of the variants described has an appreciable frequency in population databases; DPP9 is even significantly intolerant of loss-of-function variants in the exome aggregation databases, with no damaging homozygotes described [16]. The sources consulted contain no prevalence, incidence or patient series data from Portugal or Brazil, with the exception of the Brazilian woman with the recessive NLRC4 form; this absence means that such data have not been published or were not located in the sources consulted, not that the disease does not exist in these populations [33].

9.3 Special populations

The diseases of this article are, by their natural history, diseases of the paediatric age group: autoinflammation with infantile enterocolitis, NOCARH syndrome and dipeptidyl peptidase 9 deficiency manifest in the neonatal period or in the first months, and all described cases of NLRC4 enterocolitis began in infancy [13, 17, 24]. There are, however, two adult populations to recognise. The first is that of survivors, in whom the gastrointestinal inflammation attenuates or resolves but the systemic manifestations persist and require continued specialist follow-up [24]. The second is that of patients with late-onset disease due to somatic mosaicism, diagnosed between 21 and 60 years of age, whose identification depends on sensitive genetic methods and who would otherwise be classified as having uncharacterised adult autoinflammatory disease [32, 35].

Age conditions the risk of the main complication. The European recommendations for Still's disease note that very young children, particularly those under 18 months, are more likely to present with high levels of inflammation and to develop macrophage activation, and that this complication is more frequent under two years of age [8]. This observation, made in the polygenic disease, is consistent with what is seen in the monogenic forms, whose most severe presentation is neonatal.

9.4 European reference network and patient resources

The relevant European reference network is ERN RITA, dedicated to rare immunological diseases, which covers immunodeficiencies, autoinflammatory diseases and autoimmune diseases and provides a directory of centres, disease information, patient pathways and working groups on transition of care and on molecular testing [104]. For patients with the syndromic phenotype of Takenouchi-Kosaki syndrome, the European network dedicated to congenital malformations and rare intellectual disabilities is equally relevant, and for the ocular manifestations of multiple self-healing palmoplantar carcinoma the network dedicated to rare eye diseases.

Among resources for patients and professionals, the Autoinflammatory Alliance maintains a database of systemic autoinflammatory diseases, with a symptom search and a comparison chart that is a reference educational resource in these conditions, although it states that it does not replace medical assessment [105]. The natural history study of the US National Institute of Allergy and Infectious Diseases is the structured reference route for these patients for specialist assessment, with links to research and treatment protocols [98]. The phase 2 trial of the bispecific antibody against interleukin 1 beta and interleukin 18, which accepts patients from birth, is active but not recruiting, and the managed access programme for the same drug is registered but unavailable at the date of consultation, practical information relevant for those seeking access [84, 91].

10. Illustrative Clinical Case

The case described below is that of the patient identified as “Pt 1” in the series that defined NOCARH syndrome, published by Michael Lam and colleagues in the Journal of Experimental Medicine in 2019, and she is the only one of the four patients in that series who survived [13]. The patients' skin findings and the biochemical profile of the mutant protein appear in figure 1 of the article, accessible at https://pmc.ncbi.nlm.nih.gov/articles/PMC6888978/figure/fig1/, and the in vivo and ex vivo interleukin 1 beta and interleukin 18 data appear in figure 8, at https://pmc.ncbi.nlm.nih.gov/articles/PMC6888978/figure/fig8/ [13].

The patient was born to healthy unrelated parents, with no family history of genetic disease. At birth she had persistent high fever, erythematous skin lesions of the face, trunk and limbs, a distended abdomen due to hepatosplenomegaly, with a spleen 6.4 centimetres in diameter, and suspected trigonocephaly. Laboratory tests showed raised inflammatory markers and severe trilineage cytopenia, which required frequent transfusions of red cell concentrate and platelets during the first year of life. Bone marrow biopsy at disease onset revealed fibrosis with trilineage dyshaematopoiesis, characterised by a predominance of early differentiation elements in the erythroid and myeloid lineages, dysmorphic megakaryocytes in reduced numbers, increased lymphocytes and abundant CD68-positive histiocytes with focal or absent phagocytosis.

Because of the partial overlap with chronic infantile neurological, cutaneous and articular syndrome and neonatal-onset multisystem inflammatory disease, she was started on glucocorticoids and daily anakinra, with improvement of fever and rash but no effect on the cytopenias. There was failure to thrive. Granulocyte colony-stimulating factor was started, with a partial response. Reduction or discontinuation of glucocorticoids was followed by recurrence of the inflammatory symptoms. At 11 months of age, recurrent episodes of intestinal haemorrhage associated with persistent diarrhoea appeared, and Meckel's diverticulum was excluded. At two years and six months she had three episodes of generalised seizures followed by hypotonia, hypersomnia and transient hemiparesis, one of which required admission to intensive care; magnetic resonance imaging performed at the first episode showed lesions suggestive of inflammatory involvement of the central nervous system and, at the third, an additional brainstem lesion. The cerebrospinal fluid was sterile, with 18 leucocytes per cubic millimetre and no haemophagocytic cells. All episodes were treated with high-dose intravenous glucocorticoid, with a good response, and levetiracetam was started. At five years, because of incomplete control of the inflammatory state, anakinra was replaced by canakinumab.

From five years of age she developed four episodes of full haemophagocytic lymphohistiocytosis, meeting the HLH-2004 criteria. All were treated with high-dose intravenous glucocorticoids and ciclosporin A, with resolution in all but one. The last episode was severe and was associated with massive intestinal ischaemia and necrosis, which required surgical resection and ileocolostomy, and it did not respond to high-dose glucocorticoids or to interleukin-1 inhibition; the addition of emapalumab rapidly resolved the episode. At six years she underwent allogeneic haematopoietic cell transplantation from her father, a haploidentical donor, with selective depletion of αβ T lymphocytes and CD19-positive cells from the graft, intended to prevent, respectively, graft-versus-host disease and Epstein-Barr virus-associated post-transplant lymphoproliferative disease. The conditioning regimen combined treosulfan, thiotepa and fludarabine. To prevent haemophagocytosis flares, emapalumab was continued until day +28 and anakinra until day +100 after transplantation. Neutrophil recovery occurred on day +18 and platelet recovery on day +12, and successful engraftment of donor haematopoiesis was accompanied by normalisation of interleukin 1 beta and interleukin 18 production. Up to day +210 there was no inflammatory or haemophagocytosis flare.

Trio exome sequencing identified the de novo CDC42 variant c.556C>T, p.(Arg186Cys), corresponding to genomic position 1:22,417,990 in the GRCh37 reference genome and 1:22,091,497 in GRCh38, affecting an exon specific to isoform 1; the variant was absent from population aggregation databases and was predicted to be deleterious by prediction tools. Variants in genes associated with haemophagocytosis or with other immunological and haematological diseases were excluded [13, 26].

Laboratory studies in this patient documented the mechanism. Spontaneous release of interleukin 18 by unstimulated bone marrow mononuclear cells was about 300 times that of healthy donors, and levels of the cytokine in the marrow supernatant and in plasma were about 1000 times higher; stimulation with lipopolysaccharide, with or without adenosine triphosphate, did not increase production further, which indicates maximal constitutive activation of processing and secretion. Interleukin 6 showed no spontaneous release and circulating levels were comparable to those of healthy individuals. Interferon gamma was raised, especially during the haemophagocytosis episodes, and its levels correlated closely with those of CXCL9, with a coefficient of 0.82, which in turn correlated with ferritin, also with a coefficient of 0.82; interleukin 18 did not correlate with interferon gamma or with ferritin, which supports the interpretation that the cytokine creates the predisposition but does not by itself determine the timing of the episode [13].

Three aspects of this case merit comment. The first is that the management described is not routine: emapalumab was used outside its approved indication, the haploidentical transplantation with selective depletion and conditioning with treosulfan, thiotepa and fludarabine reflects the choice of a centre experienced with that platform, and none of these decisions is supported by a clinical trial in this disease [13, 81]. The second is that the course illustrates a divergence from the available guidelines: the European recommendations for macrophage activation place anakinra, ciclosporin and interferon-gamma inhibitors on the same footing as initial therapy to be considered, whereas in this case interleukin-1 blockade and ciclosporin were insufficient and only interferon-gamma blockade resolved the refractory episode, which is consistent with the observation that in Golgi-retained variants the interferon pathway is hyperactivated [8, 45]. The third is that the article refers to this patient as male in one paragraph of the results section and as female in the detailed clinical history, an inconsistency in the published text itself that is recorded here so that reading the two sections does not cause confusion [13].

11. Prognosis and Long-Term Follow-up

11.1 Prognostic data

The prognosis of these diseases depends on the phenotype and, in the severe forms, on early recognition. In NLRC4 disease, the systematic review of cases published up to 2018 attributed a mortality of 25% to the severe phenotype, with the mild phenotype showing no visceral failure; these figures come from a small denominator, subject to publication bias towards the most severe cases, and do not constitute a cohort survival estimate [60]. The individual outcomes documented illustrate the extremes: death on the twenty-third day of life from diffuse alveolar haemorrhage in the index case of the original description, death at two months in a preterm infant with high-allele-fraction mosaicism and, at the other extreme, survival to 43 and 57 years in carriers of a germline variant and of a low-fraction mosaic variant, respectively [3, 34, 35]. In survivors, the enterocolitis attenuates or resolves with age, but the systemic manifestations may persist into adulthood, and some patients maintain raised total interleukin 18 for decades without developing overt macrophage activation [24, 29].

In NOCARH syndrome, the prognosis is the most guarded of the group: three of the four patients of the original description died, at four and a half months, at six to seven months and at one and a half years, and mortality remained high in subsequent series despite inflammasome-directed treatments [13, 46]. Survival of the only patient of the founding series depended, according to the authors' interpretation, on timely neutralisation of interleukin 1 and interferon gamma followed by transplantation, and the explicit conclusion is that early recognition, rapid treatment and transplantation determine survival; the association of early-onset trilineage cytopenia with autoinflammation is the sign that should trigger this sequence [13]. In the C-terminal variants that are not retained in the Golgi, the prognosis is better: of the four patients treated with interleukin-1 blockade, none died, three went on to achieve developmental milestones and the cytopenias resolved in three, although one continued to have intermittent flares and serious infections occurred over six and eleven years of treatment [14].

In the recessive form of dipeptidyl peptidase 9 deficiency, three of the four patients of the original description required haematopoietic cell transplantation, which indicates progressive haematological disease, and the patients were aged between 6 and 14 years at the time of publication, with no deaths reported [16, 17]. In the dominant form, the only patient described died on the sixth day after transplantation, during conditioning, from staphylococcal sepsis with parainfluenza reactivation, a hyperinflammatory flare and massive capillary leak [18]. The observation that loss of progenitor cells is intrinsic and mediated by CARD8 inflammasome-dependent pyroptosis explains why transplantation is, so far, the only treatment with curative intent in this disease [19].

In Takenouchi-Kosaki syndrome, the vital prognosis is determined not by inflammation but by the malformations, the intellectual disability and the haematological and lymphatic complications; the patients described reached 18 and 22 years, with height and head circumference well below average and, in one of them, severe intellectual disability and protein-losing enteropathy [12]. The reported occurrence of immunodeficiency, autoimmunity and neoplasia in carriers of the recurrent variant, including Hodgkin lymphoma in a patient with another variant in the gene, justifies active follow-up and not merely dysmorphological follow-up [49, 66].

Some cross-cutting prognostic indicators have been systematised for haemophagocytosis and apply here. The prognosis depends on the nature of the underlying contributors, the degree of organ dysfunction and the duration of active disease; liver involvement is common and may progress to potentially fatal liver failure; multiorgan dysfunction requiring intensive care is a strong predictor of poor outcome; and central nervous system involvement is associated both with mortality and with long-term neurological sequelae in survivors, regardless of the cause [69]. Both a higher initial ferritin value and the absence of ferritin improvement on therapy are associated with a worse outcome [69]. After a successful transplant, surveillance comes to focus on the complications of the transplant itself [70].

11.2 Recommended surveillance

Surveillance comes with three caveats. The first is that no comparative study has evaluated the ideal laboratory monitoring protocol, and the intervals above derive from consensus recommendations and from the practice described in published series [69]. The second is that failure to respond to initial therapy should lead to careful re-examination of both the underlying diagnoses and the therapeutic approach, and not merely to dose increases [69]. The third is that there is no published oncological surveillance protocol for NLRP1 diseases, despite the documented susceptibility to squamous cell carcinoma, so dermatological and ophthalmological surveillance is based on clinical judgement and not on an established standard [57].

12. Summary and Final Considerations

The diseases brought together in this article are defined by a common mechanism and a common complication. The mechanism is permanent activation of an inflammasome, through gain of function of the sensor, as in NLRC4 and NLRP1, through dysregulation of a GTPase that conditions assembly of the pyrin inflammasome, as in the C-terminal CDC42 variants, or through loss of the brake that keeps the NLRP1 and CARD8 sensors inactive, as in dipeptidyl peptidase 9 deficiency. The complication is macrophage activation syndrome, the result of excessive and chronic production of interleukin 18, which acts as a co-stimulator of interferon-gamma production and creates the predisposition to episodes of hyperinflammation with cytopenias, hyperferritinaemia, coagulopathy and haemophagocytosis.

The most important clinical consequence of this pathophysiology is that these diseases mimic familial haemophagocytic lymphohistiocytosis without having its defect. The cytotoxic function of T lymphocytes and natural killer cells is typically normal, perforin expression and degranulation assays are normal, and this means that a normal result on these tests, in an infant with a haemophagocytosis picture, does not close the genetic investigation: it shifts it towards macrophage-intrinsic hyperinflammation. The laboratory test that gives the best guidance is measurement of interleukin 18, whose persistent and very marked elevation, at values not seen in the cryopyrinopathies or familial Mediterranean fever, identifies the group and indicates the cytokine to target.

The diagnosis is molecular and should be pursued early, with a broad panel or exome or genome sequencing, preferably as a trio, without waiting for conventional investigation to be exhausted, because the result changes management and prognosis. Three concrete precautions increase the yield. The first is to remember that low-allele-fraction somatic mosaicism causes disease, including disease of adult onset, and requires sensitive methods that conventional sequencing does not offer. The second is to interpret variants with functional or biochemical support, given the excess of variants of uncertain significance in NLRC4. The third is to check the nomenclature against current reference transcripts, because three notations widely cited in the literature on these diseases do not correspond to those transcripts.

In CDC42, the genetic report has to be read at the level of the amino acid, not the gene. The p.(Arg186Cys) variant predicts neonatal pancytopenia with severe autoinflammation and haemophagocytosis, requires interferon-gamma neutralisation and points towards transplantation; the p.(Cys188Tyr) and p.(Ter192CysextTer24) variants predict disease responsive to interleukin-1 blockade; the p.(Tyr64Cys) variant predicts a developmental disorder with sometimes intermittent macrothrombocytopenia, whose management is supportive with immunological and oncological surveillance. Rarely in medical genetics does the distinction between two adjacent residues have such a direct therapeutic consequence.

Treatment is entirely off-label and is organised in layers. In acute macrophage activation, high-dose glucocorticoids are the mainstay, with ciclosporin, anakinra at higher than usual doses and interferon-gamma blockade to be considered as initial therapy in severe or rapidly worsening cases, and with combinations frequently necessary. For background control, interleukin-1 blockade resolves fever and rash in most patients but frequently fails to control the cytopenias and the macrophage activation of the interleukin-18-dominated forms, and it is in this space that recombinant interleukin-18 binding protein, whose only published dose comes from compassionate use, and the bispecific antibody against interleukin 1 beta and interleukin 18, still without published trial results, are positioned. Janus kinase inhibition has a rationale in Golgi-retained CDC42 variants, with hyperactivation of the STING pathway and the type I interferon response, and produced complete remission in one patient. Haematopoietic cell transplantation is the only intervention with curative intent, indispensable in dipeptidyl peptidase 9 deficiency because of the intrinsic loss of progenitors and decisive in the most severe CDC42 forms, and it requires control of inflammation before conditioning.

The limitations of the available knowledge are considerable and should accompany any clinical decision. There is no specific guideline, consensus or surveillance protocol for these three diseases, and management rests on extrapolation from the European recommendations for macrophage activation in Still's disease and from the points to consider for haemophagocytosis. There is no population prevalence estimate for any of them, only counts of published cases, and the sources consulted contain no epidemiological data from Portugal or Brazil, with the exception of one Brazilian woman. The only trial with posted results included fifteen patients with two different diseases, does not break down the results by disease and does not publish the dose used. And the interpretation of almost all the therapeutic evidence is limited by the fact that published patients received several drugs simultaneously. What partly compensates for this fragility is the clarity of the mechanism: few rare diseases offer such a direct link between a variant, a measurable cytokine and a therapeutic target, and it is this link that underpins the central practical recommendation of this article, which is to measure interleukin 18 and to sequence early any infant with severe enterocolitis, cytopenias and recurrent hyperferritinaemia.

13. References

1. Canna SW, de Jesus AA, Gouni S, Brooks SR, Marrero B, Liu Y, et al. An activating NLRC4 inflammasome mutation causes autoinflammation with recurrent macrophage activation syndrome. Nat Genet. 2014;46(10):1140-6. doi:10.1038/ng.3089
2. Alehashemi S, Goldbach-Mansky R. Human autoinflammatory diseases mediated by NLRP3-, pyrin-, NLRP1-, and NLRC4-inflammasome dysregulation: updates on diagnosis, treatment, and the respective roles of IL-1 and IL-18. Front Immunol. 2020;11:1840. doi:10.3389/fimmu.2020.01840
3. Romberg N, Al Moussawi K, Nelson-Williams C, Stiegler AL, Loring E, Choi M, et al. Mutation of NLRC4 causes a syndrome of enterocolitis and autoinflammation. Nat Genet. 2014;46(10):1135-9. doi:10.1038/ng.3066
4. Kitamura A, Sasaki Y, Abe T, Kano H, Yasutomo K. An inherited mutation in NLRC4 causes autoinflammation in human and mice. J Exp Med. 2014;211(12):2385-96. doi:10.1084/jem.20141091
5. Online Mendelian Inheritance in Man. Familial cold autoinflammatory syndrome 4; FCAS4. MIM 616115. https://omim.org/entry/616115
6. Canna SW, Girard C, Malle L, de Jesus A, Romberg N, Kelsen J, et al. Life-threatening NLRC4-associated hyperinflammation successfully treated with IL-18 inhibition. J Allergy Clin Immunol. 2017;139(5):1698-701. doi:10.1016/j.jaci.2016.10.022
7. Weiss ES, Girard-Guyonvarc'h C, Holzinger D, de Jesus AA, Tariq Z, Picarsic J, et al. Interleukin-18 diagnostically distinguishes and pathogenically promotes human and murine macrophage activation syndrome. Blood. 2018;131(13):1442-55. doi:10.1182/blood-2017-12-820852
8. Fautrel B, Mitrovic S, De Matteis A, Bindoli S, Antón J, Belot A, et al. EULAR/PReS recommendations for the diagnosis and management of Still's disease, comprising systemic juvenile idiopathic arthritis and adult-onset Still's disease. Ann Rheum Dis. 2024;83(12):1614-27. doi:10.1136/ard-2024-225851
9. Landy E, Carol H, Ring A, Canna S. Biological and clinical roles of IL-18 in inflammatory diseases. Nat Rev Rheumatol. 2024;20(1):33-47. doi:10.1038/s41584-023-01053-w
10. Takenouchi T, Kosaki R, Niizuma T, Hata K, Kosaki K. Macrothrombocytopenia and developmental delay with a de novo CDC42 mutation: yet another locus for thrombocytopenia and developmental delay. Am J Med Genet A. 2015;167A(11):2822-5. doi:10.1002/ajmg.a.37275
11. Martinelli S, Krumbach OHF, Pantaleoni F, Coppola S, Amin E, Pannone L, et al. Functional dysregulation of CDC42 causes diverse developmental phenotypes. Am J Hum Genet. 2018;102(2):309-20. doi:10.1016/j.ajhg.2017.12.015
12. Online Mendelian Inheritance in Man. Takenouchi-Kosaki syndrome; TKS. MIM 616737. https://omim.org/entry/616737
13. Lam MT, Coppola S, Krumbach OHF, Prencipe G, Insalaco A, Cifaldi C, et al. A novel disorder involving dyshematopoiesis, inflammation, and HLH due to aberrant CDC42 function. J Exp Med. 2019;216(12):2778-99. doi:10.1084/jem.20190147
14. Gernez Y, de Jesus AA, Alsaleem H, Macaubas C, Roy A, Lovell D, et al. Severe autoinflammation in 4 patients with C-terminal variants in cell division control protein 42 homolog (CDC42) successfully treated with IL-1β inhibition. J Allergy Clin Immunol. 2019;144(4):1122-1125.e6. doi:10.1016/j.jaci.2019.06.017
15. Coppola S, Insalaco A, Zara E, Di Rocco M, Pires Marafon D, Spadaro F, et al. Mutations at the C-terminus of CDC42 cause distinct hematopoietic and autoinflammatory disorders. J Allergy Clin Immunol. 2022;150(1):223-8. doi:10.1016/j.jaci.2022.01.024
16. Harapas CR, Robinson KS, Lay K, Wong J, Moreno Traspas R, Nabavizadeh N, et al. DPP9 deficiency: an inflammasomopathy that can be rescued by lowering NLRP1/IL-1 signaling. Sci Immunol. 2022;7(75):eabi4611. doi:10.1126/sciimmunol.abi4611
17. Online Mendelian Inheritance in Man. Hatipoglu immunodeficiency syndrome; HATIS. MIM 620331. https://omim.org/entry/620331
18. Wolf C, Fischer H, Kühl JS, Koss S, Abou Jamra R, Starke S, et al. Hemophagocytic lymphohistiocytosis-like hyperinflammation due to a de novo mutation in DPP9. J Allergy Clin Immunol. 2023;152(5):1336-1344.e5. doi:10.1016/j.jaci.2023.07.013
19. Xiao T, Brewer JR, Carlino M, Han A, Takabe YJ, Lee CY, et al. Reverse genetics in humanized mice reveals CARD8-mediated pyroptosis causing pancytopenia in human DPP9 deficiency. J Clin Invest. 2026;136(18):e207530. doi:10.1172/JCI207530
20. Poli MC, Aksentijevich I, Bousfiha AA, Cunningham-Rundles C, Hambleton S, Klein C, et al. Human inborn errors of immunity: 2024 update on the classification from the International Union of Immunological Societies Expert Committee. J Hum Immun. 2025;1(1):e20250003. doi:10.70962/jhi.20250003
21. Orphanet. Periodic fever-infantile enterocolitis-autoinflammatory syndrome (ORPHA:436166). Paris: INSERM; 2026. Available from: https://www.orpha.net/en/disease/detail/436166
22. Orphanet. NOCARH syndrome (ORPHA:619363). Paris: INSERM; 2026. Available from: https://www.orpha.net/en/disease/detail/619363
23. Mertz P, Hentgen V, Boursier G, Elhani I, Calas L, Delon J, et al. [Autoinflammatory diseases associated with IL-18]. Rev Med Interne. 2025;46(3):155-63. doi:10.1016/j.revmed.2024.08.003
24. Wang Y, Lingappa S, Kang Z. Current understanding of the NLRC4 inflammasome in autoinflammation and enterocolitis. Commun Biol. 2026;9(1):611. doi:10.1038/s42003-026-10174-3
25. Online Mendelian Inheritance in Man. NLR family, caspase recruitment domain-containing 4; NLRC4. MIM 606831. https://omim.org/entry/606831
26. VariantValidator. HGVS variant normalisation (GRCh38, MANE Select transcripts). https://variantvalidator.org/
27. Bardet J, Laverdure N, Fusaro M, Picard C, Garnier L, Viel S, et al. NLRC4 GOF mutations, a challenging diagnosis from neonatal age to adulthood. J Clin Med. 2021;10(19):4369. doi:10.3390/jcm10194369
28. Jeskey J, Parida A, Graven K, Hostoffer R. Novel gene deletion in NLRC4 expanding the familial cold inflammatory syndrome phenotype. Allergy Rhinol (Providence). 2020;11:2152656720928062. doi:10.1177/2152656720928062
29. Willemsen M, Schoonbrood THM, Rodriguez E, Roosens W, De Langhe E, Schrijvers R, et al. Free IL-18 in NLRC4-associated autoinflammatory disease without macrophage activation syndrome. Rheumatology (Oxford). 2026;65(7):keag363. doi:10.1093/rheumatology/keag363
30. Asna Ashari K, Parvaneh N, Mirnia K, Ayati M, Saeedi M, Salehzadeh F, et al. Three cases of autoinflammatory disease with novel NLRC4 mutations, and the first mutation reported in the CARD domain of NLRC4 associated with autoinflammatory infantile enterocolitis (AIFEC). Pediatr Rheumatol Online J. 2024;22(1):90. doi:10.1186/s12969-024-01020-z
31. Mao Y, Xie Y. Autoinflammation with infantile enterocolitis associated with a novel CARD domain mutation in NLRC4: a case report and literature review. Front Pediatr. 2026;14:1878056. doi:10.3389/fped.2026.1878056
32. Diab F, Louvrier C, Fabre M, Lin C, Rabbaa M, Assrawi E, et al. Somatic mosaic NLRC4 variants in autoinflammatory diseases: functional characterization and correlation of mosaicism levels with disease age of onset and severity. J Clin Immunol. 2025;45(1):111. doi:10.1007/s10875-025-01907-w
33. Steiner A, Reygaerts T, Pontillo A, Ceccherini I, Moecking J, Moghaddas F, et al. Recessive NLRC4-autoinflammatory disease reveals an ulcerative colitis locus. J Clin Immunol. 2022;42(2):325-35. doi:10.1007/s10875-021-01175-4
34. Liang J, Alfano DN, Squires JE, Riley MM, Parks WT, Kofler J, et al. Novel NLRC4 mutation causes a syndrome of perinatal autoinflammation with hemophagocytic lymphohistiocytosis, hepatosplenomegaly, fetal thrombotic vasculopathy, and congenital anemia and ascites. Pediatr Dev Pathol. 2017;20(6):498-505. doi:10.1177/1093526616686890
35. Ionescu D, Peñín-Franch A, Mensa-Vilaró A, Castillo P, Hurtado-Navarro L, Molina-López C, et al. First description of late-onset autoinflammatory disease due to somatic NLRC4 mosaicism. Arthritis Rheumatol. 2022;74(4):692-9. doi:10.1002/art.41999
36. Wang L, Wen W, Deng M, Li Y, Sun G, Zhao X, et al. A novel mutation in the NBD domain of NLRC4 causes mild autoinflammation with recurrent urticaria. Front Immunol. 2021;12:674808. doi:10.3389/fimmu.2021.674808
37. Kawasaki Y, Oda H, Ito J, Niwa A, Tanaka T, Hijikata A, et al. Identification of a high-frequency somatic NLRC4 mutation as a cause of autoinflammation by pluripotent cell-based phenotype dissection. Arthritis Rheumatol. 2017;69(2):447-59. doi:10.1002/art.39960
38. Barsalou J, Blincoe A, Fernandez I, Dal-Soglio D, Marchitto L, Selleri S, et al. Rapamycin as an adjunctive therapy for NLRC4 associated macrophage activation syndrome. Front Immunol. 2018;9:2162. doi:10.3389/fimmu.2018.02162
39. Wang J, Ye Q, Zheng W, Yu X, Luo F, Fang R, et al. Low-ratio somatic NLRC4 mutation causes late-onset autoinflammatory disease. Ann Rheum Dis. 2022;81(8):1173-8. doi:10.1136/annrheumdis-2021-221708
40. Gil-Lianes J, Gariup G, Iranzo-Fernández P, Mensa-Vilaró A, Peñín-Franch A, Hurtado-Navarro L, et al. Early-onset recurrent panniculitis as a phenotype of NLRC4-associated autoinflammatory syndrome: characterization of pathogenicity of the p.Ser445Pro NLRC4 variant. Australas J Dermatol. 2023;64(2):260-7. doi:10.1111/ajd.14005
41. Volker-Touw CM, de Koning HD, Giltay JC, de Kovel CG, van Kempen TS, Oberndorff KM, et al. Erythematous nodes, urticarial rash and arthralgias in a large pedigree with NLRC4-related autoinflammatory disease, expansion of the phenotype. Br J Dermatol. 2017;176(1):244-8. doi:10.1111/bjd.14757
42. GeneGuidance. NLRC4 — gene and condition records (Autoinflammation with infantile enterocolitis, OMIM 616050, ORPHA 436166; Familial cold autoinflammatory syndrome 4, OMIM 616115, ORPHA 576349). Local geneguidance-FULL database, extracted 2026-09-26.
43. Online Mendelian Inheritance in Man. Cell division cycle 42; CDC42. MIM 116952. https://omim.org/entry/116952
44. Nishitani-Isa M, Mukai K, Honda Y, Nihira H, Tanaka T, Shibata H, et al. Trapping of CDC42 C-terminal variants in the Golgi drives pyrin inflammasome hyperactivation. J Exp Med. 2022;219(6):e20211889. doi:10.1084/jem.20211889
45. Iannuzzo A, Delafontaine S, El Masri R, Tacine R, Prencipe G, Nishitani-Isa M, et al. Autoinflammatory patients with Golgi-trapped CDC42 exhibit intracellular trafficking defects leading to STING hyperactivation and ER stress. Nat Commun. 2024;15(1):9940. doi:10.1038/s41467-024-54294-y
46. Kapp FG, Kretschmer S, Beckmann CCA, Wäsch L, Molitor A, Carapito R, et al. C-terminal variants in CDC42 drive type I interferon-dependent autoinflammation in NOCARH syndrome reversible by ruxolitinib. Clin Immunol. 2023;256:109777. doi:10.1016/j.clim.2023.109777
47. Feng S, Reygaerts T, Kong K, Wierzbowski MC, Narayanan DL, Baker PJ, et al. An autoinflammatory CDC42 variant M45L demonstrates the mechanism of pyrin inflammasome activation. Sci Immunol. 2026;11(122):eadz8733. doi:10.1126/sciimmunol.adz8733
48. Webster H. CDC42-pyrin interactions promote inflammasome activation and autoinflammatory disease. Nat Rev Rheumatol. 2026;22(11):671. doi:10.1038/s41584-026-01433-y
49. Szczawinska-Poplonyk A, Ploski R, Bernatowska E, Pac M. A novel CDC42 mutation in an 11-year old child manifesting as syndromic immunodeficiency, autoinflammation, hemophagocytic lymphohistiocytosis, and malignancy: a case report. Front Immunol. 2020;11:318. doi:10.3389/fimmu.2020.00318
50. Bucciol G, Pillay B, Casas-Martin J, Delafontaine S, Proesmans M, Lorent N, et al. Systemic inflammation and myelofibrosis in a patient with Takenouchi-Kosaki syndrome due to CDC42 Tyr64Cys mutation. J Clin Immunol. 2020;40(4):567-70. doi:10.1007/s10875-020-00742-5
51. van Wijck RTA, Swagemakers SMA, van der Spek PJ, van Hagen PM, van Daele PLA. A CDC42 stop-loss mutation in a patient with relapsing polychondritis and autoinflammation. J Clin Immunol. 2023;43(1):69-71. doi:10.1007/s10875-022-01344-z
52. Online Mendelian Inheritance in Man. Dipeptidyl peptidase IX; DPP9. MIM 608258. https://omim.org/entry/608258
53. ClinVar. NM_139159.5(DPP9):c.499G>A (p.Gly167Ser), Hatipoglu immunodeficiency syndrome, conflicting classifications. Bethesda: NCBI. https://www.ncbi.nlm.nih.gov/clinvar/ DPP9 records also consulted: NM_139159.5:c.331C>T (p.Arg111Ter), c.641C>G (p.Ser214Ter) e c.2551C>T (p.Gln851Ter).
54. Online Mendelian Inheritance in Man. Palmoplantar carcinoma, multiple self-healing; MSPC. MIM 615225. https://omim.org/entry/615225
55. Online Mendelian Inheritance in Man. Autoinflammation with arthritis and dyskeratosis; AIADK. MIM 617388. https://omim.org/entry/617388
56. Online Mendelian Inheritance in Man. Respiratory papillomatosis, juvenile recurrent, congenital; JRRP. MIM 618803. https://omim.org/entry/618803
57. GeneGuidance. NLRP1 — gene record. Local geneguidance-FULL database, extracted 2026-09-26.
58. GeneGuidance. CARD8 — gene record. Local geneguidance-FULL database, extracted 2026-09-26.
59. Orphanet. CARD8-related inflammatory bowel disease (ORPHA:714410). Paris: INSERM; 2026. Available from: https://www.orpha.net/en/disease/detail/714410
60. Rodrigues F, Hentgen V, Bachmeyer C, Kone-Paut I, Belot A, Grateau G, et al. [NLRC4 associated autoinflammatory diseases: a systematic review of the current literature]. Rev Med Interne. 2018;39(4):279-86. doi:10.1016/j.revmed.2018.02.003
61. Online Mendelian Inheritance in Man. Autoinflammation with infantile enterocolitis; AIFEC. MIM 616050. Baltimore: Johns Hopkins University. https://omim.org/entry/616050
62. Wang X, Xiao Y, Ge T, Zhang T, Li X. Autoinflammation with infantile enterocolitis induced by a heterozygous variant (c.1357C>T) in the NLRC4 gene: a case report. Front Pediatr. 2026;14:1822554. doi:10.3389/fped.2026.1822554
63. Zekre F, Laurent A, Ranchin B, Foray AP, Thilloy M, Laverdure N, et al. Occurrence of lupus-like nephritis in a girl with NLRC4 gain-of-function mutations. Kidney Int Rep. 2024;9(4):1136-8. doi:10.1016/j.ekir.2024.01.061
64. Wang Q, Ye X, Zheng W, Jin T, Zhou Y, Hu X, et al. NLRC4 gain-of-function variant is identified in a patient with systemic lupus erythematosus. Clin Immunol. 2023;255:109731. doi:10.1016/j.clim.2023.109731
65. Orphanet. NLRC4-related familial cold autoinflammatory syndrome (ORPHA:576349). Paris: INSERM; 2026. Available from: https://www.orpha.net/en/disease/detail/576349
66. GeneGuidance. CDC42 — gene and condition record (Takenouchi-Kosaki syndrome, OMIM 616737, ORPHA 487796). Local geneguidance-FULL database, extracted 2026-09-26.
67. García-Cazorla Á, Dionisi-Vici C, Saudubray JM. Disorders of cellular trafficking. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 44, sections 44.1.3 (Table 44.2) and 44.3.1.
68. Singh S, Garg A, Mandal P. Recurrent transfusions and severe infections: unmasking Hatipoglu immunodeficiency syndrome: case report and review of literature. J Pediatr Hematol Oncol. 2025;47(8):390-1. doi:10.1097/MPH.0000000000003135
69. Shakoory B, Geerlinks A, Wilejto M, Kernan K, Hines M, Romano M, et al. The 2022 EULAR/ACR points to consider at the early stages of diagnosis and management of suspected haemophagocytic lymphohistiocytosis/macrophage activation syndrome (HLH/MAS). Ann Rheum Dis. 2023;82(10):1271-85. doi:10.1136/ard-2023-224123
70. Zhang K, Astigarraga I, Bryceson Y, Lehmberg K, Machowicz R, Marsh R, Sieni E, Wang Z, Nichols KE. Familial hemophagocytic lymphohistiocytosis. 2006 Mar 22 [updated 2021 Sep 30; revised 2024 Jun 6]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1444/
71. Krei JM, Møller HJ, Larsen JB. The role of interleukin-18 in the diagnosis and monitoring of hemophagocytic lymphohistiocytosis/macrophage activation syndrome: a systematic review. Clin Exp Immunol. 2021;203(2):174-82. doi:10.1111/cei.13543
72. Girard-Guyonvarc'h C, Rodriguez E, Mueller YM, Caruso A, Katsikis PD, Gabay C, et al. Elevated serum levels of interleukin-18 discriminate Still's disease from other autoinflammatory conditions: results from the European ImmunAID cohort. RMD Open. 2025;11(2):e005388. doi:10.1136/rmdopen-2024-005388
73. Miwa Y, Kawamoto N, Miwa T, Haba A, Kadowaki S, Shiraki M, et al. Transition of serum-free IL-18 levels in systemic juvenile idiopathic arthritis measured by a novel detection method. Cytokine. 2026;201:157128. doi:10.1016/j.cyto.2026.157128
74. Shu H, Liang M. Serum IL-18 in autoinflammatory diseases diagnosis: unresolved challenges. Eur J Intern Med. 2025;138:172-3. doi:10.1016/j.ejim.2025.04.029
75. Rogani G, Erkens RGA, Bodelón A, Mocking TR, Putmans MTM, Huijsmans FMCH, et al. Persistent interleukin-18 fuels expansion of CD38+HLA-DR+CD8+ T cells in Still disease and macrophage activation syndrome. Arthritis Rheumatol. 2026. doi:10.1002/art.70319
76. Meyer L, Hines M, Zhang K, Nichols KE. X-linked lymphoproliferative disease. 2004 Feb 27 [updated 2024 May 16]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1406/
77. GeneGuidance. XIAP — gene and condition record (Lymphoproliferative syndrome, X-linked, 2, OMIM 300635, ORPHA 538934). Local geneguidance-FULL database, extracted 2026-09-26.
78. Bracaglia C, Marucci G, Del Chierico F, Russo A, Pardeo M, Pires Marafon D, et al. Microbiota transplant to control inflammation in a patient with NLRC4 gain-of-function-induced disease. J Allergy Clin Immunol. 2023;152(1):302-3. doi:10.1016/j.jaci.2023.03.031
79. European Medicines Agency. Kineret (anakinra): summary of product characteristics (EPAR product information). Amsterdam: EMA; authorised 8 March 2002, latest European Commission decision 4 September 2025. https://www.ema.europa.eu/en/documents/product-information/kineret-epar-product-information_en.pdf
80. U.S. Food and Drug Administration. KINERET (anakinra) injection, for subcutaneous use: prescribing information. Silver Spring: FDA; label effective 3 April 2026. Drugs@FDA BLA103950.
81. U.S. Food and Drug Administration. GAMIFANT (emapalumab-lzsg) injection, for intravenous use: prescribing information. Silver Spring: FDA; label effective 26 March 2026. Drugs@FDA BLA761107. https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=761107
82. ClinicalTrials.gov. Therapeutic use of tadekinig alfa in NLRC4 mutation and XIAP deficiency (NCT03113760). AB2 Bio Ltd; phase 3; completed 2 November 2023; results posted on the registry. https://clinicaltrials.gov/study/NCT03113760
83. Caorsi R, Bertoni A, Matucci-Cerinic C, Natoli V, Palmeri S, Rosina S, et al. Long-term efficacy of MAS825, a bispecific anti-IL1β and IL-18 monoclonal antibody, in two patients with systemic JIA and recurrent episodes of macrophage activation syndrome. Rheumatology (Oxford). 2025;64(3):1528-33. doi:10.1093/rheumatology/keae440
84. ClinicalTrials.gov. Study to evaluate the efficacy, safety and tolerability of MAS825 in patients with monogenic IL-18 driven autoinflammatory diseases, including NLRC4-GOF, XIAP deficiency, or CDC42 mutations (NCT04641442). Novartis; phase 2; active, not recruiting. https://clinicaltrials.gov/study/NCT04641442
85. European Medicines Agency. Ilaris (canakinumab): European public assessment report. Amsterdam: EMA; authorised 23 October 2009, latest European Commission decision 30 April 2026. https://www.ema.europa.eu/en/medicines/human/EPAR/ilaris
86. INFARMED. Ilaris (canakinumab), solution for injection 150 mg/mL: medicine information. Lisbon: INFARMED, Infomed; marketing authorisation status “Autorizado” (23 February 2017); procedure EMEA/H/C/1109/X. https://extranet.infarmed.pt/INFOMED-fo/
87. European Medicines Agency. Jakavi (ruxolitinib): European public assessment report. Amsterdam: EMA; authorised 23 August 2012. https://www.ema.europa.eu/en/medicines/human/EPAR/jakavi
88. INFARMED. Kineret (anakinra), solution for injection 100 mg/0.67 mL: medicine information. Lisbon: INFARMED, Infomed; marketing authorisation status “Autorizado” (15 November 2013); procedure EMEA/H/C/363/X. https://extranet.infarmed.pt/INFOMED-fo/
89. European Medicines Agency. Gamifant (emapalumab): refusal of marketing authorisation (EMEA/H/C/004386). Amsterdam: EMA; status “Refused”. https://www.ema.europa.eu/en/medicines/human/EPAR/gamifant
90. U.S. Food and Drug Administration. ILARIS (canakinumab) injection: prescribing information. Silver Spring: FDA; Drugs@FDA BLA125319.
91. ClinicalTrials.gov. Managed access programs for MAS825, arumakimig (NCT07889310). Novartis; status “temporarily not available” (update of 22 October 2026). https://clinicaltrials.gov/study/NCT07889310
92. Agência Nacional de Vigilância Sanitária (ANVISA). Open data: registered medicines (DADOS_ABERTOS_MEDICAMENTOS.csv) — anakinra, canaquinumabe and emapalumab [Internet]. Brasília: ANVISA. Available from: https://dados.anvisa.gov.br/dados/DADOS_ABERTOS_MEDICAMENTOS.csv [accessed 31 Oct 2026].
93. Novartis Biociências S.A. Ilaris® (canaquinumabe) solution for injection 150 mg/mL: package insert for healthcare professionals, approved by ANVISA on 26/11/2025. ANVISA registration 1.0068.1068. São Paulo: Novartis; 2025. Available from: https://portal.novartis.com.br/medicamentos/wp-content/uploads/2021/10/Bula-ILARIS-Solucao-injetavel-Medico.pdf
94. ClinicalTrials.gov. Therapeutic use of tadekinig alfa in NLRC4 mutation and XIAP deficiency as open label extension (NCT03512314). AB2 Bio Ltd; phase 3; completed 8 May 2024. https://clinicaltrials.gov/study/NCT03512314
95. ClinicalTrials.gov. A single-patient clinical trial of MAS825 in a patient with XIAP deficiency (NCT06309823). The Hospital for Sick Children; completed 17 January 2024. https://clinicaltrials.gov/study/NCT06309823
96. ClinicalTrials.gov. A study to investigate the safety and efficacy of emapalumab in patients with sJIA or AOSD developing MAS/secondary HLH (NCT03311854). Swedish Orphan Biovitrum; phase 2; completed 19 May 2020. https://clinicaltrials.gov/study/NCT03311854
97. ClinicalTrials.gov. Evaluate efficacy, safety and tolerability, PK and PD of emapalumab in children and adults with MAS in Still's or SLE (NCT05001737). Swedish Orphan Biovitrum; phase 3; completed 4 June 2025. https://clinicaltrials.gov/study/NCT05001737
98. ClinicalTrials.gov. Natural history, pathogenesis, and outcome of autoinflammatory diseases (NCT02974595). National Institute of Allergy and Infectious Diseases; recruiting. https://clinicaltrials.gov/study/NCT02974595
99. ClinicalTrials.gov. A phase II trial to evaluate the clinical efficacy, safety and tolerability of MAS825 in pediatric and adult participants with Still's disease (NCT07203001). Novartis; recruiting since 3 November 2025. https://clinicaltrials.gov/study/NCT07203001
100. Israni M, Alderson E, Mahlaoui N, Obici L, Rossi-Semerano L, Lachmann H, et al. Recommendations for transitioning young people with primary immunodeficiency disorders and autoinflammatory diseases to adult care. J Clin Immunol. 2024;45(1):57. doi:10.1007/s10875-024-01838-y
101. Orphanet. Takenouchi-Kosaki syndrome (ORPHA:487796). Paris: INSERM; 2026. Available from: https://www.orpha.net/en/disease/detail/487796
102. Orphanet. Corneal intraepithelial dyskeratosis-palmoplantar hyperkeratosis-laryngeal dyskeratosis syndrome (ORPHA:352662). Paris: INSERM; 2026. Available from: https://www.orpha.net/en/disease/detail/352662
103. He T, Huang Y, Ling J, Yang J. A new patient with NOCARH syndrome due to CDC42 defect. J Clin Immunol. 2020;40(4):571-5. doi:10.1007/s10875-020-00786-7
104. European Reference Network on Rare Immunodeficiency, Autoinflammatory and Autoimmune Diseases (ERN RITA). https://ern-rita.org/
105. Autoinflammatory Alliance. Autoinflammatory Search: the Systemic Autoinflammatory Diseases (SAID) database. https://www.autoinflammatory-search.org/