Proteasome-Associated Autoinflammatory Syndromes (CANDLE-PRAAS)

Type I interferonopathies due to proteasome and immunoproteasome dysfunction: CANDLE, Nakajo-Nishimura syndrome, JMP syndrome, PRAAS1–6 and forms with immunodeficiency

Literature and regulatory information updated: 31 October 2026.

0. Initial Reference Table

1. Introduction

1.1. Founding descriptions

The first recognisable description of this disease is Japanese. In 1939, Nakajo described two siblings, children of consanguineous parents, with erythema nodosum, elongated and thickened fingers, wasting and cardiomegaly with heart failure, and called the condition «secondary hypertrophic osteoperiostosis with pernio» [1]. In 1950, Nishimura and colleagues added similar cases, and the Japanese literature came to use the eponym Nakajo-Nishimura syndrome for a disease with onset in early childhood, pernio-like rash, high periodic fever, erythema nodosum-like lesions and myositis, followed by lipoatrophy and progressive contractures predominantly in the upper half of the body, which give a characteristic facies [1, 2]. Other Japanese reports used the designation «hereditary lipo-muscular atrophy with joint contractures, skin eruptions and hypergammaglobulinaemia» [2, 40].

In 2010 two Western descriptions appeared independently. Torrelo and colleagues gathered four children followed in Madrid and Chicago, two of them sisters, with early-onset recurrent fever, annular violaceous plaques, persistent violaceous eyelid oedema, low weight and short stature, lipodystrophy, hepatomegaly and various visceral inflammatory manifestations, with chronic anaemia and raised acute-phase reactants and transaminases; biopsy of the lesions showed atypical mononuclear infiltrates of myeloid lineage mixed with mature neutrophils, and the authors proposed the acronym CANDLE [41]. One of the four patients died in an episode of «carditis» without autopsy [2]. In the same year, Garg and colleagues described three adults aged 26 to 34 years, from two pedigrees, one from Portugal (Medical Genetics Department of the Hospital de Santa Maria, Lisbon) and the other from Mexico (Monterrey), with lipodystrophy beginning in childhood, muscle atrophy, severe joint contractures, erythematous skin lesions and microcytic anaemia, and called it JMP syndrome [40].

1.2. Molecular discovery

Also in 2010, homozygosity mapping in the two JMP families identified a single shared homozygous region of 2.4 Mb at 6p21 and, within it, the PSMB8 variant c.224C>T, p.(Thr75Met), homozygous in the three patients and heterozygous in the available parents [42]. In 2011, two Japanese groups identified the PSMB8 variant p.(Gly201Val) in the homozygous state in patients with Nakajo-Nishimura syndrome (described as G197V by Kitamura and colleagues, according to the numbering of another protein isoform) [27, 43, 44]. In 2012, Liu and colleagues confirmed that most CANDLE patients had biallelic PSMB8 variants, but noted that some had only one variant or none [45]. This residue of cases was explained in 2015, when Brehm and colleagues found variants in PSMA3, PSMB4, PSMB9 and POMP and established digenic inheritance, that is, the combination of a variant in an inducible subunit with another in a constitutive subunit, and dominant inheritance due to POMP [23].

The list of genes kept growing: POMP with truncating variants that escape nonsense-mediated decay (2018) [46], PSMG2, a chaperone for α-ring assembly (2019) [47], PSMB10 in the compound heterozygous or homozygous state (2020) [48, 49], PSMB9 with a dominant de novo variant associated with immunodeficiency (2021) [50], PSMB10 with dominant de novo variants causing severe combined immunodeficiency with features of Omenn syndrome (2024) [51], and monoallelic PSMB8 with a dominant-negative effect (2026) [13, 52].

1.3. Current nomenclature and classifications

CANDLE, JMP and Nakajo-Nishimura designate the same disease, and the designations PRAAS (proteasome-associated autoinflammatory syndrome) and ALDD have been proposed to bring them together [2, 11]. OMIM groups them in the PRAAS1 entry (256040, «and digenic forms») and numbers the remaining types by gene: PRAAS2 (POMP), PRAAS3 (PSMB4), PRAAS4 (PSMG2), PRAAS5 (PSMB10) and PRAAS6 (PSMB9); for Japanese patients it recommends the name Nakajo-Nishimura syndrome, in recognition of the original authors [1]. Orphanet uses a parent entity (ORPHA:324977) with three clinical subtypes — CANDLE, Nakajo-Nishimura and JMP — and includes the disease in the groups «type I interferonopathy» and «type I interferonopathy of childhood» [24]. The 2024 classification of the International Union of Immunological Societies (IUIS) lists «PRAAS-CANDLE» due to PSMB8 (autosomal recessive and dominant inheritance), PSMG2, PSMB10, PSMB9, PSMB4 and PRAID due to POMP among the type I interferonopathies of the autoinflammatory diseases, separates «PSMB9 deficiency (G156D)» as a dominant-negative form with immunodeficiency and places Omenn syndrome due to PSMB10 among the combined immunodeficiencies [3]. The previous classification, cited in the Oxford Desk Reference, defined autoinflammatory diseases (group VII) as diseases of abnormally increased inflammation mediated mainly by cells and molecules of innate immunity [4]. Kanazawa and colleagues proposed the term PRAAS with immunodeficiency (PRAAS-ID) for the dominant forms with infections, lymphopenia and hypogammaglobulinaemia, which Wijngaard and colleagues extended to monoallelic PSMB8 [13, 50].

The concept of type I interferonopathy, coined on the basis of Aicardi-Goutières syndrome, designates diseases in which constitutive activation of the type I interferon pathway and overexpression of interferon-stimulated genes are the cause of disease [53]. Autoinflammatory interferonopathies are defined as immune-dysregulatory diseases of monogenic or digenic cause with an interferon signature in peripheral blood and variable systemic inflammation; among them, CANDLE/PRAAS, SAVI (STING-associated vasculopathy with onset in infancy) and Aicardi-Goutières are the most frequent [5].

1.4. Definition of the entities within scope

The class brings together the proteasome diseases with an autoinflammatory phenotype. Classic PRAAS (PRAAS1, PRAAS3, PRAAS4, PRAAS5 and the digenic forms) is characterised by onset in the first weeks of life, fever, atypical neutrophilic dermatosis, panniculitis with lipodystrophy, myositis and contractures, anaemia and an interferon signature, with raised C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) [1, 5, 7]. The dominant forms (POMP, PSMB9 p.(Gly156Asp), dominant-negative PSMB8 and PSMB10) add to the autoinflammatory phenotype a combined immunodeficiency of variable severity, which at the PSMB10 extreme dominates the picture [13, 46, 50, 51]. Two neighbouring entities are covered in subsections of their own, because they share genes or mechanisms with the class without sharing the autoinflammatory phenotype: KLICK syndrome, an autosomal recessive skin disease allelic to PRAAS2, and the neurodevelopmental proteasomopathies (PSMD12, PSMC3, PSMC5 and rarer genes), with a neurological and malformative phenotype, a laboratory interferon signature and rare clinical autoinflammation [11, 14, 29]. Aicardi-Goutières syndrome enters the differential diagnosis.

1.5. Status of the reference sources

None of the reference textbooks consulted has a section dedicated to these diseases. The Saudubray handbook includes PSMB8 variants in two diagnostic orientation tables, in the list of causes of inflammatory syndrome and recurrent fever («erythema nodosum, muscle weakness and lipodystrophy») and in the list of genetic causes of skin nodules [54]. There is no GeneReviews chapter. The reference guideline is the EULAR/ACR 2021 points to consider, developed through a systematic literature review up to September 2020, two Delphi questionnaires and two consensus meetings, with four overarching principles and 17 points to consider, each with an Oxford level of evidence and a grade of recommendation by disease [5, 55].

2. Molecular Genetics and Pathophysiology

2.1. The proteasome and the immunoproteasome

The proteasome is the main non-lysosomal protein degradation system. The 20S core particle is a cylinder of four rings of seven subunits: two outer α rings (α1 to α7, encoded by PSMA1 to PSMA7) and two inner β rings (β1 to β7, PSMB1 to PSMB7); the 19S regulatory particle recognises ubiquitin-tagged proteins, unfolds them and feeds them into the 20S, forming the 26S proteasome [2, 49]. Three β subunits are catalytic: β1 (caspase-like activity), β2 (trypsin-like) and β5 (chymotrypsin-like) [2]. When cells are stimulated by interferons, and constitutively in haematopoietic cells, these subunits are replaced by the inducible forms β1i (PSMB9, LMP2), β2i (PSMB10, MECL-1) and β5i (PSMB8, LMP7), and the 19S may be replaced by the PA28 activator, forming the immunoproteasome, which alters cleavage preference, generates the peptides presented by major histocompatibility complex (MHC) class I molecules and accelerates the removal of oxidised proteins during inflammation [2, 11]. Assembly depends on chaperones: the PAC1/PAC2 heterodimer (PSMG1/PSMG2) initiates the α ring by binding the α5 and α7 subunits, and POMP (UMP1) binds the formed α ring to incorporate the β subunits sequentially, being degraded by the proteasome itself at the end of maturation; the catalytic β subunits are synthesised as pro-forms whose autocatalytic cleavage releases the active-site threonine [23, 46, 47].

2.2. From proteasome dysfunction to type I interferon

The common mechanism is a vicious circle. Proteasome dysfunction prevents the removal of ubiquitinated and oxidised proteins, which accumulate and cause proteotoxic stress; this stress induces type I interferon, which in turn increases the load of proteins to be degraded and the need for the immunoproteasome, which the cell cannot meet [2, 11]. Cold, physical or psychological stress and trivial infections increase demand and trigger inflammatory flares that can affect any organ [2]. Brehm and colleagues showed that patients' haematopoietic and non-haematopoietic cells have a type I but not type II interferon signature, whichever gene is affected, and that chemical inhibition of the proteasome or silencing of its subunits in normal cells reproduces interferon induction [23]. Patients have very high serum levels of interferon-induced protein 10 (IP-10/CXCL10), moderate elevation of MCP-1, IL-6 and IL-1 receptor antagonist, and STAT1 activation in monocytes [45]; in JMP patients, IL-6 was 7 to 19 times and IFN-γ 2.8 to 3.5 times above normal [42]. The link between proteotoxic stress and interferon runs through the unfolded protein response (UPR) and the integrated stress response: in dominant-negative PSMB8, blockade of the kinases PKR and GCN2 reduced the interferon signature of patients' T cells [52, 56, 57]. Plasmacytoid dendritic cells, the main producers of type I interferon [58], appear in aggregates in the skin lesions; rapid mobilisation of myeloid cells from the bone marrow explains the «atypical» infiltrate of immature cells [2]. Type I interferon is toxic to adipocytes, as suggested by the lipoatrophic lobular panniculitis described with IFN-β injections, which contributes to the lipodystrophy [2]. In a mouse model carrying the Japanese founder variant in the homozygous state, 20S activity was reduced in the spleen, lifespan was shortened and adipocytes were shrunken with macrophage infiltration [59]. In brain organoids derived from CANDLE patients, neuronal maturation was impaired by an excess of polyamines, resulting from the accumulation of undegraded ornithine decarboxylase; polyamines were also raised in the cerebrospinal fluid (CSF) of patients [60].

2.3. PSMB8 (PRAAS1)

PSMB8 (6p21.32) encodes β5i, the chymotrypsin-like catalytic subunit of the immunoproteasome [11, 44]. Biallelic variants cause PRAAS1, of autosomal recessive inheritance; it is the most frequent cause of CANDLE/PRAAS [1, 5]. The mechanisms differ between variants: p.(Thr75Met) alters the tertiary structure and selectively reduces chymotrypsin-like activity (it increases by 1.2 Å the distance between catalytic threonine 73 and glycine 119), also affecting assembly; p.(Gly201Val) disrupts the interface with β4, prevents incorporation of β5i, reduces the 20S and all three proteolytic activities and causes ubiquitinated proteins to accumulate; p.(Ala92Thr), p.(Lys105Gln) and p.(Met117Val) (published protein designation) reduce chymotrypsin-like activity, and p.(Lys105Gln) also prevents complete cleavage of the propeptide; p.(Cys135Ter) leads to absence of β5i and reduction of all three activities [2, 11, 27, 42]. Structural studies in yeast with the equivalent substitutions showed that the homologue of p.(Thr75Met) alters the substrate-binding channel with a moderate biogenesis defect, whereas the homologue of p.(Gly201Val) causes rearrangements that severely disrupt assembly [61].

Monoallelic variants with a dominant-negative effect were recognised in 2026: seven individuals from five families, with five distinct variants, presented with neonatal immunodeficiency with recurrent infections, B lymphopenia and hypogammaglobulinaemia requiring immunoglobulin, inflammation of variable severity (enteropathy, hepatitis, myositis, inflammatory lung disease) and leucocyte vacuolisation in blood and bone marrow; the mutant subunit is inefficiently incorporated into the complexes, compromises immunoproteasome assembly and triggers cellular stress [13]. The de novo variant p.(Gly209Arg), also described in a newborn with a response to baricitinib [62], causes steric hindrance of β5i propeptide processing, arrest of 20S maturation, protein aggregation and mitochondrial dysfunction, with severe disease [52]. A single PSMB8 variant in a patient with this phenotype cannot be read as carrier status [11]. At the time of writing OMIM records only autosomal recessive inheritance for PSMB8, whereas IUIS and GenCC (one AD submission) already recognise both patterns [3, 11, 44].

2.4. Digenic inheritance and constitutive proteasome genes: PSMA3, PSMB4 and PSMB9 (PRAAS3)

PSMB4 (1q21.3) encodes β7, a non-catalytic structural subunit important for assembly and stabilisation; PSMA3 (14q23.1) encodes α7 [2, 11]. In the Brehm series, one patient was compound heterozygous for PSMB4 variants, six patients from four families were double heterozygotes (one variant in an inducible subunit and another in a constitutive subunit) and one had a POMP variant [23]. The combinations described include PSMA3/PSMB8 (two patients with paternal p.(Thr75Met) and, respectively, de novo PSMA3 c.697_699del p.(Arg233del) and maternally inherited c.404+2T>G), PSMB9/PSMB4 and PSMB8/PSMB4 (two Irish siblings with PSMB8 p.(Lys105Gln) and PSMB4 p.(Tyr222Ter), each parent carrying one) [23, 44]. The functional consequences add up: PSMB8/PSMA3 compromises all three proteolytic activities, PSMB9/PSMB4 reduces caspase-like activity and compound heterozygosity for PSMB4 severely compromises assembly [2]. OMIM designates this form PRAAS3 (617591), with provisional mapping («?») because it rests on a single biallelic patient [7]. Other oligogenic combinations have been described: paternal PSMA5 and PSMC5 with a maternal PSMB8 variant, and a digenic form with PSMA6 in 2026 [49, 63]. A boy with two PSMB4 variants had resistant cutaneous vasculitis and developed only mild lipodystrophy after transplantation [64]. An isolated PSMB4 variant is not diagnostic, but may be half of a digenic genotype [11].

2.5. PSMB9 (PRAAS6 and digenic participation)

PSMB9 encodes β1i [11]. The de novo variant c.467G>A p.(Gly156Asp) was described in two unrelated Japanese infants with neonatal autoinflammation, myositis, liver dysfunction, splenomegaly, basal ganglia calcification, pulmonary hypertension and immunodeficiency; it disrupts the interaction between the β rings required for 20S formation and also reduces PSMB8 and PSMB10, with almost no ubiquitin accumulation; the heterozygous mouse model reproduces the proteasome defect and the immunodeficiency [50, 65]. OMIM designates it PRAAS6 (620796, AD) [10]. The variant p.(Gly165Asp), c.494G>A, described by Brehm in a digenic context, is a different variant in the same transcript, as normalisation confirms (see 2.9); the notation «c.494G>A (p.Gly156Asp)» that appears in a later article mixes the two [23, 51, 66].

2.6. POMP (PRAAS2) and the allelic disorder KLICK

Heterozygous de novo frameshift variants in the penultimate exon of POMP escape nonsense-mediated decay (NMD), produce a truncated protein that disrupts proteasome assembly and cause PRAAS2, of autosomal dominant inheritance — a mechanism the authors had not encountered in any other immunodeficiency or autoinflammatory disease [6, 46]. The phenotype combines early combined immunodeficiency, inflammatory neutrophilic dermatosis and autoimmunity; the first patient, a boy of Palestinian parents described by Mégarbané in 2002, had recurrent fever, arthralgia, periorbital oedema, nodules with leucocytoclastic vasculitis, short stature and dysmorphic features [2, 6]. POMP is also the origin of an allelic disorder without autoinflammation: homozygosity for a single-nucleotide deletion in the 5' untranslated region (c.-95delC in the original description; NM_015932.5:c.-95del in ClinVar; NC_000013.11:g.28659090del after normalisation), found in all 12 patients from 8 European families in the founding study, alters the balance of transcripts and causes KLICK, a keratoderma with ichthyosis, constricting digital bands and linear keratotic papules in the flexures, of autosomal recessive inheritance [14, 33]. An analysis limited to coding regions does not detect the KLICK variant, and a filter that treats late POMP frameshifts as simple null alleles misclassifies PRAAS2 variants [11].

The mechanism of KLICK syndrome differs from that of PRAAS2. The deletion lies 14 base pairs upstream of the annotated transcription start site, in a 19-nucleotide motif conserved among mammals, and is included only in POMP transcripts with a long 5' untranslated region, initiated at alternative sites [33]. In patients' keratinocytes, these long transcripts increased 29-fold in differentiating cells and came to represent 83% of POMP transcripts, versus 2.6% in controls, without a change in the total amount of messenger RNA; because the long 5' region forms a stable secondary structure that reduces translation, the result is less POMP protein in the suprabasal layers of the epidermis [33]. In normal skin, POMP stains uniformly from the basal to the granular layer; in patients' skin, staining is strong in the basal layer and decreases towards the surface, with abnormal distribution of the proteasome α7 and β5 subunits and an increase in the endoplasmic reticulum stress marker CHOP in the granular layer [33]. Silencing of POMP by RNA interference in keratinocytes reduced proteasome subunits and activated the unfolded protein response, which supports a model of proteasome insufficiency restricted to one phase of epidermal differentiation, with hyperkeratosis and hypergranulosis [28]. The inclusion of the syndrome among the autoinflammatory keratinisation diseases has been proposed, but there is no demonstration of innate immune activation or of a type I interferon response in patients [28]. Haplotype analysis with microsatellite markers around POMP in the eight probands revealed at least five different haplotypes, indicating a recurrent variant rather than a single founder allele [33]. All patients described up to 2020 had the same deletion in the homozygous state, and the 2017 French series described it as a hotspot mutation [28, 67].

2.7. PSMG2 (PRAAS4)

PSMG2 encodes PAC2. The index patient was compound heterozygous for NM_020232.5:c.666_667del p.(Tyr223SerfsTer2), inherited from the father, and c.675T>G p.(Asn225Lys), inherited from the mother, both absent from population databases; her fibroblasts had reduced proteolytic activities, accumulation of ubiquitin conjugates, reduced PAC1 and PAC2 and fewer assembled proteasomes [47]. OMIM records it as PRAAS4 (619183, AR, provisional) [8]. A novel PSMG2 deletion has been described in a child with disseminated bacillus Calmette-Guérin (BCG) disease [68].

2.8. PSMB10 (PRAAS5 and immunodeficiency 121)

PSMB10 encodes β2i and causes two allelic diseases with opposite inheritance [11]. Biallelic variants cause PRAAS5 (619175): the first patient, the daughter of consanguineous Algerian parents, had an annular rash from the seventh day of life, fever, faltering growth, hepatosplenomegaly, a wasted face and long thin fingers, and was homozygous for c.41T>C p.(Phe14Ser), a propeptide variant that prevents cleavage of the PSMB10 pro-form [9, 48, 66, 69]. Three unrelated Brazilian boys shared c.40_42del p.(Phe14del) on one allele, inherited from the mother in all, with a different variant on the other; the authors interpret it as a probable founder variant in Brazilian patients of Indigenous ancestry; the variants p.(Phe14del) and p.(Gly167Asp) (published protein designation) are incorporated into precursors but not into mature 20S and 26S complexes [49, 69]. The heterozygous de novo missense variants c.166G>C p.(Asp56His) and c.601G>A or c.601G>C p.(Gly201Arg) cause T-B-NK± severe combined immunodeficiency with features of Omenn syndrome (immunodeficiency 121 with autoinflammation, 620807), through a dominant-negative effect that lowers PSMB9 and PSMB10 and compromises thymic selection [12, 51]. In 2026 the de novo variants p.(Asp205Ala) and p.(Ser208Phe) (published protein designations) were added, with immunodeficiency and severe liver disease, and a splicing variant (NM_002801.4:c.56+1G>A) that segregates dominantly in a Turkish family with non-classic autoinflammatory disease and interferon signature intensity proportional to severity [70, 71].

2.9. Normalised variants

The original descriptions of PSMB4, PSMA3 and POMP were shifted to the 3' position by normalisation, as HGVS nomenclature requires; the correspondences in the table come from VariantValidator and OMIM, without manual conversion [44, 66]. The KLICK deletion lies upstream of the first nucleotide of the POMP MANE transcript (NM_015932.6), so it has no c. description on that transcript: VariantValidator rejects the c.-95 notation because it falls outside the reference sequence and returns the genomic description, and ClinVar keeps it on version NM_015932.5 [66, 72]. The PSMC3 variants, described in the source on NM_002804.4, were reprojected onto the MANE transcript NM_002804.5 from the genomic coordinate, without a change in numbering [66].

2.10. Genotype-phenotype correlations, penetrance and modifiers

The classic recessive forms share the cutaneous-lipodystrophic picture and differ mainly in severity and accessory features: Nakajo-Nishimura syndrome as described in Japan combines higher fever (>40 °C), erythema nodosum, hypergammaglobulinaemia and elongated fingers, whereas the JMP patients from Portugal and Mexico had seizures and anaemia that the Japanese patients did not; hierarchical cluster analysis separated the two groups despite the similarities [40, 43]. The dominant forms (POMP, PSMB9 p.(Gly156Asp), dominant-negative PSMB8 and PSMB10) are characterised by associated immunodeficiency — B or T lymphopenia, hypogammaglobulinaemia, a CD8 maturation defect, opportunistic infections — and may lack lipoatrophy [10, 13, 46]. Prolonged normalisation of the interferon signature on Janus kinase inhibitors has been observed only in patients with recessive PSMB8 disease [76]. Penetrance in the recessive forms is complete in the genotypes described; heterozygous parents are healthy [42, 47]. In the Turkish family with PSMB10 c.56+1G>A, expressivity varied within the family and tracked the intensity of the interferon response [71]. Additional variants may modulate the severity of infections in patients with POMP variants [5].

2.11. Allelic disorders and neurodevelopmental proteasomopathies

Besides KLICK (POMP, section 2.6) and immunodeficiency 121 (PSMB10, section 2.8), other proteasome genes cause neurodevelopmental diseases with type I interferon induction and without a CANDLE phenotype. In a 2023 review, 46 lesions in 78 individuals defined six entities in four subunits (PSMB1/β6, PSMC1/Rpt2, PSMC3/Rpt5 and PSMD12/Rpn5); about 85% of patients had heterozygous PSMD12 or PSMC3 variants, and the recessive forms, due to biallelic PSMB1, PSMC1 or PSMC3 variants, accounted for 8 of the 78 patients [29]. PSMD11, with loss of function in 10 unrelated children, and PSMC5, with 44 individuals, were added later [22, 30]. Unlike PRAAS, whose affected subunits all belong to the 20S core particle or its assembly chaperones and which as a rule require two alleles or two genes, the neurodevelopmental proteasomopathies are almost all dominant, due to de novo variants, and mainly affect the 19S regulatory particle; among the lesions described, only the PSMB1 variant p.Tyr103His (published protein designation) affects the 20S particle [29]. One proposed explanation for neuronal vulnerability is the enrichment of synapses in free 19S particles with proteasome-independent functions [29].

PSMD12 encodes Rpn5, a non-ATPase subunit of the 19S lid with a structural scaffolding function. Stankiewicz-Isidor syndrome results from haploinsufficiency: of the first 10 patients, 6 had de novo 17q24.2 deletions of 0.62 to 4 Mb, with different breakpoints in each case, and 4 had de novo truncating or splicing variants; one patient's cells had less full-length PSMD12 and accumulated high-molecular-weight ubiquitinated proteins, without loss of proteasome catalytic activity, and disruption of the zebrafish orthologue produced microcephaly, reduced convolution of the renal tubules and craniofacial anomalies [16, 34]. In the series of 24 additional patients, with truncating variants or deletions, patients' cells and CRISPR/Cas9-edited clones showed remodelling of the mTORC1 and mitophagy pathways, with induction of type I interferon-stimulated genes [77]. In 2018 the first familial transmission was described, in a family with six affected siblings and an affected father carrying a nonsense variant, together with a de novo case with another nonsense variant and periventricular nodular heterotopia [78]; in 2025, two families with six patients, four of them adults, carrying two truncating variants, showed complete penetrance with variable expressivity within and between families [79].

PSMC3 encodes Rpt5, one of the six AAA ATPases of the base of the 19S particle, which unfold the substrate and translocate it into the proteolytic chamber. In the dominant form (Ebstein-Bézieau syndrome), 15 de novo missense variants were found in 23 unrelated patients, clustered in the AAA domain in two regions: one around the recurrent variant p.(Arg304Trp), present in four patients, and another more N-terminal, with p.(Ile261Thr) in six patients [73]. Structural modelling placed five of the affected residues in loops adjacent to the substrate tunnel; four of the variants reduced the protein expressed in neuroblastoma cells by about 90%, through increased degradation or reduced translation, but the T cells of three patients had normal amounts of PSMC3, so the mechanism is not haploinsufficiency but disruption of incorporation into 26S complexes and of substrate translocation, with reduced chymotrypsin-like activity [73]. In patients' T cells, expression of interferon-stimulated genes was blocked by inhibition of the kinase PKR, an intracellular stress sensor [73]. The recessive form was described in three patients from an extended consanguineous family from the Kayseri region of Türkiye, homozygous for a deep intronic variant that leads to inclusion of a 114 bp cryptic exon with a premature stop codon; in fibroblasts, proteasome content and activity were preserved, but there was accumulation of ubiquitinated proteins, constitutive activation of the TCF11/Nrf1 proteasome recovery pathway and inability to compensate for chemical inhibition of the proteasome, and the zebrafish model reproduced the inner ear anomalies and cataract [74]. OMIM classifies this recessive form as provisional [18].

PSMC5 encodes Rpt6, another ATPase of the base of the 19S particle. A first series identified p.(Pro320Arg) in six individuals, as well as p.(Arg325Trp), a missense variant at Gln160 and a nonsense variant at Gln69, and showed that p.(Pro320Arg) weakens the binding between the 19S and 20S particles [75]. The 2025 cohort gathered 26 variants in 44 individuals (42 unrelated probands and two relatives): 33 de novo variants, six of unknown inheritance, two inherited from an affected mother and one patient with a biallelic variant (NM_002805.6:c.620C>T p.(Thr207Met) in the homozygous state); 19 of the 26 variants were missense and 16 lay in the AAA domain, with the recurrent p.(Pro320Arg) (6 of 44) and p.(Arg325Trp) (9 of 44) near the end of the domain [30]. Loss of PSMC5 function compromised proteasome activity, with protein aggregation, altered mitochondrial homeostasis and lipid metabolism and spontaneous type I interferon activation in patients' T cells, dependent on the integrated stress response [30]. In a mother and son with the variant NM_002805.6:c.773G>A p.(Arg258Gln), the mother had inherited it from her father, who was not clinically assessed [19].

PSMD11 encodes Rpn6, a lid subunit; its loss-of-function variants compromise 26S proteasome assembly and induce a persistent PKR-mediated interferon signature [22]. The rare recessive forms include the disease due to biallelic PSMB1 variants (two girls with microcephaly and severe intellectual disability) and Birk-Aharoni syndrome due to biallelic PSMC1 variants (three patients with spastic tetraplegia and chorea), both provisional in OMIM; heterozygous PSMC1 deletions have been associated with a mild developmental and language disorder in three patients [20, 21, 29]. Other proteasome genes (PSMA7, PSMD10, PSMD14) are candidates owing to enrichment of de novo variants in neurodevelopmental disorder cohorts or inclusion in microdeletions, without functional validation [29].

These diseases share with PRAAS the persistent type I interferon signature in peripheral blood, but most patients have no clinical signs of inflammation, a discordance also observed in Aicardi-Goutières syndrome and in trisomy 21 [29]. IUIS lists PSMD12 among the autoinflammatory conditions with neurodevelopmental disease with recessive inheritance, whereas OMIM records it as dominant [3, 16].

3. Clinical Spectrum and Manifestations by System

3.1. Reference cohorts

The clinical data come from small series. The founding CANDLE description gathered 4 children [41]; the JMP series, 3 adults from 2 families [40]; the first genetic CANDLE series, 9 patients from 8 families (3 Spanish, 3 Hispanic, 1 Ashkenazi Jewish and 2 white American), followed at five centres in Spain, Israel and the United States [45]; the series that established digenic inheritance, 8 patients [23]. The National Institutes of Health (NIH) expanded access programme included, between October 2011 and February 2017, 10 CANDLE patients among 18 patients with interferonopathies, with a mean age at entry of 12.5 years (1.2–24.1) [31]. A 2026 systematic review gathered 46 patients treated with Janus kinase inhibitors, with a median age of 4.5 years: rash in 100%, fever in 91.3%, lipodystrophy in 73.9%, faltering growth in 56.5%, arthralgia or arthritis in 43.4% and panniculitis in 39.1%; 73.9% had previously received corticosteroids [80]. These percentages describe published and treated patients, not a natural history cohort.

3.2. Onset and course

The disease begins in the first weeks of life: in the Liu series, most patients presented in the first 2 to 4 weeks and all by 6 months; in the NIH programme, all CANDLE patients developed symptoms in the first 2.5 weeks [31, 45]. Late presentation is rare, described up to 14 years [5]. The most common initial sign is fever or a temperature rise below 38.3 °C, daily or almost daily, with little effect on general condition; cold can trigger the fever and the lesions [2]. Skin lesions are the first clinical sign and persist throughout life, less visible after puberty; lipodystrophy begins in early childhood and is usually established before puberty; disabling joint manifestations appear in the long term [2]. Acute inflammatory flares, spontaneous or triggered, are superimposed on this chronic course and can affect almost any organ [2, 11]. In Nakajo-Nishimura syndrome, fever is periodic and high (>40 °C), with erythema nodosum from the first month to 3 years of age [43]. In JMP patients, skin lesions appeared at 3, 12 and 14 years [40].

3.3. Skin and subcutaneous tissue

There are three types of lesion [2]. Acral pernio-like lesions — oedematous plaques of intense red or violaceous colour on the nose, ears, fingers and toes — appear in newborns and infants and are not usual later; cold can trigger them, but often there is no history of exposure [2]. Annular plaques begin in infancy: oedematous erythematous or purpuric lesions with a raised border and a flat purpuric centre, appearing in crops or singly and regressing within days to weeks, leaving a purpuric macule, so that active and residual lesions coexist [2, 11]. Persistent erythematous-violaceous oedema of the eyelids and, less often, of the perioral region appears in infancy and is less visible after puberty [2, 41]. The EULAR/ACR points to consider highlight nodular rashes and violaceous annular plaques as an indication to investigate for CANDLE/PRAAS, and panniculitis, especially neutrophilic, and panniculitis-induced lipodystrophy as hallmarks of the disease [5]. Histology shows a perivascular and interstitial infiltrate of the papillary and reticular dermis extending into the hypodermis as lobular panniculitis, composed mainly of mononuclear cells with large irregular nuclei — which may suggest malignancy —, with mature neutrophils, some eosinophils and lymphocytes, leucocytoclasia without fibrinoid necrosis; immunohistochemistry reveals myeloid cells (myeloperoxidase-positive), macrophages (CD163 and CD68) and aggregates of CD123-positive plasmacytoid dendritic cells [2]. In the Brazilian patients with PSMB10 variants, biopsy showed neutrophilic dermatosis with dermal mucinosis [9]; in one of the JMP patients, biopsy showed lymphocytic panniculitis [40]. Other findings include hypertrichosis, acanthosis nigricans, hirsutism and alopecia areata [2, 45].

3.4. Lipodystrophy, growth and metabolism

Loss of subcutaneous fat is seen in most patients before 2 years of age and progresses over years, beginning in the face and extending to the trunk and upper limbs, with relative sparing of the lower limbs [2]. In the face, loss of malar fat with periorbital oedema is characteristic; in adults, eyelid and lip retraction cause pseudoproptosis and dental exposure [2, 40]. Visceral fat, the only remaining store, gives a prominent abdomen, and the internipple distance is increased [2, 23]. In the Mexican patient of the JMP series, absorptiometry showed 17.2% body fat (normal 27.2±5.3%), and whole-body magnetic resonance imaging (MRI) showed loss of fat from the head, neck, thorax and upper limbs with sparing of the lower limbs [40]. Between 40% and 80% of patients develop metabolic syndrome and hepatic steatosis, often in the first decade [5]; in the NIH programme, 6 of 10 CANDLE patients had metabolic syndrome and 7 had hyperlipidaemia at entry [31]. The consensus uses Ford's definition: at least three of five criteria — triglycerides ≥110 mg/dL, HDL cholesterol ≤40 mg/dL, waist circumference ≥ 90th centile, blood glucose ≥110 mg/dL and blood pressure ≥ 90th centile for age, height and sex [5]. The JMP patients had mild hypertriglyceridaemia and very low HDL (19–30 mg/dL), without diabetes or acanthosis [40]. Growth is delayed: 72% of patients in the NIH programme were below the 3rd centile for height and 50% below the 3rd centile for weight, with bone age delayed by a mean of 3.49±3.99 years [31]; glucocorticoids worsen height and bone mineral density and may increase weight [5]. Adult height of the JMP patients was 144 to 153 cm [40].

3.5. Muscle, joints and cartilage

Arthralgia is frequent in childhood without radiological arthritis; there may be swelling of the interphalangeal joints and, over time, hyperextensibility [2]. Most patients develop contractures of the hands and feet, sometimes very disabling [2, 5]. Myositis is common, patchy in distribution and demonstrable by MRI; chronic muscle inflammation contributes to atrophy [2, 5]. In JMP patients, radiographs showed juxta-articular osteopenia, dumbbell-shaped proximal phalanges, flexion deformities and mild acro-osteolysis, and creatine kinase (CK) reached 709 U/L [40]. Recurrent or chronic chondritis of the ears and nose may cause partial loss of the auricle and saddle nose [2].

3.6. Nervous system

Intellectual disability is not usual, but there may be mild developmental delay [2, 5]. Patients have headaches and may have aseptic meningitis or meningoencephalitis with sterile CSF pleocytosis [2, 5]. Basal ganglia calcifications occur, but are less frequent, less severe and later than in Aicardi-Goutières syndrome, and leucoencephalopathy is unusual [5]. Among the JMP patients, there were generalised seizures in two and basal ganglia calcification in one [40]. In the patient with PSMG2 variants, computed tomography (CT) showed faint bilateral basal ganglia calcifications [47].

3.7. Lungs, heart and vessels

Pulmonary hypertension without fibrosis is a potentially fatal and possibly underdiagnosed complication, of insidious onset [5]. Interstitial lung disease is rare in CANDLE/PRAAS, unlike SAVI [5]. In patients with PSMB9 p.(Gly156Asp), pulmonary hypertension was severe, and one needed venoarterial extracorporeal membrane oxygenation [50, 65]. Episodes of carditis can be fatal, as in one of the founding patients, and Nakajo's siblings had cardiomegaly with heart failure; most Japanese patients die of respiratory or heart failure [1, 2, 27].

3.8. Liver, spleen and lymph nodes

Hepatomegaly is frequent and may result from the metabolic disturbance secondary to lipoatrophy; splenomegaly and generalised lymphadenopathy reflect autoinflammatory activity [2]. Transaminases are moderately raised, owing to steatosis, inflammation or myositis [2, 5]. In dominant-negative PSMB8 disease, porto-sinusoidal vascular liver disease has been described, with nodular regenerative hyperplasia in all four patients biopsied, hepatosplenomegaly, abdominal distension and ascites in seven of eight, and one death at 25 years from complications of portal hypertension [52], and in dominant PSMB10 disease, severe liver disease [70].

3.9. Blood, immunity and autoimmunity

Chronic hypochromic microcytic anaemia is typical, sometimes severe enough to require transfusion; it is accompanied by thrombocytosis and neutrophilia [2, 40]. Cytopenias, including autoimmune cytopenias, are more frequent than in SAVI and may accompany disease activity; the patient with PSMG2 variants had autoimmune haemolytic anaemia with haemoglobin of 3.1 g/dL [5, 47]. Antinuclear antibodies appear in up to 42% of patients, and antiphospholipid antibodies have also been described [5]; in the NIH programme, 60% of CANDLE patients had autoantibodies against endothelial antigens, phospholipids, myeloperoxidase, proteinase-3 or nuclear antigens [31]. Hypergammaglobulinaemia is frequent in the Japanese forms and in JMP patients [40, 43]. There is a defect in CD8 T-cell maturation [5]. Infections can trigger macrophage activation syndrome [5].

3.10. Eye, teeth and other organs

Keratitis and episcleritis (sometimes nodular) and conjunctivitis are the ocular manifestations; one of the JMP patients had punctate corneal opacities [2, 5, 40]. Hypodontia and tooth eruption problems may require dental assessment [5]. There are episodes of aseptic epididymitis, parotitis, pneumonitis, nephritis and otitis [2]. In JMP patients, hypogonadism (small testes, female hair pattern, absent thelarche) and subclinical hypothyroidism have been described [40].

3.11. Profiles of the dominant forms and forms with immunodeficiency

In PRAAS2 due to POMP, the two patients of Poli had combined immunodeficiency with severe viral pneumonias (adenovirus, respiratory syncytial virus, parainfluenza), Pneumocystis jirovecii pneumonia, prolonged norovirus, Salmonella bacteraemia and pulmonary mycobacterial infection, with low CD8 T cells and B cells and autoimmunity, and were referred for transplantation [46]. In PRAAS6 due to PSMB9, patients had no lipoatrophy and were positive for polyomavirus [10]. In the dominant-negative form of PSMB8 p.(Gly209Arg), the eight patients (2 to 30 years), with neonatal onset, had fever, rash and systemic inflammation, recurrent infections from early on (viral in all, bacterial in five, including two Pneumocystis jirovecii pneumonias, and fungal in three), anaemia in all, persistent lymphopenia and hypogammaglobulinaemia in seven, treated with replacement immunoglobulin, and thrombocytopenia in six, suggestive of splenic sequestration; two had congenital thyroid agenesis, and one patient died at 5 years of cytomegalovirus pneumonia [52]. In other monoallelic PSMB8 variants, neonatal immunodeficiency, hypogammaglobulinaemia and organ inflammation (enteropathy, hepatitis, myositis, lung disease) predominate, with one patient undergoing lung transplantation [13]. In immunodeficiency 121 due to PSMB10, the picture is one of severe combined immunodeficiency with erythroderma, diarrhoea, alopecia, eosinophilia, opportunistic infections and T and B lymphopenia [12, 51].

3.12. Natural history and treated disease

Before Janus kinase inhibitors, the course was variable: some patients died in inflammatory flares affecting vital organs and others survived for a long time with increasing disability [2]. Of Liu's 9 patients, one died at 14 years [45]; the sister of the Portuguese JMP patient died suddenly at 8 years, and a Japanese patient compared in the JMP series died at 47 years of congestive heart failure [40]. Glucocorticoids control inflammation but not wasting [27]. On baricitinib, half of the CANDLE patients in the NIH programme achieved lasting remission, with recovery of growth and bone mineral density (see section 7) [31]; baricitinib did not improve pre-existing hepatic steatosis or prevent its development in patients with hyperlipidaemia, which points to a direct role of proteasome dysfunction in lipid metabolism [5]. Damage already established is irreversible [5].

3.13. KLICK syndrome

The first description, from 1989, gathered four members of a consanguineous Spanish family with a generalised congenital ichthyosiform dermatosis, a diffuse sclerosing palmoplantar keratoderma with deformities, pseudoainhum (constricting digital bands) and functional limitation, and multiple keratotic papules arranged in symmetrical linear cords in the flexures, with acrosyringeal keratoses; dental changes were inconsistent and there was no systemic, neurological or ophthalmological involvement [28, 81]. In 1997, a 32-year-old man, the son of non-consanguineous parents, with moderate non-bullous ichthyosis from birth, palmoplantar keratoderma with pseudoainhum, sclerosing flexion deformity of the fingers and non-inflamed longitudinal keratotic striae around the wrists, in the elbow creases and behind the knees, without dental, nail, hair or mucosal problems and with normal intellectual development, gave the syndrome its acronym; he improved with oral etretinate [14, 82].

The 12 patients from the 8 families of the genetic study, from Spain, Italy, the Netherlands, Sweden and Norway, shared mild ichthyosis, thickening of the stratum corneum of the hands and feet, hyperkeratotic plaques on the knees, wrists and axillae, circular sclerotic constrictions around the fingers, flexion deformities of the fingers and linear hyperkeratotic papules on the flexor aspects of the wrists, elbows and knees, without evident extracutaneous manifestations [33]. Onset is early, with generalised erythema and fine scaling followed by a diffuse smooth palmoplantar keratoderma, and linear and star-shaped keratoses over the large joints are the most distinctive feature [32]. In the 2025 classification of non-syndromic disorders of epidermal differentiation, the entity came to be called POMP-nEDD, described with generalised or localised skin thickening and erythema, linear keratotic papules in the antecubital and popliteal regions and palmoplantar keratoderma [15]; in the 2009 Sorèze consensus it was listed among the «other forms» of non-syndromic ichthyosis, alongside loricrin keratoderma and erythrokeratodermia variabilis [83]. A patient initially diagnosed with erythrokeratodermia or loricrin keratoderma had diffuse fine white scaling and well-demarcated, symmetrical and fixed erythematous hyperkeratotic plaques on the limbs, an unusual presentation that broadens the phenotype [28, 84].

Histology shows irregular acanthosis, hypergranulosis and hyperkeratosis, sometimes with parakeratosis, and a mild superficial perivascular lymphohistiocytic infiltrate; electron microscopy reveals numerous large, rounded keratohyalin granules in the superficial keratinocytes, a finding that underpinned the initial hypothesis of a defect in the formation of these granules [28, 33, 82]. The disease is lifelong and limited to the skin; morbidity results from functional limitation of the hands due to sclerosis and pseudoainhum and from the impact of the ichthyosis on quality of life [28, 81]. An aggressive cutaneous squamous cell carcinoma has been described in a patient with KLICK, as has a family with KLICK and autosomal recessive woolly hair with hypotrichosis simplex, two genodermatoses caused by variants in two different genes on chromosome 13 [85, 86]. By 2020 about 20 cases had been published [28].

A Brazilian case illustrates the presentation and response to treatment. A 19-year-old white young man, followed in Belo Horizonte, the son of first cousins and with a sister with a similar condition, had had scaling of the whole skin surface since birth, which progressed to generalised ichthyosis with symmetrical, linearly arranged hyperkeratotic papules and plaques in the axillae, knees, elbows and antecubital and popliteal fossae, and a diffuse transgredient palmoplantar hyperkeratosis with tapering of the fingers and constricting bands; hair, nails, mucosae, hearing and psychomotor development were normal [87]. Biopsy of the antecubital region showed orthokeratotic hyperkeratosis, papillomatosis, irregular acanthosis and hypergranulosis with large, numerous, rounded keratohyalin granules [87]. He had taken acitretin from 8 to 18 years of age, with marked improvement; stopping the retinoid was followed by recrudescence of the lesions, which led to the diagnosis, and reintroduction at a low dose again improved the condition within about a month [87]. The diagnosis was clinical and histological, without molecular study of POMP in the report [87]. The figures of the article, consulted through their legends, document the ichthyosis with linear antecubital and popliteal plaques, the transgredient keratoderma with constricting bands and the hypergranulosis [87].

3.14. Neurodevelopmental proteasomopathies

The common picture is developmental delay, present in all patients described, with predominant language delay (97%) and intellectual disability (85%), often accompanied by behavioural changes (66%), especially autism spectrum disorder, and motor delay [29]. Almost all patients (95%) have facial dysmorphism, heterogeneous and without a recognisable clinical pattern, and a substantial proportion have malformations of several systems — brain, heart, kidneys, genital tract and skeleton —, consistent with the ubiquitous expression of the proteasome; hearing loss (37% in the class) is relatively frequent, in line with the role of the proteasome in inner ear development [29]. The frequencies by gene come from internationally recruited cohorts, with variable denominators per trait, and not from population studies.

In Stankiewicz-Isidor syndrome, the four boys with point variants in the initial series all had a genital anomaly (hypospadias, micropenis, cryptorchidism), three had heart disease (septal defects, patent ductus arteriosus), two had renal anomalies (fused kidneys, ureteral duplication) and there was agenesis or hypoplasia of the thumb; among the six patients with deletions, one had seizures, four had renal anomalies, three had syndactyly of the 2nd and 3rd toes and one girl had precocious puberty [16, 34]. The series of 24 patients confirmed pre-axial hand anomalies as a feature of the syndrome and described pernio similar to that of the interferonopathies in two patients [77]. The inconsistent but evocative association of thumb anomalies with sensorineural hearing loss, together with occasional inflammatory signs (pernio, urticarial rash, congenital uveitis), makes the syndrome recognisable in some patients; congenital pancytopenia has also been described [29]. Experience in adults comes from two unrelated families: short stature, acne and obesity established from late childhood in most, altered eating behaviour, preserved sociability and new findings such as ectopic nail (onychoheterotopia), cholesteatoma, oligodontia and an ovarian teratoma, without clinical manifestations of autoinflammation [79].

In the dominant form of PSMC3, the first patient identified was a newborn girl with severe cardiac, gastrointestinal, inflammatory and immune problems; in the others, the picture was neurological and malformative, without regression, with faltering growth in 9 of 19 patients, mainly with feeding difficulties, cardiac malformations including ventricular septal defects, patent ductus arteriosus and pulmonary atresia, orofacial clefts in 2 of 19 and tumours in 2 of 19 (one craniopharyngioma and one neuroblastoma) [73]. Automated facial analysis showed heterogeneous dysmorphic features, similar only among patients with the same variant [73]. The occurrence of the two tumours raised the hypothesis of cancer predisposition, without confirmation [29]. In the recessive form, the three Turkish patients had severe congenital sensorineural deafness and early-onset cataract, developmental delay, ataxia, lower-limb polyneuropathy in two, subcutaneous calcifications at the knees and elbows and depigmentation of leg hair in two [29, 74].

In Yu-Kury syndrome, automated facial analysis suggested a recognisable facial phenotype: in 15 patients tested, 90% of pairwise combinations fell below the similarity threshold [30]. The p.(Arg325Trp) variant is associated with more severe neurodevelopmental impact, heart disease, dysmorphism, hearing loss, visual impairment and feeding difficulties, and variants outside the ATPase domain, in the initial part of the protein, are more often associated with short stature [30]. The mother and son with p.(Arg258Gln) had motor delay, autistic traits and seizures, without reported intellectual disability, which shows the variability of the phenotype [19]. A 14-year-old girl with p.(Pro320Arg) had developmental delay with regression, mild intellectual disability, hypotonia, aggressive behaviour and a small mouth, upturned nose and downslanting palpebral fissures [19, 88]. In the PSMD11 series, intellectual disability and neurodevelopmental delay were associated with recurrent obesity [22].

The natural history is that of a static developmental encephalopathy, without regression reported in the PSMD12 and PSMC3 series; regression has been described in isolated patients with PSMC5 variants [19, 29, 88]. There are no data on mortality or life expectancy, and the course in adulthood has been described only in the families with inherited PSMD12 variants [79].

4. Clinical Recognition and Warning Signs

4.1. When to suspect

In the newborn and infant, suspicion should arise in the presence of daily fever or low-grade fever with good general condition, acral pernio-like lesions without exposure to cold, violaceous or purpuric annular plaques leaving residual macules, persistent violaceous eyelid oedema, panniculitis nodules and raised acute-phase reactants without infection [2, 5]. The EULAR/ACR points to consider recommend immediate assessment — genetics, extent of inflammatory organ involvement and screening for comorbidities — whenever unexplained systemic inflammation (raised CRP or ESR and/or an interferon signature) is associated with rash, lipodystrophy or musculoskeletal, neurological, pulmonary or metabolic findings (level of evidence 4, grade C; level of agreement 9.8±0.7) [5]. In older children and adults, the warning signs are acquired lipodystrophy of the face and upper half of the body with a prominent abdomen, progressive contractures of the hands and feet, myositis, microcytic anaemia without iron deficiency to explain it, hepatosplenomegaly and early metabolic syndrome in a thin patient [2, 40]. The Saudubray handbook places PSMB8 variants among the genetic causes to consider in the presence of an inflammatory syndrome with recurrent fever and of skin nodules, alongside metabolic diseases with an inflammatory phenotype [54].

Skin biopsy may be misinterpreted as leukaemia cutis if the syndrome is not considered [2, 11]. Lack of response to conventional drugs for autoinflammation is itself a clue: colchicine and dapsone are ineffective, tumour necrosis factor (TNF) antagonists do not help and may worsen the disease, and IL-1 blockade generally fails in the interferonopathies [2, 11].

In the dominant forms, suspicion arises from early combined immunodeficiency — opportunistic infections, B lymphopenia, hypogammaglobulinaemia, low CD8 — associated with neutrophilic dermatosis, panniculitis, autoimmunity, cytopenias or vascular liver disease [13, 46, 52]. Omenn syndrome with restricted T-cell receptors and erythroderma in an infant should lead to the inclusion of PSMB10 in genetic testing [51].

In the neighbouring entities, the signs of suspicion are different. KLICK syndrome should be considered in the presence of congenital or early-childhood-onset ichthyosis with sclerosing palmoplantar keratoderma, constricting digital bands or flexion deformities of the fingers and linearly arranged keratotic papules in the flexural creases, without fever or systemic disease [32, 33]. A neurodevelopmental proteasomopathy should be considered in a child with developmental and language delay or intellectual disability associated with cardiac, renal, genital or skeletal malformations, particularly thumb anomalies with sensorineural hearing loss, and especially if there is pernio or other subtle inflammatory signs; the association of severe congenital deafness and early cataract with developmental delay points to the recessive form of PSMC3 [29, 74].

4.2. Warning and emergency signs

Signs of severity are dyspnoea, exercise intolerance or syncope (pulmonary hypertension), chest pain or signs of heart failure (carditis), severe headache with meningeal signs (aseptic meningitis), severe anaemia or haemolysis with a positive Coombs test, fever with cytopenias, hyperferritinaemia and hepatosplenomegaly (macrophage activation syndrome) and any inflammatory flare of a vital organ, which requires systemic corticosteroids and organ-directed treatment [2, 5, 47]. On Janus kinase inhibitors, a rise in creatinine with BK viruria or viraemia is a warning sign [31]. Stopping or rapidly reducing these drugs can trigger withdrawal flares [5, 76].

4.3. Newborn screening

There is no newborn screening for classic PRAAS. The dominant PSMB10 forms with severe combined immunodeficiency can be detected by newborn screening for severe combined immunodeficiency through quantification of T-cell receptor excision circles (TREC): two of the six infants in the van der Made series were identified in this way, and the authors recommend looking for PSMB10 variants in positive cases [12, 51]. In the United States, newborn screening coordinated by the Centers for Disease Control and Prevention includes severe combined immunodeficiency, whereas in Europe the number and type of diseases screened vary between countries [89].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Criteria

There are no validated clinical classification criteria. The EULAR/ACR consensus states that the diagnosis of CANDLE/PRAAS requires genetic confirmation, which enables targeted treatment, genetic counselling, screening for complications and prognosis (level 5, grade D for CANDLE; agreement 9.5±1.0) [5]. The clinical picture, skin biopsy and interferon signature raise and support the suspicion, but do not confirm the gene [5, 11]. In clinical studies, activity is measured by the daily symptom score (mean of the symptoms fever, rash, musculoskeletal pain, headache and fatigue), with the disease considered controlled below 0.5 and in remission below 0.15 with normal CRP and without corticosteroids [31, 90].

5.2. Laboratory tests

The changes are non-specific: raised ESR and CRP, thrombocytosis, chronic hypochromic microcytic anaemia, moderately raised transaminases, raised triglycerides and, less often, raised CK and aldolase; autoantibodies are usually negative, although some patients have antinuclear antibodies; immunoglobulins are normal in the classic biallelic forms, and bone marrow and lymph nodes show only reactive changes [2]. Unlike Aicardi-Goutières syndrome, CRP and ESR are typically raised in CANDLE/PRAAS and SAVI [5]. The Brazilian patients with PSMB10 variants also had raised ferritin and hypertriglyceridaemia [9]. In the dominant forms, hypogammaglobulinaemia, B lymphopenia and leucocyte vacuolisation in blood and bone marrow are orienting findings [13]. The consensus recommends a basic immunodeficiency work-up in all patients — history of infections, lymphocyte subsets and immunoglobulins (level 4, grade C) [5].

5.3. Interferon signature

A chronically raised interferon signature in peripheral blood is common to the three main autoinflammatory interferonopathies [5]. It is measured by a 28-gene score with NanoString technology or by reverse transcription followed by quantitative PCR of gene subsets, and should be repeated to demonstrate chronic elevation [5, 91]. The limitations are explicit in the consensus: the score may be negative at the diagnostic stage in mild disease or under glucocorticoids; for most assays there are no sensitivity and specificity data; few centres perform it; there is no internationally standardised methodology or reference interval, so it is interpreted against the laboratory's values; and it is not necessary for diagnosis [5]. In a retrospective study, with a defined cut-off, the score distinguished the interferonopathies from healthy controls and from cryopyrin-associated periodic syndrome with an area under the curve of 0.98 and sensitivity and specificity above 0.8 [5, 92]. By way of illustration, the 28-gene score of the NIH laboratory was −4.8±9.53 in controls, −9.37±10.97 in neonatal-onset multisystem inflammatory disease (NOMID), 738.2±754.4 in CANDLE and 495.6±225.9 in SAVI [47]. The signature also serves to monitor response to treatment and, added to the clinical criteria, to detect flares [76]. Serum IP-10 correlates with the score and with symptoms [31].

5.4. Skin and other tissue biopsy

Biopsy of a lesion that can be safely biopsied has the highest yield and may be sufficiently characteristic for early diagnosis: an infiltrate of immature myeloid mononuclear cells with partial maturation to polymorphonuclear cells and karyorrhexis, lobular panniculitis, myeloperoxidase, CD163, CD68 and CD123 aggregates [2, 5, 11]. The consensus considers that tissue sampling — lesion biopsy or CSF if neurological involvement is suspected — may support the diagnosis (level 4, grade C) [5]. Lumbar puncture is indicated in the presence of headaches [5].

5.5. Imaging

MRI best identifies white and grey matter changes, and CT, more sensitive for calcification, is considered when calcium-sensitive sequences are not available or do not show them; the risk of sedation should be weighed (level 4, grade C) [5]. In CANDLE/PRAAS, baseline brain MRI is useful for assessing complications, but is not a diagnostic prerequisite [5]. Muscle MRI shows patchy myositis and fatty infiltration [5, 47]. Echocardiography screens for pulmonary hypertension; low-dose chest CT and pulmonary function tests screen for interstitial disease [5].

5.6. Genetic testing

The consensus recommends next-generation sequencing (targeted panel, exome or genome) rather than gene-by-gene Sanger sequencing; Sanger sequencing remains acceptable in a family with a known variant or when it is the only option [5]. The genes to include are PSMB8, PSMA3, PSMB4, PSMB9, PSMB10, POMP and PSMG2 (level 4, grade C; agreement 9.8±0.6), and all proteasome genes should be assessed in a suggestive phenotype, because new genes continue to be described [5]. The genetic mimics (IKBKG, SAMD9L, RNASEH2A/B/C) should be on the panel [5]. When a single variant is found, a second variant must be sought in another proteasome gene and both parents studied to confirm digenic inheritance [5, 23].

The limitations of the test are concrete. A negative result for PSMB8 does not exclude the disease [11]. Isolated heterozygosity in PSMB8 may be carrier status, half of a digenic genotype or a dominant-negative variant, and only segregation, phenotype and functional studies resolve this [11, 13]. Frameshift variants in the penultimate exon of POMP escape NMD and should not be filtered as benign nulls, and the KLICK variant in the 5' untranslated region escapes coding-only analysis [11, 33, 46]. In the patient with PSMG2 variants, a clinical autoinflammatory disease panel that included the then known CANDLE genes was negative and the diagnosis came only from research exome sequencing [47]. A patient with a suggestive phenotype and no identified variant should be referred to a specialist or research centre, rather than the diagnosis being set aside (level 5, grade D) [5].

The report should state the gene, the versioned reference transcript (for PSMB8, NM_148919.4), the HGVS c. and p. nomenclature, the zygosity, the phase of the variants when there are two (in trans confirmed through the parents), the classification according to the ACMG/AMP criteria and, in digenic genotypes, the combination of the two variants [42, 66]. Variants of uncertain significance require family segregation and, when possible, functional studies.

5.7. Functional confirmation

Functional assays are research tools: proteasome chymotrypsin-like, trypsin-like and caspase-like activities in lymphoblasts or fibroblasts, quantification of assembled complexes by native electrophoresis, accumulation of ubiquitin conjugates, complexome profiling to demonstrate the dominant-negative effect of a monoallelic variant, and interferon score [11, 13, 27, 42]. In the PSMB9 variant p.(Gly156Asp), ubiquitin accumulation was almost undetectable, so a normal result of that assay does not exclude the disease [50].

5.8. Algorithm

5.9. Prenatal and preimplantation diagnosis

Once the causal variants in a family have been identified, prenatal diagnosis by chorionic villus sampling or amniocentesis and preimplantation genetic testing are technically possible, as for other monogenic immune diseases [4]. In the digenic forms, both variants must be known and interpretation must consider the combination of the two genes. The sources consulted do not describe cases of prenatal diagnosis specific to these diseases.

5.10. Diagnosis of KLICK syndrome and the neurodevelopmental proteasomopathies

In KLICK syndrome, the clinical diagnosis rests on the combination of ichthyosis, sclerosing palmoplantar keratoderma with constricting bands and linear keratotic papules in the flexures, and skin biopsy reinforces it through hypergranulosis with large, rounded keratohyalin granules, visible on light microscopy and confirmed by electron microscopy [33, 82, 87]. Confirmation is molecular: sequencing must cover the long 5' untranslated region of POMP, where the deletion lies, which is not detected by an analysis restricted to coding regions; targeted Sanger sequencing, or a genome, solves the problem, and the report should give the genomic description (NC_000013.11:g.28659090del, GRCh38) and the ClinVar reference, because the variant has no description on the MANE transcript [11, 33, 66, 72]. POMP is among the genes for non-syndromic ichthyoses and hereditary palmoplantar keratodermas; when a panel for these diseases is used, the laboratory report should confirm that the gene is included and that the 5' region is covered [32, 33, 83]. As all known patients are homozygous for the same variant, study of the parents confirms the heterozygosity of each and excludes a deletion in trans [33].

In the neurodevelopmental proteasomopathies, the first-line test is the exome or genome, ideally as a trio, with copy-number variant analysis: 6 of the first 10 patients with Stankiewicz-Isidor syndrome had 17q24.2 deletions of 0.62 to 4 Mb that included PSMD12, detectable by array comparative genomic hybridisation [16, 34]. The recessive form of PSMC3 is caused by a deep intronic variant, identified by genome sequencing with homozygosity mapping, and escapes the exome [74]. The de novo missense variants of PSMC3 and PSMC5 were classified by simulation of the ACMG/AMP criteria, which supported pathogenicity for 25 of the 26 PSMC5 variants, and by functional studies in patients' T cells [30, 73]. De novo inheritance is a strong argument, but transmission by parents with a mild phenotype is documented in PSMD12 and PSMC5, so a carrier parent does not exclude pathogenicity and should be examined [30, 78, 79].

Cellular biomarkers of proteasome dysfunction — aggregates of ubiquitinated proteins, activation of the NRF1 pathway, increased autophagy and mitophagy, activation of the integrated stress response and the type I interferon signature — are research tools, without clinical validation for diagnosis [29]. There is disagreement about the interferon signature in Stankiewicz-Isidor syndrome: the series of 24 patients proposed that the syndrome may be diagnosed biochemically through this signature, whereas the 2025 family series found non-specific disturbances and no patient with clinical autoinflammation [77, 79]. Automated facial analysis suggested a recognisable facial phenotype in Yu-Kury syndrome, but not in the dominant form of PSMC3 [30, 73].

6. Differential Diagnosis

6.1. Types within the class

6.2. Differential outside the class

6.3. Main differentials

The first differential is with the other interferonopathies. CANDLE/PRAAS and SAVI share the chronic interferon signature with Aicardi-Goutières syndrome, but differ in systemic inflammation, because CRP and ESR are raised in the first two and rarely in the third [5]. Nodular or violaceous annular rashes and panniculitis with lipodystrophy point to CANDLE/PRAAS; early interstitial lung disease points to SAVI; encephalopathy with leucoencephalopathy and early calcifications points to Aicardi-Goutières [5]. The genetic mimics of CANDLE recognised by the consensus are IKBKG exon 5 splicing variants (NEMO-NDAS), SAMD9L frameshift variants (SAAD) and recessive RNASEH2A/B/C variants [5].

The second differential is with the IL-1-mediated autoinflammatory diseases and periodic fevers, such as NOMID/CINCA, TRAPS and mevalonate kinase deficiency: attacks are episodic, lipodystrophy and chronic panniculitis are not part of the picture, the interferon signature is normal and the response to IL-1 blockade is good, unlike in CANDLE [2, 47, 97, 98].

The third differential is with the lipodystrophies. In congenital generalised lipodystrophy, fat loss is present from birth or appears in the first year, is accompanied by muscle hypertrophy, hypertriglyceridaemia and insulin-resistant diabetes, which develops in 25 to 35% of patients, mainly between 15 and 20 years of age, and there is no inflammation [101]; familial partial lipodystrophies, mandibuloacral dysplasia and leprechaunism also lack fever and panniculitis; Barraquer-Simons acquired partial lipodystrophy is acquired and without early systemic disease [2, 40, 100]. Cluster analysis of the JMP series showed that the picture is distinct from the progeroid laminopathies [40].

Neutrophilic and lobular panniculitides of other causes, such as α1-antitrypsin deficiency, and Sweet syndrome of infancy, whose histology may mislead, complete the cutaneous differential [2, 103]. The dominant forms with immunodeficiency have a differential with severe combined immunodeficiencies and Omenn syndrome of other causes, such as partial RAG or Artemis deficiency, which presents with faltering growth, erythroderma, lymphadenopathy, hepatosplenomegaly and severe infections [4, 51].

6.4. Differential of KLICK syndrome and the neurodevelopmental proteasomopathies

The differential of KLICK syndrome is that of the transgredient palmoplantar keratodermas with constricting bands and of the generalised congenital ichthyoses [33, 87].

The most important differential is with Vohwinkel syndrome, which also has linear hyperkeratotic lesions and a tendency to constricting bands; in KLICK syndrome, the striate lesions in the flexures, hypergranulosis with abnormal keratohyalin granules, the absence of the honeycomb pattern and autosomal recessive inheritance guide the diagnosis, and the absence of loricrin and connexin 26 variants separates it from Vohwinkel syndrome, KID syndrome and the erythrokeratodermias [87]. Presentation with well-demarcated erythematous hyperkeratotic plaques brings it close to erythrokeratodermia and loricrin keratoderma, and only study of POMP resolves the question [84]. PRAAS2, although allelic, does not enter the clinical differential: the dermatosis is neutrophilic, with leucocytoclastic vasculitis, and there is combined immunodeficiency and autoimmunity [28, 46].

The differential of the neurodevelopmental proteasomopathies is that of the syndromic neurodevelopmental disorders with malformations, for which genomic sequencing is the discriminating tool [29]. Aicardi-Goutières syndrome, which also has an interferon signature without prominent systemic inflammation, is distinguished by early encephalopathy with leucoencephalopathy and calcifications (section 6.2), whereas the magnetic resonance imaging changes described in the neurodevelopmental proteasomopathies are non-specific (delayed myelination, atrophy, hypoplasia of the corpus callosum, pons or cerebellum, pineal cysts, heterotopia) [5, 29, 30]. Compared with PRAAS, fever, panniculitis, lipodystrophy and raised acute-phase reactants are absent [29]. In the recessive form of PSMC3, the association of congenital deafness and early cataract has its own differential: Aymé-Gripp syndrome due to de novo MAF variants (cataract, deafness, intellectual disability, seizures and trisomy 21-like facies), dominant WFS1 variants with neonatal or childhood diabetes and, above all, teratogenic causes, in particular congenital infections [74].

7. Treatments and Risk Management Strategies

7.1. Principles and goals of treatment

The EULAR/ACR (European Alliance of Associations for Rheumatology/American College of Rheumatology) consensus defines the goal of treatment of type I interferonopathies as reducing systemic and organ inflammation, preventing or limiting the onset and progression of organ damage and improving quality of life (overarching principle C; level of evidence 2, grade B; agreement 9.8±0.5 on a scale of 0 to 10) [5]. Point 9 makes this goal concrete: treatment should aim for disease control or low activity, to prevent progression of damage, and in CANDLE/PRAAS this control should be maintained with the lowest possible dose of glucocorticoids (level 2, grade B for the control goal; level 4, grade C for the minimum glucocorticoid dose; agreement 9.4±1.2) [5]. Genetic confirmation is the prerequisite, because it enables targeted treatment and genetic counselling [5].

With the available strategies, inflammatory remission is achieved only in a subset of patients, described mainly in CANDLE/PRAAS; the realistic goal for the rest is reduction of inflammation and containment of damage, with therapeutic adjustments and close monitoring [5]. Established damage is irreversible, which gives priority to early treatment: established lipoatrophy and contractures do not regress when active inflammation is controlled [5, 47]. The efficacy of Janus kinase inhibitors in CANDLE/PRAAS depends on exposures higher than those used in rheumatoid arthritis, which requires weighing control of inflammation against immunosuppression, cytopenias and BK polyomavirus reactivation [76, 106].

The operational response definitions used in the reference studies serve as therapeutic targets in practice and are gathered in the following table. The daily diary score (DDS) is the daily mean, recorded by the patient or carer, of the scores of the disease symptoms (fever, rash, musculoskeletal pain, headache, fatigue and others), summed and divided by the number of symptoms assessed [31].

Treatment is organised along these lines: a Janus kinase inhibitor as the baseline targeted therapy, glucocorticoids as a bridge and at the lowest effective dose, adjustment of the inhibitor dose to weight, renal function and activity, with slow reductions and flare criteria to monitor them, blockade of the type I interferon receptor in patients with an incomplete response and haematopoietic stem cell transplantation in refractory forms or forms with combined immunodeficiency [5, 35, 76, 107]. The multidisciplinary team, rehabilitation and treatment of metabolic and organ complications complete the plan (points 13 to 15) [5].

7.2. Conventional treatment and ineffective drugs

Before Janus kinase inhibitors, no treatment was consistently effective in CANDLE [2]. Oral glucocorticoids and methotrexate gave partial improvement, and in a 2017 review methotrexate was put forward as a possible first line, on the author's opinion; non-steroidal anti-inflammatory drugs partly controlled the fever; dapsone and colchicine were ineffective; ciclosporin, azathioprine and intravenous immunoglobulin (IVIG) gave minimal or no improvement; and tumour necrosis factor (TNF) antagonists, such as etanercept, brought no benefit and were associated with worsening [2]. All 18 patients in the NIH (National Institutes of Health) expanded access programme had failed between one and six conventional or biological disease-modifying drugs, and 14 (78%) had been on chronic glucocorticoids for a mean of 5.7 years (1 to 17) [31]. Biologics targeting interleukin 1 (IL-1), TNF and interleukin 6 (IL-6) have a weak response, which is consistent with the central role of type I interferon rather than these cytokines [31, 108].

The most recent reports confirm the pattern. In four children with PSMB10 variants or with variants in PSMB8, PSMA5 and PSMC5, azathioprine, tocilizumab (anti-IL-6) and methotrexate gave only temporary improvement, with panniculitis of the heels appearing after one month of tocilizumab in one case; colchicine (0.03 mg/kg/day) with prednisolone (0.5 mg/kg/day), started at 8 months, controlled the disease for 12 months until relapse; and prednisolone (2 mg/kg/day) with methotrexate (10 mg/m², frequency not stated in the report) improved the arthritis and markers without controlling the fever or panniculitis [49]. In the cohort of eight patients with the dominant-negative PSMB8 variant p.(Gly209Arg), seven received systemic glucocorticoids (0.2 to 1 mg/kg/day) and five also received hydroxychloroquine (one patient), mycophenolate mofetil (three), methotrexate (two), infliximab (two), rituximab (one), antithymocyte globulin (one) or colchicine (one), with partial and transient responses [52]. In the 2026 systematic review, methotrexate, azathioprine, colchicine, dapsone, anti-TNF and anti-IL-6 agents were used frequently and with limited success, and one patient had a marked response to baricitinib after failure of anakinra (anti-IL-1) and tocilizumab [80].

The consensus gives priority to early treatment, because established damage is irreparable, and recommends Janus kinase inhibitors as the therapy with benefit demonstrated in open-label studies; high-dose glucocorticoids were associated with serious adverse events in the reports reviewed and do not replace targeted therapy [5, 80].

7.3. Glucocorticoids

The EULAR/ACR consensus considers glucocorticoids beneficial for the symptoms of CANDLE/PRAAS with systemic inflammation (point 12; level 4, grade C; agreement 9.0±1.3), but limits their use because of toxicity: prolonged treatment causes growth arrest, truncal obesity, arterial hypertension, glucose intolerance and osteopenia, so the lowest dose that maintains control should be sought [5]. Glucocorticoids control inflammation but not wasting or lipoatrophy, which continue to progress on prednisolone [27, 47]. In acute flares, systemic glucocorticoids may be needed, combined with specific treatment of the organ affected [2].

The doses described before Janus kinase inhibitors are high and prolonged. In the NIH programme, the median at entry was 0.44 mg/kg/day of prednisone equivalent (interquartile range 0.31–1.09), in 14 of 18 patients, maintained for a mean of 5.7 years; one of the eligibility criteria was receiving at least 0.15 mg/kg/day or having failed oral glucocorticoids [31]. In the Japanese phase 2/3 study, the mean baseline dose was 0.2 mg/kg/day [90]. In case reports, prednisolone was started at 0.5 to 2 mg/kg/day, with relapse of the lesions on dose reduction [49], and in the dominant-negative form of PSMB8 0.2 to 1 mg/kg/day was used [52]. Corticosteroid pulses were used in severe complications: in the autoimmune haemolytic anaemia of the patient with PRAAS4 and in the macrophage activation syndrome of a patient who stopped baricitinib, treated with methylprednisolone pulses and then prednisolone 30 mg/day [47, 76]. The reports do not state the pulse dose; the 2022 EULAR/ACR points to consider for haemophagocytic lymphohistiocytosis and macrophage activation syndrome give, for this situation in other diseases, intravenous methylprednisolone 10 to 30 mg/kg/day, up to 1 g, given daily, a dose applied here by extrapolation [110].

Reduction under a Janus kinase inhibitor is guided by the DDS and inflammatory markers. In the NIH programme, glucocorticoids were reduced during baricitinib dose escalation and again once the optimal dose had been reached; the median fell to 0.11 mg/kg/day (0.02–0.24), 10 of 14 patients met the criterion for successful reduction and 5 of 10 CANDLE patients stopped them permanently, without return of symptoms [31]. Catch-up growth occurred in 9 of the 13 patients with growth potential who managed to lower the dose below 0.16 mg/kg/day, and bone mineral density increased; the authors describe the effect on growth as more marked in patients who lowered the dose below this threshold [31]. In the Japanese study, in patients with Nakajo-Nishimura syndrome with a mean age of 39.2 years, the reduction was smaller: a mean of −23.5% in the five patients at the end of the primary period and −18.4% at the end of maintenance, with reduction in 3 of 5 [90]. In case reports, discontinuation took 6 months on baricitinib in a child with PSMB10 variants [49] and 14 months on ruxolitinib in the patient with PSMG2 variants, from 0.6 mg/kg/day [47]. In the dominant-negative form of PSMB8, the combination of baricitinib with anifrolumab was needed to stop glucocorticoids [52].

The flare criteria of the NIH cohort also apply to glucocorticoid reduction, and patients on high doses should be monitored for blood pressure, blood glucose, growth, bone mineral density and the ophthalmological complications of prolonged use [5, 76]. The flare criteria are not adjusted for high glucocorticoid doses, which limits their interpretation in patients who have not yet reduced them [76].

7.4. Janus kinase inhibitors

7.4.1. Mechanism and drugs

The type I interferon receptor signals through the Janus kinases, which phosphorylate the transcription factor STAT1 and induce interferon-stimulated genes; Janus kinase inhibitors reduce, in a dose-dependent manner, interferon alpha-mediated STAT1 phosphorylation and interrupt the interferon amplification loop [5, 106]. The three drugs used in CANDLE/PRAAS are first-generation: baricitinib inhibits JAK1 and JAK2, ruxolitinib inhibits JAK1 and JAK2 with slight inhibition of TYK2, and tofacitinib inhibits JAK1 and JAK3 with slight inhibition of JAK2 [80]. In patients treated with baricitinib, the concentrations blocking 50% of interferon alpha-induced STAT1 phosphorylation were lower in CD4+ and CD8+ T lymphocytes (51.1 and 43.3 nM) than in monocytes (137.7 nM); suppression tracked plasma concentration and disappeared at trough, between doses, which underpins dose splitting in children [106]. The effect translated into a dose-dependent fall in serum IP-10 (CXCL10) protein and in the 25-gene interferon score, more marked in CANDLE than in SAVI [106]. There are no published reports of the use of upadacitinib, abrocitinib or ritlecitinib in CANDLE/PRAAS [80].

7.4.2. Evidence

The EULAR/ACR consensus considers Janus kinase inhibitors beneficial for the symptoms of CANDLE/PRAAS (point 10; level 2, grade B; agreement 9.3±0.9), with baricitinib as the most used drug and ruxolitinib and tofacitinib as described options [5]. There are no randomised trials; the evidence comes from an expanded access programme, an open-label Japanese study, series and case reports, gathered in a systematic review (see the table of trials in subsection 7.12) [31, 80, 90].

In the NIH expanded access programme (NCT01724580), 18 patients (10 with CANDLE, 4 with SAVI and 4 with other interferonopathies), with a mean age at entry of 12.5 years (1.2 to 24.1), were enrolled between October 2011 and February 2017 and treated for a mean of 3.0 years (1.5 to 4.9) [31]. At the end of follow-up, 12 of 18 patients (67%) met the DDS improvement criterion, including 80% of CANDLE patients; the median DDS fell from 1.3 (0.93–1.78) to 0.25 (0.10–0.63) and 10 of 14 patients met the glucocorticoid reduction criterion [31]. Five of the ten CANDLE patients (50%) achieved remission, without symptoms, with normal CRP and without glucocorticoids; at subsequent visits, over a mean of 654 days (581 to 822), CRP was below 5 mg/L at 84.6% of visits and the interferon score was normal at 66.7% [31]. Myositis and aldolase improved (p=0.06), the height Z-score went from −4.03±2.64 to −3.19±2.33 in the 13 patients with growth potential and the bone mineral density Z-score from −3.25±1.97 to −2.20±1.36 (p<0.005) [31]. Inhibition did not correct the metabolic changes: median lipids (HDL, LDL and triglycerides) rose, pre-existing hepatic steatosis did not improve in two patients and appeared de novo in three CANDLE patients with baseline hyperlipidaemia, which suggests a mechanism of steatosis independent of interferon-mediated inflammation [31]. Three patients stopped the drug: two without a genetic diagnosis, because of lack of efficacy and because of osteonecrosis with an unsatisfactory response, and one CANDLE patient, because of BK viraemia with azotaemia [31]. The open-label design, without predefined hypotheses or adjustment for multiple comparisons, with patients as their own controls, limits the strength of these results [31].

In the Japanese phase 2/3 study (NCT04517253), the five patients with Nakajo-Nishimura syndrome had a mean age of 39.2 years (15 to 48), went through a 12-week drug-free period to measure baseline activity and an 8-week dose-adjustment period according to weight and estimated glomerular filtration rate (eGFR), with tablets or an oral suspension below 40 kg [90]. The maximum daily dose ranged from 8 to 12 mg; the mean DDS fell by 0.22 at week 20 and 0.18 at week 52, with improvement in 60% of patients versus 80% in the NIH programme, a difference the authors attribute to the higher mean age [90]. Among the study's nine patients there were three serious adverse events (acute coronary syndrome, drug-related pancytopenia and bronchopulmonary aspergillosis), and one patient with Nakajo-Nishimura syndrome stopped treatment because of drug-related Pneumocystis jirovecii pneumonia; one patient with SAVI died of intracranial haemorrhage, considered unrelated to the drug; BK virus was detected in three patients (33.3%) [90]. Adding the patients of the expanded access programme followed after the initial publication, 20 of 23 (87%) developed BK viruria [90]. The study ended in November 2024 by decision of the sponsor, citing efficacy reasons [111].

The 2026 systematic review (search up to September 2025) included 16 studies — two trials, one retrospective study, two series and eleven case reports — with 46 patients, median age 4.5 years at onset (from the neonatal period to 18 years), treated with baricitinib (73.9%), tofacitinib (23.9%) or ruxolitinib (2.1%) [80]. In the 26 patients described in full articles, the response was complete in 42.3%, significant in 11.5%, partial in 38.5% and absent in 7.6%; in the whole group of 46, which includes 20 patients described only in conference abstracts, the breakdown by drug gives a complete response in 13 (28.3%), significant in 4 (8.7%), partial in 27 (58.7%) and absent in 2 (4.3%), a category that includes death during treatment; the text of the review rounds the first three to 28.2%, 8.6% and 58.6% and records two deaths despite treatment [80]. Adverse events were mainly infectious (upper respiratory tract infections in 28.2% and BK virus infection in 19.5%), with changes in liver enzymes, lipids, creatine kinase and platelets [80]. The response categories were defined by the reviewers from qualitative descriptions, comparison between drugs was not possible and follow-up was short in most reports; the review attributes to the NIH programme a death after discontinuation for acute kidney injury and flares in nine of ten patients after dose reduction, data that do not match the original articles, which record no deaths up to the data cut-off and flares in 7 of 14 reductions in 9 patients [31, 76, 80].

7.4.3. Response according to genotype and drug

The response varies with genotype. Prolonged normalisation of the interferon signature was observed only in a subset of patients with biallelic PSMB8 variants [76]. In the dominant-negative form of PSMB8 p.(Gly209Arg), five of eight patients received Janus kinase inhibitors (three tofacitinib and two baricitinib); all improved, with resolution of skin lesions and improvement of myositis, but normalisation of inflammatory markers and of the interferon signature was only partial, unlike in the biallelic forms [52]. In one of these patients, baricitinib at 0.2 mg/kg/day did not suppress the signature and required high-dose glucocorticoids, escalation to 0.3 mg/kg/day was limited by recurrent infections, and only the combination with anifrolumab made it possible to lower baricitinib to 0.1 mg/kg/day and stop glucocorticoids [52]. A newborn with the same variant responded to baricitinib [62].

In children with biallelic PSMB10 variants, baricitinib at 3 mg/day (0.3 mg/kg/day), started at 1 year and 8 months, suppressed the skin lesions and fever, normalised liver enzymes, CRP and erythrocyte sedimentation rate (ESR) and allowed glucocorticoids to be stopped within 6 months; in another child, baricitinib at 2 mg/day (0.1 mg/kg/day), started at 6 years and combined with methotrexate, led to remission; tofacitinib at 0.9 mg/kg/day gave only a partial response in a third child, with CRP of 33 mg/L and prednisolone at 0.44 mg/kg/day at 3.5 years [49]. In a girl with variants in PSMB8, PSMA5 and PSMC5, tofacitinib was increased up to 30 mg/day (0.7 mg/kg/day) and controlled the disease only in combination with subcutaneous tocilizumab 162 mg (3.8 mg/kg) every two weeks [49]. In the infant with the dominant PSMB9 variant p.(Gly156Asp) who needed extracorporeal membrane oxygenation, tofacitinib allowed weaning before transplantation [65]. In the patient with PRAAS4 due to PSMG2, ruxolitinib, started at 8 years at 5 mg twice daily and increased with weight, controlled the inflammation and allowed prednisolone to be stopped (section 10) [47]. In the systematic review, baricitinib gave a complete response in 10 patients, significant in 3, partial in 20 and absent in 1, and tofacitinib a complete response in 3, significant in 1, partial in 6 and absent in 1; ruxolitinib, used in a single patient, gave a partial response; the authors made no formal comparison between drugs, owing to the size and heterogeneity of the samples [80].

7.4.4. Pharmacokinetics and dosing schedule of baricitinib

Baricitinib (Olumiant®) is eliminated mainly by the kidneys. In the population pharmacokinetic analysis of 18 patients in the NIH programme, treated with doses of 0.1 to 17 mg/day (0.01 to 0.82 mg/kg/day), weight influenced the volume of distribution and eGFR influenced clearance [106]. The half-life was 4.4 hours below 20 kg, 5.0 hours between 20 and 40 kg and 7.0 hours above 40 kg, versus about 10 hours in healthy adults and 12.5 hours in rheumatoid arthritis; clearance per kilogram was higher in smaller children, which requires dosing up to four times daily below 20 kg [106]. With eGFR below 120 mL/min/1.73 m², dose-normalised exposure was 52 to 59% higher than in patients of the same weight with better renal function, and one patient with eGFR between 54 and 76 mL/min/1.73 m² had an estimated half-life of 17 to 19.2 hours [106]. At clinically effective doses, the 24-hour concentration area under the curve (AUC24) had a geometric mean of about 2388 nM·h, 1.83 times that observed in rheumatoid arthritis with 4 mg once daily (about 1304 nM·h), and five CANDLE patients normalised the interferon score in a sustained manner with AUC24 between about 1500 and 2000 nM·h [106]. The mean final dose was 8.1±2.4 mg/day (0.24±0.13 mg/kg/day) [106].

The weight- and eGFR-based dosing proposal derives from these tolerated and effective doses and is the scheme to which the EULAR/ACR consensus refers to guide adjustment in the growing child [5, 106]. The programme required a weight of at least 8.5 kg and an age of at least 17.5 months, excluded patients with eGFR below 40 mL/min/1.73 m², and there are no data below 8.5 kg [106]. At the start of the programme, without paediatric pharmacokinetic data, the starting dose was 100 µg once daily, with progressive escalation; this historical dose was replaced by the following table [31, 106].

The effective doses in CANDLE/PRAAS (4 to 12 mg/day in the Kim proposal, versus 1 to 4 mg/day in the labels) are split into two to four doses, so all use is off-label and outside the conditions of the label; the reference dose is that of the Kim proposal, cited by the consensus, and the label serves as a comparison [5, 106, 112, 113]. The eGFR threshold of the proposal (120 mL/min/1.73 m²) is higher than the renal impairment thresholds of the adult labels, which the authors relate to the need for higher thresholds in children [106]. The EMA assessed doses up to 12 mg/day in 71 paediatric patients with CANDLE, related conditions, SAVI, juvenile dermatomyositis and Aicardi-Goutières syndrome, without a conclusion of efficacy owing to methodological shortcomings and with higher adverse event frequencies than in the adult indications [112]. In the Japanese study, the maximum daily dose was 8 to 12 mg [90].

During escalation, the dose is increased when the DDS remains at 0.5 or more, or symptoms of activity persist, without signs of toxicity such as a fall in haemoglobin [31]. In the Japanese study, the adjustment phase lasted 8 weeks and the dose could be modified as weight changed [90]. In the NIH cohort, the mean time to the clinically effective dose was 1.06±0.65 years [76]. The 1, 2 and 4 mg tablets can be dispersed in about 10 mL (minimum 5 mL) of room-temperature water and given orally, by gastrostomy or by nasogastric or orogastric tube, and handling of crushed tablets requires protection [113]. In Portugal, the 2 mg and 4 mg strengths are marketed [116]. The Japanese study allowed an oral suspension in patients below 40 kg [90]; the sources consulted do not describe a marketed baricitinib oral suspension. Potent inhibitors of the OAT3 transporter, such as probenecid, increase exposure and require dose reduction according to the label [113]. The labels recommend not starting the drug with an absolute lymphocyte count below 500/µL, a neutrophil count below 1000/µL or haemoglobin below 8 g/dL, and interrupting until recovery; monitoring during treatment is described in subsection 7.5 [112, 113].

7.4.5. Dose reduction, adjustment for growth and discontinuation

Reduction of the Janus kinase inhibitor dose is needed in three situations: cytopenias, BK viral load in the blood, which one seeks to keep within the safety limits used in renal transplantation, and growth of the child, because relative clearance and volume of distribution decrease with weight [76]. In the NIH cohort, between April 2014 and December 2019, there were 14 reductions in 9 of the 10 CANDLE patients, with a mean treatment duration of 6.3±2.3 years; BK viral load in blood and urine fell after the reductions, but 7 of the 14 reductions (50%) triggered a flare, five clinical and two subclinical [76]. All four reductions greater than 25% of the dose triggered a flare (p<0.05), two clinical and two subclinical; for reductions of less than 25%, the study summary gives a flare rate of 29%, and the text describes clinical flares in 2 of 10 reductions [76]. The time between reduction and meeting the flare criteria ranged from 1 to 55 days (mean 15.3±20.7), depending on the date of sampling [76]. One patient with probable BK nephropathy and raised creatinine stopped baricitinib 8 days after a reduction and developed macrophage activation syndrome, treated with methylprednisolone pulses and then prednisolone 30 mg/day [76].

In 153 visits of six patients, the flare criteria were met at 43.14% of the 51 visits with doses below those recommended and at 12.75% of the 102 visits with optimised doses (p<0.0001); the median dose was 6.00 mg/day at the low-dose visits and 9.00 mg/day at the optimised-dose visits, the median prednisone dose was 0.149 and 0.00 mg/kg/day, and the difference persisted after adjustment for glucocorticoids (p=0.03) [76]. The median interferon score was 254.46 at the low-dose visits and 45.93 at the optimised-dose visits (p<0.0001) [76]. The authors conclude that, once control has been achieved, the lowest effective dose should be sought, but that reductions should be slow, gradual and monitored, and warn against large reductions and abrupt discontinuation, because of the risk of severe rebound flares, including macrophage activation syndrome, aggravated by the drug's short half-life [76]. The EULAR/ACR consensus adds that discontinuation may cause withdrawal flares and that treatment should not be interrupted without a specific contraindication [5]. The ruxolitinib label also recommends gradual tapering, because of the risk of symptom exacerbation after interruption [117].

7.4.6. Flare criteria

The CANDLE/PRAAS flare criteria were developed from the flares observed after dose reductions, comparing each visit with a reference visit, which is the last visit with laboratory tests before the reduction period, in a phase of controlled disease [76].

Source: [76].

Of the seven clinically identified flares, six met the criteria; the patient who did not meet them had headaches, oral ulcers, musculoskeletal pain and fatigue, with lymphopenia as the only laboratory change, and his 25-gene interferon score rose from 31.21 to 236.29 (threshold 44.2) [76]. Including the interferon score increased the proportion of visits meeting flare criteria from 37.93% to 58.62% at the low-dose visits and from 10.13% to 20.25% at the optimised-dose visits, mainly in lymphopenic patients with low CRP [76]. The score fluctuates by up to 20% within the same day in patients with active disease, so it should only be used as a criterion in controlled patients; where it is not available, the criteria are applied without it [76]. The criteria serve to monitor the reduction of glucocorticoids and of the Janus kinase inhibitor, adjust the dose during growth, decide on intensification after a flare and design randomised withdrawal trials [76]. The limitations are those of a retrospective study, without a control group, with the same small cohort used to develop and to assess the criteria, without adjustment for high glucocorticoid doses and without validation in other interferonopathies [76].

7.5. Toxicity monitoring during Janus kinase inhibition

BK polyomavirus reactivation is the characteristic adverse event in this population. In the NIH programme, BK viruria and viraemia appeared in this group of patients and not in the rheumatoid arthritis cohorts; one patient with CANDLE stopped treatment because of BK viraemia with azotaemia and renal tubulointerstitial disease, and urinary and blood titres in the others remained stable over two years; in the pharmacokinetic analysis, BK viraemia appeared in 44% of patients, with low and stable titres at maintenance doses [31, 106]. In the Japanese study, BK virus was detected in 3 of 9 patients (33.3%) [90]. The consensus recommends measuring BK viral load in blood and urine and assessing renal function before starting the drug, at the baseline assessment and then at each visit, because there is no proven treatment for BK nephropathy and management depends on early detection and reduction of immunosuppression (point 11; level 4, grade C for CANDLE; agreement 9.3±1.6) [5]. Before starting, the consensus also recommends baseline renal ultrasound and a urinary protein/creatinine or albumin/creatinine ratio [5]. Other viral reactivations, such as herpes, have been described, but the data are insufficient to recommend routine antiviral prophylaxis [5]. Upper respiratory tract infections were the most frequent adverse events, with two cases of herpes zoster in 63 patient-years of exposure and no deaths in the programme [31]. One patient without a genetic diagnosis developed osteonecrosis of the femur [31]. Monitoring also includes thrombocytosis, leucopenia, anaemia, liver function tests and lipid profile [5, 112]. The baricitinib label recommends assessing lipids about 12 weeks after starting and treating dyslipidaemia according to general guidelines, screening for latent tuberculosis and viral hepatitis before starting, and not starting in the presence of significant lymphopenia, neutropenia or anaemia; the boxed warning covers serious infections, mortality, malignancy, major adverse cardiovascular events and thrombosis [112, 113]. The European label adds that, in the approved indications, the drug should only be used when there is no suitable alternative in people aged 65 years or over, with atherosclerotic disease or cardiovascular risk factors and with risk factors for malignancy [112].

7.6. Type I interferon receptor blockade

Anifrolumab is a monoclonal antibody directed against subunit 1 of the type I interferon receptor. In an open-label series of four patients with SAVI and one with CANDLE, it normalised expression of interferon-stimulated genes and gave clinical improvement beyond that obtained with the Janus kinase inhibitor alone, allowed corticosteroids to be reduced and stopped with less overall immunosuppression, and improved metabolic status and growth; in two patients with SAVI it was sufficient as monotherapy [107]. In the dominant-negative form of PSMB8 p.(Gly209Arg), two patients with an incomplete response to the Janus kinase inhibitor received anifrolumab: both normalised the interferon score, stopped glucocorticoids and reduced the high dose of the inhibitor, but kept lymphopenia, recurrent viral infections and progression of the liver disease, with worsening splenomegaly and transaminases [52]. In one of them, anifrolumab monotherapy kept the interferon score low but was followed by a flare, with rash, myositis and a rise in the chemokine CXCL9, which regressed on reintroduction of baricitinib, suggesting that the immune dysregulation of this form goes beyond the type I interferon pathway [52]. The authors propose receptor blockade as an add-on in patients with an incomplete response or with T-cell-mediated immunodeficiency [52]. Use is investigational in these diseases: the US label covers only moderate to severe systemic lupus erythematosus in adults on standard therapy, with 300 mg as a 30-minute intravenous infusion every 4 weeks or 120 mg subcutaneously weekly, and the European authorisation covers add-on therapy for active lupus in adults [118, 119]. In Portugal and Brazil there is no approved indication for interferonopathies; anifrolumab has an active registration in Brazil as Saphnelo® [120].

7.7. Haematopoietic stem cell transplantation

The proteasome is ubiquitous, so it was not evident that replacement of the haematopoietic system would correct the disease; published experience shows that it does, at least in the forms in which the phenotype depends on immune cells [35]. In a boy with PRAAS3 due to two PSMB4 variants, with resistant cutaneous vasculitis, chronic thrombocytopenia and faltering growth, a first transplant with reduced-intensity conditioning at 20 months was followed by loss of lymphoid chimerism and relapse, and a second myeloablative transplant at 3 years achieved 100% donor chimerism, with improvement of symptoms (except thrombocytopenia), catch-up growth and later appearance of mild lipodystrophy; proteasome function, protein homeostasis and the interferon-stimulated gene signature normalised [35, 64]. In two patients with autoinflammatory disease and immune dysregulation due to POMP, transplantation corrected the immunodeficiency and immune dysregulation, with 100% donor engraftment and a fall in type I interferon-regulated genes [121]. In the patient with PSMB9 p.(Gly156Asp) who required extracorporeal oxygenation, tofacitinib allowed weaning and cord blood transplantation at 7 months led to remission, without need for the drug or relapse two years later [65]. In the dominant forms of PSMB10 with severe combined immunodeficiency, transplantation is the necessary treatment, but the results were poor: of the six infants, two died in the peri-transplant period and two late, with severe inflammatory complications such as graft-versus-host disease and thrombotic microangiopathy, and the only long-term survivor benefited from revertant mosaicism [35, 51]. In the dominant-negative form of PSMB8, one patient with chronic Pseudomonas aeruginosa infections, early bronchiectasis and progression of hepatic nodular regenerative hyperplasia with splenomegaly was under assessment for transplantation at the time of publication [52]. In the 2026 series with dominant PSMB10 variants and liver disease, severe vascular disease before and after transplantation led the authors to weigh the indication with caution [70]. Transplantation is therefore an option in refractory forms and in forms with immunodeficiency, with the indication to be discussed case by case [35].

7.8. Supportive treatment and treatment of complications

Physiotherapy from an early stage prevents and limits joint contractures [2, 11]. Lipodystrophy and its metabolic consequences require endocrinology follow-up, with diet and control of lipids and blood glucose [2]. Autoimmune complications are treated conventionally: the patient with PRAAS4 and autoimmune haemolytic anaemia received transfusions, corticosteroid pulses, four infusions of rituximab and a 19-month course of azathioprine with IVIG every 4 weeks [47]. The team should include rheumatology, genetics, neurology, ophthalmology, respiratory medicine, cardiology, hepatology, gastroenterology, haematology, immunology, dermatology, endocrinology and nephrology, with access to physical medicine, wound care, psychology, bone health, physiotherapy, dentistry, dietetics, rehabilitation, orthopaedics and social support (point 13; level 5, grade D; agreement 9.9±0.3) [5]. Chronic inflammation and corticosteroids impair bone health, growth and development, which should be monitored, as should the cardiac and ophthalmological complications of prolonged corticosteroid use [5].

7.9. Vaccination, infections and COVID-19

For CANDLE/PRAAS, the consensus recommends all routine vaccines, live and inactivated, when the patient is not on immunosuppressants or glucocorticoids, case by case; on Janus kinase inhibitors or other immunosuppressants, live vaccines should be avoided (point 17; level 5, grade D; agreement 9.4±0.9) [5, 113]. There are no data indicating that routine vaccines trigger flares in CANDLE/PRAAS, and there are insufficient data on messenger ribonucleic acid vaccines [5]. At the time of the consensus, there was no evidence that the COVID-19 risk of these patients differed from that of the healthy population, and treatment of the interferonopathy should not be stopped without a specific contraindication (point 16; level 5, grade D; agreement 9.5±0.8), all the more because discontinuation may cause withdrawal flares [5].

7.10. Agents to avoid, pregnancy, perioperative care and transition

TNF antagonists, namely etanercept, should be avoided, because of ineffectiveness and reported worsening, and IL-1 blockade is not effective [2, 108]. The baricitinib and tofacitinib labels advise against combination with other Janus kinase inhibitors, with biological disease-modifying drugs and with potent immunosuppressants such as azathioprine and ciclosporin [113, 114]. Reverse transcriptase inhibitors, studied in Aicardi-Goutières syndrome with a reduction in the interferon score without demonstrated clinical benefit, have not been studied in CANDLE/PRAAS and should not be used here outside research [5, 11]. In pregnancy, baricitinib is contraindicated in the European label, because of teratogenicity in rats and rabbits and a possible effect on intrauterine bone development, requires effective contraception during treatment and for at least one week afterwards, and should not be used during breastfeeding, because it is excreted in milk in animals [112]. The sources consulted do not describe pregnancies in patients with PRAAS or specific perioperative recommendations; the general principles apply of maintaining control of inflammation, avoiding abrupt discontinuation of the Janus kinase inhibitor because of the risk of a withdrawal flare and monitoring for infections [5, 76]. Transition to adult medicine should be prepared by the multidisciplinary team, encouraging self-management skills and participation in decisions, and monitoring growth and sexual development at each visit (points 13 to 15) [5].

7.11. Regulatory status

No drug has orphan medicinal product designation or an approved indication for CANDLE, PRAAS, Nakajo-Nishimura syndrome or type I interferonopathy in the sources consulted [11, 112, 113]. In Japan, at the time of the phase 2/3 study, baricitinib was approved for rheumatoid arthritis, atopic dermatitis, alopecia areata and COVID-19 pneumonia; the sources consulted do not describe approval for Nakajo-Nishimura syndrome, which does not demonstrate that it does not exist [90].

7.12. Trials and studies

None of these studies is randomised or controlled, so the response percentages cannot be compared with one another or with placebo; the expanded access programme had no predefined hypotheses or adjustment for multiple comparisons, and patients served as their own controls [31]. The systematic review includes only published and treated patients, which favours responders, and rates its own evidence as low-level [80]. The Japanese study included nine patients and ended by decision of the sponsor [111]. The two ongoing phase 1 trials in type I interferonopathies — the inhibitor IMSB301 in Australia (NCT07364513) and BI 3000202 in Europe and the United States (NCT06878365) — explicitly exclude the proteasome genes, accepting only patients with Aicardi-Goutières syndrome, familial chilblain lupus, COPA syndrome and DNASE2 or ATAD3A variants, so they do not cover PRAAS [125, 126].

7.13. Developments in the last twelve months

The developments of 2025 and 2026 are mainly mechanistic and nosological, without new authorisations. The form with immunodeficiency due to monoallelic PSMB8 variants was described and characterised, with seven individuals from five families (published human study) [13], and the mechanism of the de novo dominant-negative variant p.(Gly209Arg) was detailed, with a partial response to Janus kinase inhibition and interferon receptor blockade and identification of the kinases PKR and GCN2 as possible targets — a clinical series of eight patients with laboratory studies, without a clinical trial [52]. The systematic review of the 46 patients treated with Janus kinase inhibitors [80], new dominant PSMB10 variants with liver disease [70], a PSMB10 splicing variant with dominant segregation in a Turkish family [71] and a new digenic inheritance involving PSMA6 [63] were published. At the preclinical level, patient-derived brain organoids showed that neuronal maturation is impaired by an excess of polyamines and that inhibition of ornithine decarboxylase corrects it in vitro [60], and a mouse model homozygous for the Japanese founder variant was created for drug testing [59]. A structural analysis compared the effect of the substitutions equivalent to p.(Thr75Met) and p.(Gly201Val) on proteasome assembly and activity [61]. None of these studies changes clinical management by itself.

7.14. Discrepancies between sources

The only specific guideline is the EULAR/ACR 2021 consensus, published in 2022 in two journals with the same content [5, 55]. No other society has its own recommendations for this disease, so there are no discrepancies between guidelines to report; the limitation acknowledged by the task force itself is the lack of high-level evidence, with several statements supported by expert opinion [5]. Between the consensus and the regulators there is a material discrepancy: the consensus recommends Janus kinase inhibitors and refers to published paediatric doses, whereas the US and European labels do not cover the indication and the EMA, in assessing the paediatric data, reached no conclusion on efficacy [5, 112]. The reference for practice is the consensus; the label serves as a comparison and defines what is authorised [5]. Between the reference textbook consulted and the guideline there is the discrepancy expected from the date: Saudubray merely cites the PSMB8 variants in orientation tables, without therapeutic recommendations [54].

7.15. Treatment of KLICK syndrome

There is no mechanism-targeted treatment or clinical trial; the approach is that of the disorders of keratinisation, led by dermatology: bathing, emollients and topical keratolytics for the ichthyosis and keratoderma, and oral retinoids in forms with functional impact [28, 87, 127]. Evidence of benefit from oral retinoids in KLICK syndrome comes from case reports: the patient of the 1997 description improved with etretinate, and the Brazilian patient improved with acitretin over ten years, relapsed on stopping and improved again on reintroduction at a low dose, which shows a suppressive rather than curative effect [82, 87]. The 2021 North American consensus on retinoids in the ichthyoses and other disorders of cornification includes KLICK syndrome among the forms with documented improvement on oral or topical retinoids (level of evidence II, grade B), and a review of the palmoplantar keratodermas cites acitretin as useful for the ichthyosis and keratoderma of this syndrome [32, 36]. Surgery for the constricting bands in KLICK syndrome has not been described in the sources consulted; in keratodermas with bands, such as Vohwinkel syndrome, oral retinoids and surgery are the options described [32].

Acitretin (Neotigason® in Portugal and Brazil; in the USA, generics, after discontinuation of Soriatane®) is an aromatic retinoid, as 10 and 25 mg capsules, taken once daily with a meal, because dietary fat increases absorption [127–129]. The reference dose for KLICK syndrome is that of the 2021 consensus, which expressly includes it; the European guidelines on congenital ichthyoses, updated in 2024, apply by extrapolation, because the syndrome is not an autosomal recessive congenital ichthyosis in the strict sense; the Portuguese SmPC and the US label serve as comparison [36, 127–129].

The doses of the consensus and the European guideline are compatible with the Portuguese SmPC in adults, but the consensus allows up to 1 mg/kg/day in children, with no paediatric ceiling of its own beyond the general maximum of 75 mg/day, which exceeds the SmPC limit of 35 mg/day in children weighing more than 35 kg; in this situation, the consensus dose is the reference and the SmPC defines the conditions of authorisation [36, 128].

Safety conditions the decision. Acitretin is teratogenic: the European guidelines require contraception and monthly pregnancy tests from one month before starting, during treatment and until 36 months after stopping, total abstinence from alcohol during treatment and for 2 months afterwards (alcohol converts acitretin into etretinate, which is eliminated very slowly) and a ban on blood donation during treatment and for the following 3 years; contraindications are pregnancy or the wish to become pregnant, severe hepatic or renal impairment, hypervitaminosis A, hyperlipidaemia and combination with tetracyclines, methotrexate, vitamin A or other retinoids [127]. Because of the long half-life, the North American consensus recommends considering a switch from acitretin to isotretinoin (0.5–1 mg/kg/day, with similar efficacy in the ichthyoses and elimination within a few months) before puberty in girls at risk of pregnancy [36]. Mucocutaneous adverse effects (cheilitis, xerosis, skin fragility, fissures of the fingertips, pruritus, photosensitivity) are common, dose-dependent and reversible [127]. Laboratory monitoring includes lipid profile and liver function tests before starting, 1–2 months later and, if normal, every 6–12 months; bone toxicity depends on dose and duration, with annual assessment of growth in children and skeletal radiographs and densitometry at the end of growth, repeated every 3–5 years or in the presence of symptoms (section 11.4) [36].

The anti-inflammatory therapies used in other autoinflammatory keratinisation diseases, such as anti-TNF or anti-IL-17 antibodies, have been proposed as a hypothesis for KLICK syndrome, without any treated patient described; modulation of the unfolded protein response is another, purely theoretical, strategy [28]. The sources consulted record no clinical trials, gene therapies or specific regulatory processes.

7.16. Management of the neurodevelopmental proteasomopathies

There is no disease-modifying treatment, guideline, consensus or GeneReviews chapter for these diseases; care is symptomatic and multidisciplinary, guided by each patient's manifestations: developmental stimulation, educational and behavioural support, management of feeding difficulties in infancy and treatment of cardiac, renal and urogenital, ophthalmological, auditory, dental and skeletal anomalies by the respective specialties [11]. No anti-interferon or proteasome-directed treatment has been described in patients [11]. Patients with pernio or other inflammatory signs have, in the sources consulted, no indication for Janus kinase inhibitors; the use of these drugs in these diseases is not documented [11, 77].

The therapeutic evidence is exclusively preclinical. In T cells from patients with PSMC3 variants, expression of interferon-stimulated genes was reduced by PKR inhibition [73]. In T cells from six patients with PSMC5 variants, the interferon score fell with the PKR inhibitor (C16), the GCN2 kinase inhibitor (A92), the STING inhibitor (H-151) and, markedly, with baricitinib, which links proteotoxic stress to spontaneous interferon activation through the integrated stress response and points to pharmacological targets [30]. The reviews propose two lines of development, blocking interferon responses upstream and restoring proteasome function, for example with pharmacological activators such as phosphodiesterase inhibitors, not yet studied in patients [29]. No interventional trial was registered on ClinicalTrials.gov for PSMD12, PSMC3, PSMC5 or Stankiewicz-Isidor syndrome; the observational Simons Searchlight registry (NCT01238250), recruiting, includes PSMD12 among the conditions studied [130]. In the last twelve months, the developments were the cohort of 44 patients with PSMC5 variants, with the description of the cellular pathways and targets mentioned above (laboratory study in patients' cells and animal models), an additional case with p.(Pro320Arg) and the first series including adults carrying inherited PSMD12 variants (clinical series) [30, 79, 88].

8. Reproductive Options and Family Counselling

8.1. Recurrence risks by inheritance pattern

The class brings together different inheritance patterns, so counselling depends on the exact genotype and not on the name of the disease [3, 5]. In the autosomal recessive forms — PRAAS1 due to biallelic PSMB8, PRAAS3 due to biallelic PSMB4, PRAAS4 due to PSMG2 and PRAAS5 due to biallelic PSMB10 —, the parents are healthy heterozygous carriers and the recurrence risk in each pregnancy is 25%, with 50% carriers among unaffected offspring [9, 42, 47]. Consanguinity was confirmed in one of the JMP families and suspected in the other, was the rule in the Japanese families described and was present in the first family with PRAAS5, so it should be asked about and explored [42, 43, 48].

In the autosomal dominant forms — PRAAS2 due to POMP, PRAAS6 due to PSMB9 p.(Gly156Asp), the dominant-negative PSMB8 variants and immunodeficiency 121 due to PSMB10 —, the variants described arose almost always de novo, with a low but not zero recurrence risk for siblings, because of the possibility of germline mosaicism in one of the parents; an affected individual transmits the variant to 50% of offspring [13, 46, 50, 51]. In the Turkish family with the PSMB10 splicing variant, segregation was dominant over generations, with four affected members and variable expressivity, which shows that these variants can be inherited [71]. Confirmation that a variant is de novo requires study of both parents.

In the digenic forms, each parent carries a variant in a different gene and the disease requires the coincidence of both; the recurrence risk is 25% for the complete combination, with 25% of offspring carrying neither variant and 50% carrying only one — a situation that, on current data, does not cause disease, although it calls for caution if the partner belongs to a family with another proteasome variant [5, 23]. In some patients the second variant was de novo, which lowers the recurrence risk [23]. The consensus insists on testing both parents to confirm digenic inheritance, precisely because the pattern determines the risk [5].

The allelic disorder KLICK, due to homozygosity for the variant in the 5' untranslated region of POMP, is autosomal recessive and is not managed as an interferonopathy; the six parents studied in the founding work were heterozygous and healthy, and the five available healthy siblings were heterozygous or non-carriers [11, 33]. The recurrence risk for each pregnancy of a carrier couple is 25%. Consanguinity was present in the original Spanish family and in the Brazilian family, but not in the 1997 patient, which is consistent with a recurrent variant, with at least five different haplotypes, rather than a single founder allele [33, 81, 82, 87]. Heterozygous carriers have no described skin manifestations [33]. The child of a person with KLICK is an obligate carrier and will only be affected if the other parent is also a carrier, a risk that depends on the frequency of the variant in the partner's population, not estimated in the sources consulted.

In the dominant neurodevelopmental proteasomopathies, most variants arise de novo (33 of 45 in Stankiewicz-Isidor syndrome, 22 of 23 confirmed in the dominant form of PSMC3, 33 of 44 in Yu-Kury syndrome), with a low recurrence risk for siblings, but not zero because of the possibility of germline mosaicism; an affected person transmits the variant to 50% of offspring [29, 30, 73]. Familial transmission is documented: 9 of 45 patients with PSMD12 variants inherited the variant, including a family with an affected father and six affected children and two families with complete penetrance and variable expressivity, and two patients with PSMC5 variants inherited it from their affected mother [29, 30, 78, 79]. Therefore, a parent carrying a variant considered causal should be examined for a mild phenotype, rather than the variant being reclassified as benign. The recessive form of PSMC3 (25% risk for each pregnancy of a carrier couple), the forms due to biallelic PSMB1 and PSMC1 variants and the patient homozygous for PSMC5 p.(Thr207Met), a variant also found in the heterozygous state in another patient in the cohort, call for caution in counselling, because the number of families is minimal [19, 29, 74].

8.2. Cascade testing and carriers

Once the variants in the index case have been identified, targeted testing extends to the parents (to establish phase, origin and possible de novo status) and to siblings, with particular attention to symptomatic siblings or those with subtle findings, given the variable expressivity described within families [71]. Heterozygous carriers of recessive variants are healthy and need no surveillance [47]. Isolated heterozygosity in PSMB8 should not be interpreted as carrier status when the child has a compatible phenotype, because it may be a dominant-negative variant or half of a digenic genotype [11, 13]. In families with a known variant, targeted Sanger sequencing remains appropriate and economically reasonable [5].

8.3. Prenatal and preimplantation diagnosis

With the familial variants known, prenatal diagnosis by molecular analysis of chorionic villi or amniocytes and preimplantation genetic testing are technically possible, as in other monogenic immune diseases, and require the analysis to be designed for both genes in digenic cases [4, 23]. The sources consulted do not describe prenatal diagnosis series in these diseases. In the dominant de novo forms, a request for prenatal diagnosis in a subsequent pregnancy is justified mainly by the hypothesis of germline mosaicism and should be framed within that low probability.

8.4. Content of counselling

The consensus states that patients, parents and siblings should have access to formal genetic counselling, that a confirmed genetic diagnosis facilitates targeted treatment, screening for complications and prognostic information, and that counselling makes it possible to start risk assessment according to the inheritance pattern of the specific variant and helps in understanding the results, the implications for one's own health, reproductive issues and the impact on relatives [5]. Counselling should include the information that established organ damage is irreversible and that early treatment aims to prevent it, which makes rapid diagnosis of a symptomatic sibling relevant [5]. In adolescents and young adults, counselling includes the contraindication of baricitinib in pregnancy and the need for effective contraception during treatment and for at least one week afterwards, as well as the recommendation not to breastfeed while on the drug [112]. Testing of asymptomatic minors in families with recessive variants has no immediate clinical utility and follows the general rules for testing minors; in families with dominant variants with immunodeficiency, early recognition has a practical consequence, because it allows immunological surveillance and possible transplantation [35, 51].

9. Special Considerations: Population Diversity and Epidemiology

9.1. What is known and what is not

There is no population-based estimate of prevalence or incidence for any form of proteasome-associated autoinflammatory syndrome. Orphanet assigns the entity a point prevalence class below 1 per 1,000,000 worldwide, with 40 documented cases, which is a count of published cases and not a measure of frequency, and applies to the group as a whole, within which biallelic PSMB8 is the most common cause and the dominant forms are rare [11, 24]. The literature consists of small series: 4 children in the founding CANDLE description, 3 adults in the JMP description, 9 patients in the first genetic series, 8 in the series that established digenic inheritance, 10 CANDLE patients in the largest treatment cohort and 46 patients in the 2026 systematic review, which pools published cases from several sources [23, 31, 40, 41, 45, 80]. Any local rate depends on which allele circulates in the population, so no figure can be generalised [11].

9.2. Japan

At the time the gene was identified, Japanese national surveillance confirmed about ten living patients with Nakajo-Nishimura syndrome, which makes this the best available approximation to a national numerator for this disease; the study used stored fibroblasts from an autopsy case for genetic analysis and obtained consent from six living patients [27]. The PSMB8 variant p.(Gly201Val) is a founder variant in Japan, with the effect demonstrated by haplotype analysis, and the patients described came from consanguineous families [27, 43, 44]. The mouse model created in 2026 reproduces this founder variant in the homozygous state [59]. The PSMB9 variant p.(Gly156Asp) was described in two unrelated Japanese infants and, independently, in a third Japanese patient [50, 65].

9.3. Iberian Peninsula, Mexico and other Hispanic populations

The PSMB8 variant c.224C>T, p.(Thr75Met), has been described in Portuguese, Mexican, Spanish and Hispanic patients. The two pedigrees of the founding JMP description were Portuguese (followed in Lisbon) and Mexican, with non-consanguineous parents originating, respectively, from a small village and a small ranch — two settings of population isolates [40, 42]. In the series of nine CANDLE patients, three were Spanish and three Hispanic, and of the four homozygotes for p.(Thr75Met) only two shared the haplotype, which led the authors to interpret the position as a mutational hotspot rather than a single founder allele [45]. The variant was not present in 750 control chromosomes [45]. The available Portuguese data are limited to this family, described at the Hospital de Santa Maria in Lisbon; there are no national series or frequency estimates for Portugal in the sources consulted [40].

9.4. Brazil

The Brazilian patients described are concentrated in PRAAS5 due to PSMB10: three unrelated boys, aged 2 to 6.5 years and children of non-consanguineous parents, with onset in the first days of life, shared the p.(Phe14del) variant, inherited from the mother in all, on one allele, with a different second variant in each; the authors interpret it as a probable founder variant in Brazilian patients of Indigenous ancestry [9, 49]. The same series included patients from Russia and the United States, with other combinations of proteasome variants [49]. There are no Brazilian frequency estimates in the sources consulted.

9.5. Other regions

Patients have been described from Israel (an Ashkenazi Jewish boy homozygous for p.(Cys135Ter)), Ireland (two siblings with a digenic PSMB8/PSMB4 genotype), Algeria (the first patient with PRAAS5), Türkiye (family with a dominant PSMB10 variant), the Middle East (the first patient with POMP variants, of Palestinian parents according to OMIM and of Lebanese origin according to Torrelo), and from the United States and Western Europe in several series [2, 6, 13, 23, 45, 48, 71]. The distribution of cases, from Japan to the Americas, Europe, North Africa and the Middle East, shows that the disease is not exclusive to any region. The absence of data from many countries reflects the absence of studies, not the absence of patients.

9.6. Special populations

The disease is essentially paediatric at diagnosis, but patients reach adulthood with accumulated disability, as shown by the three adults aged 26 to 34 years in the JMP series and the patients in the expanded access programme aged up to 24 years at entry [31, 40]. Pregnancy is not documented in the sources consulted, and the contraindication of baricitinib in pregnancy constrains reproductive planning [112]. In infants, the doses and tolerability of Janus kinase inhibitors are the practical questions, with no data below 8.5 kg in the programme that generated the dosing proposal [106].

9.7. Reference networks and patient resources

The relevant European reference network is ERN-RITA, dedicated to rare immunological diseases, including immunodeficiencies, autoinflammatory and autoimmune diseases, with a directory of centres, patient pathways, a mobility programme and its own registry [37]. Among the associations, the Autoinflammatory Alliance, in the United States, maintains information, a database of autoinflammatory diseases and support materials for patients and families [131]. Natural history studies open to recruitment offer access to specialist assessment: the European cohort of type I interferonopathies of the Institut Imagine in Paris and the natural history protocol for autoinflammatory diseases of the National Institutes of Health [123, 124]. The consensus recommends that patients without a molecular diagnosis, despite a suggestive phenotype, be referred to specialist or research centres [5].

9.8. KLICK syndrome and neurodevelopmental proteasomopathies

There are no prevalence estimates for any of these entities: Orphanet assigns KLICK syndrome an unknown prevalence class, and Stankiewicz-Isidor syndrome has no entity of its own, so the available figures are counts of published cases [11]. The KLICK patients in the genetic study came from Spain, Italy, the Netherlands, Sweden and Norway, the first family was Spanish, and cases have been described in France and Brazil, the latter in a consanguineous family from Minas Gerais [33, 67, 81, 84, 87]. The presence of the same deletion on at least five haplotypes indicates that it arose several times and is not a marker of a founder population, so the syndrome can be expected in any population [33]. The sources consulted do not describe cases in Portugal, which does not demonstrate that none exist.

The cohorts of the neurodevelopmental proteasomopathies result from international collaborations coordinated from Nantes and Greifswald, with centres in Europe and North America [30, 34, 73]. The recessive form of PSMC3 was described in a single extended consanguineous family from a village in the Kayseri region of Türkiye [74]. The predominance of de novo variants means that the distribution of these diseases follows access to genomic sequencing rather than geographical origin.

The relevant European networks are ERN-Skin, with an ichthyosis subgroup, for KLICK syndrome, and ERN ITHACA, dedicated to rare malformation syndromes, intellectual disability and other neurodevelopmental disorders, for the neurodevelopmental proteasomopathies [38, 39]. Among patient associations for ichthyosis, the Foundation for Ichthyosis & Related Skin Types, in the USA, took part in the 2021 consensus on the use of retinoids in these diseases [36]. The Simons Searchlight registry collects data from people with PSMD12 variants [130].

10. Illustrative Clinical Case

The case that established PRAAS4 due to PSMG2 illustrates the typical diagnostic journey and the response to Janus kinase inhibition [47].

A white girl presented at 4 months of age with oedema of the extremities. At 7 months, erythematous-violaceous nodules appeared on the upper and lower limbs, with a histological diagnosis of panniculitis. At 2 years, during an admission for recurrent fever episodes, generalised lipodystrophy, mild gross motor and speech delay, raised acute-phase reactants, hepatosplenomegaly and diffuse lymphadenopathy were documented; height and weight were normal and the immunodeficiency work-up showed no abnormalities. At 5 years she developed severe autoimmune haemolytic anaemia with a positive direct antiglobulin (Coombs) test and haemoglobin of 3.1 g/dL, which required three transfusions, corticosteroid pulses, four infusions of rituximab and a 19-month course of azathioprine with intravenous immunoglobulin every 4 weeks. Prednisolone, started at 5 years, improved the panniculitis and hepatosplenomegaly, but the lipoatrophy and joint contractures continued to progress. At 7 years, magnetic resonance imaging showed active myositis and muscle atrophy of the limbs, with marked fatty infiltration of the thigh muscles, and brain computed tomography revealed faint bilateral basal ganglia calcifications. On the basis of these findings, a clinical diagnosis of probable CANDLE was made at 8 years and 10 months [47].

A clinical genetic panel for autoinflammatory diseases, which included the then known CANDLE genes, was negative. Exome sequencing, under a research protocol, identified two novel variants at conserved amino acids of PSMG2: the frameshift deletion c.666_667del, p.(Tyr223SerfsTer2), inherited from the father, and the missense variant c.675T>G, p.(Asn225Lys), inherited from the mother, both confirmed by Sanger sequencing and absent from public databases (gnomAD, ESP6500, 1000 Genomes) and from 308 local exomes. The interferon signature, measured during active disease with a 28-gene NanoString assay, was 533.6, comparable to that of patients with genetically confirmed CANDLE (738.2±754.4) and with SAVI (495.6±225.9), and much higher than that of healthy controls (−4.8±9.53). The patient's fibroblasts had all three proteolytic activities reduced, accumulation of ubiquitin conjugates similar to that of disease controls, markedly reduced PAC1 and PAC2 proteins and fewer assembled proteasome complexes [47]. Figure 1 of the article documents the pedigree with the heterozygous parents, the nodular rash at 7 months, the lipoatrophy of the face and upper limbs at 7 years, the basal ganglia calcifications and the interferon score (https://pmc.ncbi.nlm.nih.gov/articles/PMC6565382/figure/F1/).

The diagnosis of CANDLE/PRAAS prompted treatment with ruxolitinib, started at 8 years at 5 mg twice daily and increased with weight gain. The improvement was marked: progression of the lipoatrophy stopped, the hepatosplenomegaly and diffuse lymphadenopathy disappeared, the contractures improved markedly except for the left ankle and the inflammatory markers normalised, which allowed prednisolone to be reduced from 0.6 mg/kg/day to discontinuation after 14 months, without the need to reintroduce it. Figure E5 of the same article compares the laboratory parameters before and 14 months after starting ruxolitinib (https://pmc.ncbi.nlm.nih.gov/articles/PMC6565382/figure/F7/) [47].

This report is not a routine protocol. The choice of ruxolitinib predates the EULAR/ACR consensus, which names baricitinib as the most used Janus kinase inhibitor and mentions ruxolitinib and tofacitinib as alternatives, and all use is off-label [5]. On the other hand, the case shows three points that the guidelines take up: a panel that does not include all the proteasome genes may be negative in a disease that the exome identifies; the interferon signature supports the suspicion but does not point to the gene; and established damage, here the lipoatrophy and the ankle contracture, does not regress with treatment, unlike active inflammation [5, 47].

11. Prognosis and Long-Term Follow-up

11.1. Prognosis

The prognosis without targeted treatment is poor, and the EULAR/ACR consensus formulates it as a principle: untreated chronic systemic and organ inflammation results in progressive damage, early morbidity and increased mortality, with damage accumulating over time and becoming manifest later in life [5]. The data come from small series. Of the nine CANDLE patients in the first genetic series, one died at 14 years [45]. One of the four patients in the founding description died in an episode of carditis [2]. The sister of the Portuguese patient in the JMP series died suddenly at 8 years, and a Japanese patient included in the comparison in that series died at 47 years of congestive heart failure [40]. In Nakajo-Nishimura syndrome, most patients die of respiratory or heart failure, and corticosteroids control inflammation without halting the wasting [27]. The siblings described by Nakajo had cardiomegaly and heart failure [1]. The course is variable: some patients die in an organ flare and others survive for a long time with increasing disability, so specialist follow-up is lifelong [2, 11]. Acute inflammatory flares, which can affect almost any organ, are the main determinant of mortality and some have been fatal [2].

With Janus kinase inhibitors, half of the CANDLE patients in the expanded access programme achieved lasting clinical remission, with normalisation of inflammatory markers and the interferon signature and permanent discontinuation of corticosteroids, and catch-up growth and recovery of bone mineral density were observed; there were no deaths in 63 patient-years of exposure in that programme [31]. Remission is described mainly in CANDLE/PRAAS, and not in the other interferonopathies [5]. Prolonged normalisation of the interferon signature was observed only in the subset with recessive PSMB8 disease [76]. In the dominant forms with severe combined immunodeficiency due to PSMB10, the prognosis is guarded even with transplantation: of the six infants described, four died [51]. Established damage — lipoatrophy, contractures, calcifications, organ loss — is irreparable, which makes the time to diagnosis a prognostic determinant [5].

11.2. Surveillance table

The table follows the monitoring proposal of the EULAR/ACR consensus, with the general rows and those specific to CANDLE/PRAAS. The frequency of visits follows clinical activity; in the absence of active disease, the consensus proposes visits every 3 months to assess activity and drug toxicity [5].

11.3. Goals and instruments of follow-up

With current strategies, inflammatory remission is achievable only in a subset of patients, so the usual goal is to reduce systemic and organ inflammation and prevent or limit progression of damage, which requires treatment adjustments and close monitoring [5]. Disease activity and burden should be measured regularly, with validated instruments, patient-reported measures and quality-of-life assessment, and with recording of school or work days lost (point 14) [5]. Growth and development are assessed at each visit (point 15) [5]. The flare criteria developed in the NIH cohort serve to monitor activity and to design dose-reduction studies, and adding the interferon score increases sensitivity for detecting flares, especially in lymphopenic patients with low CRP [76].

11.4. Prognosis and surveillance of KLICK syndrome and the neurodevelopmental proteasomopathies

KLICK syndrome is lifelong and limited to the skin; the sources consulted describe no attributable mortality, and the functional prognosis depends on the severity of the keratoderma, the constricting bands and the finger deformities, with improvement on retinoids that is lost on discontinuation [33, 87]. An aggressive squamous cell carcinoma described in one patient justifies examination of new or ulcerated lesions [85]. In the neurodevelopmental proteasomopathies, prognosis is determined by the degree of intellectual disability, epilepsy and malformations; there are no survival data, and the few adults described with PSMD12 variants developed short stature and obesity from late childhood [29, 79]. No surveillance protocol has been published for any of these diseases; the table gathers the assessments recommended for the documented risks and, in the case of acitretin, the monitoring of the 2021 consensus and the European guidelines.

12. Summary and Concluding Remarks

Proteasome-associated autoinflammatory syndromes form a class of type I interferonopathies in which loss of function of subunits or chaperones of the proteasome and immunoproteasome prevents the removal of ubiquitinated proteins, generates proteotoxic stress and induces constitutive type I interferon production. The same disease was described three times under different names — Nakajo-Nishimura syndrome in Japan from 1939, CANDLE and JMP syndrome in 2010 — before identification of biallelic PSMB8 variants showed that it was a single entity, now numbered by OMIM as PRAAS1 and extended to PRAAS2 to PRAAS6 by POMP, PSMB4, PSMG2, PSMB10 and PSMB9, with digenic forms combining an inducible and a constitutive subunit and with dominant-negative forms of PSMB8, PSMB9 and PSMB10 that add immunodeficiency to the autoinflammatory picture.

The classic phenotype begins in the first weeks of life with daily fever, often below 38.3 °C and with good general condition, acral pernio-like lesions in infancy and annular erythematous-purpuric plaques that leave residual macules, violaceous eyelid oedema and panniculitis nodules whose histology, with an immature myeloid infiltrate and karyorrhexis, can be mistaken for leukaemia cutis. The panniculitis destroys subcutaneous fat and produces the progressive lipodystrophy of the face and upper half of the body that gives the patient an unmistakable appearance; it is followed by myositis, disabling contractures of the hands and feet, microcytic anaemia, hepatosplenomegaly, faltering growth, metabolic syndrome with hepatic steatosis and, in some patients, pulmonary hypertension, basal ganglia calcifications and inflammatory organ flares, some of them fatal. C-reactive protein and erythrocyte sedimentation rate are raised, unlike in Aicardi-Goutières syndrome, and the interferon signature in blood is chronically increased, although it may be negative in mild disease or under glucocorticoids.

The diagnosis requires genetic confirmation. Skin biopsy has the highest yield among non-molecular tests, and the interferon signature supports the suspicion without pointing to the gene. Next-generation sequencing should include all proteasome genes, POMP and PSMG2, as well as the recognised mimics, because a negative PSMB8 result does not exclude the disease, an isolated heterozygous variant may be half of a digenic genotype or a dominant-negative variant, frameshift variants in the penultimate exon of POMP escape nonsense-mediated decay and the variant of the allelic disorder KLICK lies outside the coding regions. Study of the parents is part of the diagnosis, not an extra, because it establishes phase, de novo origin and digenic inheritance, and with them the recurrence risk, which ranges from 25% in the recessive and digenic forms to a low risk for siblings in the dominant de novo forms.

Treatment changed in nature with Janus kinase inhibition, which blocks signalling downstream of the interferon receptor. The conventional drugs for autoinflammation do not work here, and lack of response is itself a diagnostic clue: colchicine and dapsone are ineffective, tumour necrosis factor antagonists do not help and may worsen the disease, and interleukin-1 blockade fails. No drug has an approved indication in any jurisdiction consulted, so all targeted treatment is off-label and should be managed by experienced centres, with monitoring for BK polyomavirus reactivation, cytopenias, infections, lipids and liver function tests, and with the understanding that reducing or stopping the drug causes withdrawal flares. Haematopoietic stem cell transplantation corrects proteasome function and has resolved autoinflammation in refractory patients, but in the dominant forms with severe combined immunodeficiency the peri-transplant inflammatory complications were severe and mortality was high. Interferon receptor blockade with anifrolumab showed an additive effect in a small series and remains investigational.

Two conclusions guide practice. The first is that established damage — lipoatrophy, contractures, calcifications, loss of organ function — is irreversible, whereas active inflammation responds; this makes early recognition of the skin lesions and low-grade fever in the infant, followed by broad genetic testing, the prognostic determinant within the clinician's reach. The second is that the therapeutic evidence comes from an expanded access programme, an open-label trial with nine patients terminated by the sponsor and a review of published cases, without any randomised study; practice rests on the international consensus and on a dosing proposal derived from population pharmacokinetics, and response figures cannot be compared between studies. The disease is ultra-rare, without a prevalence estimate, with recurrent alleles identified in Japan, the Iberian Peninsula and Mexico and Brazil, and patients benefit from lifelong follow-up at a specialist centre, access to reference networks and the ongoing natural history studies.

KLICK syndrome and the neurodevelopmental proteasomopathies complete the spectrum of proteasome diseases without clinical autoinflammation. KLICK syndrome results from homozygosity for a recurrent deletion in the 5' untranslated region of POMP, which shifts production towards poorly translated transcripts and reduces POMP in the differentiated layers of the epidermis; it gives an ichthyosis with sclerosing palmoplantar keratoderma, pseudoainhum and linear keratotic papules in the flexures, without systemic disease, is diagnosed by a molecular study that covers the non-coding region and is treated with topical care and oral acitretin at the lowest effective dose, with a suppressive effect and with the monitoring proper to retinoids. The neurodevelopmental proteasomopathies, almost all dominant and de novo and concentrated in the 19S regulatory particle, combine developmental delay, intellectual disability and malformations of several organs with a laboratory interferon signature rarely accompanied by clinical inflammation; they are diagnosed by exome or genome with copy-number variant analysis, have no targeted treatment and require multidisciplinary care and counselling that takes into account the documented familial transmission.

13. References

1. OMIM. #256040 Proteasome-associated autoinflammatory syndrome 1; PRAAS1. Baltimore: Johns Hopkins University. Available from: https://omim.org/entry/256040
2. Torrelo A. CANDLE Syndrome As a Paradigm of Proteasome-Related Autoinflammation. Front Immunol. 2017;8:927. doi:10.3389/fimmu.2017.00927. PMID: 28848544.
3. 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. PMID: 41608114.
4. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 464–471 (Immunodeficiency and recurrent infection).
5. Cetin Gedik K, Lamot L, Romano M, Demirkaya E, Piskin D, Torreggiani S, et al. The 2021 European Alliance of Associations for Rheumatology/American College of Rheumatology points to consider for diagnosis and management of autoinflammatory type I interferonopathies: CANDLE/PRAAS, SAVI and AGS. Ann Rheum Dis. 2022;81(5):601-613. doi:10.1136/annrheumdis-2021-221814. PMID: 35086813.
6. OMIM. #618048 Proteasome-associated autoinflammatory syndrome 2; PRAAS2. Available from: https://omim.org/entry/618048
7. OMIM. #617591 Proteasome-associated autoinflammatory syndrome 3; PRAAS3. Available from: https://omim.org/entry/617591
8. OMIM. #619183 Proteasome-associated autoinflammatory syndrome 4; PRAAS4. Available from: https://omim.org/entry/619183
9. OMIM. #619175 Proteasome-associated autoinflammatory syndrome 5; PRAAS5. Available from: https://omim.org/entry/619175
10. OMIM. #620796 Proteasome-associated autoinflammatory syndrome 6; PRAAS6. Available from: https://omim.org/entry/620796
11. GeneGuidance. Clinical curation of the genes PSMB8, PSMB4, PSMA3, POMP, PSMB9, PSMB10, PSMG2, PSMD12, PSMC3 and PSMC5 and CANDLE/PRAAS management library. Local GeneGuidance database; extracted 2026-10-31.
12. OMIM. #620807 Immunodeficiency 121 with autoinflammation; IMD121. Available from: https://omim.org/entry/620807
13. Wijngaard R, van der Made CI, Kalkan Uçar S, Ramakrishnan G, Wang M, Brand J, et al. Monoallelic PSMB8 variants cause PRAAS with immunodeficiency through impaired immunoproteasome assembly. Am J Hum Genet. 2026;113(6):1214-1232. doi:10.1016/j.ajhg.2026.04.015. PMID: 42167218.
14. OMIM. #601952 Keratosis linearis with ichthyosis congenita and sclerosing keratoderma; KLICK. Available from: https://omim.org/entry/601952
15. Akiyama M, Choate K, Hernández-Martín Á, Aldwin-Easton M, Bodemer C, Gostyński A, et al. Nonsyndromic epidermal differentiation disorders: a new classification toward pathogenesis-based therapy. Br J Dermatol. 2025;193(4):619-641. doi:10.1093/bjd/ljaf154. PMID: 40308026.
16. OMIM. #617516 Stankiewicz-Isidor syndrome; STISS. Available from: https://omim.org/entry/617516
17. OMIM. #621539 Ebstein-Bezieau neurodevelopmental syndrome; EBNDS. Available from: https://omim.org/entry/621539
18. OMIM. #619354 ?Deafness, cataract, impaired intellectual development, and polyneuropathy; DCIDP. Available from: https://omim.org/entry/619354
19. OMIM. #621565 Yu-Kury neurodevelopmental syndrome; YKNS. Available from: https://omim.org/entry/621565
20. OMIM. #620038 ?Neurodevelopmental disorder with microcephaly, hypotonia, and absent language; NEDMHAL. Available from: https://omim.org/entry/620038
21. OMIM. #620071 ?Birk-Aharoni syndrome. Available from: https://omim.org/entry/620071
22. Deb W, Rosenfelt C, Vignard V, Papendorf JJ, Möller S, Wendlandt M, et al. PSMD11 loss-of-function variants correlate with a neurobehavioral phenotype, obesity, and increased interferon response. Am J Hum Genet. 2024;111(7):1352-1369. doi:10.1016/j.ajhg.2024.05.016. PMID: 38866022.
23. Brehm A, Liu Y, Sheikh A, Marrero B, Omoyinmi E, Zhou Q, et al. Additive loss-of-function proteasome subunit mutations in CANDLE/PRAAS patients promote type I IFN production. J Clin Invest. 2015;125(11):4196-4211. doi:10.1172/jci81260. PMID: 26524591.
24. Orphanet. Proteasome-associated autoinflammatory syndrome (ORPHA:324977); subtypes CANDLE syndrome (ORPHA:325004), Nakajo-Nishimura syndrome (ORPHA:2615) and JMP syndrome (ORPHA:324999). Available from: https://www.orpha.net/en/disease/detail/324977
25. Orphanet. ORPHA:281201 Keratosis linearis-ichthyosis congenita-sclerosing keratoderma syndrome. Available from: https://www.orpha.net/en/disease/detail/281201
26. Orphanet. ORPHA:528084 Non-specific syndromic intellectual disability (associated gene: PSMD12) and ORPHA:529962 17q24.2 microdeletion syndrome. Available from: https://www.orpha.net/en/disease/detail/528084
27. Arima K, Kinoshita A, Mishima H, Kanazawa N, Kaneko T, Mizushima T, et al. Proteasome assembly defect due to a proteasome subunit beta type 8 (PSMB8) mutation causes the autoinflammatory disorder, Nakajo-Nishimura syndrome. Proc Natl Acad Sci U S A. 2011;108(36):14914-14919. doi:10.1073/pnas.1106015108. PMID: 21852578.
28. Takeichi T, Akiyama M. KLICK Syndrome Linked to a POMP Mutation Has Features Suggestive of an Autoinflammatory Keratinization Disease. Front Immunol. 2020;11:641. doi:10.3389/fimmu.2020.00641. PMID: 32425927.
29. Cuinat S, Bézieau S, Deb W, Mercier S, Vignard V, Isidor B, et al. Understanding neurodevelopmental proteasomopathies as new rare disease entities: A review of current concepts, molecular biomarkers, and perspectives. Genes Dis. 2024;11(6):101130. doi:10.1016/j.gendis.2023.101130. PMID: 39220754.
30. Küry S, Stanton JE, van Woerden GM, Bosc-Rosati A, Hsieh TC, Bray L, et al. Investigating the neuronal role of the proteasomal ATPase subunit gene PSMC5 in neurodevelopmental proteasomopathies. Nat Commun. 2025;16(1):10545. doi:10.1038/s41467-025-65556-8. PMID: 41298377.
31. Sanchez GAM, Reinhardt A, Ramsey S, Wittkowski H, Hashkes PJ, Berkun Y, et al. JAK1/2 inhibition with baricitinib in the treatment of autoinflammatory interferonopathies. J Clin Invest. 2018;128(7):3041-3052. doi:10.1172/jci98814. PMID: 29649002.
32. Thomas BR, O'Toole EA. Diagnosis and Management of Inherited Palmoplantar Keratodermas. Acta Derm Venereol. 2020;100(7):adv00094. doi:10.2340/00015555-3430. PMID: 32147745.
33. Dahlqvist J, Klar J, Tiwari N, Schuster J, Törmä H, Badhai J, et al. A single-nucleotide deletion in the POMP 5' UTR causes a transcriptional switch and altered epidermal proteasome distribution in KLICK genodermatosis. Am J Hum Genet. 2010;86(4):596-603. doi:10.1016/j.ajhg.2010.02.018. PMID: 20226437.
34. Küry S, Besnard T, Ebstein F, Khan TN, Gambin T, Douglas J, et al. De Novo Disruption of the Proteasome Regulatory Subunit PSMD12 Causes a Syndromic Neurodevelopmental Disorder. Am J Hum Genet. 2017;100(2):352-363. doi:10.1016/j.ajhg.2017.01.003. PMID: 28132691.
35. Kanegane H, Miyamoto S, Kamiya T, Ohnishi H, Nishikomori R, Gennery AR, et al. Allogeneic hematopoietic cell transplantation for autoinflammatory disorders. Int J Hematol. 2025;122(2):190-198. doi:10.1007/s12185-025-04021-0. PMID: 40498270.
36. Zaenglein AL, Levy ML, Stefanko NS, Benjamin LT, Bruckner AL, Choate K, et al. Consensus recommendations for the use of retinoids in ichthyosis and other disorders of cornification in children and adolescents. Pediatr Dermatol. 2021;38(1):164-180. doi:10.1111/pde.14408. PMID: 33169909; PMC7984068.
37. European Reference Network on Rare Immunodeficiency, Autoinflammatory and Autoimmune Diseases (ERN-RITA). Available from: https://ern-rita.org/
38. European Reference Network for rare and complex diseases of the skin (ERN-Skin). Available from: https://ern-skin.eu/
39. ERN ITHACA — European Reference Network on Rare Malformation Syndromes, Intellectual and Other Neurodevelopmental Disorders. Available from: https://ern-ithaca.eu/
40. Garg A, Hernandez MD, Sousa AB, Subramanyam L, Martínez de Villarreal L, dos Santos HG, et al. An autosomal recessive syndrome of joint contractures, muscular atrophy, microcytic anemia, and panniculitis-associated lipodystrophy. J Clin Endocrinol Metab. 2010;95(9):E58-63. doi:10.1210/jc.2010-0488. PMID: 20534754.
41. Torrelo A, Patel S, Colmenero I, Gurbindo D, Lendínez F, Hernández A, et al. Chronic atypical neutrophilic dermatosis with lipodystrophy and elevated temperature (CANDLE) syndrome. J Am Acad Dermatol. 2010;62(3):489-495. doi:10.1016/j.jaad.2009.04.046. PMID: 20159315.
42. Agarwal AK, Xing C, DeMartino GN, Mizrachi D, Hernandez MD, Sousa AB, et al. PSMB8 encoding the β5i proteasome subunit is mutated in joint contractures, muscle atrophy, microcytic anemia, and panniculitis-induced lipodystrophy syndrome. Am J Hum Genet. 2010;87(6):866-872. doi:10.1016/j.ajhg.2010.10.031. PMID: 21129723.
43. Kitamura A, Maekawa Y, Uehara H, Izumi K, Kawachi I, Nishizawa M, et al. A mutation in the immunoproteasome subunit PSMB8 causes autoinflammation and lipodystrophy in humans. J Clin Invest. 2011;121(10):4150-4160. doi:10.1172/jci58414. PMID: 21881205.
44. OMIM. *177046 Proteasome 20S subunit beta 8; PSMB8. Available from: https://omim.org/entry/177046
45. Liu Y, Ramot Y, Torrelo A, Paller AS, Si N, Babay S, et al. Mutations in proteasome subunit β type 8 cause chronic atypical neutrophilic dermatosis with lipodystrophy and elevated temperature with evidence of genetic and phenotypic heterogeneity. Arthritis Rheum. 2012;64(3):895-907. doi:10.1002/art.33368. PMID: 21953331.
46. Poli MC, Ebstein F, Nicholas SK, de Guzman MM, Forbes LR, Chinn IK, et al. Heterozygous Truncating Variants in POMP Escape Nonsense-Mediated Decay and Cause a Unique Immune Dysregulatory Syndrome. Am J Hum Genet. 2018;102(6):1126-1142. doi:10.1016/j.ajhg.2018.04.010. PMID: 29805043.
47. de Jesus AA, Brehm A, VanTries R, Pillet P, Parentelli AS, Montealegre Sanchez GA, et al. Novel proteasome assembly chaperone mutations in PSMG2/PAC2 cause the autoinflammatory interferonopathy CANDLE/PRAAS4. J Allergy Clin Immunol. 2019;143(5):1939-1943.e8. doi:10.1016/j.jaci.2018.12.1012. PMID: 30664889.
48. Sarrabay G, Méchin D, Salhi A, Boursier G, Rittore C, Crow Y, et al. PSMB10, the last immunoproteasome gene missing for PRAAS. J Allergy Clin Immunol. 2020;145(3):1015-1017.e6. doi:10.1016/j.jaci.2019.11.024. PMID: 31783057.
49. Papendorf JJ, Ebstein F, Alehashemi S, Piotto DGP, Kozlova A, Terreri MT, et al. Identification of eight novel proteasome variants in five unrelated cases of proteasome-associated autoinflammatory syndromes (PRAAS). Front Immunol. 2023;14:1190104. doi:10.3389/fimmu.2023.1190104. PMID: 37600812.
50. Kanazawa N, Hemmi H, Kinjo N, Ohnishi H, Hamazaki J, Mishima H, et al. Heterozygous missense variant of the proteasome subunit β-type 9 causes neonatal-onset autoinflammation and immunodeficiency. Nat Commun. 2021;12(1):6819. doi:10.1038/s41467-021-27085-y. PMID: 34819510.
51. van der Made CI, Kersten S, Chorin O, Engelhardt KR, Ramakrishnan G, Griffin H, et al. Expanding the PRAAS spectrum: De novo mutations of immunoproteasome subunit β-type 10 in six infants with SCID-Omenn syndrome. Am J Hum Genet. 2024;111(4):791-804. doi:10.1016/j.ajhg.2024.02.013. PMID: 38503300.
52. Wolfgramm S, Alehashemi S, Wendlandt M, Thiel FG, de Jesus AA, Papendorf JJ, et al. A de novo dominant-negative PSMB8 mutation causes severe CANDLE/PRAAS due to arrested proteasome biogenesis. Ann Rheum Dis. 2026;85(4):715-729. doi:10.1016/j.ard.2025.10.021. PMID: 41253591.
53. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 39.1, Nucleotide and Nucleic Acid Metabolism, section 39.1.2.
54. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, Clinical Approach to Inborn Errors of Metabolism in Paediatrics, sections 1.6.2 (Dermatology, Table 1.34) and 1.6.7 (Immunology).
55. Cetin Gedik K, Lamot L, Romano M, Demirkaya E, Piskin D, Torreggiani S, et al. The 2021 European Alliance of Associations for Rheumatology/American College of Rheumatology Points to Consider for Diagnosis and Management of Autoinflammatory Type I Interferonopathies: CANDLE/PRAAS, SAVI, and AGS. Arthritis Rheumatol. 2022;74(5):735-751. doi:10.1002/art.42087. PMID: 35315249.
56. Ebstein F, Poli Harlowe MC, Studencka-Turski M, Krüger E. Contribution of the Unfolded Protein Response (UPR) to the Pathogenesis of Proteasome-Associated Autoinflammatory Syndromes (PRAAS). Front Immunol. 2019;10:2756. doi:10.3389/fimmu.2019.02756. PMID: 31827472.
57. Wolfgramm S, Alvarez FL, Thiel FG, Wendlandt M, Krüger E. Molecular mechanisms in rare proteasomopathies. EULAR Rheumatol Open. 2026;2(1):119-142. doi:10.1016/j.ero.2025.12.005. PMID: 42370079.
58. Goldbach-Mansky R, Alehashemi S, de Jesus AA. Emerging concepts and treatments in autoinflammatory interferonopathies and monogenic systemic lupus erythematosus. Nat Rev Rheumatol. 2025;21(1):22-45. doi:10.1038/s41584-024-01184-8. PMID: 39623155.
59. Hara T, Kinoshita A, Hamazaki J, Hemmi H, Kato T, Sasaki I, et al. The homozygous founder Psmb8 variant of Nakajo-Nishimura syndrome/proteasome-associated autoinflammatory syndrome causes panniculitis-associated lipoatrophy and a shortened lifespan in mice. Sci Rep. 2026;16(1):15039. doi:10.1038/s41598-026-51190-x. PMID: 42209561.
60. Winkler CW, Schwarz B, Williams K, Alehashemi S, Foliaki ST, Snow J, et al. Proteasome mutations associated with CANDLE syndrome cause altered neuronal development by dysregulating polyamine synthesis. J Clin Invest. 2026;136(19):e199140. doi:10.1172/jci199140. PMID: 42550603.
61. Huber EM, Heinemeyer W, Groll M. Proteasome-associated autoinflammatory syndromes: The impact of mutations in proteasome subunits on particle assembly, structure, and activity. Structure. 2025;33(12):2084-2094.e5. doi:10.1016/j.str.2025.09.004. PMID: 41033310.
62. López-Cano J, Monserrat-García MT, Goldbach-Mansky R, Alehashemi S, Rubio-Murillo M, Montero-Valladares C, et al. Neonatal-Onset CANDLE/PRAAS due to PSMB8 p.Gly209Arg: Clinical Course and Baricitinib Response. Int J Dermatol. 2026;65(6):1314-1316. doi:10.1111/ijd.70214. PMID: 41454462.
63. Méchin D, Tusseau M, Broly M, Sanlaville D, Viel S, Sarrabay G, et al. A new digenic inheritance of proteasome-associated autoinflammatory syndrome involving the PSMA6 gene. J Allergy Clin Immunol Pract. 2026;14(2):524-526.e1. doi:10.1016/j.jaip.2025.10.039. PMID: 41205671.
64. Verhoeven D, Schonenberg-Meinema D, Ebstein F, Papendorf JJ, Baars PA, van Leeuwen EMM, et al. Hematopoietic stem cell transplantation in a patient with proteasome-associated autoinflammatory syndrome (PRAAS). J Allergy Clin Immunol. 2022;149(3):1120-1127.e8. doi:10.1016/j.jaci.2021.07.039. PMID: 34416217.
65. Kataoka S, Kawashima N, Okuno Y, Muramatsu H, Miwata S, Narita K, et al. Successful treatment of a novel type I interferonopathy due to a de novo PSMB9 gene mutation with a Janus kinase inhibitor. J Allergy Clin Immunol. 2021;148(2):639-644. doi:10.1016/j.jaci.2021.03.010. PMID: 33727065.
66. VariantValidator. Validation and normalisation of HGVS variants, GRCh38, MANE Select. Available from: https://rest.variantvalidator.org/
67. Morice-Picard F, Jonca N, Pichery M, Mermin D, Leauté-Labrèze C, Taïeb A, et al. KLICK syndrome: recognizable phenotype and hot-spot POMP mutation. J Eur Acad Dermatol Venereol. 2017;31(3):e154-e156. doi:10.1111/jdv.13898. PMID: 27503413.
68. Bernacchia A, Garcia AL, Autoinflammatory Disease Consortium, Raccio AG, Di Giovanni D. Disseminated BCG Disease in a Child with a Novel PSMG2 Deletion. J Clin Immunol. 2024;44(6):145. doi:10.1007/s10875-024-01738-1. PMID: 38847935.
69. OMIM. *176847 Proteasome 20S subunit beta 10; PSMB10. Available from: https://omim.org/entry/176847
70. Fournier B, Poirier L, Abramowski V, Merlin E, Deutsch H, Bastard P, et al. Novel de novo dominant PSMB10 variants in three patients with immune deficiency and liver disease. J Hum Immun. 2026;2(1):e20250096. doi:10.70962/jhi.20250096. PMID: 42170594.
71. Onat UI, Bülbül A, Sigli D, Sever EA, Ugurlu S, Ozdogan AH, et al. A Novel Dominantly Segregating PSMB10 Splice-Site Variant in Familial Autoinflammatory Disease with Immunoproteasome and Interferon-Related Dysregulation. Genes (Basel). 2026;17(9):997. doi:10.3390/genes17090997. PMID: 42792891.
72. ClinVar. VCV000000116 — NM_015932.5(POMP):c.-95delC (rs112368783), pathogenic classification. National Center for Biotechnology Information; accessed 1 November 2026. Available from: https://www.ncbi.nlm.nih.gov/clinvar/variation/116/
73. Ebstein F, Küry S, Most V, Rosenfelt C, Scott-Boyer MP, van Woerden GM, et al. PSMC3 proteasome subunit variants are associated with neurodevelopmental delay and type I interferon production. Sci Transl Med. 2023;15(698):eabo3189. doi:10.1126/scitranslmed.abo3189. PMID: 37256937.
74. Kröll-Hermi A, Ebstein F, Stoetzel C, Geoffroy V, Schaefer E, Scheidecker S, et al. Proteasome subunit PSMC3 variants cause neurosensory syndrome combining deafness and cataract due to proteotoxic stress. EMBO Mol Med. 2020;12(7):e11861. doi:10.15252/emmm.201911861. PMID: 32500975.
75. Yu ZQ, Carmichael J, Collins GA, D'Agostino MD, Lessard M, Firth HV, et al. PSMC5 insufficiency and P320R mutation impair proteasome function. Hum Mol Genet. 2024;33(17):1506-1523. doi:10.1093/hmg/ddae085. PMID: 38776958.
76. Cetin Gedik K, Ortega-Villa AM, Materne G, Rastegar A, Montealegre Sanchez GA, Reinhardt A, et al. Disease flares with baricitinib dose reductions and development of flare criteria in patients with CANDLE/PRAAS. Ann Rheum Dis. 2024;83(9):1181-1188. doi:10.1136/ard-2023-225463. PMID: 38653530.
77. Isidor B, Ebstein F, Hurst A, Vincent M, Bader I, Rudy NL, et al. Stankiewicz-Isidor syndrome: expanding the clinical and molecular phenotype. Genet Med. 2022;24(1):179-191. doi:10.1016/j.gim.2021.09.005. PMID: 34906456.
78. Khalil R, Kenny C, Hill RS, Mochida GH, Nasir R, Partlow JN, et al. PSMD12 haploinsufficiency in a neurodevelopmental disorder with autistic features. Am J Med Genet B Neuropsychiatr Genet. 2018;177(8):736-745. doi:10.1002/ajmg.b.32688. PMID: 30421579.
79. Feresin A, Spedicati B, Zampieri S, Morgan A, Magnolato A, Tesser A, et al. Does It Run in Your Family? Inherited Truncating PSMD12 Variants Broaden the Phenotypic Spectrum of Stankiewicz-Isidor Syndrome. Am J Med Genet A. 2025;197(4):e63953. doi:10.1002/ajmg.a.63953. PMID: 39641441.
80. Vahabi SM, Pourgholi E, Leblebici CB, Crisler WJ, Heidari S, Tung JK. JAK inhibitor therapy in CANDLE syndrome: a systematic review of clinical outcomes in 46 patients. Eur J Pediatr. 2026;185(8):637. doi:10.1007/s00431-026-07293-8. PMID: 42547647.
81. Pujol RM, Moreno A, Alomar A, de Moragas JM. Congenital ichthyosiform dermatosis with linear keratotic flexural papules and sclerosing palmoplantar keratoderma. Arch Dermatol. 1989;125(1):103-106. doi:10.1001/archderm.1989.01670130105016. PMID: 2521286.
82. Vahlquist A, Pontén F, Pettersson A. Keratosis linearis with ichthyosis congenita and sclerosing keratoderma (KLICK-syndrome): a rare, autosomal recessive disorder of keratohyaline formation? Acta Derm Venereol. 1997;77(3):225-227. doi:10.2340/0001555577225227. PMID: 9188877.
83. Oji V, Tadini G, Akiyama M, Blanchet Bardon C, Bodemer C, Bourrat E, et al. Revised nomenclature and classification of inherited ichthyoses: results of the First Ichthyosis Consensus Conference in Sorèze 2009. J Am Acad Dermatol. 2010;63(4):607-641. doi:10.1016/j.jaad.2009.11.020. PMID: 20643494.
84. Onnis G, Bourrat E, Jonca N, Dreyfus I, Severino-Freire M, Pichery M, et al. KLICK syndrome: an unusual phenotype. Br J Dermatol. 2018;178(6):1445-1446. doi:10.1111/bjd.16318. PMID: 29315485.
85. Horev L, Murad S, Maly A, Zlotogorski A. Aggressive cutaneous squamous cell carcinoma in a patient with KLICK. J Am Acad Dermatol. 2011;64(6):e128-e130. doi:10.1016/j.jaad.2010.12.030. PMID: 21571163.
86. Horev L, Babay S, Ramot Y, Saad-Edin B, Moorad S, Ingber A, et al. Mutations in two genes on chromosome 13 resulting in a complex hair and skin phenotype due to two rare genodermatoses: KLICK and autosomal recessive woolly hair/hypotrichosis simplex. Br J Dermatol. 2011;164(5):1113-1116. doi:10.1111/j.1365-2133.2011.10229.x. PMID: 21275938.
87. Baeta IG, Pereira AC, Guedes AC, Pereira LB. Do you know this syndrome? [Você conhece esta síndrome? Síndrome KLICK]. An Bras Dermatol. 2011;86(3):605-607. doi:10.1590/s0365-05962011000300036. PMID: 21738991.
88. Vinci M, Musumeci A, Papa C, Ragalmuto A, Saccone S, Federico C, et al. Strengthening the Role of PSMC5 as a Potential Gene Associated with Neurodevelopmental Disorders. Int J Mol Sci. 2025;26(13):6386. doi:10.3390/ijms26136386. PMID: 40650163.
89. Firth HV, Hurst JA. Neonatal (newborn) screening (NS). In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 762–763.
90. Kanazawa N, Ishii T, Takita Y, Nishikawa A, Nishikomori R. Efficacy and safety of baricitinib in Japanese patients with autoinflammatory type I interferonopathies (NNS/CANDLE, SAVI, And AGS). Pediatr Rheumatol Online J. 2023;21(1):38. doi:10.1186/s12969-023-00817-8. PMID: 37087470.
91. Kim H, de Jesus AA, Brooks SR, Liu Y, Huang Y, VanTries R, et al. Development of a Validated Interferon Score Using NanoString Technology. J Interferon Cytokine Res. 2018;38(4):171-185. doi:10.1089/jir.2017.0127. PMID: 29638206.
92. de Jesus AA, Hou Y, Brooks S, Malle L, Biancotto A, Huang Y, et al. Distinct interferon signatures and cytokine patterns define additional systemic autoinflammatory diseases. J Clin Invest. 2020;130(4):1669-1682. doi:10.1172/jci129301. PMID: 31874111.
93. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 200–203 (Intracranial calcification).
94. OMIM. #615934 STING-associated vasculopathy, infantile-onset; SAVI. Available from: https://omim.org/entry/615934
95. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. p. 172–173 (Spondyloenchondrodysplasia).
96. OMIM. #607944 Spondyloenchondrodysplasia with immune dysregulation; SPENCDI. Available from: https://omim.org/entry/607944
97. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. p. 61–62 (CINCA).
98. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 37.1, Mevalonate Kinase Deficiency.
99. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 32.3, Diseases with Predominant Immunologic or Hematological Symptoms, section 32.3.2.
100. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 852–853 (Berardinelli-Seip congenital lipodystrophy syndrome).
101. Banhos MEH, Vasconcelos MC, Ribeiro EM. Lipodistrofia congénita de Berardinelli-Seip. In: Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 239–241.
102. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 35.1, Inborn Errors of the Common Pathway and of Triglyceride Synthesis and Degradation.
103. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 807–814 (Alpha-1-antitrypsin deficiency), p. 810.
104. Firth HV, Hurst JA. Alpha-1-antitrypsin deficiency. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 338–339.
105. Mukherjee S, Sahu RK, Chandra NM, Dhakad U, Kumar P. Case Report: Reconciling genomic ambiguity with clinical certainty in refractory panniculitis. Front Immunol. 2026;17:1906895. doi:10.3389/fimmu.2026.1906895. PMID: 42812382.
106. Kim H, Brooks KM, Tang CC, Wakim P, Blake M, Brooks SR, et al. Pharmacokinetics, Pharmacodynamics, and Proposed Dosing of the Oral JAK1 and JAK2 Inhibitor Baricitinib in Pediatric and Young Adult CANDLE and SAVI Patients. Clin Pharmacol Ther. 2018;104(2):364-373. doi:10.1002/cpt.936. PMID: 29134648.
107. Kretzschmar G, Páez LP, Tan Z, Wang J, Gonzalez L, Mugabo CH, et al. Normalized Interferon Signatures and Clinical Improvements by IFNAR1 Blocking Antibody (Anifrolumab) in Patients with Type I Interferonopathies. J Clin Immunol. 2024;45(1):31. doi:10.1007/s10875-024-01826-2. PMID: 39441221.
108. Hoffman HM, Broderick L. JAK inhibitors in autoinflammation. J Clin Invest. 2018;128(7):2760-2762. doi:10.1172/jci121526. PMID: 29889100.
109. Alembic Pharmaceuticals. Methotrexate tablets, prescribing information (section 2.4, Recommended Dosage for Polyarticular Juvenile Idiopathic Arthritis). ANDA 217552. Silver Spring: U.S. Food and Drug Administration; revised 05/2026. Available from: https://open.fda.gov/
110. 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):1271-1285. doi:10.1136/ard-2023-224123. PMID: 37487610.
111. Eli Lilly and Company. A Study of Baricitinib (LY3009104) in Adult and Pediatric Japanese Participants With NNS/CANDLE, SAVI, and AGS. ClinicalTrials.gov identifier: NCT04517253. Available from: https://clinicaltrials.gov/study/NCT04517253
112. European Medicines Agency. Olumiant (baricitinib): EPAR — product information. Amsterdam: EMA; updated 17 September 2026. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/olumiant
113. Eli Lilly and Company. OLUMIANT (baricitinib) tablets, prescribing information. NDA 207924. Silver Spring: U.S. Food and Drug Administration; revised 09/2026. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=207924
114. Pfizer. XELJANZ (tofacitinib) tablets and oral solution, prescribing information (section 2.4, table 2). Silver Spring: U.S. Food and Drug Administration; 2026. Available from: https://open.fda.gov/
115. European Medicines Agency. Jakavi (ruxolitinib): EPAR — product information. Amsterdam: EMA; 2025. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/jakavi
116. INFARMED. Infomed — Olumiant 2 mg e 4 mg, comprimidos revestidos por película (med_id 615663 and 615786). Lisboa: INFARMED; accessed 31 October 2026. Available from: https://extranet.infarmed.pt/INFOMED-fo/
117. Incyte Corporation. JAKAFI (ruxolitinib) tablets, prescribing information. NDA 202192. Silver Spring: U.S. Food and Drug Administration; 2026. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=202192
118. AstraZeneca. SAPHNELO (anifrolumab-fnia), prescribing information. BLA 761123. Silver Spring: U.S. Food and Drug Administration; 2026. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=761123
119. European Medicines Agency. Saphnelo (anifrolumab): EPAR — product information. Amsterdam: EMA; 2026. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/saphnelo
120. Agência Nacional de Vigilância Sanitária. Dados abertos — medicamentos registados (DADOS_ABERTOS_MEDICAMENTOS.csv). Brasília: ANVISA; accessed 31 October 2026. Available from: https://dados.anvisa.gov.br/
121. Martinez C, Ebstein F, Nicholas SK, De Guzman M, Forbes LR, Delmonte OM, et al. HSCT corrects primary immunodeficiency and immune dysregulation in patients with POMP-related autoinflammatory disease. Blood. 2021;138(19):1896-1901. doi:10.1182/blood.2021011005. PMID: 34019630.
122. Eli Lilly and Company. Compassionate Use Protocol for the Treatment of Autoinflammatory Syndromes. ClinicalTrials.gov identifier: NCT01724580. Available from: https://clinicaltrials.gov/study/NCT01724580
123. Imagine Institute. Natural History of Type 1 Interferonopathies: Insights From a European Cohort. ClinicalTrials.gov identifier: NCT07040774. Available from: https://clinicaltrials.gov/study/NCT07040774
124. National Institute of Allergy and Infectious Diseases. Natural History, Pathogenesis, and Outcome of Autoinflammatory Diseases (NOMID/CAPS, DIRA, CANDLE, SAVI, NLRC4-MAS, Still's-like Diseases). ClinicalTrials.gov identifier: NCT02974595. Available from: https://clinicaltrials.gov/study/NCT02974595
125. ImmuneSensor Therapeutics. A Phase 1b Open-Label Study of IMSB301 in Subjects With a Type 1 Interferonopathy. ClinicalTrials.gov identifier: NCT07364513. Available from: https://clinicaltrials.gov/study/NCT07364513
126. Boehringer Ingelheim. Single-arm Open-label Trial of BI 3000202 in Adult Patients With Selected Type 1 Interferonopathies. ClinicalTrials.gov identifier: NCT06878365. Available from: https://clinicaltrials.gov/study/NCT06878365
127. Mazereeuw-Hautier J, Paller AS, Dreyfus I, Sprecher E, O'Toole E, Bodemer C, et al. Management of congenital ichthyoses: guidelines of care: Part one: 2024 update. Br J Dermatol. 2025;193(1):16-27. doi:10.1093/bjd/ljaf076. PMID: 40156154.
128. INFARMED. Neotigason 10 mg e 25 mg cápsulas (acitretina) — Resumo das Características do Medicamento. Lisboa: INFARMED; approved 14-10-2024. Available from: https://extranet.infarmed.pt/INFOMED-fo/ (med_id 6012)
129. U.S. Food and Drug Administration. Drugs@FDA: NDA 019821, Soriatane (acitretin) capsules (approved 1996-10-28; brand discontinued). Label consulted: Acitretin capsules USP, Sigmapharm Laboratories, DailyMed setid a6546625-acb8-460e-b34e-f795bfb3680a, revised 2026-08-05. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=a6546625-acb8-460e-b34e-f795bfb3680a
130. ClinicalTrials.gov. NCT01238250 — Online Study of People Who Have Genetic Changes and Features of Autism: Simons Searchlight (observational study, recruiting; accessed 1 November 2026). Available from: https://clinicaltrials.gov/study/NCT01238250
131. Autoinflammatory Alliance. Available from: https://www.autoinflammatory.org/