Familial autoinflammatory diseases: genetics, pathogenesis and treatment

Purpose of reviewThe systemic autoinflammatory diseases are characterized by seemingly unprovoked inflammation, without major involvement of the adaptive immune system. This review focuses mainly on a subset of these illnesses, the hereditary recurrent fevers, which include familial Mediterranean fever, the tumor necrosis factor receptor-associated periodic syndrome, the hyperimmunoglobulinemia D with periodic fever syndrome, and cryopyrin-associated periodic syndromes. This review elucidates how recent advances have impacted diagnosis, pathogenesis, and treatment. Recent findingsMore than 170 mutations have been identified in the four genes underlying the six hereditary recurrent fevers. Genetic testing has broadened the clinical and geographic boundaries of these illnesses, given rise to the concept of the cryopyrin-associated periodic syndromes as a disease spectrum, and permitted diagnosis of compound heterozygotes for mutations in two different hereditary recurrent fever genes. Genetics has also advanced our understanding of amyloidosis, a complication of the hereditary recurrent fevers, and suggested a possible role for common hereditary recurrent fever variants in other inflammatory conditions. Recent advances in molecular pathophysiology include the elucidation of the N-terminal PYRIN domain in protein-protein interactions, the description of the NALP3 (cryopyrin) inflammasome as a macromolecular complex for interleukin-1β activation, and the identification of signaling defects other than defective receptor shedding in patients with tumor necrosis factor receptor-associated periodic syndrome. These molecular insights form the conceptual basis for targeted biologic therapies. SummaryAdvances in molecular genetics extend our ability to recognize and treat patients with systemic autoinflammatory diseases and inform our understanding of the regulation of innate immunity in humans.


Introduction
The concept of autoinflammatory disease was first proposed in 1999 to describe a group of inherited disorders characterized by episodes of seemingly unprovoked inflammation that, in contrast to the traditionally defined autoimmune diseases, lack high-titer autoantibodies or antigen-specific Tcells [1]. Two hereditary recurrent fevers (HRFs), familial Mediterranean fever (FMF, Mendelian inheritance in man [MIM] 249100) and the then newly recognized tumor necrosis factor receptor-associated periodic syndrome (TRAPS, MIM 142680), were the prototypes for this diagnostic category. The following year, this concept was extended to subsume several mendelian disorders, including other HRFs, the familial urticarial syndromes (now included among the HRFs), complement disorders such as hereditary angioedema (MIM 106100), and granulomatous disorders such as Blau's syndrome (MIM 186580) [2]. Several illnesses with a complex mode of inheritance, such as Behc xet's disease (MIM 109650) and idiopathic pulmonary fibrosis (MIM 178500), were also included among the proposed autoinflammatory diseases, and it seems reasonable to suggest that some apparently acquired disorders of inflammation, such as the syndrome of periodic fever with aphthous stomatitis, pharyngitis, and cervical adenopathy (PFAPA) [3], may also properly fall under this rubric. Subsequent advances in molecular genetics have vindicated the notion of autoinflammatory disease as a unifying concept, at both the structural and functional levels [4]. Although the gene mutated in the hyperimmunoglobulinemia D with periodic fever syndrome (HIDS, MIM 260920) [12,13] does not encode such a motif, recent data suggest that it may also impinge on the innate immune system through the regulation of interleukin-1b secretion. There are also structural and functional relationships between the HRF proteins and the proteins mutated in several other autoinflammatory disorders, including Blau's syndrome and the syndrome of pyogenic arthritis with pyoderma gangrenosum and acne (PAPA, MIM 604416).
Increased awareness of the systemic autoinflammatory diseases, coupled with the widespread availability of genetic testing, has catalyzed the evolution of our concepts of diagnosis, genotype-phenotype interaction, and the broader role of the causative genes and proteins in health and disease, while concomitant advances in our understanding of pathophysiology have allowed dramatic breakthroughs in targeted biologic therapy. This review focuses on significant advances of the past year.

Clinical genetics
Although no new HRF genes have been identified over the past year, mutational studies of cohorts of affected patients have substantially advanced our understanding of the biologic role of the relevant genes and proteins.
Areas of progress include refinement of the relationships between gene mutations and specific disease-associated clinical manifestations; analysis of the role of specific mutations and modifier factors in the risk of amyloidosis; and delineation of the relation between common gene variants and the broader spectrum of inflammatory disease. ]. The answer to this latter question may be tied to the resolution of the first two.

Amyloidosis in the hereditary recurrent fevers
Systemic amyloidosis is one of the most serious manifestations of the HRFs and is the result of the tissue deposition of misfolded fragments of serum amyloid A (SAA), one of the acute-phase reactants produced by the liver in response to systemic inflammation [48]. Most frequently, deposition occurs in the kidneys, gastrointestinal tract, adrenals, spleen, testes, and lung and sometimes in the liver, heart, and thyroid. In the precolchicine era, amyloidosis was a frequent cause of death in patients with FMF, particularly north African Jews, Turks, and Armenians. Amyloidosis in FMF can sometimes precede the development of febrile attacks (phenotype II), a phenomenon that is probably due to the persistent subclinical inflammatory state seen even in the absence of symptoms in some HRF patients [16,49-52,53 • ,54 • ].
A substantial body of literature indicates an increased risk for amyloidosis among Jewish, Arab, and Armenian patients who are homozygous for the M694V mutation [55][56][57][58][59]. In a series of more than 1000 Turkish patients for whom mutational analysis was available [60 • ], however, there was no statistically significant association between this genotype and the risk of amyloidosis. Although other smaller series from Turkey have come to the same conclusion [61,62], the explanation for the difference from other populations is not clear but could involve either differences in the frequency of modifier genes or environmental effects.
One apparently important modifier factor in amyloidosis risk in FMF is the SAA1 precursor isoform, with the a/a variant conferring increased risk [63,64]. In a recent series from Turkey, seven of 23 FMF patients with this genotype had amyloidosis vs one of 51 patients with other SAA1 genotypes [65 • ]. Significant differences were also observed in a recent study of 70 Arab patients [66 • ]. The mechanism by which this SAA1 variant increases amyloid risk is unknown, but current speculation focuses on differences in macrophage processing or intrinsic potential for fibril formation [63].
Amyloidosis also occurs relatively frequently in patients with MWS and NOMID/CINCA, as well as TRAPS. In TRAPS, susceptibility to amyloidosis appears to be increased among patients with mutations at cysteine residues [26], although patients with noncysteine mutations, most notably T50M, have been reported [67 • ]. Amyloidosis is extremely rare in HIDS, with the first case having been reported only within the past year [68 • ]. It is not clear whether the rarity of amyloidosis in HIDS, relative to FMF, TRAPS, MWS, and NOMID/CINCA, is due to an overall lower SAA burden in HIDS, to less amyloidogenic alleles at modifier genes, or to environmental factors.

Role of hereditary recurrent fever genes in inflammation
Given the relatively high frequency of certain HRF alleles in the general population, there has been considerable speculation that some of these variants may also predispose to other inflammatory phenotypes [26]. It goes without saying that in situations such as this, in which common genetic variants of HRF genes are sought in other relatively common illnesses, controls that are appropriately matched, particularly for ethnic background, are essential. Although the HRF genes may, in some circumstances, conspire with other genetic and environmental factors to cause a broader spectrum of inflammatory diseases, certain disorders may actually be less common in the HRFs. Recently a group from Turkey drew attention to the complete absence of systemic lupus erythematosus among their cohort of more than 1000 FMF patients [77]. The authors speculated that high levels of C-reactive protein typically seen in FMF patients might increase clearance of apoptotic cells and autoantigens. Although this remains an intriguing hypothesis, it underscores the potentially complicated and even reciprocal interactions among autoinflammatory and autoimmune disorders, which represent respective aberrations of the innate and adaptive arms of the immune system. The suggestion of a possible positive correlation between systemic lupus erythematosus and TRAPS in the Japanese population [78] awaits confirmation.

Pathogenesis
The elucidation of the molecular basis of the HRFs has focused attention on a group of genes encoding proteins ( Fig. 1) that regulate several critical inflammatory and apoptotic pathways. Much of the past 2 to 3 years' work has concentrated on further delineating these pathways and understanding how specific disease-associated genes cause autoinflammation.  a C-terminal CARD, through which it can interact with several downstream molecules. Although no diseaseassociated ASC mutations have been identified in HRF patients to date, it is a pivotal molecule in the pathogenesis of these diseases.
Recent biochemical evidence indicates that cryopyrin (NALP3) and ASC participate in a larger macromolecular complex termed the NALP3 inflammasome [88 •• ,89 •• ] that mediates the activation of interleukin-1b and interleukin-18. The NALP3 inflammasome activates interleukin-1b by bringing molecules of caspase-1 (interleukin-1bconverting enzyme) zymogen into proximity, thus allowing autocatalysis of its p20 and p10 subunits, which, when released, cleave prointerleukin-1b into its biologically active form. As depicted in Figure 2,  Pyrin itself also appears to play an important role in regulating interleukin-1b activation. In-vitro data suggest that pyrin competes with both cryopyrin and caspase-1 for binding to ASC [83,84]. Mice expressing a truncated, hypomorphic pyrin variant exhibit heightened sensitivity to endotoxin challenge, with increased activation of both caspase-1 and interleukin-1b. These data suggest that one function of wild-type pyrin is the suppression of inflammasome-mediated interleukin-1b production and that FMF-associated mutations may interfere with this process (Chae et al., unpublished observations). Mutations in proline serine threonine phosphatase interacting protein 1 (PSTPIP1), a protein recently shown to bind pyrin, appear to exert a dominant negative effect on this pathway [92]. Two PSTPIP1 mutations (Fig. 1) have been associated with increased pyrin binding, excessive interleukin-1b production, and a severe autoinflammatory disorder, the PAPA syndrome.
Both cryopyrin and pyrin also appear to regulate another process important in inflammation: apoptosis. The aforementioned pyrin-deficient mice exhibit a defect in leukocyte apoptosis through an interleukin-1b-independent, caspase-8-dependent pathway [83], suggesting a proapoptotic role for the wild-type protein, although in certain transfection systems it exerts an antiapoptotic effect [79,84,85]. Enforced expression of cryopyrin in HEK293T cells also induces apoptosis [84].
Depending on the cellular context, both pyrin and cryopyrin can either activate or suppress nuclear factor-kB [84,86,87,93,94], a family of transcription factors involved in the initiation and resolution of inflammation. Although the precise mechanism is still under investigation, this appears to be ASC dependent and, under some conditions, involve the inhibitor of nuclear factor-kB kinase complex [93]. Because endogenous pyrin has recently been shown to localize in the nucleus in several cell types, including synovial fibroblasts, neutrophils, and dendritic cells (but not monocytes) [95 •• ], it is also possible that pyrin may associate with one or more components of the nuclear factor-kB complex. Moreover, in the absence of ASC, a relatively rare isoform of pyrin with an inframe deletion of exon 2 also localizes in the nucleus, regardless of FMF-associated mutations [96 • ].

TRAPS: the plot thickens
Stimulation through the p55 TNF receptor can lead either to nuclear factor-kB activation or apoptosis, depending on the balance of several contextual factors. Upon receptor activation through TNF, metalloprotease-induced cleavage of the extracellular TNFRSF1A domain can limit continuous signaling at the cell surface while simultaneously creating a pool of potentially antagonistic soluble receptor (Fig. 3A). Initial studies of a family with the C52F mutation indicated impaired activation-induced receptor ' TRAPS-associated p55 mutations might also cause constitutive activation, perhaps by permitting intermolecular disulfide homodimerization and ligand-independent activation. This possibility was considered for patients with the C52F mutation in the initial description of TRAPS but appeared not to be operative [1]. Moreover, such a mechanism would appear to be inconsistent with the therapeutic effects of TNF inhibitors (vide infra). It may be fruitful, however, to reexamine this issue for a broader sampling of patients, given the heterogeneity of cleavage defects for different mutations, the observation of biochemical inflammation in TRAPS patients even between attacks [16], and the discovery of ligand-independent noncovalent interactions mediated by the first cysteine-rich domain of the p55 receptor [101]. Yet another conceptually attractive possibility relates to the recent finding that the predominant form of TNFRSF1A in human plasma is full length, probably the result of exosome-linked release of receptor [102 • ]. In light of the aforementioned defects in receptor trafficking, it is intriguing to hypothesize that TRAPS mutations might impair such a process.
From the foregoing, it appears clear that there may be multiple mechanisms leading to the TRAPS phenotype and that the pathophysiology may be heterogeneous among patients. Clarification of these issues will undoubtedly require triangulation between studies of primary cells from patients, transfected cell lines, and knock-in animal models.
Hyperimmunoglubulinemia D with periodic fever syndrome: nature's elaborate deception?
Perhaps the most enigmatic of the HRFs is HIDS. The enzyme mutated in HIDS, called mevalonate kinase, is the only HRF protein that does not include a death domain-fold motif. Mevalonate kinase catalyzes the conversion of mevalonic acid to 5-phosphomevalonic acid in the synthesis of sterols, including cholesterol, vitamin D, bile acids, and steroid hormones (Fig. 3B). Evidence is strong that HIDS is not due to excessive IgD, because there are well-documented patients who have the HIDS phenotype and MVK mutations but persistently normal IgD levels [12, [103][104][105], and, even among patients with increased serum IgD, the levels do not predictably fluctuate with attacks [106]. Moreover, the HIDS phenotype appears not to be due to a defect in cholesterol synthesis, because patients have cholesterol levels in the low-normal range, and more severe disorders of cholesterol biosynthesis do not have an autoinflammatory phenotype [107].
Currently there are two major hypotheses on the pathogenesis of HIDS: that the inflammatory attacks could result from the accumulation of mevalonic acid, the substrate for the mevalonate kinase enzyme [108 • ], or that the autoinflammation is caused by a shortage of isoprenoids, which are normally synthesized through the mevalonate pathway [109]. These latter compounds are involved in the post-translational prenylation (farnesylation or geranylation) of several important intracellular signaling molecules, including the Ras, Rho/Rac, and Rab families of small guanosine triphosphate-binding proteins.
In an in-vitro system, accentuated interleukin-1b secretion by leukocytes from HIDS patients can be reversed by the addition of farnesol or geranyl-geraniol, lending support to the second hypothesis [109].
Both the isoprenoid deficiency and mevalonate accumulation hypotheses predict a worsening of symptoms with decreased mevalonate kinase enzymatic activity. In-vitro studies of cell lines harboring wild-type or HIDS-mutant MVK indicate that the mutant enzyme functions best at 30°C, with a diminution at 37°C and further decreases at 39°C [110]. This finding may account for the triggering of HIDS attacks by immunizations and infections and may also account for the increased urinary mevalonate levels seen during HIDS attacks.

Treatment
Advances in our understanding of the biology of HRFs, coupled with the expanded armamentarium of new targeted therapies, have led to new approaches to the treatment of these disorders. Therapeutic goals include suppression of acute attacks, which are usually not life threatening but can be very disabling, and preventing long-term sequelae, such as amyloidosis and long-term neurologic/intellectual impairment in CAPS.
The most promising results of the past year involve the use of anakinra, a recombinant human interleukin-1b receptor antagonist, in patients with CAPS. FCAS patients who were pretreated with interleukin-1b receptor antagonist before cold challenge did not develop clinical symptoms or increase in acute-phase reactants [111 •• ]. Serum levels of interleukin-1b and cytokine mRNA in peripheral blood mononuclear cells were normal but highly elevated in affected parts of the skin, implicating differences in the distribution of cells contributing to disease phenotype. A complete cessation of clinical symptoms and biochemical changes was also reported in MWS patients following administration of interleukin-1b receptor antagonist [112,113 • ]. Even children with the more severe phenotype of NOMID/CINCA responded to anakinra doses of 1-2 mg/kg per day with resolution of uveitis, rash, and fever and a significant decline in cerebrospinal fluid pressure [114 • -117 • ]. The dramatic nature of the response of CAPS patients to interleukin-1 inhibition is, in a way, surprising, given the apparent role of cryopyrin in other inflammatory processes, such as nuclear factor-kB activation and apoptosis. Given the reduced life expectancy of NOMID/CINCA patients, who have a death rate of about 20% before the age of 20, it will be important to follow a larger series of these children on anakinra to monitor long-term outcome with regard to mental and physical development, as well as to determine whether early treatment can prevent joint deformities.
As noted in the previous section, interleukin-1b also appears to play a role in the pathogenesis of FMF, PAPA syndrome, and HIDS and may also be involved indirectly in the pathogenesis of TRAPS. Interleukin-1 inhibition could therefore represent a possible option as first-line or second-line treatment in these diseases. Anakinra has been reported effective in the treatment of one patient each with TRAPS and PAPA syndrome [118 There is also a substantial experience with TNF inhibitors in the HRFs, most notably the use of etanercept, the p75 TNFR:Fc fusion protein, in TRAPS. The administration of 50-75 mg per week in adults, or 0.8-1.2 mg/kg/wk in children, is effective in reducing, although not usually eliminating, clinical and laboratory evidence of inflammation [4,16], thereby allowing a dose reduction in nonsteroidal anti-inflammatory drugs or glucocorticoids. In some patients, etanercept appears to prevent amyloid formation or even reduce proteinuria in patients with amyloid nephropathy [120,121 • ]. Unfortunately, development of amyloidosis can occur even when symptoms are controlled by etanercept [122], and it is likely that monitoring of SAA levels is necessary to titrate the optimal dosage [120,121 • ].
Although HIDS very rarely leads to systemic amyloidosis, and does not share the neurologic sequelae of CAPS, attacks are frequently severe enough to warrant treatment, particularly in childhood and adolescence. To date there is no accepted therapy for HIDS, other than antipyretics and palliative measures, but pilot studies have been conducted in two areas. First, a small trial has been conducted with simvastatin, an inhibitor of 3#-hydroxy-3#methylglutaryl -coenzyme A reductase, the enzyme immediately preceding mevalonate kinase in the mevalonate pathway (Fig. 3B). It appears safe, and preliminary data suggest a possible benefit [108 • ]. A pilot study of etanercept showed substantial symptomatic improvement in two mutation-positive children with HIDS [105], although a third HIDS patient who did not respond to etanercept was recently reported by another group [123].
Interleukin-1 inhibition may represent yet another possible therapeutic strategy.
Daily oral colchicine therapy has been established as effective in preventing both the acute attacks of FMF and the development of amyloidosis. In the subset of patients who are poorly responsive to colchicine, lower colchicine concentrations were found in mononuclear cells [124 • ], suggesting that differences in responsiveness may be due to polymorphisms in transporters that control intracellular drug concentrations, such as the MDR-1encoded P-glycoprotein pump. In such patients, several adjunctive approaches are under investigation, including subcutaneous interferon-a [125,126 • ,127 • ] and biologic therapies aimed at TNF [128 • ] or interleukin-1b. Allogeneic bone marrow transplantation has recently been proposed as a treatment for refractory FMF [129], based on the predominant expression of MEFV in leukocytes [5]. Although it is possible that this approach could be effective, in nearly all cases other options exist, and the risks outweigh the potential benefits [130].

Conclusion
Identification of the genes mutated in the HRFs has led to great strides in our approach to patients with these disorders. Although substantial numbers of patients with clinical recurrent fever syndromes do not have mutations in the respective genes, the availability of genetic testing as an adjunct has led to more widespread and earlier recognition of these conditions, and recognition of important pathogenetic and therapeutic differences among patients who, 10 years ago, were largely lumped together as FMF variants. Exciting advances in molecular biology have defined new families of motifs and proteins relevant to inflammation and apoptosis, but important questions remain regarding the role of the products of the mevalonate pathway. Perhaps most notable are the great strides in therapy brought about by the happy confluence of breakthroughs in molecular pathogenesis and the new availability of targeted biologic agents. Fascinating areas for further investigation include the possible identification of additional genes that might account for patients who are currently mutation negative, the elucidation of modifier genes, the more thorough understanding of molecular pathogenesis and mechanisms of specific mutations, and a careful comparative analysis of various available treatments in multicenter trials.

References and recommended reading
Papers of particular interest, published within the annual period of review, have been highlighted as: