Dan Kastner, MD, PhD
Clinical Director, NIAMS
Chief, Genetics and Genomics Laboratory, NIAMS/NIH/DHHS
Dr. Kastner has no significant financial interest or relationships to disclose.
Release Date: March 31, 2010
Expiration Date: January 1, 2011
For CME credit,TAKE POST-TEST & EVALUATION
Objectives:
Participants will be able:
- to discuss clinical features of the hereditary periodic fever syndromes.
- to describe the pathophysiology of the diseases and what it teaches us about normal human biology.
- to distinguish one disease from the next based on clinical and genetic characteristic.
This presentation concerns the expanding spectrum of systemic autoinflammatory diseases and various findings in the genomics of inflammation. It includes a discussion about non-FDA approved uses of two medications: etanercept for TRAPS and anakinra for NOMID.
Systemic Autoinflammatory Diseases
Systemic autoinflammatory diseases are a group of disorders characterized by seemingly unprovoked inflammation. They are notable for their relative lack of high-titer autoantibodies or antigen-specific T cells.
We initially proposed a nomenclature to distinguish the disorders from the more traditional, or classical, autoimmune diseases such as lupus or rheumatoid arthritis. It has since become clear that the systemic autoinflammatory diseases tend to be disorders of the innate immune system, whereas the more traditional autoimmune diseases tend to have an impact on the adapter of the immune system. That is one of the distinctions. Another is that the systemic autoinflammatory diseases are mostly Mendelian disorders. Familial Mediterranean fever (FMF), for example, is a single-gene disorder.
At this point, we recognize autoinflammatory diseases that are either disorders of complex genetics where there are several genes involved or where there are disorders where there is a combination of genetic and environmental components. We will start out with the hereditary periodic fever syndromes. I will describe some clinical features of each and a little about what we have learned of the pathophysiology of the diseases and what it teaches us about normal human biology. Then we will end up with an actual disease that is not a periodic fever syndrome but is related to these.
Hereditary Periodic Fever Syndromes
- Familial Mediterranean fever (FMF)
- TNF receptor-associated periodic syndrome (TRAPS)
- Hyperimmunoglobulinemia D with periodic fever syndrome (HIDS)
- Familial cold autoinflammatory syndrome (FCAS)
- Muckle-Wells syndrome (MWS)
- Neonatal-onset multisystem inflammatory disease (NOMID)/chronic infantile neurologic cutaneous and articular syndrome (CINCA)
Familial Mediterranean fever (FMF)
Familial Mediterranean fever is characterized by one to three day episodes of fever. The one to three day episode is important. Periodic fever syndromes each have a distinctive duration of attack. Clinically, it helps to ask the patient about the length of the typical attack.
FMF involves inflammation of the serosal, synovial, or cutaneous tissues primarily. An episode may mimic an acute abdomen. FMF patients may have multiple scars on their abdominal wall from appendectomies, cholecystectomies, and so forth. Patients may experience pleural attacks. Much less commonly, there can be pericardial involvement or what looks like chronic arthritis. Usually the arthritis in FMF is acute and lasts for a few days; however, at least in the pre-colchicine era, chronic arthritis could be seen.
There is also a characteristic skin rash in FMF called erysipeloid erythema. It is a reddish, raised, fairly well demarcated rash that usually occurs on the dorsum of the foot, the ankle, or the lower leg and lasts from one to three days. This rash is characterized histologically with a mass influx of polymorphonuclear leukocytes to the affected anatomic compartment. More ominously, the patients can also develop systemic amyloidosis of the AA type due to the deposition of a cleavage product, a serum amyloid A (SAA), which is one of the acute phase reactants made during the acute attacks of FMF. The products of SAA can be deposited in a number of organs, including the kidneys, adrenal glands, liver, lungs, spleen, and thyroid gland. Under the electron microscope, we see disorganized fibrillar material.
My interest in FMF came about in the 1980s after seeing a patient with the disorder during the first few months of my rheumatology fellowship at the NIH. He was in his early 20s, of Armenian ancestry, and presented with a history of recurrent monthly episodes of monoarticular arthritis. He had seen numerous pediatricians, pediatric rheumatologists, and adult rheumatologists. No one had come up with the diagnosis, including me. Fortunately, there was a fellow from Israel in the lab who, when he heard this patient’s medical history, knew right away that this patient had FMF. He sees it “all the time” in Israel, where it is treated with colchicine. The patient was thus treated with colchicine, improved, and was extremely grateful.
After a year or two of thinking about this and realizing that FMF is caused by a single autosomal recessive gene, I got the idea of collecting DNA from families with FMF to do linkage studies in order to find the chromosome where the gene for FMF is located and ultimately the gene itself. With the help from my friend from Israel, I established a collaboration with Dr. Mordechai Prass and went to Israel during the summer of 1989 to collect blood from a large number of families with FMF.
Figure 1. Moroccan Jewish family in Israel. Several members are affected with FMF.
Figure 1 is a HIPAA-approved photograph of one of the families. The parents are first cousins. They are a Moroccan Jewish family. Several members are affected with FMF, some of them with amyloidosis. A number of families, including this one, allowed us to do genetic mapping studies that lead to the 1992 discovery of the gene on the short arm of chromosome-16. The gene is denoted MEFV, for Mediterranean fever. It was a novel discovery back then. The structure of it, at least the C-terminal end of the gene, encoded a set of protein domains (Figure 2). The B30.2 domain is a protein-protein interaction domain containing most of the disease associated mutations. The N-terminal end of it, which is now called the Pyrin domain after what we called the protein, is found in some 20 different proteins involved in the regulation of inflammation and apoptosis.
Figure 2. Positional cloning of MEFV, the FMF gene.
In fact, in several of the modern day populations in which FMF is common (e.g., Jewish, Arab, Armenian, and Turkish populations), the same mutations occur, suggesting a common ancestral origin. We think that these mutations probably arose in Biblical or pre-Biblical times and were disseminated into the various populations where they are found today. For example, the M694V mutation substitution of valine for methionine at position 694 is seen in the Sephardic Jewish population of North Africa, in Iraqi Jews, and among Armenians and certain Turks. The only way that it could happen with the same associated DNA fingerprint would be if there was a common ancestor. It is also interesting to know that the carrier frequency for mutations in the FMF gene is extraordinarily high in these populations. We are talking about carrier frequencies on the order of one in three to one in five. Just to put that in some perspective, the carrier frequency for cystic fibrosis—the most common lethal recessive disease in North America—is one in twenty. We think that probably there is something that has selected for these mutations in these populations, although we do not know as yet what that would be. For those people who are interested in knowing about genetic testing in FMF, at this point there are over 70 known mutations in the FMF gene as well as a number of polymorphisms. There is a hot spot for mutations in exon-10. If you order genetic testing for FMF, exon-10 is what usually gets screened thoroughly. To a lesser degree there is another area of increased numbers of mutations in exon-2. Two mutations are required for the genetic diagnosis, but only about 70% of patients with clinical FMF have two identifiable mutations. We do not know if that is because there are mutations in areas of the gene that are not screened or whether under some circumstances, maybe with modifier genes, one mutation may be sufficient. The bottom line is that clinical judgment is still required to make the diagnosis in those individuals who have only one demonstrable mutation.
With regard to the pyrin domain, it is 90 or so amino acids in length. It has a characteristic structure of six alpha helices, which is also found in death effector domains, in caspase recruitment domains. Four domains fall into a grouping called the death domains. They form a three-dimensional structure with six alpha helices, which allow an arrangement of positive and negative charges on the surface of the domain. You can see cognate interactions between, for example, pyrin domains and pyrin domains. For those who are familiar with TNF signaling, the same kind of thing happens with death domains or death effector domains. In any case, this is a domain that allows pyrin to interact with another protein. That other protein is called ASC, which stands for “apoptosis-associated spec-like protein with a CARD domain”, which is why everyone calls it “ASC”. ASC has two domains: an N-terminal pyrin domain and a C-terminal CARD domain. The pyrin domains interact with one another, and then the CARD domain of ASC can interact with the CARD domain of another protein called caspase-1, which many of you may be familiar with as IL-1β converting enzyme. That is the enzyme that converts the cleavage of pro-IL1β, which is a 31kDa protein, to biologically active IL-β, which is 17kDa in size. What does pyrin do to this? We have developed mice that have a truncated form of pyrin. When we look at the activation of IL-1β, we see more activated IL-β than in the wild type. The mice have an exaggerated body temperature response to the injection of endotoxin LPS. They become hypothermic. We think that pyrin is having a negative effect.
A macromolecular complex called inflammasome allows the activation of IL-β by bringing together a couple of molecules of pro-caspase-1. It is thought that, at least in some circumstances, pyrin may compete for binding to ASC and keep it out of this inflammasone complex. We have shown, in human pyrin, that the B30.2 domain, a pyrin and C terminal end of it, can interact with P20 and P10 subunits of caspase-1.
Our FMF patient from Baghdad who has amyloidosis does not tolerate colchicine therapy because of diarrhea from malabsorption related to his amyloidosis. He has fairly significant amyloid deposits in the gastrointestinal tract. We have used the IL1 receptor antagonist anakinra in his case instead, and have been able to control his acute phase reactants, except when he had an intercurrent infection and we had to stop the drug.
We have also made mice that totally lack pyrin, further increasing IL-1β processing. We have also looked at these mice by microarray to see whether there are other pathways besides the IL-1 pathway that might be perturbed. A principal components analysis shows that there are actually a number of different pathways that are perturbed in the knock-out mice. Among the most interesting are the interferon-γ inducible genes, several of which are markedly down-regulated. Several are very much involved in host response to intracellular pathogens. We are doing some studies with a couple of other NIAID laboratories, looking at the susceptibility (or not) of these mice to various pathogens like Listeria and Francsciella. We are also interested in the question of why is it that this B30.2 domain is where most of the mutations are found. As I mentioned, there is evidence for heterozygous selection in a number of different populations for mutations at that end of the gene. There is even evidence for selection of different B30.2 sequences over mammalian evolution. We have now introduced human B30.2 domains into mice. We have done it with the FMF-associated mutants and control wild-type, which were harder to do than the FMF-associated mutants. The mice are much smaller (the ones that have the human B30.2 domain). They have dermatitis and increased sizes of their spleens. They get ocular inflammation and inflammation of the serosal surface of the liver. They also get arthritis. Radiographically, their paws display erosive disease. Histologically, there is an influx of inflammatory cells into the joint space, as there is in FMF disease. There is speculation that perhaps bacterial products interact with the B30.2 domain and in some way allow it to open up such that it can form some sort of an inflammasome macromolecular structure that would allow IL-1 activation or perhaps other inflammatory molecule activation.
To return to a clinical focus, as I said before, patients with FMF will have one to three day attacks of fever with serositis, arthritis, or rash. Erysipeloid erythema is relatively characteristic. Renal amyloidosis is decreasing in frequency because of colchicine treatment, but it is still an important consideration. The high-risk groups are Jewish, Armenian, Arab, Turkish, and Italian. Onset is usually during childhood; however, it is often not diagnosed until adulthood. FMF is caused by a recessive mutation in the MEFV gene, which encodes the pyrin protein. Daily colchicine treatment is usually effective treatment.
A number of questions remain: What is the biologic basis of the heterozygote advantage? What is the possibility that there may be other genes involved? Could pyrin be a protein that binds to bacterial products? What is the role of pyrin in other inflammatory disorders? What are the modifier genes and how does amyloidosis work?
TNF receptor-associated periodic syndrome (TRAPS)
While we were trying to figure out the gene for FMF, we received a call from an Irish anesthesiologist at the NIH who reported knowing of some Irish woman in the Washington area who had periodic fever syndrome. We agreed to see a young woman, who we will call Jane, who reported a 14-year history of febrile episodes. Her mother, four brothers, and a cousin all have the same condition. Unlike FMF, this looked like dominant inheritance.
Jane’s attacks last three to five weeks. Her symptoms include periorbital edema, a migratory rash, and abdominal pain. We saw her one week after she delivered her baby. During the pregnancy, for the first time, she had no attacks for nine months. When we saw her, she had a high WBC count and an acute phase response. She responded to steroids, but not to colchicine—another difference between this and FMF. Her condition had been described in the literature as familial Hibernian fever, Hiberian meaning Irish. It was first described in the early 1980s in a large family of mixed Irish and Scottish ancestry.
Figure 3. The TRAPS gene on chromosome 12. From Rheumatology 3rd edition, p. 1725, 2003.
Mike McDermott, at that time a fellow in my lab, was able to map the gene for Hibernian fever to the short arm of chromosome-12 (Figure 3), showing that Hibernian fever was not just a different set of mutations from Mediterranean fever, because the genes are in completely different chromosomes. Mike returned to my lab later during a sabbatical to see whether we could find the gene for Hibernian fever by positional cloning. We had a fairly big region of about 10 million base pairs with lots and lots of genes in it. The first thing that we did—because we did not have enough families to narrow things down—was to apply the so called “embarrassment test.” The embarrassment test goes like this: What gene would have been most embarrassing if it turned out to be the one after a five-year long positional cloning effort? We figured that the worst one would be TNFRSF1A, which is the gene that encodes the P55 TNF receptor. There are actually two genes that encode TNF receptors in the human genome: one on chromosome-12 and TNFRSF1B on chromosome-1. TNFRSF1B encodes the P75 TNF receptor.
We sequenced this gene in seven different families with dominantly inherited recurrent fever. They were not all Irish in ancestry. Lo and behold, we found mutations in every family. Within a month or about a month and a half, employing the embarrassment test, we found the gene for this Irish periodic fever. Because it was also present in families of other ancestries, we renamed it, calling it TNF Receptor Associated Periodic Syndrome (TRAPS).
The P55 TNF receptor has four cysteine-rich extracellular domains. The domains have a “loop-the-loop” kind of structure. There are six cysteines in each domain. They are held together by three disulfide bonds. Actually, in our first panel of seven families, six out of the seven families had mutations at cysteines that disrupted the disulfide bonds.
Once we had several individuals with this disorder together, we began to compile a list of clinical features. Patients have long attacks lasting up to six weeks; some seem to have continuous inflammation. There is a distinct migratory rash that moved downward toward the feet rather than spreading outward. In a laparoscopic exam of a young girl who had repeated abdominal attacks, we saw adhesions in the abdominal cavity. There is also muscle involvement. MR imaging shows inflammation going down into the muscle compartment, although the inflammation is actually a fasciitis rather than a myositis. Conjunctivitis and periorbital edema is another feature. There is also AA amyloidosis with kidney biopsy showing glomerulus vessel walls stained with a monoclonal antibody against SAA.
There are currently over 50 known mutations in the P55 TNF receptor associated with dominantly inherited periodic fever, with TRAPS. About half involve cysteines in the disulfide bonds. For some of the cysteines, there are multiple mutations. You can have a cysteine substituted to arginine, serine, or tyrosine, so that it seems like that phenotype is associated a lot with cysteine mutations, although there are other mutations that do not involve cysteines, but have structural effects. There are also a couple of variants, R92Q and P46L, which are relatively common in various populations. R92Q is actually found in about one percent of the Caucasian population and P46L may be found in an even higher percentage among African-Americans.
The two variants are associated with a little more varied inflammatory phenotype than what we see with the cysteine mutations. In TRAPS, patients have inflammation nearly all of the time. Even when these patients are asymptomatic, they have elevated CRP levels. Initially, when we first discovered TRAPS, we thought (and still do, but to a lesser extent) that part of the mechanism of TRAPS was a failure to shed p55 TNF receptors from the cell surface. The idea is that normally what happens is that if you stimulate through the p55 receptor, inflammation is triggered in the white cell. Also triggered are metalloproteases which cleave the p55 receptors from the cell surface and then inhibit or prevent repeated stimulation through those receptors. In fact, the receptors are floating around out in the body fluids and can actually compete for binding the TNF with membrane-bound receptors. It made sense that if you do not have normal cleavage of these receptors of the cell surface of the patients that it might lead to a hyper inflammatory state.
We still believe that this is the case, but it actually turns out to be more complicated. There are some p55 TNF mutant receptors that do not even bind TNF. One can look at how sociable TNF receptors are using something called FRET—Fluorescence Resonance Energy Transfer. If one sets it up in such a way that you have GFP on one transfected molecule and YFP on the other, you can see whether these molecules get together. Normally, p55 TNF receptors form trimers on the cell surface. If you have wild type GFP and wild type YFP, you will get this FRET phenomenon. However, if you have H22Y, which is one of the mutations, and wild type YFP, no FRET occurs. On the other hand, H22Y can associate with itself. Mutant receptors associate with mutant receptors, and wild type receptors associate with wild type receptors; however, they do not intermix. There is a problem with aggregate formation as well. Richard Seigel’s group at the NIH has looked at the trafficking of the wild type mutant TNF receptors in transfection systems. Wild type receptors localized in the Golgi whereas mutant receptors actually get stuck in the endoplasmic reticulum. A lot of the receptors do not even make it to the cell surface.
We call our more sophisticated understanding of TRAPS the “Dud Receptor Hypothesis.”
That is, that these mutations, especially the ones with the cysteines, which distort the folding of the protein, actually have a large number of effects on mutant TNF receptors:
- reduced ligand binding
- impaired shedding (probably because the misshapen molecule is not recognizable to metalloproteases)
- trafficking defects
- defects in apoptotic signaling
- and conceivably, aggregates of receptor intracellularly may actually trigger NFκB signaling in a ligand-independent way
Nevertheless, if one uses the TNF inhibitor etanercept in patients with TRAPS (twice or three times a week), etanercept works, but the question is whether it is working in a totally specific manner, or whether it is just working because it is inhibiting a network of related cytokines. I will tell you that in the few patients that we have had to treat with something else, because they could not tolerate TNF inhibitors, they respond to anakinra.
There are numerous answered questions about TRAPS. Some patients with TRAPS-like illness do not have TNFRSF1A mutations. What is the role of common variants (R92Q and P46L) here and in other conditions? There is the whole question of pathophysiology and also treatment with IL-1 and IL-6 inhibition and the prevention of amyloidosis.
Hyperimmunoglobulinemia-D with periodic fever syndrome (HIDS)
HIDS is common in Northern Europeans, particularly among people from the Netherlands and neighboring areas of France. HIDS has been best studied by a couple of groups in the Netherlands, one of them in Nijmegen and the other at Utrecht. The syndrome is characterized by a three to seven day febrile episode. (Remember, FMF has one to three day episodes; TRAPS has a month-long episode.) Patients can have abdominal pain, arthritis or arthalgia, a diffuse rash, prominent cervical adenopathy during an attack, and aphthous ulcers. The age of onset is usually at a very early age, sometimes provoked by childhood immunizations. Amyloidosis is rare.
Two research groups from the Netherlands identified disease-causing mutations, at about the same time, in a surprising gene. The gene, on the long arm of chromosome-12, encodes mevalonate kinase, which is an enzyme involved in the synthesis of cholesterol and certain non-sterile isoprenoids (which may be involved in post-translational modification of inflammatory proteins). Measuring urine mevalonate during attacks and then between attacks shows that, although it is always elevated, it shoots up further during attacks.
In contrast, IgD levels do not correlate with the severity of the disease or with attacks. In fact, there are some patients with mutations in this gene who do not have elevated IgD levels. That makes it a little bit confusing in terms of making the diagnosis. There are actually people with hyper IgD syndrome, at least genetically, who have normal IgD levels. There has some confusion of diagnosis in HIDS. There is classic HIDS characterized by an elevated IgD, increased mevalonate levels, and mutations in the gene. Then there is a variant HIDS in which there is an increased IgD but normal mevalonate levels and no mutations in the gene. Rarely, as I mentioned, there can be mutations in the gene and increased mevalonate, but normal IgD levels. If you suspect HIDS, you should probably order serum IgD level, and either get genetic testing or biochemical testing for urinary mevalonate.
There is no specific treatment for HIDS. Colchicine, which is the treatment for FMF, and corticosteroids, which is the treatment for TRAPS, generally do not work. We have had some good luck with investigational drugs, certainly the cytokine inhibitors. Colleagues in the Netherlands have had some luck with the statins. Leukotriene inhibitors may have a role as well. There are a number of unanswered questions with HIDS similar to those with the previous diseases.
Familial cold autoinflammatory syndrome (FCAS), Muckle-Wells Syndrome, and the neonatal-onset multisystem inflammatory disease (NOMID) are three dominantly-inherited periodic fever syndromes caused by mutations in cryopyrin, a protein in the pyrin family with a pyrin domain at its N-terminus. This feature links us—at least in a homology way—with FMF.
With regard to phenotype, FCAS—also known as Familial cold urticaria—is a disease in which the patient has fever and a hives-like skin rash upon generalized exposure to cold temperatures. The rash occurs about two hours after cold exposure. An ice cube on the skin does not usually provoke an attack. Exposure has to be more generalized. The rash, although it is called cold urticaria, is not real urticaria. There are neutrophils in the skin, not mast cells. In fact, histamine levels in the blood (which would be elevated in acquired urticarias) are normal. FCAS is dominantly inherited.
Muckle-Wells syndrome is similar to FCAS in that there is fever and rash, although it is an urticarial rash. It can happen at any time. Sometimes stress induces it, but not necessarily cold. Patients also have limb pain and sensorineural hearing loss. Some patients get systemic amyloidosis.
The gene for Muckle-Wells syndrome and FCAS was mapped by Hal Hoffman at University of California, San Diego. Looking at the long arm of chromosome-1 he found a candidate gene that had a pyrin domain in it, and actually had a leucine-rich NACHT domain. The pyrin domain is also seen in the FMF protein, a part of which is in common with NOD2 that is mutated in Crohn’s disease. The candidate gene is very interesting. It turns out that mutations in this NACHT domain are associated with both Muckle-Wells syndrome and FCAS.
As these discoveries were coming to light, Raphaella Goldbach-Mansky, one of my colleagues at the NIH, referred a nine-year-old boy for possible systemic onset juvenile idiopathic arthritis. Certain aspects of the case were not typical. One was a rash that you could say might be an urticarial but certainly not a Still’s type rash. It was not cold-induced. He had papilledema and some enlargement of the cerebral ventricles. He also had peculiar knobby-looking knees. Based on this constellation, Raphaella correctly diagnosed neonatal onset multisystem inflammatory disease (NOMID). Figure 4 shows patients with NOMID, showing the rash (looking more like the urticarial rash of Muckle-Wells or cold urticaria), bony deformities of the knees related to overgrowth of the epiphysis of the long bones and of the patella.
Figure 4. Patients with NOMID have a rash, bony deformities of the knees related to overgrowth of the epiphysis of the long bones and of the patella.
In terms of the central nervous system manifestations in NOMID, there is inflammation of the meninges, appearing white in Figure 5. These patients get chronic, aseptic meningitis. They are sick nearly all the time. They can have increased intracranial pressure, papilledema, hearing loss, and eventually ventriculomegaly and cerebral atrophy. There is oftentimes blindness, deafness, mental retardation, or all in combination.
Figure 5. CNS involvement includes inflammation of the meninges.
We suspected that NOMID and Muckle-Wells syndrome were related. In one patient we sequenced the CIS1 gene which encodes cryopyrin and saw that, in fact, there was a substitution mutation. When we looked in other patients, sure enough, they had the same mutation. What’s more, we found that these patients have constitutive activation of IL-1, which we compared with a normal control. Untreated cells do not make any IL-1 beta whereas treated cells with LPS do. The patient makes IL-1 regardless of stimulation of LPS.
We figured out that NOMID is actually caused by mutations in the same gene. The three different phenotypes are cold urticaria, Muckle-Wells, and NOMID. The bottom line is that you cannot predict whether somebody is going to have NOMID or cold urticaria or Muckle-Wells based on where the mutation is located in the gene. In fact, all of these mutations—based on computer models—seem to occur on the same surface. The pyrin domain is free of mutations, and there are a couple of mutations idenitified in other regions. It is thought that these mutations somehow change the configuration of the structure such that the normally folded protein opens up and allows for the formation of the inflammasone.
Raphaella got the idea of treating these patients with anakinra, the IL-1 receptor antagonist. In the patients we conducted a pre-treatment evaluation period of up to a month, followed by a dose escalation period, then a drug withdrawal period, and then a reinstitution of the drug on open label. On pre-treatment, the patients had the urticarial rash and felt poorly. Within two or three days of starting anakinra, the rash and fever went away and the kids felt a lot better. In the subset of patients in whom we did a drug withdrawal, they went right back to square one or even worse.
The improvement in CNS status was quite remarkable. After three months of treatment, in one patient the opening pressure was 20 cm, about the upper limit of normal, down from around 30 cm before treatment. Shown in Figure 6 are MRI flare images of a patient before and after treatment, showing disappearance of the chronic meningitis. The red arrow in the neighboring image indicates the cochlea, which is inflamed before treatment and resolved with treatment. In essence, a number of IL-1 responsive genes are up regulated before treatment, down regulated after treatment, and back up again in the drug withdrawal phase.
Figure 6. On the left are flare images of a NOMID patient before treatment and after treatment, showing disappearance of the chronic meningitis. The red arrow in the neighboring image indicates the cochlea, which is inflamed before treatment and resolved with treatment.
There are a lot of unanswered questions about these cryopyrinopathies. What is the structural basis of disease severity? Why is it, even though mutations are right next to each other, that one causes FCAS and another causes NOMID? What additional genes cause these phenotypes? Not everybody with NOMID has mutation in this gene. Are there other mutations elsewhere in the gene or other genes? Why does cold exposure cause attacks of FCAS? Are all of the manifestations of NOMID/CINCA (CINCA is the European term) IL-1beta-dependent? Will early treatment with anakinra or other IL-1 inhibitors prevent all of the neurological and skeletal abnormalities of this disorder?
The list below provides a general guide for distinguishing among the hereditary periodic fevers.” TRAPS has the longest attacks; attacks in HIDS last three to seven days; attacks in FMF last one to three days; in Muckle-Wells and FCAS, about 24 hours or so.
- Duration of attacks: TRAPS > HIDS > FMF > MWS/FCAS
- Lymphadenopathy: HIDS > TRAPS > FMF > MWS/FCAS
- Conjunctivitis: MWS/FCAS, TRAPS > FMF, HIDS
- Amyloidosis: Very rare in HIDS
HIDS is associated with the most lymphadenopathy; TRAPS has a little bit less. It is usually not seen in Muckle-Wells or FCAS.) With regard to conjunctivitis, Muckle-Wells and FCAS have the most, then TRAPS. Amyloidosis is rare in HIDS.
Figure 8 demonstrates the different rashes: the erysipeloid erythema of FMF; the migratory rash of TRAPS, which can sometimes be on the trunk; the diffuse maculopapular rash of HIDS; and the urticarial rash of Muckle-Wells.
The growing family of systemic autoinflammatory disease
An additional disorder in the growing family of systemic autoinflammatory diseases is PAPA syndrome. The name refers to pyogenic arthritis with pyodermic gangrenousum and acne. It is a dominantly-inherited condition caused by mutations in PSTPIP-1: proline, serine, threonine phosphatase interacting protein. PAPA syndrome has an interesting phenotype. Figure 9 shows a pyoderma gangrenosum lesion of the forearm of one of our patients and also cystic acne. The skin completely sloughs. It is sterile unless it gets super infected. Acne is also on the face.
Figure 9. Pyroderma grangrenosum and acne in PAPA symdrom
PSTPIP-1 turns out to be a pyrin-binding protein, leading us back to the concept that pyrin, through its interactions, may have important biologic effects.
At NIAMS we have close to 2000 patients in our periodic fever cohort. A large number of genes are involved. We have identified genetic mutations for about 30 percent. We are using a resequencing chip to find addition genes.
