Late-onset tumor necrosis factor receptor–associated periodic syndrome in multiple sclerosis patients carrying the TNFRSF1A R92Q mutation




Tumor necrosis factor receptor–associated periodic syndrome (TRAPS) is an autosomal-dominantly inherited autoinflammatory disorder caused by mutations in the TNFRSF1A gene. It is characterized by episodes of autoinflammation usually associated with fever, abdominal pain, myalgia, exanthema, arthralgia/arthritis, and ocular involvement. We undertook this study to investigate the prevalence of TRAPS in patients with multiple sclerosis (MS) who reported, in addition to their neurologic symptoms, at least 2 other symptoms compatible with TRAPS.


Twenty-five unrelated MS patients were prospectively screened for TNFRSF1A mutations. In addition, blood samples from 365 unrelated MS patients and 407 unrelated Caucasian controls were analyzed to determine the R92Q carrier frequency.


Six of 25 adult MS patients (24%) with symptoms suggestive of TRAPS were found to carry the identical arginine-to-glutamine substitution at amino acid position 92 (R92Q or p.Arg121Gln) encoded by exon 4 of the TNFRSF1A gene. All R92Q heterozygotes had similar symptoms, including arthralgias/arthritis, myalgias, urticarial rash, and severe fatigue, which began before the onset of MS. In 5 of the 6 patients, we could identify family members who had TRAPS symptoms and had inherited the identical mutation. The R92Q exchange was also detected in 17 of 365 unselected MS patients (4.66%) and in 12 of 407 controls (2.95%) (P = 0.112). Three patients were heterozygous carriers of MEFV variants, in 1 patient in combination with the R92Q mutation.


Autoinflammatory syndromes and especially late-onset TRAPS should be considered in MS patients who report symptoms such as arthralgias/arthritis, myalgias, urticarial rash, and severe fatigue.

Tumor necrosis factor receptor–associated periodic syndrome (TRAPS; MIM no. 142680) is an autosomal-dominantly inherited autoinflammatory disorder resulting from mutations in the TNFRSF1A gene (MIM no. 191190) located on chromosome 12p13 (1). The median age at onset is 10 years. Although the phenotype and the clinical manifestations may vary considerably, TRAPS is usually characterized by febrile episodes lasting from several days up to some weeks (2, 3). Associated symptoms include abdominal pain, myalgia, skin rashes, and arthralgia/arthritis, as well as conjunctivitis and/or periorbital edema (4, 5). The range of clinical presentations has broadened in the last few years with reports of recurrent pericarditis, sacroiliitis, pharyngitis, sterile peritonitis, and tonsillitis, as well as neuropsychiatric manifestations such as meningitis, encephalitis, optic neuritis/papillitis, depression, or recurrent psychoses (5–10).

More than 50 different TNFRSF1A mutations have been identified, which are located in exons 2, 3, 4, and 6 (11). The R92Q substitution is by far the most frequent genetic defect observed in TRAPS patients, and it is considered to be a low-penetrance mutation (12). Symptomatic patients with the R92Q substitution present with a more heterogeneous spectrum of complaints and with a milder disease course compared with other mutations (12, 13).

Several mechanisms are discussed to explain the “hyperinflammatory state” of TRAPS patients. For example, it has been shown that some TNFRSF1A mutations lead to defective shedding of the mutant tumor necrosis factor receptor (TNFR) type I from the cell membrane into the extracellular compartment, thereby reducing the ability to neutralize TNFα by binding through the soluble protein (1). However, TNFR superfamily 1A (TNFRSF1A) shedding defects vary between cell types, and not all TRAPS-related mutations result in defective receptor shedding (14), which strongly suggests that other pathophysiologic mechanisms may also be involved in generating the autoinflammatory response.

By investigating the effect of TRAPS-associated amino acid substitutions on the synthesis and trafficking of TNFRSF1A mutants, Lobito et al (15) recently demonstrated that all the TNFRSF1A mutations studied resulted in misfolding of the extracellular receptor domain and retention of the mutated TNFRSF1A in the endoplasmic reticulum. Only the R92Q variant behaved like the wild-type receptor, suggesting a different pathomechanism. Rebelo et al (16) showed that TRAPS-associated mutations alter the 3-dimensional structure of the TNFRSF1A ectodomain differently, resulting in no or reduced cell surface expression and TNFα binding. Again, mutants with fairly conservative amino acid substitutions at position 92 (R92Q, R92E, and R92K) exhibited properties similar to the wild-type receptor, although subtle abnormalities of surface expression were noted for the R92Q mutant in an endothelial cell line.

Because central nervous system (CNS) involvement has been observed recently in TRAPS patients (10, 17), we performed a prospective study to determine the frequency of TNFRSF1A mutations in patients with multiple sclerosis (MS) who reported, in addition to their MS, episodes with at least 2 symptoms compatible with TRAPS. Affected individuals were also screened for common genetic defects resulting in 2 other hereditary autoinflammatory syndromes, familial Mediterranean fever (FMF; MIM no. 249100) and hyperimmunoglobulinemia D with periodic fever syndrome (MIM no. 260920) (3).


Patients and controls.

Two groups of patients were evaluated. Patients seen consecutively in our Institute of Clinical Neuroimmunology were included in the study as group 1 patients if they met the following criteria: 1) MS or clinically isolated syndrome suggestive of MS according to the criteria of McDonald et al (18) and 2) a history of recurrent episodes with at least 2 of the following symptoms during the attacks: fever, arthralgia/arthritis, myalgia, skin lesions, gastrointestinal involvement (diarrhea, abdominal pain, and the like), conjunctivitis, periorbital edema, or lymphadenopathy. Owing to the prospective nature of the study, clinicians were blinded to the allelic status of the patients at study entry.

Of the 200 patients with definite MS seen in our outpatient clinic between August 2005 and June 2006, 25 unrelated MS patients (4 males and 21 females; age range 22–60 years, 92% with relapsing–remitting disease and 8% with secondary chronic progressive disease) met the inclusion criteria. They were screened for mutations in exons 2, 3, 4, and 6 of the TNFRSF1A gene, in exons 2 and 10 of the MEFV gene, and in exons 9 and 11 of the MVK gene by DNA sequence analysis.

Treatment for MS (such as interferon beta [IFN beta], glatiramer acetate, or immunosuppressive therapy) was neither an inclusion nor an exclusion criterion. The 25 group 1 patients were treated for MS as follows: 10 patients received no specific therapy for MS, 7 patients were treated with IFN beta, 5 were treated with glatiramer acetate, and 3 received immunosuppressive therapy. Family members of mutation-positive patients were invited for an interview and were tested for the respective mutation, whenever possible.

Group 2 consisted of 365 unrelated MS patients (138 males and 227 females, age range 15–79 years; 37% with relapsing–remitting disease, 50% with secondary chronic progressive disease, and 13% with primary progressive disease) classified according to the criteria of McDonald et al and seen in our Institute of Clinical Neuroimmunology or in a nearby local neurologic hospital specialized in treating MS. The DNA of these patients was analyzed by restriction fragment length polymorphism (RFLP) analysis for the presence of the R92Q mutation encoded by exon 4 of the TNFRSF1A gene. Group 3 consisted of 407 unrelated Caucasian controls (203 males and 204 females; age range 1–84 years) who were also screened for the R92Q replacement by restriction digest.

All patients, family members, and control subjects gave their informed consent prior to the genetic testing. The study was approved by the local ethics committee.

DNA sequence analysis of the TNFRSF1A, MEFV, and MVK genes.

Genomic DNA was isolated from 200 μl of EDTA-preserved blood using the DNA Blood Mini Kit (Qiagen, Hilden, Germany). Exons 2, 3, 4, 6, and 7 of the TNFRSF1A gene, exons 2 and 10 of the MEFV gene, and exons 9 and 11 of the MVK gene were subsequently amplified by polymerase chain reaction (PCR) (6). Amplification products were purified with the QIAquick PCR purification kit (Qiagen) and sequenced with the ABI Prism BigDye Terminator v3.1 Ready Reaction Cycle Sequencing kit (Applied Biosystems, Foster City, CA). Electrophoresis was performed on an ABI Prism 3130 Genetic Analyzer (Applied Biosystems).

RFLP analysis for the detection of the R92Q mutation.

In order to detect the c.362G>A nucleotide substitution resulting in the R92Q exchange, exon 4 amplification products were digested with Bst NI (5′-CC↓[A or T]GG-3′; New England Biolabs, Beverly, MA) and analyzed on 2% agarose gels.

Haplotype analysis.

To determine the intragenic haplotype of the R92Q mutation carriers and their relatives, we sequenced PCR products containing single-nucleotide polymorphisms (SNPs) located in exon 1 (c.36G>A) (19), intron 4 (c.473-33C>T) (20), and intron 7 (c.740-9T>C) (12), as well as a polymorphic microsatellite marker in intron 1 (c.39+1899[GT]n[GA]n, where n is the number of variable repeats) (21).

Statistical analysis.

The carrier frequency of the R92Q mutation in MS patients (group 2) and in controls (group 3) was compared by Fisher's exact test.


Twenty-five of 200 MS patients (12.5%) reported symptoms compatible with TRAPS, and 6 of these 25 patients (5 females and 1 male) carried a c.362G>A mutation in exon 4 of the TNFRSF1A gene (carrier frequency 24%), which results in the exchange of glutamine for arginine at position 92 (R92Q) of the mature protein. Two patients were heterozygous for an amino acid substitution encoded by exon 2 of the MEFV gene (patient 5, E148Q/p.Glu148Gln; patient 9, A289V/p.Ala289Val). Patient 23 had inherited the R92Q mutation together with a V726A/p.Val726Ala replacement encoded by exon 10 of the MEFV gene from his German mother. None of the patients had a mutation in exon 9 or 11 of the MVK gene.

In all mutation-positive patients, other rheumatic diseases, such as systemic lupus erythematosus (SLE) or rheumatoid arthritis (RA), had been excluded prior to the genetic analysis. None of the patients were positive for anti-DNA antibodies, anticardiolipin antibodies, or antineutrophil cytoplasmic antibodies, and none had signs of pathologic kidney function or renal disease.

White blood cell count as well as C-reactive protein and serum amyloid A concentrations in the blood, as markers of an acute-phase response, were either not increased or only mildly elevated. They did not differ from the levels in the 17 mutation-negative patients (data not shown). The main clinical characteristics of group 1 patients are summarized in Table 1. The detailed history of the 6 R92Q carriers from group 1 is described below.

Table 1. Clinical characteristics of MS patients with additional symptoms suggestive of TRAPS*
Patient/age/sexNationalityAge at clinical onset of CNS disease, yearsAge at clinical onset of TRAPS symptoms, yearsFamily historySymptoms suggestive of TRAPSCNS diseaseMutation
  • *

    MS = multiple sclerosis; TRAPS = tumor necrosis factor receptor–associated periodic syndrome; CNS = central nervous system; HAD = hereditary autoinflammatory disease; NMO = neuromyelitis optica; CIS = clinically isolated syndrome.

  • According to the criteria of McDonald et al (18).

1/43/FGerman4038NegativeRecurrent fever, arthralgias, abdominal pain, diarrheaMSNone
2/45/FCroatian3820Positive for HADArthralgias/arthritis, fatigue, myalgias, recurrent infections, erythematous rashMSTNFRSF1A R92Q
3/46/FGerman4532Positive for HADArthralgias/arthritis, joint effusions, fatigue, erythematous rash, periorbital edemaMSTNFRSF1A R92Q
4/60/FGerman4858Positive for HADArthralgias, myalgias, severe fatigue, headacheMSNone
5/36/FGerman2833Positive for MSArthralgias, myalgias, recurrent fever, fatigueMSMEFV E148Q
6/47/FGerman4737Positive for HADArthritis, myalgias, pleuritis, conjunctivitis, recurrent feverNMONone
7/33/FCroatian3129Positive for HADPharyngitis, fatigue, urticaria-like rash, myalgias, chills, conjunctivitisCIS-MSTNFRSF1A R92Q
8/33/MGerman2729NegativeHeadache, myalgias, recurrent fever, erythemaMSNone
9/37/FGerman3737Positive for HADArthralgias/arthritis, myalgiasCIS-MSMEFV A289V
10/26/FGerman2026NegativeDiarrhea, headache, myalgias, fatigueMSNone
11/59/FGerman4243NegativeArthralgias, myalgias, fatigue, recurrent tonsillitis, infectionsMSNone
12/20/FGerman1815Positive for HADArthralgias, nonallergic urticaria-like erythema, myalgias, fatigue, headacheMSNone
13/34/FGerman3030NegativeLumbalgias, myalgias, erythematous skin rash, recurrent feverMSNone
14/36/MGerman3435NegativeArthralgias, myalgias, lymphadenopathyMSNone
15/40/FGerman4032Positive for HADArthralgias/arthritis, fatigue, myalgias, urticarial rashMSTNFRSF1A R92Q
16/40/FGerman2040Positive for MSErythema and swelling of the face, fever, myalgias, recurrent infectionsMSNone
17/50/FGerman3150NegativeArthralgias, myalgias, fatigueMSNone
18/39/FGerman2722Positive for HADLymphadenopathy, conjunctivitis, arthritis, myalgias, fatigue, headacheMSTNFRSF1A R92Q
19/22/MArabian1616NegativeMyalgias, lower back pain, fatigue, abdominal pain, diarrheaMSNone
20/52/FGerman1342NegativeArthralgias, myalgias, recurrent infections of unknown origin, episodes of feverMSNone
21/39/MGerman3536NegativeLower back pain, abdominal pain, myalgiasMSNone
22/34/FGerman3430NegativeLymphadenopathy, arthralgias, recurrent tonsillitisMSNone
23/32/MGerman/Greek2715Positive for HADArthralgias/arthritis, fatigue, periorbital edema, urticaria-like erythemaMSTNFRSF1A R92Q, MEFV V726A
24/42/FGerman3839NegativeArthralgias, arthritis, nonallergic urticaria-like erythema, fatigueMSNone
25/42/FGerman4015Positive for HADArthritis, arthralgias, episodes of fever, ankylosing spondylitisMSNone

All 6 R92Q-positive patients and 1 of the 2 patients with a single mutation in the MEFV gene had a family history suggestive of a dominantly inherited autoinflammatory syndrome. In all TRAPS cases, we identified at least 1 additional family member who was heterozygous for the R92Q mutation. In 4 of the 6 families of R92Q-positive patients, both symptomatic and asymptomatic family members were recruited for genetic testing. All individuals with symptoms suggestive of TRAPS were also found to carry the R92Q substitution. Only the father of index patient 15 was an asymptomatic carrier of the R92Q mutation (Figure 1).

Figure 1.

Pedigrees of all families with the R92Q mutation and their respective haplotypes. Numbers and letters within boxes represent the haplotype of the R92Q mutation–positive individuals. Affected individuals are indicated by solid symbols. Family members without a history of tumor necrosis factor receptor–associated periodic syndrome are indicated by open symbols. Slashes represent deceased persons. Patient numbers refer to patients described in the Results and in Table 1. Individuals without an indicated haplotype could not be tested.

Of the subsequently tested 365 MS patients, 17 (5 males and 12 females) were R92Q heterozygotes (carrier frequency 4.66%, 95% confidence interval [95% CI] 2.8–7.5), while 12 (7 males and 5 females) of the 407 Caucasian controls tested positive for the respective nucleotide substitution in exon 4 of the TNFRSF1A gene (carrier frequency 2.95%, 95% CI 1.6–5.2). There was no statistical difference between the carrier frequency in MS patients and that in controls (P = 0.112; statistical power ∼20%).

Haplotype analysis findings.

Haplotypes were constructed from 3 intragenic SNPs in exon 1, intron 4, and intron 7 of the TNFRSF1A gene as well as from a polymorphic microsatellite marker in intron 1 (12, 13, 19–21). The mutation-positive individuals of 5 families carried the identical intragenic haplotype (c.740-9T;c.473-33C;c.362A;c.39+1899[GT]15[GA]9;c.36G). In contrast, patient 2 and her mother, who were of Croatian origin, carried a different intragenic haplotype, namely, c.740-9T;c.473-33C;c.362A;c.39+1899(GT)16(GA)12;c.36A (Figure 1).

Case reports.

Patient 2.

The patient, a woman who is currently 45 years old, experienced episodes of arthralgias/arthritis and myalgias of unknown origin beginning at age 21 and disappearing after 2 years. Several years later, she developed hypoesthesias and a gait ataxia. After recurrent relapses, the diagnosis of relapsing–remitting MS was made on the basis of typical white matter lesions on cerebral magnetic resonance imaging (MRI) and the presence of oligoclonal bands on analysis of cerebrospinal fluid (CSF). Treatment with IFN beta-1a was started in 1999. During the following years, she again periodically experienced symptoms such as myalgias, arthralgias/arthritis, and severe fatigue as well as an intermittent urticaria-like rash, which increased in severity over time and caused great suffering. Rheumatologic investigations excluded autoimmune disorders such as SLE or RA. Molecular genetic testing for TRAPS revealed a heterozygous c.362G>A replacement in exon 4 of the TNFRSF1A gene, which resulted in the substitution of residue 92 of the mature protein, arginine (CGG), by glutamine (CAG). The TRAPS-related symptoms improved with nonsteroidal antiinflammatory drugs (NSAIDs) taken during the episodes.

The patient's grandmother and mother also reported recurrent episodes of arthralgias and myalgias as well as severe lumbar pain suggestive of TRAPS. So far, her mother has been tested and was shown to carry the same mutation as the patient.

Patient 3.

The patient, a 45-year-old woman, reported recurrent, unexplained episodes of arthralgias and a febrile swelling in her knees, wrists, and ankles as well as an intermittent erythema of the chest and a swelling of her eyelid since 1992. In 2004, she experienced a hemihypoesthesia and gait disturbances. A cerebral MRI showed multiple, partially contrast-enhancing periventricular lesions. Analysis of CSF revealed the presence of oligoclonal bands. In 2005, the diagnosis of relapsing–remitting MS was established on the basis of dissemination of lesions in time and space on the MRI.

After exclusion of rheumatic diseases such as SLE or Wegener's granulomatosis (WG), a DNA sequence analysis of the TNFRSF1A gene was performed which demonstrated a c.362G>A mutation in exon 4. The patient's symptoms responded well to treatment with NSAIDs when necessary.

The patient's mother had also had ailments typical of TRAPS. She had experienced episodes of severe arthritis and myalgias for >20 years before being diagnosed in 2002 as having WG, and she died in 2004. The patient's grandmother was also reported to have had episodic arthralgias, myalgias, and back pain. Recently, the patient's 26-year-old daughter also experienced episodes of arthralgias, abdominal pain, pharyngitis, and fatigue. Although she had been seen by several specialists, there were no conclusive findings. DNA analysis showed that she had inherited the TNFRSF1A R92Q mutation from her mother.

Patient 7.

The patient, a 31-year-old woman, had had episodes of tonsillitis, urticaria-like rash, conjunctivitis, and myalgias since 2002, before she presented in 2003 to our neurology department with hypoesthesia on the left side of her face associated with a painful tingling in her left arm. An MRI scan of the brain revealed several subcortical and deep white matter lesions. CSF examination showed mild pleocytosis as well as oligoclonal bands. An infectious CNS disease as well as autoimmune disorders were excluded, and a clinically isolated syndrome was diagnosed as the first manifestation of MS. Two years later, she experienced attacks of chills, myalgias, and severe fatigue, lasting several weeks.

Genetic testing showed her to be heterozygous for the TNFRSF1A R92Q mutation. Treatment with NSAIDs and corticosteroids led to an improvement of her condition.

The patient's father, who had reported having only arthralgias in the past, also tested positive for the R92Q substitution. Her brother had experienced recurrent infections, episodes of abdominal pain of unknown origin, and conjunctivitis over the previous few years. He also had a history of depression and severe anorexia 4 years previously. When he came to our department for molecular testing, he had a conjunctivitis that had affected his right eye for several weeks. He was also found to carry the R92Q mutation.

Patient 15.

The patient, a 40-year-old woman, had arthralgias/arthritis in the wrists and ankles, neck pain, and severe fatigue, as well as an intermittent urticaria-like exanthema since summer 2005. Her condition worsened, and she developed diplopia, ataxia, and a bladder disturbance in November 2005. She was referred to a local neurologic hospital, and a cerebral MRI showed 3 white matter lesions, 1 with contrast enhancement. CSF investigation revealed a normal cell count but an increased IgG index and oligoclonal bands. The diagnosis of relapsing–remitting MS was made in January 2006, when she had another relapse of disease, this time with weakness of her right arm. Her medical history revealed that 7 years previously she had had an episode of high fever of unknown origin lasting 2 weeks, which had led to a stillbirth in the 11th week of gestation. After intensive laboratory investigations had excluded rheumatic diseases, exons 2, 3, 4, and 6 of the TNFRSF1A gene were sequenced. This led to the detection of the c.362G>A mutation also found in the other patients. Therapy with NSAIDs and corticosteroids was initiated, and her symptoms improved.

The patient's paternal grandmother had had the same symptoms of severe arthralgia, myalgia, and fatigue. Her parents and brother reported no typical TRAPS symptoms, but the genetic analysis demonstrated that she had inherited the R92Q mutation from her father. Her brother did not carry the R92Q exchange.

Patient 18.

The patient, a 39-year-old woman, had a history of recurrent conjunctivitis of unknown origin in 1990, a bout of lymphadenopathy in 1991, recurrent episodes of arthralgias and joint swelling in 1994, and migraine for many years. In 1995, she was diagnosed as having relapsing–remitting MS on the basis of the clinical presentation with relapses of neurologic symptoms, typical findings on MRI of the brain, and the presence of oligoclonal bands in the CSF. Immunomodulatory therapy with IFN beta-1a was initiated in 2003. The patient subsequently developed a severe depression as well as myalgias and fatigue independently of the IFN beta-1a injections, and therapy with IFN beta-1a was stopped. Following exclusion of autoimmune disorders, a genetic analysis of the TNFRSF1A gene was performed, which demonstrated heterozygosity for the R92Q mutation.

The patient's mother was reported to have migraine, and her father had a history of depression, bronchial asthma, and sacroiliitis. Her father's deceased mother had had a history of polyarthritis. Genetic examination of the patient's parents revealed that the father carried the R92Q mutation.

Patient 23.

The patient, a 32-year-old man, reported experiencing recurrent episodes of arthritis and joint effusions of unknown origin in both knees starting at the age of 15 years. In 2001, he had severe fatigue, abdominal pain, nausea, and depression. A diagnosis of relapsing–remitting MS was made after he experienced the first clinical relapse with cervical myelitis and showed dissemination in space and time clinically and on MRI. In 2005, he began therapy with glatiramer acetate and initially tolerated it well. In 2006, his face became swollen, and periorbital edema, arthralgias/arthritis, and an urticaria-like exanthema developed. A dermatologist diagnosed a possible allergic reaction to glatiramer acetate, and the treatment was therefore stopped. However, he continued to have episodic periorbital edema, arthralgias, and urticarial exanthema, and he consulted us again.

Genetic analysis revealed the R92Q mutation encoded by exon 4 of the TNFRSF1A gene as well as a V726A substitution encoded by exon 10 of the MEFV gene. The patient's 64-year-old mother also carried both mutations. She reported having unexplained recurrent arthralgias and joint effusions in her knees for many years. Her mother had had recurrent venous thromboses and arthralgias and had died of cardiovascular disease at age 81 years. The patient's father reported no symptoms suggestive of a periodic autoinflammatory syndrome.


We identified a group of 6 patients with MS (according to the criteria of McDonald et al [18]) and late-onset TRAPS characterized mainly by arthralgias/arthritis, myalgias, urticaria-like rash, fatigue, and lack of typical febrile episodes. All these individuals were heterozygous carriers of the R92Q mutation encoded by exon 4 of the TNFRSF1A gene. One of them was also a heterozygote for the V726A mutation encoded by exon 10 of the MEFV gene, while an additional 2 MS patients had inherited the low-penetrance E148Q mutation and the A289V variant of unknown significance, both encoded by exon 2. The overall frequency of the R92Q mutation was not significantly different between MS patients and controls, but it was higher than has been reported in the past.

Our observations have the following implications: 1) TRAPS appears to be more common in the adult population than previously suspected. 2) In MS patients with recurrent symptoms such as arthralgias/arthritis, myalgias, urticarial rash, and fatigue, late-onset TRAPS should be considered as an important differential diagnosis after other autoimmune disorders such as SLE or RA have been excluded. 3) A careful evaluation of the personal and family history is important to identify MS patients with coexisting TRAPS. 4) In patients without a TNFRSF1A mutation, an MEFV gene analysis is indicated even when the individual does not belong to an ethnic group commonly affected by FMF. 5) A correct diagnosis is the prerequisite for adequate therapy in these patients and may also help to prevent side effects of MS drug treatment.

An association between TRAPS as well as FMF and inflammatory CNS disease has already been suggested. Akman-Demir et al (22) reported a 4-fold higher prevalence of FMF among patients with definite MS compared with that in the Turkish population, and they discussed a possible relationship between FMF and CNS involvement. In a study by Shinar et al (23), it was argued that mutations in the MEFV gene and especially the M694V/p.Met694Val exchange may increase the inflammatory damage, since this mutation was associated with a rapid progression to disability in non-Ashkenazi Jewish patients with MS. CNS involvement in patients with TRAPS has also been reported in the literature, but none of these patients has so far been diagnosed as having MS. A TRAPS patient with a C55R mutation reported by Rudofsky et al (10) had a single inflammatory, nondemyelinating lesion in the cerebellum, which was not consistent with MS. In another patient with the T50K mutation, the disease course and the CSF findings were consistent with MS, but the authors argued against the presence of the disease because the results of MRI of the brain were not typical (17).

In the course of our study, we also identified R92Q mutation carriers with neuropsychiatric manifestations (e.g., sacral radiculomyelitis, depression, hypoesthesias) who did not fulfill the diagnostic criteria for MS. In our experience, a careful investigation, including the disease course, laboratory and CSF examinations, evoked potentials, and cerebral and spinal MRI, is important to diagnose MS in patients with TRAPS and to distinguish “TRAPS plus MS” from “CNS-TRAPS.”

The R92Q substitution is the most frequent TRAPS-associated mutation. It has a low penetrance and appears to result in a more heterogeneous clinical spectrum compared with other mutations (5, 12, 24, 25). Currently, TRAPS is almost exclusively diagnosed in children, and affected adults are usually identified through the young index patient. However, all our mutation-positive MS patients became symptomatic mainly in adulthood and showed a late-onset form of the disease with a milder course than is usually observed in patients with periodic fever syndromes. Moreover, they all reported more or less similar symptoms, such as musculoskeletal involvement, urticaria-like rash, and severe fatigue, which were already present before the onset of MS. None of our patients had typical periodic fever flares and/or a high acute-phase response in their blood during their TRAPS episodes.

The R92Q carrier frequency of 24% in our selected group of MS patients with additional TRAPS-like symptoms is higher than that in any of the other patient groups studied thus far. For example, the incidence of the R92Q mutation was recently found to be ∼5% in patients with a clinically suspected inherited autoinflammatory syndrome or with early arthritis (12, 26). One reason for the high frequency of the R92Q mutation in our patient group could be differences in the inclusion criteria for genetic testing, which, in our study, did not require fever flares. The carrier frequency of 2.95% in our Caucasian control population is also higher than that reported in a previous study by Aksentijevich et al (1%) (12), but lower than the carrier frequency of 4.5% observed in Italian controls (24). These discrepancies might be explained by different genetic backgrounds. Nevertheless, our results suggest that TRAPS (and maybe also FMF) is an underrecognized disease entity in Central Europeans. More patients with TRAPS might be identified if fever and acute-phase response were not required as diagnostic criteria. The fact that no mutation was found in the remaining 17 MS patients, who also presented with a clinical history compatible with TRAPS, is consistent with the genetic heterogeneity of periodic fever syndromes. It is also in line with the data reported by Aksentijevich et al, who detected no TNFRSF1A mutation in 90 sporadic cases clinically suggestive of TRAPS (12).

So far, there has been only 1 report of TNFRSF1A gene variants in MS patients. In a cohort of 100 MS patients, no nucleotide substitutions were found in exon 4 (27). In our current study, the R92Q mutation was detected in 17 of 365 MS patients (4.66%), which was slightly higher than, but not significantly different from, the frequency observed in our control population (2.95%). However, a larger genetic study with enough statistical power is necessary to prove or disprove a possible relationship between the R92Q mutation and MS. For example, 2,000 controls and 2,000 MS patients would have to be analyzed to achieve a statistical power of ∼80%.

There were no distinct differences between mutation-positive and mutation-negative patients with regard to symptoms, age at onset, or course of their MS. However, in all but 1 mutation-positive patient, the family history pointed to an inherited disorder, while only 5 of the 17 mutation-negative patients had a family history of symptoms suggestive of TRAPS. TRAPS symptoms had been present before MS onset in all mutation-positive patients but had preceded MS in only 5 of the mutation-negative patients. Thus, a detailed family history as well as past medical history can help to identify MS patients with an autoinflammatory syndrome.

The pathophysiologic mechanisms by which the R92Q substitution results in TRAPS are still incompletely understood. In contrast to other TNFRSF1A mutants, the R92Q variant behaves like the wild-type receptor in in vitro studies (15, 16). Clinically, the R92Q mutation also differs from other TNFRSF1A mutants. It is associated with a more heterogeneous spectrum of symptoms, a wider range in age and in duration of attacks, and atypical clinical presentations including atherosclerotic vascular disease, extracranial venous thrombosis in Behçet's disease, and early arthritis (12, 28, 29). It was previously suggested that the R92Q mutation may exert a proinflammatory role in autoimmune diseases (2), which would be consistent with our findings in MS patients.

Haplotype analysis of R92Q-positive patients and their family members demonstrated that a common haplotype (c.740-9T;c.473-33C;c.362A;c.39+1899[GT]15[GA]9;c.36G) was associated with the R92Q substitution in 5 of the 6 families. This finding is concordant with observations by Aksentijevich et al (12) and Aganna et al (26) and again demonstrates that the R92Q mutation is relatively ancient (13). In a family of Croatian origin, however, both affected family members carried an intragenic haplotype which differed in the exon 1 SNP (A instead of G) and in the length of the GT (16 instead of 15) and GA (12 instead of 9) repeats of the polymorphic microsatellite marker in intron 1. This could suggest either that the R92Q mutation arose twice or that a recombination event occurred that involved the 2 markers in the 5′-region of the TNFRSF1A gene. Aksentijevich et al (12) have shown that there is no association of the R92Q substitution with telomeric flanking markers D12S99 and D12S356 and with centromeric flanking markers CD4, D12S1695, and D12S77 outside the TNFRSF1A gene. This rules out the possibility that a genetic defect in another gene linked to the R92Q mutation is responsible for the TRAPS symptoms in our mutation-positive patients.

Analysis of family members of R92Q carriers demonstrated transmission of systemic TRAPS symptoms together with the mutation in 5 of the 6 families, but never transmission of MS. Clearly, there must exist additional genetic and/or environmental factors that contribute to MS development in the index patients. We cannot, however, exclude the possibility that altered TNF/TNFRSF1A signaling contributes to the MS in genetically predisposed patients. This would be supported by the fact that all mutation-positive patients experienced their TRAPS episodes usually years before the onset of MS. TNFα and TNFRSF1A signaling play an important but very complex role in mediating inflammatory demyelinating diseases of the CNS (30, 31). It was shown that TNFα is elevated in the sera of patients during relapses and/or in patients showing contrast-enhanced lesions on the MRI as a sign of disease activity as compared with controls or patients with disease in remission (32, 33). In addition, TNFα can induce the expression of adhesion molecules, which guide the migration of T cells into the CNS (34). The majority of the biologic responses attributed to TNFα seem to be mediated through TNFRSF1A (35), and TNFRSF1A signaling itself appears to be important for the initiation of the inflammatory process in the CNS (30). However, TNFRSF1A could also be involved in the clearance of lymphocytes from the CNS, since impairment of the TNFRSF1A pathway led to a 50% reduction of T cell apoptosis (36).

The long-term therapeutic strategy for MS patients with TRAPS is unclear at present and needs careful evaluation. On the one hand, it is known that anti-TNFα therapy can exacerbate and even induce inflammatory demyelinating disorders of the CNS (37, 38). On the other hand, it is not known how existing immunomodulatory therapies such as IFN beta and glatiramer acetate affect coexisting TRAPS. There is 1 report of a TRAPS patient who was classified as having “CNS-TRAPS” in whom therapy with etanercept led to new relapses of neurologic symptoms and an increase of demyelinating lesions (17).

In conclusion, the results of our study have shown that a significant proportion of MS patients with “rheumatic” problems, skin lesions, and severe fatigue actually have a late-onset form of TRAPS. The described cases, together with the high frequency of the R92Q mutation in control chromosomes, suggest that TRAPS may be quite common in adults. They also illustrate the wide spectrum of symptoms associated with this particular mutation.

In order to prevent unnecessary investigations and hospitalizations, TRAPS should therefore be considered in MS patients with unexplained recurrent symptoms such as myalgia, arthralgia/arthritis, urticarial rash, and severe fatigue, even when there is no history of fever flares. Careful evaluation of the medical history and family history is also essential to identify this autosomal-dominantly inherited disorder in MS patients.


Dr. Kümpfel had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Kümpfel, Hoffmann, Lohse.

Acquisition of data. Kümpfel, Hoffmann, Rübsamen, Pöllmann, Feneberg, Lohse.

Analysis and interpretation of data. Kümpfel, Hoffmann, Rübsamen, Hohlfeld, Lohse.

Manuscript preparation. Kümpfel, Hoffmann, Hohlfeld, Lohse.

Statistical analysis. Kümpfel.

Patient care. Hoffmann.


We would like to thank all of our patients for their participation in this study. The expert technical assistance of Pia Lohse and Gabriele Simon (Department of Clinical Chemistry–Grosshadern) is also gratefully acknowledged. In addition, we would like to thank Dr. Dieter Jenne (Department of Neuroimmunology, Max-Planck-Institute of Neurobiology, Martinsried, Germany) for helpful comments, and Ms Judy Benson for carefully editing the manuscript. Special thanks go to B. H. Belohradsky (Department of Infectious Diseases and Immunology, Children's Hospital, University of Munich, Munich, Germany).