Clinical Relevance of MEFV Gene Mutations in Japanese Patients with Unexplained Fever

To the Editor:

At the beginning of 2007, we investigated the frequencies of MEFV gene mutations in Japanese patients with unexplained fever or undifferentiated arthritis to determine their role in phenotypical features of familial Mediterranean fever (FMF)-related diseases. Patients were asked to complete a questionnaire concerning fever, recurrent typical attacks of FMF, including peritonitis, pleuritis, and arthritis, and transient inflammatory response. On the basis of the Tel-Hashomer criteria¹, we divided the study subjects into 3 groups, as follows: Group 1, typical FMF (presence of 1 or more major criteria independent of the presence of minor criteria); Group 2, probable FMF (absence of major criteria and 2 or more minor criteria); Group 3, unlikely (not belonging to either Group 1 or 2). Patients who had previously been diagnosed with typical FMF were not included. All patients were first enrolled as having unexplained fever, and finally diagnosed as FMF based on clinical evidence. We stress that the overall survey for the recent clinical manifestations, including the response to colchicine, was not complete in a few patients.

Up to January 2011, we had enrolled 142 Japanese patients with unexplained fever or undifferentiated arthritis in our genetic analysis. The subjects are 86 women and 56 men, with mean age of 38.2 ± 17.8 years. As shown in Table 1, 72 (50.7%) patients had single-nucleotide polymorphisms (SNP) of exon 2 of MEFV gene and 15 (10.6%) patients had SNP of exon 3 of MEFV gene. We identified 16 patients carrying a mutation of exon 10 (M694I): 3 were homozygotes, 11 were compound heterozygotes, and 2 were heterozygotes. All patients having the M694I mutation had typical episodes of serositis or monoarthritis in addition to periodic fever, and had been newly diagnosed as typical FMF. In contrast, the prevalence of FMF in patients with MEFV exon 1, 2, or 3 SNP was markedly lower (36.0%) than that in carriers of M694I. We compared the allele frequencies among typical or incomplete FMF patients and healthy subjects (35 women, 41 men, mean age 31.5 ± 8.0 yrs). The frequencies of M694I and E84K alleles were increased in patients with typical FMF, and frequencies of E148Q, P369S, and R408Q were increased in patients with probable FMF compared to healthy subjects (p < 0.05, Fisher exact test; Table 2).

We identified 12 patients carrying E84K mutation; clinical features of these patients are listed in Table 3. Among these 12, 4 had typical episodes of serositis or synovitis and periodic fever and could be diagnosed as typical FMF (Group 1). There was remittance of clinical symptoms with colchicine therapy in these patients with typical FMF, except for 1 patient (Patient 3) who remitted spontaneously. Another 2 patients carrying E84K mutation were considered to be “probable FMF” (Group 2); one patient remitted spontaneously, and colchicine was beneficial in the other patient. The remaining 3 patients carrying E84K mutation had atypical symptoms and did not fulfill a diagnosis of FMF (Group 3). In the last group, who had been diagnosed as having definite rheumatic diseases (Group 4), E84K mutation may have contributed the modification or sustained musculoskeletal symptoms to concomitant rheumatic diseases despite optimal treatment including steroid and immunosuppressants. We also analyzed the clinical features of the patients carrying the SNP of exon 1, 2, or 3 of the MEFV gene. Similarly, a subgroup of these patients were diagnosed as having typical or probable FMF (Table 4).

In our study, all 16 patients with M694I mutation were newly diagnosed as having typical FMF and showing the higher penetration of these mutations compared to that of exon 1 (E84K), exon 2 (L110P, E148Q, R202Q, G304R), or exon 3 (P369S, R408Q) mutations. Interestingly, we found 12 patients carrying a heterozygous E84K mutation who presented heterogeneous clinical phenotypes, in contrast to M694I carriers with typical FMF. Our findings indicated that a portion of the patients carrying E84K fulfilled the diagnostic criteria for typical FMF; however, more than half of these patients had atypical symptoms. We could not find any relevant clinical similarity in patients with E84K mutation, and the clinical phenotype of these E84K carriers might differ from the homogenous FMF phenotype. Our observations suggest that the MEFV gene mutations, which are attributed mainly to FMF, may also be responsible for additional clinical manifestations that do not meet the criteria of FMF as described^2,3. A significant number of patients diagnosed as FMF have only a single mutation despite sequencing of the entire MEFV genome region or other autoinflammatory genes, and this has led to a reconsideration of the simple loss of function of the recessive model of FMF inheritance^4,5. Recently, Chae, et al⁶ demonstrated that gain-of-function pyrin mutations induce NOD-like receptor family, a pyrin domain containing 3 (NLRP3)-independent inter-leukin 1ß activation and autoinflammation. A plausible explanation might be that a subject having the MEFV single mutation carries a combination of polymorphisms that would favor more inflammation under the influence of a certain environmental factor and cross the threshold of manifesting an FMF phenotype⁷. These polymorphisms would be expected to belong to genes of the innate immune pathway⁸. Possible environmental factors are thought to be the patient’s country of origin, with a geographically related, as yet unknown pathogenesis⁹.

Our data showed a significant prevalence of FMF in Japanese patients with unexplained fever or undifferentiated arthritis; and we have to be more aware of the presence of a variant type of FMF or modification of other diseases by polymorphisms of the MEFV gene.