MEFV gene mutations in neuro‐Behçet's disease and neuro‐Sweet disease

Abstract Mediterranean fever (MEFV) gene mutations are associated with familial Mediterranean fever (FMF). Recent studies have suggested that MEFV gene mutations may act as disease modifiers in neuro‐Behçet's (NBD) disease and neuro‐Sweet disease (NSD). We investigated MEFV genes and clinical features in 17 patients with NBD or NSD. MEFV gene mutations were frequently observed (70.6%). Headaches and exertional leg pain were associated with MEFV gene mutations (P < 0.05). Moreover, higher frequency of white matter lesions without sites predilection (P < 0.05) and non‐parenchymal lesions (P < 0.05) were also observed. MEFV gene mutations may be associated with particular findings and lesion sites.


Introduction
Neuro-Behc ßet's disease (NBD) and neuro-Sweet disease (NSD) are neurological manifestations of Behc ßet's disease (BD) and Sweet disease, respectively, with occasional serious sequelae. 1,2 Considering the overlap in their symptoms, NSD and NBD can be regarded as forms of neuroneutrophilic disease (NND), a broad-spectrum disorder caused by the hyperactivity of neutrophilic cells. 3 Genetically, class I human leukocyte antigen (HLA) serotypes have been associated with BD (e.g., HLA-B51) 4 and NSD (e.g., HLA-B54 and Cw1). 2 According to recent reports, mutations in the Mediterranean fever (MEFV) gene, which is associated with familial Mediterranean fever (FMF), 5 may act as disease modifiers for NBD 6,7 and NSD. 8,9 FMF, an autoinflammatory disease, is characterized by recurrent episodes of fever, serositis, arthritis, and dermal manifestations. 5 BD and FMF share common symptoms, laboratory findings, and treatments. 10 Colchicine is the mainstay of FMF treatment and is also one of the valid therapeutic options for BD. Therefore, there may be a pathophysiological link among NBD, NSD, and FMF. Clarifying the role of MEFV gene in NBD and NSD is important for elucidating the mechanism of central nervous system (CNS) inflammation and developing treatments.
We hypothesized that HLA class I subtypes and MEFV gene mutations might serve as susceptibility, or disease-modifying factors in patients with CNS inflammation, resulting from NND. Furthermore, we hypothesized that specific clinical manifestations related to FMF 5 could indicate MEFV gene mutations in patients with NBD or NSD.
Herein, we investigated FMF-related symptoms and their associations with HLA class I alleles and MEFV gene mutations in patients with NBD or NSD. December 2018 and received confirmed NBD or NSD diagnosis. The International Consensus Recommendation criteria were used for NBD diagnosis, 1 with a diagnosis of definite NBD when all of the following three criteria were met: (1) International Study Group (ISG) criteria for BD 11 ; (2) neurological syndrome due BD and supported by relevant and characteristic abnormalities seen on neuroimaging or cerebrospinal fluid analysis; and (3) no better explanation for the neurological findings; and probable NBD when one of the following two criteria were met in the absence of a better explanation for the neurological findings: (1) neurological syndrome as in definite NBD, with systemic BD features but not satisfying the ISG criteria; (2) a non-characteristic neurological syndrome occurring in the context of ISG criteria-supported BD. NSD was diagnosed using the criteria proposed by Hisanaga et al in 2005 2 and was designated as probable NSD when all of the following three criteria were met: (1) neurological features highly systemic glucocorticoid responsive or sometimes spontaneously remitting, but frequently recurrent encephalitis or meningitis, usually accompanied by fever over 38°C; (2) dermatological featurespainful or tender, dull red erythematous plaques or nodules preferentially occurring on the face, neck, upper limbs, and upper part of the trunk; (3) other featuresabsence of cutaneous vasculitis and thrombosis, which are seen in BD; and possible NSD with any neurological manifestations, either dermatologic features or HLA association (HLA-Cw1 or B54 positive, B51 negative), and at least one item of 'other features' described above. This study followed the Clinical Study Guidelines of the Ethics Committee of Mie University Hospital and was approved by its internal review board (approval number, 1756).

MEFV gene analysis
A MEFV gene analysis was conducted in all patients. Mutations in the five hotspot regions (i.e., exons 1, 2, 3, 5, and 10) were assessed through polymerase chain reaction 12 . Exon 2 was amplified in two overlapping PCR fragments, designated as exon 2a and exon 2b. Amplified PCR products were analyzed by direct sequencing (DNA Analyzer 3730xl; Applied Biosystems, Foster City, CA, USA). Primers for the PCR and sequence analysis were as follows: Exon 1F: 5 0 -TCC TAC CAG AAG CCA GAC AG-

Symptoms and laboratory data
In addition to neurological syndromes (i.e., headache, cognitive impairment, seizure, gait disturbance, and vascular manifestations), symptoms indicative of FMF (i.e., fever, peritonitis, pleuritis or pericarditis, monoarthritis, and exertional leg pain) 5 or NND (i.e., ocular lesions, skin lesions, oral aphthae, benign genital ulcers, and pathergy reaction) 1,2 were assessed. The data of HLA genotypes and laboratory data on blood inflammatory markers, including white blood cell counts, C-reactive protein levels, serum amyloid A levels, and erythrocyte sedimentation rate, were assessed. Inflammatory markers in the cerebrospinal fluid, such as high cellularity, protein, IgG index, or interleukin-6 levels 1 were also assessed.

Data analysis
Patients were divided into a MEFV mutation positive group (MEFV+) and negative group (MEFVÀ) and compared the frequency of HLA class I genotypes between the groups. Group-wise differences in symptoms, laboratory data, and neuroimaging data were also assessed. All data analyses were performed using SPSS version 25 for Windows (IBM Corp.). To assess statistical differences between categorical variables, we used the chi-square test. The significance level was set at 0.05 for all analyses.

Results
The demographic, genetic, radiographic, and therapeutic data of patients are summarized in Table 1. Among the patients who underwent MEFV gene analysis, 12 (70.6%) displayed heterozygous or homozygous mutations. The frequencies of occurrence of HLA-B51, HLA-B54, and HLA-Cw1 were 23.5%, 29.4%, and 52.9%, respectively. The most frequent symptoms were headache (76.5%) and oral mucosal lesions (76.5%), followed by incomplete joint disorders (64.7%). MRI revealed that the lesions in the brain stem (BS) (64.7%) were most common, followed by the white matter lesions, without predilection for a particular site (58.8%).
A comparison of the clinical symptoms of the MEFV positive (+) and MEFV negative (À) groups is presented in Table 2. HLA class I serotypes were not significantly different between the MEFV+ and MEFVÀ groups. Although insignificant, MEFV+ patients had B54 alleles frequently (41.7% vs. 0%, P = 0.086) and A02 alleles less frequently (33.3% vs. 80%, P = 0.079). Common symptoms in the MEFV+ group included headaches (100% vs. 40%, P = 0.003) and exertional leg pain (66.7% vs. 0%; P = 0.012). MRI findings indicated that MEFV+ patients exhibited a higher frequency of white matter lesions without site predilection (75% vs. 20%, P = 0.036) and non-parenchymal lesions (50% vs. 0%, P = 0.049). There were no significant differences in laboratory data including inflammatory markers in blood and cerebrospinal fluid of patients classified as MEFV+ and MEFVÀ. Responses to the treatment were not significant between MEFV+ and MEFVÀ patients.

Discussion
Patients with CNS inflammation and symptoms related to NBD or NSD frequently exhibited MEFV gene mutations. Furthermore, patients with MEFV mutations commonly experienced headaches, exertional leg pain, and sustained white-matter lesions without site predilection and nonparenchymal lesions.
Single heterozygous mutations in the MEFV gene are common and were found to occur in one-quarter of patients diagnosed with FMF. 5 Our results showed that most (>70%) patients with NBD or NSD had MEFV gene mutations. This result supports the hypothesis that MEFV gene mutations may act as disease modifiers in NND. The most frequent mutations in Japanese patients with FMF included E148Q (40.2%), M694I (21.0%), L110P (18.8%), P369S (5.4%), and R408Q (5.4%). 12 In our study, the most common MEFV gene mutations were E148Q (52.9%), R202Q (11.8%), P369S (11.8%), R408Q (11.8%), E84K (5.9%), G304R (5.9%), and L110P (5.9%). Headaches and exertional leg pain are often observed in patients with FMF. 5,7,13 Some mutations reported in Japanese patients with FMF were also observed in patients in our study; therefore, we speculated that the mutations observed in our study contributed to the onset of FMFassociated symptoms, including headaches and exertional leg pain. Although serotypes of HLA class I and MEFV gene mutations showed no significant associations, B54 showed relatively high enrichment among patients with MEFV gene mutations (P < 0.10). The HLA class I B54 allele and MEFV gene mutations may additively contribute to onset of CNS inflammation.
The MRI lesions of patients who were MEFV+ disproportionately occurred in the white matter without site predilection. The typical predilection sites of NBD are brainstem and basal ganglia. 1 Non-parenchymal lesions such as cerebral venous thrombosis and acute meningeal syndrome have been reported as findings of non-parenchymal NBD. 1 White matter lesions without site predilection have been reported as one of the findings of NSD. 2 Our study suggests that patients with CNS inflammation and white matter lesions without site predilection or nonparenchymal lesions have a potential auto-inflammatory background associated with MEFV gene mutations.
While HLA is mainly associated with adaptive immune responses, the inflammasome machinery is present in numerous cell types and contributes to innate immune activation in multiple organs and in the CNS. 14 Inflammasomes are cytosolic protein complexes that sense specific infectious or host stimuli and initiate inflammatory responses through caspase activation. Neurons are known to express several inflammasome proteins, including the nucleotide oligomerization domain, leucine-rich repeat, and pyrin domain-containing protein (NLRP) 1, NLRP3, and are absent-in-melanoma 2. The NLRP3 inflammasome is the most abundant inflammasome in the CNS and is a key contributor to neuroinflammation across a broad spectrum of nervous system disorders. 14 It has been hypothesized that the pyrin protein suppresses the activation of procaspase-1 and therefore interferes with NLRP3 inflammasome activation. 15 Considering that pyrin is encoded by the human MEFV gene, 16 our results suggest that CNS inflammation and symptoms related to NBD and NSD might be associated with dysfunction of pyrindependent regulation of the inflammasome. Although this is the first cohort study investigating MEFV gene mutations in NBD and NSD, MEFV gene mutations have also been reported in multiple sclerosis. 17 Mutations of MEFV gene may be involved in CNS-inflammation in various neurological diseases.
In our study, one patient diagnosed as probable NSD with MEFV mutation was remarkably responsive to colchicine. Colchicine is a treatment for FMF caused by MEFV gene mutations. In patients with CNS inflammation with MEFV mutations, colchicine may be an effective treatment option. Large cohort studies including other diseases manifesting neurological inflammation are warranted to clarify the relationship between MEFV gene mutations and neurological inflammatory diseases.