УДК 616-056:616.34-008.87

   Keywords: familial Mediterranean fever, gut microbiota, phylogenetic analysis

   Familial Mediterranean fever (FMF; MIM249100) is a recessively inherited disorder of the inflammatory pathway, manifested by acute self-limited recurrent episodes of fever and polyserositis [1]. The Mediterranean fever gene (MEFV), responsible for the disease, has been recently identified by positional cloning [2, 3]. Pyrin, the protein product of MEFV, consists of several conserved domains, including the N-terminal pyrin domain (PYD), which is found in a number of autoinflammatory proteins involved in the regulation of inflammation and apoptosis [4]. According to recent studies, autoinflammatory genes, such as MEFV, may represent an exaggerated innate immune response to various signals in vitro, including microbial products [5]. Indeed, the CARD15/NOD2 gene product belongs to the same superfamily of proteins [6], and its mutations have been found to underlie inflammatory bowel diseases (IBD), such as Crohn's disease, in which an inappropriate immune response to components of the commensal microbiota exists [7]. In this regard, it has been proposed to investigate the composition of gut microbiota in FMF to reveal a possible contribution of commensal bacteria to the onset and maintenance of the disease.
   As a large majority of bacterial species is effectively unculturable, it is impossible for detailed examination of gut microorganisms to be achieved through traditional culture techniques. Molecular-genetic analyses of bacterial microbiota based on 16S ribosomal ribonucleic acid (rRNA) genes obviate the need for culture and have been shown to be powerful tools in determining microbial diversity in complex samples [8].
   In the present study, the fecal bacterial composition has been for the first time examined in FMF by using microbial community analysis through sequencing of 16S rDNA libraries.
   Fecal samples were collected from genetically ascertained FMF patients (12 patients in remission, 3 patients in attack periods) and 7 healthy individuals.
   DNA was extracted from fecal samples of FMF patients and healthy subjects using QIAamp DNA Stool Mini Kit (Qiagen, UK), according to the manufacturer's instructions. DNA samples were transferred to the Rowett Research Institute (UK) where 16S rDNA clone libraries were generated, and phylogenetic analysis was performed. Bacterial 16S rDNA was PCR-amplified with universal primers covering most intestinal bacterial species (Table 1). The amplicons were cloned into Escherichia coli chemically competent cells using the pCR-4 TOPO TA Cloning Kit (Invitrogen, UK), according to the manufacturer's instructions. Recombinant colonies were randomly picked and sequenced on the automated DNA-sequencer (Beckman, USA) with 926R bacterial primer (Table 1). Alignment of sequences with reference 16S rDNA gene sequences from healthy gut microflora was performed using the multiple sequence alignment programme CLUSTALX v. 1.83 [9]. Phylogenetic analyses were performed using the neighbor-joining algorithm [10]. Operational taxonomic units (OTUs) were identified by online Basic Local Alignment Search Tool (BLAST) program at the NCBI website [11], using search results of at least 99% sequence similarity.

Тable 1

Тable 2

   As shown in Table 2, Bacteroides was the most abundant group in all three cohorts, followed by the Firmicutes. The relative proportions of CFB and Firmicutes were not markedly different among the three groups; however, significant differences were detected in bacterial subgroups within these main phyla (Fig. 1A, B, C). In Fig. 1B groups of FMF patients in remission were determined according to the biodiversity in the main phylogenetic groups, demonstrating high variability, in contrast to stable composition of gut bacteria in healthy state (Fig. 1A). Particularly, there is a group (FMF2, FMF4, FMF5, FMF6, FMF9, FMF11, and FMF12) overrepresented by OTUs belonging to the CFB phylum, which amounted up to 50-55% of gut bacteria in healthy subjects (Fig. 1A) and acute FMF (Fig. 1C). In the second group of FMF remissions (FMF1, FMF7, FMF8, and FMF10) there is a substantially higher proportion of cluster IX of Propionate-producing bacteria, as compared to healthy controls. Interestingly, these bacteria tended to disappear during the attack period (Fig. 1C). The pairwise comparisons of each 16S rDNA library to every other library also revealed significant alterations in gut microbiota composition in FMF compared to the norm (Table 3). In particular, the Prevotellaceae subgroup (within CFB) was significantly low in active stage of FMF as compared to FMF remission and healthy state (16.5%, 22% and 27.6%, respectively), in contrast to Bacteroidaceae (within CFB) subgroup (30.7%, 17.8% and 21.7%, acute, remission and healthy, respectively). The Butyrate-producing Faecalibacterium group was higher in active FMF compared to both FMF remissions and controls (14.2% in attack vs. 6.5%). Gamma-proteobacteria were 0.2% and 2.1%, in healthy controls and FMF remission, respectively, and there was a complete loss of these bacteria in the acute phase. The most striking difference was observed in the Propionate-producing Acidaminococcaceae subgroup (Clostridial cluster IX within Firmicutes). These bacteria were overrepresented in remission period compared to controls (16% vs. 10%), and tended to disappear in attack (3%), found only in FMF15 (Fig. 1C). Although in the latter group the bacterial sequences were the least diverse, which might be the consequence of a general inflammatory process, however representatives of the Butyrate-producing Faecalibacterium group in attack were significantly high compared to both FMF remission and healthy state (Table 3). Butyrate, which is produced by bacterial fermentation, has been shown to reduce inflammation in experimental colitis in animal models. It reduces inflammation through an inhibitory effect on proinflammatory cytokine expression, thus demonstrating anti-inflammatory properties [12]. Such increase of butyrate producers among acute patients implies that it could correspond to a compensative response.

Тable 3