La fermentation d’oligosaccharides bovins par B. mongoliense et B. crudilactis module la virulence d’agents pathogènes et leur présence modifie la composition du microbiote infantile dans le modèle gastro-intestinal SHIME®

Introduction
Human milk oligosaccharides (HMO) allow growth of beneficial bifidobacteria improving children health [1]. With their structural similarities, bovine milk oligosaccharides (BMO) could be easily metabolised by these same bifidobacteria, as well as Bifidobacterium crudilactis, a species from bovine origin [2; 3]. Also, cell free spent media (CFSM) obtained from culture with B. crudilactis and a majority BMO, the 3’-sialyllactose (3’SL), modulated virulence genes expression of Escherichia coli O157:H7 [4].

Material and methods
The SHIME®, an in vitro gastro-intestinal tract, was inoculated with feces from young child and stabilised for 2 weeks. For treatments were successively administrated for 1 week, with a break of 1 week between each: 3’SL (I), B. crudilactis (II), 3’SL and B. crudilactis simultaneously (III) and CFSM from 3’SL and B. crudilactis culture (IV). In each section of the colon, different samplings have been done and analysed by HPLC to determinate short chain fat acids (SCFA) concentrations. Microbial populations were determinate by 16S rDNA metagenetic analysis. Impact of supernatants was evaluated on E. coli O157:H7 virulence genes expression.

Results
The results showed that SCFA levels were stable during the experiments. Metagenetic analysis showed a microbial diversity in transverse (TC) and descending colon (DC) close to feces, dominated by Bacteroides, Prevotella and Fusobacterium, while the ascending colon (AC) showed a different microbial diversity dominated by Veillonella. Treatments II and IV induced mainly a down-regulation of virulence genes: luxS, stx1, qseA in AC, DC or TC, and fliC and qseA in DC, respectively. Treatment 1 also showed a down-regulation of fliC in DC, similar to the one observed with treatment IV, but this was associated with an up-regulation of fliC, stx1 and qseA in AC or TC. Finally, treatment III showed slight upregulation of ler, fliC and qseA in AC. Diversity indices (Chao1, Simpson’s reciprocal index, eveness and PCoA) are being analized, as is the second run.

Discussion
Interesting effects have been highlighted after this first run. SCFA allowed an overview of dominant populations while metagenetic offered a more advanced investigation. The trends observed with E. coli O157:H7 virulence gene expression have to be validated with the further replicates on the SHIME® system, but these results show that treatments II and IV might have a positive effect against its pathogenicity.