Phage-mediated dispersal of biofilm and distribution of bacterial virulence genes is induced by quorum sensing.

The microbiome and the phage meta-genome within the human gut are influenced by antibiotic treatments. Identifying a novel mechanism, here we demonstrate that bacteria use the universal communication molecule AI-2 to induce virulence genes and transfer them via phage release. High concentrations (i.e. 100 µM) of AI-2 promote dispersal of bacteria from already established biofilms, and is associated with release of phages capable of infecting other bacteria. Enterococcus faecalis V583ΔABC harbours 7 prophages in its genome, and a mutant deficient in one of these prophages (i.e. prophage 5) showed a greatly reduced dispersal of biofilm. Infection of a probiotic E. faecalis strain without lytic prophages with prophage 5 resulted in increased biofilm formation and also in biofilm dispersal upon induction with AI-2. Infection of the probiotic E. faecalis strain with phage-containing supernatants released through AI-2 from E. faecalis V583ΔABC resulted in a strong increase in pathogenicity of this strain. The polylysogenic probiotic strain was also more virulent in a mouse sepsis model and a rat endocarditis model. Both AI-2 and ciprofloxacin lead to phage release, indicating that conditions in the gastrointestinal tract of hospitalized patients treated with antibiotics might lead to distribution of virulence genes to apathogenic enterococci and possibly also to other commensals or even to beneficial probiotic strains.

Isolation of highly active monoclonal antibodies against multiresistant gram-positive bacteria.

Multiresistant nosocomial pathogens often cause life-threatening infections that are sometimes untreatable with currently available antibiotics. Staphylococci and enterococci are the predominant Gram-positive species associated with hospital-acquired infections. These infections often lead to extended hospital stay and excess mortality. In this study, a panel of fully human monoclonal antibodies was isolated from a healthy individual by selection of B-cells producing antibodies with high opsonic killing against E. faecalis 12030. Variable domains (VH and VL) of these immunoglobulin genes were amplified by PCR and cloned into an eukaryotic expression vector containing the constant domains of a human IgG1 molecule and the human lambda constant domain. These constructs were transfected into CHO cells and culture supernatants were collected and tested by opsonophagocytic assay against E. faecalis and S. aureus strains (including MRSA). At concentrations of 600 pg/ml, opsonic killing was between 40% and 70% against all strains tested. Monoclonal antibodies were also evaluated in a mouse sepsis model (using S. aureus LAC and E. faecium), a mouse peritonitis model (using S. aureus Newman and LAC) and a rat endocarditis model (using E. faecalis 12030) and were shown to provide protection in all models at a concentration of 4 µg/kg per animal. Here we present a method to produce fully human IgG1 monoclonal antibodies that are opsonic in vitro and protective in vivo against several multiresistant Gram-positive bacteria. The monoclonal antibodies presented in this study are significantly more effective compared to another monoclonal antibody currently in clinical trials.