Bacteriophages are the most abundant biological entities on Earth with an estimated number of up to 1030 particles. Every bacterial cell can potentially harbor many phages. These viruses use bacteria as a mean to replicate, almost always
destroying their prokaryotic host in the process. Current knowledge states that phages ignore every cell but the strain of bacteria they have evolved to inhabit. This makes them ideal candidates to treat bacterial infections, while being harmless to
mammalian cells and even non-target bacteria.
Phages provide a valid alternative to antibiotics. Nevertheless, the effect of phages on the human immune response needs to be evaluated. Therefore we need to obtain highly purified phage preparations, free of endotoxins. The removal of these
endotoxins is necessary as they are highly immunogenic (as highlighted in ) and could influence the induced immune response. For this purpose, we evaluated the endotoxin removal efficacy of seven different endotoxin removal strategies,
described in . These strategies consisted of Endotrap HD column purification and/or CsCl density centrifugation in combination with Endotrap purification, followed by organic solvent (1-octanol) treatment, detergent (Triton X100) treatment, enzymatic inactivation of the endotoxin using alkaline phosphatase, or removal of the endotoxin using CIM monolithic anion exchange chromatography, carried out for five different phages (i.e. four P. aeruginosa phages and one S. aureus phage).
We showed that CsCl density purification of the P. aeruginosa phages, at an initial concentration of 1012 - 1013 pfu/ml, led to the strongest reduction of endotoxins, with an endotoxin removal efficacy of up to 99.6 %, whereas additional purification
methods yielded an additional endotoxin removal efficacy of 23 to 99 % on top of the initial purification, although sometimes accompanied with strong losses in phage titer.
Phage biology studies necessitate highly purified phage particles, when used in high concentration. Using the highly purified (CsCl density centrifugated, followed by Endotrap purification), endotoxin free, phage preparations, described in ,
we were able to study the immune response induced by these phages. describes a transcriptome analysis of freshly isolated peripheral blood mononuclear cells (PBMCs), isolated from one healthy individual, stimulated for 20 h with either a P. aeruginosa phage PNM lysate or its bacterial host P. aeruginosa strain 573. The phage PNM lysate was shown to increase the production of IL10, IL6, SOCS1, SOCS3, CXCL2, CXCL3 and CXCL6 and decrease the production of
LYZ, HLA-DMA, HLA-DMB, HLA-DRB1 and HLA-DRB4, CCL17, CCR1, CCR2 and CCR5. These results showed that the P. aeruginosa phage PNM lysate possess the potential to induce an immune response.
Further analysis on the potential of phages to induce an immune response was obtained by evaluating five different phages (S. aureus phage ISP and four P. aeruginosa phages (i.e. PNM, LUZ19, 14-1 and GE-vB_Pae-Kakheti25)), discussed
in . By means of specific RT-qPCR assays, we assessed the gene expression profile of PBMCs derived from six donors for twelve immunity-related genes (i.e. CD14, CXCL1, CXCL5, IL1A, IL1B, IL1RN, IL6, IL10, LYZ, SOCS3, TGFBI and
TNFA). It was established that the phages were able to induce clear and reproducible immune responses.
First, the five different phages induced a comparable immune response, although endotoxins could not be completely removed from the P. aeruginosa phage preparations. Second, the observed immune response was largely antiinflammatory, indicated by at least a fivefold up-regulation of IL1RN, IL10 and
SOCS3. Third, the observed immune response was shown to be endotoxinindependent. Addition of endotoxins to the highly purified phages did not cause an immune response comparable to the one induced by the (endotoxin containing)
phage lysate, but remained similar to that of the initial highly purified phage.
Using the observations made during these studies, we were able to construct a predictive hypothetical model on the outcome of phage therapy based on a previous model described by Hodyra-Stefaniak et al. (2015), described in . We
expanded their initial model with the inclusion of an anti-inflammatory phage. This adaptation predicts important consequences on the theoretical outcome of a phage
therapeutic intervention. Our model showed that when the phage has antiinflammatory properties, phage therapy succeeds whereas the initial model showed a failure of the phage therapeutic intervention into the removal of the bacterial
infection.