Therapeutic application of a jumbo bacteriophage against metallo-β-lactamase producing Pseudomonas aeruginosa clinical isolates

Abstract

Background: Therapeutic options against metallo-β-lactamase producing P. aeruginosa (MBL-PA) are limited due to multi-drug resistance. A jumbo phage isolated from wastewater in Greece was characterized microbiologically and genetically and evaluated for its potential as a therapeutic agent alone or in combination with antibiotics in an experimental thigh infection mouse model.

Methods: The host range of the jumbo phage vB_PaerM_AttikonH10 (AttikonH10) against 20 MBL-PA clinical isolates and 10 susceptible strains, one-step phage growth and growth curves of mid-exponential phase bacteria upon addition of the phage were analyzed. Whole-genome sequencing was performed and the de novo assembled complete phage genome was compared with other jumbo phages. In vivo pharmacokinetics in different tissues as well as the efficacy of two dosing regimens 10⁹ and 10⁶ PFU/mouse administered intraperitoneally alone and in combination with amikacin (384 mg/kg/day) was tested against an MBL-PA clinical isolate in murine thigh infection model.

Results: The phage formed small plaques in double-layer agar and demonstrated clear or semi-clear lysis in 83.3% (25/30) of P. aeruginosa clinical isolates. Growth curves showed a > 94% growth inhibition in the presence of phage even at the lowest multiplicity of infection ratio tested (10^-5). Whole genome analysis indicated a jumbo dsDNA phage with 278,406 bp (36.92% GC) belonging to Phikzvirus that is predicted to host up to 413 putative ORFs and 6 tRNA genes. No known lysogeny-associated genes, virulence factors, or antimicrobial resistance genes were identified within the genome. Phage titres 10⁴-10⁶ PFU/tissue were detected in all mouse tissues with elimination half-life of 3.4 h except in bronchoalveolar lavage where no phages were found. Only the high phage dose (10⁹ PFU/mouse) reduced bacterial load in thigh by 1.09 log₁₀ cfu/thigh compared to placebo, similar to amikacin monotherapy (1.19 log₁₀ cfu/thigh), while their combination achieved a greater reduction of 2.07 log₁₀ cfu/thigh compared to each monotherapy (p = 0.0044-0.0048).

Conclusions: The newly reported Phikzvirus jumbo phage AttikonH10 demonstrated a broad host range, strong lytic activity and synergistic effects with amikacin against MBL-PA isolates making it a candidate for phage therapy.

Keywords: Carbapenemase producing P. aeruginosa; Jumbo phages; Lytic activity; Metallo β-lactamases.

Fig. 1

Plaque assay of a lytic jumbo phage (vB_PaerM_AttikonH10) on a susceptible P. aeruginosa isolate (AUHB77) (Α) and its microbiological characteristics, including adsorption rate, adsorption time, latent phase and burst size calculated using one-step phage growth curve (B). The data shown represent the mean values from two independent replicates, and the error bars indicate the standard deviations

Fig. 2

Growth curves of jumbo bacteriophage vB_PaerM_AttikonH10 against host isolate. The experiment was performed in triplicate and bars indicate standard deviations. Horizontal dotted line represents the background optical density at 600 nm (OD₆₀₀)

Fig. 3

A VIRIDIC-generated heatmap showing pairwise intergenomic distances between Pseudomonas phage vB_PaerM_AttikonH10 and other recognized or tentative Phikzvirus phage genus representatives that have complete genomes elucidated (n = 1 + 27). Blue rectangles indicate values associated with the studied phage. Β VIRIDIC-calculated complete genome nucleotide sequence intergenomic distance Neighbour-Joining (NJ) tree. The NJ tree was drawn from a pairwise intergenomic distance matrix generated by VIRIDIC for a dataset comprising the studied Pseudomonas phage vB_PaerM_AttikonH10 and other recognized or tentative Phikzvirus phage genus representatives that have complete genomes elucidated (n = 1 + 27). The tree is unrooted and drawn to scale; the scale bar represents genome nucleotide sequence divergence percentages. Tip labels follow the format of “Genome accession|Phage”. Tip labels of bacteriophages from RefSeq are colored red. Tip label of phage vB_PaerM_AttikonH10 is in blue"

Fig. 4

Α Genome lengths and GC% of Phikzviruses. Y axis represents GC% content, X axis—genome length (bp). Points represent individual phage complete genomes. Studied Pseudomonas phage vB_PaerM_AttikonH10 and other phages of interest are colored according to the legend. B Major Capsid Protein (MCP; top) and Terminase Large subunit (TerL; bottom) phylogenetic trees with the respective aa sequences from vB_PaerM_AttikonH10 within the context of nine highest-scoring hits from other viruses found in RefSeq. The NJ trees from MCP and TerL amino acid sequence multiple sequence alignments representing the datasets comprising the MCP of the studied Pseudomonas phage vB_PaerM_AttikonH10 and other Pseudomonas-infecting viruses from RefSeq encoding homologous MCP and TerL (n = 1 + 9 in both cases, respectively) are shown. The trees are midpoint rooted. The trees are drawn to scale with the scale bar representing amino acid substitutions per site. There was a total of 734 and 716 positions in the final MSAs used for generating the MCP and TerL trees, respectively. Branch bootstrap support percentages are shown only for branches supported by at least 80% replicates (out of 1000 replicates). Tip labels follow the format of “Protein accession|Phage|Genus”. Clade comprising Pseudomonas phage vB_PaerM_AttikonH10 and other Phikzviruses, including their most-recent common ancestor node, is colored in blue. Tip label of phage vB_PaerM_AttikonH10 is highlighted in blue

Fig. 5

Genome organization and proteome content comparison of Pseudomonas phage vB_PaerM_AttikonH10 to its closest relative—phage KTN4, as well as two taxonomically recognized Phikzvirus representatives (SL2 and phiKZ). Genomes are drawn to scale; the scale bar indicates 5000 base pairs. The pseudocircular assembly of the phage vB_PaerM_AttikonH10 genome was artificially opened at the start of TerL gene during the auto-annotation. Genome representations of other phages in the comparison were re-organized as indicated to ensure collinearity with the vB_PaerM_AttikonH10. Arrows representing open reading frames point in the direction of the transcription and are color-coded based on the function of their putative product according to the legend. vB_PaerM_AttikonH10 genome was auto-annotated using the Phanotate and Pharokka combination (original annotations from downloaded GenBank files were retained for other genomes). Ribbons connect phage proteins sharing > 70% amino acid sequence similarity and are colored in lighter shades according to the predicted functional group of the respective phage vB_PaerM_AttikonH10 ORF products (see legend for details)

Fig. 6

Pharmacokinetics of vB_PaerM_AttikonH10 in vivo via IP route. The phage titre in log₁₀ PFU/mL of tissue after 1, 4, 8 and 24 h of phage administration at a dose of 10⁹ PFU/mouse (A) and 10⁶ PFU/mouse (B). The bars indicate the mean (± SD) from two independent animals

Fig. 7

In vivo activity of jumbo phage alone and in combination with amikacin in a thigh infection mouse model. Error bars indicate standard deviations. Bacterial log₁₀ cfu/thigh change after 24 h of treatment compared to bacterial load before start of treatment. Significant p values of one-way ANOVA followed by post-test are shown