Phage Therapy Related Microbial Succession Associated with Successful Clinical Outcome for a Recurrent Urinary Tract Infection

Abstract

We rationally designed a bacteriophage cocktail to treat a 56-year-old male liver transplant patient with complex, recurrent prostate and urinary tract infections caused by an extended-spectrum beta-lactamase (ESBL)-producing Escherichia coli (E. coli) (UCS1). We screened our library for phages that killed UCS1, with four promising candidates chosen for their virulence, mucolytic properties, and ability to reduce bacterial resistance. The patient received 2 weeks of intravenous phage cocktail with concomitant ertapenem for 6 weeks. Weekly serum and urine samples were collected to track the patient's response. The patient tolerated the phage therapy without any adverse events with symptom resolution. The neutralization of the phage activity occurred with sera collected 1 to 4 weeks after the first phage treatment. This was consistent with immunoassays that detected the upregulation of immune stimulatory analytes. The patient developed asymptomatic recurrent bacteriuria 6 and 11 weeks following the end of phage therapy-a condition that did not require antibiotic treatment. The bacteriuria was caused by a sister strain of E. coli (UCS1.1) that remained susceptible to the original phage cocktail and possessed putative mutations in the proteins involved in adhesion and invasion compared to UCS1. This study highlights the utility of rationally designed phage cocktails with antibiotics at controlling E. coli infection and suggests that microbial succession, without complete eradication, may produce desirable clinical outcomes.

Keywords: UTI; comparative genomics; humoral response; microbial pathogenesis; phage therapy.

Figure 1

Characterization of UCS1 Cocktail. (A) Serial dilution of UCS1 Cocktail onto lawn of UCS1 E. coli. Each tick mark represents 1 mm. (B) Killing Assay. 10⁶ CFUs of UCS1 E. coli were incubated without phage or indicated MOI of UCS1 cocktail for 4 h in 100 µL LB, and resulting CFUs counted. Error bars represent mean +/− SD n = 3 technical replicates **** p < 0.001. (C) Synogram performed as described in Gu Liu et al. 2020 [39]. Each well of a 96-well plate was inoculated with 5 × 10⁸ CFU of UCS1 and subjected to increasing concentrations of drug, phage cocktail, or both in LB and incubated/shaken at 37 °C for 24 h. Optical Density 600 nM measurements were taken every 30 min. Values are reported as % reduction in growth at 24 h compared to untreated controls.

Figure 2

Treatment and clinical sampling course for patient with ESBL-producing E. coli UTI. (Top) Patient overview, administration route (PICC line, orange), and treatments (UCS1 cocktail, green; ertapenem, gray). (Bottom) Time course for dosing regimen of UCS1 cocktail (green), ertapenem (gray), and clinical sampling of urine (yellow drops) and blood (red drops). Created with BioRender.com accessed 23 September 2021.

Figure 3

Patient humoral response to UCS1 Cocktail. (A) Equal volumes of the UCS1 cocktail mimic were mixed with phage buffer alone (No Serum), weekly patient serum samples (Neat), or dilutions in phage buffer and incubated/shaken at 37 C for 30 min. Serial dilutions from each condition were spotted onto lawns of UCS1 E. coli to determine titer. Results are reported as average +/− SD. n = 3 technical replicates. (B) Luminex assays performed on weekly sera samples. Concentrations (pg/mL) of humoral markers were measured in duplicate, averaged, and standardized to pretreatment levels before being converted to Log2 values. Heatmap and hierarchal clustering were done using HeatMapper with average linkage clustering method and Euclidean distance measurements.

Figure 4

UCS1 and UCS1.1 are nearly identical ST131 strains. (A) Purified phages HP3, HP3.1, ES17, and ES19 were tittered onto lawns of UCS1 and UCS1.1. Efficiency of Plating (EOP) was calculated using UCS1 as the reference strain (EOP = 1). n = 3 technical replicates. (B) Phylogenetic tree was generated using autoMLST and visualized using Geneious Prime. UCS1 and UCS1.1 are shown in the context of other ST131 strains from the three major subtypes (fimH22, fimH30, and fimH41) and strains from other relevant UPEC sequence types: CFT073 (ST73) and UTI89 (ST95). Nodes are annotated with Ultrafast Bootstrap values. (C) Circular genome visualization created using BLAST Ring Image Generator (BRIG). EC958 acted as reference genome. Rings from inner to outer are as follows: (1) GC Content, (2) GC skew, (3) E. coli EC958 (Red), (4) E. coli UCS1 (Blue), (5) E. coli UCS1.1, and (6) CDS features (from EC958).