Phage-specific immunity impairs efficacy of bacteriophage targeting Vancomycin Resistant Enterococcus in a murine model

Abstract

Bacteriophage therapy is a promising approach to address antimicrobial infections though questions remain regarding the impact of the immune response on clinical effectiveness. Here, we develop a mouse model to assess phage treatment using a cocktail of five phages from the Myoviridae and Siphoviridae families that target Vancomycin-Resistant Enterococcus gut colonization. Phage treatment significantly reduces fecal bacterial loads of Vancomycin-Resistant Enterococcus. We also characterize immune responses elicited following administration of the phage cocktail. While minimal innate responses are observed after phage administration, two rounds of treatment induces phage-specific neutralizing antibodies and accelerate phage clearance from tissues. Interestingly, the myophages in our cocktail induce a more robust neutralizing antibody response than the siphophages. This anti-phage immunity reduces the effectiveness of the phage cocktail in our murine model. Collectively, this study shows phage-specific immune responses may be an important consideration in the development of phage cocktails for therapeutic use.

Fig. 1. Phage cocktail reduces burden of intestinal VRE.

Five lytic bacteriophages were chosen for use in a therapeutic cocktail. A Spot titer of bacteriophage lysates on a VRE bacterial lawn. B Transmission electron microscopy of each of the five phages used in cocktail. Line displayed represents 100 nm. C Schematic of mouse model of phage therapy targeting colonized VRE. D C57Bl/6 J mice were given antibiotics for 1 week prior to inoculation with 10⁸ CFU of VRE via oral gavage. Mice were treated with daily intraperitoneal injections (IP) with either a five-phage cocktail (ɸCT) (10⁹ PFU/phage) or vehicle control for 7 days. Stool was plated to quantify VRE in colony-forming units (CFU) per gram shown with geometric mean ± standard deviation. Data are representative of three experiments with n = 5 per group. Fischer’s mixed-effects uncorrected LSD test was used between groups. Significance indicated by *p < 0.01, **p < 0.001. Source data are provided as a Source Data file.

Fig. 2. Antibody production in response to administration of bacteriophage cocktail.

Mice were treated with either ɸCT or vehicle control for 7 days, allowed to rest for 1 month, and then treated with the same phage regimen a second time. Serum was collected 1 day after completion of the secondary course (day 35). A IgG anti-phage antibodies were detected with ELISA. Mean ± SD with sigmoidal curve-fit is shown. B Serum titer was calculated using IC50 from sigmoidal curves in A. One-way ANOVA with Holm-Šídák correction was performed. C IgM anti-phage antibodies (D) Serum (diluted 1:10) from mice was incubated with purified phages (10⁹ PFU/mL) to test for neutralization of phage killing activity. Efficiency of plating (EOP) was determined by dividing titers from phage incubated with serum by phage incubated with SM buffer alone. EOP limit of detection (LOD) is shown which was calculated by dividing titers of LOD and SM buffer alone. Two-way ANOVA with Holm–Šídák correction used. E Schematic of mouse model. F Mice were treated with a primary course of either phage cocktail or vehicle. After 4-weeks, all mice were administered one ɸCT dose. Spleens were harvested 24 h after single injection and tested for presence of phage using the spot titer method. Data are representative of three experiments with n = 5 per group. For (F), Mann–Whitney test was used to compare groups. Medians with interquartile ranges are shown. Significance indicated by *p < 0.01 **p < 0.001 ***p < 0.0001. Source data are provided as a Source Data file.

Fig. 3. Immunity generated against individual siphophage and myophage administration.

Mice were treated for 7 days with daily IP injections of 10⁹ PFU of either phage ɸ45, phage ɸ53, or vehicle control then 4 weeks later administered the same 7-day phage course. Serum was collected 1 day after completion of the secondary course (day 35). Phage-specific antibodies were measured using serum from mice treated with (A) ɸ45 or (B) ɸ53 with ELISA. Mean ± SD with sigmoidal curve-fit is displayed. C Serum from these mice were then used to test neutralizing antibody function against all five phages. Two-way ANOVA was performed with Tukey correction. D Mice were treated with a 7-day course of phages ɸ45 or ɸ 53 or vehicle control. After 4-weeks, all mice were given a single dose of phage and 24 h after injection, spleens were harvested and enumerated for lytic phage. Data are representative of two experiments with n = 5 per group. For (D), a Mann–Whitney test was used to compare groups and displayed as geometric mean with 95% confidence interval Significance indicated by *p < 0.01 **p < 0.001. Source data are provided as a Source Data file.

Fig. 4. Identification of the phage protein targets of anti-phage antibodies.

A TEM images of IgG immunogold-labeled phages incubated with IgG from mice treated with ɸCT or vehicle control. Arrows refer to gold staining attachment points. B Detection of phage proteins targeted by antibodies using serum from mice treated with ɸCT, ɸ45, or vehicle control to probe whole phage proteins. L = Ladder. C Coomassie-stained SDS-page gel loaded with purified ɸ45 and ɸ53 with protein bands targeted by anti-phage antibodies numbered. Table 2 identifies number-labeled protein bands. Data are representative of two replicates with serum or IgG from combined n = 5 per group. L = Ladder. Source data are provided in Supplementary Fig. 6.

Fig. 5. Anti-phage immunity reduces effectiveness of phage treatment.

A Schematic of murine model of phage treatment targeting VRE colonization after preexposure to phage to induce anti-phage immunity. Mice were given either a course of phage cocktail or vehicle control. Two weeks later mice were administered vancomycin in drinking water. After 7 days on antibiotic water, mice were inoculated with VRE by oral gavage. Seven days after VRE exposure, mice were then treated for 7 days with either ɸCT or vehicle control administered by IP injection. Stool was collected from mice and assessed for presence of (B) VRE and (C) phage throughout the course of the study. D–G Tissues were collected and plated to enumerate lytic phage at day 13.SI = small intestine LI = large intestine. Serum was then used to probe anti-phage antibody production (H–L) with (M) IgG titers for double cocktail treatment shown calculated from IC50 from H–L. N Serum was also collected at day 13 and tested for phage neutralizing activity. Parametric data is shown as geometric mean with 95% confidence interval (5 M), while non-parametric data is displayed as median with interquartile range (5B-L, 5 N). Data are representative of two (D–G) or three (B, C, H–L) experiments with n = 5 per group. For (B, C), a mixed-effects analysis was used to compare the three groups with repeated measures. To compare groups in (D–G), Mann–Whitney test was used and shown as geometric mean with 95% confidence interval. Statistical tests for (M) include two-way ANOVA with Holm–Šídák multiple comparisons test. Asterix (*) refers to statistical comparisons between groups represented by purple circles and orange squares. Dollar sign ($) compares between green triangles and orange squares and hashtag (#) compares purple circles and green triangles. To indicate significance, one symbol p < 0.01, two symbols p < 0.001, three symbols p < 0.0001. Source data are provided as a Source Data file.