The rise, fall, and resurgence of phage therapy for urinary tract infection

ABSTRACT

In the face of rising antimicrobial resistance, bacteriophage therapy, also known as phage therapy, is seeing a resurgence as a potential treatment for bacterial infections including urinary tract infection (UTI). Primarily caused by uropathogenic Escherichia coli, the 400 million UTI cases annually are major global healthcare burdens and a primary cause of antibiotic prescriptions in the outpatient setting. Phage therapy has several potential advantages over antibiotics including the ability to disrupt bacterial biofilms and synergize with antimicrobial treatments with minimal side effects or impacts on the microbiota. Phage therapy for UTI treatment has shown generally favorable results in recent animal models and human case reports. Ongoing clinical trials seek to understand the efficacy of phage therapy in individuals with asymptomatic bacteriuria and uncomplicated cystitis. A possible challenge for phage therapy is the development of phage resistance in bacteria during treatment. While resistance frequently develops in vitro and in vivo, resistance can come with negative consequences for the bacteria, leaving them susceptible to antibiotics and other environmental conditions and reducing their overall virulence. “Steering” bacteria toward phage resistance outcomes that leave them less fit or virulent is especially useful in the context of UTI where poorly adherent or slow-growing bacteria are likely to be flushed from the system. In this article, we describe the history of phage therapy in treating UTI and its current resurgence, the state of its clinical use, and an outlook on how well-designed phage therapy could be used to “steer” bacteria toward less virulent and antimicrobial-susceptible states.

INTRODUCTION

Urinary tract infection (UTI) is the second most common bacterial infection worldwide with an estimated 400 million individuals experiencing a UTI in a given year (1, 2). These infections are a major medical burden, with an annual cost of about $3.5 billion in the United States alone (3, 4), and a cumulative lifetime risk of at least one UTI in 50% of women and 20% of men (5). While multiple Gram-negative and Gram-positive pathogens can cause UTI, the vast majority of cases are from the order Enterobacterales, with uropathogenic Escherichia coli (UPEC) alone accounting for 60%–80% of all infections across age groups (3, 6, 7).

Due to the number of cases, UTI treatment comprises a large portion of all outpatient antibiotic prescriptions (6, 8). While antibiotic treatment is still currently an effective UTI therapy for most patients, antibiotics are unable to eliminate intracellular reservoirs of bacteria that form early during infection. These reservoirs are thought to be the seed for later UTI recurrence, a phenomenon observed in approximately 30% of women within 6 months of their initial UTI (9, 10). In addition, antibiotics are often ineffective at treating infections associated with biofilm formation, as is commonly observed in catheter-associated UTI (CAUTI) and with kidney stones (11–14). Furthermore, antibiotics are poor at treating chronic infections of the prostate (15, 16). Issues with treatment inefficacy are further complicated by rising rates of antibiotic resistance among urinary pathogens. A recent study concluded that, in 2019, approximately 22,000 deaths were caused in the United States directly attributable to urinary tract infection (2, 17). These numbers are expected to increase further as urinary pathogens become more antibiotic resistant. Because of poor efficacy, antibiotic resistance, and detrimental effects of antibiotic usage on the healthy microbiome, novel non-antibiotic treatment strategies for UTI are needed.

Bacteriophages (phages), as natural predators of bacteria, are potential therapeutics. Several attributes of phage make them appealing candidates including their specificity, safety, ubiquity, self-dosing (replicating) capabilities, and potential to act synergistically with antibiotics. Since their initial discovery in the early 20th century, phages have been used for the treatment of UTI. Early studies showed some success; however, the use of bacteriophages as therapeutics rapidly dwindled in the West following the isolation and large-scale production of antibiotics in the 1940s. There were multiple reasons behind this decline, but one reason is that bacteria can rapidly develop resistance to phage under selection pressure. While this could be considered a treatment failure, it is now appreciated that mutations conferring bacterial survival from phage-mediated killing may hinder their overall fitness or virulence (18–20).

In this review, we will cover the history of phage therapy for UTI, its resurgence in the face of the rising antimicrobial resistance threat, and its current standing as a therapeutic intervention with particular focus on UTI caused by Enterobacteriaceae. We will also discuss ongoing clinical trials and how bacterial resistance to phage could be exploited for therapeutic purposes.

THE HISTORY OF PHAGE THERAPY FOR UTI

Soon after their identification in 1917 by Félix d’Hérelle, phages were explored as therapeutics for many infections including UTI. As early as 1925, patients with UTI caused by various pathogens were administered bacteriophages with varying levels of success as reviewed recently (21). Post hoc analysis of these initial studies is significantly hindered by the lack of control groups as no effective standard of care for UTI existed at that time. This limitation, coupled with poor information on dosages and composition of phage cocktails, confounds interpretation of these results. Despite this, early studies often frequently reported effective treatment, evidenced by clinical improvement or microbiological eradication, in more than 50% of patients (21–25).

Following the widespread production of antibiotics starting in the 1940s, the use of bacteriophages fell out of favor. The reasons for this are multifaceted and include scientific, socio-scientific, and personal reasons, which have been extensively covered by others (21). By the 1950s, phage therapy had largely been relegated to, at most, a supplement to antibiotic treatment in Western medicine. Bacteriophage research for the treatment of UTI continued in many Central and Eastern European and Asian countries in the following decades (26, 27). Therapeutic success was documented during this time. However, the absence of large-scale and controlled clinical trials and a lack of control in the composition and production of phage products complicate the interpretation of clinical results. Despite this, phage therapy has continued to be developed in many of these countries and is routinely used and commercialized, including in Georgia, where the Eliava Institute produces off-the-shelf and custom phage cocktails for treating many diseases including UTI (28, 29). Today, with the advent of more standards and technology around production and nucleic acid sequencing, it is easier to characterize and control phage activity and production, thereby paving the way for not only personalized use but for controlled clinical trials.

BENEFITS OF PHAGE AS A THERAPEUTIC OPTION

Bacteriophages provide multiple therapeutic advantages including that they are generally regarded as safe and can be combined into phage cocktails to expand target range and prevent resistance (30). Phages also have several features that make them particularly suited to treat UTI (Fig. 1). A few of these are discussed below.

Fig 1.

Summary of potential benefits of phage therapy in the treatment of urinary tract infection.

Ability to treat antibiotic-resistant infections

Perhaps, the greatest benefit of bacteriophages is their ability to target pathogens that have become resistant to antibiotics. A recent study estimated that there were approximately 4.95 million deaths associated with antimicrobial resistance (AMR) in the year 2019, of which, approximately 300,000 were associated with UTI (31). These numbers are expected to continue rising with the United Nations estimating that there will be 10 million deaths attributed to AMR infections by 2050 (32). While new antibiotics are needed to combat these infections, the estimated cost of $1.5 billion to develop a new antibiotic (33), coupled with the short window of time before antibiotic resistance develops, has dissuaded many pharmaceutical companies from pursuing this work. While currently unproven at scale, the genetic diversity of bacteriophages, allowing for multiple bacterial features to be targeted, and their global distribution provides an economically feasible discovery solution.

Bacteria possess several mechanisms for evading antibiotic killing (34), including limiting antibiotic uptake, pumping antibiotics out of the cell, degrading them once inside, and modulating the target of the drug. While similar mechanisms may prevent phage efficacy, there is likely to be little overlap between the mutations made to evade antibiotics and those used to evade phage.

Phage-antibiotic synergy

Compassionate use of phage therapy often occurs during concurrent antibiotic treatment which prompts the question of whether either of these two therapies directly influence the effectiveness of the other. Gu Lui et al. (35) observed that phage and antibiotics could act synergistically or antagonistically depending on the phage receptor and antibiotic mechanism of action. Parab et al. (36) and Valério et al. (37) have demonstrated that phage coupled with several different classes of antibiotics can delay the development of bacteriophage resistance. Valério et al. specifically found phages coupled with subinhibitory concentrations of bactericidal antibiotics were more potent than either individually and could limit resistance from arising. However, this study also showed that bacteriostatic antibiotics given alongside phage had no additive effect. This may be because the bacteriostatic antibiotics used interfered with DNA replication and therefore phage replication. While most studies have focused on using antibiotics and phage concurrently, some have looked at how alternating treatments may be therapeutically beneficial. Hao et al. (38) showed that resistance to colistin changed the bacterial surface charge, leading to increased phage adherence. Thus, it is reasonable to conclude that a combination of antibiotic and phage, if acting in different pathways, may not only synergize but also decrease the likelihood of resistance developing.

Biofilm disruption and prevention

In patients with indwelling urinary catheters, bacteria can form biofilms on the catheter surface resulting in catheter-associated UTI (CAUTI). Although less common than acute UTI, CAUTI is among the most commonly reported hospital-acquired infections and results in increased hospitalization duration and health care costs as well as contributes to significant morbidity and mortality (39–41). Biofilms in CAUTI are notoriously difficult to treat with antibiotics, often requiring removal of the catheter for treatment (42). Extracellular matrix produced by the biofilm prevents penetration of some antibiotics (43–45). Antibiotics that are able to penetrate the biofilm, particularly classes targeting replication or protein synthesis, are often ineffective due to the reduced metabolic activity of bacteria within the biofilm (46–48). In contrast, some bacteriophages possess factors, including cell wall degrading polysaccharide depolymerases, that facilitate bacterial killing under the physiologic conditions of biofilms (49–51). Phage eradication of biofilms has been extensively documented in vitro, with some phage having the capacity to both disrupt established biofilms (51–58) and prevent biofilm formation on catheters (53, 55, 59–62).

POTENTIAL CHALLENGES OF PHAGE THERAPEUTICS FOR UTI

While phage therapy holds much promise for the treatment of UTI, several of the factors that differentiate it from antibiotics can also complicate its clinical use. For example, while the specificity of a phage for a given bacteria is often cited as a benefit over antibiotic therapy, this feature can limit the ability of phage to be used as “off-the-shelf” therapeutics and could mean that large phage libraries are required to make custom therapeutics. This specificity is also threatened if pathogenic strains gradually drift over time, requiring new phage therapeutics to be developed. Furthermore, this specificity also selects for the development of phage resistance. While this can result in less virulent bacteria (63), it can also lead to treatment failure in the absence of bacterial clearance. In facing these challenges, more work is needed in developing phage therapeutics with wide host ranges that are difficult for bacteria to evade. Bacteriophage therapy also holds promise as being more cost effective and having a more rapid development process than the discovery and development of other antimicrobials. While the ubiquity of phages surely allows for rapid discovery, the scaling up and commercial production of phage therapeutics remains a time- and cost-intensive process when compared with large-scale antibiotic production. To combat these challenges, research into large-scale phage production and purification processes will be essential (64, 65).

MODERN USES OF PHAGE THERAPY FOR UTI IN VIVO

Phage therapy for UTI has been used since shortly after the discovery of bacteriophages. Highlighted below are some of the more recent usages of phage therapy for UTI, ranging from animal models to case reports and clinical trials in humans.

Animals

Existing animal models of UTI have proven useful for studying phage therapy in vivo and have provided insight into dosing and delivery methods for human phage therapy. Bhargava et al. (66) recently reported complete clearance of E. coli UTI in rats using a three-phage cocktail. In this study, groups treated with doses of 2 × 10⁷ and 2 × 10⁶ PFU cleared the infection with just one dose, while the lowest phage dose group of 2 × 10⁵ required additional administrations after bacteriuria returned. Similarly, Mijbel Ali et al. (67) reported that a cocktail of 25 E. coli phages could resolve infection in a chronic UTI murine model. In this study, phages were administered 10 days after bladder infection, leading to bacterial clearance from the bladder and kidneys just 24 hours after administration. This result occurred whether phages were given intraperitoneally (i.p.) or directly into the bladder.

Dufour et al. (68) utilized a bacteriophage specific for O25b-expressing E. coli to treat several infection models, including UTI. Twenty-four hours following transurethral inoculation of mice, O25b-specific bacteriophage, LM33_P1, was given i.p. Twenty-four hours post-phage treatment, bacterial burdens in the kidneys were significantly decreased in the phage-treated group; however, bacterial burdens in the bladder were not provided. Similarly, Tóthová et al. (69) gave a cocktail of two phages i.p. to mice with Cronobacter UTIs immediately following infection. Twenty-four hours later, these mice had reduced bacterial burdens in the kidneys, but bladder colonization remained similar in the treated and untreated groups. Interestingly, decreases in TNFα and MCP-1 were seen in the bladders of phage-treated mice, suggesting that the immune system was somewhat dampened. It is unknown if the dampened immune response impacted bladder clearance in the phage-treated group.

Phages have also been shown to protect mice from lethal UTI challenge. Nishikawa et al. (70) showed that phages T4 and KEP10 given i.p. immediately after lethal ECU15 E. coli infection improved survival. This response was dose dependent with higher phage doses providing increased protection; the survival rates for KEP10- and T4-treated animals given the highest phage dose, 3 × 10¹¹ PFU, were 90% and 100%, respectively, at 1 week post-infection. By contrast, all untreated animals died within the 3 days following infection. Even at the lowest phage dose, 5 × 10⁷ PFU, one phage for every one hundred bacteria, 40% of animals survived at least a week post-infection. Importantly, the researchers also mention that no negative effects of phage were seen in uninfected mice given high doses of phage.

To translate this work into widespread human use, more work is needed in the area of phage therapy for acute UTI, which makes up the vast majority of infections. Phage therapy is also not well studied in murine models of populations who frequently suffer from UTI, including elderly individuals, those with diabetes, organ transplant recipients, and those with spinal cord injuries. Additionally, little work has been done to see what effect phage therapy has on the intracellular reservoir of E. coli that can develop during infection, a mechanism which bacteria use to evade antibiotics and cause recurrent infection.

Human case reports

Several case reports on the use of bacteriophages for urinary infections have been published in recent years highlighting successes and struggles in treating chronic infections (71), infections following organ transplants (72–75), and infections on implanted urinary devices (76, 77). In addition, these case reports have shown success in the absence of microbiological cure (78, 79) as well as highlighted how phage therapy can be complicated by the development of phage resistance (28, 80, 81).

Phage therapy following organ transplant

Phage therapy has been used in cases of chronic and recurrent UTI following organ transplant. Kuipers et al. (72) presented a case report of a 58-year-old man who developed recurrent UTI with extended spectrum beta-lactamase-positive (ESBL+) Klebsiella pneumoniae shortly after kidney transplant. For over half a year, this patient was repeatedly treated with antibiotics, only for the infection to reappear after treatment cessation. During a 12-week period, the patient was given Klebsiella phages provided by Eliava Institute alongside meropenem, an antibiotic which had previously repeatedly failed to treat the infection, resulting in successful pathogen clearance. Similarly, Rostkowska et al. (73) reported on a kidney transplant patient who had several UTIs requiring hospitalization in the 15 months following transplantation. In this case, despite 29 days of phage therapy, the patient failed to clear the infection, leading to nephrectomy being performed after the conclusion of therapy. Aslam et al. (74) used a cocktail of four phages to treat a 56-year-old male patient who had a history of recurrent ESBL+ E. coli UTI who had previously undergone liver transplantation. A phage cocktail was provided intravenously twice daily for 2 weeks along with ertapenem. For at least 12 weeks following the end of treatment, the patient lacked symptomatic UTI, although urine culture remained positive.

In very rare cases, phage therapy has been used as a standalone treatment for UTI. Le et al. (75) reported a case of recurrent UTI caused by K. pneumoniae in a 70-year-old woman who had recently received a kidney transplant. These recurrent infections were caused by ESBL-producing bacteria and required multiple hospitalizations and frequent antibiotic usage. A cocktail of three phages was given to the patient twice daily via IV for 28 days without antibiotics. Despite four symptomatic K. pneumoniae UTIs in the 6 months before phage therapy, the patient experienced no UTIs in the 6 months following treatment. Furthermore, although the patient developed UTI due to K. pneumoniae approximately 200 days after phage therapy, the bacterial strains identified were all non-ESBL producers and were able to be treated with oral antibiotics.

Phage therapy for implanted urinary device infections

Khawaldeh et al. (76) reported a 67-year-old woman who, following adenocarcinoma treatment, received ureteric stents that were complicated by a symptomatic Pseudomonas aeruginosa infection. The patient received multiple courses of antibiotics and had the stents replaced over 2 years with no symptom improvement. The treatment plan was then adapted to include intravesical infusions of the Pyophage cocktail from Eliava Institute every 12 hours for 10 days combined with colistin use concurrently starting at day 5 and alone from day 10 to day 30. Bacterial cultures taken during treatment revealed that phage resulted in only a minor decrease in bacterial burdens. Despite this, following a 30-day treatment with multiple antibiotics which had previously failed to work against the pathogen, this patient was cleared of infection with urine cultures remaining sterile for more than 6 months following the completion of antibiotic therapy. Similar positive results were also shown in a 57-year-old patient with recurrent urinary tract infections caused by K. pneumoniae after undergoing nephrectomy as well as placement of a ureteral stent which was required to support the remaining kidney (77). Following an episode of sepsis for which antibiotics were successful, the ureteral stent remained colonized by K. pneumoniae. The patient received a custom-made phage therapy cocktail from Eliava Phage Therapy Center which was administered orally and intra-rectally for 3 weeks after antibiotics failed to decolonize the ureteral stent. Fifteen days after beginning phage therapy, the ureteral stent was again replaced, but this time, it was not colonized by K. pneumoniae, something that had not previously been possible with antibiotics alone. Remarkably, since the conclusion of therapy, multidrug-resistant K. pneumoniae have not been identified from the patient’s stool, urine, or new ureteral stent.

Phage therapy success in the absence of pathogen clearance

Several studies have shown apparent symptom relief following phage therapy in the absence of a clinical cure. Terwilliger et al. (78) presented a case of a patient with recurrent prostatitis with ESBL+ E. coli from 2017 to 2020. They designed a phage cocktail that included phages capable of locating the mucosal surface and others that would prevent resistance of another phage in the cocktail. The patient was given IV phage cocktail every 12 hours for 2 weeks concurrent with intravenous ertapenem. Rapidly after treatment initiation, no bacteria were detected in the urine and symptoms resolved. At 8 and 13 weeks, the patient developed asymptomatic bacteriuria; however, no antibiotics were required. Asymptomatic strains isolated at 8 and 13 weeks had many single-nucleotide polymorphisms, insertions, and deletions in genes involved in adhesion and invasion compared with the parental strains. Similar results have recently been reported by Green et al. (79) showing a failure of UTI clearance following antibiotic and phage therapy; however, the patient became symptom free, an overall favorable clinical outcome.

Bacterial phage resistance complicating treatment

The issue of phage resistance and bacterial persistence has arisen in several UTI case reports (28). In one instance, a 72-year-old woman with recurrent K. pneumoniae UTI was given a several month-long course of phage administered orally and through a vaginal suppository. Initial treatment proved unsuccessful, possibly due to prior resistance to the phage cocktail. The patient then reached out for a new custom phage preparation which was also ineffective despite showing sensitivity to the phage cocktail in vitro. While the exact causes for this treatment failure is unknown, it is possible the phage failed to reach the site of infection or the phage concentrations (10⁶–10⁸ PFU/mL) were simply too low to provide effective therapy, especially in the bladder environment where phages may be rapidly lost to urine voiding. In another case, a 66-year-old man suffering from nearly 15 years of chronic, multidrug-resistant K. pneumoniae infection, was given phage therapy (81). After the initial 2-week treatment with a single phage, the patient remained infected, with several recovered isolates now demonstrating phage resistance. To combat this, a second phage was added to the treatment and another 2 weeks of treatment was given via bladder irrigation. Despite symptom relief and negative urine cultures during treatment, shortly after the treatment ended, the patient’s infection returned. The investigators hypothesized that the bladder irrigation treatments failed to reach the infected renal pelvis and that following treatment, this bacterial community reseeded infection. By combining bladder irrigation with a percutaneous nephrostomy, investigators were able to deliver a new four-phage cocktail into both the bladder and kidney. This treatment, which was given alongside antibiotic therapy for the first week, resulted in clearance of the multidrug-resistant infection and improvement in the appearance of the bladder mucosa. Multiple phage cocktails were required in the case of a 63-year-old woman with recurrent K. pneumoniae UTI causing pyelonephritis and shock (80). This woman was given a cocktail of five phages for 5 days via bladder irrigation but, despite apparent symptom relief, the patient continued to experience low bacteriuria following therapy. These bacteria were identified as resistant against all phages tested. The investigators administered a second round of phage therapy consisting of five new phages, but resistance arose once again. In crafting their third cocktail, they added a non-active antibiotic along with a new set of phages in hopes that this would delay resistance. Unlike previous cocktails, the combination of antibiotics and phage prevented the emergence of phage resistance, and no recurrence was observed during 6 months of follow-up.

Importantly, in all human case reports discussed above, there were no reported side effects attributable to phage, an important consideration for large-scale phage therapy usage.

Human clinical trials

While there are multiple case reports of phage therapy being used to treat UTIs in humans, there are comparatively few clinical trials on the subject (Table 1). One recent cohort study (82) included 118 men with planned prostate resection who were screened for asymptomatic bacteriuria prior to the procedure. Multiple bacterial species, including Staphylococcus aureus, E. coli, Enterococcus spp., Streptococcus spp., P. aeruginosa, and Proteus mirabilis, were found during this screening. Of the 118 patients who were screened and had bacteriuria, 41% of their bacterial isolates was susceptible to the commercially available Pyophage preparation made by Eliava BioPreparations. This group was able to adapt their phage cocktail through serial passage to generate a phage cocktail capable of infecting 75% of the patient isolates. Nine of the patients with phage-sensitive bacteria (E. coli, Streptococcus spp., Enterococcus spp., or P. aeruginosa bacteriuria) were chosen to receive the new cocktail twice daily for 1 week. In six of the nine patients treated, bacterial burdens were decreased after treatment. This small cohort study provided safety data and was used to inform a larger, double-blind randomized clinical trial (83, 84). This study assigned men scheduled for prostate resection surgery who had bacteriuria to three groups in a 1:1:1 ratio. The first group received adapted Pyophage preparation intravesically twice daily for 1 week, the second received an intravesical saline placebo control, and the third received standard-of-care antibiotics. The treatment success rates, defined as clearance of all bacteria, did not differ among the three groups with success only being achieved in 18%–35% of participants among all groups. Interestingly, adverse events were relatively common in all three groups with 41% of participants in the control-treated group experiencing adverse events. As both the phage-treated and antibiotic-treated groups were not superior to the placebo group, this study reiterates that additional randomized controlled studies on patients with uncomplicated UTIs or in patients with catheter-associated UTI still are needed.

TABLE 1.

Clinical trials of UTI phage therapy registered on ClinicalTrials.gov

Study type	Status	Pathogen	Phage	Delivery method	Results	Ref.
Double-blind RCT	Completed	Multiple	Pyophage	Intravesical	Non-inferior to standard of care, non-superior to placebo	(83, 84)
Double-blind RCT	Completed	E. coli	LBP-EC01 (Crispr-Cas3 recombinant cocktail)	Intravesical	Safe in UTI patients	(85)
Double-blind RCT	Recruiting	E. coli	LBP-EC01 (Crispr-Cas3 recombinant cocktail)	Intraurethral and IV	N/A	(86)
Interventional	Active, Not Recruiting	E. coli	HP3, HP3.1, ES19	Oral, topical, intravesical	N/A	(87)
Interventional	Recruiting	Multiple	Unspecified	Intravesical	N/A	(88)

One such study is currently ongoing for the treatment of uncomplicated UTI caused by UPEC. The investigators are looking at how multiple outcomes are affected by the inclusion of a bacteriophage cocktail during antibiotic treatment (86). In this study, approximately 550 patients with acute UTI caused by E. coli that have had a previous UTI caused by E. coli will be assigned either oral trimethoprim/sulfamethoxazole or oral trimethoprim/sulfamethoxazole combined with the phage cocktail LBP-EC01. LBP-EC01 is a CRISPR-engineered bacteriophage cocktail that carries Cas3 into the bacterial cell leading to bacterial DNA degradation. This group has previously shown that this cocktail is safe in patients with UTI (85) and that this engineering approach works in vitro and in vivo in a mouse model of Clostridium difficile infection (89). The LBP-EC01 cocktail has specifically been reported to decrease urine concentrations of E. coli 100–1,000-fold compared with placebo in patients with urinary tract colonization by E. coli (90). This ongoing study will assess the pharmacokinetics and safety profile of the phage preparation, as well as clinical and microbiological success of the treatment and frequency of UTI recurrence in the 6 months following treatment. The effects of phage therapy on patients with recurrent UTI are currently being studied in another, single participant, clinical trial (87).

While promising results have been observed in animal models and in individual patients, additional work remains before the use of phage therapy can be expanded. Particularly, more data are needed on clinical and microbiological successes via clinical trials. Although Leitner et al. (83) demonstrated results that were non-inferior to antibiotics in their study, they were also non-superior to placebo bladder washing. Furthermore, a common critique of bacteriophage therapy is that bacteria possess many mechanisms for evading killing by phage. Understanding how phage treatment fails can guide new phage therapy strategies.

Safety and bioavailability of bacteriophage in the urinary tract

Phages are found everywhere, including as a part of the human urinary microbiome (91–95); thus, highly purified bacteriophages are generally regarded as safe for use in humans (96). Stacey et al. (96) recently reviewed 13 phage therapy clinical and safety trials conducted between the years 2009 and 2021 which all supported the safety of phage therapy. These trials are supported by human studies in which bacteriophages were administered through a variety of routes without leading to bacteriophage-specific side effects (72, 97–101), even after multiple administrations (71). This contrasts with antibiotics which have recently been shown to disrupt the healthy host microbiome leading to the emergence of pathogens (102) as well as having other, wide-ranging effects on the host (103–106). Recently, Grabowski et al. (103) conducted an in-depth investigation of the effects of a two-phage cocktail on mouse immunological responses compared with the antibiotics enrofloxacin and tetracycline over a 2-week period. They found that male and female mice treated with both antibiotics had reductions in T lymphocyte numbers as well as weight loss and reductions in both pro- and anti-inflammatory cytokines compared with saline-treated control mice. In contrast, male and female mice treated with the two-phage cocktail demonstrated no differences compared with the control group. In further support of their possibility for therapeutic use, following phage therapy, phages have been found accumulating in the urinary tract of both humans and mice, regardless of phage administration route (70, 71, 76, 103, 107).

PHAGE RESISTANCE AND ITS EFFECT ON THERAPY EFFICACY AND DISEASE OUTCOME

Bacteria have developed many ways to avoid being killed by phages during their evolutionary arms race. These diverse mechanisms exist at every step of infection from adsorption to bacterial cell lysis (108–110). While phage resistance may be a natural consequence of phage therapy, the mechanisms which bacteria use to evade killing by bacteriophages may reduce their overall virulence and fitness in the host (Table 2). Exploiting these trade-offs for therapeutic potential has been termed “phage steering” (111) and may be a way in which bacteriophage therapy could ultimately be successful, even in the presence of resistance to treatment.

TABLE 2.

Fitness effects of phage resistance in urinary pathogens

Pathogen	Phage resistance mutation	Effect on bacterial fitness	Ref.
In vitro
E. coli	tolC	Increased tetracycline sensitivity	(112)
E. coli	rfaC, rfaD, rfaE, rfaF, rfaH, rfaI, rfaP, gmhA	Increased colistin sensitivity	(63, 112)
S. flexneri	ompA	Increased vancomycin sensitivity	(113)
S. flexneri	gmhA, gmhC	Increased erythromycin sensitivity	(113)
P. aeruginosa	ompM	Increased ciprofloxacin, tetracycline, ceftazidime, and erythromycin sensitivity	(114)
K. pneumoniae	Not determined	Increased amikacin, gentamycin, tobramycin, ciprofloxacin, and levofloxacin sensitivity in some strains	(115)
A. baumannii	Capsule K locus	Increased susceptibility to multiple antibiotics, reduced biofilm formation	(116)
E. coli	Not determined	Reduced biofilm formation	(57)
P. aeruginosa	Not determined	Reduced colonization of urinary catheters	(117)
E. coli	rfaE, rfaH, rfaI	Reduced growth relative to wild-type strains in urine	(63)
In vivo
E. coli	rfaH, rfaI	Reduced bladder colonization 24 hours post-infection	(63)
A. baumannii	Capsule K locus	Reduced organ colonization at 8 hours post-infection	(116)

Possibly the most studied consequence of phage resistance is the sensitization of bacteria to antibiotics which were previously ineffective or impractical. While multiple mechanisms could cause this, re-sensitization often occurs due to mutations of bacterial outer membrane proteins that are important for antibiotic efflux or mutations that leave the target of the antibiotic more accessible. One class of antibiotics, polymyxins: primarily polymyxin B and polymyxin E (colistin), are cationic lipopeptides that interact with lipid A of bacterial lipopolysaccharide (LPS) (118). This results in membrane destabilization and bacterial death. However, these antibiotics can also interact with human cell membranes resulting in off-target effects, which is why they are less frequently used clinically. Multiple groups have shown that resistance to bacteriophages can come with increased susceptibility to polymyxins, particularly when phage resistance is due to modifications in bacterial LPS (63, 112).

The effect of phage-driven susceptibility to antibiotics through other outer membrane proteins has also been described. Kortright et al. (113) have shown that resistance of Shigella flexneri, common gastrointestinal pathogen and rare uropathogen, to bacteriophage A1-1 led to variable effects on antibiotic sensitivity, with some bacteria harboring mutated OmpA becoming more sensitive to the antibiotic vancomycin; other A1-1-resistant colonies became resistant by LPS modification and displayed decreased vancomycin susceptibility as well as increased susceptibility to erythromycin. OmpA is important in the establishment of intracellular bacterial reservoirs during UPEC UTI (119) which facilitate the formation of intracellular reservoirs important for re-seeding bacterial infection and which are shielded from antibiotic therapy (120, 121). Furthermore, OmpA plays important roles in immune evasion and biofilm formation (122–124). Other outer membrane proteins that are targets of phages also play important roles in antibiotic resistance. TolC is an outer membrane protein and important part of several efflux systems that has been identified as a receptor for multiple phages (125–128). Bacteria with mutant TolC are sensitized to multiple antibiotics as well as detergents (129–134). Together, this suggests phage-driven OmpA and TolC mutations may be beneficial to the success of UTI antibiotic therapy.

Capsules are produced by many urinary pathogens and are important for immune evasion, biofilm formation, and colonization of the murine urinary tract (135–140). Capsular polysaccharides are also a common target for bacteriophages. Gordillo Altamirano et al. (116) recently demonstrated that resistance to two bacteriophages specific for the occasional urinary pathogen Acinetobacter baumannii resulted in capsule loss. This loss was caused by loss-of-function mutations in the K locus, which is responsible for production, modification, and export of capsular polysaccharides (141, 142). The resulting phage-resistant strains had several fitness defects including increased susceptibility to beta lactam antibiotics, decreased biofilm formation, and as poorer colonization of several organs, including the kidney, during an in vivo infection compared with their parental strains. Likewise, Tan et al. (115) have shown that resistance to bacteriophages can result in capsule loss in K. pneumoniae, although the exact mutations underlying these changes and the bacterial fitness consequences were not investigated.

Studies have also used phage therapy targeted specifically for the bacterial proteins important for bacterial resistance to antibiotics. Chan et al. (114) screened for phages that were specific for outer membrane protein OmpM of the MexAB and MexXY antibiotic efflux pump systems in P. aeruginosa. They found one of the 42 phages tested, OMKO1, was unable to infect the OmpM mutant strain but was able to infect the parental strain and all other mutant strains it was tested on. In multiple P. aeruginosa multidrug-resistant strains, resistance to OMKO1 resulted in increased sensitivity to the antibiotics ciprofloxacin, tetracycline, ceftazidime, and erythromycin, representing antibiotics of several classes. As the authors suggest, driving resistance that leads to increased antibiotic sensitivity could open the door to the use of some antibiotics that are not normally used due to intrinsic resistance.

Mutations in outer membrane proteins can also have effects on other virulence factors used by bacteria. α-Hemolysin (HlyA) is produced by many extraintestinal pathogenic E. coli (including UPEC) and plays multiple roles during infection including a role in host cell lysis (143, 144), persistence in the bladder (145), disruption of host inflammatory signaling pathways (144), and host cell exfoliation. This virulence mechanism allows bacteria to invade deeper bladder layers and form reservoirs capable of seeding recurrent infection (121, 146–148). Bacteria with truncated LPS have decreased secretion of HlyA and, even when HlyA is produced and excreted, it is less active than HlyA from non-LPS-truncated bacteria (149–151). TolC also plays an important role in bacterial toxin production with mutants in TolC being shown to be defective in HlyA secretion (152, 153) as well as secretion of Escherichia coli heat-stable enterotoxin Ip (154).

A decreased growth rate is another bacterial factor impacted by resistance to bacteriophages in the urinary tract. UPEC-harboring phage-driven LPS truncations were shown to grow poorly in human urine despite growth similar to the wild-type strain in bacteriologic media (63). This reduced growth rate may play a clinically important role in some pathogens, with slow-growing phage-resistant P. aeruginosa resulting in decreased biofilm formation on catheters in vitro (117).

In addition to poor growth, some phage resistance mechanisms seem to leave bacteria less able to colonize the host (155). Colonization is especially important in the context of UTI, where bladder voiding can easily remove non-adherent bacteria. Truncations of LPS to evade phage predation severely attenuate colonization of the murine bladder with loss of LPS O-antigen resulting in poor colonization of a murine UTI model (63, 156) suggesting O-antigen-targeting phages may have good therapeutic potential.

By understanding the targets of the phages used therapeutically and the likely mechanisms bacteria will utilize to evade killing, we can develop phage therapeutics that drive fitness trade-offs, resulting in virulence reduction and potential improvements in patient outcomes by reducing symptoms. Furthermore, while AMR threatens our supply of antibiotics, we can utilize phage to “steer” these bacteria back to antibiotic susceptible states, prolonging the utility of our current drugs.

FUTURE DIRECTIONS

While lytic killing of bacteria is the focus of the majority of phage therapeutics, others are studying additional attributes of bacteriophage that may provide therapeutic benefits.

Beneficial host-phage interactions

Many others have incorporated the third member of the infection, the host, into therapeutic plans for phage, with the thought that phage-host interactions could provide additional therapeutic benefits. Due to the voiding of the bladder, phage in the urine has only a small time to find and kill bacteria before it is lost. A phage that can interact with the host and be retained for a longer time could have additional time to prey on bacteria. Phage that bind to the bladder epithelium could also be useful in treating recurrent UTI caused by intracellular reservoirs of bacteria. Phages have been shown to bind, gain access into, and even traverse human cells (157, 158), with Møller-Olsen et al. (159) recently demonstrating that engineered bacteriophage K1F can interact with bladder epithelial cells, be taken up, and kill E. coli intracellularly. Moller-Olsen et al. have also shown in work using cerebral endothelial cells that K1F uptake may occur without eliciting a phage-driven immune response (160). Others have demonstrated phage interactions between epithelial surfaces and bacteriophages and have shown intracellular accumulation (157, 161–163), possibly through pinocytosis (157, 163, 164) or receptor-mediated endocytosis (165, 166). If bacteriophages are capable of binding and remaining in the bladder in the absence of infection, they could even be useful as “prophylaxis” to quickly act if UTI occurs in individuals who are at a higher risk of UTI or UTI recurrence.

Furthermore, in the context of respiratory infection, others have suggested that the interplay between the host’s immune defenses and phage lysis may be essential for infection resolution and preventing the emergence of phage resistance (167), although this phenomenon has not been well studied in the setting of UTI.

The study of the interactions between phage and host bladder epithelium is certain to be aided by the development of model systems that more closely resemble the urinary environment, including “bladder-on-a-chip” and organoid models (168–170). Similar models have been used to study phage-host interactions in other organ systems and have been useful in identifying phages capable of binding to host epithelium as well as some host components (163, 171).

Using phage to modify the microbiome

The bacteria that cause UTI frequently reside in the intestines and become pathogenic when they translocate to the urinary tract. Because of this, several groups have aimed to use phages to deplete the reservoir of these potential pathogens before they can translocate. These therapies have the added benefit of having reduced off-target microbiome effects compared with antibiotics, which indiscriminately target pathogenic and non-pathogenic bacteria and which can lead to recolonization with pathogenic bacteria (172, 173). Indeed, Galtier et al. (174) demonstrated a significant decrease in E. coli intestinal burden following just one dose of a bacteriophage cocktail in a murine E. coli colonization model. This treatment, importantly, did not have as significant of an effect on the microbiome as antibiotics. Using a simulated in vitro small intestine bioreactor system, Cieplak et al. (175) found phage cocktail was able to deplete E. coli to comparable levels attained with antibiotics while mitigating damage to other bacterial populations within their bioreactor community system. While these studies show efficacy at depleting E. coli from the intestinal tract, they do not completely clear the bacteria. It is unknown whether this is due to phage resistance developing or damage to the phage from the environmental conditions within the gastrointestinal tract. As with antibiotics, phage-based microbiome manipulation may leave an environmental niche which can be populated by pathogenic bacteria. Because of this, development of new strategies for phage therapy followed by niche filling with beneficial bacteria may prove important.

Using phage as a delivery system

Another therapeutic approach being investigated is the development of genetically modified phages that serve as delivery vehicles for antimicrobial agents (176–180) and modulators of bacterial gene expression (181–183). Du et al. (180) have recently described engineered phages that encode bacteriocins and cell wall hydrolases, which are produced following bacterial infection. In this way, their engineered phages, coined heterologous effector phage therapeutics (HEPTs), serve dual functions of phage-mediated bacterial killing and production of antimicrobial compounds that are released after bacterial lysis. Du et al. used these HEPTs to target urinary pathogens E. coli, Enterococcus faecalis, and K. pneumoniae, showing that they could be engineered to target both single- and multi-genus bacterial communities. In the single-genus bacterial communities, which is the likely situation during UTI, HEPTs were also effective at preventing or delaying the development of phage resistance, which was further shown in urine samples collected from patients with UTI.

In addition to delivery of antimicrobials, phages have been proposed as useful detection and diagnostic agents via delivery of a fluorescent signals (184–188). Recently, Meile et al. (188) demonstrated how these techniques could be useful in rapid diagnosis of UTI by constructing phages specific for E. coli, Enterococcus, and Klebsiella that result in nanoluciferase gene expression following infection. Using this system, they were able to screen 206 patient samples, detecting the pathogen at levels less than 10³ bacteria per milliliter of urine with a high level of specificity and with a detection time of under 5 hours.

CONCLUSION

As the AMR threat continues to grow, new therapeutics are desperately needed for treating our most common infections, such as UTI. The development of novel therapeutics including vaccines, antimicrobials, and bacteriophage therapy have been proposed. Bacteriophages are poised to play an important role in fighting against AMR due to their therapeutic potential, long history as UTI treatment, and ability to guide virulent pathogens into less virulent states.

Biographies

Jacob Zulk obtained his undergraduate degree from the University of Minnesota and is currently a doctoral candidate in the Immunology & Microbiology Graduate Program at Baylor College of Medicine. His research focuses on phage therapeutics for the treatment of antibiotic-resistant urinary tract infections and the ways in which bacteria evade phage killing. He is the recipient of the Ruth L. Kirschstein Predoctoral Individual National Research Service Award (F31.) He was named a 2022 Chateaubriand Fellow by the French Embassy in the United States and spent eight months researching the immune response to urinary tract infection with Dr. Molly Ingersoll at Institut Cochin in Paris.

Kathryn Patras grew up in the Midwest where she received a BS in Animal Science from the University of Nebraska-Lincoln while working under Dr. Jennifer Wood. She completed her PhD at San Diego State University with Dr. Kelly Doran and postdoctoral fellowship at University of California San Diego with Dr. Victor Nizet. In 2020, she joined Baylor College of Medicine as an Assistant Professor with appointments in Molecular Virology and Microbiology and the Alkek Center of Metagenomics and Microbiome Research. The goal of her research program is to understand host-pathogen-microbiota interactions in the urogenital tract. Her group uses newly developed models to study why individuals with certain conditions, such as pregnant or diabetes, are more susceptible to urogenital infection, to test novel therapies, and to establish the functional role of the urogenital microbiota with ultimate applications to both disease pathogenesis and overall human health.

Anthony Maresso is the Melnick Endowed Chair of Virology and Microbiology at Baylor College of Medicine. His Ph.D. is in Genetics (Medical College of Wisconsin) and his postdoctoral work in Microbiology (The University of Chicago). His program is funded to study all things microbe, including mechanisms to prevent aging, new vaccine and therapeutic development, the ways in which genomes mutate, organoid engineering and biology, directed evolution, and the use of genetic libraries to decrypt the function of orphan genes. He is the founder of TAILΦR Labs, the first dedicated phage discovery and manufacturing laboratory in the United States, and PHIOGEN, a biopharma company. He has authored > 70 original papers, including a textbook on bacterial virulence. He has trained 54 Ph.D. Students, Postdoctoral Fellows, Undergraduates, Clinical Fellows, and Technicians and is the recipient of his institution’s highest honor in mentoring and teaching.