Immediately after the discovery of phages by Twort and d’Herelle, d’Herelle postulated that phages could have therapeutic applications. In 1919, he used phages to treat chickens that were infected by Salmonealla gallinarum with success ^[1][2][119,120]. Based on the successful use in animals, d’Herelle realized that phage treatment could be attempted in humans with bacterial disease. In 1921, five patients suffering from bacillary dysentery were treated successfully with the use of a phage targeting Shigella dysenteriae ^[1][3][119,121]. Furthermore, clinical trials of phage treatment in cholera in India showed a significant decrease in mortality from 62.8% in the control group to 8.1% in the group treated with phages, while d’Herelle noticed that when adding phages targeting cholera into drinking wells during outbreaks of the disease, a reduction of subsequent infections was noted ^[4][122].

After these encouraging results in animals and humans, other scientists recognized phages’ potential in prophylaxis and treatment and started targeting other infections, although with partial success. Beyond criticism regarding the quality and the design of these studies in humans, an important drawback at that time was the phages’ high specificity against certain bacteria, implying that previous recognition of the pathogen was absolutely necessary to allow successful treatment by phages ^[5][27]. In 1923, phages were successfully used for the treatment of bacteremic patients with typhoid fever. However, at that time, other scientists reportedly failed to achieve the same results using a similar patient population ^[6][7][123,124]. This inconsistency was considered to be due to the use of a phage with a very narrow spectrum. Other drawbacks included issues regarding phage production, involving contamination of the product due to imperfect filtering and purification steps. Additionally, pharmacokinetics suggested rapid phage elimination from patients’ spleen, rendering the therapeutic effect short-lived. Furthermore, the bacterial ability to rapidly mutate and develop resistance to phages probably played an important role. More specifically, bacteria have multiple antiviral mechanisms that may inhibit the entry of a phage (by blocking their interaction with the bacterial receptors or producing an extracellular matrix that may reduce the likelihood of the interaction leading to phage entry in the bacterial cell or may produce competitive inhibitors that may antagonize phage binding on the bacterial cells), may prevent DNA entry into the bacterial cell, may inactivate phage DNA even after entry, or may lead to abortion of the infection by even inducing bacterial cell self-death ^[8][125]. Due to the adaptive nature of these mechanisms, there is a possibility for the development of resistant bacterial clones during phage treatment ^[9][126]. Finally, the possibility of a lag between in vitro and in vivo experiments cannot be excluded ^{[5][10][11][12][13][14]}[23,27,127,128,129,130]. Thus, given these issues and the fact that the interest in the production of antibiotics was on the rise, interest in phage therapy waned for many decades in the West but notably persisted in the USSR and Eastern Europe. Nowadays, with the problem of the increasing antimicrobial resistance that threatens global health and sets at risk millions of hospitalized patients, interest in phage therapy is on the rise again in Western countries ^[5][27].

A series of well-designed experiments by Smith and Huggins helped to rediscover phage therapy, as they assessed several issues that had been identified as limitations in the previous sets of experiments in the first half of the 20th century and also confirmed that phage treatment is safe and efficient in animal models ^[5][27]. In the first series of experiments, Smith and Huggins showed that there was a close in vitro to in vivo correlation of phage efficacy. For experiments, they chose phage R that targets K1+ E. coli, showing the greatest results in in vitro virulence ^[15][131]. In the following experiments, they showed that a single dose of phage R was equally effective as eight doses of streptomycin ^[15][131]. Furthermore, bacterial lysate, free of phages, had no therapeutic effect in the infected animals. Intramuscular injection of phages into uninfected mice proved that they persisted inside the injected muscle as well as the spleen for four weeks after injection. However, pathogenic bacteria were cleared within 16 and 20 h after phage injection from the liver and the blood, respectively. Finally, bacterial mutants resistant to R phages were K1-, occurring at a frequency of about 0.01, while they were known to be avirulent. With this study, most of the concerns described previously were addressed, and phage treatment was presented as at least equal to or even more effective than antibiotics ^[5][27].

In another set of experiments, Smith and Huggins investigated the issue of bacterial resistance to phages, choosing an E. coli diarrhea model in calves. They developed a multi-phage treatment plan to combat the development of resistance during treatment. Before experimenting in vivo, they performed in vitro studies. Hence, they chose lytic phage B44/1, which had the ability to infect only K85+ strains of E. coli. Then, they isolated bacterial mutants that were resistant to that phage in vitro. Subsequently, they chose another phage, namely B44/3, which was able to infect bacteria resistant to the initial B44/1 phage that they used. Additionally, they selected bacterial mutants resistant to B44/3 but susceptible to B44/1. By evaluating the dynamics of phage resistance before using them in vivo, they were able to use this double-phage approach to overcome the problem of resistance evolution to phages ^[16][132]. Smith and Huggins also evaluated phage stability during treatment given orally. They noticed that although there was poor phage stability in the stomach due to an acidic environment, this could be overcome by administering calcium carbonate before the oral administration of phage ^[17][133]. Thus, this group of scientists undermined all that was previously considered as significant drawbacks for phage treatment, opening the way for further experiments and giving hope for phage application ^[5][27].

New experiments from other research groups have further elucidated questions on phage biology and therapeutics. A more recent study showed that even though rapid clearance of phages in the blood was considered a drawback, it is possible to select phage variants that have longer blood half-lives ^[18][134]. Other experiments showed that phage treatment was also efficacious against pathogens such as P. aeruginosa or A. baumannii ^[19][135]. Moreover, in another set of experiments, a phage targeting the K1+ E. coli provided 100% protection to mice, while treatment with phage not targeting K1 led to a mortality of 60%. A comparison with streptomycin revealed that the phage targeting K1 led to 9% mortality, while treatment with streptomycin was associated with 54% mortality, thus confirming the previous findings by Smith and Huggins ^[20][136].

The 21st century offers many more tools than the previous one, such as high-throughput methods for efficiently screening thousands of samples at the same time, affordable whole-genome sequencing, automated technology for microbiological techniques, etc. On the other hand, clinical studies have shifted to a different level, with a need for more adequate design for power, double-blind and randomized settings, as well as higher standards regarding safety for participants. Current technology allows deeper investigations at the time of the study or even after that, addressing clinical and biological questions that may have to do with the in vitro as well as the in vivo interaction of the phage with the target bacterium and the immune system of the human host ^[5][27].

Importantly, the well-known physiological property of very narrow targeting of phages that could be considered a limitation in some instances could be addressed with the current technology. The work by Dedrick et al. showed an example of how narrow the targeting of phages is and how phage engineering could be elaborated. In this study, a collection of more than 10,000 phages isolated using Mycobacterium smegmatis was needed for screening to allow the identification of just three useful phages that were used for the treatment of a young boy with cystic fibrosis and infection by Mycobacterium abscessus. Two of them were engineered to allow appropriate bacterial targeting ^[21][22]. Nowadays, CRISP-Cas9 technology can also be used to engineer the genomes of phages to manipulate their bacterial targeting ^[22][23][137,138].

Regarding the development of resistance to phages by bacteria, modern technologies of evolutionary biology could provide insights about appropriate phage selection. In particular, the selection of a phage targeting a specific molecule as a receptor leads to the selection of bacterial strains that do not express it. Hence, therapeutic use of phages could take advantage of the receptors that are concomitantly virulence factors. For example, the selection of strains that do not produce the virulence factor that is the target receptor for the phage might lead to bacterial survival; however, their virulence would be lower, leading to better infections outcomes ^[5][24][25][27,139,140]. Hence, the use of phages that target lipopolysaccharide (LPS) could lead to significant alterations to LPS produced by bacteria, and this might render them less virulent ^{[26][27][28][29][30][31]}[141,142,143,144,145,146]. Thus, careful selection of the type of phage to be used in the fight against infectious diseases is important and should take into consideration the evolution of phage resistance to turn it into an advantage for humans and animals.

Currently, the application of artificial intelligence in the field of phage therapeutics is a trending topic since it can facilitate the selection of phages depending on the specific characteristics of the target pathogens and the host’s profile as well ^[32][33][34][147,148,149]. Hence, machine learning, which has been implemented broadly in biology, was shown to be able to integrate enormous amounts of information from omics data to better understand the phage–host interaction. This could be used for better identification of candidate phages for human medical use ^[34][149].

Phages have been found to be effective in the management of systemic infections in several animal models. In a gut-derived model of Pseudomonas sepsis, 67% survival was noted after oral administration of phage therapy one day after the infection of mice ^[35][150]. Phage dose-dependent reduction of mortality was noted in a mouse model of bacteremia by Enterococcus faecium resistant to vancomycin ^[36][151]. In another mouse model of infection by Vibrio vulnificus, successful treatment was noted only when the administration of phage therapy was performed at the same time as the infection by the bacteria ^[37][152]. Thus, the efficacy of phage treatment seems to rely on several factors that should be taken into account, such as the dose and the time after infection.

The use of phages for the treatment of localized infections such as an abscess, an ear infection, or a burn has been proven very efficient. An intraperitoneal model of infection by P. aeruginosa in mice treated at the same time with phages showed 92% survival ^[35][150]. In another study using S. aureus leading to the formation of abscesses, administration of phages at the same time as the pathogen prevented the formation of abscesses; meanwhile, when phage treatment was given 4 days after the bacterium inoculation, a single dose of phage treatment led to 100-fold reduction of the bacterial load, and when multiple doses of phage treatment were applied, 10,000-fold reduction was observed ^[38][153].

The use of phages against pathogens infecting the gastrointestinal tract may eradicate the pathogenic bacterium without altering the normal gut flora. Phage treatment four days after inoculation of adherent-invasive E. coli to the gut of mice led to reduced bacterial colonization and was associated with a reduced likelihood of colitis development ^[39][154]. A study that used an insect model of Clostridioides difficile colonization showed that prophylactic treatment with a phage two hours before inoculation of bacteria led to 100% survival. Simultaneous administration of phage and bacteria led to a 72% survival, while phage treatment two hours after inoculation of bacteria led to 30% survival ^[40][155].

Treatment of lung infections with phage therapy could help people with chronic lung infections, such as those with cystic fibrosis, who are also at particular risk of colonization and developing infections by antibiotic-resistant microorganisms. An experiment in mice using P. aeruginosa showed that intranasal treatment with two doses of phages led to complete eradication of the bacterium when administered 24/36 or 48/60 h after initiation of the infection, while treatment at 144/156 h after an infection led to complete eradication of the infection in 70% of the animals and a significant reduction of the bacterial load in the lungs of the rest ^[41][156].

Phage therapy may also have application in veterinary medicine, with phages being used against a variety of pathogens and in a variety of hosts. Hence, phages have been effectively tested against Campylobacter jejuni in chicken ^{[42][43][44][45][46]}[157,158,159,160,161], against Salmonella enterica serovar Enteritidis in chicken ^[47][48][49][162,163,164], against S. enterica serovar Typhimurium in weaned pigs ^[50][51][165,166], against S. aureus in bovine mastitis ^[52][53][167,168], and against S. aureus in lactating dairy cattle, among others ^[54][55][169,170].

There are several studies evaluating the effect of phage treatment on specific infections. However, there are few studies evaluating the effect of the combination treatment of phages with antibiotics, with some of them being promising. Hence, in a study of broiler chickens infected with E. coli, treatment with fluoroquinolone (enrofloxacin) led to a reduction in mortality from 68% down to 3%, while treatment with phages led to a reduction in mortality to 15%. Combination treatment with fluoroquinolone and phage simultaneously led to 0% mortality ^[56][171]. Another study showed that in a rat model of endocarditis induced by P. aeruginosa, combination treatment of ciprofloxacin and phage led to a 10,000-fold reduction of bacterial load compared to treatment with ciprofloxacin or phage alone; meanwhile, the same combination showed synergy in the killing of P. aeruginosa both in vitro and in vivo ^[57][172]. Thus, even though the topic of phage treatment is promising, studies evaluating both the in vitro and in vivo effects of phages with antibiotic combination treatment are warranted.

The use of phages in humans for the treatment of infectious diseases in the modern era is limited. However, there are studies evaluating their efficacy and safety ^[5][58][27,173]. There are reports of individual patients with limited therapeutic options treated with phages due to resistance or allergies to antibiotics. A two-year-old patient with DiGeorge syndrome and allergies to several antibiotics who developed P. aeruginosa bacteremia and failed treatment with anti-pseudomonal antimicrobial agents was treated with two-phage combination with temporary blood sterilization. However, after the end of this treatment, the blood cultures became positive again ^[59][174]. It is of note, though, that in this case, the recurrence of bacteremia does not necessarily imply that the phage treatment failed since bacteremia developed in the context of persistent and uncontrolled infected fluid collections of the thoracic cavity. Thus, phage treatment of uncomplicated bacteremia could be possible; however, this has yet to be confirmed by future studies.

In another case report, a patient with a P.-aeruginosa-infected aortic graft, complicated by aorto-cutaneous fistula with purulent discharge, failed very prolonged treatment with several antibiotics. Since, at that point in time, the patient was not a candidate for a new surgery, a phage active against P. aeruginosa that was shown to have synergy with ceftazidime and was screened for lytic activity against the causative organism was applied locally in the exit point of the fistula, along with systematic administration of ceftazidime ^[24][60][139,175]. One month after phage treatment, partial graft excision and replacement took place, while all cultures were negative, and systematic ceftazidime treatment was stopped. Two years later, the infection had not relapsed in the absence of antimicrobial treatment.

Recent clinical trials in humans often show contradictory findings ^[58][173]. However, there are studies with positive results that increase optimism that carefully and thoughtfully selected phages for the right pathogen and the right route of administration could be of clinical use in the future. A phase-one, first-in-humans, open-label clinical trial of multiple ascending doses was performed in a tertiary referral center in nine patients with recalcitrant chronic rhinosinusitis using the investigational phage cocktail AB-SA01. Intranasal treatment with AB-SA01 in different doses of up to 3 × 10⁹ PFU for two weeks was shown to be a safe and well-tolerated treatment, while the preliminary efficacy observations were promising ^[61][176]. This study, even though a phase-one study in a small number of patients, implies that in the future, treatment with phages could be an alternative to classic antibiotics in patients with recalcitrant chronic rhinosinusitis.

The PhagoBurn trial was a randomized phase I/II trial where adult patients with confirmed burn infection by P. aeruginosa were recruited in French and Belgian burn centers and received treatment after randomization with either standard treatment involving 1% sulfadiazine silver emulsion cream or a cocktail of 12 natural lytic anti-P.-aeruginosa phages, and the time to an adequate sustained reduction of bacterial burden was assessed. The phage cocktail was found to decrease the bacterial burden in burn wounds at a slower pace compared to the standard of care, implying that future studies with higher concentrations of phages are required ^[62][177].

Nowadays, FDA approval has not yet been granted to phage treatment, and this could be due to concerns regarding approving a medication that is “alive” and difficult to standardize. Thus, since concerns still exist and hinder the approval of this modality, other approaches such as developing recombinant phage-derived proteins could be elaborated ^[63][64][188,189]. Even though this treatment is not yet approved, there are some clinical trials that have been registered and are currently active ^[65][66][190,191]. Clinicians in the USA who intend to use phages in patients must submit an investigational new drug application to the FDA ^[67][192]. Similarly, the European Medicines Agency (EMA) has not yet approved any bacteriophage for clinical use ^[68][193], though the draft guidelines on “quality, safety and efficacy of bacteriophages as veterinary medicines” are now undergoing an expert public consultation ^[69][194]. Moreover, a chapter on bacteriophage use in therapy will shortly be introduced in the European Pharmacopoeia ^[70][195]. There are several barriers worldwide to the production and application of phages as an alternative or a complementary therapy to classical antimicrobials. The most important one is the lack of adequate data from clinical trials set up based on widely accepted ethical standards. There are, however, countries such as Poland, Georgia, and Russia where phage therapy is used, even though there are no clearly adapted regulatory guidelines ^[71][196]. In Poland, however, treatment with phages is considered “experimental” under the Polish Law Gazette, 2011, item 1634 and article 37 of the Declaration of Helsinki ^[68][72][193,197].