Emerging applications of phage therapy and fecal virome transplantation for treatment of Clostridioides difficile infection: challenges and perspectives

Most C. difficile strains carry a set of different prophages inside their genomes [46]. These prophage-related elements have narrow host ranges in various bacterial species and can infect different C. difficile strains [67]. Typically, all known infecting phages of C. difficile belong to temperate families, including Myoviridae and Siphoviridae of Caudovirales order [68], and carry genomes of approximately 31–56 kb in length and a GC content of about 28–30% (Table 2) [69]. The interaction between phages and C. difficile strains depends on the availability of suitable receptors on bacterial host cells. There is little evidence about the phage receptors on the cell surface of C. difficile. So far, some studies have shown that surface layer (S-layer) protein, SlpA, may be a phage receptor candidate for C. difficile phages [70, 71]. Recently, Whittle et al., supported this claim and showed that SlpA of C. difficile can act as a cell surface phage receptor [72]. Each C. difficile strain contains phage-related genomic regions and carries 1 to 6 prophages. Some prophages have a large genome (> 130 kb), which can be stable as extrachromosomal DNA in C. difficile cells [67]. Hence, C. difficile genome is typically mobile and mosaic, for example, 11% of the genome of C. difficile strain 630 has been originated from MGEs [73]. Additionally, there are some reports about the diversity of prophages in different clinical C. difficile ribotypes (RTs). Based on these results, almost all discovered phages belong to myovirus, some RTs carry siphovirus prophages, and few RTs are positive for dual phage type carriage including myophages and siphophages [67, 74]. These reports have supported the coexistence of prophages with C. difficile through integration into its genome. These genetic exchanges may improve the bacterial adaptation in the GI tract by the acquisition of new traits [75]. However, C. difficile strains possess efficient defense systems to balance genetic gain and diversity, by which they can survive within phage-rich gut communities, and avoid over-acquisition of foreign genetic elements such as phages and plasmids [76]. Presently, the mechanisms of action for some of the C. difficile anti-phage defense systems are elucidated. These systems can be activated after injection of phage DNA, such as restriction-modification systems, and clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9 (CRISPR/Cas9), leading to inactivation of infection by breaking phage DNA [68, 76, 77]. Additionally, bacterial defense systems can prevent infection spread through toxin-antitoxin systems, which actuate a suicidal host response or dormancy [32]. Interestingly, endogenous prophages may express proteins, which block phage binding to bacterial surface receptors or restrict DNA injection, thus, preventing superinfection of their host [78].

C. difficile phages are involved in the development of susceptibility/virulence-associated phenotypes of their bacterial host. Some studies have demonstrated that prophages can influence the genes related to the pathogenicity of C. difficile and contribute to emergence of more virulent strains [46]. The most important impact of phages on the pathogenesis of C. difficile is their negative or positive effect on toxin expression (e.g. prophage phiSemix9P1 isolated from some C. difficile strains carries a locus encoding binary toxin (CDT)) [88]. Moreover, phiCD119 prophage can express RepR regulator that binds to TcdR promoter, leading to the repression of TcdR expression [93]. TcdR is an alternative sigma factor, which is involved in tcdA and tcdB expression by recruiting RNA polymerase to their promoters [94]. Thus, the expression of the phiCD119 RepR protein in C. difficile results in decreased toxin production. Notably, some prophages can show common features with the pathogenicity locus (PaLoc) of C. difficile strains. Accordingly, the PaLoc of some C. difficile strains can encode a phage-like holing (TcdE), which is a membrane protein and can generate pores in bacterial cell membrane and degrade the cell wall to facilitate bacteriophage release from cytoplasm [95]. Some studies have also reported that C. difficile phages can influence the expression of TcdA and TcdB [53, 93]. For example, infection of C. difficile CD274 with a temperate siphophage, called CD38-2, leads to over-expression of TcdA and TcdB up to 2-fold, indicating the role of phage in production of C. difficile toxins [85]. C. difficile phages possibly can influence toxin production either by increasing the transcription of PaLoc genes or introducing novel regulatory genes into their hosts’ genomes during lysogenic cycles [46, 68]. Encoded genes by C. difficile prophages could also impact the regulatory genes involved in the expression of surface proteins, quorum sensing (QS), and antibiotic resistance [52, 96]. For example, phiCDHM1 and related prophages carry an accessory gene regulator (Agr)- like QS gene, which influences phenotypes associated with C. difficile virulence, such as biofilm formation, oxidative resistance, and motility [74, 97, 98]. Additionally, some phages can encode a class of enzymes, known as adenosine-diphosphate-ribosyltransferases (ADPRTs), which can increase the adherence and mucosal colonization of C. difficile in the host mucosa [40, 99]. It has been also reported that phages can affect the expression of C. difficile cell wall proteins. In this regard, a recent study demonstrated that infection of C. difficile RT027 with phage CD38-2 causes 20-fold upregulation in the expression of cell wall protein CwpV [100]. Furthermore, infection with CD38-2 leads to the downregulation of genes associated with the uptake and metabolism of carbohydrates e.g. glucose, fructose and D-glucitol, in bacterial host cells [101]. Phage infection can also impact the regulation of bacterial defense systems. Another study demonstrated that infection of C. difficile RT078 with phage JD032 can change the expression of genes encoding for DNA and RNA synthesis, and suppress anti-phage systems, including toxin-antitoxin, restriction-modification, and CRISPR-Cas systems [91]. In contrast, an in vitro study demonstrated that phage øCD27 can reduce cell numbers of C. difficile and its toxin production without major effects on the composition of the gut microbiota [102]. Additionally, infection with lytic phages like CDHS-1 can decrease colonization and have negative effects on bacterial pathogenicity [100]. Therefore, lytic phages could be a valuable choice for therapeutic purposes against CDI.

The alarming rate of antimicrobial resistance has necessitated sustained research and an urgent need for new and effective alternative treatment approaches to traditional antibiotic therapy [103]. One option could be the use of phages as therapeutic agents, which can be technically effective to prevent the challenge of antibiotic resistance [14]. The application of phages to control infectious diseases in animals has been of great interest for many years. Shortly after the discovery of phages in 1915, the use of phage therapy to treat bacterial dysentery was proposed by Felix d’Herelle [19]. This hypothesis inspired the application of phages as a therapeutic tool to control bacterial infections and led to commercial production of phages in several countries until the 1940 s. However, there were some limitations in approval processes of phage-based products due to the lack of complete characterization of phages [61]. In recent decades, the increasing rise in the rate of antibiotic resistance has led to revisiting phage therapy as drug candidates. Therefore, a guideline has been proposed for the data collection on phages, which can incorporate provisions of the Food and Drug Administration (FDA) to receive approval for phages as possible therapeutics [104]. According to this guideline, an appropriate phage for therapeutic use should meet certain criteria, including having a narrow specificity range to attack specific target cells, preventing undesired lysis of commensal microbiota, the ability to replicate inside their host by hijacking host DNA replication machinery, the ability to evolve in response to host evolution, the ability to overcome some mechanisms of phage resistance in bacteria, and the inability to attack mammalian cells or having no unfavorable immune reactions [105‐107]. Additionally, the choice is limited to obligate lytic phages that do not encode any virulence factor-associated genes (e.g. toxin genes or antibiotic resistance determinants) [104]. In this regard, a phage cocktail containing LH01-Myoviridae, LL5-Siphoviridae, T4D-Myoviridae, and LL12-Myoviridae (PreforPro®) has been recently introduced as a next-generation prebiotic, which colonizes common probiotic strains such as Lactococcus, Bifidobacterium, Lactobacillus, and Bacillus subtilis and enhances their efficiency through reducing the incidence and severity of GI distress [108]. Currently, all of the natural phages selected for therapeutic purposes belong to the order Caudovirales [26], which have exhibited desirable efficacy in controlling infectious bacteria such as Pseudomonas, E. coli, and Salmonella enteritidis [109‐112].

So far, different strategies have been proposed for enhancing the efficacy of phage therapy such as the use of phage cocktails, combination of phage and antibiotics, phage-derived enzymes, and phage engineering [113]. Phage cocktail is introduced as an alternative to single-phage therapy, which can overcome the limitations of single-phage therapy and delay the development of bacterial resistance to phages. This method can bypass shortcoming of the narrow phage lysis spectrum and be used to target single or multiple bacterial pathogens [114, 115]. The use of phage-derived enzymes has been also considered in some studies, among them, phage lysins have been applied for the control of several bacteria [116, 117]. Phage lysins are a class of peptidoglycan hydrolases, which degrade the bacterial cell wall peptidoglycan [116]. These enzymes are safe and species-specific, and therefore do not damage the normal intestinal microbiome. Moreover, the use of phage-derived products (endolysins, phage tail-like particles (PTLPs), and fusion proteins) can reduce the possibility of the emergence of resistant pathogens [116, 118]. Moreover, combination of antibiotics with phage or phage-derived enzymes could show better therapeutic effects than single-phage therapy. Interestingly, some antibiotics can exhibit a synergistic effect on phage therapy through increasing the propagation of lytic phage in bacterial host, leading to acceleration of bacterial cell lysis and the release of progeny phages [55].

Another solution for enhancing the efficacy of phage therapy is the use of phage genetic engineering. Phage engineering can improve the therapeutic effects of phage therapy by expanding lysis spectrum of phages and inhibiting the emergence of resistant bacteria. The use of engineered genes encoding receptor-binding proteins (RBPs) in spikes and the use of CRISPR-Cas systems are the most common methods applied for phage engineering [113, 119]. The genetic engineering of RBPs is a powerful tool to produce broad-spectrum phages [119], whereas CRISPR-Cas systems integrate the short fragments of the phage genome into the CRISPR array (namely crRNAs), which leads to the production of complementary RNA sequences. The crRNAs guide the Cas protein complex for targeting or depredating specific foreign genetic elements [113, 120]. To sum up, these strategies can improve phage therapy outcomes and has a great prospect for preventing or treating drug-resistant bacterial infections.

In addition to refined phage therapy, the use of FVT is getting mounting attention as a new therapeutic option in recent years. So far, FVT has been used for treating a number of diseases linked to gut microbiome dysbiosis such as metabolic syndrome, type 2 diabetes (T2D), and obesity [19, 66]. Although FMT results in an almost complete restoration of the balance of microbiota in dysbiotic patients, the precise mechanism of action of FMT has yet to be fully elucidated [121]. Generally, restoring the recipient’s gut microbiota is achieved by transferring bacteria, phages and other microbes form a healthy donor following FMT [18]. Generally, fecal virus-like particles have a density similar to fecal bacteria, which is about 10⁹ per gram of feces, and more than 90% of the viral density is dominated by phages [122]. Therefore, a large number of fecal phages is transferred during FMT, which may have significant physiological effects on the overall well-being of recipients. Accordingly, some studies demonstrated that bacteriophages play an important role in successful FMT treatments through controlling the disease progression and restoring the balance of the gut microbiome [17, 123]. The procedure of FVT and its routes of administration are almost similar to FMT, albeit it is filtered to exclude intact fecal bacteria. Hence, the use of sterile fecal filtrate of a healthy donor can be administered as a refinement to FMT, which decreases the risk of invasive bacterial infections and adverse events (Fig. 2A). Recent in vivo studies have supported the capability of FVT to normalize the gut microbiota population after antibiotic therapy by affecting both the bacteriome and virome of recipients [13, 19, 124]. More specifically, in a dysbiotic state in CDI patients, a decrease in the abundance of Firmicutes (especially Clostridia), Bacteroidetes, and Actinomycetota, and an increase in the abundance of Proteobacteria (especially Gammaproteobacteria) are commonly observed [17, 125]. After FMT or FVT, the bacterial communities of recipients resemble those of donors, and the abundance of Firmicutes, Bacteroidetes, and Actinomycetota are restored. In addition to bacteriome, the abundance, diversity and richness of virome are also altered in CDI patients, in which a significantly higher Caudovirales abundance, lower Caudovirales diversity, richness and evenness, and a decreased abundance of Microviridae are reported as compared to healthy subjects [17]. Furthermore, it has been shown that FMT treatment resulted in a significant decrease in the abundance of Caudovirales, while caused an increase in Microviridae abundance (Fig. 2B) [17, 125]. These studies document the effectiveness of FVT treatment in recipients, which is probably due to phage-driven manipulation of the gut microbiota, leading to improved host metabolome and health [18, 19]. Interestingly, a recent study showed that necrotizing enterocolitis (NEC), an inflammatory disease of the small intestine, can be completely controlled by FVT through oral administration in piglets with NEC, whereas ∼42% of cases still showed the disease after administration of FMT [124]. Moreover, these researchers concluded that the application of FVT exhibited higher safety and efficacy than FMT in NEC animal model.

However, there are limited studies characterizing the role of phages in the treatment of CDI or rCDI. Thus, considering the high effectiveness of the FMT to reduce the risk of rCDI, the beneficial role of phages in CDI treatment and their impact on the gut microbiome homeostasis should not be ignored. Furthermore, the utilization of phages would be an innovative approach to combat biofilm formation especially by antibiotic-resistant C. difficile strains, which are highly difficult to eradicate using common antibiotic therapy [15, 17, 126].

FMT from healthy donors has acted as a highly efficient microbiome-based therapeutic option for treating rCDI patients with a success rate of more than 90% [127]. However, there are multiple safety issues for the use of FMT in clinical setting, which may limit its widespread application in critically ill or immunocompromised patients [128]. Additionally, due to the lack of a comprehensive and standard donor screening panel, the risk for the transfer of harmful agents from the donor microbiome cannot be fully prohibited [18]. As over mentioned, alterations in the gut microbiota composition after FMT are mostly attributed to the transferred bacterial communities. Based on recent findings obtained from various human microbiome projects, we well know that bacteria are not the only transferred component following microbiota transplants. The non-bacterial gut residents, including viruses, archaea, fungi, protozoa, and parasites, are also transplanted to different sites in the gut of the recipients post-FMT, which play vital roles in the re-establishment of a healthy microbiome [17, 129]. In a recent study by Zuo et al., it was demonstrated that phage compositions are altered in rCDI patients after FMT and resembled to those of the donor, as far as higher Caudovirales richness was reported in responders compared with those who did not respond to FMT treatment [17]. They also found that CDI patients demonstrated a significantly lower Caudovirales diversity, richness, and evenness compared with healthy household controls, suggesting that Caudovirales can play a key role in the success and effectiveness of FMT treatment. It was explained that Caudovirales can affect the efficacy of FMT in rCDI patients through directly altering the dysbiotic microbiota or by improving the bacterial colonization. Furthermore, these rCDI patients showed a lower abundance of bacteriophage Microviridae compared with healthy subjects. Notably, the enrichment of fifteen viral species of Microviridae family was reported in FMT responders. Interestingly, the intestinal phages transferred with FMT can sustainably engraft in the recipient gut microbiota for a protracted period [130]. By analyzing the long-term effects of FMT, it was shown that the phageome composition of rCDI patients who responded to FMT could resemble the phage community of healthy donors and last for 7 to 12 months after treatment [123, 131]. This long-term persistence may be due to the strong adsorption of phages to mucus and epithelial cells in the gut of the recipients [67, 132]. These findings indicate that apart from living bacterial species, other components of the microbiota, particularly bacteriophages, are contributed to the re-establishment of the gut microbiota after FMT treatment [129].

Recently, a great interest has been attracted toward C. difficile phages as an alternative to antibiotics for CDI treatment. However, the lysogenic nature of most of the C. difficile phages has significantly restricted the application of these viruses for CDI treatment. In addition, most C. difficile phages have been recovered after the induction of the host with mitomycin C, while the natural induction of prophages in CDI patients has been reported only in a study conducted by Meessen-Pinard et al. [86]. In this regard, these researchers isolated four Myoviridae phages including φMMP01, φMMP02, φMMP03, and φMMP04 from filter- sterilized stool supernatants of CDI patients. This study provides evidence that natural induction of prophage can play a role in killing C. difficile cells during episodes of CDI. It is expected that temperate phages can undertake lytic infection, and thus may still be valuable for therapeutic use. So far, different studies have been carried out for the use of temperate phages and phage-derived proteins to treat CDI. A summary of these studies is presented in Table 3.

Single-phage therapy has been performed as the first phage-based treatment for CDI in 1999. In this regard, the administration of phage CD140 significantly improved the survival of hamsters challenged with C. difficile [133]. Furthermore, the use of phage phiCD27 reduced both C. difficile vegetative cells and TcdA/TcdB production in batch fermentation and in an artificial gut model [102, 134]. Recently, the therapeutic potential of phage CDHS-1 was indicated, which targets and kills C. difficile hypervirulent RT027 strain by reducing its colonization and applying negative impacts on bacterial pathogenicity [100]. Interestingly, this phage can act more effectively in the presence of epithelial cells than when used to infect bacterial cells alone [132]. Therefore, this finding suggests that CDHS-1 has promising therapeutic potential for controlling the infection in the gut. However, there is little information about the mode of action of CDHS-1 for modulating its bacterial host genome during the infection cycle, thus further investigations are needed to help ascertain the potential therapeutic implications of this phage in the future. Moreover, single-phage therapy has some shortcomings due to the lysogenic nature of phages, their narrow host spectrum, and the emergence of phage resistance [74]. Hence, the use of phage cocktails, combination of phage and antibiotics, phage-derived enzymes, and phage engineering can be a superior strategy to overcome the pitfalls of single-phage therapy for CDI treatment (Fig. 3) [74, 98].

The effectiveness of the phage cocktail strategy to treat CDI has been investigated in both in vitro and in vivo experiments [74, 100, 136, 145]. Nale et al. examined the effectiveness of temperate phages phiCDHM1 to phiCDHM6 as an individual or cocktail against different strains of C. difficile [74]. Based on their results, multiple-phage cocktails, especially combinations including phiCDHM 1, 2, 4, and 6, could kill a broader range of C. difficile strains without regrowth than a single phage. Interestingly, there was increased phage resistance against individual phages, whereas phage resistance was limited in the application of phage cocktails. In another study, Nale et al. demonstrated that a 4-phage cocktail targeting C. difficile could reduce and prevent biofilm formation of C. difficile RT014/020 or also eliminate the bacteria [98]. The effectiveness of multiple-phage cocktails was further investigated in an in vivo model and it was shown that the use of a combination of phiCDHM 1, 2, 5, and 6 can reduce the number of spores in the cecum and colon of CDI hamsters [74]. Most importantly, the abundance of commensals, including total anaerobes, lactobacilli, and enterobacteria, was not altered by phage therapy. This suggests that phage-based treatment can disturb the microbiome structure, however, further studies are required to validate this claim.

As aforementioned, the use of phage lysins has also been proposed to overcome limitations of the lysogenic nature of phages. Endolysin is a peptidoglycan hydrolase that is encoded by lytic phages. Endolysin helps releasing of phage progeny by disrupting bacterial cell wall during the final step of viral infection. The phiCD27 endolysin (CD27L) is the first endolysin characterized in C. difficile, which has exhibited promising effect in controlling a panel of C. difficile strains over a wide range of pH conditions [40, 83, 137]. These results show that the application of phage lysins can provide specific treatment options for CDI, in particular for drug-resistant C. difficile strains [146]. Notably, the partial sequence of N-terminal portion of CD27L, and CD27L1-179, exhibited a broader lytic range than the full length itself. Also, both CD27L and CD27L1-179 were harmless to other gut commensal microorganisms [137]. PlyCD is another endolysin derived from a prophage of C. difficile 630, which shows strong lytic activity against a variety of C. difficile strains. The catalytic domain of N-terminal portion of PlyCD (PlyCD1-174) displayed higher lytic activity than the full-length [138]. A recent study demonstrated that the expression of a recombinant fusion protein containing the catalytic domain of the endolysin from phiC2, and the functional domain of the human defensin protein HD5 can help enhance the effectiveness of endolysin for CDI treatment by reducing sporulation and toxin production [140]. Additionally, the use of this fusion protein showed a lower minimum inhibitory concentration (MIC) than conventional antibiotics in vitro. Therefore, the use of genetic engineering method would be a valuable tool to improve the function of endolysins and reduce the risk of bacterial resistance.

Recently, CRISPR-Cas gene editing systems have been applied for the genetic engineering of bacteriophages. In this regard, phiCD24-2 was engineered to bear a genome that targets CRISPR-Cas3 found in C. difficile, and successfully converted lysogenic phages to lytic forms [143]. Although there was no difference between the engineered phage and wild-type phage in terms of host range or phage morphology, the engineered phage showed higher efficiency in controlling C. difficile than wild-type phage in both in vitro and in vivo models, indicating the superiority of the engineered phage for CDI treatment.

In addition to bacteriophages, PTLPs are a promising and potential alternative therapy for treating CDI. PTLPs are morphologically similar to bacteriophages, and thus may have a bacteriophage origin. Although there is no evidence for existing viral genome in PTLPs, they can kill their bacterial hosts and prevent the release of bacterial toxins. The specific lytic activity of PTLPs against C. difficile has been investigated in some studies [135, 147]. For example, Sangster et al. demonstrated that PTLPs isolated from the C. difficile RT078 strain can lyse various RT027 isolates, but did not display activity against other strains [135]. Therefore, PTLPs may have a broader host range than phages, but a narrower host range than endolysins. However, further research is necessary in order to apply any of these therapeutic tools in clinical practice.

The application of FVT for rCDI treatment has been considered in recent years. Recently, some studies have shown that treating rCDI with sterile fecal filtrate obtained from a healthy donor can alleviate the symptoms of rCDI. A pilot study successfully treated 5 of 5 rCDI patients using FVT [18]. Another study demonstrated that the use of lyophilized sterile filtrate and lyophilized donor stool can successfully help the treatment of 75% (3/4) and 80% (4/5) of the rCDI patients, respectively [144]. These results suggest that the phageome composition can play a key role in restoring the gut microbiota following FVT, and can be considered as a safer refinement than FMT [18, 19, 124]. Additionally, the application of mixed virome from several healthy donors may increase the effectiveness of FVT because of targeting a larger fraction of the recipient gut bacteria than a single virome [19]. However, the current understanding of the virome community transferred by FVT and insights about the precise interactions of these components with the gut microbiota are limited and require additional in-depth studies.