Bacteriophage endolysins as a potential weapon to combat Clostridioides difficile infection

ABSTRACT Clostridioides difficile is the leading cause of health-care-associated infection throughout the developed world and contributes significantly to patient morbidity and mortality. Typically, antibiotics are used for the primary treatment of C. difficile infections (CDIs), but they are not universally effective for all ribotypes and can result in antibiotic resistance and recurrent infection, while also disrupting the microbiota. Novel targeted therapeutics are urgently needed to combat CDI. Bacteriophage-derived endolysins are required to disrupt the bacterial cell wall of their target bacteria and are possible alternatives to antibiotics. These lytic proteins could potentially replace or augment antibiotics in CDI treatment. We discuss candidate therapeutic lysins derived from phages/prophages of C. difficile and their potential as antimicrobials against CDI. Additionally, we review the antibacterial potential of some recently identified homologues of C. difficile endolysins. Finally, the challenges of endolysins are considered with respect to the development of novel lysin-based therapies.


Introduction
Clostridioides difficile (formerly Clostridium difficile) is a Gram-positive, spore-forming obligate anaerobic bacillus, and the causative agent of healthcareassociated (HA) infectious diarrhea. 1 In recent decades C. difficile associated infection has been associated with high morbidity and mortality, in particular in Europe, the USA, Canada, and Australia. 2 Some reports recently also describe C. difficile infection (CDI) in Asia. 3,4 There are nearly 462,100 CDI cases annually in the United States and at least 12,800 fatalities. 5,6 In addition to loss of life, the treatment and management costs of CDI infection are significant, with an estimated annual cost of 800 USD million in the USA and €3,000 million in Europe. 7 Although CDI cases in the United Kingdom decreased from 55,498 to 12,275 between 2007 and 2018, 8 enormous effort has been put into CDI management strategies. 9 CDI is mediated by up to three toxins; toxin A (enterotoxin), toxin B (cytotoxin) and the less common binary CDT toxin. 10 These toxins damage the intestinal epithelium cell layer and activate the host inflammatory response that contributes to the disease pathology. The clinical symptoms range from mild to severe diarrhea and, to some instances, extend to potentially life-threatening conditions such as pseudomembranous colitis and toxic megacolon. 11 The most common treatment strategy in CDI is to administer antibiotics such as metronidazole, vancomycin, and fidaxomicin. However, recurrence of infection and treatment failure can occur. 12 Recent reports suggest there is reduced susceptibility and resistance of C. difficile toward these antibiotics. 13,14 This has driven the exploration of alternative therapies to treat infections caused by this bacterium. Several treatment alternatives have gained traction and are at various stages of development; these include new antibiotics, probiotics, fecal microbiota transplantation (FMT), antimicrobial peptides, bacteriocins and phage therapy. [15][16][17][18] The accessibility of phage genome databases and increased sequencing of phage have kindled interest in the application of phage encoded enzymes, especially endolysins, as alternative therapeutic agents.
Much has been written on the success of endolysins as targeted antimicrobials. [19][20][21][22] A number of reviews have focused on the advantages, specificity and safety of endolysin therapy. [23][24][25][26][27][28] In this work, we aim to review the existing literature on C. difficile endolysins and discuss their potency and their potential use as antimicrobials in the treatment of CDI.

Antibiotic resistance in Clostridioides difficile
Generally, CDI is initiated following disruption of the normal intestinal microbiota by antibiotics that allow C. difficile, either native or acquired, to proliferate. 29 The recommended antibiotics for primary and recurrent CDI are metronidazole, vancomycin and fidaxomicin. 1,30 Other antibiotics commonly used for bacterial infections such as cephalosporins, ampicillin, clindamycin, amoxicillin and fluoroquinolones are also associated with a higher risk for CDI. 1,31 The evolution of new ribotypes is often associated with acquisition of resistance as a result of inappropriate use of antibiotics. 14 Many of the most common C. difficile epidemic ribotypes are associated with multidrug resistance. 14,31 However, C. difficile is a sporeforming bacteria that can survive antimicrobial therapy and following germination relapse of CDI can routinely occur.
Several mechanisms of antimicrobial resistance have been identified in C. difficile. These include chromosomal resistance-associated genes, mobile genetic elements (MGEs), alterations in the antibiotic targets of antibiotics and/or in metabolic pathways, and biofilm formation. Examples of chromosomal resistance genes include those encoding β-lactamase-like proteins and penicillinbinding proteins (PBPs) that mediate resistance to the β-lactam antibiotics such as penicillin and cephalosporins. 14 The C. difficile genome contains a wide range of mobile elements. MGE-like transposons facilitate the spread of antibiotic resistance genes by the process of conjugation, transduction, and/or transformation among C. difficile and/or between C. difficile and other bacterial species. 14 Resistance to antibiotics of the macrolidelincosamide-streptogramin B (MLSB) family, tetracycline and chloramphenicol in C. difficile is thought to be associated with different transposon families. 31 Alterations in the antibiotic targets and/ or in metabolic pathways is another important route of resistance development in C. difficile, and this mechanism mediates resistance to rifamycin, fluoroquinolones, metronidazole and vancomycin. 14,31 Biofilms help pathogenic bacteria to survive unfavorable environmental stresses, including antibiotics. 32 Biofilm formation is potentially involved in metronidazole and vancomycin resistance in C. difficile. 14,31

Bacteriophage endolysins
Bacteriophages or phages are viruses that infect and kill bacteria. Bacteriophages can adopt either of two life cycles, lytic and lysogenic. Both virulent and temperate phages may enter the lytic cycle, whereas only temperate phages utilize the lysogenic cycle. 33 In most cases, the lytic cycle concludes with cell lysis that leads to cell death ( Figure 1a). In single stranded DNA/RNA phages, the genome encodes a lysis effector which inhibits peptidoglycan (PG) biosynthesis from within the bacterium. On the other hand, double stranded DNA (dsDNA) phages utilize phage-encoded muralytic enzymes called endolysins (or lysins) that lead to cell envelope disruption at the final stage of phage reproduction. There are three different lysis mechanisms in dsDNA phages. The most studied and best understood mechanism is canonical lysis, where endolysins require the help of a second phage-encoded protein called a holin to act on the PG layer. 34 Holin proteins accumulate and oligomerize in the cytoplasmic membrane (CM) in a time-controlled manner, and trigger depolarization and the formation of holes in the CM. This allows diffusion of endolysin to the membrane, facilitating the destruction of the PG layer ( Figure 1b). The second mechanism requires a special class of holins, called pinholins, which form small, heptameric channels that help to depolarize the membrane. In the third mechanism, the lysis of the outer membrane of Gram-negative hosts is facilitated by spanin proteins. 35 Spanins form a complex with outer membrane (OM) lipoprotein (o-spanin) and an integral cytoplasmic membrane protein (i-spanin) and disrupt the OM by enzymatic degradation, pore formation and inner membrane-outer membrane fusion. [36][37][38] Endolysins are 'enzybiotics', a promising class of antibiotics derived from enzymes. Endolysins possess high specificity against the genus or species infected by the phage from which they were derived, which is believed to be one of their key advantages over classical wide-spectrum antibiotics. Thus, these lytic enzymes, specific for undesired pathogenic bacteria, rarely lyse non-target bacteria, including commensals of our microbiota or 'good' bacteria in foods (e.g. starter cultures) or those in the environment. 39,40 The White house National Action Plan for Combating Antibiotic-resistant Bacteria has listed "phage-derived lysins to kill specific bacteria while preserving the microbiota" among the nontraditional therapeutics of note. 41 It is possible to expand the lytic spectrum of an endolysin by the exchange or addition of certain domains beyond the native endolysin's serovar, species, or even genus specificity. 42 Endolysins can also act synergistically when used in combination with each other (i.e. two endolysins with different cleavage specificities) or with other antimicrobial agents. 20,43 Other advantages of endolysins over traditional antibiotics are rapid host killing, low chance of resistance development, the potential to kill multi-drug-resistant bacteria, synergism with different antibacterial agents, and the ability to actively work in biofilms as well as on mucosal surfaces. [44][45][46][47][48][49] Nowadays, the emergence of multidrug-resistant pathogens has revitalized the interest in alternative therapies. Due to the unique properties and advantages of endolysins over bacteriophages, endolysins are highly ranked alternatives in eradicating drug-resistant pathogens. Different properties such as specificity and host range, 39,50-52 mode of action, 53-55 resistance development, 56,57 stability, 58,59 and pharmacokinetics 60,61 between bacteriophage and endolysins are summarized in Table 1.

Basic structure and enzymatic activity of endolysins
Generally, endolysins have a conserved biological function directed at lysing infected bacterial cells. However, constant evolutionary pressure between bacteriophage and host bacteria has resulted in significant biochemical and structural variations among endolysins. 62 Endolysins against Grampositive and Gram-negative organisms are categorized differently due to the composition of the cell walls of their targets. Gram-positive endolysins are modular in structure with one or two N-terminal enzymatically active domains (EADs) and one or more C-terminal cell-wall binding domains (CBDs), these domains are usually connected by a short linker region. 62,63 EADs contain the catalytic mechanism of endolysins, disrupting specific bonds within the bacterial peptidoglycan, whereas CBDs bind to constituents of cell walls to promote EAD localization to its target site and enhance catalytic efficiency. However, sometimes greater lytic activity has been observed in truncated endolysins containing only EADs. 20,64 Gram-negative endolysins usually have single globular catalytic domain and lack a CBD. 65,66 There have been some reports for Gram-negative endolysins that indicate a modular organization with a CBD at the N-terminus and EAD at the C-terminus which is inverse architecture typical to most Gram-positive endolysins. 67,68 PG structure is highly conserved and consists of a polysaccharide of alternating N-acetylmuramic acid (MurNAc) and N-acetylglucosamine (GlcNAc) residues, linked by a β1-4 glycosidic bond ( Figure 2). The D-lactoyl group of each MurNAc is linked with a short peptide stem, which is different between bacterial species. 69 The tetrapeptide stem found in C. difficile is L -Ala-D -Glu-A2pm-D -Ala (A2pm: 2,6-diaminopimelic acid). 70 Endolysins can recognize and digest a specific chemical bond within PG and are classified accordingly: (i) N-acetylmuramoyl-L-alanine amidases cleave the amide bond between N-acetylmuramoyl residues and L-alanine (the first amino acid of the peptide stem), (ii) L-alanoyl-D-glutamate endopeptidases target the bond between L-alanine and D-glutamate, (iii) interpeptide bridge endopeptidases digest the cross-link between peptide stems, (iv) D-glutamyl-m-DAP endopeptidase target bonds between D-glutamate and m-diaminopimelic acid, (v) N-acetyl-D-glucosaminidases hydrolyze the N-acetylglucosaminyl-β-1,4-N-acetylmuramine bond, (vi) N-acetyl-D-muramidases and lytic transglycosylases cleave the N-acetylmuramoyl −1,4-N-acetylglucosamine bond.
The chemotype of C. difficile PG is A1γ, where a meso-diaminopimelic acid (meso-A 2 pm) residue at position 3 of the peptide is directly cross-linked to a D-alanine at position 4 of the neighboring peptide. A similar structure is present in Bacillus, Listeria and some Gram-negative species. 40,70,71 However, C. difficile also have a very high percentage of 3-3 peptide cross-links between two mesodiaminopimelic acid (meso-A 2 pm) residues 69 (Figure 2b). Endolysin activity against these peptide cross-links has not yet been verified in C. difficile.

Clostridioides difficile endolysins
The use of phage therapy in CDI is limited due to the temperate nature of C. difficile bacteriophages (reported to date). In most cases low titers of lysogenic phages have been recovered after induction with mitomycin C. These phages can easily reintegrate into the host genome following the removal of inducers. 72 In the case of reported lytic phages, these may also have access the lysogenic life cycle. 73,74 Efforts have focused on alternative options and the exploitation of C. difficile phage   CD27L against the amidase containing endolysins obtained from C. difficile phage or prophage sequences was performed. After curation of the sequence cluster to remove duplicates, the sequences of five amidases (CD27L, phyCD, phiCD38, phiCD119 and CD11) were aligned and analyzed using ESPRIPT 90 for sequence conservation. Although the sequence identity is low, several conserved residues were found. The conserved four amino acid residues (His 9, Glu 26, His 84, and Glu 144) are coordinating the zinc ion and responsible for catalytic activity (Figure 3a). The threedimensional structures of major C. difficile endolysins (CD27L, CDG, phiCD211, phiCDHM11 and phiCDMMP01) have been predicted by homology modeling using SWISS MODEL, an online tool ( Figure 3b). The potential templates for target endolysins were identified based on the sequence coverage and percentage of identity between the target and template sequence, except for CD27L whose EAD and CBD three-dimensional structure are already available (PDB code 3QAY and 4CU5). Due to the presence of different catalytic groups, the folding patterns are different in amidases, glucosaminidases and NLPC_P60 containing C. difficile endolysins.

CD27L
The endolysin CD27L is derived from C. difficile bacteriophage φCD27, a temperate bacteriophage belonging to the Myoviridae family with a genome length of 50,930 bp. 78 This was the first C. difficile endolysin to be cloned and recombinantly expressed. The CD27L endolysin is 270 amino acids with an estimated molecular weight of ~30 kDa. CD27L is a modular endolysin that contains one EAD and one CBD. It has no transmembrane regions or signal peptide regions. The catalytic domain of CD27L belongs to the N-acetylmuramoyl-L-alanine amidase class (amidase 3). This lysin is active against a panel of 30 different C. difficile strains, including the hypervirulent ribotype 027, while a selection of the commensal bacteria of the GI tract are insensitive. 78 Endolysin delivery to the gut environment is a challenge, and one approach is to use a genetically modified lactic acid bacterium. Crude protein extracts from Lactococcus lactis expressing CD27L have been shown to successfully lyse target cells to a similar degree to that observed for E. coli expressed lysin. 78 It was found that CD27L lyses cells over a broad pH range, indicating it should remain active in the GI tract environment. Furthermore, the CD27L lysin catalytic domain, CD27L 1-179 , is more effective and exhibits a higher degree of specificity than the full-length endolysin. 64 However, CD27L activity has not been tested an in vivo study.

PlyCD
The lysin PlyCD is contained within a prophage in a multidrug resistance C. difficile strain, CD630. 20 PlyCD is 262 amino acids with a molecular weight of ~28 kDa. The amino acid sequence identity between PlyCD and CD27L is 33%, with 34.6% identity between the catalytic domains. Both the full-length PlyCD and the recombinantly expressed truncated PlyCD 1-174 are active against a number of C. difficile strains. Indeed, PlyCD 1-174 has significantly greater lytic activity (>4-log kill) and a broader spectrum of activity while retaining high specificity toward C. difficile versus other commensal bacterial species. Additionally, a combination of PlyCD 1-174 with a subinhibitory dose of vancomycin was significantly more successful in vitro against C. difficile compared to PlyCD 1-174 or vancomycin alone. 20 The functional capability of PlyCD 1-174 was further confirmed by an in vivo mouse model study where the PlyCD treated mice showed increased survival and a delay in morbidity and mortality rate as compared to the buffer-treated control animals. Moreover, an ex vivo treatment model of mouse colon infection showed PlyCD 1-174 functioned effectively in the presence of intestinal contents and significantly reducing colonizing C. difficile compared to buffer control. 20

CDG and CD11
The lysin proteins CDG and CD11 were identified in C. difficile DA00211 and C. difficile 630, respectively. 89 Both lysins were expressed recombinantly and were found to be effective against C. difficile cells in a dose-dependent manner, reducing levels by up to four logs within five hours.  Homology modeling was performed by Swiss-model server (https://swissmodel.expasy.org/). The predicted model of CDG, phiCD211, phiCDHM11 and phiCDMMP01 were generated using protein data bank (PDB) templates 2L47, 5WQW, 4HPE and 4FDY, respectively.
These lytic proteins were active against clinical isolates of C. difficile, while no activity was observed against Bacillus or Staphylococcal species. The biocatalytic mechanism showed that these enzymes cleave bonds between N-acetylmuramoyl and L-alanine within the cell wall PG. 89

LHD
The modular structure of lysin proteins can facilitate the modification of bacteriolytic activity, specificity, solubility, and other physicochemical properties of these proteins in order to design novel antimicrobials. A lysin, LHD, was engineered to have the catalytic domain of a lysin protein from a C. difficile bacteriophage phiC2 fused with the functional domain of a human defensin protein HD 5 by a 3-repeating unit linker; (GGGGS) 3 . 91 The reason for choosing the catalytic domain of phiC2 is that phiC2 is present in the majority of human isolates of C. difficile 92 and it may have a wide spectrum of lytic activity. On the other hand, human defensin protein HD 5 has been documented to inhibit hypervirulent C. difficile strains. 93 So, the hybrid would be more active than these individual antimicrobials. It was found that this lysin-human defensin fusion protein was active against several clinical C. difficile strains, including the epidemic 027, 078, 012, and 087 strains that are prevalent in many different regions of world. 94 The minimum inhibitory concentration (MIC) of LHD was lower than the MIC of the lysin protein LCD. The fusion protein was also active in a broader pH range (6.0, 7.0, and 8.0). In an in vivo mouse model, the LHD treated group had reduced symptoms mortality from CDI compared to the control buffer treated group. In addition, LHD significantly decreased the C. difficile spore count and toxin production in feces of the infected mice. 91

Cell wall hydrolases
Cell wall hydrolases (CWHs) are classified based on their origin as endolysins, exolysins and/or autolysins. 95,96 Some of the C. difficile phage/prophages contain putative CWH sequences that may have potent lytic activity as endolysins ( Table 2). The major catalytic domains present in C. difficile phage/prophage CWHs are NlpC/P60 and glucosaminidase. The NlpC/P60 domain is described as a superfamily and has a diverse range of catalytic activity including cleaving N-acetylmuramate-L-alanine linkages and the 4-3 linkage between D-Glu and m-DAP residues. [97][98][99] The glucosaminadases hydrolyze the glycosidic bond of the sugar backbone. The putative N-acetylglucosaminidases (EC.3.2.1.96) that are present in C. difficile phages/prophages were also present in the prophage LambdaSa2 of Streptococcus agalactiae and exhibit β-D-N-acetylglucosaminidase activity. 100

Bacterial resistance to endolysins
Bacteriophage endolysins have a unique attribute in comparison to intact phages and antibiotics, in that resistance development is an extremely rare event. Generally, antibiotics work by inhibiting essential metabolic pathways of bacteria leading to cell death. 101 However, bacteria have found ways to overcome this adverse situation by using alternative metabolic pathways. It is difficult for bacteria to find means of resistance to endolysin as they bind to and degrade highly conserved peptidoglycan targets within the cell wall. 102 Any mutations leading to endolysin resistance would be damaging to the integrity of the cell and thus a very rare event. 103 Although no attempts have been made to study lysin-resistance development in C. difficile, there are some studies using other bacterial strains that have investigated repeated lysin exposure and revealed resistance did not develop to either native or engineered phage lysins. 40,[104][105][106] Some CWHs of C. difficile have more than one catalytic domain and this theoretically lowers the chance of mutation in multiple target sites in bacteria. Similar observations have been made with S. aureus endolysins. 107

Safety and current trials
Bacteriophages are naturally an integral part of the human microbiota and thus the release of phagederived lysins is unlikely to have a harmful effect on human health. 108 The safety of phage lysins has been confirmed in a number of animal model systems. 20,47,109,110 The impact of lytic proteins on inflammatory responses and/or their toxicity has been evaluated in animal models and it was observed that the administration of some lysin proteins, for example, Cpl-1 and MV-L, triggered an immune response which ultimately resulted in the production of antibodies against this protein. 111,112 In another study, low levels of antibodies and/or cytokine production were observed in animals compared with untreated controls following endolysin treatment. 113,114 Despite the number of animal trials published, only a few lysins have undergone human clinical trials. SAL200 is the first endolysin-based therapeutic formulation against MRSA. It is derived from Staphylococcus phage SAP-1 that infects Staphylococci, including MRSA and vancomycinresistant S. aureus (VRSA) strains. Recently, protein SAL200 was evaluated in humans by intravenous infusion as part of a phase 1 clinical trial. Single ascending intravenous doses (0.1 mg/kg to 10 mg/ kg) were applied to healthy male volunteers in order to assay pharmacokinetics, pharmacodynamics, and tolerance of SAL200. 61 No serious adverse effects as well as recurrence of infection were observed in volunteers except more than three participants noticed mild and temporary effects like fatigue, headaches and myalgia. Another endolysin-based product called Staphefekt SA.100, developed by Dutch biotech company Micreos, has been available in Europe since 2017 for human use. Staphefekt SA.100 is an engineered chimeric endolysin for topical skin application that specifically binds to the cell wall of S. aureus and cleaves the cell membrane via endopeptidase and putative amidase activities. 115 In a case study on three human subjects with chronic and recurrent S. aureus-related dermatoses showed that Staphefekt™ improved the clinical symptoms, but they rapidly recurred if the treatment was ceased. Potentially due to the recolonization of S. aureus from the nose and environment. It has also been shown that the long term daily use of Staphefekt did not result in generation of bacterial resistance during chronic and recurrent S. aureus treatment. 22 A multicenter, placebo-controlled, double-blinded and randomized superiority trial study (ClinicalTrials.gov, NCT02840955) of Staphefekt showed application on the skin, targeting only S. aureus and leaving skin commensals unharmed, improves S. aureus-related skin infections, such as eczema, acne, and rosacea. 116 Staphefekt is registered as a (class 1) medical device in Europe and available as an over-the-counter treatment in the form of a cream or gels. There are several other products at different stages of clinical trials, some with promising results that will pave the way for future endolysin-based therapies. 107,117

Challenges of endolysin therapy in CDI
There are several challenges facing the commercialization of endolysins, for instance, large-scale production and formulation, targeted delivery and regulatory framework amongst others. Until now, the use of endolysins as human therapeutics has not been approved in the United States or Europe, except for the endolysin-based product Staphefekt which is marketed in the EU as a 'medical device'. 22 Although the role of phagebased lysins is already established and examined in different animal infection models, the study of C. difficile endolysins is still limited. It may be due to some experimental limitations, particularly the development of targeted enteric delivery of such endolysins. To date, only two studies were performed where C. difficile lysins were examined in animal models for clinical development. 20,91 In both studies, the targeted delivery of endolysins to the gut was by direct administration via oral gavage. Lysins are non-replicating protein molecules, with short in-vivo half-lives. 33,118 The enteral delivery of phage lytic proteins faces the challenge of maintaining enzyme activity at low pH and in the presence of proteolytic enzymes of the stomach. To avoid such obstacles, the development of novel delivery strategies is important. Nanoparticle-based anticancer drug delivery to eukaryotic cells are becoming more popular as a treatment strategy. 119,120 It also offers possibilities to effectively target bacterial cells. 121,122 Endolysins could also be encapsulated in polymeric nanoparticles to give protection from the harsh gastric environment. Another proposed innovative approach is to deliver and preserve lysins in the gastrointestinal tract via engineered lactic acid bacteria that actively excrete the endolysin during gut transit. 123,124 The C. difficile phage endolysin CD27L has been successfully expressed in Lactococcus lactis MG1363. 78 Additionally, the probiotic strain Lactobacillus johnsonii FI9785 was precisely engineered to deliver and secrete an endolysin active against C. perfringens. 124

Conclusion
Clostridioides difficile is recognized as the leading cause of nosocomial and community-acquired diarrhea associated with exposure to antibiotics. Antibiotic resistance and a high rate of recurrence limit the usefulness of current antimicrobials used for primary C. difficile infection; alternative solutions are urgently required that work effectively while maintaining the gut microbiota. Phage lytic proteins show great potential in this regard as a replacement for, or as an additional therapy to, the use of traditional antibiotics in CDI. In fact, several endolysins active against C. difficile are currently being investigated in this regard. Major features, such as high catalytic activity, modular structure, and the possibility of engineering, support the development of these novel alternatives to conventional antibiotic therapy. The possibility of resistance development to phage lytic proteins remains a real possibility worthy of further investigation. To date, preclinical and clinical studies demonstrate that endolysins are safe and very effective antimicrobials.

Disclosure of potential conflicts of interest
No potential conflicts of interest were disclosed.