Therapeutic potential of Bacillus phage lysin PlyB in ocular infections

ABSTRACT Bacteriophage lytic enzymes (i.e., phage lysins) are a trending alternative for general antibiotics to combat growing antimicrobial resistance. Gram-positive Bacillus cereus causes one of the most severe forms of intraocular infection, often resulting in complete vision loss. It is an inherently β-lactamase-resistant organism that is highly inflammogenic in the eye, and antibiotics are not often beneficial as the sole therapeutic option for these blinding infections. The use of phage lysins as a treatment for B. cereus ocular infection has never been tested or reported. In this study, the phage lysin PlyB was tested in vitro, demonstrating rapid killing of vegetative B. cereus but not its spores. PlyB was also highly group specific and effectively killed the bacteria in various bacterial growth conditions, including ex vivo rabbit vitreous (Vit). Furthermore, PlyB demonstrated no cytotoxic or hemolytic activity toward human retinal cells or erythrocytes and did not trigger innate activation. In in vivo therapeutic experiments, PlyB was effective in killing B. cereus when administered intravitreally in an experimental endophthalmitis model and topically in an experimental keratitis model. In both models of ocular infection, the effective bactericidal property of PlyB prevented pathological damage to ocular tissues. Thus, PlyB was found to be safe and effective in killing B. cereus in the eye, greatly improving an otherwise devastating outcome. Overall, this study demonstrates that PlyB is a promising therapeutic option for B. cereus eye infections. IMPORTANCE Eye infections from antibiotic-resistant Bacillus cereus are devastating and can result in blindness with few available treatment options. Bacteriophage lysins are an alternative to conventional antibiotics with the potential to control antibiotic-resistant bacteria. This study demonstrates that a lysin called PlyB can effectively kill B. cereus in two models of B. cereus eye infections, thus treating and preventing the blinding effects of these infections.

B acteriophages, or "phages, " are viruses that infect bacteria, replicate, and continue their life cycle (1,2). These bacterial parasites attach to their host cell surface, insert their DNA or RNA, and manipulate the cellular machinery for subsequent replication of viral nucleic acid and synthesis of phage proteins (3). Phage proteins self-assemble into progeny phages which exit from the host bacterium after the complete replication cycle (4,5). To exit the host bacterium, phages use lytic enzymes, or lysins, that disrupt the host cell wall through cleavage of one of the four major bonds in bacterial cell wall peptido glycan, resulting in hypotonic lysis and phage progeny release (6,7). Purified lysins, when added externally to Gram-positive bacteria, result in a similar immediate hypotonic killing. This principle of lysin-mediated bacterial cell wall disruption has been adopted to kill bacterial pathogens and is now an effective tool to fight against antibiotic-resistant bacterial pathogens (6,8,9). which they were produced (8). Phage lysins have been studied as a possible therapeutic option for Pseudomonas keratitis and Staphylococcus aureus endophthalmitis in mice, with promising outcomes and few side effects (44,45). The use of phage lysin as a treatment option for B. cereus ocular infection has not been tested or reported. Here, for the first time, we demonstrate the bactericidal activity of B. cereus-derived phage lysin (PlyB) in two ocular infection models. Our data demonstrate that PlyB has potent bactericidal properties, is not toxic or inflammogenic, and can clear B. cereus from infected eyes, improving the visual outcome of disease.
PlyB activity in different growth conditions was also assessed by turbidity assay. An overnight B. cereus culture was centrifuged, washed, and resuspended in BHI, Luria-Ber tani (LB; Sigma, St Louis, MO, USA), Dulbecco's Modified Eagle's Medium (DMEM; Gibco, Grand Island, NY, USA), Vit (PelFreez, Rogers, AR, USA), or human plasma-like media (Gibco) to an OD 600 of 2.00. An amount of 50 µL of the PBS, BHI, LB, DMEM, Vit, and plasma suspensions was plated in triplicate in a U-bottomed 96-well plate. An amount of 50 µL B. cereus phage lysin PlyB (250 µg/mL) was added to each well to a final concentration of 125 µg/mL and an OD 600 of 1.00. Plates were then incubated, and OD 600 was measured as above.

Preparation of B. cereus spores
B. cereus ATCC 14579 was streak plated on a BHI agar plate and incubated overnight at 37°C. A single colony from the plate was inoculated into 25 mL Difco Sporulation Medium (DSM). Bacteria were incubated at 37°C on a shaker at 150 rpm until midlog phase (approximately 2 hours). At log phase, B. cereus was diluted 1:10 into 250 mL of prewarmed (37°C) DSM in a 2-L flask. The cultures were incubated for 7-8 days at 37°C and 150 rpm. Culture growth was observed every 48 hours. At the end of 8 days, cultures contained 90% free spores. Cultures were then centrifuged at 10,000×g for 10 minutes and washed three times with 200 mL of cold (4°C), sterile distilled water. Pellets were resuspended in 200 mL cold distilled water and stored at 4°C overnight. The following day, resuspended cultures were washed three more times and finally resuspend in 2.0 mL of cold distilled water (54). To quantify spores, an aliquot of resuspended culture was heated at 65°C for 25 minutes, serially diluted, and plated onto BHI agar. Colony forming units (CFUs) were determined, and percent of sporulation was calculated as (number of heat resistant spores/number of total cells) × 100 (55).

Bactericidal activity of PlyB
B. cereus ATCC 14579, a B. cereus ocular isolate, B. thuringiensis, B. subtilis, B. megaterium, S. aureus, E. faecalis, and S. pneumoniae were each grown overnight, back-diluted 1:100, and grown to log phase. B. cereus ATCC 14579 was also grown to stationary phase. B. cereus spores were prepared as described above. The cultures were then centrifuged, washed, and resuspended in PBS. Log phase, stationary phase, spores of B. cereus, and other bacteria listed above (1 × 10 5 CFU) were then incubated with 125 µg/mL PlyB or PBS at 37°C for 60 minutes (53). After incubation, viable bacteria were quantified by serial dilution and plating.

Cell cytotoxicity
A cell cytotoxicity assay was performed on immortalized human retinal Muller cells (MIO-M1; a kind gift from Dr. Astrid Limb, UCL Institute of Ophthalmology, London) and human retinal pigment epithelial cells (ARPE-19; American Type Culture Collection, Manassas, VA, USA). Cells were maintained in DMEM/F-12 (Gibco), supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich Corp., St. Louis, MO, USA) and 1% penicil lin-streptomycin (Gibco) in a humidified 5% CO 2 incubator at 37°C. A Pierce Lactate Dehydrogenase (LDH) Cytotoxicity Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA) was used to determine the cytotoxicity of phage lysin PlyB. Cells were seeded (20,000 cells/100 µL) in triplicate. Phage lysin PlyB (420 µg/mL), filter-sterilized B. cereus supernatant (Sup), antibiotic gatifloxacin [GAT; 250 µg/mL (Gatifloxacin Ophthalmic Solution, Alcon Laboratories, Fort Worth, TX, USA)], and controls were added to the respective wells. Cells were then incubated for 45 minutes at 37°C in 5% CO 2 . An amount of 50 µL of each sample medium was transferred to a 96-well flat bottom plate in triplicate wells, reaction mixture was added, and the plate was incubated for 30 minutes at room temperature protected from light. The reaction was stopped by adding stop solution, and OD was measured spectrophotometrically at 490 and 680 nm using a FLUOstar Omega microplate spectrophotometer (BMG Labtech, Cary, NC, USA). Percentage of cytotoxicity was calculated by subtracting the LDH activity of negative control from sample LDH activity, divided by the total LDH activity (positive control − negative control), and multiplied by 100 (56).

Experimental B. cereus endophthalmitis
All in vivo experiments were performed with C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME, USA) following the strict guidelines and recommendations of the Guide for the Care and Use of Laboratory Animals, the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and the University of Oklahoma Health Sciences Center Institutional Animal Care and Use Committee (approved protocol 20-047). Mice were 8-10 weeks of age at the time of the experiments and were housed as described above. A combination of ketamine (85 mg/kg body weight; Ketathesia, Covetrus, Dublin, OH, USA) and xylazine (14 mg/kg body weight; AnaSed, Akorn Inc., Decatur, IL, USA) was used to sedate the mice. Mice were infected with 100 CFU B. cereus/0.5 µL BHI into the right eye using a sterile glass capillary needle, as previously described (50). At 2 hours post-infection, mice were anesthetized with isoflurane and treated with 1 µL (420 µg/mL) PlyB or 1 µL (250 µg/mL) GAT. PlyB was used directly from its stock in sterile PBS, whereas GAT was diluted in sterile PBS prior to use. A group of infected mice was left untreated to serve as a control. At 2, 4, 6, 8, and 10 hours post-infection, infected and treated eyes were harvested for quantitation of viable intraocular bacteria and analysis of ocular architecture by histology, as described below.

Experimental B. cereus keratitis
As proof of concept for treatment of corneal infections, we also tested the effectiveness of PlyB in a mouse experimental B. cereus keratitis model. C57BL/6J mice were sedated using ketamine and xylazine as previously described. After scratching the cornea with a 20-G needle, approximately 10 6 CFU B. cereus in 10 µL BHI were pipetted onto the eye. After 15 minutes, excess liquid was removed by blotting with a tissue. At 2 hours post-infection, infected eyes were treated with either 10 µL (420 µg/mL) PlyB or 10 µL (250 µg/ml) GAT. PlyB was utilized directly from its stock in sterile PBS, whereas GAT was diluted in sterile PBS before being used. A total of 5 PlyB and GAT treatments were performed at 1-hour intervals. A group of infected mice was left untreated to serve as a control. One hour after the final treatment (i.e., at 7 hours post-infection), infected and treated eyes were harvested for quantitation of viable intraocular bacteria and analysis of ocular architecture by histology, as described below.

Ocular bacterial quantitation and histology
For quantitation of viable B. cereus, harvested, infected, and/or treated eyes from euthanized mice were homogenized in 400 µL PBS with sterile 1-mm glass beads (BioSpec Products, Inc., Bartlesville, OK, USA). An amount of 20 µL of each eye homoge nate was track diluted 10-fold in PBS. Dilutions and the remaining homogenates were plated onto BHI agar plates for quantitation (50). The limit of detection for quantifying B. cereus in eye homogenates was 1 CFU.
For histology, infected and/or treated eyes were harvested from euthanized mice, incubated in high-alcoholic fixative for 4 hours, and then transferred to 70% ethanol. Eyes were embedded in paraffin, sectioned, and stained with hematoxylin and eosin (33).

Statistics
Statistical analysis was performed by GraphPad Prism 9 (GraphPad Software, Inc., La Jolla, CA, USA). The Mann-Whitney test was used for statistical comparisons, unless otherwise specified. Differences between groups were taken to be statistically significant at P value of <0.05 (51,56).

PlyB is active against B. cereus ATCC 14579
Phage lysins from Gram-positive bacteriophage like PlyB have an N-terminal catalytic domain and a C-terminal binding domain ( Fig. 1A) (17,59). The binding domain constitutes about half the protein and is responsible for binding the lysin to its cell wall target. The catalytic domain possesses a hydrolytic activity (muramidase, amidase, endopeptidase, or glucosaminidase) to cleave one of the four major bonds essential for peptidoglycan integrity, resulting in bacterial hypotonic lysis (14). Phage lysins typically carry one of the indicated enzymatic activities (Fig. 1A); PlyB is a muramidase based on sequence homology (60). To evaluate lytic activity in vitro, various concentrations of PlyB were incubated with B. cereus ATCC 14579, and bacterial turbidity (OD 600 ) was monitored over time. The lytic activity of PlyB on B. cereus was observed by a gradual decrease in turbidity of the corresponding PlyB concentration in microgram per milliliter (Fig. 1B), suggesting that the lytic activity of PlyB functions in a dose-dependent manner. Within 30 seconds, turbidity was reduced by approximately 63% at the 125 µg/mL concentration. At this same concentration of PlyB, turbidity was reduced by 96% within 10 minutes and 99.9% within 60 minutes, suggesting near-complete lysis of the bacterial cells. We next investigated the bactericidal activity of PlyB toward B. cereus ( Fig. 2A). Within 2.5 minutes, B. cereus CFUs were reduced >2 log 10 , and within 5 minutes, CFUs were reduced >5 log 10 .
The composition of bacterial cell walls varies with their life cycle and may impact the effectiveness of antimicrobials (61). Since B. cereus can exist in the environment as a vegetative bacillus or an inactive spore, we compared PlyB activity against different growth phases of B. cereus (Fig. 2B). Logarithmic phase, stationary phase, and spore forms of B. cereus were incubated with 125 µg/mL PlyB for 60 minutes and plated for viability. Logarithmic and stationary phase B. cereus were reduced by >5 log 10 with PlyB. However, we observed <1 log 10 reduction in spore viability after PlyB treatment. These results suggest that PlyB was highly active against vegetative forms of B. cereus but was not effective against spores. Together, these results suggest that PlyB is a rapid bactericidal antimicrobial against B. cereus.  Fig. 3). The viability of B. subtilis was reduced to only 0.7 log 10 (P = 0.1032, Fig. 3), while the effect of PlyB on the viability of B. megaterium, S. aureus, E. faecalis, and S. pneumoniae was minimal (Fig. 3). Together, these results suggest that the target specificity of PlyB was limited to members of the BCSL group.

PlyB is bactericidal under different physiological conditions
We tested the activity of PlyB in different media, including vitreous humor, representing an ex vivo ocular environment. Harvested bacteria were resuspended in PBS, BHI, LB, DMEM, rabbit vitreous, or human plasma-like medium (plasma) and incubated with 125 µg/mL PlyB or PBS only for 10 minutes. Incubation with PlyB resulted in significant reductions of the OD 600 in all tested conditions compared with controls without PlyB (P = 0.0079). The OD 600 of PlyB-incubated B. cereus in PBS, BHI, LB, DMEM, and rabbit vitreous was similar (P > 0.05, Fig. 4). However, compared with PlyB incubation in PBS, the OD 600 of PlyB-incubated B. cereus in human plasma-like medium was somewhat greater, suggesting a slightly weaker but effective bactericidal activity of PlyB against B. cereus under plasma-like conditions. These results indicate that PlyB is active in several

PlyB is not cytotoxic
We tested the safety of PlyB by analyzing potential cytotoxic and hemolytic activities against human retinal cells and sheep red blood cells. Using an LDH assay, we compared the cytotoxicity of PlyB, GAT (negative control), and the overnight culture Sup of B. cereus (positive control) on retinal pigment epithelial cells (ARPE-19) and human retinal Muller cells (MIO-M1). PlyB and GAT exhibited no cytotoxicity, comparable with the negative control (P > 0.05), whereas the B. cereus overnight Sup was significantly cytotoxic to both RPE and Muller cells (P = 0.0022) as expected ( Fig. 5A and B). When we tested the hemolytic activity of PlyB using sheep erythrocytes, we found that the 18 hours superna tants of B. cereus were highly hemolytic, whereas PlyB and GAT exhibited no hemolytic activity (Fig. 5C). Together, these results suggest that PlyB would be safe to use in an ocular environment, as it lacks measurable cytotoxic and hemolytic activity under these in vitro conditions.

PlyB does not activate TLR2 and TLR4 innate pathways
Lysins target bacterial peptidoglycan and, depending on dose, could fragment the cell wall. Bacterial cell walls and their fragments are highly immunogenic and may trigger inflammatory responses (58). We tested whether PlyB or PlyB-incubated bacteria activated TLR2 and TLR4 in hTLR2 or hTLR4 reporter cell line assays. We also tested the innate pathway activation potential of GAT and GAT-incubated B. cereus, as well as purified B. cereus envelope as a positive control. PlyB alone did not activate either of the TLR pathways comparable with the negative controls (P > 0.09, Fig. 6A and B). Activation of TLR2 and TLR4 by PlyB-or GAT-incubated B. cereus was significantly lower compared with their respective positive control (P < 0.0022, Fig. 6C and D). TLR2 activation by PlyBincubated B. cereus was significantly lower compared with the TLR2 activation by GATincubated B. cereus (P = 0.0317, Fig. 6C). We found no difference in TLR4 activation between PlyB-incubated and GAT-incubated B. cereus (P = 0.0931, Fig. 6D). We also tested Research Article mSphere an oxidized phospholipid (OxPAPC) with the agonists mentioned above to further assess inhibition of TLR2 and TLR4 activation under these conditions. TLR2 activation by Pam3Csk4 (+ve TLR2 control), PlyB-incubated B. cereus, and GAT-incubated B. cereus was significantly reduced in OxPAPC-incubated groups (P < 0.0079, Fig. 6C). Similarly, TLR4 activation by LPS (+ve TLR4 control), PlyB-incubated B. cereus, and GAT-incubated B. cereus was significantly reduced in OxPAPC-incubated groups (P < 0.0022, Fig. 6D). These findings suggest that PlyB is not a TLR2 and TLR4 agonist and that TLR activation and potential inflammation caused by bacterial debris could be prevented by using a TLR inhibitor. At 2 hours post-treatment (4 hours post-infection), some bacterial growth was observed in both the control and PlyB groups, but no bacteria were isolated from the GAT-treated group. However, by 6 hours post-infection (4 hours post-treatment), a 5 log 10 reduction in bacterial count was observed in both the PlyB and GAT treatment groups. While the bacterial counts in the eyes of the untreated animals continued to rise at 6 and 8 hours post-treatment (8 and 10 hours post-infection), the bacterial counts in the PlyB-and GATtreated eyes remained undetectable (Fig. 7A). Since significant tissue damage occurs after 6 hours post-infection in this model, we analyzed eyes by histology at 8 hours post-treatment (Fig. 7B). Untreated eyes at this time were severely inflamed with partially detached retinas. Untreated eyes and PlyB-or GAT-treated eyes presented with significant fibrin accumulation in the anterior chamber. The posterior compartments of the untreated eyes were severely inflamed compared with that of PlyB-and GAT-treated eyes. PlyB-and GAT-treated eyes also had intact retinas with distinct retinal layers comparable with that of uninfected eyes. A closer examination showed the presence of B. cereus and inflammatory cells near the optic nerve area and in the midvitreous of the untreated eyes, which were absent in both treatment groups.

PlyB significantly reduced B. cereus in endophthalmitis and keratitis models, preventing infection-associated damage
Next, we investigated whether phage lysin could be an effective alternative for treating keratitis, which, in severe cases, requires multiple instances of topical administra tion of drugs. Keratitis was established in C57BL/6J mice by topical administration of 10 µL of 10 6 CFU B. cereus onto each scratched cornea. After 2 hours post-infection, eyes were treated with a single drop of PlyB or were left untreated. At 8 hours post-treatment, eyes were harvested, and bacterial load was determined. These experiments revealed that a single topical administration of PlyB was unable to reduce the bacterial load (data Research Article mSphere not shown), suggesting a need for repeated administration similar to routine clinical practices used to treat bacterial keratitis. Therefore, starting at 2 hours post-corneal infection, eyes were treated hourly with either PlyB or GAT for a total of five treatments. After 7 hours post-infection, eyes were harvested, and bacterial load was determined. B. cereus CFUs in PlyB-or GAT-treated eyes were reduced by >4 log 10 and >5 log 10 , respectively, compared with untreated eyes (Fig. 8A, P < 0.0022). No statistical difference was seen between the PlyB-and GAT-treated groups (Fig. 8A, P = 0.1818). Histological analysis revealed B. cereus in ulcerations within the corneal epithelium and stroma in untreated eyes, whereas bacteria were absent in sections of eyes treated with PlyB or GAT (Fig. 8B). The anterior chambers of untreated infected eyes had significant fibrin deposits compared with that of uninfected eyes. Fibrin deposition appeared to be greater in GAT-treated eyes compared with PlyB-treated eyes (Fig. 8B). Corneal epithelial layers in the untreated eyes were damaged and absent in some areas. In PlyB-and GAT-treated eyes, the corneal epithelial layers were relatively intact, but some stromal edema was observed (Fig. 8B). Taken together, these results demonstrated that PlyB lysin alone was effective in killing B. cereus in both the endophthalmitis and keratitis models and was as effective as GAT when administered intravitreally or topically in these models, respectively.

DISCUSSION
Humankind is facing a critical period of returning to the preantibiotic era due to the emergence of pathogenic bacteria resistant to most, if not all, currently available antibiotics (63). This emerging threat has elevated the pursuit of alternative anti-infection modalities as a top priority of medicine and biotechnology (64,65). Recombinant phage lysins gained momentum in the early 21st century as an alternative to commercially available antibiotics (6,14). New lysins against Gram-positive pathogens have been identified and purified, and nanogram quantities of lysin kill bacteria seconds after contact (9,14). No known biological compounds, except chemical agents, kill bacteria as quickly (16). This advancement of a new therapeutic alternative to fighting bacterial and (B) human retinal Muller cells (MIO-M1). PlyB was not cytotoxic to these cells. Data represent the mean ± SEM of percent of cytotoxicity for N >5 samples, *P = 0.0022, ns >0.05. (C) PlyB and GAT were tested for hemolytic activity on erythrocytes, and both were non-hemolytic (two-way analysis of variance, P > 0.05). .

Research Article mSphere
infection has been effective in a number of animal models (8,18,53,66). However, the potential use of phage lysin as an alternative treatment option for ocular infection is very recent, and there has been a scarcity of studies in this field.
The accidental presence of B. cereus in the ocular environment can be devastating for a patient's vision and quality of life (33). Due to its rapid progression compared with other ocular pathogens, the time to intervene therapeutically in a case of B. cereus endophthalmitis is short. Therefore, finding an appropriate antimicrobial alternative that works quickly, especially on drug-resistant microbes, has always been a high priority to prevent vision loss. Here, we tested the bactericidal effectiveness of a novel B. cereus phage lysin (PlyB) under laboratory conditions and two different murine ocular infection that result in rapid vision loss (33,38,47,68,69). We first tested the effectiveness of PlyB in vitro and observed a >5 log 10 reduction of CFUs in 5 minutes, verifying previous studies and suggesting a very potent killing activity of PlyB against B. cereus. B. cereus is a spore-forming bacterium and can exist in various forms in its life cycle (62). In the lag phase, B. cereus adapts to its environment, and cells have been shown to increase in volume (70). During logarithmic phase, B. cereus divides exponentially and is most susceptible to antibiotic activities. During stationary phase, the cell wall is highly cross-linked, which reduces membrane fluidity. During sporulation, bacteria enter an intricate phase of cell differentiation events that leads to the formation of a dormant spore, which is resistant to harsh environments and certain antimicrobials (70,71). The cell-wall-binding domain of PlyG has been reported to reach its substrate in the peptidoglycan only during germination of B. anthracis spores and not the spores alone. As such, PlyG kills B. anthracis vegetative cells and germinating spores (72). In the current study, PlyB reduced both logarithmic phase and stationary phase B. cereus by >5 log 10 CFU but did not reduce the number of nongerminating spores. This suggests that PlyB is active against replicating on metabolically active B. cereus when the cell wall is exposed but not on metabolically inactive spores, where the spore coat blocks its access to the peptidoglycan.
Commercially available antibiotics have a broad spectrum of activity and can also disrupt the normal microbial flora (73). One benefit of phage lysins over antibiotics is their genera or species specificity (15, 74); however, lysins with broader activity have been reported. PlySs2 developed from a S. suis phage has activity against a wide range of Gram-positive pathogens and in vivo efficacy against methicillin-resistant S. aureus and S. pyogenes (75,76). A lysin from P. aeruginosa demonstrated substantial bactericidal activity against Pseudomonas, Klebsiella, Enterobacter, and other Gram-negative bacterial strains in vitro (53). Here, PlyB demonstrated high bactericidal activity against B. cereus and B. thuringiensis, causing a reduction of more than 5 log 10 . However, it did not exhibit bactericidal effects against B. subtilis, B. megaterium, S. aureus, E. faecalis, and S. pneumoniae, suggesting that PlyB is highly specific toward members of the BCSL group.
The intraocular environment is pH neutral and contains approximately 98%-99% water with trace amounts of hyaluronic acid, glucose, anions, cations, ions, and collagen (77). Both aqueous and vitreous humor form an excellent growth medium for some microorganisms, but only when an appropriate starting inoculum is present (52,78,79). Phage lysins are relatively stable and active in a wide range of pH, salt, and urea concentrations. With few exceptions, the lytic activity of phage lysins directed to most Gram-positive bacteria is retained independent of the culture medium (6,14,17). Except for PlySs2, a phage lysin active against S. aureus which has successfully completed phase 2 human trials (19), no phage lysin has yet been commercially ready for human use. It is essential to determine bactericidal and cytotoxic properties prior to use in a particu lar disease condition (80). PlyB showed bactericidal activity in different culture media, including ex vivo rabbit vitreous and human plasma-like medium as mimics for the intraocular and bloodstream environments, respectively. PlyB did not lyse erythrocytes and was not cytotoxic toward human retinal Muller cells and retinal pigment epithelial cells. Together, these results suggest that PlyB would not be toxic in an intraocular environment if used as a therapeutic. Because the eye is an immune-privileged site, foreign antigens can trigger an immune response that could interfere with visual systems (43). Lysins are proteins that could stimulate an immune response when administrated via mucosal, intravenous, or ocular routes (6,14). In the current study, PlyB did not activate TLR2 or TLR4, TLRs that impact B. cereus endophthalmitis pathogenesis (48,49). This suggests that PlyB would not trigger inflammation in vivo by those pathways. Compared with antibiotics that target bacterial cell walls or replication machinery, phage lysins specifically target the integrity of the bacterial cell wall (6,9,14,16,17,66). Antibiotic-treated fragmented bacteria release highly inflammatory pathogen-associated molecular patterns (81). Antibiotic-induced inflammation is a general phenomenon in animal models of otitis media and meningi tis (82,83), and the antibiotic-induced systemic release of bacterial components may initiate an inflammatory cascade, resulting in septic shock and multiple organ failure (84). The B. cereus cell wall is highly inflammogenic and its components contribute to its intraocular virulence (23,34,49,58). GAT interrupts bacterial DNA replication, rapidly kills bacteria, and is a widely used broad-spectrum antibiotic approved for ocular use (85). Since antibiotics and phage lysin kill the pathogen, leaving the dead cell either intact or fragmented, we investigated the immune activation of PlyB-incubated and GAT-incubated B. cereus. TLR2/4 activation by PlyB-incubated B. cereus was less than that induced by GAT-incubated B. cereus. Also, innate activation by PlyB-incubated or GAT-incubated B. cereus could be controlled by inhibiting TLR2/4 activation with OxPAPC. In this regard, we previously reported that OxPAPC reduced the TLR2-and TLR4-driven inflammatory responses and disease severity during Bacillus endophthalmitis (34).
The therapeutic efficacy of phage lysin has been studied in animal models of pneumonia, endocarditis, and sepsis (6,(86)(87)(88). Endophthalmitis and keratitis are two dangerous bacterial infections of the eye caused primarily by Gram-positive pathogens (20). B. cereus endophthalmitis requires intravitreal administration of antibiotics to kill the bacteria (33), while the treatment of B. cereus keratitis requires the administration of topical antibiotics (28). Bacteriophage suspensions have been used topically to treat staphylococcal conjunctivitis and blepharitis, and although better visual outcomes were reported, no details on bacterial load, pathology, or visual function were reported (89)(90)(91)(92). Bacteriophages have also been used to treat patients with traumatic bacterial keratitis and purulent corneal ulcers, with a more rapid improvement in inflammation and pain than patients treated with gentamicin (90,93). The study of phage lysin efficacy in ocular infections has been limited. S. pneumoniae endolysin (MSlys) rapidly killed ocular S. pneumoniae strains in vitro regardless of the isolation source, strain genotypes, and encapsulation status; however, its effectiveness in vivo was not reported (94). A chimeric endolysin derived from the Ply187 prophage showed potent antimicro bial activity in a mouse model of S. aureus endophthalmitis (45). B. cereus intraocular infections can be more severe than infections caused by other ocular pathogens due to its rapid growth, which ultimately compromises vision and endangers the structure of the globe (33). Here, PlyB had potent bactericidal activity and sterilized B. cereus-infec ted mouse eyes when administrated intravitreally during the early stages of infection. A single dose of PlyB reduced the B. cereus load by >5 log 10 relatively quickly. PlyB treatment of B. cereus endophthalmitis preserved ocular structures, and distinct retinal and corneal layers were maintained. PlyB was also a potent bactericidal agent when administered topically to treat experimental B. cereus keratitis. In contrast to a single administration of PlyB to treat endophthalmitis, hourly administration of PlyB was required to significantly reduce the bacterial load in the B. cereus keratitis model. In both infection models, cellular infiltration was not prevented, but inflammation was reduced compared with that observed in untreated eyes. Our results demonstrated that the use of PlyB in treating B. cereus endophthalmitis and keratitis was as potent as GAT in killing B. cereus and greatly improved the clinical outcome of both infections.
After the first report in 2001 (95), numerous studies have demonstrated the potential of phage lysin to treat different bacterial infections in vivo. Until recently, ocular infection was missing from the scope of lysin's therapeutic potential. Since antibiotics do not discriminate between pathogenic and commensal organisms, alternative therapeutics are warranted to minimize collateral damage to the protective microbiota. To date, no harmful side effects have been reported following topical or systemic administration of phage lysin in preclinical trials in vivo and in phase 1 and 2 human trials of S. aureus bacteremia and endocarditis (19). PlyB was a very potent bactericidal agent against members of the BCSL group in laboratory conditions and against B. cereus in two different ocular infection models, indicating its potential effectiveness. Since B. cereus ocular infection is a rapidly developing multifactorial event and is often refractory to available treatment options (33), a synergistic treatment approach is needed to kill the pathogen and prevent damaging inflammation and vision loss. Further investigation is required to demonstrate the effectiveness of phage lysins in therapeutic modalities that include antibiotics and anti-inflammatory drugs in treating blinding bacterial ocular infections, especially those caused by MDR pathogens.