Deimmunized Lysostaphin Synergizes with Small-Molecule Chemotherapies and Resensitizes Methicillin-Resistant Staphylococcus aureus to β-Lactam Antibiotics

There is an urgent need for novel agents to treat drug-resistant bacterial infections, such as multidrug-resistant Staphylococcus aureus (MRSA). Desirable properties for new antibiotics include high potency, narrow species selectivity, low propensity to elicit new resistance phenotypes, and synergy with standard-of-care (SOC) chemotherapies.

ABSTRACT

There is an urgent need for novel agents to treat drug-resistant bacterial infections, such as multidrug-resistant Staphylococcus aureus (MRSA). Desirable properties for new antibiotics include high potency, narrow species selectivity, low propensity to elicit new resistance phenotypes, and synergy with standard-of-care (SOC) chemotherapies. Here, we describe analysis of the antibacterial potential exhibited by F12, an innovative anti-MRSA lysin that has been genetically engineered to evade detrimental antidrug immune responses in human patients. F12 possesses high potency and rapid onset of action, it has narrow selectivity against pathogenic staphylococci, and it manifests synergy with numerous SOC antibiotics. Additionally, resistance to F12 and β-lactam antibiotics appears mutually exclusive, and, importantly, we provide evidence that F12 resensitizes normally resistant MRSA strains to β-lactams both in vitro and in vivo. These results suggest that combinations of F12 and SOC antibiotics are a promising new approach to treating refractory S. aureus infections.

INTRODUCTION

The threat posed by antibiotic-resistant bacteria is continuing to grow (1), and predictions for the impact on the future of global health are dire (2). In the United States alone, there are almost three million drug-resistant bacterial infections and 36,000 deaths each year (3). Methicillin-resistant Staphylococcus aureus (MRSA) remains the single most dangerous of these pathogens, causing at least 13,000 annual deaths in the United States. Both hospital- and community-acquired MRSA are prevalent, and this prolific bacterium can manifest a variety of infections, including superficial skin and soft-tissue infections (4), sepsis (5), endocarditis (6), pneumonias (7), osteomyelitis, and more (8, 9). During the second half of the 20th century, continuous development of new antibacterial chemotherapies kept MRSA largely in check, but in almost every case MRSA developed resistance shortly after each drug’s introduction (10). This trend ultimately led to the current crisis point (11). Currently, vancomycin, daptomycin, and linezolid represent standard-of-care (SOC) treatments for invasive MRSA infections (12, 13), but even these powerful drugs are now being undermined by the evolution of resistance in S. aureus (14–19). As a result, there is an urgent need for alternative approaches to combat the growing threat MRSA poses to public health.

Lysostaphin (LST) is an antibacterial enzyme with exceptional potency against S. aureus both in vitro and in vivo (20). LST was cloned more than 30 years ago (21), and it has an archetypical lysin structure composed of a cell wall binding domain, which localizes the enzyme to the S. aureus cell wall, and a catalytic domain, which cleaves the cross-linking pentaglycine bridge in S. aureus peptidoglycan. The hydrolysis of glycine-peptide bonds in the cell wall leads to rapid bacterial lysis via osmotic stress (22). Although LST exhibits extraordinary antibacterial properties, it is highly immunogenic in both animals and human subjects (20, 23–25). Anti-LST antibodies are suspected of driving immune complex-associated toxicity in preclinical models (20), and, more generally, antidrug antibodies can undermine biotherapeutic efficacy and threaten patient safety (26). To capitalize on LST’s full therapeutic potential, there is a need to mitigate its inherent immunogenicity (27, 28).

While chemical modification with polyethylene glycol (PEG) is a common strategy to shield biotherapeutics from detrimental immune responses, PEGylation has been shown to compromise LST antibacterial activity (20), and PEGylation is more generally thought to be incompatible with lysin biotherapeutic function (29). As an alternative approach, we have successfully deleted constituent LST T cell epitopes via mutagenesis, yielding the variant F12, which bears 14 amino acid substitutions. We reported previously that F12 silences human T cell activation in ex vivo immunoassays and dampens the antidrug antibody response in human HLA transgenic mice (30). In the same report, F12’s low immunogenicity was further shown to contribute to enhanced anti-MRSA efficacy relative to the wild-type LST counterpart.

These prior results demonstrate that F12 achieved the dual objectives of mitigating antidrug immune surveillance while maintaining high-level anti-MRSA activity. However, the most direct translational path for F12, or any lysin, is in combination with SOC antibiotics. Here, we describe a systematic analysis of F12’s antibacterial activity both in vitro and in vivo, including investigations of synergy with chemotherapeutics. The results provide evidence that F12 is a promising new agent for treating highly refractory infections by multidrug-resistant S. aureus, and they further suggest that F12 treatment could resensitize MRSA strains to β-lactam antibiotics, which are currently contraindicated for MRSA infections.

RESULTS

Species selectivity of F12 and wild-type lysostaphin.

As a preliminary test of F12’s spectrum of activity, MIC assays were performed on a panel of staphylococcal species as well as the Gram-negative bacteria Escherichia coli and Pseudomonas aeruginosa (Table 1). Wild-type lysostaphin (LST), bearing an S126P mutation to facilitate expression in Pichia pastoris (31), was tested in parallel as a comparative benchmark. F12 possessed strong antibacterial activity against S. aureus, Staphylococcus pseudintermedius, Staphylococcus arlettae, and Staphylococcus carnosus (MIC range, 16 to 900 ng/ml), with a relatively small 4- to 35-fold increase in MIC compared to that of LST. Notably, three of those four staphylococcal species, excepting S. carnosus, are known human pathogens (9, 32, 33). In contrast, F12 activity against Staphylococcus lentus and Staphylococcus equorum was 750-fold lower than that of LST, and its MIC for Staphylococcus gallinarum exceeded the assay range. Interestingly, the latter three staphylococci are best known for colonizing dairy animals, horses, and chickens, respectively (34–36). Neither F12 nor LST exhibited detectable activity against Staphylococcus epidermidis, Staphylococcus capitis, or Staphylococcus simulans. The first is an important part of normal human skin flora, and its peptidoglycan structure differs from that of S. aureus, rendering it largely insensitive to lysostaphin (37, 38). S. capitis produces ALE-1, a close homolog of lysostaphin, as well as its immunity factor, and cross-resistance to lysostaphin is known (39). S. simulans is the native source of lysostaphin, and it is resistant due to the presence of the lysostaphin immunity factor (LIF) (40). Finally, as expected, neither F12 nor LST manifested detectable activity against the Gram-negative bacteria E. coli or P. aeruginosa. In summary, F12 and LST exhibit similar spectra of activity against staphylococci that are pathogenic in humans, whereas F12 is substantially less active than LST against staphylococci that preferentially colonize animals. These results support F12’s intended use in human patients (30).

TABLE 1.

Bactericidal selectivity of lysostaphins

Bacteria	MIC (ng/ml)
	LSTa	F12b
S. aureus	4 ± 1	140 ± 40
S. carnosus	4 ± 0	16 ± 0
S. pseudintermedius	40 ± 10	300 ± 100
S. arlettae	110 ± 50	900 ± 50
S. lentus	8 ± 0	6,000 ± 2,000
S. equorum	40 ± 10	30,000 ± 10,000
S. gallinarum	1,500 ± 500	>125,000
S. captis	>64,000	>125,000
S. epidermidis	>64,000	>125,000
S. simulans	>64,000	>125,000
E. coli	>64,000	>125,000
P. aeruginosa (PAO1)	>64,000	>125,000
P. aeruginosa (FRD1)	>64,000	>125,000

^aLST is wild-type lysostaphin.

^bF12 is deimmunized lysostaphin variant F12.

Breadth of anti-S. aureus activity.

The primary target bacterium for F12 therapeutic applications is S. aureus, particularly drug-resistant strains, so its activity was next tested by MIC assays against a panel of 25 S. aureus isolates (Fig. 1). MICs were 500 ng/ml or less for the entire panel of isolates, with values for three strains ranging below 100 ng/ml. Importantly, F12 manifested potent growth inhibition for numerous strains that are resistant to the standard-of-care (SOC) MRSA antibiotics vancomycin, daptomycin, and linezolid, suggesting the ability to treat even the most serious cases of antibiotic-refractory infections. More generally, F12’s broad inhibitory capacity for S. aureus clinical isolates mirrors that previously reported for LST (20).

FIG 1.

F12 MICs for panel of S. aureus isolates. F12 MIC values were measured in caMHIIB medium as technical triplicates in two to four independent experiments. Means and standard deviations for all results are shown. DRSA, daptomycin resistant; VRSA, vancomycin resistant; LRSA, linezolid resistant. Data for strains in light blue are taken from reference 30.

In vivo efficacy in a murine bacteremia model.

As a more biologically relevant test of therapeutic potential, F12 was evaluated in a murine model of systemic bacteremia (28, 30). C57BL/6J mice were given a lethal intraperitoneal (i.p.) challenge of MRSA clinical isolate USA400, and 1 h postinfection, animals were treated with a single subcutaneous (s.c.) injection of drug or phosphate-buffered saline (PBS) buffer control. F12 was first evaluated in a dose-response study, where animals were treated with a range of 62.5 to 500 µg of enzyme (Fig. 2A). Both 500 and 250 µg of F12 rescued 80% of infected mice, whereas the 125-µg dose rescued only 40% of animals. The lowest dose of 62.5 µg F12 was similar to that of the PBS vehicle control, with all animals succumbing to the infection in the first 24 h. Having identified the range of F12 efficacious dosing, it was next compared in a head-to-head study with LST at a single dose of 200 µg per mouse (Fig. 2B). Notably, only 20% of LST-treated animals survived beyond 45 h postinfection, whereas animals treated with F12 had an 80% overall survival rate. These results provide strong evidence that F12 manifests high levels of therapeutic efficacy, even exceeding that of its nondeimmunized LST counterpart in the murine bacteremia model.

FIG 2.

Murine bacteremia model in C57BL/6J mice. (A) Dose-response analysis of F12, with enzyme dose in microgram per mouse, as noted (N = 5 per group). Control mice received PBS sham treatment (N = 3). Survival curves are significantly different (P < 0.0001, log rank test). (B) Head-to-head comparison of F12 (blue) and LST (orange) at 200 µg per mouse (N = 5 per group). Control mice received PBS sham treatment (N = 3). Survival curves are significantly different (P = 0.0005, log rank test).

Antibacterial activity in serum.

F12’s significantly enhanced in vivo efficacy compared to LST was unexpected given the fact that its in vitro MIC potency toward S. aureus was approximately 30-fold lower than that of LST (Table 1). In an effort to reconcile this discrepancy, in vitro MICs were reassessed in 50% serum, a more biologically relevant matrix. LST potency was largely unchanged by the presence of serum, whereas F12’s potency increased more than 10-fold relative to Mueller-Hinton II broth cation-adjusted (caMHIIB) medium (Fig. 3). While serum is a complex mixture of proteins and other biological factors, one possible mechanism for serum’s influence on F12 potency was interaction with complement components. To test this hypothesis, F12 and LST MIC assays were redone using 50% heat-inactivated serum in which complement components were inactivated (Fig. 3). The lack of any significant difference in MICs between normal and heat-inactivated sera suggested that complement did not contribute to F12’s improved effectiveness in this biological medium. The mechanistic details notwithstanding, F12’s enhanced MIC potency in serum suggested a partial basis for the variant’s enhanced in vivo efficacy relative to the wild type.

FIG 3.

F12 in vitro potency in synthetic medium versus serum. F12 (blue) and LST (orange) MIC values in standard caMHIIB medium (flat pattern), 50% human serum (checkered pattern), and heat-inactivated 50% serum (crosshatched pattern). Shown are means and standard deviations from three independent experiments. Both the enzyme and assay medium resulted in significant differences (two-way ANOVA). F12 in caMHIIB was significantly less potent than all other treatments (P < 0.0001, Benjamini, Krieger and Yekutieli method, with Q = 0.05). No other conditions exhibited significant differences. F12 and LST activity in caMHIIB and 50% serum are from reference 30.

In vitro synergy with antibacterial chemotherapies.

LST is known to exhibit synergy with a wide variety of antibacterial chemotherapies both in vitro and in vivo (20), suggesting the potential for powerful combination treatments in the clinic. To determine if F12 also manifests synergy with SOC antibiotics, microdilution checkerboard assays (41, 42) were performed with F12 and a panel of small-molecule chemotherapies (Table 2). Similar to previous reports for LST (43, 44), F12 exhibited synergy with seven different beta-lactam antibiotics (amoxicillin-clavulanate, cefazolin, cefoperazone, cephalexin, penicillin G, nafcillin, and ampicillin-sulbactam), yielding especially strong synergy with cefazolin and cefoperazone (fractional inhibitory concentration index [FICI] of 0.046 and 0.094, respectively). F12 demonstrated more modest, but detectable, synergy with the quinoline antibiotic ciprofloxacin (FICI of 0.4) and the anti-MRSA SOC antibiotic linezolid (FICI of 0.5). Conversely, F12 showed only additive activity with daptomycin and vancomycin, two additional anti-MRSA SOC antibiotics, although the daptomycin FICI was borderline synergistic (FICI of 0.57).

TABLE 2.

F12 in vitro synergy with antibacterial chemotherapeutics

Antibiotic	caMHIIB medium		50% Human serum
	FICI	Effect	FICIa	Effect
Rifampin	0.015 ± 0.004	Synergy	ND
Cefazolin	0.046 ± 0.001	Synergy	ND
Cefoperazone	0.094 ± 0	Synergy	ND
Cephalexin	0.11 ± 0.02	Synergy	ND
Penicillin G	0.11 ± 0.02	Synergy	ND
Amoxicillin-clavulanate	0.17 ± 0.04	Synergy	ND
Ampicillin-sulbactam	0.22 ± 0.04	Synergy	0.34 ± 0.04	Synergy
Nafcillin	0.23 ± 0.04	Synergy	0.17 ± 0.02	Synergy
Ciprofloxacin	0.40 ± 0.10	Synergy	ND
Linezolid	0.5 ± 0	Synergy	0.34 ± 0.04	Synergy
Daptomycin	0.57 ± 0.08	Additive	0.80 ± 0.30	Additive
Vancomycin	1	Additive	1	Additive

^aND, not determined.

Given the unexpected increase of F12 MIC potency in serum, as described above (Fig. 3), potential serum effects on antibiotic synergy were a point of interest. Therefore, checkerboard assays were reassessed for a subset of the chemotherapeutics using 50% serum as the growth medium (Table 2). Only subtle increases and decreases in FICI were observed in serum. Thus, serum neither strongly enhanced nor suppressed in vitro synergy with chemotherapeutics.

Synergy with cefazolin in time-kill assays.

The strong F12 synergy with cefazolin in checkerboard assays motivated an examination of cefazolin-F12 combinations in time-kill assays (Fig. 4). MRSA USA400 was treated with 4 μg/ml cefazolin or 100 ng/ml F12 as single agents, each of which represented sub-MICs (cefazolin MIC, 23 ± 9 μg/ml; F12 MIC, 140 ± 40 ng/ml). After 3 h of treatment, cefazolin achieved a maximum 1-log reduction in number of CFU, which was largely sustained through the end of the experiment. F12 achieved a 2-log reduction in CFU numbers in just 1 h, but the bacteria began rebounding at 3 h, yielding a final 1-log reduction. In sharp contrast, the combination of cefazolin and F12, each at the same respective concentrations as the single-agent treatments, yielded a 6-log reduction in CFU by 1 h, and viable counts remained below the detection limit for the duration of the study. Thus, the strong synergy observed between cefazolin and F12 in checkerboard assays manifested as rapid and potent killing for the drug combination in time-kill assays.

FIG 4.

Time-kill assays for F12, cefazolin, and a combination thereof. MRSA USA400 was incubated in caMHIIB medium with F12 alone (100 ng/ml, open blue squares), cefazolin alone (4 μg/ml, closed black circles), or a combination of the two treatments (closed blue circles), where each individual agent was used at sub-MICs. Viable bacterial counts were determined as a function of time.

F12 and β-lactam resistance are mutually exclusive.

One highly desirable quality for new antibiotics is a low propensity to elicit new resistance phenotypes. While S. aureus is known for facile development of resistance to LST, it has been shown that LST resistance renders S. aureus, including MRSA, hypersusceptible to β-lactam antibiotics (41). This clinically relevant characteristic was evaluated by forcing F12 resistance via serial passage of MRSA USA400 in caMHIIB containing subinhibitory concentrations of the enzyme. After each overnight outgrowth, the cultures’ MICs were determined for both F12 and the β-lactam nafcillin (Fig. 5A). MRSA USA400 resistance to F12 showed a relatively steady increase throughout the 9-day experiment, ultimately yielding a 500-fold loss of potency. At the same time, the strain’s nafcillin resistance began to decrease on day 6, reaching a final MIC of 250 ng/ml on day 9. Note that the nafcillin MIC for a standard methicillin-sensitive S. aureus strain (ATCC 29233) is 120 to 500 ng/ml, and by comparison here the starting nafcillin MIC for MRSA USA400 was 8,000 ng/ml, which is above the accepted nafcillin resistance breakpoint (4,000 ng/ml) (45). Therefore, MRSA USA400 was effectively resensitized to nafcillin during this study. Thus, F12 resistance is mutually exclusive of resistance to the β-lactam nafcillin (Fig. 5B). Interestingly, sequencing of the femA and femB genes from this F12-resistant isolate showed no mutations relative to the original strain, and an LST-resistant isolate generated separately also showed no femA or femB mutations. These results show that the lysostaphin resistance mechanism observed here is distinct from the well-known femA or femB deletion mechanism (20), but regardless of the underlying genetics, the mutual exclusion of F12 and β-lactam resistance could have important implications for F12 combination therapies in the clinic.

FIG 5.

S. aureus resistance to lysostaphin and beta-lactam antibiotics is mutually exclusive. F12 resistance was induced in MRSA USA400 by 9 days of serial passage in caMHIIB synthetic medium containing escalating concentrations of subinhibitory enzyme. After each serial passage, the MIC of the culture in caMHIIB medium was determined separately for both F12 and nafcillin. (A) MIC values for F12 (closed blue circles) and nafcillin (open black circles) measured daily during the resistance induction experiment. The final nafcillin MIC was 250 ng/ml on day 9. (B) Nafcillin versus F12 MIC values for individual cultures from the resistance induction experiment. The sharp elbow pattern demonstrates that S. aureus resistance to F12 and nafcillin is mutually exclusive. Data in panel A are averages from two independent resistance induction experiments where MIC values were measured daily in duplicate for both experiments. Data in panel B are from one of the two independent experiments.

In vivo synergy with antibacterial chemotherapies.

The above-described results suggested that F12 synergizes strongly with β-lactam antibiotics and is additionally capable of resensitizing MRSA to β-lactam drugs. To evaluate this desirable attribute in a more clinically relevant fashion, F12 and cefazolin were tested in a difficult-to-treat rabbit model of MRSA infective endocarditis (IE). F12 monotherapy was given as a single 40-mg/kg of body weight intravenous (i.v.) dose, and it resulted in small but statistically significant reductions in bacterial burden in aortic vegetations, kidneys, and spleens (Fig. 6). Notably, F12 monotherapy reduced bacterial burden in the blood by 2.6 logs and reduced mortality to 20%, compared to 95% mortality for no treatment controls. Cefazolin monotherapy administered intramuscularly (i.m.) at 100 mg/kg three times daily (TID) yielded results similar to those of F12 monotherapy, although the mortality rate in this group was higher, at 40%. In contrast, the combination therapy yielded striking in vivo efficacy. For several animals, bacterial burden was below the limit of detection in the blood (4/5 rabbits), spleen (3/5), kidney (3/5), and cardiac vegetations (2/5). On average, bacterial burden was reduced in the blood, spleen, kidney, and cardiac vegetations by 4, 4, 5, and 6 logs, respectively, and the combination therapy afforded 100% survival of treated animals. These results demonstrate that F12 exhibits strong synergy with cefazolin in vivo, and, taken together with the in vitro studies, they provide further evidence that F12 effectively resensitizes MRSA to β-lactam antibiotics.

FIG 6.

F12, cefazolin, and combination therapy in a rabbit model of left-sided infective endocarditis. Combination dosing matched single-agent dosing, with F12 administered once on day 1. N = 5 rabbits per treatment group and N = 6 for the control group. Mortality rates were 95% control, 40% cefazolin, 20% F12, and 0% combination therapy. (A to D) Viable bacterial counts recovered from cardiac vegetations (A), blood (B), kidney (C), and spleen (D). Lines and bars are means and standard deviations. The hashed horizontal line represents the limit of detection, and open symbols indicate sterilizing efficacy, i.e., no bacteria detected. In each tissue, all treatment groups exhibited significantly lower numbers of CFU than the respective control, and the combination therapy yielded significantly lower numbers of CFU in all tissues than in both monotherapies, with the exception of similar numbers of blood CFU compared to those of cefazolin monotherapy (2-way ANOVA, Benjamini, Krieger and Yekutieli method with Q = 0.05). Data for the F12 monotherapy group are taken from reference 30.

DISCUSSION

There is an urgent need for novel therapeutics to treat drug-resistant bacteria, and MRSA in particular continues to be a prolific multidrug-resistant pathogen that manifests outsized impacts on human health (1, 3, 11, 46). Lysins represent the vanguard of new biotherapies targeting MRSA infections (47), but these innovative antibacterial biocatalysts are of nonhuman origin and suffer from detrimental immunogenicity in humans. The immunogenicity of LST, a prototypical lytic enzyme, has been addressed via mutagenic deletion of its constituent T cell epitopes, yielding the deimmunized variant F12 (30). The studies reported here sought to probe the antibacterial activity of F12, with a focus on assessing its ability to complement SOC antibiotics.

Because synergy with antibiotics is desirable for any prospective lysin therapy (47), a key objective of this work was analysis of F12 synergy with various antibacterial chemotherapies. Wild-type LST is known to be synergistic with a wide array of antibiotics (20), and, similarly, F12 was found to manifest in vitro synergy with several classes of antibiotics, including a rifamycin, a fluoroquinolone, and an oxazolidinone. F12 also exhibited strong synergy with β-lactams in checkerboard and time-kill assays, and this attribute translated into near-sterilizing efficacy of an F12-cefazolin combination therapy toward MRSA IE in rabbits. Importantly, while both F12 and the phage lysin CF-301 are known to synergize strongly with daptomycin in this refractory rabbit IE model (30, 48), the results shown here demonstrate that a normally ineffective β-lactam can be rendered highly efficacious against MRSA when combined with F12. These results are similar to prior analyses of LST (41). In total, the β-lactam synergy experiments presented here are evidence that F12 has the potential to restore β-lactam activity against otherwise resistant MRSA strains. The potential to breathe new life into β-lactam therapies, as they relate to MRSA, represents an intriguing and potentially valuable proposition for physicians and patients.

An additional advantage of F12 and β-lactam combination treatments may be suppression of new resistance phenotypes. We showed here that forced evolution of F12 resistance causes MRSA USA400 to become hypersusceptible to the β-lactam nafcillin. The mutual exclusion of F12 and nafcillin resistance is reminiscent of similar observations seen with wild-type lysostaphin (20, 41). In those prior studies, it was found that S. aureus can develop lysostaphin resistance via mutational inactivation of the femA gene (20). The FemA peptidyltransferase adds glycines two and three in the pentaglycine interpeptide bridge of S. aureus peptidoglycan (49, 50), and inactivation of FemA results in monoglycine cross-bridges. This eliminates both a lysostaphin binding target and the catalytic substrate, thereby imparting high-level lysostaphin resistance (20, 41). Conversely, MRSA resistance to β-lactams is imparted by penicillin-binding protein 2a (PBP 2a) (51), encoded by the mecA gene. This transpeptidase has a low affinity for β-lactams, thereby conveying resistance, but it can only perform transpeptidation and cross-linking using a pentaglycine substrate; PBP 2a cannot cross-link monoglycine bridges. Thus, in lysostaphin-resistant strains, which have monoglycine crossbridges, transpeptidation occurs via the chromosomally encoded PBPs, which are inherently susceptible to β-lactam inhibition. Interestingly, the F12-resistant strains generated here did not encode mutations in either FemA or FemB peptidyltransferases, suggesting an alternative mechanism of resistance. It is noteworthy that neither of the comparably LST-resistant strains, generated separately, encode FemA or FemB mutations. Thus, the unknown resistance mechanism is likely not unique to F12. Recent reports have pointed to alternative mechanisms of lysostaphin resistance, including wall teichoic acids (52, 53) as well as the fmhA and fmhC gene products, which incorporate serine into the normal pentaglycine crossbridge, conferring lysostaphin resistance but β-lactam sensitivity (54). Regardless of the underlying mechanism, the observed mutual exclusion of F12 and β-lactam resistance provides a basis for development of F12 combination therapies. Such resistance-suppressing combinations, together with previously reported serum-mediated suppression of F12 resistance (30), suggests a high barrier to emergent F12 resistance in the clinic.

Lastly, given the higher MIC value of F12 versus LST in standard caMHIIB assay medium, it was surprising that F12 manifested better-than-wild-type efficacy in a murine model of acute bacteremia. This observation prompted reanalysis of F12 in vitro potency using serum as a biologically relevant matrix. Mirroring its enhanced in vivo efficacy, F12 activity increased more than 10-fold in serum, whereas wild-type LST activity remained unchanged. The latter result is consistent with prior reports on the tolerance of wild-type LST toward serum (55). In contrast, serum’s complement-independent enhancement of F12 activity is reminiscent of serum effects on the phage lysin CF-301, where this phage lysin was shown to synergize with serum lysozyme and serum albumin (56).

In total, the systematic in vitro and in vivo analyses conducted here provide a preliminary rationale for clinical use of F12 in combination with SOC antibiotics. Benefits of this strategy might include synergistic activity yielding improved outcomes, suppression of new resistance phenotypes, and the potential for resensitization of MRSA to β-lactam antibiotics. F12’s ability to complement SOC antibiotic therapies and its capacity for safe and effective repeated dosing (30) suggest an opportunity to develop a best-in-class combination therapy for highly refractory or recurrent MRSA infections.

MATERIALS AND METHODS

Chemicals and reagents.

Unless otherwise stated, disposable materials and common chemicals were purchased from Fisher Scientific (Pittsburgh, PA, USA), VWR International (Radnor, PA), and Research Products International (Mount Prospect, IL). Miller's LB broth, tryptic soy broth (TSB), ampicillin, cefazolin, cefoperazone, cephalexin, penicillin G, gentamicin, and vancomycin were purchased from Research Products International. BBL Mueller-Hinton II broth cation-adjusted (caMHIIB) was purchased from Becton, Dickinson, and Company. Bovine serum albumin (BSA), human serum (from human male AB plasma, U.S. origin, sterile filtered), nafcillin, ciprofloxacin, and sulbactam were purchased from Sigma-Aldrich. Amoxicillin clavulanate was purchased from VWR International. Daptomycin and linezolid were purchased from R&D Systems (Minneapolis, MN).

Bacterial strains.

Strains used these studies are summarized in Table 3. Unless otherwise stated, all bacterial strains were stored as 20% glycerol stocks at −80°C.

TABLE 3.

Bacterial strains used in this study

Strain	Source	Strain no.a
S. aureus MW2 (USA400)	ATCC	BAA-1707
S. aureus SA113	ATCC	35556
S. aureus Seattle 1945	ATCC	25923
S. aureus JE2 (USA300)	University of Nebraska Medical Center	JE2
S. aureus SA LinR #12	BEI Resources	NR-45924
S. aureus H2138	BEI Resources	NR-46062
S. aureus AIS 080003	BEI Resources	NR-46419
S. aureus AIS 1000505	BEI Resources	NR-46420
S. aureus ALC 3425-1	Ambrose Cheung	ALC 3425-1
S. aureus ALC 3425-3	Ambrose Cheung	ALC 3425-3
S. aureus ALC 4051	Ambrose Cheung	ALC 4051
S. aureus ALC 5240	Ambrose Cheung	ALC 5240
S. aureus ALC 6334	Ambrose Cheung	ALC 6334
S. aureus ALC 6424	Ambrose Cheung	ALC 6424
S. aureus ALC 6445	Ambrose Cheung	ALC 6445
S. aureus ALC 7236	Ambrose Cheung	ALC 7236
S. aureus isolates SA1 to SA10	Lyticon LLC	SA1 to SA10
S. carnosus DSM 20501	ATCC	51365
S. arlettae	NRRL	14764
S. equorum	NRRL	14765
S. simulans	NRRL	B-2628
S. gallinarum	NRRL	B-14763
S. capitis	NRRL	B-14752
S. lentus LRA 003.11.78	ATCC	49574
S. epidermidis	This study	NA
S. pseudintermedius	This study	NA
E. coli DH5α	Thermo Fisher	18265017
P. aeruginosa (PAO1)	George O’Toole	PAO1
P. aeruginosa (FRD1)	George O’Toole	FRD1

^aCatalog or repository number is provided. NA, not applicable.

Protein expression and purification.

Recombinant Pichia pastoris strains expressing an S126P point mutant of LST, used here as a wild-type control, and F12 have been described elsewhere (30, 31). The expression and purification of LST and F12 followed a previously described procedure (31). In brief, a single colony of recombinant P. pastoris harboring pPIC9-LST or pPIC9-F12 was picked and grown in 10 ml YPD medium (1% yeast extract, 2% peptone, 2% dextrose) at 30°C for 48 h. The entire culture was transferred into a 3-liter Applikon Bioreactor (Applikon Biotechnology) and grown in 1.5 liters of basal low-salt medium (0.5 g CaSO₄·2H₂O, 5 g K₂SO₄, 2 g MgSO₄ [4 g MgSO·7H₂O], 1 g KOH, 20 ml H₃PO₄, 30 ml glycerol, 11 ml PTM1 [standard trace salts recipe], 11 ml 500× biotin, pH 5.0) at 30°C. Upon observing a dissolved oxygen (DO) spike after approximately 16 to 20 h, 50% glycerol-PTM1-biotin solution was fed at approximately 12.5 ml per h, adjusting the feed rate to maintain DO at 10% or higher. After the glycerol feed was complete, 100 ml YPD feed medium (20 g peptone and 10 g yeast extract mixture) was added as a single bolus, and the pH control was set to 6.0. After 2 h of growth, the cells were induced for 72 h at 20°C by feeding methanol solution at a rate sufficient to maintain the DO at 10% or higher. Following induction, the reactor was harvested and centrifuged at 10,000 rpm for 20 min. The supernatant was collected and diluted 10-fold with 20 mM NaHPO₄ buffer, pH 7.5, and the diluted supernatant was loaded onto a GE SP Sepharose fast flow column (100 ml), which was preequilibrated with 20 mM NaHPO₄ buffer, pH 7.5. Following a 10-column-volume (CV) wash with the same buffer, LST or F12 was eluted with a 0 to 250 mM NaCl gradient. Fractions containing concentrated LST or F12 were identified by SDS-PAGE, combined, and dialyzed into HIC binding buffer [20 mM NaHPO₄, 1 M (NH₄)₂SO₄, pH 7.0]. This solution was then loaded onto an HIC phenyl HP column (20 ml) preequilibrated with HIC binding buffer. Following a 10-CV wash with the same buffer, LST or F12 was eluted with a 1 to 0 M (NH₄)₂SO₄ gradient in 20 mM NaHPO₄, pH 7.0. Fractions containing concentrated enzymes were identified by SDS-PAGE, pooled, and dialyzed into phosphate-buffered saline (PBS; 137 mM NaCl, 2.6 mM KCl, 10 mM Na₂HPO₄, 1.7 mM KH₂PO₄, pH 7.4). The final preparations were approximately 95% pure based on SDS-PAGE analysis.

MIC assay.

The MIC assay was performed in a 96-well CELLSTAR tissue culture plate (Greiner Bio-One) with a lid, clear and sterile (VWR International), by following a standard modification of CLSI methods, which employ a synthetic cation-adjusted MHIIB growth medium (caMHIIB) (57, 58). In brief, purified LST or F12 was serially diluted 1:2 into each well using deionized water supplemented with 0.2% BSA in a total volume of 50 μl. For bacterial preparation, a single bacterial colony from a freshly streaked TSB agar plate was picked, grown in TSB medium at 37°C overnight to saturation, subcultured 1:100 into 5 ml of fresh TSB medium, and grown at 37°C to an optical density at 600 nm (OD₆₀₀) of 0.6 to 0.9. After centrifugation, cell pellets were resuspended in fresh 2× caMHIIB medium and diluted to a final concentration of approximately 10⁶ CFU per ml. Each well was then inoculated with 50 μl of the bacterial working stock to yield approximately 50,000 CFU in a total volume of 100 μl per well. The microtiter plate was then incubated at 37°C without shaking, and the MIC values were recorded as the lowest enzyme concentration at which there was no visible growth at 24 h. Triplicate MICs were measured in two to four independent experiments. Means and standard deviations are reported.

Serial induction of resistance.

Serial induction of resistance was performed by following a set of sequential MIC assays as previously described (59). Briefly, MRSA strain USA400 was cultured in either caMHIIB medium or 50% human serum at 37°C with the presence of antibiotic (LST, F12, or nafcillin), in a format identical to that of the MIC assay described above. Subsequently, the outgrown wells having the highest concentration of antibiotic were selected and inoculated into fresh TSB medium and grown at 37°C overnight to saturation. The MIC assay was then repeated using this overnight culture, such that the bacteria were exposed to incrementally elevated concentrations of each antibiotic. This procedure was repeated for the duration of the experiment (14 days in human serum) or until the MIC exceeded 1 mg/ml. Resistance induction was performed in two parallel experiments, and the corresponding means and standard deviations from resulting MIC values are reported. The femA and femB genes were sequenced from the original USA400 stock strain, the F12-resistant isolate, and an LST-resistant isolate that was generated separately, but no mutations were found for either resistant isolate.

Checkerboard synergy assay.

The checkerboard assay was performed in a 96-well microtiter plate by following a previously described method (41, 60). In brief, LST or F12 was serially diluted down the columns of the plate and a small-molecule antibiotic was serially diluted across the rows of the plate. Dilutions were 1:2 in 50 μl deionized water supplemented with 0.2% BSA. A MRSA USA400 culture was prepared as described for the MIC assay, and 50 μl of bacterial culture was added into each well of the plate, yielding 50,000 CFU in a total volume of 100 μl. For testing cefazolin, cefoperazone, and cephalexin, caMHIIB medium was supplemented with 2% NaCl. The microtiter plates were incubated at 37°C without shaking for 24 h, and wells with no visible growth were recorded. Each assay was done as three separate experiments; the corresponding means and standard deviations are reported.

Each drug combination’s minimum fractional inhibitory concentration (FIC) value and fractional inhibitory concentration index (FICI) were calculated using the following methods: (i) FIC = (MIC of an agent in the combination)/(MIC of the agent alone) or (ii) FICI = (FIC of agent 1) + (FIC of agent 2). FICI values of less than 0.5 or higher than 4 are defined as synergy or antagonism, respectively. FIC values between 0.5 and 4 are defined as additive (61).

Time-kill assays.

Time-kill assays were performed using the method described by CLSI guidelines (62). In brief, 10⁶ CFU/ml MRSA USA400 was inoculated into 25 ml caMHIIB medium containing subinhibitory F12 (100 ng/ml), subinhibitory cefazolin (4 μg/ml), or a combination thereof. The flasks were cultured at 37°C with continuous shaking. One hundred-microliter aliquots were withdrawn at the following time points: 0 min, 15 min, 30 min, 60 min, 2 h, 3 h, 4 h, 5 h, and 6 h. The aliquots were 10-fold serially diluted in test tubes of sterile saline (0.9% NaCl), and diluted samples were spread onto TSB agar plates. After overnight incubation, agar plates with 20 to 200 colonies were used to determine colony counts, and the bacterial density in the corresponding culture was calculated. Time-kill kinetic curves were plotted from the bacterial density and time data. Each assay was performed three times; the mean and standard deviation bacterial counts are reported.

In vivo model studies.

All animal protocols were approved by the Institutional Animal Care and Use Committee of Dartmouth College (Hanover, NH) or the Institutional Animal Care and Use Committee of the Lundquist Institute at Harbor-UCLA Medical Center, as appropriate, in accordance with the Association for the Assessment and Accreditation of Laboratory Animal Care Guidelines.

Murine bacteremia model.

A murine model of systemic MRSA infection was used to compare the in vivo efficacy of LST and F12 against MRSA USA400. Six- to 8-week-old female C57BL/6J mice were challenged with intraperitoneal injection of 2 × 10⁸ CFU of MRSA USA400 (in 3% mucin). After 1 h, the mice were treated by subcutaneous injection of sterile PBS (untreated control group), 200 µg of LST in sterile PBS, or 200 µg F12 in sterile PBS, respectively, and animals were monitored daily throughout the study. The F12 dose-response study followed a similar protocol, but F12 dosing was varied from 62.5 µg to 500 µg per mouse. N = 5 for groups, except the PBS control groups, which were composed of N = 3 due to the high reproducibility of the control in numerous prior studies. Significant differences were determined by a log rank test.

Rabbit infective endocarditis model.

A well-characterized rabbit transcarotid artery-to-left-ventricle catheter-induced aortic valve endocarditis model (63), using MRSA USA400, was used to evaluate the efficacy of F12 alone and in combination with cefazolin. At 48 h after catheter placement, animals were challenged i.v. with ∼2 × 10⁵ CFU, an inoculum of USA400 that induces infective endocarditis in >95% of catheterized animals (ID₉₅). At 24 h postinfection, animals were randomized into one of four treatment groups: (i) control without treatment and sacrificed at 24 h postinfection (the time of therapy initiation); (ii) F12 at 40 mg/kg, i.v., once; (iii) cefazolin alone (100 mg/kg, i.m., TID, for 3 days); or (iv) F12 plus cefazolin, each dosed as described above. At 24 h after the last treatment, animals were humanely euthanized, and cardiac vegetations, kidneys, spleen, and blood were sterilely removed and quantitatively cultured. Bacterial burden was calculated as the log₁₀ mean number of CFU/g of tissue (or ml of blood) for each group. Significant differences were determined by two-way analysis of variance (ANOVA) for treatment and tissue type using the Benjamini, Krieger, and Yekutieli method. The false discovery rate was set to a Q of 0.05.