A Novel, Broad-Spectrum Antimicrobial Combination for the Treatment of Pseudomonas aeruginosa Corneal Infections

Bacterial keratitis causes significant blindness, yet antimicrobial resistance has rendered current treatments ineffective. Polymyxin B-trimethoprim (PT) plus rifampin has potent in vitro activity against Staphylococcus aureus and Pseudomonas aeruginosa, two important causes of keratitis. Here we further characterize this combination against P. aeruginosa in a murine keratitis model.

ABSTRACT

Bacterial keratitis causes significant blindness, yet antimicrobial resistance has rendered current treatments ineffective. Polymyxin B-trimethoprim (PT) plus rifampin has potent in vitro activity against Staphylococcus aureus and Pseudomonas aeruginosa, two important causes of keratitis. Here we further characterize this combination against P. aeruginosa in a murine keratitis model. PT plus rifampin performed comparably to or better than moxifloxacin, the gold standard, suggesting that the combination may be a promising therapy for bacterial keratitis.

TEXT

Bacterial keratitis (corneal infection) is a devastating disease that is responsible for 2 million cases of blindness annually (1, 2). Staphylococcus aureus and Pseudomonas aeruginosa are among the most important causes of keratitis (3, 4), requiring immediate, broad-spectrum antimicrobial treatment to prevent vision loss. In that regard, fluoroquinolones are widely employed; however, increasing resistance has rendered this class of drugs less effective (5). With limited available alternatives, there is a critical need for novel ophthalmic antimicrobial therapeutics (6–8).

Polymyxin B-trimethoprim (PT) is a current ophthalmic antimicrobial that, despite low levels of circulating resistance (5), is rarely utilized to treat keratitis because of slow bactericidal activity (9, 10). To develop an improved PT-based combination for the treatment of keratitis, we recently conducted a large-scale screen utilizing a library containing FDA-approved agents, to identify compounds that, when combined with PT, would demonstrate synergistic, broad-spectrum antimicrobial activity. The results revealed PT plus rifampin as a promising combination for further investigation (11).

PT plus rifampin displays potent bactericidal activity that rivals or exceeds that of the fluoroquinolone moxifloxacin, a current standard keratitis treatment. Importantly, the potent antimicrobial activity of PT plus rifampin extended to a S. aureus in vivo keratitis model (11). However, to justify further development of PT plus rifampin as a clinical therapeutic, we sought to demonstrate in vivo efficacy toward P. aeruginosa corneal infections.

P. aeruginosa strains in this study included PAO1, which is a fluoroquinolone-sensitive laboratory strain, and RW37, which is a clinical keratitis isolate collected from the University of Rochester Medical Center. RW37 is fluoroquinolone sensitive; the MICs, determined using standard MIC assays (12), for moxifloxacin (Alcon), levofloxacin (Sigma-Aldrich), and ciprofloxacin (Sigma-Aldrich) were 1.2 μg ml⁻¹, 1 μg ml⁻¹, and 0.5 μg ml⁻¹, respectively. A moxifloxacin-resistant strain was created by culturing RW37 cells in medium containing 2.44 μg ml⁻¹ moxifloxacin (2×MIC). Cells were subsequently passaged three times to yield RW38, a fluoroquinolone-resistant derivative with a moxifloxacin MIC of 27 μg ml⁻¹ (22-fold increase). New point mutations in gyrB (E133G) and parC (L328P, L415P, G579D, A605G, and V638C) were identified in RW38 via PCR and DNA sequencing, suggesting that moxifloxacin resistance in RW38 can be attributed to mutations within known fluoroquinolone targets (13), as opposed to being efflux mediated.

To evaluate the in vitro activity of PT plus rifampin toward RW37 and RW38, fractional inhibitory concentration (FIC) measurements were performed in a checkerboard format, as described previously (14). The FIC index was then calculated as follows: (MIC of drug A in combination/MIC of drug A alone) + (MIC of drug B in combination/MIC of drug B alone) = FIC. Synergism was defined as FIC of <0.5, additive as FIC of 0.5 to 1.0, neutral as FIC of 1 to 4, and antagonistic as FIC of >4 (14). The MICs of commercial PT (Sandoz Pharmaceuticals) and rifampin (Sigma-Aldrich) (alone) toward PAO1 were 0.41 μg ml⁻¹ and 15.6 μg ml⁻¹, respectively; in combination, the MIC values decreased 2-fold and 8-fold to 0.21 μg ml⁻¹ and 1.9 μg ml⁻¹, respectively, resulting in a FIC index of 0.637 (Table 1). The MIC values of PT and rifampin (alone) for RW37 and RW38 were identical at 0.313 μg ml⁻¹ and 32 μg ml⁻¹, respectively. However, in combination, the MIC values for PT and rifampin decreased 2-fold to 0.156 μg ml⁻¹ and 256-fold to 0.125 μg ml⁻¹, respectively, resulting in a FIC index of 0.502, approaching a synergistic effect.

TABLE 1.

MIC and FIC results for PT and rifampin, alone and in combination

Strain	MIC (μg ml−1)				FIC index
	Alone		Combination
	PT	Rifampin	PT	Rifampin
PAO1	0.41	15.6	0.21	1.9	0.637
RW38	0.313	32	0.156	0.125	0.502
RW37	0.313	32	0.156	0.125	0.502

Next, the in vivo efficacy of PT plus rifampin toward P. aeruginosa was evaluated. Four- to 6-week-old female C57BL/6 mice were obtained from Charles River Laboratories (Washington, MA) and housed according to a protocol approved by the University of Rochester Council on Animal Research. Mice were anesthetized with subcutaneous 100 mg kg⁻¹ ketamine (Par Pharmaceutical, Chestnut Ridge, NY) and 10 mg kg⁻¹ xylazine (Akorn, Inc., Lake Forest, IL) and topical 0.5% proparacaine (Akorn). A 27-gauge needle carrying a single colony of P. aeruginosa was used to create three parallel 1-mm scratches across the central right cornea, followed by inoculation with a 5-μl suspension of overnight culture containing 10⁷ PAO1 or RW38 cells. Slit-lamp biomicroscopy was used to assign a disease severity score using the following previously validated metric (15): 0, no opacity; 1, opacity of <4 mm; 2, opacity of >4 mm; 3, 100% opacity; 4, corneal perforation.

Treatments were administered every 6 h, beginning 6 h postinoculation (11). Treatment groups (5 animals each) included balanced salt solution (BSS), moxifloxacin (5.4 mg ml⁻¹), PT (1,000 units polymyxin B and 1 mg ml⁻¹ trimethoprim), rifampin (0.5 mg ml⁻¹), and PT plus rifampin (1,000 units polymyxin B and 1 mg ml⁻¹ trimethoprim plus 0.5 mg ml⁻¹ rifampin). At 72 h postinoculation, mice were euthanized and the eyes homogenized in a bead mill homogenizer for CFU enumeration.

The clinical severity scores of P. aeruginosa PAO1 across all treatment groups were similar (Fig. 1A). With respect to bacterial burden, however, PT treatment (alone) resulted in a 3-log decrease in CFU, compared to BSS or rifampin (alone), while PT plus rifampin and moxifloxacin resulted in complete eradication of bacteria (Fig. 1B). Representative slit-lamp images for each treatment are shown in Fig. 1C. Of note, while corneal opacity often correlates with active infection, these changes may also represent the host immune response (16) and may correlate with current or prior disease severity.

FIG 1.

Antimicrobial efficacy of PT plus rifampin against P. aeruginosa in a murine keratitis model. (A) Disease severity scores for animals inoculated with PAO1 following 72 h of treatment with BSS, moxifloxacin, rifampin, PT, or PT plus rifampin. (B) Numbers of CFU recovered from animals inoculated with PAO1 following 72 h of treatment with BSS, moxifloxacin, rifampin, PT, or PT plus rifampin. (C) Representative slit-lamp images of each treatment group after 72 h for eyes inoculated with PAO1. RIF, rifampin. (D) Disease severity scores for animals inoculated with RW38 following 72 h of treatment with BSS, moxifloxacin, rifampin, PT, or PT plus rifampin. (E) Numbers of CFU recovered from animals inoculated with RW38 following 72 h of treatment with BSS, moxifloxacin, rifampin, PT, or PT plus rifampin. Each symbol represents an individual animal, with the averages indicated by horizontal lines. The lower limit of detection (5 × 10² CFU) is indicated by horizontal dashed lines. Asterisks indicate statistical significance based on Student's t test. *, P < 0.15; **, P < 0.05.

Given increasing fluoroquinolone resistance among P. aeruginosa ocular isolates, the ability of PT plus rifampin to effectively combat a fluoroquinolone-resistant in vivo infection was evaluated. Across all groups, PT plus rifampin demonstrated the greatest decrease in clinical severity (Fig. 1D). Indeed, corresponding CFU counts correlated well with severity scores (Fig. 1E). While moxifloxacin resulted in a 4-log CFU decrease, treatment with PT plus rifampin resulted in complete eradication of bacteria in all but one animal tested. Taken together, these results suggest that PT plus rifampin can effectively treat a fluoroquinolone-resistant clinical P. aeruginosa infection.

The success of PT plus rifampin toward resistant isolates may be attributed to the targeting of multiple bacterial molecular pathways. While polymyxin B acts as a cell wall detergent, trimethoprim targets the dihydrofolic acid pathway and rifampin inhibits bacterial RNA polymerase (17–19). Indeed, this cocktail strategy has been utilized to treat a wide variety of bacterial infections, such as Enterococcus endocarditis, in which cell wall inhibitors are routinely combined with aminoglycosides, which inhibit protein synthesis (20).

The combination of PT plus rifampin has been shown previously to be a potent antibiotic effective against P. aeruginosa and S. aureus in in vitro studies, as well as against S. aureus in an in vivo model of keratitis (11). The data presented here extend the efficacy of PT plus rifampin to include P. aeruginosa in a murine model of keratitis, thus bolstering the combination as a promising new antimicrobial for the treatment of corneal infections.