Natamycin and Voriconazole Exhibit Synergistic Interactions with Nonantifungal Ophthalmic Agents against Fusarium Species Ocular Isolates

The in vitro activities of two antifungal drugs in combination with four nonantifungal ophthalmic agents were evaluated using a broth microdilution method and a collection of eight Fusarium ocular isolates that exhibited resistance to both natamycin (MICs, 14 to 32 μg/ml) and voriconazole (MICs, 4 to >128 μg/ml). Synergistic and indifferent interactions were observed for natamycin and 5-fluorouracil and natamycin with timolol dependent on the Fusarium isolate tested.

ABSTRACT

The in vitro activities of two antifungal drugs in combination with four nonantifungal ophthalmic agents were evaluated using a broth microdilution method and a collection of eight Fusarium ocular isolates that exhibited resistance to both natamycin (MICs, 14 to 32 μg/ml) and voriconazole (MICs, 4 to >128 μg/ml). Synergistic and indifferent interactions were observed for natamycin and 5-fluorouracil and natamycin with timolol dependent on the Fusarium isolate tested. Isolate-dependent synergistic and indifferent interactions were also observed for natamycin with EDTA and natamycin with dorzolamide. Synergistic or indifferent interactions were observed for voriconazole with timolol and voriconazole with 5-fluorouracil depending on Fusarium isolate. Taken together, these data suggest that commonly used ophthalmic agents enhance the in vitro activity of antifungal drugs against drug-recalcitrant ocular fungal pathogens.

INTRODUCTION

Filamentous fungal keratitis is a potentially devastating infection of the eye that can rapidly lead to the loss of vision. This condition occurs most commonly in the setting of corneal trauma or contact lens wear in patients living in tropical regions of the world. Epidemiological data suggest that most cases of fungal keratitis are caused by either Aspergillus or Fusarium species (1). Fusarium isolates often exhibit high-level drug resistance or tolerance to contemporary antifungal drugs (2) and may produce devastating infections even when treated with a drug that in vitro susceptibility profiles suggest should be efficacious (3). Thus, traditional first-line antifungal therapies are often ineffective as monotherapy in the management of Fusarium keratitis, potentially resulting in corneal ulceration and blindness (4).

In recent years, topical voriconazole and natamycin have been proposed to be first-line therapies in the setting of filamentous fungal keratitis. Earlier studies suggested that voriconazole was the superior agent with respect to both outcomes (5) and in vitro susceptibility profiles (6, 7). However, more recent results from the mycotic ulcer treatment trial (MUTT) published in 2013 demonstrated the superiority of natamycin over voriconazole as monotherapy in filamentous fungal keratitis (8). Specifically, natamycin was shown to be superior to voriconazole in the setting of infections caused by Fusarium, while no significant differences between the two agents were observed in the setting of non-Fusarium infections. However, even with topical antifungal therapy, approximately one in six study subjects with fungal keratitis experienced a corneal perforation or required a therapeutic penetrating keratoplasty (TPK).

Many prior studies have evaluated the efficacy of antifungals when combined with other agents, such as antibiotics (9, 10) or other antifungals (11, 12). However, to the best of our knowledge, studies to evaluate possible synergistic interactions between topical antifungal agents and ophthalmic agents that are without well-characterized intrinsic antifungal or antimicrobial properties are rare. In this study, we evaluated the interactions between two first-line topical antifungal drugs (natamycin and voriconazole) and four nonantifungal ophthalmic drugs (5-fluorouracil [5FU], dorzolamide, EDTA, and timolol). We hypothesized that since these medications are known to impact eukaryotic cells, they might alter the biology of Fusarium species through a mechanism(s) which enhances the activity of contemporary antifungal therapies.

RESULTS AND DISCUSSION

The MICs of all compounds used in this study are presented in Table 1. Notably, the evaluated Fusarium species isolates demonstrated marked resistance to both natamycin (MICs, 14 to 32 μg/ml) and voriconazole (MICs, 4 to >128 μg/ml). 5-Fluorouracil and EDTA were capable of inhibiting the growth of the Fusarium isolates alone at high concentrations, while dorzolamide and timolol exhibited no antifungal activity, even at the maximum doses evaluated (5.6 and 6.8 mg/ml, respectively). The MIC of pure dimethyl sulfoxide (DMSO) was 137.6 mg/ml and was higher than the concentration of DMSO present in the medium at the MIC of 5-fluorouracil (between 17.2 mg/ml [3%] and 68.8 mg/ml [12%]). These findings imply that the antifungal activity associated with 5-fluorouracil is likely beyond that which can be attributed to the solvent alone. However, we cannot completely rule out the possibility of a subtle effect of the DMSO on the natamycin-5FU assays.

TABLE 1.

MIC results for Fusarium species ocular isolates^a

Isolate	Antifungal MIC (μg/ml)		Nonantifungal MIC (μg/ml)
	NAT	VOR	5FU	DRZ	EDTA	TIM
F. solani
132.02	16	16	1,880	>5,575	403	>6,800
06-0487	26	>128	3,750	>5,575	819	>6,800
06-0330	26	>128	2,500	>5,575	2,041	>6,800
06-0133	29	>128	2,500	>5,575	819	>6,800
06-0111	32	>128	2,500	>5,575	1,222	>6,800
06-0110	29	>128	2,500	>5,575	2,041	>6,800
F. oxysporum
06-0342	26	>128	5,000	>5,575	2,041	>6,800
06-0197	14	4	3,750	>5,575	1,222	>6,800

^aNAT, natamycin; VOR, voriconazole; 5FU, 5-fluorouracil; DRZ, dorzolamide; TIM, timolol.

For the combination of antifungal and nonantifungal agents, the fractional inhibitory concentration index (FICI) associated with each drug-drug combination was calculated as follows: FICI = (MIC of drug A in the combination/MIC of drug A alone) + (MIC of drug B in the combination/MIC of drug B alone).

Using thresholds that have been described previously (13), an FICI of <0.5 represented a synergistic interaction between drugs, an FICI of ≥0.5 and <4.0 represented an indifferent interaction, and an FICI of ≥4.0 represented an antagonistic interaction. The FICIs for all antifungal-nonantifungal pairs are presented in Table 2. Synergistic interactions were observed for natamycin and timolol (mean FICI, 0.45) and natamycin and 5-fluorouracil (mean FICI, 0.46), and the combined MICs and FICIs calculated are presented in Table 3 for natamycin-timolol and in Table 4 for natamycin-5FU. Indifferent interactions were observed for voriconazole and EDTA (mean FICI, 0.57), natamycin and EDTA (mean FICI, 0.63), voriconazole and 5-fluorouracil (mean FICI, 0.64), voriconazole and timolol (mean FICI, 0.67), and natamycin and dorzolamide (mean FICI, 0.68). Additional indifferent interactions were observed for voriconazole and dorzolamide (mean FICI, 2.00). Natamycin in combination with DMSO yielded an indifferent interaction (mean FICI, 1.29), as did voriconazole in combination with DMSO (mean FICI, 0.81).

TABLE 2.

FICI results for Fusarium species ocular isolates

Isolate	FICIa
	Natamycin plus:				Voriconazole plus:
	5FU	DRZ	EDTA	TIM	5FU	DRZ	EDTA	TIM
F. solani
132.02	0.61	0.75	1.00	0.52	0.72	2.00	1.00	0.94
06-0487	0.56	0.75	0.67	0.45	0.75	2.00	0.38	0.53
06-0330	0.40	0.64	0.66	0.41	0.59	2.00	0.63	0.69
06-0133	0.34	0.69	0.75	0.41	0.53	2.00	0.52	0.58
06-0111	0.44	1.00	0.38	0.52	0.59	2.00	0.53	0.66
06-0110	0.40	0.53	0.47	0.41	0.69	2.00	0.52	0.66
F. oxysporum
06-0342	0.40	0.63	0.52	0.41	0.63	2.00	0.45	0.39
06-0197	0.55	0.53	0.56	0.44	0.58	2.00	0.55	0.94

^a5FU, 5-fluorouracil; DRZ, dorzolamide; TIM, timolol. Bold values indicate that the FICI is in the synergistic range (<0.5).

TABLE 3.

MICs for the synergistically interacting antifungal and nonantifungal ophthalmic agents natamycin and timolol

Strain	MIC (μg/ml) of each agent alone		MIC (μg/ml) of each agent in the combination		FICI	Avg FICI
	Natamycin	Timolol	Natamycin	Timolol
F. solani
132.02	16	>6,800	8	425	0.53	0.52
			4	3,400	0.50
06-0487	32	>6,800	16	213	0.52	0.45
			8	1,700	0.38
06-0330	32	>6,800	16	213	0.50	0.41
			8	850	0.31
06-0133	32	>6,800	16	53	0.50	0.41
			8	850	0.31
06-0111	32	>6,800	16	213	0.52	0.52
06-0110	32	>6,800	16	53	0.50	0.41
			8	850	0.31
F. oxysporum
06-0342	32	>6,800	16	213	0.52	0.41
			8	850	0.31
06-0197	16	>6,800	8	53	0.50	0.44
			4	1,700	0.38

TABLE 4.

MICs for the synergistically interacting antifungal and nonantifungal ophthalmic agents natamycin and 5-fluorouracil

Strain	MIC (μg/ml) of each agent alone		MIC (μg/ml) of each agent in the combination		FICI	Avg FICI
	Natamycin	5-Fluorouracil	Natamycin	5-Fluorouracil
F. solani
132.02	16	1,250	8	78	0.56	0.59
			2	625	0.63
06-0487	16	2,500	8	156	0.56	0.56
			4	625	0.50
			2	1,250	0.63
06-0330	32	2,500	16	10	0.50	0.40
			8	156	0.31
			4	625	0.38
06-0133	32	2,500	16	10	0.50	0.34
			8	78	0.28
			4	313	0.25
			2	625	0.31
06-0111	32	2,500	16	10	0.50	0.44
			8	156	0.31
			4	625	0.38
			2	1,250	0.56
06-0110	32	2,500	16	10	0.50	0.40
			8	156	0.31
			4	625	0.38
F. oxysporum
06-0342	32	5,000	16	78	0.52	0.40
			8	313	0.31
			4	1,250	0.38
06-0197	16	5,000	8	156	0.53	0.55
			4	1,250	0.50
			2	2,500	0.63

These data suggest that the antifungal activity of both natamycin and voriconazole is increased in vitro when they are used in combination with nonantifungal ophthalmic agents that are already used for the treatment of other ophthalmic conditions, such as squamous cell carcinoma, glaucoma, and band keratopathy. Of note, all of the nonantifungal drugs used in this study were evaluated at concentrations that were either equal to or less than the concentrations present in commercially available ophthalmic formulations, suggesting that supraphysiologic doses would not be required in order to obtain a clinical effect. Timolol is an FDA-approved commercially available medication, while ETDA and 5- fluorouracil are compounded for ocular use. Of the four nonantifungal drugs evaluated, 5-fluorouracil, EDTA, and timolol yielded the most promising results. 5-Fluorouracil is a potent chemotherapeutic agent that halts DNA replication via inhibition of thymidylate synthase and is also the active component of flucytosine, a prodrug used in combination with amphotericin B to treat cryptococcal meningitis (14–18). EDTA is a chelating agent that forms stable complexes with polyvalent cations, such as Ca²⁺ and Fe³⁺ (19). Timolol is a competitive antagonist of β-adrenergic receptors located at various sites throughout the human body, including the ciliary epithelium of the eye (20, 21). Additional studies are needed to clarify how these agents increase the antifungal properties of either natamycin (which binds to ergosterol within the plasma membrane and inhibits membrane transport [22]) or voriconazole (which disrupts ergosterol biosynthesis via cytochrome P450 [CYP450]-dependent 14α-sterol demethylase inhibition (23)).

While the mechanism of action of 5FU is known, it is possible that 5-fluorouracil may additionally inhibit CYP450-associated enzymes in Fusarium, as it has been observed to inhibit CYP450 2C9 in humans (24). This may contribute to the indifferent interaction observed between 5-fluorouracil and voriconazole and the synergistic interactions observed between 5-fluorouracil and natamycin (25). With respect to EDTA, prior work has suggested that this compound may interact synergistically with both natamycin and various azoles (26), although the mechanism that underpins this phenomenon remains unknown. It has been demonstrated that EDTA alters numerous aspects of metabolism in Cryptococcus neoformans, including biofilm formation, capsular polysaccharide remodeling, and extracellular vesicular secretion (27). An influence of EDTA on cell membrane synthesis could potentially explain the indifferent interactions observed when it is combined with either natamycin or voriconazole.

To the best of our knowledge, timolol has not previously been described as having intrinsic antifungal activity, which is consistent with the findings from the current study. It is not clear that timolol should interact with any fungal enzyme targets, as orthologues of human β₁- and β₂-adrenergic receptors (ADRB1 and ADRB2, respectively) are not annotated in fungi (28). However, the beta blocker propranolol has been shown to inhibit the phospholipase activity, hyphal morphogenesis, and pathogenesis of Candida albicans in vitro and in murine models of infection (29, 30). Moreover, a prior study has demonstrated that carvedilol (another β-adrenergic receptor antagonist) alters cell membrane permeability in the filamentous fungus Cunninghamella echinulata (31). It is unclear if timolol impacts Fusarium membrane permeability, given the marked structural difference from carvedilol, but one possible explanation for the observed synergy between timolol and natamycin involves increased fungal cell membrane permeability. Alternatively, it is possible that the synergistic interaction observed between natamycin and timolol is mediated via the maleic acid component of timolol maleate. Indeed, maleic acid has previously demonstrated antimicrobial properties against Enterococcus faecalis (32). However, we did evaluate maleic acid in combination with natamycin and concluded that it was not sufficient (FICIs, 0.58 to 0.69) to account for the synergistic interactions observed between natamycin and timolol. Taken together, our data and the limited data on beta blockers and antifungal activity suggest that additional research should explore this family of compounds in the context of fungal infections.

Importantly, we also evaluated the interaction both between timolol and voriconazole and between timolol and natamycin against five Aspergillus fumigatus isolates to assess the potential for synergistic or antagonistic interactions. Antagonistic interactions with A. fumigatus could prove devastating in the setting of fungal keratitis, especially if fungal culture results were not immediately available at the time of pharmacotherapy. These A. fumigatus isolates were uniformly more susceptible to both natamycin (MICs, 4 μg/ml) and voriconazole (MICs, 0.25 to 1 μg/ml) than the Fusarium isolates evaluated, and there was no evidence of antagonism between either natamycin and timolol (FICI, 1.00) or voriconazole and timolol (FICI, 1.00). These results suggest that in the setting of fungal keratitis caused by A. fumigatus, the addition of timolol to either natamycin or voriconazole therapy would likely neither improve nor worsen the therapeutic effect of these antifungal agents. The observed interaction between timolol and either natamycin or voriconazole thus appears to be fungus dependent, with Fusarium isolates demonstrating increased susceptibility to these antifungal agents in the presence of timolol and A. fumigatus isolates demonstrating effectively no change in antifungal susceptibility as a function of timolol.

In conclusion, the present study demonstrates in vitro synergistic interactions between first-line antifungal agents used in the treatment of fungal keratitis and nonantifungal ophthalmic agents used in the treatment of other ocular diseases. These findings suggest that the addition of a nonantifungal agent to either natamycin or voriconazole may increase the antifungal properties associated with these drugs. While the underlying mechanisms are unclear, these findings are particularly important in the context of infections caused by Fusarium species, which are often intrinsically resistant to contemporary ocular antifungal drugs. In addition, because the selected nonantifungal agents have already demonstrated safety when applied to the ocular surface, combined formulations could be developed and deployed quickly. A potential limitation of the present study is that it involves exclusively in vitro experiments, and it has been shown with Fusarium species that in vitro susceptibility does not necessarily correlate with the in vivo clinical response (33). Therefore, future work should evaluate the translation of these preliminary in vitro findings to in vivo animal or human trials.

MATERIALS AND METHODS

A total of eight Fusarium species ocular isolates were evaluated in the present study. Fusarium solani strains 06-0110, 06-0111, 06-0133, 06-0330, and 06-0487 and Fusarium oxysporum strains 06-0197 and 06-0342 were obtained from the University of California at San Francisco-Proctor Foundation. F. solani strain DUMC 132.02 (also known as Maren Klich strain 2748) was kindly obtained from Wiley Schell, Duke University Medical Center. Species-level identifications were affirmed via sequencing of the nuclear ribosomal internal transcribed spacer (ITS) and translation elongation factor 1-alpha (TEF) (34, 35). Fungal isolates were cultured on potato dextrose agar (PDA; comprised of Difco potato dextrose broth and Difco granulated agar [both from Becton, Dickinson, Franklin Lakes, NJ]) for 72 h at 30°C and then subcultured in liquid potato dextrose broth (PDB; Difco potato dextrose broth [Becton, Dickinson]) for 72 h with vigorous shaking at 30°C to produce microconidia. Liquid cultures were filtered with Miracloth to remove mycelia, and culture filtrates were centrifuged (3,000 rpm for 5 min) and washed twice with phosphate-buffered saline (PBS; Mediatech Inc., Manassas, VA). Microconidia were then resuspended in RPMI 1640 (Mediatech Inc.) to a density of 5 × 10⁵ conidia/ml. One hundred microliters of this suspension was added to each well of a flat-bottomed 96-well plate. Five Aspergillus fumigatus isolates (AF100_9.2, AF100_3A, AF293, ATCC 46455, and CEA10) were also evaluated for experiments involving interactions between timolol and either natamycin or voriconazole. The AF100 isolates came from the sputum of a cystic fibrosis patient at Dartmouth Hitchcock Medical Center. Culture conditions were as described above for Fusarium, except that all steps were performed at 37°C in a 5% CO₂ environment.

Stock solutions of natamycin (1.0 mg/ml; Sigma-Aldrich, St. Louis, MO), voriconazole (25.0 mg/ml; Sigma-Aldrich), and 5-fluorouracil (40.0 mg/ml; Sigma-Aldrich) were solubilized in DMSO (Alfa-Aesar, Haverhill, MA), while stock solutions of dorzolamide hydrochloride (22.3 mg/ml; Sigma-Aldrich), Na₄ EDTA (52.0 mg/ml; MP Biomedicals, Santa Ana, CA), and timolol maleate (27.2 mg/ml; Sigma-Aldrich) were solubilized in distilled water. All serial dilutions were made in distilled water. Fifty microliters of one antifungal agent and 50 μl of one nonantifungal ophthalmic solution were added to each well of the 96-well plate containing the 5 × 10⁵-conidia/ml suspension described above in a checkerboard pattern, with rows containing decreasing concentrations of the antifungal agent and columns containing decreasing concentrations of the nonantifungal agent. Ninety-six-well plates were sealed with Parafilm and incubated at 37°C for 3 days. Fungal growth was assessed visually via microscopy, and the concentration in the well in which a lack of fungal growth was observed was determined to be the MIC.