Small molecules restore azole activity against drug-tolerant and drug-resistant Candida isolates

Abstract

Each year, fungi cause more than 1.5 billion infections worldwide and have a devastating impact on human health, particularly in immunocompromised individuals or patients in intensive care units. The limited antifungal arsenal and emerging multidrug-resistant species necessitate the development of new therapies. One strategy for combating drug-resistant pathogens is the administration of molecules that restore fungal susceptibility to approved drugs. Accordingly, we carried out a screen to identify small molecules that could restore the susceptibility of pathogenic Candida species to azole antifungals. This screening effort led to the discovery of novel 1,4-benzodiazepines that restore fluconazole susceptibility in resistant isolates of Candida albicans, as evidenced by 100-1,000-fold potentiation of fluconazole activity. This potentiation effect was also observed in azole-tolerant strains of C. albicans and in other pathogenic Candida species. The 1,4-benzodiazepines selectively potentiated different azoles, but not other approved antifungals. A remarkable feature of the potentiation was that the combination of the compounds with fluconazole was fungicidal, whereas fluconazole alone is fungistatic. Interestingly, the potentiators were not toxic to C. albicans in the absence of fluconazole, but inhibited virulence-associated filamentation of the fungus. We found that the combination of the potentiators and fluconazole significantly enhanced host survival in a Galleria mellonella model of systemic fungal infection. Taken together, these observations validate a strategy wherein small molecules can restore the activity of highly used anti-infectives that have lost potency. IMPORTANCE In the last decade, we have been witnessing a higher incidence of fungal infections, due to an expansion of the fungal species capable of causing disease (e.g., Candida auris), as well as increased antifungal drug resistance. Among human fungal pathogens, Candida species are a leading cause of invasive infections and are associated with high mortality rates. Infections by these pathogens are commonly treated with azole antifungals, yet the expansion of drug-resistant isolates has reduced their clinical utility. In this work, we describe the discovery and characterization of small molecules that potentiate fluconazole and restore the susceptibility of azole-resistant and azole-tolerant Candida isolates. Interestingly, the potentiating 1,4-benzodiazepines were not toxic to fungal cells but inhibited their virulence-associated filamentous growth. Furthermore, combinations of the potentiators and fluconazole decreased fungal burdens and enhanced host survival in a Galleria mellonella model of systemic fungal infections. Accordingly, we propose the use of novel antifungal potentiators as a powerful strategy for addressing the growing resistance of fungi to clinically approved drugs.

Keywords: 1,4-benzodiazepines; Candida albicans; azole resistance; azole tolerance; drug combinations.

Fig 1

Chemical structures of control compounds (A) and the set of 26 1,4-benzodiazepines (1,4-BZDs) (B) tested. The most active 1,4-BZDs are indicated in blue.

Fig 2

Screening strategy for 1,4-benzodiazepines (1,4-BZDs) and control compounds. (A) Fluconazole (FLC) susceptibility profiles of Candida albicans isolates SC5314, SP-945, and P60002 on disk diffusion assays, including their mean RAD₅₀, RAD₂₀, and fraction of growth at 20% inhibition (FoG₂₀) levels. (B) A lawn of C. albicans cells was plated on YPD agar and grown in the presence of a drug disk (FLC, 25 µg) for 48 hours. The size of the radius at 50% and 20% growth inhibition (RAD₅₀ and RAD₂₀, respectively) was measured as a proxy for drug susceptibility. The FoG₂₀ was measured as a proxy for tolerance. Susceptibility and tolerance were determined using diskImageR. (C) Following 48 hours of growth, the FLC disk was replaced with a glucose disk (G, 8 mg) and the plates were grown for an additional 48 hours. For drug combinations, YPD plates contained 1,4-BZDs at 100 µM. Plates show example profiles of tolerant and resistant strains with FLC alone (left, with DMSO used as control). Plates with FLC and potentiators (right) illustrate fungistatic or fungicidal effects of different drug combinations following additional growth on glucose disks.

Fig 3

Impact of the seven 1,4-benzodiazepines (1,4-BZD)/fluconazole (FLC) drug combinations on FLC susceptibility, tolerance, and cell survival of Candida albicans. Compounds were screened using FLC (25 µg) disk diffusion assays on YPD plates containing 100 µM 1,4-BZDs or the same volume of DMSO (vehicle). Panels show FLC susceptibility (RAD₅₀ and RAD₂₀, (A) and tolerance [fraction of growth at 20% drug inhibition (FoG₂₀), (B)] following 48 hours of growth. After 48 hours, FLC disks were replaced with glucose disks (G, 8 mg), grown for an additional 48 hours, and susceptibility and tolerance were quantified again. Heatmaps show average values for three or more biological replicates. (C–D) Checkerboard assays showing the impact of drug combinations of FLC (starting at 128 µg/mL) with the seven 1,4-BZDs and sortin2 (starting at 100 µM) on FLC susceptibility (MIC₉₀, C) and tolerance (SMG, D) for C. albicans isolates SC5314, SP-945, and P60002. (E) FICI₉₀ scores for combinations of potentiators and FLC based on MIC₉₀ values. Left-side table indicates the type of FLC-potentiator interaction based on the FICI score. Heatmap shows average values from three to four biological replicates.

Fig 4

Time-kill assays in which Candida albicans strains were cultured in YPD media with DMSO (control), fluconazole (FLC; 128 µg/mL) or drug combinations of FLC (128 µg/mL) and 1,4-benzodiazepines (1,4-BZDs) (100 µM) over 24 hours. Cells were plated at 0, 4, 8, 12, and 24 hours during growth for colony forming unit (CFU) determination (A). The percent of live cells incubated with different treatments is shown over time relative to the starting inoculum measured at 0 hour (B). Plots show an average of three independent experiments, error bars show ± SEM.

Fig 5

Impact of potentiators on Candida albicans cell growth. (A, B) Growth curves of C. albicans isolates (SC5314, SP-945, and P60002) on YPD at 30°C and 37°C in the presence of DMSO, sortin2 (100 µM), or 1,4-benzodiazepines (1,4-BZDs) (100 µM). Lines show average values from three biological replicates, error bars represent ± SEM. (C) Heatmap shows relative doubling times of C. albicans cells in the presence of potentiators, normalized to DMSO controls.

Fig 6

Cytotoxicity of 1,4-benzodiazepines (1,4-BZDs) on human peripheral blood mononuclear cells (PBMCs, A). Lytic activity of the 1,4-BZDs was assessed with red blood cells (RBCs, B). PBMCs and RBCs were incubated with DMSO, PBS, Triton (positive control), or 1,4-BZDs at different concentrations, and treatment toxicity was inferred by quantifying LDH release for PBMCs (A) or by measuring hemolytic titers for RBCs (B). Assays utilized blood samples from two donors, each performed with two technical replicates, error bars show ± SEM.

Fig 7

Changes in FLC susceptibility (MIC₅₀ and MIC₉₀, (A) and tolerance (SMG, B) of other yeast species using combinations of potentiators (100 µM) and fluconazole (FLC) (from 0 to 256 µg/mL). Values are shown relative to DMSO controls. R, FLC-resistant isolates (MIC₅₀ ≥16 µg/mL for Candida glabrata and MIC₅₀ ≥4 µg/mL for all other species); T, FLC-tolerant isolates (SMG ≥0.5). Heatmaps show average values from three to four biological replicates. (C) Lines indicate phylogenetic relationships between different species, adapted with permission from Ref. (49).

Fig 8

1,4-Benzodiazepines (1,4-BZDs) potentiate other azole antifungals. Heatmaps show drug susceptibility (RAD₅₀ and RAD₂₀, A) and tolerance [fraction of growth at 20% inhibition (FoG₂₀), B] when Candida albicans isolates were plated with combinations of potentiators and other azole drugs. Compounds were screened using disk diffusion assays on YPD plates containing 1,4-BZDs (100 µM), sortin2 (100 µM), or DMSO (no drug) and disks of itraconazole (ITC, 50 µg), ketoconazole (KCA, 10 µg), posaconazole (POS, 5 µg), or voriconazole (VOR, 1 µg). Heatmaps show average values from three biological replicates.

Fig 9

1,4-Benzodiazepines (1,4-BZDs) potentiate inhibitors of ergosterol and sphingolipid biosynthesis. (A) Terbinafine (TERB) and fenpropimorph (FEN) target ergosterol biosynthesis via Erg1 and Erg24/Erg2, respectively, while myriocin (MYO) targets sphingolipid biosynthesis via Lcb1. (B) Changes in FLC, TERB, FEN, or MYO susceptibility (MIC₅₀) and tolerance (SMG) when Candida albicans isolates were treated with combinations of 1,4-BZDs (at 100 µM) and serial dilutions of fluconazole (FLC) (0 to 128 µg/mL), TERB (0 to 32 µg/mL), FEN (0 to 32 µg/mL), or MYO (0 to 128 µg/mL), relative to DMSO controls. FLC data are included for reference. Deletion mutants of Upc2, the central regulator of ergosterol synthesis genes, and Erg5, a cytochrome P450 enzyme catalyzing the last step in ergosterol biosynthesis, were also tested with FLC and 1,4-BZDs; symbols adjacent to strain names denote their corresponding lineage/WT controls, all constructed in the SC5314 background. All values are shown relative to DMSO controls and indicate the average from three to four biological replicates. (C) C. albicans cells stained with the BODIPY 493/503 dye following 4-hour incubation with either DMSO, FLC (128 µg/mL), MYO (32 µg/mL), or combinations of FLC (128 µg/mL) and 1,4-BZDs (100 µM). Images show representative cells for the respective conditions, at 60X magnification. The relative dilution and distribution of yeast cells on the slides necessitated the creation of composite images to ensure the best representation of the phenotypes. Scale bar, 10 µm.

Fig 10

Impact of drug combinations on Candida albicans mutants of drug transporters and their regulators. (A) Schematic shows transporters from the classes of ABC (ATP-binding cassette, Adp1, Cdr1, Cdr2, and Mdl2) and MSF (major facilitator superfamily, Flu1 and Mdr1) as well as their corresponding transcription factor regulators (Tac1, Mrr1, and Upc2). (B) Impact of 1,4-benzodiazepines (1,4-BZDs) on the fluconazole (FLC) susceptibility (MIC₅₀) and tolerance (SMG) of different single and combined deletion mutants. Symbols denote the corresponding WT strains for each mutant, all constructed in the SC5314 background. Heatmaps show average values from three biological replicates, relative to corresponding DMSO controls. (C) Impact of 1,4-BZDs on intracellular rhodamine 6G (R6G) accumulation. (D-E) R6G fluorescence intensity levels were measured after 4-hour incubation with R6G (1 µg/mL) and treatment with either DMSO (no drug), sortin2 (100 µM), 1,4-BZDs (100 µM), or beauvericin (10 µg/mL, as positive control). Strains examined include SC5314, SP-945, and P60002 (D), as well as cdr1, cdr2, cdr1 cdr2 mutants with the corresponding WT parental strain (E). Heatmaps show average R6G values from three or more biological replicates, relative to corresponding DMSO controls.

Fig 11

Impact of drug combinations on Candida albicans filamentation. (A) Cell morphologies of strains SC5314, SP-945, and P60002 treated with either DMSO, fluconazole (FLC) (10 µg/mL), 1,4-benzodiazepines (1,4-BZDs) (PA158 or PA162, 100 µM), or a combination of FLC (10 µg/mL) and 1,4-BZDs (100 µM). Cells were incubated in YPD at 37°C for 2 hours prior to imaging. Images show representative cells for the respective conditions, at 40X magnification. Scale bar, 10 µm. (B) The fraction of each morphology was quantified for at least 200 cells per condition from four biological replicates. Asterisks indicate significant differences relative to DMSO, t-test, *P < 0.05. Error bars represent ± SEM. (C) Impact of 1,4-BZDs (100 µM) on FLC susceptibility (MIC₅₀) and tolerance (SMG) for mutants affecting cell morphology (see schematic on the right side) and corresponding parental strain. Changes in MIC₅₀ and SMG are shown relative to levels observed with DMSO treatment. Heatmaps show average values from three to six biological replicates.

Fig 12

Impact of drug combinations on Galleria mellonella survival during systemic fungal infection. (A) Larvae of G. mellonella were systemically infected with Candida albicans isolates (3 × 10⁵ cells per larva) and subsequently treated with either no drug (DMSO as vehicle), fluconazole (FLC, 20 µg/mL), sortin2 (50 µM), 1,4-benzodiazepines (1,4-BZDs) (PA158 or PA162, 50 µM), or a combination of FLC (20 µg/mL) and 1,4-BZDs (50 µM) in volumes of 10 µL/larva. Additional control groups of larvae were systemically injected with PBS and treated with either DMSO, FLC, sortin2, or 1,4-BZDs in the absence of fungal infection (B). Survival of larvae infected with SC5314 (C), SP-945 (D), or P60002 (E) was monitored daily for 10 days for three to five groups of larvae with 12 larvae per group for each condition (n = 36–60 larvae per group). Asterisks indicate significant differences for comparisons between FLC and drug combinations (FLC vs FLC + sortin2, FLC vs FLC + PA158, FLC vs FLC + PA162), colored according to the corresponding lines, based on Log-rank (Mantel–Cox) tests. All statistical comparisons between each two groups are included in Table S1E. (F-H) Fungal burdens from larvae treated with single agents or drug combinations recovered before or at 4 days post infection (n = 12 larvae/group). Gray dots show larvae that had succumbed to infection before or at 4 days, white dots show larvae sacrificed on day 4. Error bars represent ± SEM, asterisks indicate significant differences based on unpaired t-tests, * P < 0.05; ** P < 0.01.