A screen to identify antifungal antagonists reveals a variety of pharmacotherapies that induce echinocandin tolerance in Candida albicans

ABSTRACT

Through screening a comprehensive collection of drugs approved for human use, we identified over 20 that oppose the antifungal activity of the echinocandins upon the infectious yeast, Candida albicans. More detailed evaluation of five such drugs, including the atypical antipsychotic aripiprazole and the tyrosine kinase inhibitor ponatinib, indicated they promote C. albicans survival following exposure to the echinocandin antifungals. The activity of the five selected antagonists was dependent upon the Mkc1p mitogen-activated protein kinase pathway; however, ponatinib was paradoxically shown to suppress phosphorylation and therefore activation of Mkc1p itself. Components of several other signaling pathways are also required, including those of calcineurin and casein kinase-2, suggesting the observed antagonism required much of the cell wall stress responses previously described for C. albicans. Transcriptome analysis revealed that the antagonists stimulated the expression of genes involved in xenobiotic and antifungal resistance and suppressed the expression of genes associated with hyphal growth. Thus, the echinocandin antagonistic drugs modulate C. albicans physiology in ways that could impact its pathogenicity and/or response to therapeutic intervention. Finally, a mutant lacking the Efg1p transcription factor, which has a central role in the activation of C. albicans hyphal growth, was found to have intrinsically high levels of echinocandin tolerance, suggesting a link between modulation of morphogenesis-related signaling and echinocandin tolerance.

INTRODUCTION

The Centers for Disease Control and Prevention estimates the direct healthcare costs of treating fungal disease to be over $7 billion each year in the U.S.A. alone (1). Infections involving fungal invasion to the deeper organs are of greatest concern and predominantly occur in immunocompromised individuals. Several Candida spp. account for more than 75% of disseminated fungal infections in the U.S.A. (2, 3), with attributable mortality rates of 35%–75% (4–6). They are also a common cause of mucosal diseases, including oral and vaginal thrush, with Candida albicans being the most prevalent and virulent (7). The echinocandins are the first-line therapy for invasive Candida infections and disrupt fungal cell wall synthesis through inhibiting β (1,3)-D-glucan synthase (8). However, nearly a third of patients with invasive candidiasis fail to respond to treatment with an echinocandin (9–12), and outcomes are significantly worse for neutropenic patients (13). The incidence of echinocandin resistance is <0.5% for C. albicans (14, 15); thus, genetically encoded antifungal resistance does not account for the majority of treatment failures (16–19). Several host-related factors have been proposed to account for the discordance between in vitro susceptibility tests and patient outcomes (18). Certainly, rapid diagnosis and administration of an appropriate antifungal agent can reduce patient mortality (20, 21). The severity of patient immunosuppression may also impact the response to antifungal therapy (19, 22). Either way, many treatment failures remain unexplained. Given the limited therapeutic options available, it is critical to identify and address the causes of therapeutic failure to improve patient outcomes.

As natural inhabitants of the human gastrointestinal (GI) and reproductive tracts, several medically important Candida spp. are routinely exposed to medications and other pharmacologically active substances consumed by their human host. Given the fundamental conservation of metabolic and signaling modules between eukaryotic genera, it is likely that many drugs designed to modulate human physiology may dysregulate analogous pathways in resident or infecting fungi (23–26). Yet, the influence of most medications upon fungal physiology, pathogenicity, and antifungal susceptibility remains largely uncharacterized. Moreover, it is unclear if such drug-fungus interactions have the potential to influence the incidence or outcomes of mycotic disease. We and others have previously reported that a wide variety of drugs approved for human use can diminish the capacity of the azole antifungals to inhibit the growth of pathogenic Candida spp. (27, 28). We also identified a smaller collection that moderates C. albicans sensitivity to the echinocandin, caspofungin (27). These interactions are potentially important determinants of clinical outcome, as chemically induced antifungal resistance occurring within a patient receiving an antagonistic medication would not be detected by in vitro antifungal susceptibility testing methods. However, the drug collections screened in previous studies only offered partial coverage of the current pharmacopeia. Additionally, the physiological impact of these drug-fungus interactions and their potential to impact interaction with the mammalian host remain unknown. The purpose of this study was to identify medications that alter C. albicans sensitivity to the echinocandins and examine the mechanisms by which they act.

MATERIALS AND METHODS

Growth conditions

C. albicans was grown on yeast extract-peptone-dextrose (YPD) agar plates at 30°C and incubated in YPD liquid medium unless otherwise stated. Select strains were graciously provided either by Dr. Brian M. Peters or Dr. Todd B. Reynolds. Kinase-defective mutants were kindly provided by Dr. Damian Krystan (29).

Drug stocks

Stock solutions of each antifungal (caspofungin, micafungin, and anidulafungin), as well as each compound (aripiprazole, cinacalcet, haloperidol, netupitant, and ponatinib), were prepared at 10 mM in dimethyl sulfoxide (DMSO) and diluted to required working concentrations. The Food and Drug Administration (FDA) collection was purchased from SelleckChem, with all 2,701 compounds dissolved in the respective solvents listed (either DMSO or water) at 10 mM in 384-well plates. Each was further diluted, and 1 µL volumes were dispensed into the flat-bottomed 384-well polystyrene plates used for the chemical screens.

Antifungal antagonism screens

Assays were performed similarly as described (28). The wells of the 384-well flat-bottomed assay plates were seeded with 1 µL of the 1 mM stock solutions of each library compound in DMSO or with DMSO alone (control). SC5314 was grown overnight in YPD medium at 30°C, and the cells were harvested by centrifugation, washed twice in phosphate-buffered saline (PBS), and resuspended at ∼6.25 × 10⁴ cells/mL in Roswell Park Memorial Institute (RPMI) 1640 medium buffered to pH 7 with [3-(N-morpholino)propanesulfonic acid] 2% glucose supplemented with the indicated concentration of caspofungin. A 39 µL volume of the cell-plus-antifungal drug suspension was then dispensed into each well of the drugged 384-well plates to achieve approximately 1,000 cells/well. Control wells had antifungal drug but no test/library compound (antifungal alone) or had neither the antifungal nor the library compound (minus drug/untreated control). The final concentration of DMSO was 0.55% in all wells, with each library compound supplied at a final concentration of 5 µM. After 72 hours of incubation at 35°C, growth was measured as OD_600nm without shaking. Wells with visible precipitation were excluded from analysis. Growth was compared to the antifungal alone and untreated control and expressed as percent growth. The growth of each drug-treated well was also compared to antifungal-alone wells, and the Z-score was calculated (30). “Hits” were identified as compounds with a Z-score of ≥3 in two independently run screens.

Antifungal susceptibility assays

Antifungal susceptibility testing was performed using the broth microdilution method as described in Clinical and Laboratory Standards Institute document M27-A3 (16), with minor modifications. Each echinocandin was diluted in DMSO, resuspended in Roswell Park Memorial Institute medium adjusted to pH 7 (RPMI-pH 7) (2% glucose) at twice the final concentration, and serially diluted. C. albicans strains were grown overnight in YPD at 30°C, resuspended at 1 × 10⁴ cells/mL in RPMI-pH 7, and 100 µL was transferred to wells of a round-bottomed 96-well plate containing an equal volume of diluted echinocandin solution. The final concentration of DMSO was 0.5% for all treatments, with drug-free control wells having DMSO alone. Plates were incubated without shaking for 72 hours at 35°C, with plates scanned every 24 hours using an EPSON Perfection v.700 Photo scanner. Experiments were performed in biological duplicate.

Growth kinetic assays

Growth curve assays were set up in 96-well plates using RPMI-pH 7 (2% glucose). C. albicans strains were grown overnight in YPD at 30°C, and the cell density was adjusted to 1 × 10⁴ cells/mL in the appropriate medium for the growth kinetic assays. An aliquot of cell suspension (100 µL) was mixed with an equal volume of twice the desired drug concentration of each antagonist. Cells were then incubated at 35°C inside a BioTek Cytation 5 plate reader with shaking for 48 hours, and OD_600nm was read every 30 min. Data were then analyzed via GraphPad Prism software. These assays were repeated in biological duplicate.

Cell survival and cell damage assays

For cell survival assays, the C. albicans parental strain GP1 and strain CAI4 + pKE1 NLUC were grown overnight in YPD at 30°C. Cells were washed and subcultured at 1 × 10⁶ cells/mL in 10 mL of RPMI-pH 7 (2% glucose) supplemented with 0.5% DMSO (vehicle), 0.4 µM, or 1.6 caspofungin alone or in the presence of 5 µM aripiprazole, cinacalcet, haloperidol, netupitant, or 2.5 µM ponatinib. Cells were incubated at 35°C for 6 hours plus or minus drug with shaking. After incubation, cells were spun down at 3,000 revolutions per minute (rpm) for 5 min and washed twice with sterile dH₂O. Cells were then resuspended in sterile dH₂O and serially diluted, and 100 µL of cell solutions was plated onto YPD plates. Plates were then incubated for 48 hours at 30°C, and cell viability was measured by counting colony-forming units (CFUs). For cell damage assays, cells were grown under the same conditions, and a portion of supernatant was collected for measuring luciferase activity. Experiments were repeated in biological triplicate.

For cell damage assays, supernatant was collected from the cells grown in the previously listed condition. Luciferase-based assays were performed as previously described (31). Experiments were performed in biological triplicate, and statistical significance was calculated via one-way analysis of variance (ANOVA) test.

RNA sequencing analysis

SC5314 was grown in YPD medium at 30°C overnight, then subcultured at 1 × 10⁶ cells/mL into 50 mL RPMI-pH 7 medium supplemented with either 5 µM aripiprazole, cinacalcet, haloperidol, netupitant, 2.5 µM ponatinib, or with 0.5% DMSO (vehicle control). Cells were then incubated at 35°C for 6 hours with shaking. Cells were then harvested at 4°C; the supernatant was removed; and cells were frozen at −80°C. Total cellular RNA was extracted using the hot phenol method (32). Novogene provided RNA library preparation and sequencing analysis as a fee-for-service. Messenger RNA was purified from total RNA using poly-T oligo attached magnetic beads. After fragmentation, the first-strand cDNA was synthesized using random hexamer primers, followed by the second-strand cDNA synthesis using either deoxyuridine triphosphate for directional library or deoxythymidine triphosphate for non-directional library. The library was checked with Qubit and real-time PCR for quantification, and a bioanalyzer was used for size distribution detection. Libraries were then pooled and sequenced on Illumina platforms. Genes were mapped to the SC5314 haploid genome assembly 22. Drug-responsive genes were identified as those significantly upregulated or downregulated compared to their respective vehicle controls, with significant gene expression identified as either >1 or <−1 log fold (P value of <0.05, adjusted P value of <0.05 to account for false discovery rate). Samples were prepared in independent biological duplicate, and log₂ fold was converted to fold change and averaged, and the average value was converted to log₂ fold for the term “AVG log₂ fold.”

SC5314 strain was inoculated in YPD and incubated overnight at 30°C. The next day, SC5314 was subcultured into 75 mL of pre-warmed RPMI-pH 7 (2 % glucose) supplemented with vehicle, 5 µM of aripiprazole, cinacalcet, haloperidol, netupitant, and 2.5 µM ponatinib alone or in combination with caspofungin. SC5314 was washed with PBS and diluted to an OD_600nm of 0.25 in RPMI-pH 7 medium using a spectrophotometer (BioPhotometer from Eppendorf). Flasks were incubated at 35°C with 250 rpm with shaking. After 2 hours of growth, cells were harvested at 4,000 rpm for 30 min, and the supernatant was removed. The cells were resuspended in PBS, transferred into a 1.5 mL screw-capped tube, and harvested at 13,000 rpm for 10 min. Tubes were immediately stored at −80°C.

Cell were removed from −80°C, thawed on ice, and then resuspended in 120 mL buffer containing (300 mM NaCl, 10 mM Tris, pH 8.0, 0.1% NP-40, and 10% glycerol) and containing 1× protease inhibitor cocktail, phosphatase inhibitor cocktail set II (Millipore Sigma, 1:100 dilution), and 1 mM dithiothreitol (33). Glass beads equivalent to the size of a pellet were added into the tube, and the tube was kept in the bead beater Bullet Blender Gold (Next Advance) operated for five cycles of 1 min each, with 1 min incubations on ice for each cycle. After cell lysis, whole cell extract was collected in a fresh 1.7 mL Eppendorf tube and kept on ice. Protein quantification was performed by Bradford assay (Thermo Fisher Scientific, cat #1856210), and protein concentration was calculated comparing the result to a standard curve.

Thirty micrograms of protein sample was mixed with 4× loading buffer (from Bio-Rad, cat #1610747), and protein samples were heated for 5 min at 100°C in a thermocycler. Protein samples were cooled to room temperature and loaded onto a 10% SDS-PAGE gel. The gel was run at a constant 100 V until the dye front reached the bottom of the gel. Proteins were transferred onto nitrocellulose membrane (0.2 µM from Bio-Rad, cat #162-0112) using wet transfer for 1 hour at a constant voltage of 100 V. After transfer, membranes were rinsed with distilled water and stained with Ponceaus S stain (from Thermo Fisher Scientific, cat #A40000279), and blots were imaged after rinsing with distilled water. Blots were then incubated in blocking buffer in Tris-buffered saline with 0.1% Tween 20 (TBST) detergent, pH 7.5 (blocking grade blocker from Bio-Rad cat #170-6404) (TBST diluted from 10× stock from Bio-Rad cat #170-6435) for 1 hour at room temperature. Blots were then incubated overnight with rabbit antiphospho-p44/42 antibody (from Cell Signaling, cat #4370) diluted to 1:2,000 in TBST with (5% non-fat skimmed milk powder in TBST, pH 7.5) at 4°C with shaking overnight. The next day, primary antibodies were removed, and blots were washed with TBST for 10 min three times. After the final wash, blots were incubated for 1 hour at room temperature with horseradish peroxidase-conjugated goat antirabbit IgG (H + L) (from Invitrogen, cat #31460) diluted to 1:10,000 in TBST with 5% non-fat skimmed milk powder. After incubation, non-bound antibodies were removed, and blots were washed thrice as above. Finally, the signal was detected using ECL (SuperSignal West Femto Luminol/Enhancer Solution and SuperSignal West Femto Stable Peroxide Buffer, Thermo Fisher Scientific), and images were captured using Molecular Imager ChemiDoc XRS+ with Image Lab Software (from Bio-Rad). Experiments were performed in biological quadruplicate, and statistical significance was calculated via one-way ANOVA test.

RESULTS

Several drugs approved for human use oppose the antifungal activity of echinocandins upon C. albicans

To identify medications that potentially modulate the efficacy of echinocandin therapy, we conducted a simple screen. C. albicans (strain SC5314) was suspended in RPMI medium supplemented with approximately 4× the minimum inhibitory concentration (MIC) (0.78 µM) of caspofungin and dispensed into 384-well plates arrayed with a collection of 2,701 pharmacologically active small molecules, most of which are approved for human use by the FDA, the European Medicines Agency, and other approval agencies. Each compound was provided to a final concentration of 5 µM, and control wells contained either no drugs or antifungal (DMSO vehicle growth control), or caspofungin alone. Given the cidal mode of action of the echinocandins, we measured fungal growth in each well as OD_600nm after 72 hours of incubation at 35°C. This strategy was intended to identify drugs that promote fungal survival in the presence of caspofungin (i.e., tolerance), as well as any that induce outright resistance (i.e., insensitivity). Compounds were called hits if they restored fungal growth in the presence of caspofungin (Z-score ≥3) in each of two independent replicate screens. A total of 22 hits, corresponding to 21 distinct chemical entities, were identified as opposing the antifungal activity of caspofungin upon C. albicans (Table 1). This included (i) aripiprazole, an atypical antipsychotic that we had previously reported to diminish the antifungal activity of the azole antifungals upon C. albicans (34); (ii) cinacalcet, a calcimimetic used to treat hyperparathyroidism; (iii) the antipsychotic haloperidol; (iv) the antiemetic netupitant; and (v) ponatinib, a multi-targeted tyrosine-kinase inhibitor used to treat leukemia. These five drugs were selected as representative hits for further evaluation, and their activity was compared.

TABLE 1.

Caspofungin antagonists identified from the SelleckChem FDA collection^a,b

AVG % growthc	AVG Z-scoresc (CAS alone)	Product name	Human target	C. albicans homolog (% ID)
43.9	29.9	Ponatinib (AP24534)	Bcr-Abl:FGFR:PDGFR:VEGFR	Approx. 40 kinase targets (26%–35%)
45.1	28.4	Haloperidol	5-Hydroxytryptamine receptor 2C, d(2) dopamine receptor	NA, NA
29.4	15.1	Sanguinarine chloride	PP2C	Cch1p (20%)
22.3	13.5	Cinacalcet HCl	Extracellular calcium-sensing receptor	NA
3.4	13.2	Oxaliplatin	DNA	NA
18.9	11.5	Aripiprazole	5-Hydroxytryptamine receptor 2C, d(2) dopamine receptor	NA, NA
20	11.4	Netupitant	Neuropeptide substance-P receptor	NA
21.5	11.1	Sanguinarine	PP2C	Cch1p (20%)
13.7	9.5	Fesoterodine fumarate	Muscarinic acetylcholine receptor M3	NA
14.1	8.9	Proanthocyanidins	NLRP3-inflammasome among other immune inflammatory response pathways	NA
11.2	7.8	Costunolide	Telomerase	NA
12.3	6.6	Vorapaxar	Proteinase-activated receptor 1	NA
9.5	6.4	Bambuterol HCl	Beta-2 adrenergic receptor	NA
10.1	6.3	Methyldopa	Aromatic-L-amino acid decarboxylase, alpha-2A adrenergic receptor	NA, NA
10.9	6.0	Acarbose	Maltase-glucoamylase, sucrase-isomaltase, pancreatic alpha-amylase	Gca1p (33%–34%), Gca2 (33%–39%), Rot2p (28%)
8.5	5.0	Suprofen	Prostaglandin G/H synthase 1, prostaglandin G/H synthase 2	NA, NA
8.2	5.0	Nicardipine HCl	Voltage-dependent L-type calcium channel subunit alpha-1C, beta-2, alpha-2/delta-1, and subunit-1D	Cch1p (21%)/Syg1p (20%), NA, NA, Cch1p (21%)
7.2	4.8	Biotin (vitamin B7)	Vitamin	NA
6.3	3.9	Iopamidol	Diagnostic imaging agent for X-rays	NA
5.7	3.6	Yangonin	Cannabinoid CB1 receptor	NA
5.4	3.4	Clindamycin alcoholate	Large ribosomal subunit uL1 (Staphylococcus aureus)	NA
5.1	3.2	Bethanechol chloride	Muscarinic acetylcholine receptor M3, M5	NA, NA

^a Wild-type C. albicans strain SC5314 was dispensed in RPMI-pH 7 2% glucose supplemented with 4× MIC of CAS into 384-well plates arrayed with the SelleckChem FDA collection to a final concentration of 5 µM. Plates were incubated for 72 hours at 35°C, and growth quantified as OD_600nm. Growth is expressed relative to untreated (DMSO vehicle alone/minus antifungal) control wells (% growth). Hits were called as compounds restoring growth (Z-score of ≥3 compared to antifungal-alone control) in two independently run experiments.

^b CAS, caspofungin; NA, non-applicable; PP2C, protein phosphatase 2C.

^c Average of two independently conducted screens.

Initial dose-response assays confirmed that all five drugs opposed the antifungal activity of caspofungin, although the degree of protection conferred by each varied. In all cases, their protective effect upon C. albicans was progressively more apparent at later time points (i.e., 48 and 72 hours) (Fig. 1A). When read after 24 hours of incubation, caspofungin MICs were comparable for all treatments, but residual growth was apparent at concentrations above the MIC in the presence of all five antagonists. Additionally, with the exception of ponatinib, the other four antagonists appear to preferentially restore C. albicans growth at the highest concentrations of caspofungin and therefore act to induce, or enhance, the previously described paradoxical growth observed at high concentrations of this antifungal (35). These data indicate that the antagonists identified do not substantially alter antifungal potency (i.e., MIC) but rather diminish the cidal capacity of caspofungin and promote tolerance. Subsequent checkerboard analysis revealed that many of the antagonists are protective at much lower concentrations than the 5 µM concentrations used in the initial screens (Fig. S1). Given that four of the five antagonists studied lack stand-alone antifungal activity, it was not possible to calculate fractional inhibitory concentration indices (36, 37). However, Bliss independence scores were calculated by assessing the maximum positive Bliss deviation, which corresponds to the most antagonistic interaction with caspofungin (38) (Table 2), and revealed substantial antagonistic effects over specific concentration ranges of each pairwise combination, with up to 65% growth restoration versus the minus antifungal control, by the 72 hour time point.

Fig 1.

Several medications alter Candida albicans tolerance of the echinocandins. Antifungal susceptibility assays were performed with C. albicans SC5314 in RPMI-pH 7 (2% glucose) and twofold dilutions of caspofungin (A), micafungin (B), or anidulafungin (C) in the presence of either 5 μM of aripiprazole (ARI), 5 μM cinacalcet (CINA), 5 μM haloperidol (HALO), 5 μM netupitant (NEPU), 2.5 μM ponatinib (PONA), or vehicle (DMSO). Plates were incubated at 35°C and imaged after 24, 48, and 72 hours. Images are representative of assays performed in biological duplicate.

TABLE 2.

Bliss independence analysis for antagonist and caspofungin interaction^a

Antagonist	Min Bliss deviation (%)	Max Bliss deviation (%)
Aripiprazole	−1.957 ± 1.287	53.301 ± 8.270
Cincalcet	−1.715 ± 0.452	59.623 ± 18.986
Haloperidol	−0.836 ± 0.827	65.886 ± 3.649
Netupitant	−2.377 ± 2.595	14.753 ± 9.119
Ponatinib	−0.943 ± 1.515	24.539 ± 14.352

^a Checkerboard assays were performed with caspofungin and each of the five antagonists with C. albicans SC5314 in RPMI-pH 7 (2% glucose). Plates were incubated for 72 hours (at 35°C); growth was quantified as OD_600nm and expressed relative to the minus drug/vehicle control wells (% growth). Bliss volumes were calculated for each combination of antagonist and caspofungin as the observed deviation from the expected growth, i.e., growth expected based on the combined activity of each individual agent at the same concentration. The minimum (min) and maximum (max) deviations observed for each drug combination are reported for the average ± standard deviation of three independently conducted checkerboard assays.

All five of the selected drugs also enhanced SC5314 growth in the presence of supra-MIC concentrations of micafungin and anidulafungin, the two other echinocandins approved for human use (Fig. 1B and C). All five also oppose the antifungal activity of caspofungin on C. albicans strains TW1 (39) and American Type Culture Collection 10231 (Fig. 2A and B). However, in the case of anidulafungin, the antagonists caused an overt increase in MIC rather than enhanced paradoxical growth. Thus, their antagonistic activity is neither strain nor echinocandin specific. Interestingly, at least three of the five antagonists (aripiprazole, cinacalcet, and haloperidol) increased the concentration of caspofungin required to suppress Candida parapsilosis growth (Fig. S2A), while both cinacalcet and netupitant also enhanced caspofungin tolerance in Candida tropicalis (Fig. S2B), although the effects were less pronounced than those observed for C. albicans. Thus, at least a subset of the echinocandin antagonists identified affect the response of multiple medically important Candida spp.

Fig 2.

Echinocandin antagonist activity is not strain specific. Caspofungin susceptibility assays were performed with C. albicans strain TW1 (A) or American Type Culture Collection 10231 (B) in RPMI-pH 7 (2% glucose) supplemented with 5 μM aripiprazole (ARI), 5 μM cinacalcet (CINA), 5 μM haloperidol (HALO), 5 μM netupitant (NEPU), 2.5 μM ponatinib (PONA), or vehicle (DMSO). Plates were incubated at 35°C and imaged after 24, 48, and 72 hours. Images are representative of assays performed in biological duplicate.

Antagonists diminish cidality and induce echinocandin tolerance

To further characterize the mode by which the selected antagonists protect C. albicans from the echinocandins, we conducted a more thorough assessment of cell growth, injury, and viability. Growth kinetics in the presence of inhibitory concentrations of caspofungin revealed that the antagonists permit growth but at much slower rates than in the absence of the antifungal (Fig. 3). Furthermore, their effect is more profound at high concentrations of the antifungal. This is consistent with them inducing a form of echinocandin tolerance rather than insensitivity, specifically, enhancing the paradoxical effect.

Fig 3.

Echinocandin antagonists partially restore Candida albicans growth in the presence of caspofungin. SC5314 was seeded RPMI-pH 7 (2% glucose) medium supplemented with vehicle (DMSO, A), 5 μM aripiprazole (B), 5 μM cinacalcet (C), 5 μM haloperidol (D), 5 μM netupitant (E), or 2.5 μM ponatinib (F) and a range of caspofungin (CAS) concentrations. Plates were incubated at 35°C for 48 hours, and OD_600nm was measured every 30 min. Results are representative of assays performed in biological duplicate.

A C. albicans strain expressing a small cytoplasmic luciferase (NanoLuc, Promega) (40) was used to compare the overall levels of damage sustained by populations of cells. Following 4 hours of caspofungin exposure at either 4× or 16× the MIC, NanoLuc released into the culture supernatant was approximately 100-fold higher than mock-treated controls (Fig. 4A), indicating substantial levels of cell injury and/or lysis. Comparable levels of NanoLuc were released in the presence of each of the antagonistic drugs, suggesting they do not diminish the damage sustained following caspofungin exposure per se. Cell viability was also quantified as CFUs after 4 hours of exposure to 4× MIC of caspofungin. As expected, 4× MIC of caspofungin profoundly reduced C. albicans viability (Fig. 4B). The presence of four antagonists improved cell survival in the presence of caspofungin by approximately 2-5-fold, although this difference was not statistically significant due to substantial variation in CFUs between experiments. Nonetheless, these data suggest that the antagonists protect a subpopulation of C. albicans cells from the cidal effects of the echinocandins and permit continued proliferation of the surviving cells at a much-reduced rate.

Fig 4.

Antagonists do not diminish cell damage but may enhance Candida albicans survival of echinocandin exposure. (A) A C. albicans wild-type strain expressing NanoLuc (NLUC) was subcultured at 5 × 10⁶ cells/mL into RPMI-pH 7 (2% glucose) supplemented with vehicle (0.5% DMSO), 5 μM aripiprazole (ARI), 5 μM cinacalcet (CINA), 5 μM haloperidol (HALO), 5 μM netupitant (NEPU), or 2.5 μM ponatinib (PONA) for 2 hours at 35°C, then caspofungin was added at 0×, 4×, and 16× MIC and further incubated for 4 hours at 35°C. The culture supernatant was collected, and NLUC release was quantified in each sample as relative luminescence units (RLU). (B) Cell viability for each sample described in panel A was quantified as colony-forming units (CFUs) after plating samples onto YPD agar. Representative data are shown from three independent experiments. See the main text for further details.

Echinocandin antagonistic drugs require Mkc1p

To provide insight into the molecular mechanisms underlying echinocandin antagonism, a collection of 91 deletion mutants lacking genes encoding kinases, or kinase-related products, was probed (29). Initially, the collection was screened to identify gene deletion strains with caspofungin sensitivity. The first screen identified five sensitive mutants with reduced growth compared to wild type after 48 hours in the presence of subinhibitory concentrations of caspofungin (0.25× MIC) (Fig. 5; Table 3). A further eight grew in the presence of supra-growth inhibitory concentrations (4× or 16× MIC), including mutants lacking Hog1p, as well as the Pbs2p MAPKK and Ssk2p MAPKKK, which regulate Hog1p activity (41) (Fig. 5; Table 4). A third screen identified 15 mutants as deficient in paradoxical growth, measured after 72 hours of incubation with 64× MIC of caspofungin (Fig. 5; Table 5). These varied phenotypes (sensitivity as well as the capacity to survive and replicate in the presence of various concentrations of caspofungin) indicate that kinase-mediated signaling is a critical determinant of C. albicans cell fate following echinocandin exposure, with both pro-survival and pro-cidal functions evident.

Fig 5.

Candida albicans deletion mutants with altered echinocandin sensitivity. Ninety-one C. albicans gene deletion strains were suspended in RPMI-pH 7 (2% glucose) supplemented with 0, 0.25×, 1.0×, 4.0×, 16.0×, or 64.0× MIC caspofungin. Plates were then incubated at 35°C, and growth was compared visually after 24, 48, and 72 hours. Those that had reduced or no growth compared to the wild-type control at 0.25× MIC were classified as sensitive. Mutants lacking or with reduced paradoxical growth versus the wild-type control were identified at 64× MIC after 72 hours. Finally, tolerant mutants were identified as those with increased growth versus the wild-type in the presence of 1×, 4×, and/or 16× MIC. In each case, the phenotype was observed in each of two independently conducted experiments.

TABLE 3.

Caspofungin-sensitive Candida albicans mutants

Mutant	Pathwaya
sln1Δ/Δ	Cell wall biosynthesis
prk1Δ/Δ	Putative protein serine/threonine kinase
cr_02210wΔ/Δ	Putative serine/threonine protein kinase, involved in the regulation of biofilm formation
stt4Δ/Δ	Phosphatidylinositol-4-kinase, forms a complex with Ypp1p and Efr3p that is required for phosphatidylinositol-4-phosphate in plasma membrane
cka1Δ/Δ	Putative alpha subunit (catalytic subunit) of protein kinase CK2

^a As described in Candida Genome Database.

TABLE 4.

Caspofungin-tolerant Candida albicans mutants

Mutant	Pathwaya,b
c1_05370cΔ/Δ	Putative serine/threonine kinase
hog1Δ/Δ	MAP kinase of osmotic-, heavy metal-, and core stress response; role in the regulation of response to stress
cka2Δ/Δ	Catalytic alpha-subunit of protein kinase CK2, interaction with the calcineurin pathway
pbs2Δ/Δ	MAPK kinase, role in osmotic and oxidative stress responses
ssk2Δ/Δ	MAP kinase kinase kinase, regulates Hog1 activation and signaling
snf1Δ/Δ	Functional homolog of Saccharomyces cerevisiae Snf1p, which regulates sugar metabolism
sok1Δ/Δ	Protein kinase required for degradation of Nrg1p
cr_06040w_aΔ/Δ	Ortholog of S. cerevisiae Sat4p, amphotericin B induced, clade-associated gene expression

^a As described in Candida Genome Database.

^b MAPK, mitogen-activated protein kinase.

TABLE 5.

Candida albicans mutants lacking paradoxical growth with caspofungin

Mutant	Pathwaya
c1_03450cΔ/Δ	Protein kinase-related protein, possibly involved in sterol transport between the plasma membrane and ER
ckb1Δ/Δ	Regulatory subunit of protein kinase CK2 (casein kinase II), beta subunit
chk1Δ/Δ	Histidine kinase; two-component signaling, cell wall synthesis
rim11Δ/Δ	Ortholog of S. cerevisiae Rim11p, a protein kinase involved in meiosis and sporulation in S. cerevisiae
ckb2Δ/Δ	Regulatory subunit of protein kinase CK2 (casein kinase II)
bck1Δ/Δ	Ortholog of S. cerevisiae Bck1p, MAP kinase kinase kinase of the cell integrity pathway
mkc1Δ/Δ	MAP kinase; role in biofilm formation, contact-induced invasive filamentation, systemic virulence in mouse, cell wall structure/maintenance
kin2Δ/Δ	Protein with similarity to S. cerevisiae Kin2p, transcription is positively regulated by Tbf1p
cr_02210wΔ/Δ	Putative serine/threonine protein kinase, involved in the regulation of biofilm formation
stt4Δ/Δ	Phosphatidylinositol-4-kinase, forms a complex with Ypp1p and Efr3p that is required for phosphatidylinositol-4-phosphate in the plasma membrane
cka1Δ/Δ	Putative alpha subunit (catalytic subunit) of protein kinase CK2

^a As described in Candida Genome Database.

Additional screens were performed with 4× MIC of caspofungin in the presence of each of the five antagonistic agents, non-responsive mutants identified (i.e., no antagonism) and follow-up dose responses conducted with caspofungin ± antagonist to confirm the phenotype. Between 8 and 13 mutants were confirmed to have reduced or no antagonistic responses for each antagonist (Fig. 6; Table 6). Substantial overlap was observed—with a total of 10 mutants identified as non-responsive to at least 4 of the five antagonists tested—suggesting shared underlying mechanisms. Most conspicuous were strains lacking BCK1, MKK2, and MKC1 that encode components of the same MAP kinase signaling module, which has an established role in cell wall maintenance, caspofungin responses, and induction of hyphal growth through activation of the Cph1p transcription factor (29, 42). The importance of Cph1p was confirmed by the failure of a cph1∆/∆ mutant to exhibit caspofungin antagonism with any of the five agents tested (Fig. 7). To determine if any of the antagonistic drugs affect activation of the Mkc1p pathway, extracts from SC5314 cells treated with each agent were probed with antiphospho-Mkc1p, but no major differences were apparent (data not shown). The presence of 0.25× MIC of caspofungin activated Mkc1p phosphorylation as expected (Fig. 8). Ponatinib substantially suppressed caspofungin-induced Mkc1p phosphorylation; however, the other four drugs had little effect. Thus, while the Mkc1p pathway is necessary for echinocandin antagonism with all five of the selected drugs, only one of the five agents had clear effects on its activity. The echinocandin antagonistic activity of four of the five drugs tested was also reduced in a cek1Δ/Δ mutant lacking an ERK-family protein kinase, which also has an established role in cell wall stress responses (43). However, the effect was less dramatic than observed for the Mkc1p-deficient mutant, being most obvious for ponatinib (Fig. 6 and 9). In addition, mutants lacking the CST20 and STE11 encoded kinases that function upstream of Cek1p were unaffected (Fig. S3), further indicating this pathway is less important.

Fig 6.

The Mkc1p MAP kinase pathway is required for echinocandin antagonism. Caspofungin susceptibility of wild-type (SN95) (A), bck1Δ/Δ (B), mkk2Δ/Δ (C), and mkc1Δ/Δ (D) C. albicans strains were compared in RPMI-pH 7 (2% glucose) supplemented with either 5 μM aripiprazole (ARI), 5 μM cinacalcet (CINA), 5 μM haloperidol (HALO), 5 μM netupitant (NEPU), 2.5 μM ponatinib (PONA), or vehicle (DMSO). Plates were incubated at 35°C and imaged after 24, 48, and 72 hours. Images are representative of assays performed in biological duplicate.

TABLE 6.

Candida albicans kinase mutants non-responsive to echinocandin antagonists

Mutant	Antagonists affected	Gene descriptiona
swe1Δ/Δ	ARI, CINA, HALO, NEPU, PONA	Putative protein kinase with a role in the control of growth and morphogenesis
ckb1Δ/Δ	ARI, CINA, HALO, NEPU, PONA	Regulatory subunit of protein kinase CK2 (casein kinase II), beta subunit
ckb2Δ/Δ	ARI, CINA, HALO, NEPU, PONA	Regulatory subunit of protein kinase CK2 (casein kinase II), beta′ subunit
mkc1Δ/Δ	ARI, CINA, HALO, NEPU, PONA	MAP kinase; role in biofilm formation, contact-induced invasive filamentation, systemic virulence in mouse, cell wall structure/maintenance
stt4Δ/Δ	ARI, CINA, HALO, NEPU, PONA	Phosphatidylinositol-4-kinase, forms a complex with Ypp1p and Efr3p that is required for phosphatidylinositol-4-phosphate in the plasma membrane
mkk2Δ/Δ	ARI, CINA, NEPU, PONA	Ortholog of S. cerevisiae Mkk2p, MAP kinase kinase involved in signal transduction
rad53Δ/Δ	ARI, CINA, HALO, PONA	Protein involved in the regulation of DNA damage-induced filamentous growth, putative component of cell cycle checkpoint
cek1Δ/Δ	ARI, CINA, HALO, PONA	ERK-family protein kinase; required for wild-type yeast-hypha switch, mating efficiency, virulence in mice; Cst20p-Hst7p-Cek1p-Cph1p MAPK pathway
bck1Δ/Δ	ARI, CINA, HALO, NEPU	Ortholog of S. cerevisiae Bck1p, MAP kinase kinase kinase of the cell integrity pathway
prk1Δ/Δ	ARI, CINA, HALO, NEPU	Putative protein serine/threonine kinase
kin2Δ/Δ	ARI, CINA, HALO	Protein with similarity to S. cerevisiae Kin2p
sak1Δ/Δ	ARI, CINA	Serine/threonine protein kinase, acts as an upstream activating factor for the SNF1 complex that regulates responses to nutrient stress
hsl1Δ/Δ	ARI, CINA	Probable protein kinase involved in the determination of morphology during the cell cycle of both yeast-form and hyphal cells via regulation of Swe1p and Cdc28p

^a As described in Candida Genome Database.

Fig 7.

The Cph1p transcription factor is required for echinocandin antagonism in Candida albicans. Caspofungin sensitivity of wild-type (SC5314) and (A) and cph1Δ/Δ (B) strains of C. albicans was compared in RPMI-pH 7 (2% glucose) supplemented with either 5 μM aripiprazole (ARI), 5 μM cinacalcet (CINA), 5 μM haloperidol (HALO), 5 μM netupitant (NEPU), 2.5 μM ponatinib (PONA), or vehicle (DMSO). Plates were incubated at 35°C and imaged after 24, 48, and 72 hours. Images are representative of assays performed in biological duplicate.

Fig 8.

Ponatinib suppresses echinocandin-induced Mkc1p phosphorylation. (A) C. albicans SC5314 was grown for 2 hours in RPMI medium supplemented with 0.5% DMSO (vehicle), 0.05 μM caspofungin (CAS), 2.5 μM ponatinib (PONA), or both drugs in combination. Cell extracts were prepared, and levels of phospho-Mk1cp were quantified through immunoblot analysis with antiphospho-p44/42 and chemiluminescent substrate. (B) Ponceau S-stained membrane was used as a loading control. (C) Levels of phospho-Mk1cp detected were quantified by volumetric analysis. Data are the mean ± standard deviation of four independently performed experiments. Statistical significance was calculated using one-way ANOVA with Dunnett’s post-test. ****P < 0.0001.

Fig 9.

The Ceklp kinase is not required for echinocandin antagonism. The caspofungin susceptibility of wild-type (SN95) (A) and cek1Δ/Δ (B) C. albicans mutants was compared in RPMI-pH 7 (2% glucose) supplemented with either 5 μM aripiprazole (ARI), 5 μM cinacalcet (CINA), 5 μM haloperidol (HALO), 5 μM netupitant (NEPU), 2.5 μM ponatinib (PONA), or vehicle (DMSO). Plates were incubated at 35°C and imaged after 24, 48, and 72 hours. Images are representative of assays performed in biological duplicate.

A core set of genes associated with xenobiotic response and hyphal growth is responsive to echinocandin antagonistic drugs

To provide additional insight into the mechanisms underlying the protective activity of the five selected antagonists, as well as their overall impact upon C. albicans physiology, we determined how each affects global patterns of gene transcription. SC5314 was grown at 35°C in RPMI medium supplemented with 5 µM of each antagonist (2.5 µM for ponatinib, which is growth inhibitory at 5 µM), or 0.5% DMSO (vehicle control); total RNA was extracted and subjected to high-throughput sequencing. Antagonist-responsive genes were defined as those in which the relative transcript read frequency was increased or decreased by ≥2-fold (P < 0.05) in the presence of drug versus vehicle control in each of two independent experiments. Between 590 and 42 gene transcripts were responsive to each individual drug with several altered by multiple drugs, indicating at least some similarities in their effects upon C. albicans physiology. We identified 17 transcripts as increased by at least four of the five antagonists, seven of which were associated with fluconazole resistance or regulation by the Tac1p transcription factor (Table 7; Table S1), including CDR2, which encodes an ATP-driven multi-drug efflux pump. An additional 10 of 43 transcripts elevated by any three of the five drugs were also associated with response or sensitivity to the azole antifungals or Tac1p-dependent regulation, including that of CDR1, encoding a second ATP-driven multi-drug efflux pump. Of the 22 transcripts repressed in the presence of at least four of the antagonists, nine are either required for C. albicans hyphal growth or exhibit hyphal-specific expression, including that of ECE1, which encodes candidalysin (44); ALS3 and HWP1, both hyphal cell wall-associated adhesins that mediate attachment to mammalian cells (45, 46); and UME6, a transcription factor that promotes sustained hyphal growth (47). A further 11 of 41 transcripts downregulated by any three antagonists were also associated in some way with hyphal growth, including CHT2 encoding a chitinase previously shown to be suppressed following echinocandin exposure (48). Many of these morphogenesis-related genes are known to be induced following activation of the Ras-cAMP-PkA-Efg1p-based signaling module, which plays a major role in the regulation of the yeast-hypha transition (49–51). Additionally, two transcripts previously identified as the most responsive to loss of cAMP production (52) were also affected, with ASR1 induced by aripiprazole, cinacalcet, and netupitant and suppressed by ponatinib and ASR2 induced by aripiprazole, cinacalcet, netupitant, and ponatinib. Curiously, the elevated transcription of the ASR1 and ASR2 genes is indicative of enhanced cAMP signaling, while suppression of the ECE1, HWP1, ALS3, and several other gene transcripts is consistent with diminished signaling through this pathway.

TABLE 7.

Candida albicans gene transcripts responsive to four or more echinocandin antagonists tested^ab

	Log2 of average fold change
Gene	Aripiprazole	Cinacalcet	Haloperidol	Netupitant	Ponatinib
RTA3	3.423	3.706	1.818	4.198	2.059
WH11	2.809	3.507	1.984	3.878	2.661
PGA26	−2.107	−2.912	−3.453	−2.580	−2.667
ALS3	−2.545	−4.114	−4.510	−4.839	−1.146
SOU2	3.336	3.436	3.420	2.476	NA
orf19.6502	2.249	2.483	1.961	2.382	NA
RSN1	1.684	1.663	2.670	1.465	NA
GCY1	1.436	1.419	1.719	1.389	NA
LTV1	−1.660	−1.542	−1.349	−1.237	NA
orf19.5710	−1.232	−1.617	−1.369	−1.737	NA
SEO1	−1.889	−2.155	−1.872	−2.312	NA
orf19.3475	−1.724	−1.847	−3.245	−1.588	NA
DEF1	−2.213	−2.071	−1.843	−2.361	NA
UME6	−2.695	−3.967	−2.154	−4.213	NA
IHD1	−3.511	−5.346	−2.929	−4.296	NA
ECE1	−3.211	−5.788	−2.523	−5.915	NA
HWP1	−3.118	−5.969	−2.933	−5.876	NA
CFL11	−4.302	−4.847	−4.012	−4.788	NA
CDR2	5.776	6.432	NA	6.186	4.105
orf19.4886	2.527	3.187	NA	2.959	2.313
OYE23	1.828	2.208	NA	2.751	2.596
ASR2	2.287	2.772	NA	3.218	1.074
orf19.4216	2.077	2.556	NA	2.254	1.857
orf19.86	2.034	2.093	NA	2.503	1.654
HSP12	1.881	2.355	NA	2.131	1.565
orf19.4907	1.927	1.987	NA	2.438	1.227
IFE1	1.498	1.855	NA	2.240	1.880
SOD5	−1.790	−2.456	NA	−2.170	−2.946
orf19.6200	−3.348	−3.614	NA	−4.799	−3.497
orf19.3406	NA	−1.193	−1.382	−1.206	−1.811
PHO89	NA	−2.685	−2.514	−2.941	−1.850
GIT1	NA	−2.433	−3.220	−2.046	−2.393

^a NA, non applicable (i.e., no significant change in transcript abundance compared to vehicle-treated cells).

^b Upregulated gene transcripts are shown in bold typeface.

Echinocandin antagonists suppress C. albicans hyphal growth

Given the impact of the echinocandin antagonists upon morphogenesis-responsive transcripts, the influence of each upon C. albicans hyphal growth was examined. All five echinocandin antagonists suppressed C. albicans hyphal growth on either 10% fetal bovine serum or M199 agar to varying extents (Fig. 10A and B), with haloperidol and ponatinib having the greatest effect. In liquid RPMI medium at 35°C, aripiprazole moderately suppressed hyphal formation, as we have previously reported (34), reducing the proportion of cells forming hyphae as well as the length of those that form. Haloperidol and ponatinib also significantly reduced hyphal length under these conditions (Fig. 10C), while netupitant and cinacalcet had no obvious effect. To further probe the connection between the cAMP-PkA-Efg1p morphogenesis-related signaling pathway and echinocandin antagonism, an efg1∆/∆ mutant was utilized. Dose-response experiments revealed that the efg1∆/∆ mutant had elevated levels of paradoxical growth compared to wild-type that was unaffected by any of the aforementioned five antagonists (Fig. 11). The mutant also exhibited substantial levels of tolerance to both anidulafungin and micafungin (data not shown). These data indicate that the morphogenesis-related transcription factor Efg1p confers susceptibility to caspofungin. Furthermore, the data are consistent with the five echinocandin antagonists acting through suppression of Efg1p activity.

Fig 10.

Several echinocandin antagonists suppress Candida albicans hyphal growth. (A and B) SC5314 was suspended at 2 × 10⁶ cells/mL, and 2 μL was spotted to 10% fetal bovine serum (A) or M199 (B) agar supplemented with either 5 μM aripiprazole, 5 μM haloperidol, 5 μM netupitant, 5 μM cinacalcet, 2.5 μM ponatinib, or vehicle (DMSO). Colonies were then imaged after 96 hours at 37°C. (C) SC5314 was suspended in liquid RPMI medium supplemented with the same drug concentrations as in A and B. After 6 hours of incubation, cells were fixed in 10% formalin, and cell morphology was observed microscopically. Images are representative of assays performed in biological triplicate.

Fig 11.

Deletion of EFG1 induces echinocandin tolerance in Candida albicans. The caspofungin sensitivity of wild-type (SC5314, (A), efg1Δ/Δ (B), and efg1Δ/Δcph1Δ/Δ (C) C. albicans strains was compared in RPMI-pH 7 (2% glucose) supplemented with either 5 μM aripiprazole (ARI), 5 μM cinacalcet (CINA), 5 μM haloperidol (HALO), 5 μM netupitant (NEPU), 2.5 μM ponatinib (PONA), or vehicle (DMSO). Plates were incubated at 35°C and imaged after 24, 48, and 72 hours. Images are representative of assays performed in biological duplicate.

DISCUSSION

The studies described herein have revealed a substantial number of drug-fungus interactions that directly or indirectly affect C. albicans susceptibility to the ordinarily cidal echinocandin antifungals. Several points are notable: the MIC determined at the 24 hour time point was not (in most cases) significantly altered by the antagonistic drug, but rather residual growth observed above the MIC at 48 and 72 hours was dramatically enhanced. Based on this phenotype, as well as viability counts following caspofungin exposure, we have classified the phenomenon as a form of tolerance that promotes survival of the fungal cells, rather than resistance. For four of the five antagonists examined in detail, this also appeared consistent with an enhanced form of the paradoxical effect previously described at high concentrations of the echinocandins. However, paradoxical growth is principally observed with caspofungin, less commonly with micafungin or anidulafungin (35, 53), and is more apparent in some C. albicans isolates over others. In contrast, the antagonistic activities we have described affect C. albicans response to all three echinocandins, and their activity does not seem restricted to specific isolates. Thus, it remains uncertain how the drug-induced echinocandin antagonism observed within this study is qualitatively related to the previously described paradoxical phenotype.

Our findings suggest that the antagonism observed depends upon a complex response requiring multiple signaling pathways that have been fairly well characterized with respect to their roles in regulating yeast-hypha morphogenesis and C. albicans response to cell wall stress (43, 54, 55). Specifically, the Mkc1p-mitogen-activated protein kinase (MAPK) signaling module, which acts upon the Cph1p transcription factor (56), is crucial for the echinocandin tolerance induced by the antagonists. CKB1 and CKB2, both encoding regulatory subunits of casein kinase 2, were also identified as non-responsive to the caspofungin antagonists. Signaling through the calcium-dependent calcineurin (CN) phosphatase pathway is known to be important for fungal survival upon exposure to both azole and echinocandin antifungals (57). Blocking CN signaling with either cyclosporin A or FK506 or using a C. albicans cnb1∆/∆ mutant, lacking the B regulatory subunit of CN, completely eliminates the antagonistic activity of all five drugs (Fig. S4 and S5). In contrast, four of the five antagonists examined (except haloperidol) promote caspofungin tolerance in a crz1∆/∆ mutant (Fig. S5), lacking one of the major CN-responsive transcription factors (58). Thus, Crz1p-independent CN signaling is also required for echinocandin antagonism. Unexpectedly, we observed that a C. albicans mutant lacking the Hog1p MAP kinase, also with established roles in cell wall stress responses, is highly caspofungin tolerant. However, aripiprazole, cinacalcet, netupitant, and ponatinib all further elevate the growth of this mutant in the presence of supra-MIC concentrations of caspofungin (Fig. S6), suggesting they act independently of Hog1p. Together, these results indicate that the echinocandin antagonists tested act through a complex mechanism that requires both Mkc1p, CK2, and Crz1p-independent functions of CN. The similarities in the pathway requirements of the five antagonists studied in detail suggest a largely shared molecular mechanism, which is surprising, given the diversity of their chemical structure and of their targets within the mammalian host. No clear C. albicans homologs of the known targets of aripiprazole, cinacalcet, haloperidol, or netupitant could be identified (Table 1), while the ABL1-encoded target of ponatinib shares only general homology to a large number of kinases encoded by the C. albicans genome. Thus, the capacity to sense these unrelated small molecules and produce a seemingly common functional response with respect to the tolerance phenotype is cryptic and suggests the presence of xenobiotic sensing modules and/or receptors capable of promiscuous interactions with small molecules. Indeed, analysis of transcriptional responses indicated that a significant number of echinocandin antagonists affect the transcription of Tac1p-responsive genes. The transcription factor Tac1p and close relatives from other species can directly bind several medically important molecules, most notably fluphenazine (59), to activate transcription of genes encoding the Cdr1p and Cdr2p efflux pumps. We also recently reported that a substantial proportion of drugs identified as fluconazole antagonists require Tac1p for activity (28), suggesting that this response module is able to sense a much broader array of molecular species than previously appreciated. However, in contrast to the fluconazole antagonists, most of which required Tac1p for activity, Tac1p is not required for the echinocandin antagonism described herein (Fig. S7), and its activation is therefore likely incidental. Given that the echinocandins are believed to engage β(1,3)-D-glucan synthase at the cell surface, it is unsurprising that Tac1p-mediated activation of CDR1 and CDR2 transcription, or of the MDR1-encoded efflux pump (Fig. S8), does not account for the observed antagonism. As such, Tac1p activation seems to define a more general response to encountering xenobiotics. Nonetheless, it is clear that C. albicans possesses systems that function to detect and respond to an impressive variety of molecular entities.

The suppression of hyphal growth and hyphal-associated gene transcription by the five antagonists studied raises questions of how cellular morphotype may affect C. albicans echinocandin sensitivity. Future work should explore if modulation of yeast-hypha morphogenesis can provide a strategy for C. albicans to mitigate the consequences of echinocandin exposure and determine if yeast and hyphal forms are similarly susceptible to echinocandin-induced cell death.

A major rationale for conducting the antagonism screen was to identify factors that can help explain the often-unpredictable response of individual patients to echinocandin therapy; i.e., could patients treated with one or more antagonistic medications be less responsive to antifungal treatment and/or more prone to recurrence? In a broader context, this study raises questions of how biologically active molecules, including the medications consumed by colonized or infected individuals, affect fungal physiology and thus potentially the equilibrium of the host-pathogen interaction. Drug-fungus interactions, for example, could alter the outcome of infection by: (i) causing profound physiological dysfunction that compromises fungal viability, fitness, or pathogenicity; (ii) changing the surface characteristics of the fungus in ways that alter its immunogenicity, binding/activation of complement, or sensitivity to host-derived antimicrobial peptides; or (iii) altering fungal sensitivity to antifungal drugs, as is the focus of this study. However, the highly variable, often transient, and combinatorial nature of medications usually provided to patients at risk of invasive fungal infections, specific clinical circumstances (including the nature and severity of immune dysfunction), site of infection, causative species, and potentially genetic factors of both the host and pathogen are likely to make it difficult, if not impossible, to discern the impact of any individual drug upon antifungal efficacy based on clinical observations alone. These factors are also likely to provide significant constraints on the power of clinical studies to reliably detect the impact of individual medications upon therapeutic efficacy, or the outcome of infection, especially given the relatively small numbers of infected patients that may be treated with a given antagonistic medication. Nonetheless, these limitations do not mean that such drug-fungus interactions have no impact, and it is possible they may help explain the often-poor correlation between the results of in vitro antifungal susceptibility tests and clinical efficacy. A more reliable strategy may be to use animal models of infection that enable control of many of the confounding variables.

Several key factors are likely to determine if the antagonistic agents identified within this study affect echinocandin efficacy within mammalian tissue. First, it is important to consider the effects of each drug at therapeutically relevant drug concentrations, which—in the context of disseminated infection—most investigators will take to be unbound blood serum concentrations from patients on a standard dosing regimen. Based on this parameter, only aripiprazole alters C. albicans echinocandin sensitivity within or at least close to clinically relevant concentration ranges (Table S2) (60–62). However, it is the available concentration of the antifungal as well as the antagonistic agent at the site of infection that is most relevant, and tissue concentrations can differ markedly from unbound serum concentrations. Additionally, acute inflammation, tissue dissolution, and protein digestion by fungal enzymes within foci of infection could have further effects, and drug concentrations at the site of fungal infection are unknown. Finally, Candida spp. naturally colonize the GI tract, where the concentration of orally administered drugs (or those excreted via the GI tract) is potentially much higher. Either way, the antagonists identified through this study can provide valuable chemical probes to better understand the molecular mechanisms that affect fungal survival following echinocandin exposure.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.00484-25.

Supplemental figures. aac.00484-25-s0001.pdf.

Fig. S1 to S4.

Supplemental material. aac.00484-25-s0002.pdf.

Fig. S5 to S8.

Supplemental tables. aac.00484-25-s0003.pdf.

Tables S1 to S3.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.