Fitness trade-offs restrict the evolution of resistance to amphotericin B

Abstract

The evolution of drug resistance in microbial pathogens provides a paradigm for investigating evolutionary dynamics with important consequences for human health. Candida albicans, the leading fungal pathogen of humans, rapidly evolves resistance to two major antifungal classes, the triazoles and echinocandins. In contrast, resistance to the third major antifungal used in the clinic, amphotericin B (AmB), remains extremely rare despite 50 years of use as monotherapy. We sought to understand this long-standing evolutionary puzzle. We used whole genome sequencing of rare AmB-resistant clinical isolates as well as laboratory-evolved strains to identify and investigate mutations that confer AmB resistance in vitro. Resistance to AmB came at a great cost. Mutations that conferred resistance simultaneously created diverse stresses that required high levels of the molecular chaperone Hsp90 for survival, even in the absence of AmB. This requirement stemmed from severe internal stresses caused by the mutations, which drastically diminished tolerance to external stresses from the host. AmB-resistant mutants were hypersensitive to oxidative stress, febrile temperatures, and killing by neutrophils and also had defects in filamentation and tissue invasion. These strains were avirulent in a mouse infection model. Thus, the costs of evolving resistance to AmB limit the emergence of this phenotype in the clinic. Our work provides a vivid example of the ways in which conflicting selective pressures shape evolutionary trajectories and illustrates another mechanism by which the Hsp90 buffer potentiates the emergence of new phenotypes. Developing antibiotics that deliberately create such evolutionary constraints might offer a strategy for limiting the rapid emergence of drug resistance.

Figure 1. Mechanisms of AmB resistance in Candida.

(A) Alignment of reads from whole-genome sequencing of C. albicans wild-type strain SC5314 and AmB-resistant clinical isolate ATCC 200955 demonstrates transposon insertion in ERG2 in the clinical isolate. The ERG2 locus is shown. Colored reads are indicative of mate-pairs that do not both map to the same chromosome, but instead one end to ERG2 and the other end to the TCA2 locus (elaborated in Figure S1A). Reads were visualized with the integrative genomics viewer (IGV) . (B) Alignment of selected strains from whole-genome sequencing of in vitro–evolved AmB-resistant series identifies causal mutations. Mutations in ERG6 ORF are highlighted, and the corresponding amino acid changes are indicated. Strain #1, 0 generations (founder); Strain #2, 60 generations; Strain #3, 120 generations; Strain #4, 240 generations. Two segments of IGV visualization for ERG6 were joined to allow visualization of both mutations in one image; point of joining indicated by “::”. (C) Spectrophotometric analysis of sterols reveals lack of C5–C6∶C7–C8 conjugation in AmB-resistant clinical isolates as well as laboratory-generated erg2 and erg3 erg11 mutants. Sterols were isolated by saponification and heptane extraction and analyzed spectrophotometrically between 240 and 300 nm, following established methods . (D) AmB susceptibility of clinical isolates and laboratory-generated mutants in every nonessential gene in the latter half of the ergosterol biosynthesis pathway (after cyclization of squalene to lanosterol). Mutants were generated in SN152 strain background using HIS1, LEU2, and ARG4 markers . AmB susceptibility was determined by microplate dilution in RPMI at 37°C for 24 h, repeated in duplicate; growth was normalized to wild-type in the absence of AmB.

Figure 2. AmB-resistant strains critically depend on high levels of Hsp90 function for survival.

(A) Mild inhibition of Hsp90 not only reverses AmB resistance but selectively kills AmB-resistant isolates. Spot assays (5-fold serial dilutions) of wild-type (SC5314), fluconazole-resistant (Isolate #2, [61]), and AmB-resistant (ATCC 200955) C. albicans on RPMI media containing antifungals and/or Hsp90 inhibitors at the indicated concentrations. (B) Increased dependence on Hsp90 for AmB resistance is conserved in a C. tropicalis clinical isolate. Spot assays of C. tropicalis reference strain MYA-3404 (AmB-sensitive) and ATCC 200956 (AmB-resistant) on media containing AmB, geldanamycin, or both compounds. (C) Isogenic, laboratory-generated AmB-resistant mutants also show an increased dependence on Hsp90. Spot assay of laboratory-generated mutants on AmB and geldanamycin. Fluconazole-resistant erg3 mutant and AmB-resistant clinical isolate are provided for reference. Full MIC data in liquid culture for all strains are provided in Figure S4. (D) In vitro selection for AmB resistance leads to hypersensitivity to Hsp90 inhibition. 24-hour MIC80 (drug concentration reducing growth by 80 percent) of geldanamycin, radicicol, and AmB for each isolate from in vitro evolution series, tested in YPD at 30°C. Note the discontinuity in the left y-axis due to the dramatic decrease in Hsp90 inhibitor MIC. (E) Hsp90 inhibition is cidal to AmB-resistant strains. Viability assays of wild-type and AmB-resistant C. albicans strains in the presence of the Hsp90 inhibitors geldanamycin or radicicol. Strains were incubated in liquid culture for 24 h with the drugs and then plated for surviving colony-forming units. Drugs were used at 5 µM in all strains. Error bars indicate SEM; each measurement was performed in duplicate.

Figure 3. Constitutive stress response activation in AmB-resistant strains.

(A) AmB resistant mutants constitutively express diverse stress response genes at high levels in the absence of any external stressors. qRT-PCR profiling of stress response genes in wild-type, fluconazole-resistant (erg3), and AmB-resistant mutants of C. albicans grown in rich (YPD) media with no added stressors (left panel), compared with wild-type strains treated with a variety of acute stressors known to activate these stress response pathways (right panel). Strains were grown to mid-log phase in YPD, and RNA was isolated by established methods. For stress treatment of wild-type strains (right panel), the indicated stressors are: AmB, 1 µg/mL AmB; NaCl, 0.3 M NaCl; Fe Chelator, 500 µM bapthophenanthroline disulfonate; DPTA-NO, 2 mM DPTA-NONOate; Ca²⁺, 150 mM CaCl₂; Peroxide, 10 mM tert-butyl peroxide. Expression levels of diverse stress response genes were quantified and normalized to four internal control genes: TDH3, TEF3, ACT1, and RPP2B; the mean value obtained from these four normalizations was used. To generate quantitative comparisons for color visualization, the relative expression level of each gene in each sample was divided by its maximal expression level observed (in any of the deletion strains or stress conditions; see Materials and Methods for further description). (B) Calcineurin and PKC pathways are differentially required for survival of AmB-resistant strains. Growth of AmB-resistant laboratory mutants and clinical isolates, as well as wild-type controls, in the presence of small molecule stress response inhibitors at the following concentrations: Geldanamycin and Radicicol, 2 µM; FK-506, 5 µg/mL; Cyclosporin A, 5 µg/mL; Enzastaurin and Cercosporamide, 5 µg/mL. Growth was assayed after 24 h; values indicate means of duplicate measurements, normalized to wild-type in DMSO. (C) Hog1, but not calcineurin, is required for wild-type levels of AmB tolerance in the absence of ergosterol biosynthesis mutations. Wild-type, cnb1, and hog1 strains, as well as wild-type with 1 or 5 µM geldanamycin, were tested for AmB MIC by microplate dilution. Growth was assayed after 48 h to highlight differences in drug tolerance between strains; values indicate means of duplicate measurements, normalized to the wild type in the absence of AmB. (D) Hog1, Cnb1, and high levels of Hsp90 are required for the de novo emergence of AmB-resistant colonies. ERG2/erg2 heterozygotes from wild-type, cnb1, or hog1 backgrounds were plated at a density of 8×10⁶ cells per plate on media containing 0.4 µg/mL AmB. The wild-type was also plated on media containing AmB and 2.5 µM geldanamycin. Plates were photographed after 2 d.

Figure 4. AmB-resistant strains are hypersensitive to the stresses of the host environment.

(A) AmB-resistant strains are sensitive to very high febrile temperatures and extremely sensitive to oxidative stress, especially at elevated temperature. Wild-type and AmB-resistant strains were spotted by serial dilution on RPMI media with or without tert-butyl peroxide at the concentrations indicated. (B–E) AmB-resistant strains are hypersensitive to stresses encountered in the host environment. Wild-type and AmB-resistant strains were grown in RPMI media containing 2 mM of hypochlorous acid (B) or 4 mM of the nitric oxide donor DPTA-NONOATE (C), two neutrophil-secreted products that are the critical final effectors of anti-Candida immunity. Sensitivity to iron deprivation was tested by growth in RPMI+500 µM of the iron chelator bathophenanthrolinedisulfonic acid (D). No increase in sensitivity to the antimicrobial peptide Calprotectin (10 µg/mL) (E) was observed. Growth values were obtained by normalization to wild-type growth in the same condition. All mutant strains were significantly more sensitive than wild-type to the tested concentrations of hypochlorous acid, DPTA-NONOate, and BPS at 37°C (**p<0.01, two-tailed Student's t test), but not to calprotectin. Values indicate the mean of two independent experiments of three replicates each; error bars indicate SEM. (F) AmB-resistant strains proliferate more slowly than wild type in serum at elevated temperature. Wild-type and AmB-resistant strains were inoculated in 100% fetal bovine serum at 39.5°C, and viable colony-forming units were determined by plating dilutions on YPD after 24 and 48 h. Error bars indicate SEM. (G) AmB-resistant strains are hypersensitive to killing by neutrophils. Human neutrophils were isolated from whole blood (see Materials and Methods) and activated by treatment with recombinant TNF-α. Candida were added to wells containing neutrophils at a 1∶1 effector∶target ratio and incubated at 37°C for 6 h, at which point neutrophils were lysed and Candida growth was quantitated with Alamar blue. Percent growth for each strain was calculated as the fraction of growth in the presence of neutrophils to growth in the absence of neutrophils; mutants were normalized to growth of the WT on each plate. The experiment was performed with a total of three biological replicates from two separate days; error bars indicate mean and SEM.

Figure 5. AmB-resistant strains are defective in filamentation and tissue invasion.

(A) AmB-resistant mutants fail to properly induce filamentous growth upon stimulation. Wild-type strains and resistant mutants were grown in YPD and then inoculated into RPMI media containing 10% fetal bovine serum at 37°C. Strains were analyzed by DIC microscopy after 2 and 4 h. (B) AmB-resistant strains cause much less damage to endothelial monolayers than wild type. Monolayers of HUVECs (human umbilical vein endothelial cells) were established and infected with C. albicans. After 6 h, endothelial cell cytotoxicity was assayed by quantifying LDH release, using uninfected cells as a negative control and cells lysed with 1% Triton X-100 as 100% lysis control. Data were pooled from two independent experiments with six replicate wells each; error bars indicate SEM. All mutant strains were significantly less cytotoxic than wild type (*p<0.05, **p<0.01, two-tailed paired Student's t test).

Figure 6. AmB-resistant strains are avirulent in a mammalian model of disseminated candidiasis.

(A) AmB-resistant mutants do not cause morbidity even when injected into mice at a high inoculum. Wild-type, erg3, and AmB-resistant mutants were grown into log phase for 5 h in YPD, counted, and 4×10⁶ cells of each strain were injected into the tail vein of 7–9-wk-old Balb/c mice (n = 8–14 mice per strain). Mice were monitored for weight loss and sacrificed after a >20% drop in body weight or appearance of morbidity. All surviving mice were sacrificed after 12 d. (B) AmB-resistant strains are unable to colonize the mouse kidney. Colonies were counted after homogenization and plating in duplicate of viable Candida from one kidney of each infected mouse at the time of sacrifice. As wild-type, erg3, and other mutants survived for different periods of time before sacrifice, direct comparisons cannot be made between these groups based on CFU values. However, the extremely low CFU values from AmB-resistant strains are typically indicative of sterilization. (C) AmB-resistant strains do not damage the mouse kidney. Kidneys from each infected mouse were fixed and stained with periodic acid-Schiff stain to visualize Candida and kidney pathology. Wild-type and erg3 strains demonstrated filamentous growth and severe kidney pathology, which was even greater in the erg3 strain. Examples of sites where fungi are observed are highlighted with a black arrow; fungi appear as long filaments with a purple color. No viable Candida were seen in sections from mice infected with AmB-resistant strains. Scale bar, 50 µm.